Browse the corpus

Introduction Cerebral ischemia accounts for 80 % of all human strokes and has a major impact on the public health [1]. It causes primary neuronal death in the ischemic region and leads to delayed neuronal degeneration in the penumbra [2]. Stroke disables the patients more than it kills. This fact has led a recent effort to develop strategies for neural repair after stroke. Several small animal models of stroke have been developed to identify mechanisms of cerebral ischemia for developing novel recanalyzing, neuroprotective, neuroregenerative, or anti-inflammatory drugs at a preclinical level [3, 4]. Among them, middle cerebral artery occlusion (MCAo) by an intraluminal filament technique is the most widely used method [5]. Another approach that is technically simpler involves induction of cerebral ischemia and infarction in the cortical vasculature of rats by photochemical reaction triggered by systemic administration of rose bengal (disodium tetraiodo-tetrachloro-fluorescein) and focal illumination of the brain [6]. Illumination leads to production of singlet oxygen via dye triplet energy transfer, which in turn induces peroxidative damage to the endothelium and a vasoconstriction. Consequently, platelet aggregation is produced, with the development of a thrombus and vascular occlusion, and a distal territory ischemia is formed [6–8]. Rats with small infarct volume induced by this method have low mortality, and even performed on aged rats, the mortality of rats with photothrombotic stroke remained much lower than those with global, permanent stroke [9, 10]. An animal model of cerebral ischemia, providing both reproducibility and precise control of lesion size, is critical for translational studies. MCAo produces large ischemic lesions of varying size, which may be considered as a burden to research conducting. Further, MACo lacks the participation of platelet aggregation which is primary initiator of clinical ischemic events [11].

cibility and precise control of lesion size, is critical for translational studies. MCAo produces large ischemic lesions of varying size, which may be considered as a burden to research conducting. Further, MACo lacks the participation of platelet aggregation which is primary initiator of clinical ischemic events [11]. The photothrombotic model not only overcomes the drawbacks of MCAo but can also be used to characterize inflammatory response and apoptosis following thrombosis [12–14] and to monitor structural and functional plasticity of neurons [15].

cibility and precise control of lesion size, is critical for translational studies. MCAo produces large ischemic lesions of varying size, which may be considered as a burden to research conducting. Further, MACo lacks the participation of platelet aggregation which is primary initiator of clinical ischemic events [11]. The photothrombotic model not only overcomes the drawbacks of MCAo but can also be used to characterize inflammatory response and apoptosis following thrombosis [12–14] and to monitor structural and functional plasticity of neurons [15]. Current managements of acute stroke include restoring cerebral blood flow (CBF) to ischemic penumbral area by thrombolytic therapy, interventional procedures, surgery, and/or cell-based therapy to enhance tissue repair and functional recovery after ischemic stroke [16–18]. Noninvasive imaging modalities, such as magnetic resonance imaging (MRI) and positron emission tomography (PET), are promising for investigating the evolution of stroke, confirming the validity of models, monitoring the effect of revascularization interventions, and evaluating the efficacy of novel therapeutic drugs [19–23]. MRI has been used in several preclinical studies of stroke using MCAo model [8, 20, 21]. When ischemic brain injury occurs, glucose metabolism changes in the infarct and peri-infarct regions. This was demonstrated in earlier reports using 14C-2-deoxy-D-glucose (14C-2-DG) autoradiography for the study of photochemically induced ischemic stroke [24] and MCAo model [25]. However, method of 14C-2-DG with autoradiography is limited for in vivo longitudinal evaluation of glucose metabolism on the same animal. Until recently, 18F-2-deoxy-glucose (FDG) PET imaging was used to evaluate glucose metabolism in transient and permanent MCAo models of rats [26, 27]. The powerful imaging tools of 18F-FDG/PET and MRI offer the information of metabolic changes, lesion structure, edematous and CBF status, respectively. In order to characterize the temporal evolution of stroke lesion induced by photothrombotic method in 14 days, 18F-FDG/PET, T2-weighted image (T2WI), perfusion-weighted image (PWI), and diffusion-weighted image (DWI) were performed to measure cerebral glucose metabolism, edematous lesion, tissue perfusion, and water diffusion, respectively.

ze the temporal evolution of stroke lesion induced by photothrombotic method in 14 days, 18F-FDG/PET, T2-weighted image (T2WI), perfusion-weighted image (PWI), and diffusion-weighted image (DWI) were performed to measure cerebral glucose metabolism, edematous lesion, tissue perfusion, and water diffusion, respectively. This image information was further correlated with the cellular and molecular analysis including tissue viability, morphological changes, inflammatory response, astrocyte scar formation, neovascularization, and blood-brain barrier (BBB) permeability, performed by 2, 3, 5-triphenyl tetrazolium chloride (TTC), hematoxylin-eosin (H&E) staining, immunohistochemistry (IHC) staining, and ex vivo Evans blue (EB) imaging in the photochemically induced stroke model of rat.

ry response, astrocyte scar formation, neovascularization, and blood-brain barrier (BBB) permeability, performed by 2, 3, 5-triphenyl tetrazolium chloride (TTC), hematoxylin-eosin (H&E) staining, immunohistochemistry (IHC) staining, and ex vivo Evans blue (EB) imaging in the photochemically induced stroke model of rat. Materials and Methods Animals and Study Design Seven-week-old male Sprague-Dawley rats were kept under standardized condition (12–12-h light-dark cycle, with free access to food and water). The animals received serial MRI and 18F-FDG PET/computed tomography (CT) imaging at day 1, 3, 7, and 14 after photothrombotic stroke induction. For TTC, H&E, IHC, and EB staining, rats were sacrificed at day 1, 3, 7, and 14 after stroke induction, and intact brain was carefully removed for subsequent manipulation. All animal experiments in this study were conducted according to the guidelines set by the National Laboratory Animal Center and approved by the Institutional Animal Care and Use Committee of National Yang-Ming University and Macau University of Science and Technology. Reporting of this work complies with Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines.

cording to the guidelines set by the National Laboratory Animal Center and approved by the Institutional Animal Care and Use Committee of National Yang-Ming University and Macau University of Science and Technology. Reporting of this work complies with Animal Research: Reporting in Vivo Experiments (ARRIVE) guidelines. Photochemically Induced Stroke Rats were temporarily anesthetized with isoflurane (induction 3.0 % in air) followed by intramuscular injection of Dexdomitor and Zoletil 50 mixture (1:1) (Orion, Finland and Virbac, France, respectively) with a dose of 100 μl per 300 g body weight and placed in a stereotaxic frame (Narishige Instruments, Tokyo, Japan). After a small incision was made on the scalp, a craniotomic window (3 mm × 6 mm) was made over the somatosensory cortex with the center at the coordinate of 1 mm rostrally from the bregma and 3.5 mm lateral to the midline. A laser beam of 1.5 mm diameter and 532 nm wavelength (GPD105-M-12, Onset Electro-Optics, Taiwan) was stereotactically positioned at the middle of the craniotomic window and illuminated for 20 min. During the first 2 min of illumination, rose bengal (2 ml/kg body weight, concentration: 10 mg/ml saline) was slowly injected through the tail vein. Two control groups were performed following full procedure except for laser illumination or rose bengal injection. All rats after stroke induction were able to survive until they were sacrificed at the end point in this study.

ml/kg body weight, concentration: 10 mg/ml saline) was slowly injected through the tail vein. Two control groups were performed following full procedure except for laser illumination or rose bengal injection. All rats after stroke induction were able to survive until they were sacrificed at the end point in this study. Regional Cerebral Vasculature Examination with Laser Speckle Contrast Imaging Laser speckle contrast imaging (LSCI) is a technique in which coherent light incidence on a surface produces a reflected speckle pattern that is related to the underlying movement of optical scatters, such as red blood cells, indicating blood flow [28]. Before and after thrombosis was induced by photochemical method, regional cerebral vasculature was examined by LSCI. Briefly, a moorFLPI-2 Full-Field Laser Perfusion Imager (Moor Instruments, Axminster, UK) was placed at the center of cranial window on the somatosensory cortex where the laser beam illuminated. Before and after laser illumination, the laser speckle imaging was acquired with 25-Hz sampling frequency, 1 frame/s, 580 × 752 pixels resolution, and zoom size of 5.6 mm × 7.5 mm.

(Moor Instruments, Axminster, UK) was placed at the center of cranial window on the somatosensory cortex where the laser beam illuminated. Before and after laser illumination, the laser speckle imaging was acquired with 25-Hz sampling frequency, 1 frame/s, 580 × 752 pixels resolution, and zoom size of 5.6 mm × 7.5 mm. TTC Assay and H&E Staining At day 1, 3, 7, and 14 after initiation of photothrombotic stroke, selected rats (n = 3 at each day) were sacrificed by overdose injection of pentobarbital for tissue viability and histopathological examinations. Following sacrifice, fresh brain was removed from the skull, washed in iced phosphate buffer saline (PBS), and placed in a brain mold. Coronal sections of 1 mm in thickness were cut through the cerebrum and placed in 2 % TTC (Sigma) for 15 min in a 37 °C incubator for the macroscopic determination of tissue viability. All sections were photographed for delineation of infarct as revealed with TTC. The brain sections were then fixed with 10 % formalin and processed with H&E staining for microscopic examination.

cerebrum and placed in 2 % TTC (Sigma) for 15 min in a 37 °C incubator for the macroscopic determination of tissue viability. All sections were photographed for delineation of infarct as revealed with TTC. The brain sections were then fixed with 10 % formalin and processed with H&E staining for microscopic examination. Immunohistochemical Staining Paraffin-embedded rat brain tissues with thrombotic stroke at day 1, 3, 7, and 14 (n = 3 at each time point) were cut into 5-μm section slices. After 1-h blocking (10 % normal serum, 1 % bovine serum albumin [BSA], and 0.025 % TritonX-100 in tris-HCL buffered solution [TBS]), sections were incubated overnight at 4 °C with the following primary antibodies (pre-diluted in TBS containing 1 % BSA): anti-alpha smooth muscle actin (αSMA, 1:100, ab7817, Abcam), anti-von Willebrand Factor (vWF, 1:100, ab6994, Abcam), anti-glial fibrillary acidic protein (GFAP, 1:200, ab53554, Abcam), anti-CD68 (1:100, mab6564, Abnova), and anti-NeuN (1:200 #52283, Arigobio). After being rinsed with TBS-0.1 % Tween-20, tissue sections were detected using horseradish peroxidase-conjugated secondary antibodies and the DAKO Dual Link system (DAKO, K4065) with 2 % 3,3-diaminobenzidine. The images were photographed by the Aperio Image Scope 12.3 (Leica).

, and anti-NeuN (1:200 #52283, Arigobio). After being rinsed with TBS-0.1 % Tween-20, tissue sections were detected using horseradish peroxidase-conjugated secondary antibodies and the DAKO Dual Link system (DAKO, K4065) with 2 % 3,3-diaminobenzidine. The images were photographed by the Aperio Image Scope 12.3 (Leica). T2WI, DWI, and PWI Examinations by Magnetic Resonance Imaging BioSpec-70/30 7T system (Bruker, Ettlingen, Germany) using a birdcage head-coil of 75 mm inner diameter for radio frequency (RF) transmission and a 20 mm diameter surface coil for reception was used for MRI experiment. The same rat at day 1, 3, 7, and 14 after thrombotic stroke was anesthetized with initial inhalation of 4 % isoflurane for 3 min and maintained with 2 % isoflurane in a mixture of 20 % oxygen and 80 % room air. Prior to imaging, rats under anesthesia were placed in the stereotaxic holder of MRI machine equipped with a heating system to maintain body temperature and a pressure detector to monitor respiration. MRI data sets consisting TWI, DWI, and PWI were acquired at corresponding time points. T2WIs were acquired with a multislice multiecho Carr, Purcell, Meiboom, Gill (CPMG) sequence: repetition time (TR) = 2500 msec, echo time (TE) = 33 msec, 16 echoes, field of view (FOV) = 25 × 25 mm2, slice thickness = 1 mm, matrix = 256 × 256. Eight coronal and eight horizontal slices were acquired covering a volume extending 10 mm in the rostrocaudal direction. Coronal slices were centered around the infarction lesion, whereas horizontal slices were aligned with the skullcap. DWIs were recorded with a multislice Stejskal-Tanner spin-echo sequence: TR = 4000 msec, echo time = 22 msec, field of view = 25 × 25 mm2, slice thickness = 1 mm, matrix = 128 × 128. Data were recorded in the same coronal and horizontal slices as chosen for T2WI. Two sets of images were acquired with two different diffusion-encoding gradient strengths (b = 30, 1000 s/mm2) in the rostrocaudal direction. PWIs were performed with pulsed arterial spin labeling (PASL) technique using a flow-sensitive alternating inversion-recovery echo planar imaging (FAIR-EPI) sequence with matrix = 96 × 96, FOV = 25 × 25 mm2, inversion recovery time (TIR) = 30 to 2300 msec, number of TIR values = 22, recovery time = 10,000 msec, TE/TR > = 10/18,000 msec. The T1 changes between slice selective inversion sequence and nonselective inversion sequence were used for CBF quantification [29].

-EPI) sequence with matrix = 96 × 96, FOV = 25 × 25 mm2, inversion recovery time (TIR) = 30 to 2300 msec, number of TIR values = 22, recovery time = 10,000 msec, TE/TR > = 10/18,000 msec. The T1 changes between slice selective inversion sequence and nonselective inversion sequence were used for CBF quantification [29]. The infarct volume and CBF ratio were measured by using PMOD software (version 3.0; PMOD Technologies). The CBF difference indexes (CBFDI) of the infarct lesion, peri-infarct region, ipsilateral remote cortex, and hippocampus were calculated as follows: CBFDI = (CBF of the mirror site of contralateral hemisphere–CBF of the selected region in the affected hemisphere)/CBF of the mirror site of contralateral hemisphere. Evaluation of Cerebral Glucose Metabolism by MicroPET/CT Imaging

The infarct volume and CBF ratio were measured by using PMOD software (version 3.0; PMOD Technologies). The CBF difference indexes (CBFDI) of the infarct lesion, peri-infarct region, ipsilateral remote cortex, and hippocampus were calculated as follows: CBFDI = (CBF of the mirror site of contralateral hemisphere–CBF of the selected region in the affected hemisphere)/CBF of the mirror site of contralateral hemisphere. Evaluation of Cerebral Glucose Metabolism by MicroPET/CT Imaging 18F-FDG-PET/CT was performed to assess the evolutional change of cerebral glucose utilization. After a 12-h fasting, thrombotic stroke rats (n = 5) at 1, 3, 7, and 14 days were anesthetized with isoflurane and intravenously administered approximately 37 MBq (~1 mCi) of 18F-FDG. Sixty minutes later, microPET/CT images were acquired for 30 min using a FLEX X-PET and X-O small animal imaging system (GE Healthcare) with the spatial resolution of 1.6 mm and the voxel size of 0.4 mm × 0.4 mm × 1.2 mm. CT images were acquired with 256 projections over 2 min for attenuation correction and anatomy landmarks. PET data were reconstructed using 3D ordered subset expectation maximization (OSEM) method. CT images were reconstructed using a cone-beam reconstruction algorithm. PET and CT images were co-registered using commercial software (Visage Imaging) with 72-μm isotropic CT spatial resolution and 2 mm for PET imaging. For quantitative analysis, volume of interests (VOIs) were drawn on the infarct lesion, one-voxel width surrounding the lesion, and remote region on the ipsilateral cortex. VOIs were also selected at the mirror sites of the contralateral hemispheres. Percent-injected dose of 18F-FDG per c.c. of brain tissue (%ID/cm3) was obtained from each VOI. The metabolic difference index (MDI) of each VOI was calculated with the following formula: MDI = (%ID/c.c. of contralateral side VOI − %ID/c.c. of lesion VOI)/%ID/c.c. of contralateral side VOI.

e contralateral hemispheres. Percent-injected dose of 18F-FDG per c.c. of brain tissue (%ID/cm3) was obtained from each VOI. The metabolic difference index (MDI) of each VOI was calculated with the following formula: MDI = (%ID/c.c. of contralateral side VOI − %ID/c.c. of lesion VOI)/%ID/c.c. of contralateral side VOI. Assessment of Blood-Brain Barrier by EB Dye Staining Disruption of BBB was evaluated at day 1, 3, 7, and 14 after photothrombotic stroke induction. Briefly, 2 h after EB dye (Sigma, 10 mg/ml in saline, 2.5 ml/kg rat weight) was injected via tail vein, animals were sacrificed with an overdose of pentobarbital injection (RMB, Animal Health Ltd., UK). The brains were then removed from the skulls immediately after death. To detect the presence of EB, the intact brain was imaged using in vivo fluorescence and bioluminescence imaging system (IVIS 50, Perkin Elma, UK) with following steps and image acquisition settings: the brains were placed at the center of imaging field, and images were acquired for 2 s using Cy5.5 band pass filter channel (excitation/emission wavelength: 615~665 nm/695~770 nm). Region of interest (ROI) selection and quantification were performed using living image software 3.2 (IVIS Imaging System, Perkin Elma, UK).

e brains were placed at the center of imaging field, and images were acquired for 2 s using Cy5.5 band pass filter channel (excitation/emission wavelength: 615~665 nm/695~770 nm). Region of interest (ROI) selection and quantification were performed using living image software 3.2 (IVIS Imaging System, Perkin Elma, UK). Statistical Analysis The numerical data were reported as means ± standard deviation (SD). Statistical analysis was carried out with the SPSS for windows software package (release 13.0, SPSS Inc., Chicago IL). One-way ANOVA and Tukey post hoc test were used to compare the MDIs and fluorescent signal intensity of EB staining between different time points after stroke. A significant difference was considered if the p value was less than 0.05.

out with the SPSS for windows software package (release 13.0, SPSS Inc., Chicago IL). One-way ANOVA and Tukey post hoc test were used to compare the MDIs and fluorescent signal intensity of EB staining between different time points after stroke. A significant difference was considered if the p value was less than 0.05. Results Cerebral Vascular Occlusion by Photothrombotic Induction Photochemical induction of embolic stroke was completed by 532-nm LASER illumination at somatosensory cortex through the cranial window upon rose bengal infusion. Vasoconstriction and blockade of blood flow were examined by LSCI before and after stroke induction. As shown in Fig. 1, the surface vessel network was clearly appeared before induction in each group. Thirty minutes after stroke induction, remarkably vanishing of blood vessel network was observed in rats injected with rose bengal through the cranial window where the LASER beam illuminated at, while hyperemic blood vessel network was presented in the group of only illuminated by laser beam and there was no change of blood flow in the group of sole injection of rose bengal. The rapid and massive coagulation of vessels within the illumination area features the pathomechanism of this method which is not similar to that in clinical ischemic stroke usually caused by an embolus or two.Fig. 1 Laser speckle contrast imaging (LSCI) before and after photochemical induction of stroke. Cerebral vasculature occlusion was presented after photothrombotic stroke. Either illumination of laser beam or sole injection of rose bengal did not block cerebral blood flow as shown on LSCI

ly caused by an embolus or two.Fig. 1 Laser speckle contrast imaging (LSCI) before and after photochemical induction of stroke. Cerebral vasculature occlusion was presented after photothrombotic stroke. Either illumination of laser beam or sole injection of rose bengal did not block cerebral blood flow as shown on LSCI Tissue Viability and Morphological Characteristics of Infarct Lesion Tissue viability and morphologic change of the photochemically induced stroke were evaluated by TTC and H&E stains at day 1, 3, 7, and 14 post-induction. Results of TTC and H&E staining at each time point were shown in Fig. 2. TTC-negative region which represents the necrotic and non-viable tissue was well demonstrated on the first day after stroke induction, and gradually decreased in size afterward. At day 3, in addition to the shrinkage of ischemic lesion, the boundary between TTC-positive and TTC-negative regions became blurred, as compared to a sharp, clear dividing line at day 1, indicating an increasing number of cells which were migrating and infiltrating to the ischemic margin from outside the region. The volume of the infarct lesions at day 3 to day 14 reduced gradually and remarkably compared to that of the initial infarct volume at day 1. TTC-negative region became hardly observed at day 14 (Fig. 2a). Microscopic examination of the H&E-stained tissues showed a typical stroke-induced liquefactive necrosis at day 1, dilated vessels at day 3, infiltration of cells into the surrounding area of ischemia at days 3 and 7, and neovascular formation at day 14 (Fig. 2b, c).Fig. 2 Longitudinal TTC and H&E staining at day 1, 3, 7, and 14 after induction of photochemical thrombosis. TTC-negative area, observed at the site where laser illuminated, was largest in size at day 1, then decreased with time, and was almost not detectable at day 14 (a). b–c Morphological change in H&E-stained tissue sections. It showed a typical stroke-induced liquefactive necrosis at day 1, dilated vessels at day 3, infiltration of cells into the surrounding area of ischemia at days 3 and 7, and neovascular formation at day 14

d with time, and was almost not detectable at day 14 (a). b–c Morphological change in H&E-stained tissue sections. It showed a typical stroke-induced liquefactive necrosis at day 1, dilated vessels at day 3, infiltration of cells into the surrounding area of ischemia at days 3 and 7, and neovascular formation at day 14 The Change of Inflammatory Cells and Astrocytes in Infarct Region As a marker of macrophage and microglia, CD68 expression represents the evolution of inflammatory response post-stroke [30]. Figure 3a shows the presence of CD68-positive cells in the border zone of infarction lesion at day 7, and the cell number increased in the lesion at day 14. IHC staining using antibody against GFAP showed the infiltration of astrocytes into infarct boundary at day 7, and interdigitating compact astrocyte scar surrounding the lesion core was observed at day 14 (Fig. 3b). In addition, NeuN-expressing mature neuronal cells were absent in stroke region throughout 14 days after induction (Fig. 3c). These results indicated that the cellular evolutional changes in the lesion were mainly the inflammatory response and astrocyte scar formation, instead of neurogenesis.Fig. 3 Immunohistochemical staining of GFAP, CD68, and NeuN at day 1, 3, 7, and 14 after stroke induction. At day 7 post-stroke, the result of staining indicated that GFAP- and CD68-positive cells accumulated at the peri-infarct zone and formed a belt surrounding the infarct lesion at day 14. No NeuN expression was observed in ischemic zone throughout 14 days. Scale bar 100 μm

at day 1, 3, 7, and 14 after stroke induction. At day 7 post-stroke, the result of staining indicated that GFAP- and CD68-positive cells accumulated at the peri-infarct zone and formed a belt surrounding the infarct lesion at day 14. No NeuN expression was observed in ischemic zone throughout 14 days. Scale bar 100 μm Angiogenesis and Vasculogenesis in Infarct Lesion As shown in Fig. 4, there were scattered vWF (a marker of endothelial cell) positive vessels in infarct boundary at day 3, and the number of capillaries and αSMA (a marker of vascular smooth muscle cells) positive vessels significantly increased along the infarct margin at day 7. Further, neovascularization was detected in the core of infarct lesion at day 14. These results suggested that angiogenesis and vasculogenesis occurred in the regions of infarct and peri-infarct lesion within 2 weeks after photochemically induced stroke.Fig. 4 Immunohistochemical staining of vWF and αSMA at day 1, 3, 7, and 14 after stroke induction. Scattered vWF–vessels were observed in the margin of infarct zone at day 3, and the number of vWF- and αSMA-positive vasculature began to increase at day 7. At day 14, neovascularization notably occurred in the core of infarct zone

unohistochemical staining of vWF and αSMA at day 1, 3, 7, and 14 after stroke induction. Scattered vWF–vessels were observed in the margin of infarct zone at day 3, and the number of vWF- and αSMA-positive vasculature began to increase at day 7. At day 14, neovascularization notably occurred in the core of infarct zone The Dynamic Evaluation of Brain Edema and CBF To monitor the evolutional brain edema and hemodynamic change of photothrombotic stroke, T2WI, DWI, and PWI were performed on rats at day 1, 3, 7, and 14 after stroke induction. The ischemic lesion showed a hyperintensity in T2WI with sharp margin at day 1, and the signals gradually declined afterward (Fig. 5a). Quantitative analysis indicated that the lesion volume declined from 11.4 mm3 at day 1 to 0.7 mm3 at day 14 (Fig. 5b). In DWI, restriction of water diffusion was observed at the margin of stroke shown as an intense halo at day 1 and day 3, but not in the core of the lesion which showed a hypointense signal instead. At day 7, the halo was getting smaller. The hypointense lesion almost completely disappeared at day 14 (Fig. 5a). As T2WI and DWI have been the reliable methods for investigation of vasogenic edema and cytotoxic edema, respectively [20, 21, 23], gradual reduction of hypointense signal of T2WI and DWI suggested the improvement of brain edema, which contributed to decrease of infarct volume. PWI at day 1 after photochemical induction demonstrated a region of compromised CBF on arterial spin labeling (ASL) corresponding to the photothrombotic infarct lesion (6 ± 51 ml/100 g/min vs 200 ± 95 ml/100 g/min of contralateral cortex), and slightly decreased CBF in the remote cortex of ipsilateral hemisphere. The CBF compromised zone was getting smaller with time. At day 14, a tiny infarct core was still noted, suggesting a persisted damage of brain tissue. CBF of the ischemic region recovered to 195 ± 55 ml/100 g/min (Fig. 5c). CBFDI of the infarct region (zone A) and the peri-infarct region (zone B) declined rapidly at day 7. CBFDIs of the infarct region and the peri-infarct region showed no significant difference with the contralateral cortex at day 7 and day 14, respectively. CBFDI of ipsilateral remote cortex (zone C) and hippocampus (zone D) closed to zero means CBF almost identical to that of the contralateral hemisphere throughout the whole study of 14 days (Fig. 5d).

he peri-infarct region showed no significant difference with the contralateral cortex at day 7 and day 14, respectively. CBFDI of ipsilateral remote cortex (zone C) and hippocampus (zone D) closed to zero means CBF almost identical to that of the contralateral hemisphere throughout the whole study of 14 days (Fig. 5d). In summary, the MRI results revealed progressive shrinkage of stroke volume at the periphery of the infarct lesion early after stroke and almost complete recovery of CBF within 2 weeks after stroke induction. Moreover, the remarkable improvement of CBF in infarct lesion and peri-infarct region at day 7 might be related to the synchronous enhancement of angiogenesis and vasculogenesis.Fig. 5 Sequential images of T2WI, DWI, and PWI in stroke lesion on the right somatosensory cortex and CBF analysis. a Images of T2WI, DWI, and PWI at day 1, 3, 7, and 14 after ischemia induction demonstrated the evolution of vasogenic edema and cytotoxic edema with time. b Evolutional changes and CBFDI, which was calculated from the regional CBF obtained from the PWI of the infarct lesion (zone A), peri-lesional region (zone B, one-pixel width surrounding the infarct lesion), ipsilateral remote cortex (zone C), and hippocampus (zone D). It suggests that CBF of zone B improved more significantly than that of zone A from day 7 after induction

om the regional CBF obtained from the PWI of the infarct lesion (zone A), peri-lesional region (zone B, one-pixel width surrounding the infarct lesion), ipsilateral remote cortex (zone C), and hippocampus (zone D). It suggests that CBF of zone B improved more significantly than that of zone A from day 7 after induction Glucose Metabolism of Stroke Lesion by 18F-FDG PET Imaging To investigate the metabolic change of the photochemically induced brain stroke with time, 18F–FDG PET imaging was performed at day 1, 3, 7, and 14 after induction of stroke. As shown in Fig. 6a, a remarkably reduced 18F-FDG uptake was noted at the site of infarct lesion and moderately reduced uptake in the remote region of ipsilateral hemisphere at day 1 following induction of stroke. At day 3, the cortical metabolic defect partially recovered in the outer region of the ischemic zone and almost completely recovered in the remote cortex. At day 14, the 18F-FDG uptake at the infarct site almost completely recovered. MDI was calculated to monitor the change of glucose utilization in the infarct lesion (zone A), in the region at one-pixel width surrounding the lesion (zone B), in the remote region of the ipsilateral cortex (zone C), as well as in the hippocampus (zone D). MDIs of zones A and B were decreasing with time, indicating that the glucose consumption was gradually restored at both the infarct region and the margin (Fig. 6b). There was significant decrease in MDIs with time in zone A and zone B in 14 days (ANOVA, p < 0.05). From day 7 to day 14, MDI of zone A decreased by 10 % compared to 20 % decrease in zone B, indicating that the recovery of glucose metabolism at the margin of the infarct is better than that in the infarct core. There was no significant statistical difference of MDI observed for 14 days in zones C and D (ANOVA, p > 0.05), suggesting no obvious compromise of glucose metabolism observed in the ipsilateral remote cortex and hippocampus. Using photothrombotic method, infarct induction at somatosensory cortex did not interfere the glucose utilization in the remote ipsilateral cortex and hippocampus and contralateral cortex as well.Fig. 6 Sequential 18F-FDG PET imaging of rats with photochemical thrombosis on the right somatosensory cortex. a Transaxial tomographic images at day 1, 3, 7, and 14 after stroke induction.

cortex did not interfere the glucose utilization in the remote ipsilateral cortex and hippocampus and contralateral cortex as well.Fig. 6 Sequential 18F-FDG PET imaging of rats with photochemical thrombosis on the right somatosensory cortex. a Transaxial tomographic images at day 1, 3, 7, and 14 after stroke induction. b Metabolic difference index (MDI) at ROIs of infarct zone (zone A), one-pixel width surrounding the lesion (zone B, one-pixel width surrounding the infarct lesion), remote region of the ipsilateral cortex (zone C), and hippocampus (zone D) showed at each time point. There was significant decrease in MDIs with time in zone A and zone B in 14 days (p < 0.05). Stroke rats, n = 5; shame rats, n = 3

width surrounding the lesion (zone B, one-pixel width surrounding the infarct lesion), remote region of the ipsilateral cortex (zone C), and hippocampus (zone D) showed at each time point. There was significant decrease in MDIs with time in zone A and zone B in 14 days (p < 0.05). Stroke rats, n = 5; shame rats, n = 3 Evolution of BBB Permeability in Photothromboic Stroke Evidence indicates that BBB disruption occurs after ischemic stroke and leads to the leakage of normally excluded substances into interstitial fluid of the brain. To evaluate the extent of BBB disruption and its restoration, we performed the EB staining and ex vivo fluorescence imaging at day 1, 3, 7, and 14 post-stroke. Figure 7a demonstrated an infarction core surrounded by an EB-stained zone in the somatosensory cortex at day 1 after photochemical induction. Subsequently, the leakage of BBB gradually increased at day 3 and began to restore at day 7; a small EB-stained lesion still remained at day 14. The ex vivo fluorescence imaging showed a high intensity fluorescence corresponding to the EB-stained ischemic lesion. The fluorescence signals were much less intense and less extensive at day 14. Quantitative analysis of EB fluorescence at the ROIs of ischemic region showed a markedly reduced signal intensity at day 14 compared to the highest intensity at day 3 (Fig. 7b, p < 0.05). Fluorescent signal at peri-infarct region was much lower than that at infarct core throughout 7 days and greatly reduced to a background level while the infarct core still showed a high signal at day 14. These results suggest a favorable self-repair of BBB in 14 days after photothrombotic stroke.Fig. 7 Detection of blood-brain barrier disruption by EB staining. a Representative images of EB staining and fluorescence imaging of rat brain at day 1, 3, 7, and 14 after ischemia induction. b Quantitative analysis of EB fluorescence at the ROIs of ischemic region (n = 3 at each time point). It suggested that the leakage of BBB was most remarkable at day 3 and was subsequently decreasing with time

e images of EB staining and fluorescence imaging of rat brain at day 1, 3, 7, and 14 after ischemia induction. b Quantitative analysis of EB fluorescence at the ROIs of ischemic region (n = 3 at each time point). It suggested that the leakage of BBB was most remarkable at day 3 and was subsequently decreasing with time Discussion In the present study, we have successfully assessed the evolutional changes of cellular viability in the lesion, infarct volume, brain edema, infiltration of inflammatory cells and astrocytes, neovascularization, CBF, glucose metabolism, and BBB permeability (Fig. 8). To the best of our knowledge, this is the first study conducted using 18F-FDG/PET and MRI to longitudinally evaluate the metabolic and hemodynamic changes of the brain in the rat model of photochemically induced cerebral infarction.Fig. 8 Evolutional characterization of photochemically induced stroke. The changes of tissue viability, infarct lesion volume, brain edema, inflammatory cells and astrocyte infiltration, neovascularization, CBF, glucose metabolism, and BBB permeability at each time point were summarized

ally induced cerebral infarction.Fig. 8 Evolutional characterization of photochemically induced stroke. The changes of tissue viability, infarct lesion volume, brain edema, inflammatory cells and astrocyte infiltration, neovascularization, CBF, glucose metabolism, and BBB permeability at each time point were summarized The method applied in current study was categorized as “end-artery occlusion in the cortex” described by Chen et al. [23]. It is the classical and simplest method of photochemically induced stroke model. Although this method can be achieved with intact skull, however, to avoid the scatter and reflection of light when illuminating through skull, a cranial window was made before illumination. In this way, the infarction is anticipated to be more sharply demarcated and more consistent in lesion size than using intact skull. Also, the volume of the ischemic lesion can be controlled by manipulating the intensity of irradiating light and the size of irradiated zone in the cranial window [11]. In the current study, a laser beam of 532 nm wavelength was applied instead of arc beam irradiated system with 560 nm wavelength. Less intensity of the irradiating light causing less severity of the induced stroke may account for the short period of ipsilateral hypometabolism and lack of compromised glucose utilization in contralateral cortex.

aser beam of 532 nm wavelength was applied instead of arc beam irradiated system with 560 nm wavelength. Less intensity of the irradiating light causing less severity of the induced stroke may account for the short period of ipsilateral hypometabolism and lack of compromised glucose utilization in contralateral cortex. To confirm the success of photochemical induction, we applied LSCI through the cranial window of the rat before and after thrombosis was induced to assess the occlusion of cerebral vasculature caused by photochemically induced thrombosis. Laser speckle imaging is useful to assess cerebral vasculature and CBF noninvasively with high temporal and spatial resolution [28]. It has been used to study changes of CBF after distal middle cerebral artery ligation in mice [31]. Upon photochemical induction, LSCI demonstrated blockade of the majority of blood vessels in the illuminated region, indicating successful induction of cerebral ischemia and/or infarction.

al and spatial resolution [28]. It has been used to study changes of CBF after distal middle cerebral artery ligation in mice [31]. Upon photochemical induction, LSCI demonstrated blockade of the majority of blood vessels in the illuminated region, indicating successful induction of cerebral ischemia and/or infarction. In the evaluation of tissue viability and morphologic change in infarct lesion, a wedge-shaped TTC-negative area appeared at day 1 after stroke induction and the size gradually reduced through day 3 to day 7 and was eventually not detectable at day 14. The result of H&E stain showed increased number of infiltrated cells in infarct region since day 3, and some cells were detected in the infarct core at day 14. The identity of these infiltrated cells was further confirmed to be macrophages, glial, endothelial, and smooth muscle cells, but not neuronal cells by IHC staining. Infiltration of cells into the infarct region caused the gradual shrinkage of TTC-negative area.

d some cells were detected in the infarct core at day 14. The identity of these infiltrated cells was further confirmed to be macrophages, glial, endothelial, and smooth muscle cells, but not neuronal cells by IHC staining. Infiltration of cells into the infarct region caused the gradual shrinkage of TTC-negative area. Inflammation is strongly linked to the development of stroke. Macrophages and microglia, two important types of inflammatory cells in response to stroke, have significant effect on removing debris and activating inflammatory cascades involved in repairing tissue damage [32, 33]. In the present study, IHC staining showed that CD68 positive macrophages/microglia and GFAP-expressing astrocytes notably accumulated along the boundary of infarct lesion at day 14, forming a belt consisting of inflammatory cells and astrocytes. Simultaneously, some microglia/macrophages and astrocytes were detected in the core of infarct lesion. Previous reports suggested that astrocytes produce pro-inflammatory cytokines and chemokines which subsequently recruit microglia/macrophages for dead cell clearance [34, 35]. However, excessive inflammation may have adverse effect, i.e., inducing free radicals which are harmful to neuronal cells [36]. Also, excessive thickness of astrocyte-associated scar blocks neuronal cells from migrating into the injured site. This evidence further supports our result in which neuronal cells were not detectable in the ischemic lesion throughout 14 days after stroke induction.

ee radicals which are harmful to neuronal cells [36]. Also, excessive thickness of astrocyte-associated scar blocks neuronal cells from migrating into the injured site. This evidence further supports our result in which neuronal cells were not detectable in the ischemic lesion throughout 14 days after stroke induction. Vasogenic edema examined by T2WI is strongly correlated with BBB disruption after stroke [37, 38]. This was also observed in present study. Hyperintensity in T2WI at day 1 and day 3 post-stroke was paralleled with the intensity of EB fluorescence, indicating the linkage between vasogenic edema and BBB permeability in photothrombotic stroke. Serial DWI in the current study revealed an intense halo located at the periphery of the stroke lesion seen on T2WI. This hyperintense halo seen on DWI indicated the zone with cytotoxic edema and began to gradually reduce since day 3 post-stroke. MR images of photochemical stroke in a previous report also showed early increase in T2 signal and decreased diffusion of water, indicating the simultaneous development of substantial vasogenic edema and ischemic infarction [39]. This pattern is different from that seen in human stroke, where infarcts develop with cytotoxic edema, followed by a vasogenic edema which is delayed by several hours [3].

in T2 signal and decreased diffusion of water, indicating the simultaneous development of substantial vasogenic edema and ischemic infarction [39]. This pattern is different from that seen in human stroke, where infarcts develop with cytotoxic edema, followed by a vasogenic edema which is delayed by several hours [3]. The PWI demonstrated a region of absent CBF at the infarct site as well as the edematous region. CBFDI revealed progressive restoration of CBF in the infarct zone and returned to normal at day 14. CBF in the peri-infarct zone was less compromised than that in the infarct zone and returned to normal at day 7. The status of CBF was further confirmed and well correlated with IHC results which showed dilated vessels close to the infarct core at day 3 and neovascularization at day 7 to 14.

turned to normal at day 14. CBF in the peri-infarct zone was less compromised than that in the infarct zone and returned to normal at day 7. The status of CBF was further confirmed and well correlated with IHC results which showed dilated vessels close to the infarct core at day 3 and neovascularization at day 7 to 14. In current study, progressive recovery of glucose utilization was noted at the infarct lesion, peri-infarct zone, and ipsilateral remote cortex. CBF recovery and progressive cell infiltration subsequently causing enhancement of glucose utilization may partially account for the rapid recovery of 18F-FDG uptake in the infarct zone and its vicinity. Analysis of the MDI is more sensitive than visual inspection of hypometabolic lesion. MDI at day 14 revealed relatively higher restoration of glucose consumption at the peri-infarct zone than that in the infarct core (10 vs 20 %, respectively). Several preclinical studies of small animal ischemic stroke models have consistently revealed an increased 18F-FDG uptake in the peri-infarct regions due to the effects of activation of glucose transporters, hexokinase, and neuroinflammation [40]. Previous study which used the model of photothrombotic MCAo in rats showed that 18F-FDG uptake in the peri-ischemic areas was comparable to the normal brain regions at day 1 and 3 and notably elevated at days 7 and 14 [41]. In the present study, glucose utilization in the peri-infarct region was decreased compared to that in the normal region throughout 14 days after stroke induction. However, progressive recovery of glucose utilization in this region was noted. These different results of glucose metabolism in peri-infarct area might attribute to different animal stroke models, time of evaluation, and imaging techniques.

ompared to that in the normal region throughout 14 days after stroke induction. However, progressive recovery of glucose utilization in this region was noted. These different results of glucose metabolism in peri-infarct area might attribute to different animal stroke models, time of evaluation, and imaging techniques. It is commonly agreed that no salvageable penumbral tissue exists in the model of photochemically induced stroke. This disadvantageous phenomenon of this model is different to that of other embolic models where penumbra is much more like those in human stroke and much larger. Hilger and colleagues named peri-infarct zone of photothrombotic stroke as “region-at-risk,” where a little viable cells, low energy metabolism, and vasogenic edema existed [42]. In present study, the number of inflammatory cells and astrocytes, as well as neovascularization, gradually increased in peri-infarct region. At day 14, a belt consisting of inflammatory cells and astrocytes was observed surrounding the infarct lesion. In addition, vasogenic edema and cytotoxic edema progressively reduced, and CBF and glucose metabolism gradually improved. The leakage of EB dye at the peri-infarct zone significantly reduced than that in the infarct core. These data suggests the recoverable BBB leakage, neovascularization, CBF, and glucose metabolism at peri-infarct region rather than at infarct core.

ogressively reduced, and CBF and glucose metabolism gradually improved. The leakage of EB dye at the peri-infarct zone significantly reduced than that in the infarct core. These data suggests the recoverable BBB leakage, neovascularization, CBF, and glucose metabolism at peri-infarct region rather than at infarct core. Although we have longitudinally characterized the pathophysiological changes of photothrombotic brain ischemia by multiple imaging modalities and cellular immunostaining, some limitations should be addressed in the current study. First, photochemically induced occlusion occurred in vessels within the irradiation area, where the mechanism is different from that of clinical ischemic stroke which was usually caused by an embolus or two. Only a little or no local collateral flow/reperfusion and ischemic penumbra occur in this model. Despite that the cellular changes in this model are still evolving and follow a similar trend to those of other occlusive models, the pathomechanism is different from those seen in human stroke. Second, evaluation was not performed at earlier time points (3–12 h post-stroke) due to poor animal condition early after stroke induction for 18F-FDG/PET and MRI examination. Sequential assessment of the evolutional change of photochemically induced rat stroke model during the first 24 h following stroke will be carried out in the future study. Third, we did not perform quantitation of blood flow of regional cerebral vasculature by LSCI because the measurement of CBF is limited to the superficial cortex. Fourth, the experiments of multi-modality imaging were performed on the same animal as possible as we could for serial imaging protocol. However, to avoid being anesthetized for a long period of time on the same animal, PET and MRI imaging were not performed on the same day. Despite these drawbacks and limitations of this model, the advantages of easy manipulation, highly reproducible, lesion controllable, and low mortality have made this method comprehensively applied. Further, the consistent lesion made in the cortex where cellular changes evolve and follow a similar trend to those of the occlusive models makes it a suitable model for the study of neurorestorative therapy of pharmaceuticals or stem cells. In the scope of translational purpose, this method had been applied on animals other than rodents, such as rabbits and pigs [43–46].

e cellular changes evolve and follow a similar trend to those of the occlusive models makes it a suitable model for the study of neurorestorative therapy of pharmaceuticals or stem cells. In the scope of translational purpose, this method had been applied on animals other than rodents, such as rabbits and pigs [43–46]. Conclusion The evolution of photochemically induced stroke model in rats has been longitudinally characterized for 14 days by 18F-FDG/PET, MRI, IVIS, histopathology, and immunohistochemistry examination. Within 14 days after stroke induction, we found the occurrence of early brain edema, infiltration of inflammatory cells and astrocytes, neovascularization in the infarct core and peri-infarct zone, improvement of CBF, glucose metabolism, and BBB permeability. These serial changes characterized in this study provide better understanding of cerebral ischemia and are highly beneficial to the development of therapeutic strategies for ischemic stroke.

ytes, neovascularization in the infarct core and peri-infarct zone, improvement of CBF, glucose metabolism, and BBB permeability. These serial changes characterized in this study provide better understanding of cerebral ischemia and are highly beneficial to the development of therapeutic strategies for ischemic stroke. Abbreviation BBB, blood-brain barrier; BSA, bovine serum albumin;14C-2-DG, 14C-2-Deoxy-D-Glucose;CBF, cerebral blood flow; CBFDI, difference indexes;CPMG, Carr, Purcell, Meiboom, Gill; CT, computed tomography; DW, diffusion weighted; DWI, diffusion-weighted image (imaging); EB, Evans blue;18F-FDG, 18F-2-deoxy-glucose; FOV, field of view; GFAP, glial fibrillary acid protein; H&E, hematoxylin-eosin; IHC, immunohistochemistry; IVIS, In Vivo Imaging System; LSCI, Laser speckle contrast imaging;MCAo, middle cerebral artery occlusion; MDI, metabolic difference index; MRI, magnetic resonance imaging; OSEM, ordered subsets expectation maximization; PBS, phosphate buffer saline; PET, positron emission tomography; PWI, perfusion-weighted image (imaging); RF, radio frequency; ROI, Region of Interest; S.D., standard deviation; αSMA, alpha smooth muscle actin; TBS, tris-HCL buffered solution; TIR, inversion recovery time; TTC, 2, 3, 5-triphenyl tetrazolium chloride; TE, echo time; TR, repetition time; T2W, T2 weighted; T2WI, T2-weighted image (imaging); VOIs, volume of interests; vWF, von Willebrand Factor;%ID/cm3, percent injected dose per c.c; Nai-Wei Liu and Chien-Chih Ke contributed equally to this work

Abbreviation BBB, blood-brain barrier; BSA, bovine serum albumin;14C-2-DG, 14C-2-Deoxy-D-Glucose;CBF, cerebral blood flow; CBFDI, difference indexes;CPMG, Carr, Purcell, Meiboom, Gill; CT, computed tomography; DW, diffusion weighted; DWI, diffusion-weighted image (imaging); EB, Evans blue;18F-FDG, 18F-2-deoxy-glucose; FOV, field of view; GFAP, glial fibrillary acid protein; H&E, hematoxylin-eosin; IHC, immunohistochemistry; IVIS, In Vivo Imaging System; LSCI, Laser speckle contrast imaging;MCAo, middle cerebral artery occlusion; MDI, metabolic difference index; MRI, magnetic resonance imaging; OSEM, ordered subsets expectation maximization; PBS, phosphate buffer saline; PET, positron emission tomography; PWI, perfusion-weighted image (imaging); RF, radio frequency; ROI, Region of Interest; S.D., standard deviation; αSMA, alpha smooth muscle actin; TBS, tris-HCL buffered solution; TIR, inversion recovery time; TTC, 2, 3, 5-triphenyl tetrazolium chloride; TE, echo time; TR, repetition time; T2W, T2 weighted; T2WI, T2-weighted image (imaging); VOIs, volume of interests; vWF, von Willebrand Factor;%ID/cm3, percent injected dose per c.c; Nai-Wei Liu and Chien-Chih Ke contributed equally to this work Acknowledgements This research was supported by the grants: NSC 102-2314-B-010-038-MY3, NSC 102-2627-M-010-003 (Ministry of Science and Technology), MOHW105-TDU-B- 211-134-003 (Department of Health), V105C-038 (Taipei Veterans General Hospital), 105AC-BI1 (BMIRC, National Yang-Ming University). The authors thank the technical support from Molecular and Genetic Imaging Core, Taiwan Mouse Clinic (MOST 104-2325-B-001-011) which is funded by the National Research Program for Biopharmaceuticals (NRPB) at the Ministry of Science and Technology (MOST) of Taiwan, and Ms. Tsuey-Ling Jan. for the assistance of preparing the manuscript This research was supported by the grants: The Science and technology development fund, Macau (FDCT No. 089/2012/A3 & 106/2014/A3), NSC 102-2314-B-010-038-MY3, NSC 102-2627-M-010-003 (Ministry of Science and Technology), MOHW105-TDU-B- 211-134-003 (Department of Health), V105C-038 (Taipei Veterans General Hospital), 105 AC-BI1 (BMIRC, National Yang-Ming University). The authors thank the technical support from Molecular and Genetic Imaging Core, Taiwan Mouse Clinic (MOST 104-2325-B-001-011) which is funded by the National Research Program for Biopharmaceuticals (NRPB) at the Ministry of Science and Technology (MOST) of Taiwan, and Ms. Tsuey-Ling Jan. for the assistance of preparing the manuscript.

the technical support from Molecular and Genetic Imaging Core, Taiwan Mouse Clinic (MOST 104-2325-B-001-011) which is funded by the National Research Program for Biopharmaceuticals (NRPB) at the Ministry of Science and Technology (MOST) of Taiwan, and Ms. Tsuey-Ling Jan. for the assistance of preparing the manuscript. Authors’ Contributions YZ and RSL conceived and designed the experiment. NWL, CCK, KCC, DTT, JSL, YYC, IAC, TWH, YJH, and CWC performed the experimental work. BHY, WSH, and CCK analyzed the data. NWL and CCK wrote the manuscript. Compliance with Ethical Standards Conflicts of Interest All authors declare no conflict of interest. Ethical Approval All applicable national and institutional guidelines for the care and use of animals were followed. This article does not contain any studies with human participants performed by any of the authors.

tive albumin levels were calculated based on densitometry analysis. The mean albumin level of the sham group was normalized to 1.0. c Recorded brain water content at 72 h post-ICH. All data are displayed as means ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 Additionally, brain water content was examined to explore the effects of melatonin treatment on ICH-induced brain edema. No significant differences were noted in the contralateral cortex, contralateral basal ganglia, or cerebellum in these four groups (Fig. 2c). However, in the ipsilateral cortex and ipsilateral basal ganglia of these four groups, the ICH group showed significant increases in water content compared to the sham group, while melatonin treatment significantly impaired the brain water content. These results indicated that melatonin treatment is able to ameliorate brain injury (including neurological behavioral impairment, BBB disruption, and brain edema) after ICH.

owed significant increases in water content compared to the sham group, while melatonin treatment significantly impaired the brain water content. These results indicated that melatonin treatment is able to ameliorate brain injury (including neurological behavioral impairment, BBB disruption, and brain edema) after ICH. Melatonin Inhibited Oxidative Stress in Brain Tissues at 72 h After ICH To explore the effects of melatonin in oxidative stress after ICH, the protein levels of NOX-1 and NOX-2, which are indicators of oxidative stress, were detected by western blot analysis at 72 h after ICH. The results showed that these two indicators were significantly increased in the ICH group compared to the sham group, while no significant difference between the ICH and ICH + vehicle groups was noted. However, melatonin treatment significantly decreased expression levels of these two indicators when compared to the ICH + vehicle group (Fig. 3a–c).Fig. 3 Oxidative stress indicator expression levels in brain tissues at 72 h after ICH. a Western blot analysis shows protein levels of NOX-1 and NOX-2 in the sham, ICH, ICH + vehicle, and ICH + melatonin groups. b, c Relative NOX-1 and NOX-2 expression level calculations based on densitometry analysis. The mean values of NOX-1 and NOX-2 within the sham group were normalized to 1.0. All data are displayed as a mean ± SEM, with *P < 0.05 and & P < 0.05 deemed significant difference. NS, no significant difference compared to the sham group (n = 6). d Effects of melatonin treatment on brain ROS levels at 72 h post-ICH. The mean ROS value of the sham group was normalized to 1.0. All data are displayed as a mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference. NS, no significant difference compared to the sham group (n = 6)

sham group (n = 6). d Effects of melatonin treatment on brain ROS levels at 72 h post-ICH. The mean ROS value of the sham group was normalized to 1.0. All data are displayed as a mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference. NS, no significant difference compared to the sham group (n = 6) Additionally, the ROS levels in these four groups were also measured at 72 h after ICH induction (Fig. 3d). The results showed that the ROS level was significantly increased in the ICH group compared to the sham group, while melatonin treatment remarkably reduced ROS level in brain tissues. These results suggested that melatonin can inhibit oxidative stress in brain tissues after ICH.

successful pellet-reaching attempts, indicating that a higher intensity of training may be beneficial. Since the effect was most pronounced on day 14 and less evident on day 28, our experimental data argue for continued motor rehabilitation for stroke patients in order to maintain a sustained effect on motor function. Summary Although studies generally support the concept that motor rehabilitation is associated with improved outcomes after stroke, few experimental studies have specifically tested whether focused rehabilitation improves other motor outcomes. To our knowledge, our study represents the first translational evidence that task-specific training after stroke generalizes to a different motor task. Electronic Supplementary Material ESM 1 (DOCX 15 kb) . Electronic supplementary material The online version of this article (doi:10.1007/s12975-016-0519-x) contains supplementary material, which is available to authorized users. Acknowledgements This project was supported by the Swiss National Science Foundation Marie Heim-Vögtlin program, the UZH Filling the Gap foundation, and the P & K Pühringer Foundation. Authors’ Contributions M.E., P.B., and O.B. performed the research and analyzed the data. S.W. designed the research. M.E., S.W., and A.L. wrote the manuscript. All authors reviewed and edited the manuscript. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflict of interest.

Introduction About 60% of stroke survivors suffer from motor disability 6 months after stroke [1, 2]. By training of motor skills, rehabilitation aims to maximize patients’ functional independence and quality of life. The physiological mechanisms of training interventions are incompletely understood, especially their generalization, i.e., how and how much improvement in the specific task trained generalizes to other movements. These mechanisms need to be explored in animal models to optimize and develop treatments. In rodents, post-stroke motor rehabilitation by pellet-reaching training improves pellet-reaching success [3]. This is accompanied by reorganization in motor cortex regions controlling the affected limb [4], e.g., an increase in dendritic complexity [5, 6]. The issue of generalization of trained to other tasks has not been addressed in animal models of post-stroke recovery. The present study investigated whether motor training by pellet reaching translates into improvement in other motor tasks in a rat stroke model. The transient middle cerebral artery occlusion (MCAO) was chosen for stroke induction, because the lesion is not confined to the motor cortex but has a variable spread towards adjacent cortical and subcortical areas, similar to human stroke.

aching translates into improvement in other motor tasks in a rat stroke model. The transient middle cerebral artery occlusion (MCAO) was chosen for stroke induction, because the lesion is not confined to the motor cortex but has a variable spread towards adjacent cortical and subcortical areas, similar to human stroke. Materials and Methods All experiments were performed in accordance with the guidelines and regulations approved by the Federal Veterinary Office of Switzerland (Veterinary Office of the Canton of Zurich). Adult male Sprague Dawley rats (280 to 310 mg body weight) were used. The experimental setup is shown in Fig. 1. Out of 46 animals, 5 animals died or had to be euthanized prematurely due to massive infarction. Five rats were excluded because of insufficient stroke induction as judged by less than 15 s to remove the sticky tape on day 1 after MCAO. Analysis of sensorimotor scores and stroke lesion size was performed by an investigator blinded to group assignment.Fig. 1 Flow of the experiments. a Experimental schedule. b Photos illustrating rats during pellet-reaching training. c Representative MRI-T2 images from the rehabilitation and no rehabilitation group 28 days after MCAO

nsorimotor scores and stroke lesion size was performed by an investigator blinded to group assignment.Fig. 1 Flow of the experiments. a Experimental schedule. b Photos illustrating rats during pellet-reaching training. c Representative MRI-T2 images from the rehabilitation and no rehabilitation group 28 days after MCAO Middle Cerebral Artery Occlusion During surgery, rats were anesthetized using facemask inhalation of 1.5 to 2.5% isoflurane in a 2:1 N2O:O2 atmosphere. Animals were subjected to 60-min MCAO as described previously [7]. Laser Doppler flowmetry (Moor Instruments Ltd., Millwey, UK) was used to confirm the occlusion of the MCA. The probe was fixed on the skull in the left MCA territory and rats with less than 30% drop in cerebral blood flow were excluded from the studies. The body temperature was monitored throughout surgery using a rectal probe and maintained at 37 °C with a normothermic blanket (Harvard Apparatus, Edenbridge, Kent, UK). Pre-training, Motor Skill Learning, and Motor Rehabilitation Training procedures were conducted as previously described [8]. More details are available as supplemental data. One daily session consisting of 100 trials or a maximum time of 45 min was performed for each animal. Reaching performance was measured by counting the number of successful reaches.

and Motor Rehabilitation Training procedures were conducted as previously described [8]. More details are available as supplemental data. One daily session consisting of 100 trials or a maximum time of 45 min was performed for each animal. Reaching performance was measured by counting the number of successful reaches. Animals were randomized into the rehabilitation (“rehab”) or no rehabilitation (“no rehab”) group on day 4 after MCAO. Five days after MCAO (D5), animals of the rehab group received daily pellet-reaching sessions for 7 days until D12. The number of successful reaches was tested in all animals at D5 and D12 after MCAO. Sensorimotor Testing Sensorimotor function was evaluated using the adhesive tape removal test as well as a composite observational neurological score (see supplemental methods) [7]. Magnetic Resonance Imaging Methods On day 28 after MCAO, magnetic resonance imaging (MRI) was carried out on a 4.7-T rodent scanner (see also supplemental material).

Sensorimotor Testing Sensorimotor function was evaluated using the adhesive tape removal test as well as a composite observational neurological score (see supplemental methods) [7]. Magnetic Resonance Imaging Methods On day 28 after MCAO, magnetic resonance imaging (MRI) was carried out on a 4.7-T rodent scanner (see also supplemental material). Statistical Analysis A sample size calculation (alpha 0.05, power 0.8) was performed based on previous sticky tape removal test data in rats after MCAO. We calculated a minimum sample size of 16 animals per group to show an effect size of 0.9. Power calculation was performed using G Power Software (version 3.1.5). Statistical analyses were done using SPSS v12.0 for Windows. All values are given as mean ± standard error of mean (s.e.m.). For group comparisons, either the two-sided independent sample t test or the nonparametric Mann-Whitney U test was used depending on data distribution. A p value <0.05 was considered significant.

Statistical analyses were done using SPSS v12.0 for Windows. All values are given as mean ± standard error of mean (s.e.m.). For group comparisons, either the two-sided independent sample t test or the nonparametric Mann-Whitney U test was used depending on data distribution. A p value <0.05 was considered significant. Results Before stroke surgery, both groups achieved a similar reaching performance (38.1 ± 7.1% success rate for rehab versus 32.5 ± 6.2% for no rehab, Fig. 2a). Five days after stroke, all animals had considerable problems in obtaining pellets with the impaired limb (2.1 ± 1.0% in rehab and 2.0 ± 0.9% in no rehab group). As expected, animals that received daily rehabilitation from D5 to D12 after MCAO achieved a higher pellet-reaching success rate at D12 (23.0 ± 5.5% versus 4.2 ± 2.6% p < 0.01; Fig. 2a). MRI stroke lesion analysis showed no difference between the two groups (165 ± 41.9 mm3 in rehab versus 176 ± 38.7 mm3 in no rehab animals). In the composite neurological score, trained and control animals where similarly affected after stroke (Fig. 2b). In the sticky tape test, deficits to perceive and remove the tape on the right side were noted in all animals at D1 after MCAO (Fig. 2c). However, animals with rehabilitation performed faster in the sticky tape removal task (motor component) on D14 than animals without rehabilitation (12 ± 2.6 s versus 38 ± 10.2 s; p = 0.007). A high-cumulative number of pellet reaches during training was negatively correlated with the time to remove the sticky tape on D14 (R = −0.68, p < 0.01, Fig. 2d).Fig. 2 Pellet-reaching training improved motor function after MCAO. a Success rate for reaching task of rehab (n = 18) and no rehab rats (n = 18) from pre-stroke to D12. b Composite neurological score. c Sticky tape test: latency to remove (left) and contact (right) a sticky tape applied to the right forepaw on all time points tested. The gray-shaded box in a–c indicates duration of rehabilitation. No repeated measures ANOVA was performed because of the short (7d) intervention period. Lower left: latency to remove sticky tape at D14 in rehab versus no rehab rats. d Correlation between cumulative number of pellet reaches and latency to remove the sticky tape at D14 in rehab rats

dicates duration of rehabilitation. No repeated measures ANOVA was performed because of the short (7d) intervention period. Lower left: latency to remove sticky tape at D14 in rehab versus no rehab rats. d Correlation between cumulative number of pellet reaches and latency to remove the sticky tape at D14 in rehab rats Discussion Although intense rehabilitation incorporating exercise, forelimb constraint therapy and task-specific pellet reaching have been shown to enhance performance in a similar but not identical task-specific test (tray-reaching) in rats after brain injury, our study demonstrates that a task-oriented motor rehabilitation algorithm alone can improve motor function in a task that is substantially different from the trained one [9]. Daily pellet-reaching training enhanced motor ability for sticky tape removal.

al task-specific test (tray-reaching) in rats after brain injury, our study demonstrates that a task-oriented motor rehabilitation algorithm alone can improve motor function in a task that is substantially different from the trained one [9]. Daily pellet-reaching training enhanced motor ability for sticky tape removal. Motor rehabilitation by pellet-reaching training is focused on recovery of highly specific functions (skilled grasping ability) and requires intensive training and practice of the impaired function [10]. As potential mechanisms mediating the beneficial effects of motor rehabilitation, pro-plastic and neurotrophic factors such as brain-derived neurotrophic factor (BDNF) and insulin-like growth factor (IGF-1) have been indicated [11]. The control group did not show significant recovery in the skilled reaching task. Neither the sensory component of the sticky tape test nor the composite neurologic score was influenced by motor rehabilitation in our model. This argues for specific effects on motor recovery and against other functional domains, such as sensation or neglect involved in the observed regain of function. The improvement in sticky tape removal was correlated with more successful pellet-reaching attempts, indicating that a higher intensity of training may be beneficial. Since the effect was most pronounced on day 14 and less evident on day 28, our experimental data argue for continued motor rehabilitation for stroke patients in order to maintain a sustained effect on motor function.

Introduction Blood-brain barrier (BBB) breakdown is a major contributing factor to ischemic brain injury or hemorrhagic transformation (HT) and often leads to poor outcomes in acute ischemic stroke patients receiving recombinant tissue plasminogen activator (rt-PA) treatment [1, 2]. Tight-junction proteins such as claudin, junctional adhesion molecule, and occludin play essential roles in maintaining BBB integrity. Measurements of permeability-surface area product (PS), an indicator of BBB permeability, by computer tomography perfusion (CTP) imaging have been successfully applied for early identification of BBB damage and HT development in acute stroke patients [3, 4]. Sodium tanshinone IIA sulfonate (STS) is a water-soluble derivative of tanshinone IIA, a main bioactive component isolated from the roots of the Chinese herb Salviae miltiorrhiza Bunge (Danshen) [5]. STS has been widely used for treatments of cardiovascular and cerebrovascular diseases in China [6, 7]. In mice with cerebral ischemia, STS could protect BBB and had a patent in China [8, 9]. Here, we hypothesized that STS could work as a BBB protective agent that help acute ischemic stroke patients who received thrombolysis treatment recover better. In this study, we used CTP-derived PS to reveal whether treatment with STS could reduce BBB leakage in acute ischemic stroke patients receiving rt-PA thrombolysis and sought to investigate the underlying mechanisms.

ctive agent that help acute ischemic stroke patients who received thrombolysis treatment recover better. In this study, we used CTP-derived PS to reveal whether treatment with STS could reduce BBB leakage in acute ischemic stroke patients receiving rt-PA thrombolysis and sought to investigate the underlying mechanisms. Materials and Methods Participants This single-centered, randomized, double-blinded prospective study was approved by the ethics committee of Drum Tower Hospital, Medical School of Nanjing University. An entry was made in the Chinese clinical trial registry (ChiCTR-ONRC-14004659). The inclusion criteria were as follows: (1) acute ischemic stroke patients receiving rt-PA treatment, (2) at age between 18 to 80 years old, and (3) willing to participate in all follow-up neurologic and imaging examinations. We excluded patients with presence of platelet abnormalities (PLT < 100 or >300 × 109/L), severe bleeding disorders, contraindications to iodinated contrast agent, a history of severe renal failure, or known or suspected infection. Informed consents were obtained from all participants involved in this study. Randomization sequences were computer generated. The flow chart of patient cohort selection is shown in Fig. 1.Fig. 1 Flow chart showing patient cohort selection in this study

history of severe renal failure, or known or suspected infection. Informed consents were obtained from all participants involved in this study. Randomization sequences were computer generated. The flow chart of patient cohort selection is shown in Fig. 1.Fig. 1 Flow chart showing patient cohort selection in this study Treatments After rt-PA thrombolysis, STS (60 mg per day) was intravenously administrated daily to patients in the STS group for 10 days, while patients in the placebo group received equivalent administrations of saline. All participants received aspirin after thrombolysis 24 h for acute ischemic stroke. Neurologic Assessment Neurologic function outcomes were assessed using the National Institutes of Health Stroke Scale (NIHSS) score at 0, 24 h, and 10 days; activities of daily living (ADLs) score at 0 h and 10 days; and the modified Rankin Scale (mRS) at 90 days. CTP Scan Protocol and Image Analysis CT perfusion and non-contrast CT scans were performed on a 64-slice CT scanner (Discovery CT750 HD, GE Healthcare, Milwaukee, WI, USA). CTP started with intravenous injection of 50 mL of iodinated contrast agent (350 mg/mL, Omnipaque, GE Healthcare, Shanghai, China) followed by a saline flush of 45 mL at 5 mL/s. CTP scans began after a delay of 5 s from contrast injection, and the following technical settings were applied: 80 kVp, 150 mAs, temporal sampling rate of 2 s for 60 s, and total axial coverage of 40 mm at 5-mm slice thickness.

, Omnipaque, GE Healthcare, Shanghai, China) followed by a saline flush of 45 mL at 5 mL/s. CTP scans began after a delay of 5 s from contrast injection, and the following technical settings were applied: 80 kVp, 150 mAs, temporal sampling rate of 2 s for 60 s, and total axial coverage of 40 mm at 5-mm slice thickness. All CTP images were analyzed by experienced radiologists who were blinded to this research. For the PS maps and data, manually drawn region of interests (ROIs) in the ipsilateral hemisphere were compared to those in the contralateral hemisphere. PS values were calculated using CTP software (CT Kinetics, GE Healthcare, China). The corresponding ROIs for the contralateral side were generated automatically by mirroring the ipsilateral ROIs, and the relative PS (rPS) was defined as ipsilateral/contralateral ROIs. Measurement of BBB Damage Biomarkers Venous blood were collected at 0, 24 h, and 10 days, and the separated serum was stored in aliquots at −80 °C until biochemical analysis. Serum levels of matrix metalloproteinase (MMP)-2 (R&D, Minneapolis, IL, USA), MMP-9 (R&D, Minneapolis, IL, USA), tissue inhibitor of metalloproteinase (TIMP)-1 (R&D, Minneapolis, IL, USA), claudin-5 (CusaBio, Wu han, China), and zonula occluden (ZO)-1 (CusaBio, Wu han, China) were measured using commercially available ELISA kits according to the manufacturer’s instructions.

inneapolis, IL, USA), MMP-9 (R&D, Minneapolis, IL, USA), tissue inhibitor of metalloproteinase (TIMP)-1 (R&D, Minneapolis, IL, USA), claudin-5 (CusaBio, Wu han, China), and zonula occluden (ZO)-1 (CusaBio, Wu han, China) were measured using commercially available ELISA kits according to the manufacturer’s instructions. Study Outcomes The outcomes were the integrity of blood-brain barrier by measurement of PS, MMP-9, MMP-2, TIMP-1, claudin-5, and ZO-1 and neurologic improvement by the score on NIHSS and the mRS at 90 days. Safety outcome measures the incidence of adverse event. Statistical Analysis All results were analyzed using SPSS (SPSS version 22.0, Chicago, Illinois, USA). The data were shown as mean ± standard deviation (SD) or medians with interquartile ranges for continuous variables and proportions for categorical variables. Continuous variables were compared using t tests, and categorical variables were analyzed using the Pearson χ 2 test or Fisher’s exact test. P value <0.05 was considered statistically significant.

ard deviation (SD) or medians with interquartile ranges for continuous variables and proportions for categorical variables. Continuous variables were compared using t tests, and categorical variables were analyzed using the Pearson χ 2 test or Fisher’s exact test. P value <0.05 was considered statistically significant. Results Patient Characteristics From February 2014 to February 2016, 46 patients receiving rt-PA were included: 23 patients were treated with STS and 23 with placebo as control. Three patients had uncompleted treatment due to transferring to another department, and one patient was lost follow-up, thus leaving 21 patients in each group for clinical and prognostic analysis. General clinical characteristics (e.g., age, sex, risk factors, and NIHSS scores) and the distribution of baseline NIHSS at admission were similar (Table 1 and Fig. 2). In the 42 patients, 12 were unavailable to attend the second CT examinations, leaving 16 and 14 patients for imaging analysis in the STS and placebo groups, respectively.Table 1 Patients characteristics STS (N = 21) Placebo (N = 21) F/χ2

Results Patient Characteristics From February 2014 to February 2016, 46 patients receiving rt-PA were included: 23 patients were treated with STS and 23 with placebo as control. Three patients had uncompleted treatment due to transferring to another department, and one patient was lost follow-up, thus leaving 21 patients in each group for clinical and prognostic analysis. General clinical characteristics (e.g., age, sex, risk factors, and NIHSS scores) and the distribution of baseline NIHSS at admission were similar (Table 1 and Fig. 2). In the 42 patients, 12 were unavailable to attend the second CT examinations, leaving 16 and 14 patients for imaging analysis in the STS and placebo groups, respectively.Table 1 Patients characteristics STS (N = 21) Placebo (N = 21) F/χ2 P value Average age (years) 63.81 ± 9.872 63.62 ± 11.561 1.321 0.954 Male (%) 14 (66.7%) 14 (66.7%) 0.000 1.000 Hypertension 12 (57.1%) 17 (81.0%) 2.785 0.181 Diabetes 3 (14.3%) 5 (23.8%) 0.618 0.694 Hyperlipidemia 7(33.3%) 6 (28.6%) 0.111 1.000 Hyperhomocysteinemia 5 (23.8%) 4 (19.0%) 0.141 1.000 Atrial fibrillation 4 (19.0%) 2 (9.5%) 0.778 0.663 Coronary heart disease 3 (14.3%) 1 (4.8%) 1.105 0.606 Previous cerebral infarction 2 (9.5%) 3 (14.3%) 0.227 1.000 Smoking 5 (23.8%) 3 (14.3%) 0.618 0.238 Alcohol use 6 (28.6%) 2 (9.5%) 2.471 0.238 At admission SBP (mmHg) 153.48 ± 17.665 154.81 ± 21.558 1.896 0.828 DBP (mmHg) 86.62 ± 14.333 85.29 ± 13.050 0.527 0.754 Glu (mmol/L) 7.93 ± 2.595 8.66 ± 4.070 6.424 0.492 Time to rt-PA (min) 214.62 ± 43.246 198.71 ± 42.717 0.021 0.238 NIHSS 8.05 ± 4.522 7.38 ± 5.220 0.427 0.661 ADL 54.29 ± 25.801 53.33 ± 27.034 0.054 0.908

mission SBP (mmHg) 153.48 ± 17.665 154.81 ± 21.558 1.896 0.828 DBP (mmHg) 86.62 ± 14.333 85.29 ± 13.050 0.527 0.754 Glu (mmol/L) 7.93 ± 2.595 8.66 ± 4.070 6.424 0.492 Time to rt-PA (min) 214.62 ± 43.246 198.71 ± 42.717 0.021 0.238 NIHSS 8.05 ± 4.522 7.38 ± 5.220 0.427 0.661 ADL 54.29 ± 25.801 53.33 ± 27.034 0.054 0.908 SBP systolic blood pressure, DBP diastolic blood pressure Fig. 2 The distribution of baseline NIHSS between the two groups Clinical Outcomes A battery of neurologic function assessments was used to investigate whether patients with acute ischemic stroke could benefit from STS after rt-PA thrombolysis. Results showed that in whole 42 patients (21 STS and 21 placebo), there were more patients with a 90-day mRS score ≤1 (an excellent functional outcome) in the STS group than that in the placebo group. Though there is no statistical significance, STS did decrease the incidence of HT during hospitalization. No significant difference was found in the means of NIHSS and ADL score at 10 days after rt-PA treatment between the two groups (Table 2 and Fig. 3).Table 2 Patient clinical outcomes All enrolled patients STS (N = 21) Placebo (N = 21) F/χ2 P value 10 day-NIHSS 2.81 ± 2.64 4.10 ± 3.81 1.853 0.211 10 day-ADL 82.62 ± 20.89 72.38 ± 25.18 1.016 0.159 90 day-mRS ≤ 1 16 (76.19%) 9 (42.86%) 6.462 0.028* HT during hospitalization 2 (9.52%) 5 (23.81%) 1.543 0.410 *p<0.05 Fig. 3 Distribution of mRS in the STS and placebo groups at 3 months after treatment. STS had more patients with 90-day mRS score ≤1 in the whole 42 patients

P value 10 day-NIHSS 2.81 ± 2.64 4.10 ± 3.81 1.853 0.211 10 day-ADL 82.62 ± 20.89 72.38 ± 25.18 1.016 0.159 90 day-mRS ≤ 1 16 (76.19%) 9 (42.86%) 6.462 0.028* HT during hospitalization 2 (9.52%) 5 (23.81%) 1.543 0.410 *p<0.05 Fig. 3 Distribution of mRS in the STS and placebo groups at 3 months after treatment. STS had more patients with 90-day mRS score ≤1 in the whole 42 patients Neuroimaging of BBB Permeability between STS and Placebo Groups CTP was performed to reveal the effects of STS treatment on BBB integrity. Twenty-four hours after rt-PA treatment, a baseline CTP was performed for each patient and there was no significant difference in BBB permeability (ipsilateral or contralateral and rPS) between the two groups. The follow-up CTP was performed at 10 days after the treatments. Moreover, the STS group showed significantly lower levels of ipsilateral PS and rPS than did the placebo group at 10 days (Table 3 and Fig. 4).Table 3 BBB permeability measured by CTP-derived PS Parameters STS (N = 16) Placebo (N = 14) P value Baseline Ipsilateral PS value (ml/100 g/min) 0.373 ± 0.062 0.395 ± 0.073 0.371 Contralateral PS value (ml/100 g/min) 0.175 ± 0.025 0.179 ± 0.022 0.703 rPS 2.128 ± 0.219 2.225 ± 0.380 0.394 PS region area (cm2) 2.586 ± 2.461 2.893 ± 3.131 0.766 10 days after STS or placebo Ipsilateral PS value (ml/100 g/min) 0.266 ± 0.083 0.332 ± 0.079 0.034* Contralateral PS value (ml/100 g/min) 0.170 ± 0.170 0.172 ± 0.017 0.761 rPS 1.548 ± 0.393 1.910 ± 0.345 0.013* PS region area (cm2) 1.773 ± 1.563 2.287 ± 2.214 0.464 Infarct volumes (cm3) 2.020 ± 1.762 2.415 ± 3.083 0.676

0 days after STS or placebo Ipsilateral PS value (ml/100 g/min) 0.266 ± 0.083 0.332 ± 0.079 0.034* Contralateral PS value (ml/100 g/min) 0.170 ± 0.170 0.172 ± 0.017 0.761 rPS 1.548 ± 0.393 1.910 ± 0.345 0.013* PS region area (cm2) 1.773 ± 1.563 2.287 ± 2.214 0.464 Infarct volumes (cm3) 2.020 ± 1.762 2.415 ± 3.083 0.676 Fig. 4 BBB-PS maps from the STS and placebo groups. Quantitative PS maps of both groups at baseline (a, c) and 10 days after STS or placebo treatment (b, d). At day 10, the patient with STS treatment showed a significant decline in BBB-PS when compared to the placebo Serum BBB Damage Biomarkers Serum levels of MMP-9, MMP-2, TIMP-1 and tight-junction proteins, including claudin-5 and ZO-1, in 30 patients (16 STS and 14 placebo) were measured to demonstrate BBB damage. The STS group had lower MMP-9 (633.352 vs 750.739 ng/ml, p = 0.036, Fig. 5a) and claudin-5 (337.822 vs 407.763 pg/ml, p = 0.026, Fig. 5b) levels but higher TIMP-1 expression (520.652 vs 459.567 ng/ml, p = 0.040, Fig. 5c) than did the placebo group at 10 days after thrombolytic therapy. However, at acute phase (0 and 24 h), there were no significant differences in all biomarkers between the two groups.Fig. 5 Serum levels of BBB damage biomarkers. Serum MMP-9 (a), claudin-5 (b), and TIMP-1 (c) protein levels were measured at different time points using ELISA in the STS (n = 16) and placebo groups (n = 14). STS showed a lower MMP-9 and claudin-5 and higher TIMP-1 expression after 10 days treatment

tween the two groups.Fig. 5 Serum levels of BBB damage biomarkers. Serum MMP-9 (a), claudin-5 (b), and TIMP-1 (c) protein levels were measured at different time points using ELISA in the STS (n = 16) and placebo groups (n = 14). STS showed a lower MMP-9 and claudin-5 and higher TIMP-1 expression after 10 days treatment Discussion Current rt-PA thrombolytic therapy could augment BBB disruption in the acute stroke patients, which increase ischemic brain injury or HT [10]. Therefore, therapeutic strategies designed to alleviate BBB damage are needed to improve clinical outcomes of rt-PA thrombolysis [11]. Prognosis of rt-PA-treated patients with mRS ≤1 (an excellent functional outcome) was reported differently, from 42.7% to 54.6% at 3 months [12–14]. As Tsivgoulis reported, mRS ≤1 scores was 42.7%, which is similar to our control group (mRS ≤ 1 scores was 42.86%). Effect of rt-PA therapy is affected by many factors, such as treatment time since stroke onset, age, stroke severity, BBB integrity, and so on. Our study found that there were more patients with a 90-day mRS score ≤1 in the STS group compared with the placebo group.

lar to our control group (mRS ≤ 1 scores was 42.86%). Effect of rt-PA therapy is affected by many factors, such as treatment time since stroke onset, age, stroke severity, BBB integrity, and so on. Our study found that there were more patients with a 90-day mRS score ≤1 in the STS group compared with the placebo group. To investigate whether the neuroprotection of STS is associated with decreasing BBB disruption, PS was detected by dynamic contrast-enhanced CT [10, 15, 16]. Recently, CTP-derived PS quantification has been used to predict BBB permeability and HT for acute stroke patients because of its good reproducibility in hemodynamic measurement, wide accessibility, and relatively low cost [2, 4, 10]. The neuroimaging results from this study showed that STS treatment could reduce BBB-PS to ameliorate BBB damage and benefit clinical outcomes post-rt-PA thrombolysis in acute stroke patients. Pathologically, release of tight junction adhesion molecules to blood circulation is associated with compromised BBB integrity in ischemic stroke [3]. Tanshinone IIA (precursor of STS) treatment has been previously shown to diminish BBB breakdown in experimental model of ischemic stroke [8, 9] and autoimmune encephalomyelitis [17]. Similarly, in this study, we found that STS treatment decreased the levels of BBB damage biomarkers, MMP-9 and claudin-5, in acute ischemic stroke patients after intravenous thrombolysis. From our observations, STS seemed to reduce the effects of BBB disruption by inhibiting MMP-9 activity and increasing expression of TIMP-1.

ly, in this study, we found that STS treatment decreased the levels of BBB damage biomarkers, MMP-9 and claudin-5, in acute ischemic stroke patients after intravenous thrombolysis. From our observations, STS seemed to reduce the effects of BBB disruption by inhibiting MMP-9 activity and increasing expression of TIMP-1. Next, we evaluated the infarct volumes and found no difference between the STS group (2.020 ± 1.762 cm3) and the placebo group (2.415 ± 3.083 cm3) after 10 days of treatment (p = 0.676). It suggested that STS did not reduce BBB damage through decreasing infarct volumes. Furthermore, to study whether STS decreases cerebral hemorrhagic transformation, STS did reduce the trend of cerebral hemorrhagic transformation, but it was not significant, which may be correlated to the small sample research. STS could inhibit peripheral inflammatory cells into brain after stroke by suppressing BBB injury, which protects brain from immuno-inflammation and improves recover of stroke patients. How STS protects the BBB from injury after stroke remains unclear. STS might improve BBB damage by suppressing astrocyte-mediated inflammation or decreasing brain microvascular endothelial cell apoptosis after stroke. Together, the neuroimaging and pathological evidence of our study could delineate the mechanistic pathway of STS treatment in improving BBB dysfunction post-rt-PA thrombolysis.

How STS protects the BBB from injury after stroke remains unclear. STS might improve BBB damage by suppressing astrocyte-mediated inflammation or decreasing brain microvascular endothelial cell apoptosis after stroke. Together, the neuroimaging and pathological evidence of our study could delineate the mechanistic pathway of STS treatment in improving BBB dysfunction post-rt-PA thrombolysis. However, we had several limitations in the current work. First, to better elucidate the relation between BBB damage biomarkers and PS, MMPs and tight junction proteins obtained from cerebrospinal fluid (CSF) examination by lumbar puncture and a correlation study are needed in the future. Second, lack of BBB-PS measurements from CTP imaging at 90 days could not fully evaluate the effects of STS treatment on BBB repair. In addition, our study had a small sample size although the randomization has been applied. These questions will be addressed in our future work. In conclusion, both neuroimaging and serum biomarkers of BBB damage in this study demonstrated that acute ischemic stroke patients might benefit from STS treatment by ameliorating/diminishing BBB disruption following rt-PA thrombolysis. Biying Ji, Fei Zhou, and Lijuan Han contributed equally to this article.

However, we had several limitations in the current work. First, to better elucidate the relation between BBB damage biomarkers and PS, MMPs and tight junction proteins obtained from cerebrospinal fluid (CSF) examination by lumbar puncture and a correlation study are needed in the future. Second, lack of BBB-PS measurements from CTP imaging at 90 days could not fully evaluate the effects of STS treatment on BBB repair. In addition, our study had a small sample size although the randomization has been applied. These questions will be addressed in our future work. In conclusion, both neuroimaging and serum biomarkers of BBB damage in this study demonstrated that acute ischemic stroke patients might benefit from STS treatment by ameliorating/diminishing BBB disruption following rt-PA thrombolysis. Biying Ji, Fei Zhou, and Lijuan Han contributed equally to this article. Acknowledgements Xu Yun designed the experiments and edited the manuscript. Ji Biying, Li Shanshan, Li Jingwei, and Zhang Xin collected patients and blood samples. Ji Biying performed ELISA test. Zhou Fei, Fan Haijian, and Yang Jun scanned and analyzed the CTP. Han Lijuan and Ji Biying wrote the manuscript. Wang Xiaoying and Chen Xiangyan revised this paper. All authors contributed to the data analysis and the manuscript preparation. We thank Brad Peterson for editing the English.

i Biying performed ELISA test. Zhou Fei, Fan Haijian, and Yang Jun scanned and analyzed the CTP. Han Lijuan and Ji Biying wrote the manuscript. Wang Xiaoying and Chen Xiangyan revised this paper. All authors contributed to the data analysis and the manuscript preparation. We thank Brad Peterson for editing the English. Compliance with Ethical Standards Funding This research was supported by the National Natural Science Foundation of China (81230026, 81630028, 81601016, 81171085), the Natural Science Foundation (BE2016610, BK20160119) of Jiangsu Province, the National Key Research and Development Program of China (2016YFC1300500-504). Conflict of Interest The authors declare that they have no conflict of interest. Ethical Approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed Consent Informed consent was obtained from all individual participants included in the study. Disclosures None. None of the authors are financially tied to the patent mentioned above.

d at 72 h after ICH induction (Fig. 3d). The results showed that the ROS level was significantly increased in the ICH group compared to the sham group, while melatonin treatment remarkably reduced ROS level in brain tissues. These results suggested that melatonin can inhibit oxidative stress in brain tissues after ICH. Melatonin Impaired Inflammation in Brain Tissues at 72 h Post-ICH Induction To determine whether melatonin inhibits inflammation after ICH, western blot analysis was used to measure the protein level of MMP-9 in brain tissues, which is an indicator of inflammation. The results showed that, compared to the sham group, the level of MMP-9 was significantly increased in the ICH group, while melatonin treatment obviously decreased its level compared to the ICH + vehicle group (Fig. 4a, b).Fig. 4 Inflammatory cytokines expression levels in brain tissues at 72 h post-ICH. a Western blot analysis examines MMP-9 in the sham, ICH, ICH + vehicle, and ICH + melatonin groups. b Relative MMP-9 expression level calculation based on densitometry analysis. The mean values of MMP-9 in the sham group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6. c–f TNF-α and IL-1β levels in the CSF and serum at 72 h post-ICH. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

s mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6. c–f TNF-α and IL-1β levels in the CSF and serum at 72 h post-ICH. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 Furthermore, the levels of the pro-inflammatory cytokines (IL-1β and TNF-α) in the CSF and serum from rats in these four groups were elevated by ELISA. The results confirmed that IL-1β and TNF-α levels were significantly increased in the ICH group compared to the sham group. However, melatonin treatment significantly decreased the levels of these two cytokines compared to the ICH + vehicle groups (Fig. 4c–f).

Introduction Cerebral intraventricular hemorrhage (IVH) continues to be a serious complication of preterm birth, resulting in a high incidence of neurodevelopmental impairment, including cerebral palsy and intellectual disability [1]. During the past decades, neurological impairment following very preterm birth has primarily been considered to originate in cerebral white matter lesions [2, 3], but recent findings have also linked neurological deficits of preterm birth to cerebellar abnormalities [4, 5]. Prevalence of cerebellar injury has been described to be as high as 58% in infants with cerebral palsy following IVH and preterm birth [6]. From gestational weeks 20 to 40, the cerebellum undergoes an unparalleled growth with a volumetric increase from approximately 1 to 25 cm3 [7]. This rapid growth renders the cerebellum very sensitive to injury [8, 9]. Cerebellar underdevelopment may ensue from a direct cerebellar injury, such as hemorrhage or infarction, or from a secondary effect related to damage at a remote but connected area of the brain [10]. Cerebellar hypoplasia has repeatedly been shown to be associated with supratentorial IVH in very preterm infants and is a potential component in neurological disability [9, 11, 12]. Of note, the severity of IVH is linked to the degree of impaired cerebellar development in preterm infants, with cerebellar volume at term age being inversely correlated with increasing severity of IVH [7].

h supratentorial IVH in very preterm infants and is a potential component in neurological disability [9, 11, 12]. Of note, the severity of IVH is linked to the degree of impaired cerebellar development in preterm infants, with cerebellar volume at term age being inversely correlated with increasing severity of IVH [7]. In clinical studies, MRI at term age shows infratentorial hemosiderin deposits in 70% of preterm infants with IVH and disrupted cerebellar development. The deposits are prominent not only on the cerebellar surface but also on the surface of the brain stem and in the region of the fourth ventricle. This hemosiderin deposition is the most predictive factor for impairment in cerebellar development and thus is suggested as a plausible causal mechanism of cerebellar hypoplasia following preterm IVH [9].

only on the cerebellar surface but also on the surface of the brain stem and in the region of the fourth ventricle. This hemosiderin deposition is the most predictive factor for impairment in cerebellar development and thus is suggested as a plausible causal mechanism of cerebellar hypoplasia following preterm IVH [9]. The neurotoxicity of cell-free hemoglobin (Hb) and its metabolites has been reported after intraventricular, intraparenchymal, and subarachnoid hemorrhage (SAH) [13–20]. Cell-free Hb and its metabolites free heme, iron, reactive oxygen species (ROS), and free radicals can be highly damaging to cells, lipids, proteins, and DNA through oxidative modification, fragmentation, and cross-linking [21–23]. Cell-free Hb and its metabolites can induce cytotoxic, oxidative, and inflammatory pathways in the cerebrospinal fluid (CSF) and choroid plexus ependyma leading to tissue damage and cell death following preterm rabbit pup IVH [17–19]. Furthermore, a high accumulation of cell-free Hb in the periventricular white matter has been observed following hemorrhage in the rabbit pup IVH model [20].

flammatory pathways in the cerebrospinal fluid (CSF) and choroid plexus ependyma leading to tissue damage and cell death following preterm rabbit pup IVH [17–19]. Furthermore, a high accumulation of cell-free Hb in the periventricular white matter has been observed following hemorrhage in the rabbit pup IVH model [20]. In this study, we have completed the first investigation of the exposure of the developing cerebellum to cell-free Hb following preterm IVH and the potentially damaging effect on cerebellar development. Furthermore, we report on the protective effects of the Hb scavenger haptoglobin (Hp) following intraventricular administration. Results show that after IVH, key cell populations of the developing cerebellum are exposed to cell-free Hb, which may be central in the pathophysiological events leading to cerebellar underdevelopment.

e, we report on the protective effects of the Hb scavenger haptoglobin (Hp) following intraventricular administration. Results show that after IVH, key cell populations of the developing cerebellum are exposed to cell-free Hb, which may be central in the pathophysiological events leading to cerebellar underdevelopment. Materials and Methods Animals The study was approved by the Swedish Animal Ethics Committee in Lund. We used the well-established preterm rabbit pup model of glycerol-induced IVH as previously described [24]. The study included 59 rabbit pups from 9 litters delivered at gestational day 29 (full term corresponding to 32 days) [25, 26]. A half-breed between New Zealand White and Lop was used. The pups were delivered by caesarean section after the does were anesthetized with i.v. propofol (5 mg/kg) and with local infiltration of the abdominal wall using lidocaine with adrenaline (10 mg/ml + 5 μl/ml, 20–30 ml). After delivery, the pups were dried, weighed, and placed in an infant incubator set to a temperature of 34–35 °C and ambient humidity. At 2 h of age, the pups were hand-fed with 2 ml (100 ml/kg/day) of kitten milk formula (KMR; PetAg Inc., Hampshire, IL, USA) using a 3.5 French feeding tube and fed every 12 h increasing each meal by 1 ml. At 2 h of age, the pups were injected intraperitoneally with 50% (v/v) sterile glycerol (6.5 g/kg; Teknova, Hollister, CA, USA) to induce IVH. Ultrasound imaging of the brain was performed at 6 h of age to grade the severity of the IVH and detect SAH and daily thereafter using the VisualSonics Vevo 2100 (VisualSonics Inc., ON, Canada) with a MS-550D 40 MHz transducer. Animals with IVH at 6 h were included in the IVH group, and those without detectable IVH at all time points were used as controls (denoted as sham control). The reproducibility and accuracy of high-frequency ultrasound in this animal model have been described previously [24].

, Canada) with a MS-550D 40 MHz transducer. Animals with IVH at 6 h were included in the IVH group, and those without detectable IVH at all time points were used as controls (denoted as sham control). The reproducibility and accuracy of high-frequency ultrasound in this animal model have been described previously [24]. Intraventricular Injections After the initial ultrasound examination at 6 h of age, pups with IVH (presence of blood within distended lateral ventricles and no sign of parenchymal involvement) were randomized into one of the following three groups: IVH, IVH + Hp, or IVH + Vehicle. Pups in the IVH + Hp and IVH + Vehicle groups received an ultrasound-guided intraventricular injection at 8 h of age of either 20 μl of human Hp (50 mg/ml, Bio Products Laboratory, London, UK) or 20 μl of vehicle solution (9 mg/ml NaCl, Fresenius Kabi, Lake Zurich, IL, USA), using 27 G Hamilton syringes (Hamilton Robotics, Reno, NV, USA). The efficacy and accuracy of this method have been described previously [26]. The animals were euthanized at the following time points: 72 h (P0, corresponding to term-equivalent postnatal day 0), 120 h (P2, corresponding to term-equivalent postnatal day 2), or 192 h (P5, corresponding to term-equivalent postnatal day 5) of age. Cerebellar tissues were sampled and processed as described below. An overview of the study design is given in Fig. 1.Fig. 1 Study outline. A diagram summarizing the experimental procedure. The experiment consisted of the following steps: preterm delivery of rabbit pups by caesarean section, induction of IVH by intraperitoneal glycerol administration, verification of IVH or sham control by the use of high-frequency ultrasound, randomization into study groups, intraventricular administration of Hp or vehicle solution, termination of pups, and collection of cerebellar tissue. For details about each step, see “Materials and Methods”

aperitoneal glycerol administration, verification of IVH or sham control by the use of high-frequency ultrasound, randomization into study groups, intraventricular administration of Hp or vehicle solution, termination of pups, and collection of cerebellar tissue. For details about each step, see “Materials and Methods” Tissue Collection and Processing Following sedation with isoflurane inhalation, perfusion fixation of the brain was performed at P0 and P5 by cardiac cannulation following thoracotomy and infusion of 0.9% saline followed by 4% paraformaldehyde (PFA, buffered with phosphate buffer saline (PBS) 0.1 M, pH 7.4). After completed perfusion, the cerebrum and cerebellum were carefully extracted from the skulls and immersed in 4% PFA for a total of 48 h. SAH was confirmed in all pups with IVH with visible presence of hemorrhagic CSF covering the cerebellar cortex. None of the control pups exhibited macroscopic signs of SAH. A change to fresh PFA was performed after 3–6 h. Thereafter, the tissues were dehydrated, cleared, infiltrated with paraffin, and embedded in paraffin blocks. The cerebellum was sectioned into 4-μm sections (Leica, RM2255 Microtome) in the parasagittal plane at the level of the dentate nucleus and mounted on microscope slides and dried at 37 °C for 12–16 h. None of the cerebellar samples in pups with IVH or in control pups exhibited signs of primary cerebellar hemorrhage. Prior to antibody staining for immunohistochemistry (IHC), the sections were rehydrated, followed by heat-induced antigen retrieval at 90–95 °C for 20 min either in boric acid buffer (pH 8.0) for labeling of Ki67, calbindin, and Iba1 or in citric acid (pH 6.0, with 0.05% Tween 20 or 0.2% Triton X) for 10–20 min for immunofluorescence labeling of Hb and Hp.

istochemistry (IHC), the sections were rehydrated, followed by heat-induced antigen retrieval at 90–95 °C for 20 min either in boric acid buffer (pH 8.0) for labeling of Ki67, calbindin, and Iba1 or in citric acid (pH 6.0, with 0.05% Tween 20 or 0.2% Triton X) for 10–20 min for immunofluorescence labeling of Hb and Hp. Immunofluorescence Labeling Immunofluorescence labeling of Hb was performed to investigate the presence and distribution of both encapsulated erythrocytes and cell-free Hb within the cerebellum. Double immunofluorescence labeling of Hb together with human Hp was performed to simultaneously visualize Hb and Hp to elucidate whether the intraventricularly injected human Hp could reach the cerebellar brain regions containing Hb (preferentially the cell-free Hb) in the IVH rabbit pups.

ree Hb within the cerebellum. Double immunofluorescence labeling of Hb together with human Hp was performed to simultaneously visualize Hb and Hp to elucidate whether the intraventricularly injected human Hp could reach the cerebellar brain regions containing Hb (preferentially the cell-free Hb) in the IVH rabbit pups. In brief, the immunofluorescence labeling protocol was carried out as described below. Following antigen retrieval, sections were immersed in PBS (2 × 5 min), encircled with silicon (PAP-pen, Sakura, Tokyo, Japan), and then blocked with 1% bovine serum albumin (BSA) in PBS containing 0.05% Triton X (PBST × BSA) for 60 min at room temperature (RT). This step was followed by 16 h of incubation at 4 °C with either one of the primary antibodies or a mixture of the two primary antibodies diluted in PBST × BSA. All antibody incubations were performed in a moisture chamber. Primary antibodies used were against Hb, made in goat (diluted 1:500), and against human Hp, made in chicken (diluted 1:1000), both from GenWay Biotech, Inc. (San Diego, CA, USA) and diluted in PBST × BSA. Sections were then rinsed in PBS (3 × 3 min), followed by incubation for 60 min at RT with one secondary antibody made against goat IgG or with a mixture of secondary antibodies made against goat IgG and chicken IgY (diluted 1:200 in PBST × BSA). The secondary antibodies were both affinity-purified Fab2 fragments for multi-labeling, made in donkey (Jackson ImmunoResearch, West Grove, PA, USA). The anti-chicken IgY was conjugated with Alexa Fluor 488 (AF488) and the goat IgG conjugated with Rhodamine Red (rhodamine). Sections were then rinsed in PBS (3 × 3 min) and incubated in DAPI (0.1 μM, diluted in PBS, Invitrogen, Rockford, IL, USA) for 30 min at RT. After being rinsed in PBS (3 × 3 min), sections were mounted (Fluoroshield, Abcam, England, ab104135) and cover-slipped. All animal groups were always processed together in the same immunolabeling experiment.

PBS (3 × 3 min) and incubated in DAPI (0.1 μM, diluted in PBS, Invitrogen, Rockford, IL, USA) for 30 min at RT. After being rinsed in PBS (3 × 3 min), sections were mounted (Fluoroshield, Abcam, England, ab104135) and cover-slipped. All animal groups were always processed together in the same immunolabeling experiment. Antibody Control for Immunofluorescence Labeling Antibody specificity tests were performed on parallel sections in all labeling experiments; in these tests, the primary antibodies were excluded from the labeling protocol (Fig. 1 in the Data supplement). This control confirmed that the visualized and documented Hb and Hp immunofluorescence labeling (Fig. 2) was caused by binding of the respective primary antibodies and was not the result of binding of secondary antibodies or autofluorescence. All tested samples showed no Hb or Hp labeling within the cerebellum (see Fig. 1 in the Data supplement). Autofluorescence was solely obtained from whole cell bodies, from erythrocytes/RBCs in the subarachnoid space preferentially, and occasionally from some neuronal cell bodies. Thus, the antibody controls showed that both primary and secondary antibodies bind to their targets in the immunofluorescence labeling protocol applied here, supporting their specific detection of rabbit Hb and human Hp, visualized as extracellular (cell-free) and in whole erythrocytes.Fig. 2 Immunofluorescence labeling of Hb and the administered human Hp. Representative images are from rabbit pups at P0. Images illustrate the detected immunofluorescence labeling, performed by double immunofluorescence labeling of Hb (red) and Hp (green) together with a DAPI nuclear staining (blue), in animals with no IVH (Control), in animals with IVH (IVH), and in animals with IVH that received human Hp injections (IVH + Haptoglobin). Antibody specificity tests showed that the antibodies against Hb and human Hp bound to their true targets (see Fig. 1 in the Data supplement). a–d Control animal: Images b and c show the lack of Hb and Hp labeling and the autofluorescence mainly from whole erythrocytes (RBCs) restricted to the subarachnoid space and some blood vessels (d). e–h IVH: In pups with IVH, the Hb labeling (red) was extensive, widely distributed in the molecular layer and white matter and to some degree in the EGL.

c show the lack of Hb and Hp labeling and the autofluorescence mainly from whole erythrocytes (RBCs) restricted to the subarachnoid space and some blood vessels (d). e–h IVH: In pups with IVH, the Hb labeling (red) was extensive, widely distributed in the molecular layer and white matter and to some degree in the EGL. Whole erythrocytes in the subarachnoid space surrounding the cerebellar lobuli were also intensely labeled and gave rise to green autofluorescence (g), observed as yellowish in the merged image (h). Hb labeling intermingled with dense nuclear regions (intense DAPI staining) appears as pink (bottom images). i–l IVH + Haptoglobin: j and k show immunofluorescence labeling of Hb (red) and human Hp (green) following intraventricular injection of Hp at E29. j shows the widespread distribution of cell-free Hb (red), corresponding to that in IVH animals (f), and the domination coexistence of Hp in K (green), primarily in the molecular layer, white matter, and the EGL as shown in the merged image (l). Hp labeling was scarce in the subarachnoid space (k and l), in which Hb labeling of RBCs was extensive (j and l). Thus, the cell-free Hb and Hp are clearly distinguishable from the cell body–associated Hb labeling and autofluorescence. Scale bar = 50 μm. m HO-1 mRNA expression in the cerebellum was investigated at P0 following IVH. Following IVH, heme-degrading protein HO-1 mRNA was upregulated (IVH, dark gray bar, n = 7) as compared to the controls (n = 5). mRNA expression for HO-1 was normalized against GAPDH and is given as fold change. The fold change values were calculated by normalizing against samples from control pups. Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between no IVH and IVH at P0 were analyzed using the Mann–Whitney U test

O-1 was normalized against GAPDH and is given as fold change. The fold change values were calculated by normalizing against samples from control pups. Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between no IVH and IVH at P0 were analyzed using the Mann–Whitney U test Analyses of the Distribution of Hb and Its Relation to Hp The anatomical distribution of the double immunofluorescence-labeled Hb and Hp was analyzed using a wide-field epi-fluorescence microscope (Olympus IX73, Shinjuku, Tokyo, Japan). Analysis of the double labeling was performed by switching between the specific filter sets used for each fluorophore, DAPI for cell nuclei (blue), rhodamine for Hb (red), and AF488 for Hp (green), together with digital image documentation (Olympus DP80). The separate images for each channel (fluorophore) were merged for detailed analyses of double labeling to identify them as co-existing or not (see representative images in Fig. 2). To ensure sole detection of primary antibody binding, i.e., excluding detection of autofluorescence or of nonspecific secondary antibody binding, the detection level (threshold) for each channel was always set from sections with antibody controls (see above) and from sections from control animals that had been taken through the whole labeling protocol. Analyses and digital imaging were performed with the preset detection levels (detection intensities solely from specific labeling) for each channel. The relatively strong autofluorescence from cell bodies, mainly from RBCs, could be clearly separated from the non-cell body-associated, cell-free, and widely distributed Hb in IVH animals and together with Hp in IVH animals that received Hp (see Fig. 2).

ion intensities solely from specific labeling) for each channel. The relatively strong autofluorescence from cell bodies, mainly from RBCs, could be clearly separated from the non-cell body-associated, cell-free, and widely distributed Hb in IVH animals and together with Hp in IVH animals that received Hp (see Fig. 2). Immunohistochemistry of Cerebellar Development and Reactive Microgliosis To investigate the effect of IVH on the cerebellum of preterm rabbit pups, IHC labeling against the following antigens was performed: (1) Ki67, to evaluate cellular proliferation; (2) calbindin, to evaluate Purkinje cell development and maturation; and (3) Iba1, to evaluate microglial activation. Qualitative and quantitative analysis at P0, P2, and P5 were performed. Briefly, the protocol was as follows. After antigen retrieval and rinsing in PBS, sections were incubated with primary antibodies (diluted in PBS + 5% normal goat serum, Jackson ImmunoResearch, 005-000-121) for 1 h at RT. Primary antibodies were made against rabbit Ki67 (mouse IgG anti-Ki67, Dako, Copenhagen, Denmark), calbindin (mouse IgG anti-calbindin, DBS, Pleasanton, CA), and Iba1 (rabbit IgG anti-Iba1, Biocare, Concord, CA). Sections were then rinsed in PBS (3 × 2 min). To detect the primary antibody, sections were incubated with either BrightVision rabbit/horseradish peroxidase (HRP) or BrightVision mouse/HRP (DPVR110HRP or DPVM110HRP, both from Immunogen) for 30 min at RT. Sections were then rinsed in Tris (0.05 M, pH 7.6, 3 × 2 min). To visualize the HRP conjugations, sections were incubated with a diaminobenzidine (DAB; 50 mg DAB, Sigma, dissolved in 100 ml Tris buffer, pH 7.6, 3 × 2 min) and 100 μl of hydrogen peroxide (Merck, prepared just prior to incubation) solution was added for 5 min at RT. After rinsing in Tris (3 × 2 min), hematoxylin staining of cell nuclei (Mayers HTX, Bio-Optica) was performed for 5 s, after which the sections were dehydrated and slides were then mounted with coverslips (X-Tra-Kitt, Medite, Burgdorf, Germany). Antibody specificity tests were performed on parallel sections to confirm that the visualized immunostaining was specific for the primary antibodies. In these tests, the primary antibodies were excluded from the labeling protocol (Fig. 2 in the Data supplement). Analysis and image documentation for the results of qualitative and quantitative analysis (see below) of IHC labeling were performed with a bright-field microscope (Leica DMRX), equipped with a digital camera (Leica MC120HD).

e primary antibodies were excluded from the labeling protocol (Fig. 2 in the Data supplement). Analysis and image documentation for the results of qualitative and quantitative analysis (see below) of IHC labeling were performed with a bright-field microscope (Leica DMRX), equipped with a digital camera (Leica MC120HD). Measurement of the width (μm) of the proliferative external granular layer (EGL), as determined by Ki67-positive cells, was performed in four predefined regions. These regions were the inner and outer portions of lobule V and the inner and outer portions of lobule IX, respectively, as illustrated in Fig. 3 in the Data supplement. These regions were chosen because they represent regions with possible maturational differences in EGL proliferation and subsequent width. Measurements were performed with a bright-field microscope (Leica DMRX), using a ×40 dry objective lens. The average of the four respective measured widths was calculated for each pup. Using the Leica Q500 image analysis system of the microscope, the areas of Iba1- and calbindin-positive stained cells were respectively determined in relation to the cerebellar white matter area and the area of the molecular layer. Thus, both positive Iba1 and calbindin staining were expressed as percentage positive area in relation to, respectively, a standardized area of the cerebellar white matter and of the molecular layer. Nonspecific background staining was taken into consideration with respect to a setup threshold.

of the molecular layer. Thus, both positive Iba1 and calbindin staining were expressed as percentage positive area in relation to, respectively, a standardized area of the cerebellar white matter and of the molecular layer. Nonspecific background staining was taken into consideration with respect to a setup threshold. For mRNA analysis, the rabbit pups were euthanized with intracardiac thiopental injection at P0. The brain was dissected out of the skull and cerebellar tissue collected, snap-frozen, and stored at −80 °C until further analysis as described below.

of the molecular layer. Thus, both positive Iba1 and calbindin staining were expressed as percentage positive area in relation to, respectively, a standardized area of the cerebellar white matter and of the molecular layer. Nonspecific background staining was taken into consideration with respect to a setup threshold. For mRNA analysis, the rabbit pups were euthanized with intracardiac thiopental injection at P0. The brain was dissected out of the skull and cerebellar tissue collected, snap-frozen, and stored at −80 °C until further analysis as described below. RNA Isolation and Real-Time PCR Total RNA was extracted from the cerebellar tissue of the rabbit pups using the NucleoSpin RNA/protein extraction kit as described by the manufacturer (Macherey-Nagel, Neumann-Neander, Düren, Germany). The optical density ratio (OD at 260 nm/280 nm) of extracted RNA samples was always approximately 2.0. Reverse transcription was performed according to the manufacturer’s instructions on 1 μg total RNA using iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). The RT2 qPCR Primer Assay (primer from QIAGEN, Germantown, MD, USA) was used to quantify mRNA expression of heme oxygenase 1 (HO-1), and expression was analyzed using iTaq Universal SYBR Green Supermix (Bio-Rad). Amplification was performed as described by the manufacturer (Bio-Rad) for 40 cycles in an iCycler Thermal Cycler (Bio-Rad), and data were analyzed using iCycler iQ Optical System Software (Bio-Rad). Data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH, primer from QIAGEN), with fold change values calculated by normalizing against control animals.

manufacturer (Bio-Rad) for 40 cycles in an iCycler Thermal Cycler (Bio-Rad), and data were analyzed using iCycler iQ Optical System Software (Bio-Rad). Data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH, primer from QIAGEN), with fold change values calculated by normalizing against control animals. Statistics Statistical analysis was performed with IBM SPSS Statistics version 22. Results are presented as medians (ranges) and displayed as box plots. Comparisons between unrelated groups were performed with the Mann–Whitney U test as appropriate. Comparisons between multiple groups were made using the Kruskal–Wallis test followed by pairwise comparison with significance values adjusted for multiple comparisons. P values <0.05 were considered significant. Results Extensive Presence of Cell-Free Hb in the Cerebellum Following IVH Immunofluorescence labeling of Hb was evaluated at P0 and revealed extensive deposition of RBCs in the subarachnoid space surrounding the cerebellar lobuli following IVH, which was not observed in control animals (control and IVH in Fig. 2). Labeled Hb was widespread within the cerebellum and not associated with cell bodies (IVH in Fig. 2). Extensive deposition of radially oriented cell-free Hb was observed in the deeper cerebellar layers, in the molecular layer, and in the white matter. Relatively low amounts of cell-free Hb molecules were observed in the EGL and primarily in lobules in immediate proximity to large deposits of RBCs in the subarachnoid space.

nsive deposition of radially oriented cell-free Hb was observed in the deeper cerebellar layers, in the molecular layer, and in the white matter. Relatively low amounts of cell-free Hb molecules were observed in the EGL and primarily in lobules in immediate proximity to large deposits of RBCs in the subarachnoid space. To further investigate the indicated widespread presence of cell-free Hb in P0 IVH pups shown by immunofluorescence labeling, we performed RT-PCR analysis of mRNA expression in cerebellar tissue of the major heme-degrading protein heme-oxygenase 1 (HO-1). At P0, the HO-1 mRNA expression levels were tenfold higher in IVH pups compared to controls (Fig. 2m).

idespread presence of cell-free Hb in P0 IVH pups shown by immunofluorescence labeling, we performed RT-PCR analysis of mRNA expression in cerebellar tissue of the major heme-degrading protein heme-oxygenase 1 (HO-1). At P0, the HO-1 mRNA expression levels were tenfold higher in IVH pups compared to controls (Fig. 2m). EGL Proliferation Following IVH The total width of the EGL comprises an outer proliferative portion where the granule cell precursors (GCPs) divide and a deeper portion where the granule cells differentiate [3]. The width of the outer proliferative (Ki67-positive) portion of the EGL was measured and compared between groups (Fig. 3a, b). The median (range) widths of the proliferative EGL were 36.0 (42–26), 36.0 (42–26), and 22.0 (27–19) μm, respectively, at P0, P2, and P5 in the IVH pups and 40.0 (49–31), 36.5(42–30), and 30.0 (39–23) μm, respectively, in the control pups. The median proliferative EGL width was significantly smaller in pups with IVH compared to control pups at P5 (P = 0.017) with a clear tendency at P0 (P = 0.08) (Fig. 3b).Fig. 3 Reduced width of proliferative EGL following preterm IVH. a Images of EGL of the developing cerebellum from which quantitative measurements were made of the proliferative width in the respective groups at the time points studied. The image shows the Ki67-positive outer portion of the EGL where proliferation of granule cell precursors occurs and the deeper portion, which hosts the differentiation of granule cell precursors to mature granule cells. Scale bar = 50 μm. b GCP proliferation in the outer portion of the EGL of the developing cerebellum was investigated following IVH by Ki67 staining. Measurement of the width of proliferative EGL was done in cerebellar tissue sections of both sham controls (control, white bars; n at P0 = 6, n at P2 = 6, n at P5 = 5) and IVH pups (dark gray bars; n at P0 = 5, n at P2 = 6, n at P5 = 6) at P0, P2, and P5. Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test

at P0 = 6, n at P2 = 6, n at P5 = 5) and IVH pups (dark gray bars; n at P0 = 5, n at P2 = 6, n at P5 = 6) at P0, P2, and P5. Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test Purkinje Cell Maturation Following IVH Staining of calbindin, a calcium-binding protein, was used to evaluate Purkinje cell maturation in the molecular layer of the cerebellar cortex. IVH pups had smaller neuronal cell bodies and underdeveloped dendritic processes compared to control pups at P0, P2, and P5, respectively (Fig. 4a). Purkinje cell calbindin labeling was calculated and graded using densitometry, which showed that calbindin-labeled Purkinje cells at P0 and P2 had a significantly lower area in the IVH pups compared to controls (Fig. 4b; P0, P = 0.015; P2, P = 0.026); however, for smaller cells at P5, the differences were not statistically significant (P = 0.247). The smaller size in IVH animals indicated a reduced Purkinje cell differentiation and maturation in IVH animals.Fig. 4 Impaired Purkinje cell maturation following preterm IVH. a Immunostaining of calbindin, a calcium-binding protein, was used as a marker of Purkinje cell development in the molecular layer of the developing cerebellum. Calbindin stains are seen as brown to dark brown. Decreased calbindin immunoreactivity was observed in IVH pups (brown) compared to controls (intense dark brown). Observation of neuronal morphology revealed smaller neuronal cell bodies and underdeveloped Purkinje dendrites in IVH pups compared to controls at postnatal time points of P0, P2, and P5. ML molecular layer, PC Purkinje cell, DT dendrites, CB cell bodies; scale bar = 50 μm. b Grading of Purkinje cell development by measurement of percentage area of positive calbindin staining was done in cerebellar tissue sections of both control (white bars; n at P0 = 6, n at P2 = 6, n at P5 = 5) and IVH pups (dark gray bars; n at P0 = 6, n at P2 = 6, n at P5 = 6) at P0, P2, and P5, as described in “Materials and Methods.” Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test

IVH pups (dark gray bars; n at P0 = 6, n at P2 = 6, n at P5 = 6) at P0, P2, and P5, as described in “Materials and Methods.” Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test Microglial Response in Cerebellar White Matter Following IVH Iba1 immunoreactivity was investigated to evaluate cerebellar white matter microglial response following IVH (Fig. 5a). At P0 and P2, IVH pups compared to control pups showed a significantly higher area of Iba1 immunoreactivity, based on cells with amoeboid morphology corresponding to activated microglia (P0, P = 0.009; P2, P = 0.004; see Fig. 5b). Microglial activation was less marked at P5 (P = 0.247) in both groups and did not differ significantly between groups.Fig. 5 Microglial activation in the cerebellar white matter following preterm IVH. a Immunolabeling to confirm upregulation of Iba1 (seen as brown to dark brown) expression, a marker of microglial activation was used as a qualitative marker of reactive microglia cellular response in the white matter of the developing cerebellum. Increased Iba1 immunoreactivity was observed in IVH pups compared to controls at P0, P2, and P5. Observation of microglial morphology revealed an amoeboid shape with long processes in the IVH pups. Scale bar = 50 μm. b Measurement of percentage area of positive Iba1 staining was done in cerebellar tissue sections of both control (white bars; n at P0 = 6, n at P2 = 6, n at P5 = 5) and IVH pups (dark gray bars; n at P0 = 5, n at P2 = 6, n at P5 = 6) at P0, P2, and P5, as described in “Materials and Methods.” Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test

IVH pups (dark gray bars; n at P0 = 5, n at P2 = 6, n at P5 = 6) at P0, P2, and P5, as described in “Materials and Methods.” Results are presented as box plots displaying medians and 25th and 75th percentiles. Statistical differences between groups for respective time points were analyzed using the Mann–Whitney U test Hp Distribution Following Intraventricular Administration At P0, the presence of Hp and its distributional relation to cell-free Hb was investigated in all groups by means of double immunofluorescence labeling (Fig. 2). Hp labeling was detected only in IVH pups that received intraventricular (human) Hp at 8 h of age. No Hp labeling was detected in control pups or in pups with IVH receiving only vehicle. In IVH pups receiving human Hp, the Hp immunolabeling was widely distributed throughout large parts of the cerebellum. Double immunofluorescence labeling of Hp and Hb in these pups displayed a high degree of co-existence of human Hp and Hb in most regions, including the molecular layer and white matter (Fig. 2, IVH + Hp). Similar to labeling of cell-free Hb, labeling of Hp was relatively low in the EGL.

t large parts of the cerebellum. Double immunofluorescence labeling of Hp and Hb in these pups displayed a high degree of co-existence of human Hp and Hb in most regions, including the molecular layer and white matter (Fig. 2, IVH + Hp). Similar to labeling of cell-free Hb, labeling of Hp was relatively low in the EGL. Reduced Cerebellar Damage Following Hp Administration The group of pups receiving intraventricular administration of Hp following IVH (IVH + Hp), displayed an improved Purkinje cell maturation at P0 compared to both IVH + Vehicle pups and IVH pups (Fig. 6a–d). These findings included both a higher intensity of calbindin immunoreactivity and relatively larger neuronal cell bodies with more developed dendritic processes (Fig. 6a–d). Results from quantification of Purkinje cell development by calbindin staining densitometry showed an increased staining in the IVH pups following intraventricular Hp administration (Fig. 6e; Control, IVH + Hp, P = 1.00; Control, IVH + Vehicle, P = 0.024).Fig. 6 Intraventricular Hp administration protects against impaired Purkinje cell development following preterm IVH. a–d Following intraventricular Hp administration at P0, a higher intensity of calbindin immunoreactivity, relatively larger Purkinje cell bodies, and developed dendrites were observed in the Hp-administered IVH pups as compared to pups with IVH only or vehicle-treated IVH pups. Scale bar = 50 μm. e Grading of Purkinje cell development by measurement of percentage area of positive calbindin staining was done in cerebellar tissue sections at P0 of control pups (white bars, n = 6), IVH pups (dark gray bars, n = 6), and following intraventricular injection of Hp in pups with IVH (IVH + Hp, gray bars, n = 6) or vehicle solution (IVH + Vehicle, light gray bars, n = 4). Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between IVH + Hp vs. control and IVH + Vehicle vs. control were analyzed using the Kruskal–Wallis test followed by pairwise comparison with significance values adjusted for multiple comparisons

cle, light gray bars, n = 4). Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between IVH + Hp vs. control and IVH + Vehicle vs. control were analyzed using the Kruskal–Wallis test followed by pairwise comparison with significance values adjusted for multiple comparisons Furthermore, Hp administration restored the arrested cell proliferative activity in the outer portion of the EGL at P0 following IVH, as shown by the width of the proliferative part of the EGL in the respective treatment groups (described in Fig. 7a–d). The median (range) widths of the proliferative EGL were 39 (48–32) μm in the IVH + Hp pups, 30.5 (36–26) μm in the IVH + Vehicle pups, 36.0 (42–26) μm in the IVH pups, and 40.0 (49–31) μm in the control pups (Fig. 7e; Control, IVH + Hp, P = 0.93; Control, IVH + Vehicle, P = 0.038).Fig. 7 Intraventricular Hp administration protects against reduction in width of proliferative EGL following preterm IVH. a–d Following intraventricular Hp administration at P0, a higher intensity of Ki67 immunoreactivity was observed in the Hp-administered IVH pups as compared to pups with IVH only or vehicle-treated IVH pups. Scale bar = 20 μm. e Measurement of the width of Ki67-positive proliferative EGL was performed in cerebellar tissue sections at P0 of control pups (white bars, n = 6), IVH pups (dark gray bars, n = 5), and following intraventricular injection of Hp in pups with IVH (IVH + Hp, gray bars, n = 5) or vehicle solution (IVH + Vehicle, light gray bars, n = 4). Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between IVH + Hp vs. control and IVH + Vehicle vs. control were analyzed using the Kruskal–Wallis test followed by pairwise comparison with significance values adjusted for multiple comparisons

cle, light gray bars, n = 4). Results are presented as box plots displaying medians and 25th and 75th percentiles. Differences between IVH + Hp vs. control and IVH + Vehicle vs. control were analyzed using the Kruskal–Wallis test followed by pairwise comparison with significance values adjusted for multiple comparisons Discussion In this study, we show that IVH in the preterm rabbit pup is followed by an extensive deposition of blood products, specifically cell-free Hb, in the cerebellar cortex and white matter. This event is accompanied by a decrease in neuronal cell proliferation and a delay in Purkinje cell maturation. Intraventricular administration of the cell-free Hb scavenger Hp resulted in a high co-existence of administered Hp with cell-free Hb within the cerebellum. Furthermore, administered Hp partially reversed the cerebellar damage, indicating that cell-free Hb and its metabolites are causal in cerebellar underdevelopment. To the best of our knowledge, this work is the first animal study to evaluate cerebellar exposure to blood products and their role in cerebellar impairment following preterm IVH.

ered Hp partially reversed the cerebellar damage, indicating that cell-free Hb and its metabolites are causal in cerebellar underdevelopment. To the best of our knowledge, this work is the first animal study to evaluate cerebellar exposure to blood products and their role in cerebellar impairment following preterm IVH. Following preterm IVH, there is a deposition of extravasated blood into the CSF of the intraventricular space. This deposition is followed by hemolysis of RBCs, leading to a release of cell-free Hb. Physiologically, cerebral CSF produced by the choroid plexus of the ventricular system passes through the fourth ventricle and enters the subarachnoid space, resulting in an immediate interface with the cortex of the developing cerebellum [27, 28]. Consequently, there is a strong physiological support for CSF containing extravasated blood reaching the cerebellum following cerebral IVH, as evidenced in this study by the visible presence of hemorrhagic CSF surrounding cerebellar tissue at termination of pups with IVH.

e cortex of the developing cerebellum [27, 28]. Consequently, there is a strong physiological support for CSF containing extravasated blood reaching the cerebellum following cerebral IVH, as evidenced in this study by the visible presence of hemorrhagic CSF surrounding cerebellar tissue at termination of pups with IVH. In the rabbit pup model, the spontaneous vessel rupture and the subsequent sequence of events leading to IVH mimics the situation in the human preterm infant quite well. It has been suggested that many of the effects observed in this model are related to the administered glycerol, including decreased proliferation leading to cerebellar hypoplasia [29]. Of importance in this study, as well as in our previous work, all pups including controls received the same dose of intraperitoneal glycerol, which rules out the possibility that the present findings in IVH pups are related to the administered glycerol.

ing decreased proliferation leading to cerebellar hypoplasia [29]. Of importance in this study, as well as in our previous work, all pups including controls received the same dose of intraperitoneal glycerol, which rules out the possibility that the present findings in IVH pups are related to the administered glycerol. Using Hb immunofluorescence and as demonstrated by autofluorescence, we identified an extensive deposition of RBCs and cell-free Hb in the subarachnoid space enveloping the cerebellar lobules following IVH. Cell-free Hb reached the innermost layers of the cerebellar cortex at P0 and was extensively deposited in the molecular layer and white matter of the cerebellum but to a much lesser extent in the EGL. Cell-free Hb within the EGL was basically found only in cerebellar lobules in immediate proximity to large deposits of RBCs in the subarachnoid space, possibly serving as a source of the cell-free Hb. In conjunction with the radial orientation of the Hb molecules, the high amount of cell-free Hb in the molecular layer and white matter suggests additional sources beyond the CSF in the subarachnoid space. Speculatively, the source of cell-free Hb could be via the roof of the fourth ventricle and transfer through the cerebellar peduncles to the white matter of the cerebellum.

ules, the high amount of cell-free Hb in the molecular layer and white matter suggests additional sources beyond the CSF in the subarachnoid space. Speculatively, the source of cell-free Hb could be via the roof of the fourth ventricle and transfer through the cerebellar peduncles to the white matter of the cerebellum. Cell-free Hb and its metabolites, e.g., heme and iron, are well described to act as sources of ROS and free radicals, which are causal initiators of oxidative damage to cells and tissues [30]. We have previously shown that cell-free Hb and its metabolites, i.e., methemoglobin and heme, are potent inducers of pro-inflammatory pathways in choroid plexus epithelium and in astrocytes [17–19]. The extensive presence of cell-free Hb in the cerebellar white matter following IVH in this study was accompanied by clear signs of microglial activation in corresponding white matter regions, marked by increased expression of Iba1 antigen and an activated morphology in the IVH group (Fig. 5). This result suggests that deposited cell-free Hb may induce a microglial pro-inflammatory response with possible adverse effects on immature oligodendrocyte proliferation and maturation and subsequent cerebellar white matter damage. In the current study, using immunofluorescence and immunohistochemistry, we could not distinguish between different forms of oxidized Hb, e.g., oxyHb and metHb, and thus cannot conclude whether the effects observed are caused by oxyHb, metHb, or the downstream metabolites heme and iron.

cerebellar white matter damage. In the current study, using immunofluorescence and immunohistochemistry, we could not distinguish between different forms of oxidized Hb, e.g., oxyHb and metHb, and thus cannot conclude whether the effects observed are caused by oxyHb, metHb, or the downstream metabolites heme and iron. Our finding that cell-free Hb is extensively deposited in the molecular layer of the cerebellum is a cause for concern because this layer constitutes the environment for Purkinje cell maturation. The Purkinje cells are intrinsically sensitive to oxidative stress and essential for establishing the cerebellar circuitry, which is vital for impulse transmission in the cerebellum [31–34]. In addition, mature Purkinje cells also play a vital role in the development of the EGL by sourcing GCPs with sonic hedgehog protein, an important mitotic growth factor vital to their proliferation [35, 36]. Consequently, exposure of the molecular layer to cell-free Hb not only will have neurotoxic effects on Purkinje cells but also will further impair the development of the EGL. The EGL of the developing cerebellum serves as a germinal center where GCPs proliferate and subsequently differentiate into mature granule cells. Granule cells are important for the structural integrity of the cerebellum; in addition, during their migration to form the granular layer, they transmit certain excitatory signals needed for the differentiation and maturation of the Purkinje cells. Thus, exposure of the developing cerebellum to cell-free Hb may lead to damaging effects not only to the cellular architecture but also to the functional integrity of the cerebellum, subsequently causing cerebellar underdevelopment.

excitatory signals needed for the differentiation and maturation of the Purkinje cells. Thus, exposure of the developing cerebellum to cell-free Hb may lead to damaging effects not only to the cellular architecture but also to the functional integrity of the cerebellum, subsequently causing cerebellar underdevelopment. To evaluate the possible effects of impaired Purkinje cell support and direct exposure to the hemorrhage, we performed metric analysis of Ki67 staining to evaluate EGL cell proliferation and thus pathological cellular senescence. Cellular senescence in this context can be seen as a process by which damage to tissue causes a decrease in metabolism leading to arrest of cell proliferation and recruitment of phagocytic immune cells to help in tissue renewal [37]. Measurements of the EGL (Fig. 3a) showed that IVH caused a significant decrease in the width of the proliferative portion of the EGL at P0 and P5 (Fig. 3b). This result is a clear indication that IVH-related processes cause impairment of the proliferative activity of the EGL. The postnatal time points P0 to P5 studied in the preterm rabbit pup correspond to the gestational ages of 25 to 35 weeks in humans, a period characterized by intense cell proliferation in the outer portion of the EGL [38]. In the human preterm infant, the width of the proliferative EGL decreases from 30 gestational weeks onwards as the GCPs mature into granule cells and leave the EGL to form the internal granular layer [3]. This timing corresponds well to our observations in the rabbit pup with EGL proliferative width in control pups showing a decrease in width from P0 to P5 (Fig. 3b).

liferative EGL decreases from 30 gestational weeks onwards as the GCPs mature into granule cells and leave the EGL to form the internal granular layer [3]. This timing corresponds well to our observations in the rabbit pup with EGL proliferative width in control pups showing a decrease in width from P0 to P5 (Fig. 3b). Cell-free Hb may cause damage to the cerebellum in a number of different ways. Following hemolysis, release of excess cell-free Hb may lead to the formation of heme and free iron, increasing the concentration of redox-active iron in the extracellular environment. Both heme and free iron have a pro-oxidative damaging effect on cells, and iron overload has been reported to cause cerebral damage following IVH [39, 40]. Indeed, reduction in iron overload attenuated development of hydrocephalus and brain damage in a rodent model of neonatal germinal matrix hemorrhage [41]. In addition to its redox-related effects, cell-free Hb also acts as a redox-active damage-associated molecular pattern (known as DAMP) molecule that perturbs the innate immune homeostasis by triggering Toll-like receptor signal transduction pathways and causing pro-inflammatory damage to cells [42–44]. In this study, we investigated the causal importance of cell-free Hb in the impairment of Purkinje cell maturation and in the arrest of EGL cell proliferation by administering the Hb-scavenging protein Hp intraventricularly following detection of IVH. Hp binds to cell-free Hb, forming an inert Hb–Hp complex, which then channels the Hb molecules for intracellular degradation via CD163-mediated endocytosis [45, 46]. Intracellularly, the enzyme HO-1 breaks down heme to bilirubin and CO, both of which have antioxidant and vasodilatory benefits [47]. By forming a tight complex with cell-free Hb, Hp stabilizes and shields heme iron within the hydrophobic pocket of Hb, thereby preventing its cytotoxic and pro-oxidative effect [48]. The removal of cell-free Hb from the extracellular environment through its complex formation with Hp could thus reduce interaction with signal-transducing receptors of cells in the brain innate immune system and reduce exposure to excess iron and to heme-induced toxicity.

ng its cytotoxic and pro-oxidative effect [48]. The removal of cell-free Hb from the extracellular environment through its complex formation with Hp could thus reduce interaction with signal-transducing receptors of cells in the brain innate immune system and reduce exposure to excess iron and to heme-induced toxicity. A neuroprotective role of induced endogenous Hp following intracerebral hemorrhage has been documented [49]. The induction of Hp was necessary because of very low levels of endogenous Hp in the human brain. In a previous study, the resting state capacity of the intrathecal Hb–Hp complex clearance was found to be 50,000-fold lower than that in the circulation in the adult. The system was quickly saturated during SAH with a residual inability to deal with cell-free Hb, clearly indicating an insufficient Hb scavenging capacity within the brain [50]. In view of this, we administered human Hp intraventricularly, which resulted in an extensive presence of Hp in the cerebellum. Hp was not detected in animals that did not receive exogenous Hp. The Hp labeling was specific for the administered human Hp, i.e., completely absent in sham-injected IVH pups as in IVH and control pups, thus excluding endogenous Hp as a source of the positive Hp labeling.

an extensive presence of Hp in the cerebellum. Hp was not detected in animals that did not receive exogenous Hp. The Hp labeling was specific for the administered human Hp, i.e., completely absent in sham-injected IVH pups as in IVH and control pups, thus excluding endogenous Hp as a source of the positive Hp labeling. Our double immunofluorescence of Hb and Hp showed that the injected Hp reaches the same cerebellar areas as cell-free Hb and that the two are extensively co-localized in these regions. Hb and administered Hp co-existed in several regions of the cerebellum, mainly within the molecular layer and white matter and to a lesser degree in the EGL. Congruent with the anatomical co-existence of Hp and Hb, results showed that Hp administration partially reduced the Purkinje cell maturational arrest caused by IVH, represented by calbindin immunoreactivity showing a higher intensity of labeling, relatively larger cell bodies, and more extensive dendritic processes in pups receiving Hp as compared to the other IVH groups. Furthermore, Hp administration counteracted the decreased development of the proliferative region of the EGL following IVH and increased the proliferative width almost to the level of the control pups.

y larger cell bodies, and more extensive dendritic processes in pups receiving Hp as compared to the other IVH groups. Furthermore, Hp administration counteracted the decreased development of the proliferative region of the EGL following IVH and increased the proliferative width almost to the level of the control pups. Conclusion In this study, we showed that IVH in the preterm rabbit pup is followed by an extensive deposition of cell-free Hb in cerebellar cell layers and white matter. This exposure to cell-free Hb was associated with microglial activation, an arrest in neuronal cell proliferation, and a delayed Purkinje cell maturation. Intraventricular administration of the cell-free Hb scavenger Hp partially blocked these effects, suggesting that cell-free Hb and its downstream metabolites are causal in cerebellar impairment following IVH. In terms of future clinical application, these results suggest that removal or scavenging of Hb metabolites following IVH, for instance by administered Hp, may reduce subsequent cerebellar impairment. BSA, bovine serum albumin; EGL, external granular layer; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GCP, granular cell precursor; Hb, hemoglobin; Hp, haptoglobin; IVH, intraventricular hemorrhage; PBS, phosphate buffer saline; PFA, paraformaldehyde; SAH, subarachnoid hemorrhage Electronic supplementary material

Conclusion In this study, we showed that IVH in the preterm rabbit pup is followed by an extensive deposition of cell-free Hb in cerebellar cell layers and white matter. This exposure to cell-free Hb was associated with microglial activation, an arrest in neuronal cell proliferation, and a delayed Purkinje cell maturation. Intraventricular administration of the cell-free Hb scavenger Hp partially blocked these effects, suggesting that cell-free Hb and its downstream metabolites are causal in cerebellar impairment following IVH. In terms of future clinical application, these results suggest that removal or scavenging of Hb metabolites following IVH, for instance by administered Hp, may reduce subsequent cerebellar impairment. BSA, bovine serum albumin; EGL, external granular layer; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GCP, granular cell precursor; Hb, hemoglobin; Hp, haptoglobin; IVH, intraventricular hemorrhage; PBS, phosphate buffer saline; PFA, paraformaldehyde; SAH, subarachnoid hemorrhage Electronic supplementary material Supplementary Fig. 1 Antibody specificity of the immunofluorescence labeling of Hb and Hp. Antibody specificity tests on cerebellar sections from rabbit pups showed that the immunofluorescence labeling is the result of specific binding of the endogenous rabbit Hb and administered human Hp (i.e., not endogenous) to their corresponding epitopes (see Fig. 2). This inference is further supported by the lack of Hb labeling in control animals and of Hp labeling in control animals as well as in IVH animals that did not receive injections of human Hp (i.e., no labeling of endogenous Hp). These tests also showed that the endogenous tissue fluorescence could be concluded to arise only from cell bodies, preferentially from whole erythrocytes located in the arachnoid space (arrows in B and D), which was even more pronounced in IVH animals (see Fig. 2). Thus, the detected extracellular Hb and Hp can be considered to represent a specific detection and visualization of their distribution in the cerebellum. The antibody control sections were processed for double immunofluorescence labeling (see also Fig. 2) with the only difference that the primary antibody incubation was excluded from the protocol (i.e., no anti-Hb or anti-human Hp antibodies). Antibody specificity control sections were used in every labeling experiment to eliminate the risk for false interpretation of fluorescence caused by nonspecific secondary antibody binding or endogenous cell/tissue autofluorescence. These sections were also used during the analyses to ensure that the “threshold” for the fluorescence detection level visualized only immunofluorescence from secondary antibodies bound to anti-Hb and anti-human Hp antibodies, i.e., not background levels of fluorescence from nonspecific binding of secondary antibodies. The images show a representative section from an animal with IVH that received Hp injection. Nuclear counterstaining was performed with DAPI (blue in A and D).

econdary antibodies bound to anti-Hb and anti-human Hp antibodies, i.e., not background levels of fluorescence from nonspecific binding of secondary antibodies. The images show a representative section from an animal with IVH that received Hp injection. Nuclear counterstaining was performed with DAPI (blue in A and D). D is a merged image from all visualized channels (A–C) of DAPI together with the fluorophore visualization used for Hb (red) and Hp (green) immunofluorescence of the used secondary antibodies targeting the anti-rabbit Hb and injected anti-human Hp (B and C). B and C show the lack of binding by the used secondary antibodies to cell bodies or other extracellular targets comparable to their presented targeting of our anti-rabbit Hb and anti-human Hp antibodies. Thus, the immunofluorescence labeling presented is most likely due to secondary antibody binding to the primary antibodies, which bind to the rabbit Hb and human Hp epitopes, respectively. 20 μm (GIF 158 kb) High resolution image (TIFF 24909 kb)

D is a merged image from all visualized channels (A–C) of DAPI together with the fluorophore visualization used for Hb (red) and Hp (green) immunofluorescence of the used secondary antibodies targeting the anti-rabbit Hb and injected anti-human Hp (B and C). B and C show the lack of binding by the used secondary antibodies to cell bodies or other extracellular targets comparable to their presented targeting of our anti-rabbit Hb and anti-human Hp antibodies. Thus, the immunofluorescence labeling presented is most likely due to secondary antibody binding to the primary antibodies, which bind to the rabbit Hb and human Hp epitopes, respectively. 20 μm (GIF 158 kb) High resolution image (TIFF 24909 kb) Supplementary Fig. 2 Antibody specificity of the immunohistochemical labeling of calbindin, Ki67, and Iba1. No immunolabeling or background staining was observed in sections when primary antibodies were omitted from the immunohistochemical labeling protocol (A–D). Images illustrate the staining with only anti-mouse secondary antibodies conjugated with BrightVision-HRP, used for calbindin and Ki67 labelings, in a P5 control animal (A) and in a rabbit pup with IVH (B). C and D illustrate staining achieved when using the anti-rabbit secondary antibodies conjugated with BrightVision-HRP for Iba1 labelings, in a P5 control animal (C) and a P5 rabbit pup with IVH (D). Scale bar = 50 μm (GIF 447 kb) High resolution image (TIFF 24909 kb)

Supplementary Fig. 2 Antibody specificity of the immunohistochemical labeling of calbindin, Ki67, and Iba1. No immunolabeling or background staining was observed in sections when primary antibodies were omitted from the immunohistochemical labeling protocol (A–D). Images illustrate the staining with only anti-mouse secondary antibodies conjugated with BrightVision-HRP, used for calbindin and Ki67 labelings, in a P5 control animal (A) and in a rabbit pup with IVH (B). C and D illustrate staining achieved when using the anti-rabbit secondary antibodies conjugated with BrightVision-HRP for Iba1 labelings, in a P5 control animal (C) and a P5 rabbit pup with IVH (D). Scale bar = 50 μm (GIF 447 kb) High resolution image (TIFF 24909 kb) Supplementary Fig. 3 An overview of the cerebellar lobuli. The image is a pictorial representation of the cerebellar lobules used for the EGL analysis. It shows the four predefined regions from which the metric analysis of the width of the proliferative EGL was done. These regions were the inner (designated as in) and outer portions (designated as out) of lobule V EGL germinal region and the inner (designated as in) and outer portions (designated as out) of lobule IX EGL germinal region. Scale bar = 50 μm (GIF 367 kb) High resolution image (TIFF 24909 kb) Electronic supplementary material The online version of this article (doi:10.1007/s12975-017-0539-1) contains supplementary material, which is available to authorized users.

Supplementary Fig. 3 An overview of the cerebellar lobuli. The image is a pictorial representation of the cerebellar lobules used for the EGL analysis. It shows the four predefined regions from which the metric analysis of the width of the proliferative EGL was done. These regions were the inner (designated as in) and outer portions (designated as out) of lobule V EGL germinal region and the inner (designated as in) and outer portions (designated as out) of lobule IX EGL germinal region. Scale bar = 50 μm (GIF 367 kb) High resolution image (TIFF 24909 kb) Electronic supplementary material The online version of this article (doi:10.1007/s12975-017-0539-1) contains supplementary material, which is available to authorized users. Acknowledgements This work was supported by the Swedish Research Council, governmental ALF research grants to Lund University and Lund University Hospital, the European Commission (FP7, Project 305485 PREVENT-ROP), the Crafoordska Foundation, the Greta and Johan Kock Foundation, the Alfred Österlund Foundation, the Erasmus+ programme of the European Union (Framework Agreement Number: 2013-0040), and the Fanny Ekdahls Foundation. The authors wish to acknowledge Carin Sjölund, Lund University, for excellent technical assistance. Compliance with Ethical Standards Conflict of Interest All authors declare that they have no conflict of interest. Ethical Approval All applicable national and institutional guidelines for the care and use of animals were followed.

Introduction Intracerebral hemorrhage (ICH) is a common acute nervous system disease with high mortality and disability, accounting for ~15% of all patients with stroke. Currently, an effective treatment modality for ICH is not available, with some patients resorting to hematoma evacuation, which is not satisfactory [1]. Seventy-five percent of patients that survive after sustaining ICH have varying degrees of motor, sensory, language, and other advanced neural function defects. While significant progress has been made to investigate the mechanisms of brain injury after ICH, an effective clinical treatment which can significantly improve ICH prognosis is still unavailable [1–5]. Previous studies indicated that ICH-induced brain injury is not only due to the hematoma mass effect and the potential hematoma expansion (they are the main causes of primary brain injury) but also due to secondary brain injury (SBI) [6]. Therefore, exploring how to reduce SBI and promote neural function recovery became the primary focus of researchers. The mechanisms contributing to SBI are very complex and mainly related to the following aspects: oxidative stress, neuronal death (including apoptosis and necrosis), inflammation, reactive oxygen species (ROS) generation, mitochondrial dysfunction, and so on. These mechanisms resulted in multiple pathological events in SBI, such as blood-brain barrier (BBB) integrity damage, brain edema, and brain injury. As many researches reported that these mechanisms are related to each other, but they have not been fully illustrated.

S) generation, mitochondrial dysfunction, and so on. These mechanisms resulted in multiple pathological events in SBI, such as blood-brain barrier (BBB) integrity damage, brain edema, and brain injury. As many researches reported that these mechanisms are related to each other, but they have not been fully illustrated. Melatonin (N-acetyl-5-methoxytryptamine) is a type of indolamine derived from tryptophan secreted mainly by the pineal gland, which has high biological availability and easily crosses the BBB to enter the brain parenchyma [7]. Recent studies showed that melatonin has high antioxidant properties: scavenging ROS; protecting mitochondrial oxidoreductase, superoxide dismutase (SOD), and other important proteins and enzymes which can alleviate DNA oxidative damage [8]; and reducing inflammation [9–14]. Additionally, previous research investigated the mechanisms of the protective effects of melatonin in brain injury caused by middle cerebral artery occlusion (MCAO) in a rat model of cerebral ischemia/reperfusion (I/R) [15]. Our previous study also showed that melatonin has beneficial effects against early brain injury after subarachnoid hemorrhage (SAH) in a rat model [16]; however, there are few reports that focus on the effects of melatonin in ICH. Therefore, the aim of this study was to explore the effects and mechanisms of melatonin on ICH-induced SBI.

also showed that melatonin has beneficial effects against early brain injury after subarachnoid hemorrhage (SAH) in a rat model [16]; however, there are few reports that focus on the effects of melatonin in ICH. Therefore, the aim of this study was to explore the effects and mechanisms of melatonin on ICH-induced SBI. Methods Experimental Design, ICH Procedure, and Treatment Strategy Adult male SD rats weighing 320–360 g (Animal Center of the Chinese Academy of Sciences, Shanghai, China) were randomly and equally divided into the following four groups: sham group, ICH group, ICH + vehicle group, and ICH + melatonin group (n = 12 in each group), using the table of random numbers by a technician who did not take part in this research. At 72 h after ICH, as before, all rats were examined for behavioral impairment, and then immediately, all rats were sacrificed, and blood and cerebrospinal fluid from each rat were collected. Six rats per group were sacrificed for western blot analysis, immunofluorescence analysis, TUNEL and Nissl staining, and ROS tests, and another six for brain edema in each group. For behavioral impairment and brain edema detection, the observers did not know the component of infusion or the group of rats. For western blot analysis, BBB permeability and ROS assay show quantitative results; each n represents data collected from one independent experiment using one rat; combined data from at least one independent experiment using six different rats are shown. For all the immunofluorescence analysis and TUNEL and Nissl staining, representative images from at least three independent experiments using six different rats are shown. The n is always defined as number of rats in every figure legend (Fig. 1).Fig. 1 Experimental designs. a Experiment 1 was designed to investigate the effect of melatonin treatment on ICH-induced secondary brain injury (SBI) in vivo. b Experiment 2 was designed to investigate the effect of melatonin treatment on ICH-induced SBI in vitro. c Coronal sections of brain tissues after 72 h post-ICH induction

erimental designs. a Experiment 1 was designed to investigate the effect of melatonin treatment on ICH-induced secondary brain injury (SBI) in vivo. b Experiment 2 was designed to investigate the effect of melatonin treatment on ICH-induced SBI in vitro. c Coronal sections of brain tissues after 72 h post-ICH induction The autologous whole blood model was utilized as previously described [17] due to its ability to closely simulate clinical ICH. SD rats were anesthetized intraperitoneally with 4% chloral hydrate, with additional chloral hydrate administered if needed based on tail pinch response. Once fully anesthetized, the specimens were set in a stereotactic apparatus frame (Shanghai Ruanlong Science and Technology Development Co., Ltd., China). Autologous whole blood (80 μl) was drawn by cardiac puncture and injected slowly (5 min) unilaterally with a microliter syringe into the right basal ganglia. The position of the basal ganglia was 3.5 mm lateral to the midline, 1.5 mm posterior to the bregma, and 5.5 mm ventral to the cortical surface. To prevent reflux, the needle stayed in place for an additional 5 min, and the scalp was then sutured. During the entire surgery, the rat was placed supine on a heating blanket to maintain body temperature between approximately 37 ± 0.5 °C [18, 19].

m posterior to the bregma, and 5.5 mm ventral to the cortical surface. To prevent reflux, the needle stayed in place for an additional 5 min, and the scalp was then sutured. During the entire surgery, the rat was placed supine on a heating blanket to maintain body temperature between approximately 37 ± 0.5 °C [18, 19]. As described in previous studies [20], melatonin (Sigma, USA) was dissolved in absolute ethyl alcohol and diluted with 0.9% normal saline. A dose of melatonin (5 mg/kg), which was determined based on animal body weight, was injected intraperitoneally at 1, 24, and 48 h after ICH induction, with the animals sacrificed at 72 h after ICH. Vehicle-treated animals received an equal volume of the vehicle, which was also injected intraperitoneally.

ormal saline. A dose of melatonin (5 mg/kg), which was determined based on animal body weight, was injected intraperitoneally at 1, 24, and 48 h after ICH induction, with the animals sacrificed at 72 h after ICH. Vehicle-treated animals received an equal volume of the vehicle, which was also injected intraperitoneally. Isolation and Treatment of Primary Cortical Neuron The procedure and protocol for primary cortical neuron isolation has been depicted in our previous report [21]. Briefly, cortical tissues were isolated from fetal SD rat brains at 18 days of gestation and treated with papain (100 mg/ml; Worthington, USA) for 10 min at 37 °C. Dissociated neurons were plated at a density of 20,000 cells/cm2 onto plates (Corning, USA) precoated with 0.1 mg/ml poly-d-lysine (Sigma, USA), cultured in Neurobasal medium supplemented with 2% B-27 and 0.5 mM GlutaMAX TM-I (all from GIBCO, USA), and maintained at 37 °C under humidified conditions and 5% CO2. Cells were maintained for 14–19 days, with half of the media exchanged for fresh media every 3 days. The neurons were then divided into four groups for the Annexin V and PI staining and mitochondrial membrane permeability transition pore (MPTP) assay as follows: control group; OxyHb (30 μM) treatment for 1 h; OxyHb + vehicle; and pretreatment with melatonin (60 μM) for 1 h, thorough rinsing, and OxyHb treatment for 1 h.

e neurons were then divided into four groups for the Annexin V and PI staining and mitochondrial membrane permeability transition pore (MPTP) assay as follows: control group; OxyHb (30 μM) treatment for 1 h; OxyHb + vehicle; and pretreatment with melatonin (60 μM) for 1 h, thorough rinsing, and OxyHb treatment for 1 h. ELISA At 72 h post-ICH, blood and cerebrospinal fluid (CSF) were collected prior to sacrifice by puncturing the heart and foramen magnum. The CSF of all rats were immediately centrifuged for 30 min (4 °C, 12,000g), and the blood samples of all rats were centrifuged for 5 min (4 °C, 1000g); then, their supernatants were collected to measure TNF-α and IL-1β levels using specific ELISA kit (Bio-Swamp, China) according to the manufacturers’ instructions.

um. The CSF of all rats were immediately centrifuged for 30 min (4 °C, 12,000g), and the blood samples of all rats were centrifuged for 5 min (4 °C, 1000g); then, their supernatants were collected to measure TNF-α and IL-1β levels using specific ELISA kit (Bio-Swamp, China) according to the manufacturers’ instructions. ROS Assay The levels of ROS in brain tissues were detected by the Reactive Oxygen Species Assay Kit (Beyotime, China). The collected tissues were first homogenized and centrifuged at 12,000g for 10 min at 4 °C. The supernatants were used for the ROS assay. ROS concentrations were evaluated using the oxidant-sensitive probe 2,7-dichlorofluorescein diacetate (DCF-DA) according to the manufacturer’s instructions. A fluorometric microplate reader (FilterMax F5, Molecular Devices, Sunnyvale, USA) with excitation and emission at 485 and 530 nm, respectively, was used to measure the fluorescence intensity, and the samples of all rats were detected in at least one dependent experiment. The concentrations of ROS were expressed as fluorescence intensity per milligram protein, and the results of all groups were normalized to the sham group and served as the relative levels of oxidative stress.

ure the fluorescence intensity, and the samples of all rats were detected in at least one dependent experiment. The concentrations of ROS were expressed as fluorescence intensity per milligram protein, and the results of all groups were normalized to the sham group and served as the relative levels of oxidative stress. Brain Water Content and Behavioral Tests As described in a previous study, at 72 h post-ICH induction, rats were injected intraperitoneally with 4% chloral hydrate, with the intact brain tissues removed immediately [20]. The brain tissues were divided into two hemispheres along the midline, with each hemisphere divided into two parts containing the cortex and basal ganglia. The obtained samples and the cerebellum were then divided into the following five groups: contralateral cortex (Cont-CX), contralateral basal ganglia (Cont-BG), ipsilateral cortex (Ipsi-CX), ipsilateral basal ganglia (Ipsi-BG), and cerebellum (CB; all groups: n = 6). The brain tissues were immediately weighed with an electronic analytical balance and the wet weight was recorded. The brain tissues were then dried in an electric thermostatic drier at 100 ± 5 °C for 72 h until the sample weights were consistent to obtain the dry weight and calculated as follows: water content of brain tissues = (wet weight − dry weight) / (wet weight) × 100%.

ic analytical balance and the wet weight was recorded. The brain tissues were then dried in an electric thermostatic drier at 100 ± 5 °C for 72 h until the sample weights were consistent to obtain the dry weight and calculated as follows: water content of brain tissues = (wet weight − dry weight) / (wet weight) × 100%. Behavioral testing was performed at 1 h before sacrifice. All the rats in each group were examined using a previously published scoring system and monitored for appetite, activity, and neurological defects (details shown in Table 1) [22].Table 1 Neurobehavioral evaluation Category Behavior Score Appetite Finished meal 0 Left meal unfinished 1 Scarcely ate 2 Activity Walk and reach at least three corners of the cage 0 Walk with some stimulations 1 Almost always lying down 2 Deficits No deficits 0 Unstable walk 1 Impossible to walk 2 BBB Injury BBB permeability was assessed on the basis of albumin extravasation [23]. Generally, albumin concentration in the brain is very low because of the existence of BBB, but the content of albumin in brain tissues increases obviously once BBB is damaged. Therefore, the changes of albumin concentration can serve as an indicator to estimate the degree of BBB injury [24, 25]. The western blot analysis was used to test the protein levels of albumin in brain tissues of rats in each group.

, but the content of albumin in brain tissues increases obviously once BBB is damaged. Therefore, the changes of albumin concentration can serve as an indicator to estimate the degree of BBB injury [24, 25]. The western blot analysis was used to test the protein levels of albumin in brain tissues of rats in each group. Nissl Staining After coronal sections had been deparaffinized and rehydrated, the slides were stained in toluidine blue solution for 40 min at 50–60 °C. After clearing in distilled water, the slides were gradually dehydrated for 3 min in successive baths of ethanol, with one pass each in 70, 80, and 95% and two passes in 100%. All slides were then given two 5 min passes in 100% dimethylbenzene and coverslips were applied with neutral balsam. Finally, the numbers of surviving neurons per ×400 field within the hippocampal CA1 were counted. TUNEL Staining TUNEL staining was utilized to detect cellular apoptosis in the brain tissues around the hematoma from all of the groups according to the manufacturer’s protocol (In Situ Cell Death Detection Kit, Roche, Germany). The TUNEL-positive cells in the brain tissues around the hematoma were observed and analyzed using a fluorescence microscope (Olympus Co., Japan).

llular apoptosis in the brain tissues around the hematoma from all of the groups according to the manufacturer’s protocol (In Situ Cell Death Detection Kit, Roche, Germany). The TUNEL-positive cells in the brain tissues around the hematoma were observed and analyzed using a fluorescence microscope (Olympus Co., Japan). Western Blot Analysis Western blot analysis was performed as described previously [26]. Briefly, brain samples around the hematoma were collected, homogenized, and lysed separately in ice-cold RIPA lysis buffer (Beyotime, China). The samples were then centrifuged for 10 min (4 °C, 12,000g). The supernatants were collected immediately and the protein concentrations were determined using a bicinchoninic acid (BCA) kit (Beyotime, China) according to the manufacturer’s instructions. The protein samples (60 μg/lane) were separated by 10 or 12% SDS polyacrylamide gel and electrotransferred to nitrocellulose filter membranes (Millipore, USA). The membranes were blocked with 5% skim milk for 1 h at room temperature and then incubated with primary antibodies overnight at 4 °C. The membranes were then washed with TBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 2 h at room temperature. The protein bands were visualized using enhanced chemiluminescence (ECL), and the relative protein quantity was determined using ImageJ software (National Institutes of Health, USA).

nes were then washed with TBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody for 2 h at room temperature. The protein bands were visualized using enhanced chemiluminescence (ECL), and the relative protein quantity was determined using ImageJ software (National Institutes of Health, USA). For the release of cytochrome c into the cytoplasm, we isolated the mitochondria from the protein extraction of the brain tissues by Cell Mitochondria Isolation Kit (Beyotime, China), and then western blot analysis was used to measure cytochrome c in residual protein sample. The primary antibodies used include the following: matrix metalloproteinase (MMP)-9 (Abcam, USA), NADPH oxidase (NOX)-1 (Santa Cruz, USA), NOX-2 (Abcam, USA), heme oxygenase (HO)-1 (Santa Cruz, USA), NAD(P)H quinone oxidoreductase (NQO) 1 (Santa Cruz, USA), Bcl-2 (Abcam, USA), BAX (Abcam, USA), cleaved caspase 3 (Abcam, USA), cleaved poly(ADP-ribose) polymerase (c-PARP) (Abcam, USA), X-ray repair complementing defective repair in Chinese hamster cells (XRCC) 1 (Santa Cruz, USA), gamma histone H2AX (γ-H2AX) (Abcam, USA), cytochrome c (Santa Cruz, USA), and β-tubulin (Santa Cruz, USA) as a loading control. HRP-conjugated anti-IgG (Santa Cruz, USA) was used as secondary antibodies.

PARP) (Abcam, USA), X-ray repair complementing defective repair in Chinese hamster cells (XRCC) 1 (Santa Cruz, USA), gamma histone H2AX (γ-H2AX) (Abcam, USA), cytochrome c (Santa Cruz, USA), and β-tubulin (Santa Cruz, USA) as a loading control. HRP-conjugated anti-IgG (Santa Cruz, USA) was used as secondary antibodies. Immunofluorescence (IF) Staining The brain tissues was embedded in paraffin and sectioned at 4 μm, and immunofluorescence staining for myeloperoxidase (MPO), p53-binding proteins (53BP) 1, CD14, and CD68 was performed. The sections were incubated with primary MPO, 53BP1, CD14, and CD68 (all diluted in 1:100; Santa Cruz, USA) antibodies overnight at 4 °C. Secondary antibody (Life Technologies, USA, 1:300 dilution) was added and incubated for 1 h at 37 °C, and the sections were then washed three times with PBST. After final washing, sections were protected with coverslips, with the nucleus visualized with DAPI (Southern Biotech, USA). The brain tissues around the hematoma were observed and analyzed using a fluorescence microscope (Olympus Co., Japan), and the relative intensities were determined using ImageJ software (National Institutes of Health, USA).

s were protected with coverslips, with the nucleus visualized with DAPI (Southern Biotech, USA). The brain tissues around the hematoma were observed and analyzed using a fluorescence microscope (Olympus Co., Japan), and the relative intensities were determined using ImageJ software (National Institutes of Health, USA). Annexin V and PI Staining In Vitro After various treatments, neurons were trypsinized with 0.25% trypsin (without EDTA) and centrifuged at 300g for 5 min, and the cell pellet was resuspended in 500 μl binding buffer, with 5 μl Annexin V and 5 μl PI (Beyotime, China) added. After incubation for 20 min in the dark at 37 °C, the cells were analyzed by flow cytometry (FACS Calibur, BD, USA) and at least 20,000 events per sample were recorded. JC-1 and TMRM Staining In Vitro Tetrechloro-tetraethylbenzimidazol carbocyanine iodide (JC-1) and tetramethylrhodamine methyl ester perchlorate (TMRM) staining were used to detect neuron MPTP opening according to the manufacturer’s protocol (Mitochondrial membrane potential assay kit with JC-1, Beyotime, China; TMRM, Santa Cruz, USA) [27–29]. Pretreated neurons were washed three times with PBS, followed by the additional 1 ml JC-1 working solution per sample and incubated at 37 °C for 20 min. After incubation, the neurons were washed twice with JC-1 staining buffer and coverslipped with neutral balsam.

JC-1, Beyotime, China; TMRM, Santa Cruz, USA) [27–29]. Pretreated neurons were washed three times with PBS, followed by the additional 1 ml JC-1 working solution per sample and incubated at 37 °C for 20 min. After incubation, the neurons were washed twice with JC-1 staining buffer and coverslipped with neutral balsam. TMRM was dissolved in DMSO and diluted with PBS, with 1 ml TMRM solution added to each sample and incubated at 37 °C for 20 min. After incubation, the neurons were washed three times with PBS and coverslipped with neutral balsam. These results were visualized using a fluorescence microscope (Olympus Co., Japan), and the relative fluorescence intensities of JC-1 and TMRM were determined using ImageJ software (National Institutes of Health, USA). Statistical Analysis All data were expressed as mean ± SEM and GraphPad Prism 6.0 was adopted for all statistical analyses. Data sets were tested for normality of distribution with Kolmogorov-Smirnov test. Data groups (two groups) with normal distribution were compared using two-sided unpaired Student’s t test, and the Mann-Whitney U test was used for nonparametric data. P < 0.05 was considered as statistically significant difference.

yses. Data sets were tested for normality of distribution with Kolmogorov-Smirnov test. Data groups (two groups) with normal distribution were compared using two-sided unpaired Student’s t test, and the Mann-Whitney U test was used for nonparametric data. P < 0.05 was considered as statistically significant difference. Results Melatonin Attenuated Neurological Behavior Impairment, BBB Disruption, and Brain Water Content in Brain Tissues After ICH To define the effects of melatonin in neurological behavioral impairment after ICH, we performed behavioral testing at 1 h before sacrifice. At 72 h after ICH induction, the rats showed severe neurological behavioral impairment compared to the sham group. After receiving an intraperitoneal melatonin injection, the impairment was ameliorated (Table 2). We further evaluated the levels of albumin in each group, which is an important hallmark of BBB disruption. In the ICH group, albumin levels in the brain tissues were significantly increased when compared to the sham group, while melatonin treatment significantly reduced the albumin levels induced by ICH (Fig. 2a, b).Table 2 Clinical behavioral scores in each group Group (n = 12) Scores (mean ± SEM) Sham 0.333 ± 0.211 ICH 3.167 ± 0.307* ICH + vehicle 3.500 ± 0.224 ns ICH + melatonin 2.500 ± 0.342# *P < 0.05 vs. sham group; ns, no significant difference vs. ICH group; #P < 0.05 vs. ICH + vehicle group

Results Melatonin Attenuated Neurological Behavior Impairment, BBB Disruption, and Brain Water Content in Brain Tissues After ICH To define the effects of melatonin in neurological behavioral impairment after ICH, we performed behavioral testing at 1 h before sacrifice. At 72 h after ICH induction, the rats showed severe neurological behavioral impairment compared to the sham group. After receiving an intraperitoneal melatonin injection, the impairment was ameliorated (Table 2). We further evaluated the levels of albumin in each group, which is an important hallmark of BBB disruption. In the ICH group, albumin levels in the brain tissues were significantly increased when compared to the sham group, while melatonin treatment significantly reduced the albumin levels induced by ICH (Fig. 2a, b).Table 2 Clinical behavioral scores in each group Group (n = 12) Scores (mean ± SEM) Sham 0.333 ± 0.211 ICH 3.167 ± 0.307* ICH + vehicle 3.500 ± 0.224 ns ICH + melatonin 2.500 ± 0.342# *P < 0.05 vs. sham group; ns, no significant difference vs. ICH group; #P < 0.05 vs. ICH + vehicle group Fig. 2 Evaluation of blood-brain barrier (BBB) disruption and brain water content of brain tissues at 72 h post-ICH induction. a Western blot analysis examines the albumin level of the sham, ICH, ICH + vehicle, and ICH + melatonin groups. b Relative albumin levels were calculated based on densitometry analysis. The mean albumin level of the sham group was normalized to 1.0. c Recorded brain water content at 72 h post-ICH. All data are displayed as means ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

erum from rats in these four groups were elevated by ELISA. The results confirmed that IL-1β and TNF-α levels were significantly increased in the ICH group compared to the sham group. However, melatonin treatment significantly decreased the levels of these two cytokines compared to the ICH + vehicle groups (Fig. 4c–f). To explore infiltration of inflammatory cells in the brain tissues around the hematoma after ICH, the indicators (CD14, CD68, and MPO) of the inflammatory cells were examined by immunofluorescence staining. The results showed that, relative to the sham group, the positive ratios of these indicators were increased in the ICH groups. Nevertheless, these positive ratios were reduced significantly following melatonin treatment (Fig. 5a–f). These findings suggested that melatonin treatment can inhibit inflammation in brain tissues after ICH.Fig. 5 Immunofluorescence (IF) staining for the identification of inflammatory cells around the hematoma in brain tissues at 72 h post-ICH. Representative IF staining to identify CD14-positive (a), CD68-positive (c), and MPO-positive (e) cells (green or red), with the nuclei fluorescently labeled with 4,6-diamino-2-phenyl indole (DAPI, blue); scale bar = 32 μm. Percentage of CD14-positive (b), CD68-positive (d), and MPO-positive (f) cells around the hematoma in the brain tissues. Arrows indicate CD14-positive, CD68-positive, and MPO-positive cells. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

e of CD14-positive (b), CD68-positive (d), and MPO-positive (f) cells around the hematoma in the brain tissues. Arrows indicate CD14-positive, CD68-positive, and MPO-positive cells. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 Melatonin Reduced DNA Damage in Brain Tissues at 72 h Post-ICH Induction The protein levels of γ-H2AX and XRCC1, two indicators of DNA damage, were detected by western blot analysis, and the results showed that, compared to the sham group, they were obviously increased in the ICH group, while no significant difference was noted between the ICH and ICH + vehicle groups. Furthermore, the levels of γ-H2AX and XRCC1 were significantly reduced by melatonin treatment when compared to the ICH + vehicle group (Fig. 6a–c).Fig. 6 DNA damage indicator expression and immunofluorescence (IF) staining to visualize DNA-damaged cells around the hematoma in brain tissues 72 h post-ICH induction. a Western blot analysis examining XRCC1 and γ-H2AX levels in the sham, ICH, ICH + vehicle, and ICH + melatonin groups. b, c Relative XRCC1 and γ-H2AX levels were calculated based on densitometry analysis. The mean values of XRCC1 and γ-H2AX within the sham group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6. d Representative IF staining to identify 53BP1-positive cells (green) and with nuclei fluorescently labeled with 4,6-diamino-2-phenyl indole (DAPI, blue). Arrows indicate 53BP1-positive cells; scale bar = 32 μm. e Percentage of 53BP1-positive cells around the hematoma in the brain. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

abeled with 4,6-diamino-2-phenyl indole (DAPI, blue). Arrows indicate 53BP1-positive cells; scale bar = 32 μm. e Percentage of 53BP1-positive cells around the hematoma in the brain. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 53BP1 is also an indicator of DNA damage. The numbers of 53BP1-positive cells in brain tissues around the hematoma were increased significantly in the ICH group compared to the sham group, and no significant difference was noted between the ICH and ICH + vehicle groups. However, the ratio of 53BP1-positive cells was reduced remarkably in the ICH + melatonin group compared to the ICH + vehicle group (Fig. 6d, e). These results suggested that melatonin can alleviate ICH-induced DNA damage in brain tissues.

and no significant difference was noted between the ICH and ICH + vehicle groups. However, the ratio of 53BP1-positive cells was reduced remarkably in the ICH + melatonin group compared to the ICH + vehicle group (Fig. 6d, e). These results suggested that melatonin can alleviate ICH-induced DNA damage in brain tissues. Melatonin Promoted Antioxidant in Brain Tissues at 72 h After ICH The results of western blot analysis demonstrated that, compared with the sham group, the protein levels of HO-1 and NQO1, two indicators of antioxidant, were significantly increased in the ICH group, while there was no significant difference between the ICH and ICH + vehicle groups. Meanwhile, melatonin treatment increased HO-1 and NQO1 levels compared to the ICH + vehicle group (Fig. 7a–c). The results suggested that melatonin plays an important role in alleviating oxidative stress by promoting antioxidant in brain tissues after ICH.Fig. 7 Melatonin increased antioxidant indicator expressions in brain tissues at 72 h after ICH. a Western blot analysis displayed the expressions of HO-1 and NQO1 in the sham, ICH, ICH + vehicle, and ICH + melatonin groups. Relative HO-1 (b) and NQO1 (c) levels were calculated based on densitometry analysis. The mean values of the HO-1 and NQO1 within the sham group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

onin groups. Relative HO-1 (b) and NQO1 (c) levels were calculated based on densitometry analysis. The mean values of the HO-1 and NQO1 within the sham group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 Neuroprotective Effects of Melatonin Against the Neuron Death Induced by ICH We used Nissl staining to further assess whether melatonin has the protective effects on neurons in the CA1 region of the hippocampus after ICH. The sham group hardly had any neuronal death. The numbers of surviving neurons in the ICH group were significantly reduced to that in the sham group, but melatonin treatment remarkably increased it. There was no significant difference between the ICH and ICH + vehicle groups (Fig. 8a, b). These results showed that melatonin was against neuron death in brain tissues after ICH.Fig. 8 Neuroprotective effects of melatonin against ICH-induced neuronal death. a Representative Nissl staining sections of the hippocampus and hippocampal CA1 region in rats. Arrows indicate surviving neurons; scale bar = 32 μm. B quantitative analysis of the numbers of surviving neurons per ×400 field in the hippocampal CA1 region. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

l CA1 region in rats. Arrows indicate surviving neurons; scale bar = 32 μm. B quantitative analysis of the numbers of surviving neurons per ×400 field in the hippocampal CA1 region. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 Melatonin Inhibited Apoptosis in Brain Tissues After ICH Both In Vivo and In Vitro To explore the effects of melatonin in cell apoptosis after ICH, the expression levels of cleaved caspase 3 and cleaved PARP, two indicators of apoptosis, were tested by western blot analysis. The results suggested that, compared to the sham group, there was a significant increase in the ICH group, while this upregulation was obviously decreased by melatonin treatment. Another indicator of apoptosis, the ratio of Bcl-2/BAX, was decreased in the ICH group compared to the sham group, but melatonin treatment increased the ratio of Bcl-2/BAX in brain tissues after ICH (Fig. 9a–d). Microscopically, TUNEL staining was utilized to explore the effects of melatonin treatment on apoptosis in the brain tissues at 72 h post-ICH induction. The numbers of TUNEL-positive cells in the brain tissues around the hematomas were significantly increased in the ICH groups compared to the sham group, while a reduction was noted in the ICH + melatonin group (Fig. 9e, f).Fig. 9 Melatonin inhibited apoptosis in brain tissues at 72 h after ICH. a Western blot analysis shows c-PARP, Bcl-2, Bax, and cleaved caspase 3 levels in the sham, ICH, ICH + vehicle, and ICH + melatonin groups. b–d Relative c-PARP, cleaved caspase 3 levels, and the ratio of Bcl-2/Bax were calculated based on densitometry analysis. The mean values of c-PARP, cleaved caspase 3, and the ratio of Bcl-2/Bax in the sham group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6. e Double staining for terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL, green) and 4,6-diamino-2-phenyl indole (DAPI, blue). Arrows indicate TUNEL-positive cells; scale bar = 32 μm. f Percentage of TUNEL-positive cells around the hematoma in the brain tissues. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

and 4,6-diamino-2-phenyl indole (DAPI, blue). Arrows indicate TUNEL-positive cells; scale bar = 32 μm. f Percentage of TUNEL-positive cells around the hematoma in the brain tissues. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 In the in vitro experiments, after various treatments, neurons were digested to generate cell suspension, stained with Annexin V and PI, and examined by flow cytometry. The results showed that the apoptotic ratio was higher in the OxyHb treatment group relative to the control group. Additionally, the apoptotic ratio of neurons in the OxyHb + melatonin group was significantly decreased compared to the OxyHb + vehicle group (Fig. 10a, b). These results indicated that melatonin can inhibit apoptosis in neurons induced by ICH both in vivo and in vitro.Fig. 10 Melatonin inhibited OxyHb-induced neuronal apoptosis. a Neuronal apoptosis in various groups were detected via PI and Annexin V double staining and flow cytometry analysis in vitro. b Bar graphs showing different conditions of neurons in various groups: PI−/Annexin V− represents survival neurons, PI+/Annexin V− represents necroptotic neurons, PI−/Annexin V+ represents apoptotic neurons, and PI+/Annexin V+ represents neurons with mixed damage. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 considered significant; NS, no significant difference, n = 6

I−/Annexin V− represents survival neurons, PI+/Annexin V− represents necroptotic neurons, PI−/Annexin V+ represents apoptotic neurons, and PI+/Annexin V+ represents neurons with mixed damage. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 considered significant; NS, no significant difference, n = 6 Melatonin Suppressed the MPTP Opening and Mitochondrial Damage After ICH Both In Vivo and In Vitro To define the effects of melatonin treatment in mitochondrial damage in brain tissues after ICH, the levels of cytochrome c in the cytoplasm, an indicator of mitochondrial damage, were detected by western blot analysis. The results demonstrated that, compared with the sham group, the levels of cytochrome c in the cytoplasm were obviously increased in the ICH group. However, melatonin reduced the levels of cytochrome c in the cytoplasm compared to the ICH + vehicle groups (Fig. 11a, b). These results suggested that melatonin plays an important role in protecting mitochondria injury after ICH in vivo.Fig. 11 Melatonin suppressed mitochondrial damage indicator levels in brain tissues at 72 h after ICH and mitochondrial permeability transition pore (MPTP) opening in vitro. a Western blot analysis showed cytochrome c expression level in the sham, ICH, ICH + vehicle, and ICH + melatonin groups in vivo. b Relative cytochrome c levels were calculated based on densitometry analysis. The mean cytochrome c level of the sham group was normalized to 1.0. All data are displayed as a mean ± SEM, with *P < 0.05 and & P < 0.05 considered significant. NS, no significant difference, n = 6. The primary cultured neurons were stained by JC-1 (c) and TMRM (e) after OxyHb treatment and were then observed using fluorescence microscopy. Arrows indicate neurons; scale bar = 32 μm. The relative fluorescence intensity of JC-1 (d) and TMRM (f) was calculated using ImageJ. The mean fluorescence intensity values in the control group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6

32 μm. The relative fluorescence intensity of JC-1 (d) and TMRM (f) was calculated using ImageJ. The mean fluorescence intensity values in the control group were normalized to 1.0. All data are displayed as mean ± SEM, with *P < 0.05 and & P < 0.05 deemed as significant difference; NS, no significant difference, n = 6 JC-1 and TMRM are ideal fluorescent probes widely used to detect mitochondrial membrane potential (△Ψm). The mitochondrial transmembrane potential makes some lipophilic cationic fluorescent dyes such as JC-1 and TMRM to bind to the mitochondrial matrix, and the enhancement or decrease of fluorescence indicates the increase or decrease of electrical negativity of the mitochondrial inner membrane. MPTP opening was evaluated by observing the relative fluorescence intensity of JC-1 (Fig. 11c, d) and TMRM (Fig. 11e, f) in neurons in vitro. MPTP opening increased significantly in the OxyHb group compared to the sham group, while it decreased in the OxyHb + melatonin group compared with the OxyHb + vehicle group. These results indicated that melatonin can offer mitochondrial protection by decreasing OxyHb-induced MPTP opening in neurons in vitro.

in vitro. MPTP opening increased significantly in the OxyHb group compared to the sham group, while it decreased in the OxyHb + melatonin group compared with the OxyHb + vehicle group. These results indicated that melatonin can offer mitochondrial protection by decreasing OxyHb-induced MPTP opening in neurons in vitro. Discussion The mechanisms contributing to SBI after ICH are complex and mainly attributed to the following: mechanical rupture of nerve and glial cells due to the space-occupying effect of the hematoma, excitatory amino acid toxicity, free radical induced destruction of intracranial cells via increased levels of ROS and other small molecules, inflammation induced by intravascular inflammatory cells (neutrophils, macrophages, and so on) and activated microglia in brain tissues, and apoptosis induced by a variety of mechanisms [6, 30–32].

id toxicity, free radical induced destruction of intracranial cells via increased levels of ROS and other small molecules, inflammation induced by intravascular inflammatory cells (neutrophils, macrophages, and so on) and activated microglia in brain tissues, and apoptosis induced by a variety of mechanisms [6, 30–32]. Brain edema is a major and severe pathological event induced by ICH, with the mechanisms contributing to its formation being quite complex [33, 34]. The pathological process of ICH-induced brain edema mainly included the following: the space-occupying effects of the hematoma, the hydrostatic pressure generated during blood clot formation and retraction, destruction of the BBB, a coagulation cascade and the formation of thrombin, the dissolution of red blood cells and the toxic effect of hemoglobin, and secondary cerebral ischemia/reperfusion injury [35–39]. In this study, we found that melatonin significantly reduced ICH-induced brain edema. Furthermore, melatonin reduced the albumin concentration in the ICH group, thus indicating that melatonin can rehabilitate the BBB integrity. This rehabilitation of the BBB integrity may be an important contributing factor to the reduction of brain edema following melatonin treatment.

y reduced ICH-induced brain edema. Furthermore, melatonin reduced the albumin concentration in the ICH group, thus indicating that melatonin can rehabilitate the BBB integrity. This rehabilitation of the BBB integrity may be an important contributing factor to the reduction of brain edema following melatonin treatment. During the inflammatory reaction in brain tissues after ICH, the most basic signs are the activation of microglia and the infiltration of inflammatory cells [40]. Previous studies found that white blood cells and macrophages released from the hematoma infiltrated and blocked microblood vessels thereby reducing cerebral perfusion and damaging the BBB [41, 42]. TNF-α, IL-1β, and other cytotoxic molecules caused damage to neurons [43] and were involved in the process of SBI [44]. Previous studies showed that ischemia and hypoxia occurred in the brain tissues around the hematoma after ICH, which activated microglia, inflammatory cells, vascular endothelial cells, and so on, and then led to increased expression of MMP-9. MMP-9 caused vascular matrix degradation and disrupted BBB, thus contributing to brain edema [45–50]. Our study showed that melatonin can decrease inflammatory cytokine (TNF-α and IL-1β) levels and reduce inflammatory cell infiltration in brain tissues, thereby alleviating the inflammatory reaction after ICH (Fig. 12).Fig. 12 Proposed mechanisms underlying the positive therapeutic effects of melatonin treatment on SBI induced by ICH. c-caspase 3 cleaved caspase 3, c-PARP cleaved poly(ADP-ribose) polymerase (PARP), NOX NADPH oxidase, TNF tumor necrotic factor, MMP-9 matrix metalloproteinase 9, IL-1β interleukin-1β, MPO myeloperoxidase, ROS reactive oxygen species, HO heme oxygenase, NQO1 NAD(P)H quinone oxidoreductase 1, XRCC X-ray repair complementing defective repair in Chinese hamster cells, MPO myeloperoxidase, 53BP1 p53-binding proteins 1

ase, TNF tumor necrotic factor, MMP-9 matrix metalloproteinase 9, IL-1β interleukin-1β, MPO myeloperoxidase, ROS reactive oxygen species, HO heme oxygenase, NQO1 NAD(P)H quinone oxidoreductase 1, XRCC X-ray repair complementing defective repair in Chinese hamster cells, MPO myeloperoxidase, 53BP1 p53-binding proteins 1 Oxidative stress played an essential role in the occurrence of ICH-induced brain edema and SBI. Following ICH induction, a large number of free radicals were released, and ROS led to membrane lipid peroxidation and protein and DNA oxidative damage [51–53]. HO-1 protected cells and tissues during oxidative stress as an important member of antioxidant [54–56]. NQO1 used NADH or NADPH as electron donors to catalyze the reduction of chinone compounds, thereby avoiding the formation of unstable semiquinone compounds [56, 57]. In this study, we found that melatonin increased the expressions of HO-1 and NQO1 after ICH, and these results indicated that melatonin can promote antioxidant in brain tissues after ICH. The major function of the NOX family was the production of ROS [58, 59]. Following ICH induction, the expression of NOX-1 and NOX-2 were upregulated and contributed to brain injury. Our study showed that increasing of NOX-1 and NOX-2 after ICH could be ameliorated by melatonin. Furthermore, melatonin decreased ROS levels in brain tissues after ICH, thereby alleviating oxidative stress.

59]. Following ICH induction, the expression of NOX-1 and NOX-2 were upregulated and contributed to brain injury. Our study showed that increasing of NOX-1 and NOX-2 after ICH could be ameliorated by melatonin. Furthermore, melatonin decreased ROS levels in brain tissues after ICH, thereby alleviating oxidative stress. Apoptosis was the main mechanism of early tissue injury in the region surrounding the hematoma after ICH. There were many factors which induced cell apoptosis after ICH, such as free radical cascade reaction, inflammation, cytokine stimulation, and the induction of thrombin and blood components. There were also various genes that regulated neuronal apoptosis, including BAX, which promoted apoptosis, and Bcl-2, which inhibited apoptosis [60, 61]. When the expression of Bcl-2 increased, the heterodimer Bcl-2-Bax formed and inhibited cell apoptosis. On the contrary, when the expression of Bax increased, the form of homodimer Bax/Bax was also increased and promoted cell apoptosis. The ratio of Bcl-2/Bax can reflect whether the cells tend to undergo apoptosis or survive after stimulation. When the level of Bcl-2 was increased while Bax was decreased, the ratio of Bcl-2/Bax increased, and cells tended to survive [62, 63]. In this study, we found that the ratio of Bcl-2/Bax in brain tissues decreased post-ICH but this decrease was alleviated by melatonin. At the start of cellular apoptosis, PARP was cleaved into two fragments by caspase 3, thereby inactivating PARP and leading to apoptosis [64, 65]. Neurons in the hippocampal CA1 region were closely related to learning and memory function of human and mammal and sensitive to ischemia. In recent years, researchers found that secondary cerebral ischemia probably occurred in the hippocampal CA1 region post-ICH. The delayed death of neurons in the hippocampal CA1 region after transient cerebral ischemia was the process of apoptosis [66, 67]. This study showed that melatonin can reduce apoptotic cell numbers around the hematoma in brain tissues after ICH.

rebral ischemia probably occurred in the hippocampal CA1 region post-ICH. The delayed death of neurons in the hippocampal CA1 region after transient cerebral ischemia was the process of apoptosis [66, 67]. This study showed that melatonin can reduce apoptotic cell numbers around the hematoma in brain tissues after ICH. The mitochondria, a eukaryotic organelle, consists of a bilayer that aids in ATP production, and is the site of intracellular oxygen free radical production and the main target of oxygen free radicals [68, 69]. Mitochondrial dysfunction can lead to a variety of intracellular signaling cascades, oxidative stress, and apoptosis, which play a crucial role in the progress of almost all diseases [70]. There was evidence that mitochondrial damage induced by cerebral I/R injury was directly related with neuronal apoptosis [71–73]. The mitochondria also have other important functions, such as ROS production, regulating cellular redox potentials and signal transduction, and controlling cellular apoptosis and gene expression [73, 74]. Following ICH induction, mitochondrial injury occurred as the result of MPTP opening; a variety of proteins including cytochrome c were released into the cytoplasm, being an important event that led to cell apoptosis [75, 76]. Melatonin reduced mitochondrial dysfunction by upregulating antioxidants, thus inhibiting MPTP opening. This study suggested that melatonin treatment significantly inhibited apoptosis and mitochondrial damage, with these findings substantiated both in vitro and in vivo. Additionally, ICH-induced DNA damage was also an important cause of SBI. When DNA was broken, 53BP1 and γ-H2AX were synthesized and released, with γ-H2AX providing an index of DNA damage and 53BP1 promoting DNA repair [77–79]. This study showed that melatonin treatment can alleviate ICH-induced DNA damage.

and in vivo. Additionally, ICH-induced DNA damage was also an important cause of SBI. When DNA was broken, 53BP1 and γ-H2AX were synthesized and released, with γ-H2AX providing an index of DNA damage and 53BP1 promoting DNA repair [77–79]. This study showed that melatonin treatment can alleviate ICH-induced DNA damage. We also noticed that there are three papers from a group reporting that melatonin reduced oxidative stress, and provided brain protection after ICH, but it did not change the extent of brain edema or neurologic deficits in short-term outcomes [20, 80, 81]. These differences may due to the methods of induction of ICH (we used autologous whole blood, but they used bacterial collagenase in induction of ICH in rats). In collagenase ICH model, bacterial collagenase induced excessive inflammation response in the brain tissues; this should not be ignored in neuroprotection-related studies. In this study, we researched the neuroprotective effects of melatonin in autologous blood injection ICH model for the first time. Taking these researches together, using multiple models may be more appropriate in the study of brain injury following ICH.

es; this should not be ignored in neuroprotection-related studies. In this study, we researched the neuroprotective effects of melatonin in autologous blood injection ICH model for the first time. Taking these researches together, using multiple models may be more appropriate in the study of brain injury following ICH. One of the limitations of the present study is that it only covered a short period of time; thus, the long-term effects of melatonin on ICH remain unclear. Furthermore, only the effects of melatonin on inflammation, oxidative stress, apoptosis, and other brain injury were examined, but the mechanisms were not characterized. The relationship between inflammation, oxidative stress, and apoptosis in ICH thus remains unclear. Moreover, the direct effect of melatonin on the inflammatory reaction was not examined herein, although the reduction in inflammatory response was possible due to the reduction in the oxidative stress response. According to previous reports, receptors of melatonin are as follows: MT1 (on cell membrane), MT2 (on cell membrane), melatonin receptor type 1c, quinone reductase 2 enzyme (MT3 receptor, a detoxification enzyme), retinoid-related orphan nuclear hormone receptor, and GPR50 (X-linked melatonin-related orphan receptor) [82]. To date, there were only a few studies reporting that MT1 and MT2 receptors were expressed in the brain tissues in normal SD rats; however, whether expression levels of these two receptors were changed and other receptors of melatonin were expressed in brain tissues in rats after ICH have not been reported [83]. So it is necessary to make further investigations to identify whether the effects of melatonin in brain injury after ICH in rats were mediated by MT1 receptor, MT2 receptor, and (or) other receptors. Meanwhile, some researchers thought that melatonin maybe a pleiotropic agent that is capable of interfering with oxidative stress, cell apoptosis, inflammation, and so on, all of which would be useful in treating common pathological events taking place in ICH and other disease [84]. Thus, the main purpose of this study was investigating the effects of melatonin in secondary brain injury after ICH. Of course, the major mechanisms and the key targets of melatonin involved in melatonin-induced neuroprotective effects showed in this study would be explored in our future work.

H and other disease [84]. Thus, the main purpose of this study was investigating the effects of melatonin in secondary brain injury after ICH. Of course, the major mechanisms and the key targets of melatonin involved in melatonin-induced neuroprotective effects showed in this study would be explored in our future work. In conclusion, we found that melatonin treatment can alleviate SBI and protect brain tissues after ICH by impacting apoptosis, inflammation, oxidative stress, DNA damage, brain edema, and BBB damage and reducing mitochondrial membrane permeability transition pore opening. This study indicated that melatonin may become an important treatment against mitochondrial dysfunction and ICH-induced disabilities. Zhong Wang and Feng Zhou contributed equally to this work. Acknowledgements This work was supported by Suzhou Key Medical Center (Szzx201501) and grants from the National Natural Science Foundation of China (No. 81571115), Scientific Department of Jiangsu Province (No. BL2014045), Suzhou Government (No. SZS201413, SYS201608, and LCZX201601), and Jiangsu Province (No. 16KJB320008). Authors’ Contributions G.C. and H.S. conceived and designed the study, including quality assurance and control. Z.W. and F.Z. performed the experiments and wrote the paper. Y.D. and X.T. designed the study’s analytic strategy and wrote the paper. C.L. helped conduct the literature review and prepare the Materials and Methods section of the text. H.L. reviewed and edited the manuscript. All authors read and approved the manuscript.

W. and F.Z. performed the experiments and wrote the paper. Y.D. and X.T. designed the study’s analytic strategy and wrote the paper. C.L. helped conduct the literature review and prepare the Materials and Methods section of the text. H.L. reviewed and edited the manuscript. All authors read and approved the manuscript. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no competing interests. Ethics All animal experimental procedures were approved by the First Affiliated Hospital of Soochow University and performed in accordance with the guidelines of the National Institutes of Health on the care and use of animals. Animals were housed in an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC)-approved animal quarters in our hospital with a controlled temperature of 22 °C and 12-h light-dark cycles.

Introduction The prevalence of dementia is expected to triple by 2050, a major threat to the world’s public health. Vascular dementia makes up to 20% of the cases of dementia, and mixed dementia (vascular and Alzheimer’s) is estimated to be up to 50% of the cases of dementia [1]. Moreover, cerebral ischemia worsens Alzheimer’s disease (AD) and triggers its clinical expression [2]. Vascular contributions to cognitive impairment and dementia (VCID) is a term that encompasses the vascular factors and vascular pathology that underlie the clinical spectrum from mild cognitive impairment to dementia. VCID is a major research focus of the National Institute of Neurological Diseases and Stroke [3, 4]. Other than controlling the comorbid vascular risk factors, there is no known effective treatment for VCID. However, clinical observational studies strongly suggest that physical exercise is effective at reducing progression of cognitive decline and dementia [5]. Reduction of cerebral blood flow (CBF) may be the key precipitating event in AD and VCID [1]. Hypoperfusion appears to be an early finding that plays a pathophysiological role in the development of white matter (WM) damage [6]. Low blood flow by MRI perfusion or MRI arterial spin labeling (ASL) is predictive of white matter lesions [7, 8]. A penumbra exists around white matter lesions that expand in relation to low CBF [8]. An animal model for VCID is essential to test new interventions.

ole in the development of white matter (WM) damage [6]. Low blood flow by MRI perfusion or MRI arterial spin labeling (ASL) is predictive of white matter lesions [7, 8]. A penumbra exists around white matter lesions that expand in relation to low CBF [8]. An animal model for VCID is essential to test new interventions. In a recent review of all mouse models for VCID, Bink et al. determined the mouse bilateral carotid artery stenosis (BCAS) model to be the most valid [9]. This model reproduces the WM damage, cerebral hypoperfusion, inflammation, BBB damage, and cognitive deficits of the human condition [10, 11]. There are also fibrinoid changes in the small vessels of the brain with gliosis and disruption of aquaporin polarization [12]. With these small vessel changes, the BCAS model may be useful to test interventions to treat small vessel disease of the brain [12]. Remote limb ischemic conditioning (RIC) is the simple, inexpensive, and safe use of repetitive inflation of a blood pressure (BP) cuff on the arm or leg to protect distant organs such as the brain from ischemic injury. Chronic RIC (C-RIC) is the repetitive use of daily RIC for periods of weeks or months. RIC shares common mechanisms with physical exercise and may be viewed as an “exercise equivalent” [13, 14]. We previously showed in a BCAS mouse model that C-RIC for 2 weeks increased CBF in a sustained fashion, reduced WM damage, improved cognitive performance, and reduced accumulation of amyloid-beta 42 protein (Aβ42) in the brain [15].

chanisms with physical exercise and may be viewed as an “exercise equivalent” [13, 14]. We previously showed in a BCAS mouse model that C-RIC for 2 weeks increased CBF in a sustained fashion, reduced WM damage, improved cognitive performance, and reduced accumulation of amyloid-beta 42 protein (Aβ42) in the brain [15]. However, whether RIC can induce long-term neuroprotection and how long RIC would need to be applied are not known. This study aimed to determine if C-RIC can induce long-term (at 6 months) neurovascular protection, and if 1 month of daily RIC is as effective as 4 months in a BCAS model. Our secondary aim was to determine whether C-RIC-induced neuroprotection is attributed to RIC-induced vascular remodeling and angiogenesis in the brain, similar to what is seen with chronic physical exercise. Materials and Methods Animal Models and Experimental Groups The animals were housed at Augusta University’s animal care facility, which is approved by the American Association for Accreditation of Laboratory Animal Care. This study was conducted in accordance with the National Institute of Health guidelines for the care and use of animals in research, and all protocols were approved by the institutional animal care and use committee. All the STAIR (Stroke Therapy Academic Industry Roundtable) and RIGOR recommendations and guidelines regarding randomization, blinding, and statistical analysis were followed in this study [16, 17].

care and use of animals in research, and all protocols were approved by the institutional animal care and use committee. All the STAIR (Stroke Therapy Academic Industry Roundtable) and RIGOR recommendations and guidelines regarding randomization, blinding, and statistical analysis were followed in this study [16, 17]. Determine the Effect of Chronic RIC on Cognitive Impairment, Functional Outcomes, and CBF Forty male mice were randomized into four groups: (1) sham (operated group for procedures of BCAS), (2) BCAS and sham RIC, (3) BCAS treated with daily RIC for 1 month post BCAS surgery (BCAS + RIC-1MO, N = 10), and (4) BCAS treated with daily RIC for 4 months post BCAS surgery (BCAS + RIC-4MO, N = 10). The outcomes were assessed by a blinded investigator. Cognitive function, functional outcomes, and CBF changes were considered as the primary outcomes and were determined at 4 and 6 months after BCAS, with either 1MO or 4MO RIC. Thereafter, the animals were sacrificed, the blood was collected for plasma nitrite estimation, and the brains were isolated and dissected for IHC (Schematic representation of the study plan in Supplemental Fig. 1).

ed as the primary outcomes and were determined at 4 and 6 months after BCAS, with either 1MO or 4MO RIC. Thereafter, the animals were sacrificed, the blood was collected for plasma nitrite estimation, and the brains were isolated and dissected for IHC (Schematic representation of the study plan in Supplemental Fig. 1). Bilateral Carotid Artery Stenosis Surgical Procedure BCAS was performed as previously described [10, 15]. In brief, animals were anesthetized with 2% isofluorane and both common carotid arteries (CCAs) were exposed by a midline cervical incision. Customized microcoils, specially designed to mimic a VCID model in the mice which was made of steel wire with an inner diameter of 0.18 mm, was twined by rotating around both right and left CCA at 30 min interval. Non-Invasive RIC Therapeutic Methods Non-invasive RIC therapy method was performed as published elsewhere [18, 19] with a programmable cuff or its sham procedure with a cuff that did not inflate or deflate (see detail in “ Supplemental Methods ”) [15]. Bilateral RIC therapy was performed in both hind limbs simultaneously (4 cycles × 5 min duration of each cycle inflate and deflate × 5 min interval between each cycle) daily for 1 month (BCAS+RIC-1MO) or 4 months (BCAS+RIC-4MO) after 1 week from BCAS surgery.

te or deflate (see detail in “ Supplemental Methods ”) [15]. Bilateral RIC therapy was performed in both hind limbs simultaneously (4 cycles × 5 min duration of each cycle inflate and deflate × 5 min interval between each cycle) daily for 1 month (BCAS+RIC-1MO) or 4 months (BCAS+RIC-4MO) after 1 week from BCAS surgery. Cerebral Blood Flow by Laser Speckle Contrast Imager Cerebral blood flow (CBF) was measured using high-resolution Laser Speckle Contrast Imager (LSCI) (PSI system, Perimed Inc.) at different time points as indicated in the figure and as previously reported by us [15]. Mice were placed on a warming pad and thermostatically controlled at around 37 °C to avoid the effect of body temperature during the measurement of CBF. Estimation of Nitrite in Plasma Plasma NOx (NO2 − + NO3 −) levels were measured by NO-specific chemiluminescence, as described previously [20] with slight modification. Briefly, 100 ul of plasma were mixed with twice volumes of ethanol (100%) and kept at −20 C for 40–60 min followed by centrifuged at 13000 rpm for 10 min to remove protein as pellet. The supernatant (100 ul) was taken and injected to measure nitrite. The nitrite levels were measured by NO Analyzer 280i (GE Analytical Instruments, CO, USA). The level of nitrite was expressed in nanomolar.

0%) and kept at −20 C for 40–60 min followed by centrifuged at 13000 rpm for 10 min to remove protein as pellet. The supernatant (100 ul) was taken and injected to measure nitrite. The nitrite levels were measured by NO Analyzer 280i (GE Analytical Instruments, CO, USA). The level of nitrite was expressed in nanomolar. Behavioral Test: Functional Outcomes and Cognitive Test An investigator who was blinded to the experimental design and behavioral test or treatment including NOR test, [21] Y-maze test, [22] beam walk, [23] and wire hanging test [24] (Detailed procedures are explained in the “Supplemental Methods.”). All groups were examined at 4 and 6 months after sham surgery, BCAS surgery, and with RIC (1MO or 4MO). NOR and Y-maze test were performed for cognitive test where beam walk test required mice to balance on a wooden beam to evaluate any gait abnormality and wire hanging test for muscular or motor function.

Methods.”). All groups were examined at 4 and 6 months after sham surgery, BCAS surgery, and with RIC (1MO or 4MO). NOR and Y-maze test were performed for cognitive test where beam walk test required mice to balance on a wooden beam to evaluate any gait abnormality and wire hanging test for muscular or motor function. Histological and Immunohistochemical Assessment Immunohistochemistry was performed according to the protocol previously described [15] with slight modification. Both paraffin and cryo-sections were used for immunohistochemistry and immunofluorescence with similar anatomic features (see the “Supplemental Methods” for details). The primary antibodies were anti-myelin basic protein (MBP) (C-16 clone, SC-13914, Santa Cruz, CA, USA; 1:100 dilution); rat anti-platelet endothelial cell adhesion molecule [PECAME-1 (CD31), BD no. 550274, USA,1:200 dilution], with mouse monoclonal anti-α-smooth muscle actin [α-SMA, SC-53142, Santa Cruz, CA, USA; 1:50 dilution]; and anti-platelet-derived growth factor [PDGFR-β (958): SC-432, Santa Cruz, CA, USA; 1:200 dilution], isolectin B4 conjugates (Invitrogen, molecular probe Life Technology, IB4 no. 121411). For biotinylated immunostaining, the brain sections were incubated in the anti-MBP primary antibodies and using the avidin–biotin–peroxidase complex method with diaminobenzidine (DAB) as the chromogen. The immunostaining was carried out using the ABC kit system (Vector, Burlingame, CA, USA). After staining, the sections were counterstained with Harris hematoxylin (cat. no. HHS35-1L; Sigma) for few seconds. The sections were then dehydrated rapidly through ethanol and xylene and mounted with VectaMount medium (Vector).

he immunostaining was carried out using the ABC kit system (Vector, Burlingame, CA, USA). After staining, the sections were counterstained with Harris hematoxylin (cat. no. HHS35-1L; Sigma) for few seconds. The sections were then dehydrated rapidly through ethanol and xylene and mounted with VectaMount medium (Vector). Assessment of New Collateral Formation and Angiogenesis: Micro-CT and BriteVu Methods Twenty-seven male mice were randomized into three groups: (1) sham, (2) BCAS and sham RIC (BCAS), and (3) BCAS treated with daily RIC for 3 weeks post BCAS surgery (BCAS+RIC). Three weeks after BCAS, blood was collected either from the eye (retro orbital) for flow cytometry analysis of endothelial progenitor cells (EPCs) and macrophages, or from the heart plasma nitrite estimation. Animals were transcardially perfused with heparinized saline to flush out the blood, followed immediately by freshly prepared BriteVu™ solution (Scarlet Imaging, LCC; Murray, UT, USA) according to manufacturer instruction. Carcasses were kept on ice at 4 °C overnight to ensure solidification of BriteVu dye prior to imaging. After 24 h, brains were isolated, fixed with 4% PFA for 48 h, and then saved in 70% alcohol till the time of imaging. SkyScan 1174 (Bruker micro-CT/formerly known as SkyScan, USA) is used for imaging. SkyScan is associated with a full-range software that provides fast volumetric reconstruction of 2D/3D quantitative analysis and realistic 3D visualization of cerebrovasculature. Outcomes: (1) full 3D picture of cerebrovasculature and (2) quantification of vascular number, vascular volume, vascular density, lumen diameter, and formation of new collaterals.

e software that provides fast volumetric reconstruction of 2D/3D quantitative analysis and realistic 3D visualization of cerebrovasculature. Outcomes: (1) full 3D picture of cerebrovasculature and (2) quantification of vascular number, vascular volume, vascular density, lumen diameter, and formation of new collaterals. Evaluation of EPCs and Macrophages Whole blood (WB) was collected using heparinated microtubes as described previously with slight modification [25, 26]. One hundred fifty milliliters of WB were then incubated with antibodies for EPC markers including CD31, CD34, VEGFR2, and surface markers for M1/M2 macrophages, CD11b, F4/80, and CD206 (eBioscience, USA) for 20 min on ice in the dark. After washing was completed, cells were fixed and permeabilized using fix/perm concentrate (eBioscience, USA) before incubation with antibodies for intracellular staining of TNFα (for M1 macrophages) and IL10 (for M2 macrophages). Samples were then washed and run through a four-color flow cytometer (FACSCalibur, BD Biosciences), and data were collected using the CellQuest software. Samples were double-stained with control IgG and cell markers to assess any spillover signal of fluorochromes. Proper compensation was set to ensure that the median fluorescence intensities of negative and positive cells were identical and then was used to gate the population. Gating excluded dead cells and debris using forward and side scatter plots. To confirm the specificity of primary antibody binding and rule out non-specific Fc receptor binding to cells or other cellular protein interactions, negative control experiments were conducted using isotype controls matched to each primary antibody’s host species, isotype, and conjugation format.

d side scatter plots. To confirm the specificity of primary antibody binding and rule out non-specific Fc receptor binding to cells or other cellular protein interactions, negative control experiments were conducted using isotype controls matched to each primary antibody’s host species, isotype, and conjugation format. Statistical Analysis All statistical analysis was performed using SAS 9.4, and statistical significance was assessed using a significance level of 0.05. To examine differences between the groups (sham, BCAS, BCAS+RIC-1MO, BCAS+RIC-4MO) for nitrites, a one-way ANOVA was used. If the overall test for the one-way ANOVA was statistically significant a Tukey–Kramer multiple comparison procedure was used to examine differences between the four groups. For cerebral blood flow, beam walk, hanging wire, NOR, and Y-maze measures, a repeated measures mixed model was used to examine differences between the four groups over time (CBF measurement times: baseline, post surgery, 4 months, 6 months; all other outcomes measurement times: 4 months, 6 months). Main effects of group and time as well as the two-factor interaction between group and time were included in the model. For CBF, an auto-regressive order 1 correlation structure fit the data best and was used to account for the correlation between measurement times. For the beam walk, hanging wire, NOR discrimination index, and Y-maze measures, an unstructured correlation structure was used as there were only two measurement times. Of statistical interest was the F-test for the two-factor interaction between group and time, and if it is statistically significant, it indicates that the change in the outcome over time is different for the four groups. To examine differences between groups within measurement time and between measurement times within group, a Bonferroni adjustment to the overall alpha level was used to control for the multiple post hoc pair-wise tests as not all pair-wise tests were of interest. For CBF, the Bonferroni adjusted alpha is 0.0010 and for the other outcomes, the Bonferroni adjusted alpha is 0.0031.

and between measurement times within group, a Bonferroni adjustment to the overall alpha level was used to control for the multiple post hoc pair-wise tests as not all pair-wise tests were of interest. For CBF, the Bonferroni adjusted alpha is 0.0010 and for the other outcomes, the Bonferroni adjusted alpha is 0.0031. Results RIC Improved Cerebral Blood Flow C-RIC therapy after BCAS increased CBF compared to sham RIC (Fig. 1a–e; Supplemental Fig. 2). CBF was significantly increased in both 1MO and 4MO RIC therapy as compared to BCAS-sham RIC. Moreover, 1MO of RIC was as effective as 4MO RIC in improving CBF at 6 months.Fig. 1 Measuring of CBF changes by Laser Speckle Contrast Imager (LSCI), remote ischemic conditioning (RIC) (1-MO and 4-MO therapy at 4MO and 6MO) increases cerebral blood flow (CBF) in bilateral carotid artery stenosis (BCAS) mice. a Mice underwent BCAS and were randomized to RIC daily for 4 months (long with dark blue arrow, top row), RIC daily for 1 month (short with dark blue arrow, second row), or sham RIC (green arrow, third row). The bottom row shows mice with sham BCAS surgery (no coils). CBF was measured at 4 and 6 months (Supplemental Fig. 2) in all mice. Daily RIC for 1 month produced similar increases in CBF to daily RIC for 4 months. Red indicates higher blood flow. b–e Absolute value of cerebral perfusion in perfusion unit (PU) at different time points where pre (prior to BCAS), post (after BCAS and prior to RIC), at 4 months (4MO) and at 6 months (6MO); pre and post sham operation, BCAS with sham RIC; and BCAS+RIC with 1MO and 4MO. Using repeated measures mixed models, there is no significant difference between CBF in the 1MO and 4MO therapy groups but both groups are significantly higher than the BCAS sham group at 4MO and at 6MO (N = 7 to 10/groups; a p < 0.0001 vs BCAS+sham-RIC; a p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO)

with 1MO and 4MO. Using repeated measures mixed models, there is no significant difference between CBF in the 1MO and 4MO therapy groups but both groups are significantly higher than the BCAS sham group at 4MO and at 6MO (N = 7 to 10/groups; a p < 0.0001 vs BCAS+sham-RIC; a p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO) Fig. 2 Using a repeated measures mixed model for time of exploration across (a) and (b),(a) time of exploration (TN) spent with the novel object at 4 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. The RIC for 1MO and 4MO groups shows no significant difference but is significantly different than BCAS (sham RIC). Values are indicated as mean ± SE. a p < 0.001 vs sham; b p < 0.001 vs BCAS+RIC-1MO; c p < 0.01 vs BCAS+RIC-4MO. (b) TN spent with the novel object at 6 months (6MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. Values are indicated as mean ± SE. a p < 0.001 vs sham; b p < 0.001 vs BCAS+RIC-1MO; c p < 0.01 vs BCAS+RIC-4MO. Using a repeated measures mixed model for the discrimination index across (c) and (d), (c) the discrimination index (DI) at 4 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO.The BCAS+RIC with 1MO and 4MO groups show no significant difference from one another but both are significantly different from the BCAS sham RIC. Values are indicated as mean ± SE. a p < 0.0001 vs sham; b p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO. (d) The discrimination index (DI) at 6 months (6MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. There is no difference between 1MO and 4MO RIC groups, but both are significantly different than BCAS (sham RIC). Values are indicated as mean ± SE. a p < 0.0001 vs sham; b p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO

discrimination index (DI) at 6 months (6MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. There is no difference between 1MO and 4MO RIC groups, but both are significantly different than BCAS (sham RIC). Values are indicated as mean ± SE. a p < 0.0001 vs sham; b p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO RIC Improved Cognition Through Enhanced Spatial and Working Memory Both 1MO and 4MO RIC improved cognition. Results from NOR test (for spatial memory) (Fig. 2a–d) and Y-maze test for working memory (Supplemental Fig. 3A–D) showed that the sham group is more attracted toward novel objects compared to BCAS. The BCAS group had significantly shorter exploration time for novel object (TN) and less discrimination index (DI). This indicates that BCAS causes impairment of discriminative ability in mice. RIC therapy for either 1-MO or 4-MO significantly restored the TN and DI scores. Moreover, RIC therapy for either 1MO or 4-MO significantly increased the entries’ alternations in the arms of Y-maze compared to the sham group tested at 4 and 6 months after BCAS.Fig. 3 RIC prevents motor/muscular impairment after BCAS. Using a repeated measures mixed model for beam walk across (a) and (b), (a) time of crossing on a beam at 4 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. Values are indicated as mean ± SE. a p < 0.0001 vs sham, BCAS+RIC-1MO, and BCAS+RIC-4MO. (b) Time of crossing on a beam at 6 months (6MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. Values are indicated as mean ± SE.a p < 0.0001 vs sham, BCAS+RIC-1MO, and BCAS+RIC-4MO. Using a repeated measures mixed model for the hanging wire test across (c) and (d), (c) cord-wire hanging test for muscular impairment at 4 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. Values are indicated as mean ± SE. a p = 0.0008 vs sham; b p < 0.0001 vs BCAS+RIC-1MO and BCAS+RIC-4MO. (d) Cord-wire hanging test for muscular impairment at 6 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. There is no difference between 1MO and 4MO RIC groups, but both significantly are different than BCAS (sham RIC). Values are indicated as mean ± SE. a p < 0.0001 vs sham; b p = 0.0008 vs BCAS+RIC-1MO; c p = 0.0001 BCAS+RIC-4MO

cular impairment at 6 months (4MO) after sham operation, BCAS, and BCAS+RIC with 1MO and 4MO. There is no difference between 1MO and 4MO RIC groups, but both significantly are different than BCAS (sham RIC). Values are indicated as mean ± SE. a p < 0.0001 vs sham; b p = 0.0008 vs BCAS+RIC-1MO; c p = 0.0001 BCAS+RIC-4MO RIC Improved Muscular/Motor Function Animals subjected to BCAS spent more time to cross a beam in comparison to the sham control animals (Fig. 3a, b). RIC therapy for either 1MO or 4MO significantly reduced the time spent by the animals to cross the beam as compared to BCAS groups. Moreover, muscular impairment or motor impairment was assayed by wire hanging test (Fig. 3c, d). BCAS groups spent significantly less time suspending their body to the hanging wire compared to sham. RIC therapy for 1MO or 4MO significantly increased the hanging wire time compared to the BCAS group.

ared to BCAS groups. Moreover, muscular impairment or motor impairment was assayed by wire hanging test (Fig. 3c, d). BCAS groups spent significantly less time suspending their body to the hanging wire compared to sham. RIC therapy for 1MO or 4MO significantly increased the hanging wire time compared to the BCAS group. RIC Enhanced Angiogenesis and Arteriogenesis RIC therapy significantly increased capillary density, angiogenesis, and arteriogenesis as indicated by increased expression of CD31 and α-SMA, compared to the BCAS group (Fig. 4a–d; Supplemental Figs. 4 and 5). However, there was no significant difference between RIC-1MO and RIC-4MO therapy groups. Moreover, RIC therapy increased the expression and colocalization of pericytes with cerebral blood vessels as indicated by increased expression of PDGFR-B and IB4, respectively (Supplemental Fig. 6A). Chronic treatment also showed to increase the capillary diameter (Supplemental Fig. 6B).Fig. 4 RIC promotes angiogenesis and arteriogenesis. a, c Representative photomicrographs of single immunofluorescence for CD31 (red, (a)) and α-SMA (green, (c)) or double immunofluorescence (Supplemental Figs. 4 and 5 for CD31 and α-SMA with DAPI) for vessels in the striatum (caudoputamen) of each indicated group at 6 months with or without RIC therapy (scale bar for CD31 and α-SMA = 20 μm/20×). b, d The quantitative analysis shows of capillary density at 6 months in each indicated group (N = 4 to 6/group; b p < 0.01 vs sham; a p < 0.001 vs BCAS+RIC-1M/4MO)

API) for vessels in the striatum (caudoputamen) of each indicated group at 6 months with or without RIC therapy (scale bar for CD31 and α-SMA = 20 μm/20×). b, d The quantitative analysis shows of capillary density at 6 months in each indicated group (N = 4 to 6/group; b p < 0.01 vs sham; a p < 0.001 vs BCAS+RIC-1M/4MO) Fig. 5 RIC facilitates cerebrovascular angioarchitecture with 3 weeks treatment after BCAS. (a) Representative 3D images showing the whole cerebrovascular angioarchitecture from a top, side, and bottom view of sham, BCAS, and BCSA+RIC groups of mice brain. Histogram showing vascular volume percentage (b) and number of vessels (c) for linear space between vessels, density, and lumen thickness (Supplemental Fig. 9A–C). N = 8/group, a p < 0.01 vs sham; b p < 0.05 vs BCAS

ar angioarchitecture from a top, side, and bottom view of sham, BCAS, and BCSA+RIC groups of mice brain. Histogram showing vascular volume percentage (b) and number of vessels (c) for linear space between vessels, density, and lumen thickness (Supplemental Fig. 9A–C). N = 8/group, a p < 0.01 vs sham; b p < 0.05 vs BCAS Fig. 6 RIC therapy activates endothelial progenitor cells (EPCs) and increases M2/M1 macrophages in the blood 3 weeks treatment. (a) Flow cytometry graphs showing a significant increase the EPCs count, as indicated by increased expression of CD31, VEGFR2, and CD34. (b) Dropped CBF in BCAS group results in recruitment of macrophages in response to chronic ischemia. RIC therapy decreased the level of inflammatory M1 macrophages while it enhanced the level of anti-inflammatory M2 macrophages in the blood (as indicated by the expression of CD11b, F4/80; CD68, TNFα; and CD206, IL-10). However, a high level of circulating M1 macrophage was counted in BCAS groups, indicating high inflammatory burden. RIC therapy activated circulating EPCs, thus reduced the vascular injury and protects ischemic brain. (c) RIC therapy with 1-MO or 4-MO showed a trend but insignificant increase in plasma nitrite levels at 6 months compared to BCAS group. However, RIC therapy for 3 weeks post BCAS significantly increased the plasma nitrite levels compared to BCAS sham RIC (d). Values are indicated as mean ± SD. a p = 0.0092 vs sham; b p = 0.0044 vs BCAS+RIC

-MO or 4-MO showed a trend but insignificant increase in plasma nitrite levels at 6 months compared to BCAS group. However, RIC therapy for 3 weeks post BCAS significantly increased the plasma nitrite levels compared to BCAS sham RIC (d). Values are indicated as mean ± SD. a p = 0.0092 vs sham; b p = 0.0044 vs BCAS+RIC RIC Reduced White Matter Damage and Myelin Basic Protein (MBP) The white matter degeneration in the corpus callosum was tested by Klüver–Barrera staining. The intensity in Klüver–Barrera staining was significantly reduced after BCAS compared to sham animals. RIC therapy for 1 or 4 months significantly reversed the BCAS-induced white matter damage (Supplemental Fig. 7A–B). MBP staining significantly decreased in cortical and hippocampal CA1 field region (Supplemental Fig. 8A–B), and this was reversed with RIC therapy in both BCAS+RIC1-MO and BCAS+RIC4-MO groups.

am animals. RIC therapy for 1 or 4 months significantly reversed the BCAS-induced white matter damage (Supplemental Fig. 7A–B). MBP staining significantly decreased in cortical and hippocampal CA1 field region (Supplemental Fig. 8A–B), and this was reversed with RIC therapy in both BCAS+RIC1-MO and BCAS+RIC4-MO groups. RIC Improves the Cerebrovascular Angioarchitecture and Increases NO Production The use of BriteVu staining enabled the 3D visualization of the complete tree of the cerebrovasculature. BCAS caused a significant decline in number and volume of cerebral vessels. RIC significantly induced angiogenesis and collaterals formation as indicated by the increase in the vessels number and volume (Fig. 5a–c; Supplemental Fig. 9A–C). In support with this, results from the flow cytometry studies showed a corresponding increase in the EPC count (as indicated by increased expression of CD31 and VEGF-R2) and in macrophage expression and polarization (as indicated by increased expression of CD11b, F4/80, and CD206) with RIC therapy (Fig. 6a, b). RIC therapy for 3 weeks post BCAS significantly increased the plasma nitrite levels compared to BCAS without RIC. However, RIC therapy for 1-MO and 4-MO showed a trend but insignificant increase in plasma nitrite levels at 6 months compared to the BCAS group (Fig. 6c, d). However, plasma nitrite level in the BCAS group was significantly decreased as compared to the sham group. Probably the short half-life of NO and long delay after RIC was finished caused NO levels to decrease.

trend but insignificant increase in plasma nitrite levels at 6 months compared to the BCAS group (Fig. 6c, d). However, plasma nitrite level in the BCAS group was significantly decreased as compared to the sham group. Probably the short half-life of NO and long delay after RIC was finished caused NO levels to decrease. Discussion RIC can be performed before acute cerebral ischemia (preconditioning), during acute ischemia (per-conditioning), or after reperfusion (post conditioning). There is a large body of evidence that acute remote ischemic conditioning is a powerful cerebroprotectant in acute focal cerebral ischemia models [27, 28]. The current study proposes C-RIC as a therapeutic paradigm in chronic mild ischemia. The repetitive use of RIC for weeks or months can be an analogue to long-term daily exercise. There are fewer published studies and less data on C-RIC in chronic cerebral ischemia. C-RIC administered for 6 months was effective in reducing recurrent stroke and TIA in two small randomized clinical trials in patients with intracranial atherosclerosis (ICAS) [29, 30]. There was increased CBF by SPECT in one trial suggesting that C-RIC increased CBF in human patients. There is also a small clinical trial suggesting that C-RIC may reduce progression of white matter disease in patients with VCID. However, we have little preclinical data to support its use clinically, or to better understand its mechanisms. The BCAS mouse model has been offered as a model to test interventions in cerebral small vessel disease and VCID, and we now describe findings that support translation to the bedside.

e in patients with VCID. However, we have little preclinical data to support its use clinically, or to better understand its mechanisms. The BCAS mouse model has been offered as a model to test interventions in cerebral small vessel disease and VCID, and we now describe findings that support translation to the bedside. First, our results show that both 1 month C-RIC (1-MO) and 4 months of C-RIC (4-MO) are equally effective in improving long-term (6 months) CBF, working and spatial memory, and in improving balance and motor skills compared with sham animals. This is a crucial piece of data suggesting that shorter term use of C-RIC (1 month) in patients with chronic ischemia may be as effective as longer term treatment (4 months). However, we exercise caution in extrapolating these findings to human VCID/small vessel disease, as the BCAS model has a known start and onset time. In VCID patients, the course of disease is insidious and there is no precise start time. Nevertheless, in our preclinical model, C-RIC appears to be efficient in improving the long-term CBF and long-term cognitive and motor function, likely by inducing angiogenesis and collateral remodeling. Our findings may have some implications to use C-RIC in ICAS where there is hypoperfusion, albeit from intracranial rather than extracranial stenosis. While the clinical trials in ICAS have used 6 months or 300 days of C-RIC, our results show that shorter durations may be effective.

iogenesis and collateral remodeling. Our findings may have some implications to use C-RIC in ICAS where there is hypoperfusion, albeit from intracranial rather than extracranial stenosis. While the clinical trials in ICAS have used 6 months or 300 days of C-RIC, our results show that shorter durations may be effective. Second, C-RIC for 3 weeks induces cerebral vascular remodeling, angiogenesis, arteriogenesis, and new collateral formation. These structural changes in the angioarchitecture may underlie the improvement in CBF. Beyond the functions of supplying oxygen and nutrients, endothelial cells have trophic functions for neurons and oligodendrocytes. This angiogenesis and improved CBF are likely protective to oligodendrocytes and neurons. This increase in cerebral angiogenesis is similar to what is seen in the rodents’ models of daily physical exercise for 3–4 weeks [31–34]. There are strong parallels between RIC and physical exercise. Both involve shear stress to the vasculature and upregulation of endothelial NOS (eNOS) and increased plasma nitrite [35, 36]. Third, similar to exercise, C-RIC increases circulating EPCs, which may be playing a role in angiogenesis. While there is controversy over whether EPCs incorporate directly into growing vessels or provide trophic effect to vessels, they are associated with angiogenesis and vascular health. Further studies are needed to define their participation in angiogenesis in our model.

PCs, which may be playing a role in angiogenesis. While there is controversy over whether EPCs incorporate directly into growing vessels or provide trophic effect to vessels, they are associated with angiogenesis and vascular health. Further studies are needed to define their participation in angiogenesis in our model. Fourth, C-RIC is associated with increases in plasma nitrite, which suggests a role for increased NO bioavailability in mediating vascular remodeling. While once regarded as an inert molecule and oxidative end product of NO metabolism, research over the last decade has shown that nitrite serves a “storage” pool of NO derived from endogenous eNOS [37–39]. Nitrite circulates in the blood associated with RBC/hemoglobin and is reduced to NO in areas of hypoxemia, mediating hypoxic vasodilatation [38, 40]. While NO is normally limited to a paracrine effect due to its very short half-life, the eNOS/nitrite/NO system provides a distant “endocrine” effect of NO. We showed in this study that increased NO bioavailability is associated with angiogenesis and collateral remodeling. Moreover, NO is associated with mobilization of endothelial progenitor cells.

o a paracrine effect due to its very short half-life, the eNOS/nitrite/NO system provides a distant “endocrine” effect of NO. We showed in this study that increased NO bioavailability is associated with angiogenesis and collateral remodeling. Moreover, NO is associated with mobilization of endothelial progenitor cells. We have demonstrated that C-RIC is effective in promoting angiogenesis and improving long-term functional outcome. The major risk factor for dementia and VCID is age. It is paramount to demonstrate these findings in aged mice of both sexes, and we are conducting series of these studies in our group. C-RIC is a safe and well-tolerated intervention that may be useful in patients with VCID. Further studies are needed to determine the dosing and to develop biomarkers. Both EPCs and plasma nitrite may serve as useful biomarkers. Electronic Supplementary Material ESM 1 (PDF 5671 kb) Abbreviations 1MOOne month 4MOFour months 6MOSix months BCASBilateral carotid artery stenosis CCAsCommon carotid arteries CBFCerebral blood flow RICRemote ischemic conditioning C-RICChronic remote ischemic conditioning GFAPGlial fibrillary acidic protein IFImmunofluorescence IHCImmunohistochemistry LSCILaser Speckle Contrast Imager MBPMyelin basic protein NONitric oxide VCIDVascular contributions to cognitive impairment and dementia WMWhite matter Electronic supplementary material The online version of this article (doi:10.1007/s12975-017-0555-1) contains supplementary material, which is available to authorized users.

IHCImmunohistochemistry LSCILaser Speckle Contrast Imager MBPMyelin basic protein NONitric oxide VCIDVascular contributions to cognitive impairment and dementia WMWhite matter Electronic supplementary material The online version of this article (doi:10.1007/s12975-017-0555-1) contains supplementary material, which is available to authorized users. Compliance with Ethical Standards Funding This work was supported by the NIH/NINDS R21NS090609-01A1. We are special thankful to Mr. Richard Goodman, Hatteras Instruments, Cary, NC, in accepting a proposed design and making a multichannel non-invasive programmable remote ischemic conditioner and for a generous gift to our laboratory. Conflict of Interest The authors declare that they have no conflict of interest. Ethical Approval All applicable institutional and national guidelines for the care and use of animals were followed.

Introduction Brain ischemia is a leading cause of morbidity and mortality [1]. Although thrombolysis for stroke is an accepted treatment, fewer than 5% of stroke patients are treated with tissue plasminogen activator. New therapeutic strategies are needed for stroke patients. While cerebral ischemia induces pro-inflammatory cytokines, resulting in the expansion of brain damage, it stimulates inherent neurogenesis in the subventricular zone (SVZ) and promotes the migration of newly formed neuronal progenitor cells toward the ischemic area [2]. This process may help to repair the brain and reconstruct neural networks. In rodents [3], monkeys [4], and humans [5], the SVZ, located in the lateral wall lining the lateral ventricle, harbors the largest population of neural stem cells capable of generating new neurons, astrocytes, and oligodendrocytes. While enhanced neurogenesis and angiogenesis in the SVZ and subgranular zone have been documented in adult rodent brains after focal ischemia [6], as most of the newly formed cells promptly undergo apoptosis, their effect on functional recovery is limited [2]. The administration of pioglitazone (PGZ), a peroxisome proliferator-activated receptor gamma (PPARγ) agonist, before cerebral ischemia induction protected rodents against ischemic brain damage [7, 8]. Its beneficial effects were associated with a decrease in the expression of interleukin-6 (IL-6) and caspase-3 and the improvement of neurological function in male rats [7]. However, there are few studies regarding the effects of post-ischemic treatment with PGZ.

protected rodents against ischemic brain damage [7, 8]. Its beneficial effects were associated with a decrease in the expression of interleukin-6 (IL-6) and caspase-3 and the improvement of neurological function in male rats [7]. However, there are few studies regarding the effects of post-ischemic treatment with PGZ. As estrogen deficiency and menopause are thought to be major risk factors for cerebral ischemia, in an earlier study [8], we compared the effects of pre-treatment with PGZ on cerebral ischemia in oophorectomized (OVX) and non-OVX rats. We demonstrated that PGZ prevented ischemic brain damage in OVX rats when it was administered before stroke and that this was associated with the upregulation of anti-apoptotic and survival genes via the trans-activation of STAT3 and PPARγ in the peri-infarct region. This suggests its contribution to neurogenesis. However, it remained to be determined whether post-ischemic treatment with PGZ contributes to neurogenesis and neurological improvement after an ischemic insult.

n of anti-apoptotic and survival genes via the trans-activation of STAT3 and PPARγ in the peri-infarct region. This suggests its contribution to neurogenesis. However, it remained to be determined whether post-ischemic treatment with PGZ contributes to neurogenesis and neurological improvement after an ischemic insult. The local release of PGZ promoted wound healing [9]; the regulation of monocytes/macrophages by PGZ inhibited plaque destabilization and rupture in ApoE−/− mice [10], suggesting that post-stroke treatment with PGZ may be useful. Based on findings in other [7, 9, 10] and our studies [8], we hypothesized that the activation of PPARγ elicited by the administration of PGZ after stroke may be beneficial due to not only its anti-inflammatory effects but also its promotion of neurogenesis. As neurogenesis is regulated by the stem cell niche [11], we focused on this issue. To examine whether PGZ is associated with neuro- and angiogenesis, we prepared female rats whose bone marrow (BM) cells were replaced with BM cells from green fluorescent protein-transgenic (GFP+BM) rats before oophorectomy (OVX/GFP+BM rats). We first examined the neuroprotective effects elicited by post-ischemia treatment with PGZ in male and OVX/GFP+BM rats and then investigated its effects on neurogenesis in OVX/GFP+BM rats.

M) cells were replaced with BM cells from green fluorescent protein-transgenic (GFP+BM) rats before oophorectomy (OVX/GFP+BM rats). We first examined the neuroprotective effects elicited by post-ischemia treatment with PGZ in male and OVX/GFP+BM rats and then investigated its effects on neurogenesis in OVX/GFP+BM rats. Here, we show that even post-ischemia, early-phase treatment with PGZ inhibits the infarct size and alleviates neurological deficits in male and in OVX/GFP+BM rats. We demonstrate that the decrease in the pro-inflammatory cytokine IL-6 and in M1-like macrophages associated with an increase in PPARγ after ischemia in male rats. We also document that in OVX/GFP+BM rats, PGZ activated innate stem cells in the SVZ and recruitment of GFP+BM stem cells with an increase in PPARγ and then increased the expression of Akt, MAP2, and VEGF in the cortical peri-infarct area, leading to neurogenesis.

with an increase in PPARγ after ischemia in male rats. We also document that in OVX/GFP+BM rats, PGZ activated innate stem cells in the SVZ and recruitment of GFP+BM stem cells with an increase in PPARγ and then increased the expression of Akt, MAP2, and VEGF in the cortical peri-infarct area, leading to neurogenesis. Materials and Methods Our study was approved by the Ethics Committee of the Institute of Biomedical Sciences, Tokushima University Graduate School, and conducted in accordance with current RIGOR guidelines [12, 13] and the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. We purchased male and female Wistar rats from Charles River Laboratories Japan Inc. (Yokohama, Japan). They were housed in conventional rat cages in a temperature- and humidity-controlled room (about 23 °C and 50%, respectively) under a 12-h inverted light/dark cycle and were fed standard chow. Anesthesia was with 2% isofluorane in 30% oxygen and 70% nitrous oxide. The experiments were reported according to the “Animal Research: Reporting of In Vivo Experiments (ARRIVE)” guidelines to improve the design, analysis, and reporting of research using animals—maximizing information published and minimizing unnecessary studies. All procedures were performed by investigators blinded to the treatment that rats had undergone according to a protocol approved by the Animal Care Committee of Tokushima University Medical School. For randomization, we used random-number tables; the treatment groups were identified by ear punch and the cages were labeled.

procedures were performed by investigators blinded to the treatment that rats had undergone according to a protocol approved by the Animal Care Committee of Tokushima University Medical School. For randomization, we used random-number tables; the treatment groups were identified by ear punch and the cages were labeled. Whole-Body Irradiation and BM Transplantation in Female Rats The BM donors were 6-week-old GFP-transgenic rats purchased from Japan SLC, Inc. (Hamamatsu, Japan). They were killed under deep anesthesia by cervical dislocation. BM was obtained by flushing the femora and tibiae with sterile phosphate-buffered saline (PBS). BM cells were suspended in PBS, washed several times, counted, and resuspended at 4 × 107 cells/ml. All procedures were as described elsewhere [14]. At the age of 7 weeks, the Wistar female rats underwent whole-body irradiation with a single 10-Gy dose. Within 24 h, thereafter, they were injected with 300 μl of the GFP-labeled BM cell suspension via the tail vein.

Whole-Body Irradiation and BM Transplantation in Female Rats The BM donors were 6-week-old GFP-transgenic rats purchased from Japan SLC, Inc. (Hamamatsu, Japan). They were killed under deep anesthesia by cervical dislocation. BM was obtained by flushing the femora and tibiae with sterile phosphate-buffered saline (PBS). BM cells were suspended in PBS, washed several times, counted, and resuspended at 4 × 107 cells/ml. All procedures were as described elsewhere [14]. At the age of 7 weeks, the Wistar female rats underwent whole-body irradiation with a single 10-Gy dose. Within 24 h, thereafter, they were injected with 300 μl of the GFP-labeled BM cell suspension via the tail vein. Animals Females received BM cells from GFP rats who were subjected to bilateral OVX at the age of 10 weeks (OVX/GFP+BM rats). Based on our and other earlier studies [7, 8], 7-week-old male and 13-week-old OVX/GFP+BM Wistar rats weighing 250–270 g and 280–300 g, respectively, were subjected to 90-min middle cerebral artery occlusion-reperfusion (MCAO-R). One group of randomly selected rats was injected intraperitoneally (i.p.) with 2.5 mg/kg/day PGZ; the other was the vehicle control (VC). In male and OVX/GFP+BM rats, we recorded the effects of PGZ delivered in the early phase after MCAO-R (day 0) and once a day for 7 (males) and 14 consecutive days (OVX/GFP+BM rats).

One group of randomly selected rats was injected intraperitoneally (i.p.) with 2.5 mg/kg/day PGZ; the other was the vehicle control (VC). In male and OVX/GFP+BM rats, we recorded the effects of PGZ delivered in the early phase after MCAO-R (day 0) and once a day for 7 (males) and 14 consecutive days (OVX/GFP+BM rats). Besides examining the response to early-phase post-ischemia treatment with PGZ in OVX/GFP+BM rats, we also assessed its effect on post-MCAO-R regeneration on days 7–14. The OVX/GFP+BM rats were randomly sub-divided into two groups: one group received 2.5 mg/kg PGZ i.p. once a day for 14 consecutive days after MCAO-R, the other served as the VC. PGZ, a gift from Takeda Pharmaceutical Co., was dissolved in dimethylsulfoxide (DMSO) and diluted (× 3) with saline just before i.p. injection (0.4 ml/kg body weight). VC rats were injected DMSO dilution at the same concentration and volume as PGZ. Focal Cerebral Ischemia During all surgical procedures, the rats were under anesthesia with 2% isofluorane in 30% oxygen and 70% nitrous oxide; their rectal temperature was monitored with a thermometer (KN-91, Natsume) and maintained at 37 ± 0.5 °C with a warming plate.

Besides examining the response to early-phase post-ischemia treatment with PGZ in OVX/GFP+BM rats, we also assessed its effect on post-MCAO-R regeneration on days 7–14. The OVX/GFP+BM rats were randomly sub-divided into two groups: one group received 2.5 mg/kg PGZ i.p. once a day for 14 consecutive days after MCAO-R, the other served as the VC. PGZ, a gift from Takeda Pharmaceutical Co., was dissolved in dimethylsulfoxide (DMSO) and diluted (× 3) with saline just before i.p. injection (0.4 ml/kg body weight). VC rats were injected DMSO dilution at the same concentration and volume as PGZ. Focal Cerebral Ischemia During all surgical procedures, the rats were under anesthesia with 2% isofluorane in 30% oxygen and 70% nitrous oxide; their rectal temperature was monitored with a thermometer (KN-91, Natsume) and maintained at 37 ± 0.5 °C with a warming plate. For 90-min MCAO, we inserted an intraluminal filament as described elsewhere [15, 16]. To block major collateral flow, the pterygopalatine artery was ligated at its origin. The internal and common carotid artery were transiently occluded with loosely tied 3–0 silk sutures; a silicon-coated 4–0 nylon thread was introduced into the external carotid artery and advanced into the internal carotid artery to occlude the proximal orifice of the MCA. To confirm MCAO, we used a laser-Doppler flow probe (Unique Medical, Osaka, Japan) to measure the blood flow at the temporal bone surface at a site 1 mm posterior to the bregma and 3 mm inferior to the temporal line. MCAO reduced the blood flow to 20–30% of the baseline.

tery to occlude the proximal orifice of the MCA. To confirm MCAO, we used a laser-Doppler flow probe (Unique Medical, Osaka, Japan) to measure the blood flow at the temporal bone surface at a site 1 mm posterior to the bregma and 3 mm inferior to the temporal line. MCAO reduced the blood flow to 20–30% of the baseline. Rats with successful MCAO-R consistently exhibited circling behavior, decreased resistance to lateral push, forelimb flexion, and shoulder adduction. We excluded around 10% of the rats because MCAO-R was incomplete. Blood glucose levels were determined in whole venous blood with an automatic glucose meter (Accu-check Aviva blood glucose meter, Roche Diagnostics, Tokyo, Japan). The blood pressure was measured by telemetry (Data Science Inc., MN55126, USA) before, during, and after MCAO-R and recorded using the Dataquest Advanced Research Technologies Acquisition program (Unique Medical).

d with an automatic glucose meter (Accu-check Aviva blood glucose meter, Roche Diagnostics, Tokyo, Japan). The blood pressure was measured by telemetry (Data Science Inc., MN55126, USA) before, during, and after MCAO-R and recorded using the Dataquest Advanced Research Technologies Acquisition program (Unique Medical). Measurement of the Infarct Volume The brains were extracted and equal 2-mm-spaced slices and six coronal blocks were prepared immediately using a brain matrix (Bioresearch Center, Nagoya, Japan). The samples did not contain olfactory tissue or tissue from the cerebellum. All but the 3rd coronal sections were immersed in a 2,3,5-triphenyltetrazolium chloride (TTC) solution in PBS to detect the infarct area. We identified the area surrounding the infarct area in the frontal cortex as the peri-infarct area. The tissue samples were stored at − 80 °C until Western blot analysis and determination of the mRNA level. The extent of ischemic infarction was traced manually, and the integrated volume was calculated using NIH 1.36b Image J software. Artifacts from brain edema were eliminated by applying the indirect measurement method based on the contralateral brain volume.

til Western blot analysis and determination of the mRNA level. The extent of ischemic infarction was traced manually, and the integrated volume was calculated using NIH 1.36b Image J software. Artifacts from brain edema were eliminated by applying the indirect measurement method based on the contralateral brain volume. Neurological Assessment Neurological deficits were assessed by an examiner blinded to the treatment the rats had undergone. We modified the neurological scoring system of Huang et al. [17] and Chen et al. [18] and recorded our findings as 0 = normal; 1 = forelimb or hindlimb flexion, head turned > 10 to the vertical axis within 30 s after raising the rat by the tail, inability to walk straight on the floor, and grasping the side of the beam during the beam balance test; 2 = circling toward the paretic side, one limb falling off the beam; 3 = falling to the paretic side, hugging the beam, and two limbs falling off the beam; 4 = attempting to balance on the beam but falling off (> 40 s); 5 = attempting to keep balance on the beam but falling off (> 20 s); and 6 = falling off the beam without attempting to balance or hang on to the beam (< 20 s). The rats were evaluated immediately after successful MCAO, 24 h after MCAO-R, and again on days 1, 3, and 7 (males) and on days 1, 3, 7, and 14 (OVX/GFP+BM rats) after treatment with PGZ or VC. The total maximum score was 12.

); and 6 = falling off the beam without attempting to balance or hang on to the beam (< 20 s). The rats were evaluated immediately after successful MCAO, 24 h after MCAO-R, and again on days 1, 3, and 7 (males) and on days 1, 3, 7, and 14 (OVX/GFP+BM rats) after treatment with PGZ or VC. The total maximum score was 12. Quantitative Real-Time PCR Total RNA obtained from the peri-infarct area was isolated with the BioRobot EZ1 and EZ1 universal tissue kit (Qiagen, Tokyo, Japan). RNA was converted to cDNA using the transcript first-strand cDNA synthesis kit (Qiagen). Quantitative real-time PCR assay of each sample was on Light Cycler FastStart DNA Master SYBR Green I and Roche LightCycler 2.0 (Roche Diagnostics, Tokyo, Japan) instruments. Primers for GAPDH were from Roche and used according to the manufacturer’s directions. The other primers were: for rat IL-6, forward (F), 5′-TCT CAG GGA GAT CTT GGA AAT G-3′, reverse (R), 5′-TAG AAA CGG AAC TCC AGA AGA C-3′; for rat TNF-α, (F), 5′-CCC AAC AAG GAG GAG AAG T-3′, (R), 5′-CGC TTG GTG GTT TGC TAC-3′; for rat IL-1β, (F), 5′-TGC AGG CTT CGA GAT GAA C-3′ (R), 5′-AGC TCA TGG AGA ATA CCA CTT G-3′; for rat VEGF, (F), 5′-CACATAGGAGAGATGAGCTT-3′, (R), 5′-CTGGCTTTG TTCTATCTTTC-3′. The amplified product was separated on 1.5% agarose gels containing EtBr solution (Wako, Osaka, Japan) and visualized on an ultraviolet transilluminator. The results were normalized to the expression of GAPDH mRNA.

GG AGA ATA CCA CTT G-3′; for rat VEGF, (F), 5′-CACATAGGAGAGATGAGCTT-3′, (R), 5′-CTGGCTTTG TTCTATCTTTC-3′. The amplified product was separated on 1.5% agarose gels containing EtBr solution (Wako, Osaka, Japan) and visualized on an ultraviolet transilluminator. The results were normalized to the expression of GAPDH mRNA. Immunohistochemistry The rats were transcardially perfused with 4% paraformaldehyde in PBS on ice. Their brains were fixed and 6-μm-thick frozen sections were mounted on Matsunami adhesive silane (MAS)-coated glass slides (Matsunami Glass, Tokyo, Japan), blocked with serum-free protein (DakoCytomation), and then the slides were incubated with primary antibodies diluted with Canget signal immunostain (Toyobo, Osaka, Japan). The antibodies were rabbit polyclonal antibody against PPARγ (Abcam, Tokyo, Japan), GFAP, Nestin, MAP2, (Santa Cruz Biotechnology), and antibodies against GFP (Cell Signaling Technology). We used mouse monoclonal antibody against neuronal nucleus (NeuN) (Millipore, Tokyo, Japan), Musashi-1 (AbD Serotec), CD16, CD68 (Santa Cruz Biotechnology), CD31 (MAB1393, CHEMICON, MA), caspase-3, GFP, and Nestin (Cell Signaling Technology). The tissue samples were mounted with Vectashield (Vector Laboratories Inc., Burlingame, CA). Visualization was with Alexa Fluor 594 donkey anti-rabbit IgG or 488 goat anti-mouse IgG (Molecular Probes, Eugene, OR); the slides were examined under a fluorescence microscope (KEYENCE, BZ-X710, Osaka, Japan). To examine the specificity of immunoreactivity, the primary antibody was omitted to provide a nonspecific control. A parallel set of tissue sections was subjected to hematoxylin and eosin staining to identify the infarct core and the peri-infarct region. We counted all cells positive for Musashi-1 and GFP in the peri-infarct area. Images were captured at ×10 magnification under the microscope. In each animal, we randomly selected two areas containing positive cells in 150 × 150-μm fields around the peri-infarct area. Tissue samples from four rats in each group were analyzed by densitometry of the positive cells using BZH-3A (KEYENCE) analysis software.

were captured at ×10 magnification under the microscope. In each animal, we randomly selected two areas containing positive cells in 150 × 150-μm fields around the peri-infarct area. Tissue samples from four rats in each group were analyzed by densitometry of the positive cells using BZH-3A (KEYENCE) analysis software. Bromodeoxyuridine (BrdU) Labeling in OVX/GFP+BM Rats We recorded the time course of proliferating cells in the brain after cerebral ischemia by pulse-labeling. BrdU, a thymidine analogue that is incorporated into the DNA of dividing cells during the S-phase, was used for mitotic labeling (Sigma Chemical) [19]. We applied a cumulative labeling method to examine the population of proliferating cells in OVX rats exposed for 14 days to cerebral ischemia and injected BrdU (50 mg/kg, i.p.) every 4 h for 12 h before killing the rats on the 1st, 3rd, 7th, or 14th day after the induction of cerebral ischemia (n = 6 each). BrdU-positive cells were detected immunohistochemically using sheep polyclonal antibody BrdU (LifeSpan BioSciences Inc.) as the primary and donkey anti-sheep IgG (Molecular Probes, Eugene, OR) as the secondary antibody.

ng the rats on the 1st, 3rd, 7th, or 14th day after the induction of cerebral ischemia (n = 6 each). BrdU-positive cells were detected immunohistochemically using sheep polyclonal antibody BrdU (LifeSpan BioSciences Inc.) as the primary and donkey anti-sheep IgG (Molecular Probes, Eugene, OR) as the secondary antibody. Western Blot Analysis in OVX/GFP+BM Rats Brain tissue in the peri-infarct area was homogenized and sonicated in RIPA buffer (Thermo Scientific, Rockford, IL) containing phosphatase and protease inhibitors (Roche, Tokyo, Japan) and centrifuged. Total protein in the supernatant was measured with the BCA protein assay kit (Pierce, Rockford, IL). Protein was separated by 7.5 or 12% SDS-PAGE and transferred to a polyvinylidenedifluoride membrane. After blocking with 5% skim milk or BSA in Tris-buffered saline solution-Tween 20 (T-TBS), the membrane was incubated with the primary antibodies in Canget signal immunostain or T-TBS. The same primary antibodies as used for the immunohistochemical studies, rabbit polyclonal antibody against p-Akt (Cell Signaling Technology), MAP2, and mouse monoclonal anti-β-actin (Sigma, Tokyo Japan) were used. After incubation with horseradish peroxidase-conjugated secondary antibodies (GE Healthcare, Buckinghamshire, UK), signals were detected by chemiluminescence using an ECL-plus kit (GE Healthcare). Images were analyzed with Image Quant LAS 4000 mini (GE Healthcare) and Image J software and quantified as the relative increase over the controls after normalization with β-actin.

econdary antibodies (GE Healthcare, Buckinghamshire, UK), signals were detected by chemiluminescence using an ECL-plus kit (GE Healthcare). Images were analyzed with Image Quant LAS 4000 mini (GE Healthcare) and Image J software and quantified as the relative increase over the controls after normalization with β-actin. Statistical Analysis Power estimates were calculated based on α = 0.05, 1-β = 0.8, and a surgery-related drop-out rate of around 10% to obtain group sizes appropriate for detecting an effect size of 0.4 based on a preliminary experiment using G*Power 3.1. The infarct volume and the neurological score were analyzed with Student’s t test and Man-Whitney U test, respectively; their correlation was assessed by the Spearman’s rank-correlation coefficient. The mRNA expression levels were determined with analysis of variance (ANOVA) followed by Scheffe’s test for three-group comparisons. Statistical analyses were performed using IBM SPSS Statistics 22. Data are shown as the mean ± SD. Differences were considered statistically significant at p < 0.05.

relation coefficient. The mRNA expression levels were determined with analysis of variance (ANOVA) followed by Scheffe’s test for three-group comparisons. Statistical analyses were performed using IBM SPSS Statistics 22. Data are shown as the mean ± SD. Differences were considered statistically significant at p < 0.05. Results Male Rats: Treatment with PGZ in the Early Post-ischemia Phase Reduced the Cerebral Infarct Size and Ameliorated Neurological Deficits by Inhibiting Pro-inflammatory Responses We first assessed the effects of PGZ treatment in the early phase after experimental cerebral ischemia. Compared to the VC males, on days 1–7 after the ischemic insult, PGZ rats manifested a lower neurological score and earlier recovery of the body weight loss (Fig. 1a, b). The infarct volume was smaller than in VC rats (Fig. 1c) and correlated with the neurological score (Fig. 1d). Next, to address the mechanisms underlying the effects of PGZ, we examined its anti-inflammatory effects against brain ischemic injury. In VC rats, the mRNA level of IL-6, IL-β, and TNFα was increased 3 h after ischemia induction and augmented at 24 h (Fig. 2a–c). In PGZ rats, the mRNA level of IL-6 (Fig. 2a) but not of IL-β and TNFα was significantly decreased at 3 and 24 h (Fig. 2b, c). Immunohistochemically, the expression of PPARγ was higher in PGZ than VC rats; CD16- and CD68-positive cells were fewer and the expression of caspase-3 was lower (Fig. 2d). PPARγ was localized in CD31-positive cells. CD16-positive cells were Iba-1 or CD68 positive. Treatment with PGZ in the early phase post-ischemia appeared to exert beneficial effects through anti-apoptosis and anti-inflammatory response effects elicited by the expression of PPARγ.Fig. 1 Effects of PGZ against brain damage in male rats. Post-ischemia treatment with PGZ or vehicle was performed immediately after MCAO induction on day 0 and once a day for 7 consecutive days. The neurological scores (a), body weight (b), and infarct volume (c) were recorded in PGZ- and vehicle-treated male rats. The infarct volume was recorded as a percentage of the contralateral hemisphere using Image J software (each group n = 12). The correlation between the infarct volume and neurological deficits was assessed by Spearman’s rank-correlation coefficient (d). Each bar represents the mean ± SD. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. MCAO-R middle cerebral artery occlusion-reperfusion, VC vehicle control, PGZ 2.5 mg/kg pioglitazone

. The correlation between the infarct volume and neurological deficits was assessed by Spearman’s rank-correlation coefficient (d). Each bar represents the mean ± SD. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. MCAO-R middle cerebral artery occlusion-reperfusion, VC vehicle control, PGZ 2.5 mg/kg pioglitazone Fig. 2 The mRNA level of pro-inflammatory cytokines and representative immunohistochemistry findings in PGZ- and VC-treated male rats. The mRNA level of the pro-inflammatory cytokines IL-6 (a), TNFα (b), and IL-1β (c) was assessed by quantitative real-time PCR assay and normalized by GAPDH. Data obtained 3 and 24 h after MCAO-R are shown. Data are the mean ± SD from eight rats per group. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. The expression of PPARγ, CD16, CD68, and caspase-3 was examined in the peri-infarct region of PGZ-treated and VC rats 24 h after MCAO (d). PPARγ- and CD16-positive cells were co-localized with VEGF-, Iba-1- or CD68-positive cells

mean ± SD from eight rats per group. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. The expression of PPARγ, CD16, CD68, and caspase-3 was examined in the peri-infarct region of PGZ-treated and VC rats 24 h after MCAO (d). PPARγ- and CD16-positive cells were co-localized with VEGF-, Iba-1- or CD68-positive cells OVX/GFP+BM Rats: Increased PPARγ Expression Elicited by PGZ Was Associated with Neuroprotection and the Activation of Resident Stem Cells in the SVZ and the Recruitment of BM Cells Our earlier study in ovariectomized (OVX) rats treated with PGZ before ischemia induction suggested neurogenesis in the post-ischemic phase [8]. To examine the role in neuroprotection and neurogenesis of early-phase PGZ treatment after ischemia induction, we used OVX/GFP+BM rats whose BM cells were replaced by BM cells from GFP rats. While 75% of the VC-OVX/GFP+BM rats died within 14 days after MCAO-R, all OVX/GFP+BM rats treated with PGZ (PGZ rats) survived (n = 12 in each group). On days 1, 3, 7, and 14 after MCAO-R, the cerebral infarct volume was significantly smaller in PGZ than VC rats (p < 0.05, Fig. 3a) and their neurological deficits were significantly less pronounced (p < 0.05, Fig. 3b), indicating neuroprotection by early-phase PGZ treatment. There was no significant difference in the cerebral blood flow, blood glucose levels, and blood pressure between the PGZ and VC rats during and after MCAO-R (data not shown).Fig. 3 Effects of post-ischemia treatment with PGZ in OVX/GFP+BM rats. PGZ or vehicle (VC) was administered immediately after MCAO induction on day 0 and once a day for 14 consecutive days. The infarct volume (a) was recorded in OVX/GFP+BM rats treated with PGZ or VC (each group n = 12). The neurological score (b) was assessed as described in Materials and Methods. Each bar represents the mean ± SD *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. Representative immunohistochemistry findings for PPARγ, Musashi-1, and GFP (d) in the SVZ after MCAO-R

M rats treated with PGZ or VC (each group n = 12). The neurological score (b) was assessed as described in Materials and Methods. Each bar represents the mean ± SD *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. Representative immunohistochemistry findings for PPARγ, Musashi-1, and GFP (d) in the SVZ after MCAO-R Compared to the contralateral non-ischemic side of VC rats, on days 3 and 7 post-MCAO-R, the expression of PPARγ and of Musashi-1- and GFP-positive cells was increased in the SVZ on the infarct side of PGZ rats (Fig. 3c). The expression of PPARγ in the SVZ seemed to be associated with an increase in innate stem cells and the recruitment of allogeneic BM-derived stem cells.

ats, on days 3 and 7 post-MCAO-R, the expression of PPARγ and of Musashi-1- and GFP-positive cells was increased in the SVZ on the infarct side of PGZ rats (Fig. 3c). The expression of PPARγ in the SVZ seemed to be associated with an increase in innate stem cells and the recruitment of allogeneic BM-derived stem cells. OVX/GFP+BM Rats: PPARγ Activation Promoted the Proliferation of Stem Cells in the SVZ and their Translocation to the Peri-infarct Region As shown in Fig. 4a, on the 7th day after ischemia induction, nestin-positive cells were detected in the SVZ and the cortical peri-infarct area of PGZ rats; the cells seemed to migrate from the SVZ to the peri-infarct region. The number of translocated Musashi-1-positive- and of GFP-positive cells was significantly higher in PGZ than VC rats (p < 0.05, Fig. 4b) and associated with the increase in NeuN- and GFAP-positive cells. These cell populations seemed to include both resident stem cells in SVZ and recruitment of BM-derived stem cells replaced by allogeneic BM cells. We immunohistochemically assessed the incorporation of the cell proliferation marker BrdU in the stem cells. The presence of many BrdU-labeled Musashi-1- and GFP-positive cells was evidence for the proliferation of stem cells in the peri-infarct region (Fig. 4c). PPARγ activation may promote not only the proliferation of stem cells in the SVZ but also the migration of proliferated stem cells from the SVZ to the cortical peri-infarct region.Fig. 4 Stem cells migration from the SVZ into the cortical peri-infarct region promoted by PGZ in OVX/GFP+BM rats. Representative nestin-positive cells in the SVZ and the cortical peri-infarct region on day 7 after ischemia induction in PGZ-treated OVX/GFP+BM rats (a). Cells positive for Musashi-1, GFP, NeuN, and GFAP were detected in the infarct- and the peri-infarct area (b). Musashi-1- and GFP-positive cells were counted in 150 × 150-μm fields. The total cell number (n) in each area was 140–160 cells for DAPI. Each bar represents the mean ± SD from two areas in each of four rats. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. Representative Musashi-1- and GFP-positive cells incorporated BrdU (c)

d GFP-positive cells were counted in 150 × 150-μm fields. The total cell number (n) in each area was 140–160 cells for DAPI. Each bar represents the mean ± SD from two areas in each of four rats. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test. Representative Musashi-1- and GFP-positive cells incorporated BrdU (c) OVX/GFP+BM Rats: Proliferated Neuronal Stem Cells Differentiated into Blood Vessels, Neurons, and Glia in the Cortical Peri-infarct Region Via the Upregulation by PGZ of the Survival Signal Pathway Lastly, we examined the effects of PPARγ activation by PGZ on neurogenesis. Some of the Musashi-1- and GFP-positive cells were NeuN, GFAP, or CD-31 positive (Fig. 5a, b). BrdU-labeled cells were also NeuN, GFAP, or CD-31 positive and some of them were mature, and seen in neurons, glia, and blood vessels. Extended axons were also observed (Figs. 5a–c, 6), suggesting that PGZ promoted the differentiation into mature neurons, glia, and blood vessels in the cortical peri-infarct region.Fig. 5 Stem cells differentiation into blood vessels, neurons, and glia via the upregulation of the survival signal pathway induced by PGZ in OVX/GFP+BM rats. In the peri-infarct region, double immunohistochemical staining for Musashi-1 (a), GFP (b), and BrdU (c) shows co-localization with NeuN, GFAP, and CD31, respectively. The mRNA level of VEGF (d) was analyzed by quantitative PCR. The protein expression of VEGF, p-Akt, and MAP2 in the peri-infarct region (e) was detected on Western blots and analyzed using LAS 4000 and Image J software. Data are the mean ± SD from eight rats per group. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test

e mRNA level of VEGF (d) was analyzed by quantitative PCR. The protein expression of VEGF, p-Akt, and MAP2 in the peri-infarct region (e) was detected on Western blots and analyzed using LAS 4000 and Image J software. Data are the mean ± SD from eight rats per group. *p < 0.05 vs. VC by ANOVA followed by Scheffe’s test Fig. 6 Schematic role of PGZ administered in the early phase after cerebral ischemia. Activation of PPARγ by PGZ inhibits pro-inflammatory responses and promotes neurogenesis in the peri-infarct region. Some NeuN (green)-, GFAP (green)-, or CD31 (red)-positive cells were mature neurons, extended axons, glias, and blood vessels Examination of the gene and protein expression of survival- and proliferation-related molecules elicited by treatment with PGZ showed that on days 7 and 14, the mRNA level of VEGF was increased in the peri-infarct region (p < 0.05, Fig. 5d). Consequently, the expression of VEGF, MAP2, and p-Akt was significantly higher in PGZ than VC rats (p < 0.05, Fig. 5e). These findings suggest that the upregulation by PGZ of survival signaling pathways was associated with the proliferation of neuronal stem cells and their differentiation into mature neurons, glia, and blood vessels.

he expression of VEGF, MAP2, and p-Akt was significantly higher in PGZ than VC rats (p < 0.05, Fig. 5e). These findings suggest that the upregulation by PGZ of survival signaling pathways was associated with the proliferation of neuronal stem cells and their differentiation into mature neurons, glia, and blood vessels. Discussion We first demonstrate that post-ischemia treatment with PGZ in early phase prevented the expansion of brain injury and attenuated neurological deficits in male and OVX/GFP+BM rats. We showed that the neuroprotective effects of PGZ in the early post-ischemia phase were associated with an anti-inflammatory response involving a decrease in IL-6 and M1-type macrophages in male rats. Furthermore, we found that in OVX/GFP+BM rats PGZ promoted the proliferation of both innate stem cells in the SVZ and BM-derived stem cells and their translocation from the SVZ into the cortical peri-infarct region on days 7–14 of post-ischemia. This was associated with the high expression of VEGF, MAP2, and p-AKT and an increase in PPARγ in the peri-infarct area and the differentiation of stem cells into mature cells, resulting in the alleviation of neurological dysfunction. Our findings suggest that early-phase treatment with PGZ, even after experimentally induced cerebral ischemia, contributed to a good stroke outcome in rats via its anti-inflammatory effects and the elicitation of neurogenesis (Fig. 6).

m cells into mature cells, resulting in the alleviation of neurological dysfunction. Our findings suggest that early-phase treatment with PGZ, even after experimentally induced cerebral ischemia, contributed to a good stroke outcome in rats via its anti-inflammatory effects and the elicitation of neurogenesis (Fig. 6). Lambertsen et al. [20] who studied experimental and human stroke, reported that within 1–6 h post-ischemia, the protein level of IL-1β and TNFα but not of IL-6 was increased in the brain and cerebrospinal fluid and persisted at high levels until 24 h. In our animal study, the mRNA level of IL-6, TNFα, and IL-1β was increased 24 h after ischemia induction; this was associated with the presence of M1-like macrophages and the expansion of cerebral infarction although the mRNA level of IL-6 was low during the first 3 h. PGZ suppressed the expression of IL-6 at 24 h without affecting TNFα and IL-1β. During the early post-ischemic phase in male rats, the different effects of PGZ against these molecules may be associated with their different expression profile after ischemia and the timing of PGZ treatment. Nakashiro [10] reported that PGZ decreased the expression of circulating inflammatory monocytes. As PGZ inhibited the expression of CD16+/CD68+ M1-like macrophages in our rats, we think that in the early phase of post-ischemia, it inhibited the recruitment of M1-like monocyte-derived macrophages from the perivascular area into the brain. According to Patzer et al. [21], IL-6-expressing microglia/macrophages in the brain were activated in the initial post-ischemic stage and immunohistochemical and Western blot analysis showed that PGZ reduced the expression of IL-6. Thus, the down-regulation of IL-6 in the early post-ischemic phase may be crucial for a reduction in ischemic brain damage.

-6-expressing microglia/macrophages in the brain were activated in the initial post-ischemic stage and immunohistochemical and Western blot analysis showed that PGZ reduced the expression of IL-6. Thus, the down-regulation of IL-6 in the early post-ischemic phase may be crucial for a reduction in ischemic brain damage. Although we do not know how PGZ increases the expression of PPARγ in areas adjacent to the SVZ, elsewhere [8], we demonstrated that treatment with PGZ before ischemia induction upregulated STAT3 and increased the expression of anti-apoptotic bcl-2 and VEGF. IL-6 is associated with the JAK/STAT pathway and elicits a pro-inflammatory response in the early phase; in the late phase, it is associated with neurotrophic effects [22]. The regulation of IL-6 by PGZ may exert ambivalent beneficial effects in the early and late phase after stroke.

n of anti-apoptotic bcl-2 and VEGF. IL-6 is associated with the JAK/STAT pathway and elicits a pro-inflammatory response in the early phase; in the late phase, it is associated with neurotrophic effects [22]. The regulation of IL-6 by PGZ may exert ambivalent beneficial effects in the early and late phase after stroke. Neuronal stem cells are primordial, multipotent, self-renewing cells that give rise to differentiated progeny within all neuronal and glial lineages. They continue to produce new neurons throughout life in the SVZ and the dentate gyrus of the hippocampus [19, 23]. However, most of the newly formed cells promptly undergo apoptosis, possibly due to unfavorable conditions after cerebral ischemia and the lack of adequate trophic support [6]. In contrast, PGZ-induced stem cells seem to be less affected by ischemic conditions. The increased expression of MAP2 and p-Akt elicited by PGZ and the upregulation of anti-apoptotic genes [8] might contribute to the activation of survival signaling pathways to nourish the stem cells. Thus, the activation of PPARγ by PGZ after the ischemic insult may not only accelerate the proliferation of resident stem cells in the SVZ and the recruitment of BM-derived stem cells but may also contribute to their differentiation in the peri-infarct area.

ion of survival signaling pathways to nourish the stem cells. Thus, the activation of PPARγ by PGZ after the ischemic insult may not only accelerate the proliferation of resident stem cells in the SVZ and the recruitment of BM-derived stem cells but may also contribute to their differentiation in the peri-infarct area. Neurotrophic factors regulate the survival, proliferation, and differentiation of cells in the central nervous system [6, 23]. VEGF, identified as an angiogenic and vascular permeability factor, is recognized as a neurotrophic factor [24] whose role depends on its temporal and spatial profiles. During the acute phase of ischemic injury, the upregulation of VEGF in cerebral vessels increases the permeability of the blood-brain-barrier, thereby exacerbating ischemic cell damage [25]. At a later stage after stroke, it triggers angiogenesis, promotes the blood supply to the brain, and accelerates neurogenesis [26]. Despite the high expression of PPARγ elicited by PGZ in both the early and late phase after the ischemic insult imposed on our rats, the VEGF expression pattern was different in these phases. The upregulation of VEGF on days 7 and 14 may have contributed to angiogenesis. The phosphatidylinositol-3 kinase (PI3)/Akt signaling pathway plays a central role in regulating the growth, proliferation, and survival of cells under physiological and pathophysiological conditions. Its activation protects vascular function [27], promotes cell survival, and suppresses apoptosis. As did others [28, 29], we found that treatment with PGZ even after an ischemic insult increased the expression of p-Akt and MAP2 in the peri-infarct region. Therefore, elicitation of angiogenesis and neurogenesis may be attributable to the upregulation of survival signaling pathways by PGZ.

, and suppresses apoptosis. As did others [28, 29], we found that treatment with PGZ even after an ischemic insult increased the expression of p-Akt and MAP2 in the peri-infarct region. Therefore, elicitation of angiogenesis and neurogenesis may be attributable to the upregulation of survival signaling pathways by PGZ. Although attempts to promote neurogenesis by injecting exogenous stem cells have remained unsuccessful, the combined delivery of PGZ and exogenous stem cells may help to promote neurogenesis. We need further studies to determine whether PGZ helps to promote the survival and proliferation of exogenous stem cells. Kernan et al. [30] reported that 4.8-year PGZ treatment of patients with insulin resistance, ischemic stroke, and transient ischemic attacks reduced the risk for recurrent stroke or myocardial infarcts but increased the risk for weight gain, edema, and bone fracture. Therefore, the condition of patients treated with PGZ must be monitored carefully, and its administration should be short term.

sulin resistance, ischemic stroke, and transient ischemic attacks reduced the risk for recurrent stroke or myocardial infarcts but increased the risk for weight gain, edema, and bone fracture. Therefore, the condition of patients treated with PGZ must be monitored carefully, and its administration should be short term. In conclusion, we provide new evidence that PGZ treatment in the early phase after an experimentally induced ischemic insult ameliorated neurological dysfunction and suppressed the infarct size in male and OVX/GFP+BM rats. We show that its beneficial effects were associated with the induction of anti-inflammatory responses in the early and with the elicitation of angiogenesis and neurogenesis in the late post-ischemic phase. Treatment with PGZ in the early post-ischemic phase may help to limit ischemic brain damage and to alleviate neuronal deficits. We are continuing to assess the potential role of PGZ in the activation of post-stroke PPARγ because such findings may lead to treatments that improve the outcomes in stroke patients. Funding This work was supported by a Grant-in-Aid for Scientific Research [JSPS KAKENHI Grant Number JP15K10306] and a Grant-in-Aid for the Strategic Young Researcher Overseas Visits Program for Accelerating Brain Circulation from the Japan Society for the Promotion of Science [JSPS Grant Number JPS2407]. The original version of this article was revised: The second author (Keiko T. Kitazato) was missing and was added to the list of authors. A correction to this article is available online at https://doi.org/10.1007/s12975-017-0589-4.

Funding This work was supported by a Grant-in-Aid for Scientific Research [JSPS KAKENHI Grant Number JP15K10306] and a Grant-in-Aid for the Strategic Young Researcher Overseas Visits Program for Accelerating Brain Circulation from the Japan Society for the Promotion of Science [JSPS Grant Number JPS2407]. The original version of this article was revised: The second author (Keiko T. Kitazato) was missing and was added to the list of authors. A correction to this article is available online at https://doi.org/10.1007/s12975-017-0589-4. Change history 11/26/2017 In the original publication of the article, the second author (Keiko T. Kitazato) was missing. Compliance with Ethical Standards Ethical Approval Our animal study was approved by the Ethics Committee of the Institute of Biological Medicine of the University of Tokushima Graduate School and followed all applicable international, national, and/or institutional guidelines for the care and use of animals. Conflict of Interest The authors declare that they have no conflicts of interest.

n of tissue iron content [24, 25], assessment of functional changes [26], and quantification of contrast agent concentration [27]. Concerning acute stroke QSM has been shown to be able to assess vessel function and oxygen metabolism in patients with acute stroke [18, 28] as well as in animal models of the disease [29]. In the present study, we investigated the potential of QSM and MR frequency mapping to assess the evolution of vascular and tissue changes in the mouse brain after transient middle cerebral artery occlusion (tMCAO). We acquired high-resolution GRE, DWI, and multi-slice multi-echo imaging data of mice brains at different time points after reperfusion. On the post-processed QSM and MR frequency maps, magnetic susceptibility and frequency were quantified in prominent vessels and brain tissues in both the ischemic and contralateral hemisphere side. We also evaluated the time courses of the occurrence of regional contrast changes on the frequency, magnetic susceptibility, apparent diffusion coefficient (ADC), and T 2 relaxation time constant maps. Immunohistochemical analyses were performed to assess underlying vascular pathology.

nd contralateral hemisphere side. We also evaluated the time courses of the occurrence of regional contrast changes on the frequency, magnetic susceptibility, apparent diffusion coefficient (ADC), and T 2 relaxation time constant maps. Immunohistochemical analyses were performed to assess underlying vascular pathology. Methods Animals All procedures conformed to the national guidelines of the Swiss Federal act on animal protection and were approved by the Cantonal Veterinary Office Zurich (Permit Number: 18-2014 and 49-2011). All procedures fulfilled the ARRIVE guidelines on reporting animal experiments. Animals were housed in a temperature-controlled room in individually ventilated cages, containing up to five animals per cage, under a 12-h dark/light cycle. Paper tissue was given as environmental enrichment. Access to pelleted food (3437PXL15, CARGILL) and water was provided ad libitum.

Introduction Magnetic resonance imaging (MRI) is an important aid for physicians in the diagnosis and management of patients with acute stroke [1], providing multiple useful contrasts for assessing hemodynamic function as well as extent and severity of brain injury. In case of ischemic stroke, magnetic resonance angiography, for instance, can identify occlusion of a parent artery [2], whereas perfusion-weighted imaging (PWI) informs about regional disturbances of cerebral blood supply in hyperacute and acute ischemic stroke [3]. Diffusion-weighted imaging (DWI) has been shown to depict the ischemic lesion in the hyperacute, acute, and subacute stage after an ischemic insult [4–7]. Analyzing T 1 and T 2 relaxation times has also been used to assess ischemic damage [8, 9].

ces of cerebral blood supply in hyperacute and acute ischemic stroke [3]. Diffusion-weighted imaging (DWI) has been shown to depict the ischemic lesion in the hyperacute, acute, and subacute stage after an ischemic insult [4–7]. Analyzing T 1 and T 2 relaxation times has also been used to assess ischemic damage [8, 9]. Bulk magnetic susceptibility is a fundamental physical property representing a materials’ tendency to interact with and distort an applied magnetic field. By applying gradient (recalled) echo (GRE) magnetic resonance-based techniques, such as T 2*-weighted imaging [10, 11], phase imaging [12, 13], and susceptibility-weighted imaging (SWI) [14, 15], it is possible to assess qualitatively magnetic susceptibility variations in the brain. Regarding acute stroke MRI, T 2*-weighted imaging and SWI are used to detect cerebral microbleeds and hemorrhages [16], where SWI is also used to identify areas of hypoperfusion and to detect acute intravascular emboli [1]. Furthermore, asymmetrical veins between ischemic and normal brain tissues have been demonstrated with SWI, which may add information about local oxygen metabolism [17–19].

t cerebral microbleeds and hemorrhages [16], where SWI is also used to identify areas of hypoperfusion and to detect acute intravascular emboli [1]. Furthermore, asymmetrical veins between ischemic and normal brain tissues have been demonstrated with SWI, which may add information about local oxygen metabolism [17–19]. More recently, quantitative susceptibility mapping (QSM) has been introduced as a promising post-processing technique based on GRE data. QSM utilizes the small magnetic field variations arising from the underlying tissue magnetic susceptibility distribution to compute quantitative maps. It provides complementary anatomical contrast of the brain [20, 21], supports identification and characterization of brain lesions [22, 23], but also enables quantification of tissue iron content [24, 25], assessment of functional changes [26], and quantification of contrast agent concentration [27]. Concerning acute stroke QSM has been shown to be able to assess vessel function and oxygen metabolism in patients with acute stroke [18, 28] as well as in animal models of the disease [29].

reporting animal experiments. Animals were housed in a temperature-controlled room in individually ventilated cages, containing up to five animals per cage, under a 12-h dark/light cycle. Paper tissue was given as environmental enrichment. Access to pelleted food (3437PXL15, CARGILL) and water was provided ad libitum. Study Design and Ischemia Model Seventeen male C57Bl6/J mice (Janvier, France, weight range 20–25 g, age range 8–10 weeks) were used. Anesthesia was initiated by using 3% isoflurane (Abbott, Cham, Switzerland) in a mixture of O2 (200 ml/min) and air (800 ml/min) and maintained with 1.5–2% isoflurane. Prior to surgery, a local analgesic (lidocaine, 0.5%, 7 mg/kg) was administered subcutaneously. Temperature was controlled during the surgery and kept constant at 36.5 ± 0.5 °C with a feedback-controlled heating pad system. The surgical procedure was carried out as described [30, 31]. The middle cerebral artery was occluded for 1 h. After surgery, buprenorphine was administered as subcutaneous injection every 6–8 h on the day of surgery (Temgesic, 0.1 mg/kg b.w) and supplied thereafter via the drinking water (1 mg/kg) for 36 h. tMCAO animals were assessed with MRI at 2 h (n = 3), 4 h (n = 4), 6 h (n = 3), 12 h (n = 4), 24 h (n = 3), and 48 h (n = 6) after reperfusion, with the majority of animals being measured at two time points. ADC maps of all investigated mice were inspected. Animals were analyzed when a lesion was present on the ADC maps.

tMCAO animals were assessed with MRI at 2 h (n = 3), 4 h (n = 4), 6 h (n = 3), 12 h (n = 4), 24 h (n = 3), and 48 h (n = 6) after reperfusion, with the majority of animals being measured at two time points. ADC maps of all investigated mice were inspected. Animals were analyzed when a lesion was present on the ADC maps. Magnetic Resonance Imaging MRI measurements were acquired on a Bruker PharmaScan 47/16 (Bruker BioSpin GmbH, Ettlingen, Germany) operating at 4.7 T and equipped with a cryogenic transmit-receive RF coil [32]. During MRI, mice were spontaneously breathing under isoflurane anesthesia (1.5%). Body temperature was monitored with a rectal temperature probe (MLT415, ADInstruments, Spechbach, Germany) and kept at 36 ± 0.5 °C using a warm water circuit integrated into the animal support. Anatomical reference data acquired in coronal and sagittal orientations served for accurate positioning of the animal’s head. Global first-order shimming followed by fieldmap-based local shimming was performed on the mouse brain using the automated MAPshim routine to reduce field inhomogeneities.

Magnetic Resonance Imaging MRI measurements were acquired on a Bruker PharmaScan 47/16 (Bruker BioSpin GmbH, Ettlingen, Germany) operating at 4.7 T and equipped with a cryogenic transmit-receive RF coil [32]. During MRI, mice were spontaneously breathing under isoflurane anesthesia (1.5%). Body temperature was monitored with a rectal temperature probe (MLT415, ADInstruments, Spechbach, Germany) and kept at 36 ± 0.5 °C using a warm water circuit integrated into the animal support. Anatomical reference data acquired in coronal and sagittal orientations served for accurate positioning of the animal’s head. Global first-order shimming followed by fieldmap-based local shimming was performed on the mouse brain using the automated MAPshim routine to reduce field inhomogeneities. For DWI, a two-dimensional (2D) multi-segment spin echo sequence with echo planar imaging readout (SE-EPI) was used. The scan parameters were field-of-view (FOV) = 17 mm × 14 mm, acquisition matrix = 128 × 128, nominal in-plane voxel size = 133 μm × 109 μm, 12 slices of 1 mm thickness, and an interslice distance = 1.3 mm, number of segments = 4, echo time (TE) = 27.5 ms, and repetition time (TR) = 3000 ms. Diffusion-encoding was applied in x-, y-, and z-direction (gradient pulse duration = 4 ms, gradient pulse separation = 14 ms) with b values of 100, 200, 400, 600, 800, and 1000 s/mm2, respectively. The acquisition time was 3 min and 48 s.

er of segments = 4, echo time (TE) = 27.5 ms, and repetition time (TR) = 3000 ms. Diffusion-encoding was applied in x-, y-, and z-direction (gradient pulse duration = 4 ms, gradient pulse separation = 14 ms) with b values of 100, 200, 400, 600, 800, and 1000 s/mm2, respectively. The acquisition time was 3 min and 48 s. To extract T 2 relaxation time constant of the brain tissue, a 2D Carr-Purcell-Meiboom-Gill multi-slice multi-echo sequence was applied with FOV = 20 mm × 20 mm, acquisition matrix = 100 × 100, nominal in-plane voxel size = 200 μm × 200 μm, 14 echoes with TE1 = 12 ms and an inter-echo time = 12 ms, TR = 2783 ms, and four averages. The acquisition time was 14 min and 6 s. For frequency mapping and QSM, a 3D multi-echo GRE sequence was applied using a FOV = 25.6 mm × 25.6 mm × 8 mm and an acquisition matrix = 256 × 256 × 80, resulting in an effectively isotropic spatial resolution of 100 μm × 100 μm × 100 μm. Four echoes were recorded (TE1–4 = 4.5/10.5/16.5/22.5 ms) with TR = 100 ms, flip angle = 15°, monopolar echo readout and no averaging. The acquisition time was 25 min and 36 s. Data Processing ADC maps were calculated on a pixel-by-pixel basis with linear regression analysis using the model function: 1 ln(S(b)/S0)=−b⋅ADC, where S(b) is the measured signal intensity at a specific b value (b) and S 0 the signal intensity in the absence of a diffusion gradient (b = 0).

For frequency mapping and QSM, a 3D multi-echo GRE sequence was applied using a FOV = 25.6 mm × 25.6 mm × 8 mm and an acquisition matrix = 256 × 256 × 80, resulting in an effectively isotropic spatial resolution of 100 μm × 100 μm × 100 μm. Four echoes were recorded (TE1–4 = 4.5/10.5/16.5/22.5 ms) with TR = 100 ms, flip angle = 15°, monopolar echo readout and no averaging. The acquisition time was 25 min and 36 s. Data Processing ADC maps were calculated on a pixel-by-pixel basis with linear regression analysis using the model function: 1 ln(S(b)/S0)=−b⋅ADC, where S(b) is the measured signal intensity at a specific b value (b) and S 0 the signal intensity in the absence of a diffusion gradient (b = 0). The T2 relaxation time was computed by fitting the spin echo magnitude signal, S, at each TE, to a mono-exponential decay function for each pixel using Paravision software (Bruker): 2 S(TE)=S0⋅exp(−TET2), where both S 0, the signal at TE = 0, and T2, the irreversible transverse relaxation time, are fit parameters.

1 ln(S(b)/S0)=−b⋅ADC, where S(b) is the measured signal intensity at a specific b value (b) and S 0 the signal intensity in the absence of a diffusion gradient (b = 0). The T2 relaxation time was computed by fitting the spin echo magnitude signal, S, at each TE, to a mono-exponential decay function for each pixel using Paravision software (Bruker): 2 S(TE)=S0⋅exp(−TET2), where both S 0, the signal at TE = 0, and T2, the irreversible transverse relaxation time, are fit parameters. Single-channel GRE magnitude images were combined using the sum-of-squares method [33], whereas single-channel GRE phase images were combined by taking the argument of the complex summed single-channel images after subtracting the channel-dependent phase offset estimated in the center of the 3D volume of the first echo [34]. Quantitative susceptibility maps were computed based on these combined phase images. To this end, the combined phase images for each echo were unwrapped using a 3D best-path algorithm [35], divided by (2π · TE) to obtain the Larmor frequency variation in Hz, and then combined across the different TEs in an optimized way that takes into account the local echo time-dependent contrast-to-noise ratio of the Larmor frequency images [36]. Background frequency contributions were eliminated using sophisticated harmonic artifact removal for phase data (SHARP) [37], with ten different spherical kernels with varying radii ranging from 100 to 1000 μm [38], and employing a regularization parameter for truncated singular value decomposition of 0.05. Susceptibility mapping was performed based on SHARP-processed frequency images using homogeneity enabled incremental dipole inversion (HEIDI) [20].

ifferent spherical kernels with varying radii ranging from 100 to 1000 μm [38], and employing a regularization parameter for truncated singular value decomposition of 0.05. Susceptibility mapping was performed based on SHARP-processed frequency images using homogeneity enabled incremental dipole inversion (HEIDI) [20]. Volume-of-Interest Analysis Significant deviations from the signal distribution were identified in the ischemic hemisphere compared to the unaffected, contralateral hemisphere on all contrasts using Paravision (Bruker) and MRIcron. For quantitative evaluation, volumes-of-interest (VOIs) were drawn around three individual vessel structures that appeared prominent on both the ipsilateral side and the contralateral side if visible, and around one lateral ventricle using MRIcron (www.sph.sc.edu/comd/rorden/mricron). In addition, VOIs were drawn around areas of regional contrast change, as well as the remaining ipsi- and contralateral striatum and cortex. The extent of regional contrast changes was calculated edema corrected as described [39].

ble, and around one lateral ventricle using MRIcron (www.sph.sc.edu/comd/rorden/mricron). In addition, VOIs were drawn around areas of regional contrast change, as well as the remaining ipsi- and contralateral striatum and cortex. The extent of regional contrast changes was calculated edema corrected as described [39]. Immunohistochemistry After MRI, few mice were used for immunohistochemistry: 12 h (n = 2), 24 h (n = 3), and 48 h (n = 2) after reperfusion. Animals were deeply anesthetized by intraperitoneal injection of ketamine/xylazine/acepromazine maleate (100/20/3 mg/kg body weight) and decapitated. Brains were immediately removed and snap-frozen in 2-methylbutane (Sigma-Aldrich, Switzerland) cooled with dry ice to − 30 °C and stored at − 80 °C until processing. They were then thawed and fixed in 4% paraformaldehyde (PFA) for 48 h, then trimmed (coronal section) and embedded routinely in paraffin wax. Consecutive sections (3–5 μm) were prepared and, after antigen retrieval via incubation in citrate buffer (pH 6.0) for 20 min at 98 °C, incubated, with rabbit anti-mouse collagen IV (Cat # 2150-1470, AbD Serotec, dil 1:200) for 15–18 h at 4 °C. Subsequently, they were incubated with Envision rabbit, Dako.

paraffin wax. Consecutive sections (3–5 μm) were prepared and, after antigen retrieval via incubation in citrate buffer (pH 6.0) for 20 min at 98 °C, incubated, with rabbit anti-mouse collagen IV (Cat # 2150-1470, AbD Serotec, dil 1:200) for 15–18 h at 4 °C. Subsequently, they were incubated with Envision rabbit, Dako. Statistical Analysis Statistical analysis was performed using SigmaPlot 12.5 (Systat Software, San Jose, CA). Frequency values and susceptibility differences of vessels were compared with a Mann-Whitney rank sum test, whereas comparisons between different brain regions were performed using an analysis of variance, followed by Holm-Sidak post hoc test for multiple comparisons. Lesion volumes between different contrasts were compared with Student’s t test. Results All post-processed images are made available in a data repository (10.6084/m9.figshare.5630071.v1).

Statistical Analysis Statistical analysis was performed using SigmaPlot 12.5 (Systat Software, San Jose, CA). Frequency values and susceptibility differences of vessels were compared with a Mann-Whitney rank sum test, whereas comparisons between different brain regions were performed using an analysis of variance, followed by Holm-Sidak post hoc test for multiple comparisons. Lesion volumes between different contrasts were compared with Student’s t test. Results All post-processed images are made available in a data repository (10.6084/m9.figshare.5630071.v1). Prominent Vessels Within the Ischemic Hemisphere on Frequency and Susceptibility Maps One mouse was excluded from the analysis because no lesion was visible on ADC maps. GRE data of the brain were inspected for the different time periods after reperfusion. Prominent vessels on background-corrected frequency maps and magnetic susceptibility maps of the ischemic hemisphere revealed high frequency and magnetic susceptibility values (Figs. 1 and 2, white arrows). On the frequency maps, they often appeared as white structures surrounded by a dark rim and were mainly found ipsilateral, in the territory supplied by the middle cerebral artery (MCA). On the contralateral hemisphere, vessel-like structures were occasionally observed, but were only faintly visible against tissue background (for example, Fig. 4, 6 h after reperfusion). Moreover, ipsilateral prominent vessels appeared larger in diameter than comparable vessels on the contralateral side. Furthermore, an increased number of prominent vessels were found in the ischemic hemisphere of mice imaged at 12, 24, and 48 h after reperfusion compared to mice imaged at 2, 4, and 6 h after reperfusion.Fig. 1 Display of representative axial background-field corrected frequency and quantitative susceptibility maps (QSM) of the ischemic hemisphere of a mouse of a tMCAO mouse after 2 and 6 h of reperfusion. For both contrasts, three cross-sections containing the ischemic territory (approximately bregma 0.14 and − 0.82 mm) are shown. Only few prominent vessels are seen with high MR frequencies and increased magnetic susceptibilities (white arrows). Lesions showing decreased frequencies are also discernable (enclosed by white dotted line)

sts, three cross-sections containing the ischemic territory (approximately bregma 0.14 and − 0.82 mm) are shown. Only few prominent vessels are seen with high MR frequencies and increased magnetic susceptibilities (white arrows). Lesions showing decreased frequencies are also discernable (enclosed by white dotted line) Fig. 2 Display of three cross-sections of frequency and quantitative susceptibility maps (QSM) of the ischemic hemisphere of a tMCAO mouse after 12 and 48 h of reperfusion. Prominent vessels with increased magnetic susceptibility (white arrows) occurred more frequently compared to shorter time intervals of reperfusion. Lesions were discernable on both frequency and susceptibility maps (white dotted line) which increased in size

hemisphere of a tMCAO mouse after 12 and 48 h of reperfusion. Prominent vessels with increased magnetic susceptibility (white arrows) occurred more frequently compared to shorter time intervals of reperfusion. Lesions were discernable on both frequency and susceptibility maps (white dotted line) which increased in size VOI analysis revealed significantly higher frequency values at 2, 4, and 48 h after reperfusion as well as higher differences in magnetic susceptibility (relative to CSF) at 2 and 4 h after reperfusion in ipsilateral vessels compared to vessels in the contralateral hemisphere (Fig. 3a, b).Fig. 3 Quantitative analysis of MRI data and assessment of vessels with immunohistochemistry. a, b VOI analysis of MR frequency values and magnetic susceptibilities in prominent vessels in the ischemic ipsilateral and contralateral hemisphere at different reperfusion intervals, respectively. Bar graphs represent mean ± SD. *p < 0.05 compared to contralateral side. c, d Representative anti-collagen IV (basal membrane) immunohistochemistry of the brain 24 h after reperfusion. c Larger vessels are dilated (asterisk) and endothelial cells in capillaries are swollen and the vessel lumen are narrowed (insert) on the ipsilateral, d while this is not observed at the contralateral side. Bar = 100 μm. Insert bar = 20 μm

n IV (basal membrane) immunohistochemistry of the brain 24 h after reperfusion. c Larger vessels are dilated (asterisk) and endothelial cells in capillaries are swollen and the vessel lumen are narrowed (insert) on the ipsilateral, d while this is not observed at the contralateral side. Bar = 100 μm. Insert bar = 20 μm Examination of brain sections after collagen IV staining demonstrated dilation of larger vessels on the ipsilateral compared to the contralateral side (Fig. 3c). In addition, capillaries showed swollen endothelial cells and narrowed vessel lumen (Fig. 3c, zoom in) compared to contralateral capillaries in equivalent locations, for which vascular lumen was not affected (Fig. 3c, d). Detection of Tissue Changes on Frequency and Susceptibility Maps Inspection of the background-field corrected frequency maps revealed ipsilateral tissue areas of decreased frequency values at all investigated time points (Figs. 1 and 2, dotted line). The contrast changes were more apparent at 24 and 48 h after reperfusion compared to earlier time points. Similarly, in QSM areas of low magnetic susceptibility, values were observed at 24 and 48 h after reperfusion, while at earlier time points, such areas were only occasionally visible (Figs. 1 and 2). Areas of contrast change confined within the MCA territory; however, the location and extent largely varied within the groups of animals investigated at an individual time point as well as between the different time points.

fusion, while at earlier time points, such areas were only occasionally visible (Figs. 1 and 2). Areas of contrast change confined within the MCA territory; however, the location and extent largely varied within the groups of animals investigated at an individual time point as well as between the different time points. Quantitative analysis demonstrated significantly different frequency values between the lesion areas of regional tissue changes and the contralateral striatum at all time points, while for differences in magnetic susceptibility, a statistical difference was seen at 48 h after reperfusion only (Fig. 4). Values of the cortex and striatum in the ipsilateral hemisphere (excluding the area of regional tissue change) were not statistically different from the values of the corresponding areas on the contralateral side for both contrasts.Fig. 4 Quantitative analysis of MRI data of different brain regions. a Exemplary VOIs selected on the frequency maps. Cortex and striatum were identified on the ischemic hemisphere (striped pattern) and the contralateral hemisphere (dotted pattern), while excluding areas of markedly reduced frequencies (blue areas). VOI analysis of b MR frequency values and c differences in magnetic susceptibilities (to CSF) in different brain regions at different reperfusion intervals. Bar graphs represent mean ± SD. *p < 0.05 compared to contralateral side

here (dotted pattern), while excluding areas of markedly reduced frequencies (blue areas). VOI analysis of b MR frequency values and c differences in magnetic susceptibilities (to CSF) in different brain regions at different reperfusion intervals. Bar graphs represent mean ± SD. *p < 0.05 compared to contralateral side Extent of Regional Contrast Changes for Different MRI Contrasts We measured the extent of contrast changes on frequency and quantitative susceptibility maps and compared it to the cerebral lesion volumes obtained on ADC and T 2 maps (Figs. 5 and 6). Among all MRI contrasts, the ischemic lesion became first visible on ADC maps as areas of decreased ADC at 2 h after reperfusion followed by a steady growth of the lesion until 24 h after reperfusion (Figs. 5 and 6a). On T 2 maps, a lesion with increased T 2 values became the earliest discernable at 6 h after reperfusion with growth until 24 h after reperfusion (Figs. 5 and 6b). The final cerebral hemispheric lesion volumes determined on ADC maps at 48 h after reperfusion were not significantly different from the volumes obtained from T 2 maps (mean ± SD, 47.4 ± 15.2% ADC vs. 48.3 ± 12.8% T 2; p = 0.914). In contrast, regions of contrast change as seen on frequency and quantitative susceptibility maps varied considerably in extent and temporal trajectory. Areas of decreased frequency values and magnetic susceptibility were first discernable at 2 h after reperfusion (Figs. 5 and 6c, d). Growth of these regions was only minor on frequency and quantitative susceptibility maps until 48 h after reperfusion, and the extent was at all time points lower compared to the lesion volumes seen on ADC and T 2 maps. The final extent after 48 h after reperfusion was significantly lower on quantitative susceptibility compared to frequency maps (19.1 ± 12.0% frequency vs. 4.7 ± 2.4% QSM; p = 0.003).Fig. 5 Contrast changes observed on parametric maps of ADC, T 2 relaxation time constant, background-field corrected MR frequency, and magnetic susceptibility (QSM) following 1 h of tMCAO and different intervals of reperfusion. The ischemic lesion first appears on the ADC map as an area with significantly reduced ADC at 2 h after MCAO, before it becomes apparent by increased T 2 values on the T 2 map at 6 h after MCAO. Regional but comparatively smaller contrast changes are discernable on MR frequency and quantitative susceptibility maps from 2 h after reperfusion onwards

ars on the ADC map as an area with significantly reduced ADC at 2 h after MCAO, before it becomes apparent by increased T 2 values on the T 2 map at 6 h after MCAO. Regional but comparatively smaller contrast changes are discernable on MR frequency and quantitative susceptibility maps from 2 h after reperfusion onwards Fig. 6 Temporal evolution of the different MRI contrasts after 1 h of tMCAO and different periods of reperfusion. Percentage changes of the extent of regional contrast change (edema corrected) from a region-of-interest analysis of maps of a ADC, b T 2 relaxation time constant, c background-field corrected MR frequency, and d magnetic susceptibility (QSM). Displayed are mean ± standard deviations Discussion In the current study, high-resolution MR frequency and quantitative susceptibility maps of the mouse brain were generated after 1 h of tMCAO and different time points after reperfusion. On both maps, prominent vessels were seen with increased frequencies and magnetic susceptibilities. In addition, we observed evolutions of regional contrast, which differed in appearance from those seen on ADC and T 2 relaxation time constant maps.

nerated after 1 h of tMCAO and different time points after reperfusion. On both maps, prominent vessels were seen with increased frequencies and magnetic susceptibilities. In addition, we observed evolutions of regional contrast, which differed in appearance from those seen on ADC and T 2 relaxation time constant maps. Increased Susceptibility of Prominent Cerebral Vessels Appearance of prominent vessels in the brain of patients with ischemic stroke on SW images and quantitative susceptibility maps have been described recently [17–19]. Their occurrence has been attributed to an increase in the oxygen extraction fraction (OEF) and correlation to misery perfusion indicated by PWI [18]. The increased oxygen extraction leads to higher deoxyhemoglobin concentrations in veins, which increase magnetic susceptibility locally [15]. It has therefore been suggested that these prominent vessels demarcate the penumbra that can be salvaged by vessel recanalization and that SWI can be used to predict infarct growth [17, 19].

d oxygen extraction leads to higher deoxyhemoglobin concentrations in veins, which increase magnetic susceptibility locally [15]. It has therefore been suggested that these prominent vessels demarcate the penumbra that can be salvaged by vessel recanalization and that SWI can be used to predict infarct growth [17, 19]. In the current study, we also demonstrated the occurrence of prominent vessels in the mouse brain following transient ischemia and reperfusion. Vessels were better visible, appeared wider, and had significantly higher MR frequency values and larger differences of magnetic susceptibility relative to CSF on the ipsilateral side compared to vessels in the contralateral unaffected hemisphere (Figs. 1, 2, and 3). Prominent vessels were mainly observed in the surroundings of the core of the lesion. Reperfusion is associated in part by incomplete blood flow of the microvasculature (no-reflow phenomenon). Examination of brain section stained for the basal membrane demonstrated vasodilation of larger vessels in the ischemic ipsilateral hemisphere, particularly at 24 and 48 h after reperfusion, together with swollen capillaries with narrowed lumen. These findings are in agreement with previous studies which described capillary constriction and impaired capillary reflow as a consequence of pericyte contraction [40, 41]. Thus, our observation of prominent vessels around the lesion after restoration of cerebral blood flow is compatible of lower tissue oxygen availability due to compressed capillaries in the ischemic tissue, which is compensated by an increased oxygen extraction in the surrounding area. As an incomplete blood supply is contributing to the propagation of the ischemic lesion, treatments are required that ameliorate pericyte constrictions and capillary compression to fully restore tissue function after reperfusion. The role of prominent vessels after reperfusion need to be further investigated. Thus, prominent vessels are an important indicator of underlying microvascular pathology and QSM and frequency maps might be used as a proxy of underlying microvascular pathology.

ression to fully restore tissue function after reperfusion. The role of prominent vessels after reperfusion need to be further investigated. Thus, prominent vessels are an important indicator of underlying microvascular pathology and QSM and frequency maps might be used as a proxy of underlying microvascular pathology. Detection of Tissue Changes on Frequency and Susceptibility Maps We also observed decreased frequency values and smaller susceptibility differences relative to CSF in brain tissue regions, which has so far not been reported in patients with ischemic stroke. One reason for this could be that we performed our study at a field strength of 4.7 T, which is higher than those commonly used in clinical studies (usually 1.5 T). As the effects of magnetic susceptibility are proportional to the applied magnetic field, we are thus more sensitive regarding detectability of differences in magnetic susceptibility between tissues. Furthermore, we used a higher spatial resolution (100 μm) compared to clinical studies (usually > 1 mm), which again helps to delineate small morphological features. The accompanying loss in signal-to-noise ratio was compensated for by using higher field strength and a cryogenic transmitter-receive RF coil [32]. However, also differences in principle function between our animal model and typical clinical scenarios, including patient selection and stroke etiology, type and location may be responsible for the lack of this observation in patients.

or by using higher field strength and a cryogenic transmitter-receive RF coil [32]. However, also differences in principle function between our animal model and typical clinical scenarios, including patient selection and stroke etiology, type and location may be responsible for the lack of this observation in patients. The appearance of regions with decreased frequency values and smaller susceptibility differences was more apparent at 24 and 48 h after reperfusion compared to the earlier investigated time points, but was highly variable between individual animals. Moreover, the growth rate and extent of these areas were significantly smaller on frequency and quantitative susceptibility maps compared to ADC and T 2 maps. The underpinnings of these changes on susceptibility and frequency maps are currently unclear. Since cerebral ischemia is followed by microstructural changes at different stages with edema formation, cell death, and tissue destruction and phagocytosis of debris resulting in cavitation with surrounding gliosis [42–44], this complex chain of processes is difficult to capture with single frequency or susceptibility values. On the other hand, compartmental water shifts due to cytotoxic and vasogenic edema can be detected by DWI and T 2 mapping, respectively [6, 7]. Interestingly, we did not observe a spatial congruence between regions of increased T 2 and reduced ADC values with areas of reduced MR frequency values and differences in magnetic susceptibility. Thus, both contrasts do likely not represent the ischemic lesion as distinct to T 2-weighted imaging and ADC, where a good correlation to the histopathological lesion has been demonstrated [45]. It might be speculated that in the regions, which are at the core of the lesion, oxygen is not extracted [46], which increases oxygenated hemoglobin concentrations in those areas and produces a diamagnetic shift of susceptibility [47]. Clearly, the relevance of these speculations and the associated cellular underpinnings need to be investigated in further studies.

ich are at the core of the lesion, oxygen is not extracted [46], which increases oxygenated hemoglobin concentrations in those areas and produces a diamagnetic shift of susceptibility [47]. Clearly, the relevance of these speculations and the associated cellular underpinnings need to be investigated in further studies. Taken together, we were able in our study to identify characteristic changes of magnetic susceptibility in the mouse brain after transient ischemia followed by reperfusion. Since QSM can be reconstructed from GRE data, there is no extra penalty in acquisition time to reconstruct such maps. In addition, several approaches for accelerating GRE acquisitions have been proposed for human imaging [48–51]. Currently, clinical application of QSM is hampered due to its numerical complexity and computational cost associated with the post-processing procedure. More recently, novel algorithms have been proposed, which allow rapid online reconstruction of susceptibility maps directly after data acquisition, enabling instant evaluation by medical personal [52]. Thus, QSM appears promising as a useful post-processing tool to evaluate GRE data for the diagnosis and follow-up of patients with ischemic stroke. Funding This study was supported by the Swiss National Science Foundation (Grant PZ00P3_136822) and the Hartmann-Müller Foundation. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflict of interest. Ethical Approval All applicable international and national guidelines for the care and use of animals were followed.

Introduction Detection of paroxysmal atrial fibrillation (PAF) after an ischemic stroke is of paramount importance since patients with AF benefit from oral anticoagulation more than from standard antiplatelet therapy [1, 2]. The diagnosis of PAF is often challenging, as it may remain undetected on conventional ECG and a Holter ECG as well as short-term continuous cardiac monitoring (CCM) provided by stroke unit (SU) monitors [3, 4]. Current guidelines recommend the use of CCM at least for the first 24–72 h after stroke [1, 2]; however, there is at present no consensus on the optimal method and duration of CCM in SU settings [5]. Moreover, conventional bedside monitors are usually withheld after the first days as they severely restrict the patient mobility and are often poor tolerated. The search for PAF in SU can be further pursued with longer term CCM using mobile telemetry devices, repeat Holters, and patch-type devices [5]. However, access to these techniques is currently limited by their economic costs, the lack of standardized approach, and organizational burden. Thus, a large number of stroke patients with undiagnosed PAF may not receive an optimal antithrombotic therapy.

e telemetry devices, repeat Holters, and patch-type devices [5]. However, access to these techniques is currently limited by their economic costs, the lack of standardized approach, and organizational burden. Thus, a large number of stroke patients with undiagnosed PAF may not receive an optimal antithrombotic therapy. The ability to ascertain those patients at higher AF risk within the first days after stroke could improve the patient selection for longer term CCM throughout the SU stay. Clinical features (such as age, sex, the CHA2-DS2-VASc score) and electrocardiographic or echocardiographic (left atrial volume, premature atrial complexes, PR interval) predictors for PAF have been assessed. However, their clinical effectiveness is unclear [6]. SRA clinic is a telemedicine service based upon RR interval dynamic analysis providing automated AF detection for SU patients [7]. The stroke risk analysis (SRA) algorithm had been demonstrated [8, 9] able to identify abnormalities suggestive of increased PAF risk in patients with a previously diagnosed PAF but at sinus rhythm during the analysis with a 60% sensitivity and 99% specificity. Therefore, the SRA algorithm may be used to identify those acute stroke patients at higher risk for forthcoming PAF episodes. In this study, we applied SRA analysis on CCM obtained within the initial 48 h from SU admission. We hypothesized that this technique would identify patients at higher risk for PAF detected during SU hospitalization.

SRA clinic is a telemedicine service based upon RR interval dynamic analysis providing automated AF detection for SU patients [7]. The stroke risk analysis (SRA) algorithm had been demonstrated [8, 9] able to identify abnormalities suggestive of increased PAF risk in patients with a previously diagnosed PAF but at sinus rhythm during the analysis with a 60% sensitivity and 99% specificity. Therefore, the SRA algorithm may be used to identify those acute stroke patients at higher risk for forthcoming PAF episodes. In this study, we applied SRA analysis on CCM obtained within the initial 48 h from SU admission. We hypothesized that this technique would identify patients at higher risk for PAF detected during SU hospitalization. Methods Within an unsponsored investigator-initiated observational study approved by the local ethical committee, we investigated unselected acute ischemic stroke and TIA patients consecutively admitted in a single stroke center. Enrollment was retrospective from January 2014 to January 2015 and then prospective until December 2016. Inclusion criteria were as follows: stroke/TIA diagnosis confirmed by a neurologist, hospital admission within 24 h from symptom onset, availability of initial 48-h CCM, and availability of standard ECG or a 24-h Holter ECG or further CCM after the first days of monitoring. Inability to give consent and the presence of implantable rhythm control devices were exclusion criteria. For all patients, a clinical history was collected and National Institutes of Health Stroke Scale (NIHSS) admission and CHA2-DS2-Vasc scores calculated. Standard investigations comprised brain CT/MRI, echocardiography, and carotid ultrasound study. Additional ECGs, a 24-h Holter ECG, and further CCM monitoring were performed according to the clinician’s discretion. Patients treated with IV thrombolysis were admitted in the ICU department for the first 24–48 h and initiated their initial CCM when transferred into the SU.

cardiography, and carotid ultrasound study. Additional ECGs, a 24-h Holter ECG, and further CCM monitoring were performed according to the clinician’s discretion. Patients treated with IV thrombolysis were admitted in the ICU department for the first 24–48 h and initiated their initial CCM when transferred into the SU. After SU admission, CCM for consecutive 48 h was immediately initiated using standard bedside monitors (Intellivue MP40, Philips). After 48 h, CCM was withheld to permit mobilization unless further monitoring was deemed clinically necessary in patients with more severe/unstable clinical conditions. CCM ECG tracks were sent via a secure internet connection to the SRA server (Apoplex Medical Technologies, Pirmasens, Germany) for the assessment of RR interval dynamics calculated on overall 48-h recordings. The details of SRA data processing have been previously described [8, 9]. Briefly, the AF risk is associated with the presence of premature atrial complexes, atrial tachycardia, and other ectopic activities. These abnormalities alter atrioventricular nodal conduction and result in changes in the ventricular response that are often not detected by conventional linear assessment of heart rate variability. The SRA algorithm assesses the AF risk in three steps. QRS complexes are first identified on ECG to create a RR interval list. RR intervals are normalized by dividing them by the mean of the two corresponding RR intervals [(Ri − Ri + 1) ∕ (Ri + Ri + 1)]. These RR intervals are then used to calculate various, mostly nonlinear, mathematical parameters such as the following: standard deviation of the minor and major axis of the Poincaré plots and their ratio; RR fluctuation based upon different analyses of consecutive RR intervals; the number of premature atrial complexes without sinus nodal reset. Finally, entropic analysis of RR intervals regularity is applied.

ers such as the following: standard deviation of the minor and major axis of the Poincaré plots and their ratio; RR fluctuation based upon different analyses of consecutive RR intervals; the number of premature atrial complexes without sinus nodal reset. Finally, entropic analysis of RR intervals regularity is applied. The SRA algorithm classified initial 48-h CCM tracks as (1) no presence of AF and low risk for AF; (2) no AF and high risk for AF; and (3) presence of manifest episodes of AF. In the case of detected AF, the SRA service provided source ECG tracks for clinical confirmation of AF. SRA clinic detection of manifest AF episodes includes only episodes > 30 s. In the presence of manifest AF on ECG data, SRA clinic has a 99% sensitivity and specificity compared to Holter ECG [9]. At discharge, for each patient, the heart rhythm was adjudicated by a cardiologist as sinus rhythm/PAF/AF based on all available ECGs, a Holter ECG, and CCM tracks (including the SRA algorithm provided data).

In the presence of manifest AF on ECG data, SRA clinic has a 99% sensitivity and specificity compared to Holter ECG [9]. At discharge, for each patient, the heart rhythm was adjudicated by a cardiologist as sinus rhythm/PAF/AF based on all available ECGs, a Holter ECG, and CCM tracks (including the SRA algorithm provided data). Statistical Analysis In univariate analysis, we compared the risk factors that were different between patients at low risk of AF, at high risk of AF, and with manifest AF, using ANOVA for continuous data (age, NIHSS, QTc interval) and the Pearson chi-square test for categorical data. Pairwise post hoc comparisons were performed with the Bonferroni correction. We subsequently compared risk factors for subsequent diagnosis of AF in the subgroup of patients without manifest AF on SRA clinic assessment. Logistic regression was then performed in the subgroup of patients without manifest AF on SRA clinic assessment with presence of AF as the dependent and SRA clinic assessment as low or high risk, previous diagnosis of PAF/AF and CHA2D2-Vasc score as the independent variables. Because of the limited number of patients with a further diagnosis of AF, we did not further adjust for other factors. The analyses were performed using SPSS 21.0 (IBM). A p value of 0.05 was considered significant.

nt as low or high risk, previous diagnosis of PAF/AF and CHA2D2-Vasc score as the independent variables. Because of the limited number of patients with a further diagnosis of AF, we did not further adjust for other factors. The analyses were performed using SPSS 21.0 (IBM). A p value of 0.05 was considered significant. Results Between January 2014 and December 2016, 259 patients were considered for inclusion in the study. Of these, 59 were excluded for the following reasons: hospital admission later than 24 h from symptom onset (8); lack of early CCM (23); presence of ICDs (10); refusal/inability to participate in the study (18). Two hundred patients (99 retrospective, 40% females, mean age 71 ± 16 years) with ischemic stroke (n = 187) and TIA (n = 13) were finally enrolled. Stroke/TIA etiologic TOAST classification of enrolled patients was as follows: large artery atherosclerosis 11 (5.5%) patients; lacunar 20 (10%); other determined etiology 9 (4.5%); cardioembolic 62 (31%); cryptogenic (more than one cause or incomplete evaluation) 13 (6.5%); embolic stroke of undetermined source 85 (42.5%). One hundred and fifty-six patients (77%) initiated CCM within the same day of symptom onset while 44 patients treated with I.V thrombolysis and first admitted to the ICU initiated CCM 24 to 48 h from hospital admission.

han one cause or incomplete evaluation) 13 (6.5%); embolic stroke of undetermined source 85 (42.5%). One hundred and fifty-six patients (77%) initiated CCM within the same day of symptom onset while 44 patients treated with I.V thrombolysis and first admitted to the ICU initiated CCM 24 to 48 h from hospital admission. Table 1 presents a summary of baseline characteristics of the patients. There were no differences in the clinical and instrumental findings between the retrospective and prospective patients cohort. On initial CCM, AF risk was considered low in 111 (55.5%) patients and high in 52 (26%) while 37 patients (18.5%) had manifest AF. Significant differences among the three groups were found with respect to age, risk factors (hypertension, smoking, CHA2DS2-VASC score), stroke severity, thrombolysis rates, and left atrial volume. The low-risk group patients were younger compared to the high-risk (mean difference, 15.3 years, 95% CI 10–20) and the manifest AF groups (15 years, 95% CI 9–21). The manifest AF group had a higher baseline NIHSS compared to the low (p = 0.001) and high AF risk groups (p = 0.001). There was no difference in NIHSS between low and high risk AF groups. The CHA2-DS2-Vasc scores and the QTc values were higher in the high AF and manifest AF groups compared to the low-risk group (both p = 0.001), but there were no significant differences between the high-risk and the manifest AF groups. There were no differences in the rate of previously diagnosed PAF, while the use of class I–III–IV antiarrhythmic drugs was higher in high-risk and AF patients. The mean SU stay was 9.7 ± 4 days without differences between the three groups.Table 1 Patient characteristics

etween the high-risk and the manifest AF groups. There were no differences in the rate of previously diagnosed PAF, while the use of class I–III–IV antiarrhythmic drugs was higher in high-risk and AF patients. The mean SU stay was 9.7 ± 4 days without differences between the three groups.Table 1 Patient characteristics Patient characteristics Low AF risk, N = 111 High AF risk, N = 52 Manifest AF, N = 37 p Age (years) 65 ± 14 80 ± 8 80 ± 12 < 0.001 Sex, female 39 (35) 22 (42) 19 (51) 0.202 Current smoking 20 (18) 1 (2) 3 (8) 0.009 Hypertension 70 (63) 42 (81) 33 (89) 0.003 Diabetes 15 (13) 12 (23) 8 (22) 0.249 Dyslipidemia 44 (40) 17 (33) 10 (27) 0.338 Previous AF/PAF diagnosis 6 (5) 7 (13) 5 (13) 0.140 Antiarrhythmic drugs 24 (22) 22 (42) 18 (50) 0.001 CHA2-DS2-Vasc Median 4.5 5 6 25th percentile 3 5 5 < 0.001 75th percentile 6 6 6 NIHSS 4 ± 5 5 ± 6 11 ± 8 0.009 QTc (ms) 430 ± 34 448 ± 29 462 ± 33 < 0.001 Thrombolysis 17 (15) 6 (12) 12 (33) 0.023 Left atrial volume Normal (< 40 mL/m2) 65 (59) 21 (40) 4 (11) Enlarged (40–45 mL/m2) 11 (10) 7 (13) 4 (11) < 0.001 Severely enlarged (> 45 mL/m2) 8 (7) 13 (25) 15 (41) Not available 27 (24) 11 (21) 14 (38) Left atrial diameter Normal (< 39 mm) 33 (30) 12 (29) 3 (8) 0.095 Enlarged (39–50 mm) 40 (36) 19 (45) 15 (41) Severely enlarged (> 50 mm) 5 (5) 5 (12) 4 (11) Not available 33 (30) 6 (14) 15 (41) Final heart rhythm Sinus rhythm 110 (99) 32 (61.5) 0 PAF 1 (1) 20 (38.5) 12 (32) < 0.001 Permanent/persistent AF 0 0 25 (68)

Left atrial diameter Normal (< 39 mm) 33 (30) 12 (29) 3 (8) 0.095 Enlarged (39–50 mm) 40 (36) 19 (45) 15 (41) Severely enlarged (> 50 mm) 5 (5) 5 (12) 4 (11) Not available 33 (30) 6 (14) 15 (41) Final heart rhythm Sinus rhythm 110 (99) 32 (61.5) 0 PAF 1 (1) 20 (38.5) 12 (32) < 0.001 Permanent/persistent AF 0 0 25 (68) During SU stay, AF was evident in 58 (29%) patients, in 33 (16.5%) as paroxysmal, and in 25 (12.5%) as permanent/persistent. AF was detected in 1/110 low-risk patients (0.9%; 95% CI 0–4.9%) whereas 20/52 (38.5%; 95% CI 25–52%) of high-risk patients had a diagnosis of AF. In the detection of AF, SRA exhibited 100% sensitivity and 97% specificity compared to CCM. Those six patients with AF on SRA not confirmed on source CCM were classified as high-risk for AF. In univariate analysis, the presence of PAF among patients categorized as low- or high-risk of AF was associated with SRA category, age, previous PAF/AF history, use of antiarrhythmic drugs, CHA2-DS2-Vasc, and QTc duration (Table 2). In an exploratory multivariate model, after correction for CHA2DS2-VASC score and previous AF/PAF diagnosis, SRA risk score remained a significant predictor for a final diagnosis of AF (Table 3).Table 2 Comparison between patients with and without paroxysmal atrial fibrillation (PAF) Patient characteristics PAF, N = 21 No AF, N = 142 p

In univariate analysis, the presence of PAF among patients categorized as low- or high-risk of AF was associated with SRA category, age, previous PAF/AF history, use of antiarrhythmic drugs, CHA2-DS2-Vasc, and QTc duration (Table 2). In an exploratory multivariate model, after correction for CHA2DS2-VASC score and previous AF/PAF diagnosis, SRA risk score remained a significant predictor for a final diagnosis of AF (Table 3).Table 2 Comparison between patients with and without paroxysmal atrial fibrillation (PAF) Patient characteristics PAF, N = 21 No AF, N = 142 p Age (years) 81 ± 9 67 ± 15 <0.001 Sex, female 8 (38) 53 (37) 0.946 Current smoking 1 (5) 20 (14) 0.234 Hypertension 17 (81) 95 (67) 0.195 Diabetes 3 (14) 24 (17) 0.763 Dyslipidemia 7 (33) 54 (38) 0.678 Previous AF/PAF diagnosis 5(24) 8 (6) 0.004 Antiarrhythmic drugs 11 (52) 35(25) 0.009 CHA2DS2-Vasc Median 5 4.5 25th percentile 4 3 0.013 75th percentile 6 6 NIHSS 7 ± 7 4 ± 5 0.146 QTc (ms) 458 ± 30 432 ± 34 0.002 Thrombolysis 3 (15) 20 (14) 0.922 Left atrial volume Normal (< 40 mL/m2) 9 (50) 77 (72) Enlarged (40–45 mL/m2) 4 (22) 14 (13) 0.175 Severely enlarged (> 45 mL/m2) 5 (28) 16 (15) Not available 3 35 Left atrial diameter Normal (< 39 mm) 2 (18) 43 (41) 0.296 Enlarged (39–50 mm) 8 (73) 51 (50) Severely enlarged (> 50 mm) 1(9) 9 (9) Not available 10 39 SRA risk score High-risk score 20 (95.2%) 32 (22.5%) < 0.001 Table 3 Multivariate analysis Patient characteristics Hazard Ratio Lower 95% CI Higher 95% CI p

Age (years) 81 ± 9 67 ± 15 <0.001 Sex, female 8 (38) 53 (37) 0.946 Current smoking 1 (5) 20 (14) 0.234 Hypertension 17 (81) 95 (67) 0.195 Diabetes 3 (14) 24 (17) 0.763 Dyslipidemia 7 (33) 54 (38) 0.678 Previous AF/PAF diagnosis 5(24) 8 (6) 0.004 Antiarrhythmic drugs 11 (52) 35(25) 0.009 CHA2DS2-Vasc Median 5 4.5 25th percentile 4 3 0.013 75th percentile 6 6 NIHSS 7 ± 7 4 ± 5 0.146 QTc (ms) 458 ± 30 432 ± 34 0.002 Thrombolysis 3 (15) 20 (14) 0.922 Left atrial volume Normal (< 40 mL/m2) 9 (50) 77 (72) Enlarged (40–45 mL/m2) 4 (22) 14 (13) 0.175 Severely enlarged (> 45 mL/m2) 5 (28) 16 (15) Not available 3 35 Left atrial diameter Normal (< 39 mm) 2 (18) 43 (41) 0.296 Enlarged (39–50 mm) 8 (73) 51 (50) Severely enlarged (> 50 mm) 1(9) 9 (9) Not available 10 39 SRA risk score High-risk score 20 (95.2%) 32 (22.5%) < 0.001 Table 3 Multivariate analysis Patient characteristics Hazard Ratio Lower 95% CI Higher 95% CI p Previous AF/PAF diagnosis 4.27 0.84 21.7 0.080 CHA2DS2-VASC score 0.96 0.56 1.63 0.870 High-risk score SRA 70.1 7.8 632 0.000

Age (years) 81 ± 9 67 ± 15 <0.001 Sex, female 8 (38) 53 (37) 0.946 Current smoking 1 (5) 20 (14) 0.234 Hypertension 17 (81) 95 (67) 0.195 Diabetes 3 (14) 24 (17) 0.763 Dyslipidemia 7 (33) 54 (38) 0.678 Previous AF/PAF diagnosis 5(24) 8 (6) 0.004 Antiarrhythmic drugs 11 (52) 35(25) 0.009 CHA2DS2-Vasc Median 5 4.5 25th percentile 4 3 0.013 75th percentile 6 6 NIHSS 7 ± 7 4 ± 5 0.146 QTc (ms) 458 ± 30 432 ± 34 0.002 Thrombolysis 3 (15) 20 (14) 0.922 Left atrial volume Normal (< 40 mL/m2) 9 (50) 77 (72) Enlarged (40–45 mL/m2) 4 (22) 14 (13) 0.175 Severely enlarged (> 45 mL/m2) 5 (28) 16 (15) Not available 3 35 Left atrial diameter Normal (< 39 mm) 2 (18) 43 (41) 0.296 Enlarged (39–50 mm) 8 (73) 51 (50) Severely enlarged (> 50 mm) 1(9) 9 (9) Not available 10 39 SRA risk score High-risk score 20 (95.2%) 32 (22.5%) < 0.001 Table 3 Multivariate analysis Patient characteristics Hazard Ratio Lower 95% CI Higher 95% CI p Previous AF/PAF diagnosis 4.27 0.84 21.7 0.080 CHA2DS2-VASC score 0.96 0.56 1.63 0.870 High-risk score SRA 70.1 7.8 632 0.000 Discussion In this study, electrocardiographic RR dynamics analysis, performed in the first 48 h from stroke unit admission, was able to stratify patients at high or low short-term risk for PAF. Patients at low risk with SRA analysis were unlikely to develop AF during the stay in the SU, whereas about 1/3 patients identified as high risk had a subsequent diagnosis of PAF. It is noteworthy that our analysis did not require additional investigations but were obtained using already available data, i.e., the CCM provided by conventional stroke unit monitors.

nlikely to develop AF during the stay in the SU, whereas about 1/3 patients identified as high risk had a subsequent diagnosis of PAF. It is noteworthy that our analysis did not require additional investigations but were obtained using already available data, i.e., the CCM provided by conventional stroke unit monitors. Our results are in accordance with the previously published studies. In our cohort, the total AF rate was 29%, while PAF was detected in 16.5% of cases. In a recent global survey [10], the AF prevalence among acute stroke patients was 28% while in two systematic reviews, the expected yield of detecting PAF in a standard SU setting was estimated to be 14.7 [1, 2]. Bettin et al. [11] applied SRA analysis to 106 patients with acute stroke of unknown etiology. These patients were followed up with an implantable loop recorder. On long-term monitoring, 13 patients with PAF were detected, 4% in the low-risk group and 33% in the high-risk group.

setting was estimated to be 14.7 [1, 2]. Bettin et al. [11] applied SRA analysis to 106 patients with acute stroke of unknown etiology. These patients were followed up with an implantable loop recorder. On long-term monitoring, 13 patients with PAF were detected, 4% in the low-risk group and 33% in the high-risk group. Rizos et al. [12] investigated 136 acute stroke patients applying SRA analysis to 1–2 h ECG recordings obtained in emergency room. The patients were then followed with CCM for ≥ 48 h after SU admittance. SRA stratified 70/136 patients as low risk and in 7/70 of these PAF was subsequently detect at CCM (this accounting for a false negative rate of 10%). In our sample, only 1/111 patient of the low-risk group developed PAF. This difference between the two studies may depend on the fact that we stratified the PAF risk analyzing with SRA 48 h of CCM rather than 1–2 h of ECG recordings. In fact, RR dynamics, as Rizos et al. [12] found, can consistently vary within contiguous hours (Fig. 1), and thus, the evaluation of longer time periods might provide a more accurate PAF risk. Interestingly, 14/34 patients classified at high risk in the study were then diagnosed with PAF, a similar percentage as in our sample (41 vs 38%).Fig. 1 Examples of RR interval variability during continuous cardiac monitoring. Each plot represents 1 h of monitoring (the Lorentz plots: each RR interval is plotted as a function of the preceding RR interval). a Transitioning from low to high AF risk. b Transitioning from high risk to manifest AF. c Transitioning from high risk to low risk. AF atrial fibrillation, s seconds, Risk SRA clinic risk grade for atrial fibrillation. 0 = low risk; 3 = high risk; 4 = manifest AF (see the “Methods” section)

the preceding RR interval). a Transitioning from low to high AF risk. b Transitioning from high risk to manifest AF. c Transitioning from high risk to low risk. AF atrial fibrillation, s seconds, Risk SRA clinic risk grade for atrial fibrillation. 0 = low risk; 3 = high risk; 4 = manifest AF (see the “Methods” section) The main limitation of this study was the lack of a predefined standard amount of ECG recordings for AF detection. After the first 48 h CCM, the use of additional ECGs, a Holter ECG, or further CCM for the detection of PAF was not standardized and left to the discretion of the treating physicians. Some patients (for example those with more severe strokes) may have received longer CCM and more ECGs/Holter ECGs than others and this may have led to increased opportunities for PAF detection. We did not report the duration of the detected PAF episodes. As pointed out by Kishore et al., in clinical acute stroke studies PAF is frequently poorly defined and there is a lack in a clear distinction between AF/PAF as well as in reporting the length of PAF episodes. The estimation of AF burden might be more important in primary than in secondary stroke prevention. Finally, our multivariate model must be considered exploratory as the relatively small sample size and several factors were associated in univariate analysis with the presence of AF.

reporting the length of PAF episodes. The estimation of AF burden might be more important in primary than in secondary stroke prevention. Finally, our multivariate model must be considered exploratory as the relatively small sample size and several factors were associated in univariate analysis with the presence of AF. In conclusion, using standard CCM monitoring and application of a commercial software algorithm, we have found that acute ischemic stroke patients with increased RR dynamics are at higher risk of PAF during stroke unit admission. This information could be of value to decide on longer term monitoring within the stroke unit hospitalization. Compliance with Ethical Standards This was an investigator-initiated, non-funded, unsponsored study. All procedures performed in this study were performed in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the present study. Conflict of Interest The authors declare that they have no conflict of interest.

Introduction Aneurysmatic subarachnoid hemorrhage (SAH) is a form of hemorrhagic stroke and presents a frequent clinical picture in neurointensive care [1–3]. While the hemorrhage itself and associated global cerebral hypoxia induce fatal brain injury in some patients, clinical outcome in patients surviving the initial hemorrhagic event is thought to be largely determined by cerebral hypoperfusion and large vessel vasospasm (CV) [1–6], which occur in the days and weeks following the bleeding event due to pathophysiological events induced by the subarachnoid hematoma [4]. However, more recent clinical studies have reported discrepancies between CV, cerebral hypoperfusion, and neurological outcome [7–11], attenuating the association of CV with an unfavorable outcome.

the days and weeks following the bleeding event due to pathophysiological events induced by the subarachnoid hematoma [4]. However, more recent clinical studies have reported discrepancies between CV, cerebral hypoperfusion, and neurological outcome [7–11], attenuating the association of CV with an unfavorable outcome. Similar to the disease in humans, mice develop CV and impaired cerebral perfusion after experimental induction of SAH [12–22]. Although these changes occur with faster temporal dynamics in mice than humans—peaking 3 days after SAH in mice [16, 23, 24] compared to 7 to 10 days in humans [3]—murine models have become an important tool for basic research on the pathophysiology of SAH. The majority of experimental studies to date have focused on CV [12–20]. By contrast, little data on cerebral perfusion in mice after SAH have been published [20]. Furthermore, the relation between CV, cerebral hypoperfusion, and neurological deficits in murine SAH models remains unclear despite the fact that, from the clinical perspective mentioned above [5–10], cerebral hypoperfusion may be the most important parameter concerning unfavorable outcome. Thus, better knowledge of the link between vasospasm, hypoperfusion, and neurological deficits in the murine model could benefit future experimental studies. In the present study, we set out to (i) characterize cerebral perfusion in a murine endovascular filament perforation model of SAH, (ii) assess the contribution of CV to changes in cerebral perfusion, and (iii) investigate whether neurological disability is linked to CV and changes in cerebral perfusion.

ental studies. In the present study, we set out to (i) characterize cerebral perfusion in a murine endovascular filament perforation model of SAH, (ii) assess the contribution of CV to changes in cerebral perfusion, and (iii) investigate whether neurological disability is linked to CV and changes in cerebral perfusion. Methods Ethics, Animals, and Housing Conditions The animal experiments were approved by the responsible animal care committee (Landesuntersuchungsamt Rheinland-Pfalz, G12-1-093) and carried out in accordance with the German Animal Welfare Act (TierSchG). All applicable international, national, and institutional guidelines for the care and use of animals were followed. We used male C57BL6 mice (Charles River, Cologne, Germany; age, 11–14 weeks). No other inclusion or exclusion criteria were defined. Mice were kept under controlled environmental conditions (12-h dark/light cycle, 23 ± 1 °C, 55 ± 5% relative humidity) with free access to food (Altromin, Germany) and water. Body weight was recorded daily as a general marker of well-being. A neuroscore (0 to 29 points, with 0 indicating no neurological deficit and 29 indicating severe neurological disability) was determined 1 day prior to and 24 and 72 h after SAH or sham surgery, as previously described [25, 26]. Neuroscores were determined by an investigator blinded to the treatment.

ker of well-being. A neuroscore (0 to 29 points, with 0 indicating no neurological deficit and 29 indicating severe neurological disability) was determined 1 day prior to and 24 and 72 h after SAH or sham surgery, as previously described [25, 26]. Neuroscores were determined by an investigator blinded to the treatment. In a separate set of experiments (data not shown), we observed that (i) mean arterial pressure measured daily before and 3 days after SAH did not differ significantly between SAH and sham groups and that (ii) intracranial pressure (ICP), after its peak during SAH induction, returned to nearly baseline levels only 3 h post-insult, indicating that cerebral perfusion pressure was similar between SAH and sham animals at 3, 24, and 72 h after SAH induction. These findings were consistent with previous reports [27]. Anesthesia and Murine Model of SAH Measurements of cerebral perfusion, SAH induction, and transcardiac perfusion were performed in anesthetized animals. Anesthesia was induced with 4% (v/v) isoflurane for 1 min and maintained with 2% (v/v) isoflurane in spontaneously breathing mice [28]. Body temperature was monitored and maintained at 37 °C with a heating pad during all procedures. For analgesia, buprenorphine (Indivior, Slough, Berkshire, UK, 0.1 mg/kg body weight) was injected subcutaneously twice daily starting at the time of induction of SAH or sham surgery. SAH was induced under continuous monitoring of ICP, as described previously [28]. Animals were randomized to the SAH or sham group prior to surgery.

a, buprenorphine (Indivior, Slough, Berkshire, UK, 0.1 mg/kg body weight) was injected subcutaneously twice daily starting at the time of induction of SAH or sham surgery. SAH was induced under continuous monitoring of ICP, as described previously [28]. Animals were randomized to the SAH or sham group prior to surgery. Laser SPECKLE Contrast Imaging (LSCI) We used a laser perfusion imager (MoorFLPI-2-blood flow imager, Cologne, Germany) to visualize cerebral cortical perfusion of the whole convexity through the intact calvaria before induction of SAH and after 15 min, 3, 24, and 72 h. The animal’s skull was immobilized in a stereotaxic frame (Stoelting CO., IL, USA). A midline incision was made to expose the calvaria. Perfusion images and corresponding photographs were acquired every second for 60 s. After these measurements, the skin was closed using prolene 6.0 (Ethicon, Norderstedt, Germany). Figure 1 illustrates the experimental setting.Fig. 1 Determination of cerebral cortical perfusion. a The perfusion measurement: The animal is mounted on a stereotaxic frame. After the skin incision, the laser SPECKLE camera is placed over the animal to acquire perfusion images. b The evaluation of cerebral perfusion (upper image: photograph, lower image: flux image visualizing cerebral cortical perfusion). A region of interest of 7 mm2 is placed on the left parietal region to measure perfusion flux values

ncision, the laser SPECKLE camera is placed over the animal to acquire perfusion images. b The evaluation of cerebral perfusion (upper image: photograph, lower image: flux image visualizing cerebral cortical perfusion). A region of interest of 7 mm2 is placed on the left parietal region to measure perfusion flux values Analysis of Cerebral Perfusion of the Left MCA Territory LSCI data were evaluated using Moor review software (moorFLPI Full-Field Laser Perfusion Imager Review Version 4.0). A mean image was calculated from the 60 perfusion images. Mean flux values were determined from a region of interest (ROI) of 7 mm2 placed on the vascular territory of the left MCA [29] (see Fig. 1). Perfusion was evaluated by an investigator blinded to the treatment.

l-Field Laser Perfusion Imager Review Version 4.0). A mean image was calculated from the 60 perfusion images. Mean flux values were determined from a region of interest (ROI) of 7 mm2 placed on the vascular territory of the left MCA [29] (see Fig. 1). Perfusion was evaluated by an investigator blinded to the treatment. Perfusion, Intravascular Casting, Microcomputed Tomography (Micro-CT) of Murine Brains, and Volumetric Analysis of CV After LSCI measurement 72 h after SAH, animals were sacrificed by a perfusion casting procedure with a radiopaque compound (Microfil MV-122, Flow Tech Inc., Carver, Massachusetts, USA), as previously described [28, 30]. Brain samples were scanned with a micro-CT (μCT40, Scanco Medical AG, Brüttisellen, Switzerland) with power settings of 70 kVp and 114 μA and a voxel size of 20 μm. Resulting DICOM data were imported to Amira® software, version 5.4.2 (FEI Visualization Sciences Group, Hillsboro, OR, USA). The intracranial vascular tree was reconstructed and the vessel volume of a defined 2.5-mm vessel segment of the MCA distal of the carotid T was calculated as previously described [28]. Volumetric analysis was selected on the basis of a methodological study, which compared vessel volumes and diameters in mice with SAH-induced vasospasm [28] and showed that the analysis of vessel diameters can determine vasospasm-induced vascular changes with a greater sensitivity. Figure 2 illustrates the method. CV was evaluated by an investigator blinded to the treatment.Fig. 2 Determination of CV. a, b Determination of cerebral vessel volumes: the cerebrovascular tree is virtually reconstructed from Dicom data after micro-CT scanning of the brain (a). Afterwards, a defined vessel segment is selected and the vessel volume is calculated as a marker for CV (b)

or blinded to the treatment.Fig. 2 Determination of CV. a, b Determination of cerebral vessel volumes: the cerebrovascular tree is virtually reconstructed from Dicom data after micro-CT scanning of the brain (a). Afterwards, a defined vessel segment is selected and the vessel volume is calculated as a marker for CV (b) Statistics Data are presented as mean ± standard error of the mean. The Mann–Whitney U test was used for statistical testing. Correlation analysis was calculated using Pearson’s correlation coefficients. The level of p < 0.05 was considered statistically significant. Group sizes were chosen after power analysis of pilot data (not shown) using Sigma Plot version 12.5 (Systat Software Inc., San Jose, CA, USA) with alpha of 0.05, power of 0.8, and expected differences standard deviations in vessel volumes of 0.003 ± 0.002 μL and in perfusion values of 25 ± 15% between SAH and sham groups.

p sizes were chosen after power analysis of pilot data (not shown) using Sigma Plot version 12.5 (Systat Software Inc., San Jose, CA, USA) with alpha of 0.05, power of 0.8, and expected differences standard deviations in vessel volumes of 0.003 ± 0.002 μL and in perfusion values of 25 ± 15% between SAH and sham groups. Results Murine Model of SAH Fifteen animals were included. Five were assigned to the sham group and ten to the SAH group. Two SAH animals died (days 1 and 3 after SAH) and were thus excluded from analysis. Mean duration of surgery was similar between SAH and sham groups (SAH, 35.8 ± 3.0 min; sham, 37.0 ± 3.2 min). Baseline ICP was not significantly different between SAH and sham animals (SAH, 9.3 ± 1.8 mmHg; sham, 12.0 ± 2.6 mmHg). SAH induction resulted in a sharp increase in ICP with a peak at 71.8 ± 3.5 mmHg, whereas, during vascular insertion of the filament during sham surgery, we observed an ICP of 13.2 ± 2.1 mmHg. Loss of body weight was not significantly different between SAH and sham animals (before surgery, SAH 25.1 ± 0.5 g, sham, 26.4 ± 1.1 g; postop day 1, SAH 22.9 ± 0.7 g, sham, 23.8 ± 1.0 g; postop day 2, SAH 21.4 ± 0.7 g, sham, 22.8 ± 1.2 g; postop day 3, SAH 21.3 ± 0.6 g, sham, 22.7 ± 1.4 g). Neuroscores were higher for SAH animals (SAH vs. sham, 0.4 ± 0.4 vs. 0.0 ± 0.0 (d0); 5.6 ± 0.9 vs. 2 ± 0.4 (d1); 9.3 ± 1.6 vs. 5.8 ± 1.5 (d3)), although the difference reached statistical significance (p < 0.05) only on day 1. All of the SAH animals, but none of the sham animals, showed a subarachnoid hematoma.

1.4 g). Neuroscores were higher for SAH animals (SAH vs. sham, 0.4 ± 0.4 vs. 0.0 ± 0.0 (d0); 5.6 ± 0.9 vs. 2 ± 0.4 (d1); 9.3 ± 1.6 vs. 5.8 ± 1.5 (d3)), although the difference reached statistical significance (p < 0.05) only on day 1. All of the SAH animals, but none of the sham animals, showed a subarachnoid hematoma. SAH Causes Cerebral Hypoperfusion Flux values determined before induction of SAH were similar between SAH and sham animals. Induction of SAH induced a significant decrease in cortical perfusion in SAH animals compared to values before SAH (SAH, 35.7 ± 3.1%, sham, 101.4 ± 10.2%, p < 0.01). In SAH animals, perfusion partly recovered after 3, 24, and 72 h but remained significantly lower compared to that in sham (SAH vs. sham, 3 h, 85.0 ± 8.6 vs. 121.9 ± 13.4, p < 0.05; 24 h, 75.3 ± 4.6 vs. 110.6 ± 11.4%, p < 0.01; 72 h, 81.8 ± 4.8 vs. 108.5 ± 14.5%, n.s.; Fig. 3).Fig. 3 SAH induces cerebral hypoperfusion. a An anatomical image and representative perfusion images of the same animal before SAH and 15 min, 3, 24, and 72 h after SAH. b, c The course of cerebral perfusion after SAH and sham surgery. Note that perfusion is significantly impaired after SAH compared to sham (*p < 0.05, **p < 0.01). d, e The correlation between cerebral perfusion and neuroscore at 24 and 72 h

SAH Causes Cerebral Hypoperfusion Flux values determined before induction of SAH were similar between SAH and sham animals. Induction of SAH induced a significant decrease in cortical perfusion in SAH animals compared to values before SAH (SAH, 35.7 ± 3.1%, sham, 101.4 ± 10.2%, p < 0.01). In SAH animals, perfusion partly recovered after 3, 24, and 72 h but remained significantly lower compared to that in sham (SAH vs. sham, 3 h, 85.0 ± 8.6 vs. 121.9 ± 13.4, p < 0.05; 24 h, 75.3 ± 4.6 vs. 110.6 ± 11.4%, p < 0.01; 72 h, 81.8 ± 4.8 vs. 108.5 ± 14.5%, n.s.; Fig. 3).Fig. 3 SAH induces cerebral hypoperfusion. a An anatomical image and representative perfusion images of the same animal before SAH and 15 min, 3, 24, and 72 h after SAH. b, c The course of cerebral perfusion after SAH and sham surgery. Note that perfusion is significantly impaired after SAH compared to sham (*p < 0.05, **p < 0.01). d, e The correlation between cerebral perfusion and neuroscore at 24 and 72 h SAH Induces Vasospasm We assessed CV after perfusion and casting of the mice 72 h after SAH or sham surgery. This time point was selected to determine the impact on cortical perfusion at the peak of CV [16, 23, 24]. We analyzed the vessel volume of a 2.5-mm vessel segment of the left MCA distal of the carotid T, which is a highly sensitive parameter for evaluating vasospasm [28]. Vessel volumes of the left MCA were significantly lower in SAH animals compared to sham animals (SAH, 5.6 ± 0.6 nL; sham, 8.3 ± 0.5 nL, p < 0.01), indicating CV in the SAH group (Fig. 4).Fig. 4 CV and cerebral perfusion after SAH. a Vessel volumes of a 2.5-mm segment of the MCA distal of the carotid T after SAH and sham surgery. Note that the vessel volumes are significantly lower in SAH animals, indicating CV. b, c The correlation of MCA vessel volume with perfusion and neuroscore. d, e Exemplarily show the reconstructed vascular tree and cerebral perfusion in a sample without CV (d) and with CV (e), in which CV was not associated with impaired perfusion

e vessel volumes are significantly lower in SAH animals, indicating CV. b, c The correlation of MCA vessel volume with perfusion and neuroscore. d, e Exemplarily show the reconstructed vascular tree and cerebral perfusion in a sample without CV (d) and with CV (e), in which CV was not associated with impaired perfusion Correlation of Cerebral Perfusion, Vasospasm, and Neuroscore Cerebral perfusion was determined in a ROI representing the vascular territory of the left MCA [29, 31]. To assess the influence of vasospasm on cerebral perfusion, we correlated vessel volumes of sham operated animals (normal vessel diameter) and SAH animals (vasospasm) with cortical perfusion images on day three. There was no significant correlation between the parameters (r = 0.34, p = 0.25). Interestingly, in some cases, pronounced MCA vasospasm was not associated with cortical hypoperfusion of the MCA territory (Fig. 4). To assess whether impaired cerebral perfusion and vasospasm were linked to neurological disability, we correlated cerebral perfusion with the corresponding neuroscore. There was a moderate correlation of perfusion with neuroscore (24 h: r = − 0.58, p < 0.05; 72 h: r = − 0.44, p = 0.14), which, however, did not reach statistical significance at 72 h (Fig. 3). There was no significant correlation between vessel volume and neuroscore (r = − 0.21, p = 0.50, Fig. 4).

esponding neuroscore. There was a moderate correlation of perfusion with neuroscore (24 h: r = − 0.58, p < 0.05; 72 h: r = − 0.44, p = 0.14), which, however, did not reach statistical significance at 72 h (Fig. 3). There was no significant correlation between vessel volume and neuroscore (r = − 0.21, p = 0.50, Fig. 4). In the next step, we used SAH data alone to determine if vessel diameter was correlated with cortical perfusion in animals with vasospasm to determine if differences in degree of vasospasm resulted in alterations of cortical perfusion. In this data set, no correlation between vasospasm, perfusion, and neuroscore was present. However, the power to assess this relationship may have been limited by the small group size and relatively high homogeneity in vessel volume between SAH animals.

degree of vasospasm resulted in alterations of cortical perfusion. In this data set, no correlation between vasospasm, perfusion, and neuroscore was present. However, the power to assess this relationship may have been limited by the small group size and relatively high homogeneity in vessel volume between SAH animals. Discussion To the best of our knowledge, this study was the first to analyze the relationship between CV, cerebral perfusion, and neurological impairment in a murine model of SAH. We determined cortical cerebral perfusion of the whole convexity through the intact calvaria using laser SPECKLE contrast imaging. In this way, we were able to analyze the course of cerebral perfusion at several time points within the first 72 h of SAH induction and to correlate perfusion of the MCA vascular territory with CV of the MCA and quantitative neuroscore. The most important findings of our study are (i) the lack of a clear correlation between CV and perfusion or neuroscore and (ii) the correlation between cortical perfusion and neuroscore, which became weaker at later time points. Taken together, these data indicate that factors other than CV play a major role in the alteration of cerebral perfusion after SAH, and that local cerebral malperfusion, rather than CV, determines neurological impairment.

e correlation between cortical perfusion and neuroscore, which became weaker at later time points. Taken together, these data indicate that factors other than CV play a major role in the alteration of cerebral perfusion after SAH, and that local cerebral malperfusion, rather than CV, determines neurological impairment. In human patients, CV typically occurs within 14 days of SAH. As CV can lead to cerebral infarction, it is thought to be associated with a poor outcome [1–3]. Therefore, CV has become a key target of novel pharmaceutical agents in recent years. However, a number of studies achieved reduction of angiographic vasospasm without the expected improvement of clinical outcome [7–11]. In contrast, a study with nimodipine [32] achieved significant improvement in outcomes without significant changes in rate of CV. The interpretation of these findings is complicated because, in a clinical setting, factors such as adverse effects of pharmaceutical agents or rescue therapies could conceal the beneficial effect of amelioration of CV. Nevertheless, these clinical data imply that factors other than CV play an important role in outcome. Taking our experimental observations into account, neurological impairment appears to be linked to local changes in perfusion rather than to large vessel vasospasm of, for example, the MCA. Another important aspect of our study is that neurological impairment more weakly correlated with cortical perfusion at 72 h compared with that at 24 h. This could indicate that with increasing time after SAH, the role of cortical perfusion as a key driver of neurological disability declines. This is in line with clinical studies [4, 5], which have shown the high relevance of cerebral hypoperfusion early after SAH for unfavorable outcome. Therefore, future SAH treatment research should focus on causes of and strategies for cerebral perfusion, starting at an early time point after the bleeding event.

clines. This is in line with clinical studies [4, 5], which have shown the high relevance of cerebral hypoperfusion early after SAH for unfavorable outcome. Therefore, future SAH treatment research should focus on causes of and strategies for cerebral perfusion, starting at an early time point after the bleeding event. CV depicted on the angiogram is one factor, but not the primary factor, contributing to cerebral hypoperfusion. Studies conducted in recent years have identified additional factors that can occur independently of CV and impair cerebral microcirculation. In an autopsy series, Stein et al. observed microthrombi in patients after SAH [33]. In another study, Uhl et al. [34] described narrowing of microvessels early after SAH. These pathological changes at the level of microvessels were also recently described in a murine SAH model [21, 22]. Given that cerebrovascular resistance is primarily determined by vasoconstriction of microvessels [35], narrowing and thrombosis of microvessels has the potential to markedly affect cerebrovascular resistance and perfusion independently of CV. Therefore, we assume that increased cerebrovascular resistance caused by changes at the level of the microcirculation is the most likely factor responsible for cerebral hypoperfusion.

rowing and thrombosis of microvessels has the potential to markedly affect cerebrovascular resistance and perfusion independently of CV. Therefore, we assume that increased cerebrovascular resistance caused by changes at the level of the microcirculation is the most likely factor responsible for cerebral hypoperfusion. Finally, we want to address the limitations of this study. Cerebral perfusion is dependent on cerebral perfusion pressure. In line with data by Feiler et al. [27], we determined the two key factors of intracranial pressure and arterial blood pressure in a separate set of animals and did not find significant differences at days one and three after SAH. To limit the distress imposed upon the animals, we did not perform these analyses in the current study and so cannot present ICP and blood pressure data. We therefore cannot rule out the possibility that low cerebral perfusion pressure contributed to impaired cortical perfusion. Secondly, we measured cortical perfusion at different time points, whereas vasospasm was only determined 72 h after SAH. In addition, perfusion was examined in vivo, whereas vasospasm was analyzed ex vivo. The time point of 72 h was chosen because our data and other studies have reported that vasospasm peaks 72 h after induction of SAH in the murine endovascular filament perforation model [16, 23, 24]. Although we cannot rule out that a correlation between vasospasm and cortical perfusion may have been present at earlier time points, the lack of correlation at 72 h indicates that large vessel vasospasm is of limited relevance in the murine model. Third, it should be noted that perfusion levels exceeded the preoperative value after 3, 24, and 72 h in the sham group. These changes are most likely related to the isoflurane anesthesia, which is reported to have a cerebral vasodilatory effect [36]. It is therefore important to mention that the experimental conditions were similar between SAH and sham groups in this study. Finally, we only examined whether cortical perfusion correlated with vessel volume using sham animals for normal vessel diameter and SAH for vasospasm. Our study does not allow conclusions on whether different degrees of SAH-induced vasospasm correlate with hypoperfusion or neurological impairment. Although there are differences between the SAH animals, the model used here is designed to produce hemorrhage with high reproducibility. A larger number of animals would be required to assess correlations within the SAH group.

ent degrees of SAH-induced vasospasm correlate with hypoperfusion or neurological impairment. Although there are differences between the SAH animals, the model used here is designed to produce hemorrhage with high reproducibility. A larger number of animals would be required to assess correlations within the SAH group. One last note of caution is warranted. Our data imply that CV is only of minor importance with regard to cerebral perfusion and neurological impairment in the murine model, suggesting that other factors play a prominent role, like elevated cerebrovascular resistance or microthrombi formation at the level of microcirculation. However, in contrast to humans, territorial infarctions as a result of CV after SAH occur only rarely in mice. This could indicate that, in the murine model, the contribution of CV to cerebral perfusion is less important, while the contribution of microcirculatory changes is more important, compared to human cases. Furthermore, in a clinical setting, aggressive treatment of CV may be the only option to improve cerebral hypoperfusion in cases of threatening cerebral infarction [1–3]. Therefore, although our findings underline the importance of future research efforts towards amelioration of cerebral perfusion by targeting CV-independent mechanisms, CV monitoring and, in some cases, aggressive treatment of CV remain a cornerstone of SAH therapy.

ion in cases of threatening cerebral infarction [1–3]. Therefore, although our findings underline the importance of future research efforts towards amelioration of cerebral perfusion by targeting CV-independent mechanisms, CV monitoring and, in some cases, aggressive treatment of CV remain a cornerstone of SAH therapy. Parts of this study are from the doctoral thesis of S. Kunzelmann, presented to the Medical Faculty of the Johannes Gutenberg-University of Mainz. The authors would like to thank Dr. Alicia Poplawski (Institute of Medical Biostatistics, Epidemiology and Informatics, University Medical Center, Mainz, Germany) for statistical advice and Stefan Kindel (Department of Neurosurgery, University Medical Center, Mainz, Germany) for the illustration in Fig. 1. Funding The study was supported by a grant of the Medical Center of the Johannes Gutenberg University Mainz (Stufe I Foerderung, grant to A.N.). The funder had no role in the design or conduct of this research. Compliance with Ethical Standards Statement on the Welfare of Animals The animal experiments were approved by the responsible animal care committee (Landesuntersuchungsamt Rheinland-Pfalz, G12-1-093) and carried out in accordance with the German Animal Welfare Act (TierSchG). All applicable international, national, and institutional guidelines for the care and use of animals were followed. Conflict of Interest The authors declare that they have no conflict of interest.

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is a relatively rare hereditary small vessel disease resulting in neurological deficits [1]. Stroke and transient ischaemic attack (TIA) are reported in the majority with symptomatic CADASIL, driving presentation in 71% in a large case series, migraine is common (23%) and life expectancy is reduced [1, 2]. CADASIL can also contribute to vascular dementia. The exact prevalence of CADASIL is unknown as it is under-recognised even in specialist stroke settings, but the minimum prevalence is estimated at 2–5 in 100,000 [1]. Previous CADASIL studies have generally involved genotyping for patients presenting with repeated strokes, with several missense (amino-acid substituting) mutations in the NOTCH3 gene reported [3]. Many, often rare, mutations have been linked to CADASIL. However, population-based estimates of the predictive value of NOTCH3 missense mutations in community samples are unknown. With increasing availability of genotyping, including in clinical biobanks and direct to consumers, data are needed on the true penetrance of NOTCH3 mutations.

utations have been linked to CADASIL. However, population-based estimates of the predictive value of NOTCH3 missense mutations in community samples are unknown. With increasing availability of genotyping, including in clinical biobanks and direct to consumers, data are needed on the true penetrance of NOTCH3 mutations. We aimed to estimate cerebrovascular outcomes associated with all missense (amino-acid altering) likely pathogenic variants in NOTCH3 in 451,424 participants of European descent in the UK Biobank (UKB) volunteer study. UKB community volunteers consented for genotyping and baseline (2006–2010) physiological measures, including systolic and diastolic blood pressure (BP), plus demographics and medical history [4]. The cohort was linked to national hospital inpatient and death certificate records and followed up for a mean of 7 years (maximum 10). We searched UKB Affymetrix Axiom array genotype data with imputation (v3, ~ 93million variants) for all amino-acid substituting variants in NOTCH3 (n = 131), excluding low imputation quality (< 80%) and in linkage disequilibrium (R2 > 0.3), leaving 22 variants for follow-up. We used the Variant Effect Predictor (http://grch37.ensembl.org/Homo_sapiens/Tools/VEP, accessed: 17th May 2018) tools to prioritise variants with potential pathogenic effects; this process identified rs201680145 (p.Arg1231Cys: 0.04% mutation prevalence in UKB) and rs35769976 (p.Ala1020Pro; 0.96% mutation prevalence in UKB) only. Logistic regression tested cross-sectional associations between genotype and baseline BP, prevalent stroke or TIA, and prevalent Coronary Heart Disease (CHD). We used Cox proportional hazards regression models for all-cause mortality and Fine and Gray competing risk models for incident stroke, with all-cause mortality as the competing risk. Those with prior strokes were included in the analysis of incident strokes as CADASIL is linked to recurrent strokes. Models were adjusted for age, sex and technical genetic covariates (including genetic principal components to account for population admixture). Analyses were conducted using Stata v14.1. The rs201680145 variant has been reported as pathogenic for CADASIL by two sources [5]. There were 176 UKB participants heterozygous for the missense A allele of rs201680145 (p.Arg1231Cys/p.R1231C, prevalence 0.04% of 451,424) and no homozygous participants. Mean age was 56.8 years (SD 7.7, median 58, IQR 13).

a v14.1. The rs201680145 variant has been reported as pathogenic for CADASIL by two sources [5]. There were 176 UKB participants heterozygous for the missense A allele of rs201680145 (p.Arg1231Cys/p.R1231C, prevalence 0.04% of 451,424) and no homozygous participants. Mean age was 56.8 years (SD 7.7, median 58, IQR 13). rs201680145 heterozygotes had modest increases in baseline systolic (3.12 mmHg; 95% confidence interval 0.28 to 5.96) and diastolic blood pressure (BP) (2.02 mmHg, 0.39 to 3.64), compared to no mutations. Figure 1 shows the percentage of subjects developing incident stroke or TIA, by NOTCH3 variant genotype. During the follow-up period, four of 176 A allele rs201680145 heterozygotes experienced incident strokes or TIAs (2.27%), compared to 0.88% (3874/442,613) with no mutation: sub-HR 3.21, 1.20 to 8.56; p = 0.020. Incident CHD or all-cause mortality was not associated with rs201680145.Fig. 1 Graph of percentage with incident stroke or TIA during follow-up by NOTCH3 variant genotype There were 8596 heterozygotes for rs35769976 and only 44 G allele homozygotes (0.01% of 451,424), so we grouped homozygotes and heterozygotes. The mean age of the sample was 56.8 years (SD 8.0, median 58, IQR 12). Participants with at least one rs35769976 G allele had a slightly increased diastolic BP (coefficient 0.24, 0.0079 to 0.47; p value = 0.043). There were no strokes or TIAs in the G allele homozygotes and 77 in heterozygotes. The rs35769976 variant was not associated with incident stroke or TIA, CHD or all-cause mortality.

ticipants with at least one rs35769976 G allele had a slightly increased diastolic BP (coefficient 0.24, 0.0079 to 0.47; p value = 0.043). There were no strokes or TIAs in the G allele homozygotes and 77 in heterozygotes. The rs35769976 variant was not associated with incident stroke or TIA, CHD or all-cause mortality. In a European Ancestry community volunteer biobank, we have shown that only two potentially pathogenic NOTCH3 missense variants were ascertained in array genotyping data with imputation. rs201680145, previously reported as pathogenic by two sources [5], was associated with elevated systolic and diastolic BP, plus substantially increased risk of incident stroke or TIA. However, the great majority of rs35769976 carriers did not have a stroke or TIA, suggesting that this mutation is associated with a milder clinical course. If these mutations were associated with high rates of serious strokes, we would expect prevalence of the mutations to be lower at older ages, but neither mutation was associated with age in UK biobank. A limitation of our analysis is that the maximum age at end follow-up was 80 years. However, as the cohort included those between 41 and 70 years at baseline, with up to 10-year follow-up, the important time for early strokes is captured. In addition, while 10-year follow-up is longer than in most reported CADASIL studies, data on even longer follow-ups would be helpful. This would also allow analysis of repeat events in individuals, as numbers with recurrent stroke or TIA in this cohort were too small to analyse.

ant time for early strokes is captured. In addition, while 10-year follow-up is longer than in most reported CADASIL studies, data on even longer follow-ups would be helpful. This would also allow analysis of repeat events in individuals, as numbers with recurrent stroke or TIA in this cohort were too small to analyse. This study suggests that potentially pathogenic NOTCH3 missense mutations seen in large European ancestry array genotyped biobanks have relatively limited CADASIL risk profiles. Caution is needed in responding to incidental data on these mutations in community volunteer samples. More work is needed to establish whether co-factors (including higher blood pressures) modify risks of clinical outcomes. Funding This report is independent research supported by the National Institute for Health Research (NIHR Doctoral Research Fellowship, Dr. Jane Masoli, DRF-2014-07-177). Compliance with Ethical Standards Informed Consent This research has been conducted using the UK Biobank resource under Application Number 14631. We thank UK Biobank participants and coordinators for this dataset. Informed consent was obtained from all individual participants included in the study. Conflict of Interest The authors declare that they have no conflict of interest. Disclaimer The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.

i Yang in the Department of Pharmacy for providing technical help with the HPLC analysis, Kailong Zhu in the Department of Anesthesiology for assisting with the animal behavioral scoring after tMCAO, and Ting Gu in the Department of Anesthesiology for her kind assistance in isolating the primary neurons and astrocytes. Funding This study was supported by the National Natural Science Foundation of China (81420108013, 81471110, and 81671184), the National Basic Research Program of China (2014CB543202), and the Young Scholar Research Grant of Chinese Anesthesiologist Association (220160900001). Compliance with Ethical Standards Conflict of Interest The authors declare that they have no conflicts of interest. Ethical Approval All applicable international and national guidelines for the care and use of animals were followed.

). Percent of ki67+/HuNu+ drNPCs at Day 8, 8.5% ± 1.5 (aCSF) and 11.0% ± 2.0 (HAMC); and Day 32, 6.8% ± 1.3 (aCSF) and 8.2% ± 1.0 (HAMC). d IHC reveals that drNPCs (HuNu+) remain at the boundary of the stroke injury (demarcated by NeuN+ cells) following transplantation into the stroke-injured cortex 4 days post-stroke. e Surviving drNPCs at 32 days post-stroke mostly remain undifferentiated (Sox2, Nestin). The majority of differentiated drNPCs primarily gave rise to astrocytes (GFAP), while a smaller population gave rise to immature neurons (TUJ1). No drNPCs differentiated to mature neurons (NeuN) or oligodendrocytes (Olig2). Data are represented as mean ± SEM, a Arrowheads indicate Ki67+/HuNu+ cells. e Arrowheads indicate colocalization between markers. a, d Scale bars = 100 μm (b), n = 4 for both vehicles at day 8 and n = 7 for both vehicles at day 32. cn = 4 for both vehicles at day 8, and n = 6 for aCSF and n = 3 for HAMC at day 32 (e) Scale bar = 50 μm, at least three brains were analyzed per marker and representative images were used

Introduction Stroke is diagnosed based on patient history, neurological exam, and brain imaging. Differential diagnosis can be difficult particularly for distinguishing ischemic stroke (IS) from intracerebral hemorrhage (ICH) when imaging is unavailable in the acute setting. Thus, an accurate, inexpensive, and rapid blood test would be useful. Blood transcriptomes show promise as diagnostic biomarkers and have provided insight into understanding the nature of the immune response following human stroke [1–6]. However, these studies have investigated only a portion of the protein coding transcriptome, since they have used 3′-biased microarrays to measure blood mRNA expression [1–6]. Though these studies demonstrated proof-of-principle, most of the stroke transcriptome which is comprised of all alternatively spliced isoforms remains unstudied in stroke. The importance of alternative splicing is supported by evidence implicating it in the pathogenesis of many diseases [7, 8] .

Introduction Ischemic strokes are the most common type of stroke. Occlusion of cerebrovascular blood flow resulting in a lack of glucose and oxygen delivery to the brain results in rapid cell death and impaired neural function within the affected regions. The resulting functional deficits have a significant impact on an individual’s quality of life and current treatment strategies offer limited success [1–5]. Most available therapies for stroke focus on restoring blood flow and neuroprotection, which have a limited therapeutic window. Cell-based interventions to repair the stroke-injured brain and promote functional recovery have demonstrated some therapeutic efficacy [6–9]. However, a number of challenges including the identification of an appropriate cell type for transplantation that circumvents immune rejection, tumorigenicity, ethical concerns, misguided or misdirected growth, and limited availability in terms of cell isolation and expansion remain [10–13].

therapeutic efficacy [6–9]. However, a number of challenges including the identification of an appropriate cell type for transplantation that circumvents immune rejection, tumorigenicity, ethical concerns, misguided or misdirected growth, and limited availability in terms of cell isolation and expansion remain [10–13]. Neural precursor cells (NPCs), comprised of neural stem cells and their progeny, have the capacity to differentiate into neural specific cell types, making them good candidates to repair the stroke-injured brain. Although their underlying mechanism of action is not entirely clear, NPCs have demonstrated efficacy in treating several models of stroke, resulting in improved outcomes, including better functional performance, decreased glial scarring, and reduced extent of injury [14–21]. However, harvesting human NPCs is challenging due to their limited availability and location within the brain. Other potential sources of human NPCs are those derived from reprogrammed cells, such as induced pluripotent stem cells (iPSCs), which offer an autologous cell source for transplantation. Unfortunately, iPSCs pose concerns for clinical application because of their acquired pluripotent state during reprogramming in addition to the length of time and complexity required to generate sufficient numbers of cells.

duced pluripotent stem cells (iPSCs), which offer an autologous cell source for transplantation. Unfortunately, iPSCs pose concerns for clinical application because of their acquired pluripotent state during reprogramming in addition to the length of time and complexity required to generate sufficient numbers of cells. With clinical translation in mind, we examined the therapeutic potential of a population of human cells that have been directly reprogrammed from somatic cells to NPCs, without passing through a pluripotent state during reprogramming. We address important considerations that may influence transplant success, including the transplant vehicle [22], and the sex of the stroke-injured mice, which has not been adequately studied to date despite the observation that males and females are differentially responsive to stroke injury [12, 23–26]. Furthermore, we explore the importance of cell survival for recovery and investigate changes in synaptogenesis as a mechanism underlying cell-mediated effects.

d mice, which has not been adequately studied to date despite the observation that males and females are differentially responsive to stroke injury [12, 23–26]. Furthermore, we explore the importance of cell survival for recovery and investigate changes in synaptogenesis as a mechanism underlying cell-mediated effects. Using a preclinical model of cortical stroke, we demonstrate that human directly reprogrammed neural precursor cell (drNPC) transplants delivered during the subacute phase of stroke are sufficient to elicit motor recovery irrespective of recipient sex and transplant vehicle. The observed functional recovery was not correlated with the extent of glial scarring or lesion volumes and did not require long-term xenograft survival. Furthermore, drNPC transplants appear to promote synaptogenesis, as indicated by increased expression of the presynaptic vesicle protein synaptophysin in the ipsilesional hemisphere of transplanted brains. These findings suggest that NPCs may indirectly promote functional recovery by influencing the surrounding tissue, making drNPCs a promising population of cells to treat the stroke-injured brain.

increased expression of the presynaptic vesicle protein synaptophysin in the ipsilesional hemisphere of transplanted brains. These findings suggest that NPCs may indirectly promote functional recovery by influencing the surrounding tissue, making drNPCs a promising population of cells to treat the stroke-injured brain. Materials and Methods Study Design The experimental design was a controlled laboratory experiment. Male and female animals were used and separated into groups via random assignment by a blinded third party until appropriate numbers of samples were achieved for each group. Behavioral analysis was conducted by an observer blinded to the treatment groups. Tissue and cellular outcomes were evaluated by three separate observers blinded to the experimental groups. For functional analysis, 10–15 mice per group were analyzed. Using a sensitivity analysis on G*Power (version 3.1) with power = 0.8, we determined the treatment effect size to be ~ f(V) = 0.65 (η2partial = 0.3) when analyzing all four treatment groups, f(V) = 0.42 when comparing between drNPC and vehicle groups, and f(V) = 0.93 when comparing between brains that had surviving drNPCs and those that did not. For tissue outcome comparisons between mice that received drNPCs or vehicle alone, the effect size was always above d = 0.9 (Gliosis, 0.97; Lesion Volume, 1.51; Synaptophysin, 1.60) with a power = 0.8. All effect size (sensitivity) calculations were based on Cohen’s d [27]. Excluded animals were not considered for power and sensitivity analyses.

ons between mice that received drNPCs or vehicle alone, the effect size was always above d = 0.9 (Gliosis, 0.97; Lesion Volume, 1.51; Synaptophysin, 1.60) with a power = 0.8. All effect size (sensitivity) calculations were based on Cohen’s d [27]. Excluded animals were not considered for power and sensitivity analyses. Animals Immunocompromised Fox Chase SCID/Beige (8–16 weeks old; CB17.Cg-PrkdcscidLystbg-J/Crl; Charles River Laboratories, Wilmington, MA) mice were singly housed on a 12-h light/dark cycle with food and water provided ad libitum for the duration of testing, starting 3 days prior to stroke, until sacrifice. A total of 87 mice (establishment of stroke, n = 13, [sex not tracked]; long-term deficit analysis for ET-1 stroke, n = 13 [7 males, 6 females]; confirmation of stable measures in long-term testing of naïve mice, n = 8 [6 males, 2 females]; therapeutic evaluation of drNPC transplants, n = 53 [30 males, 23 females]) were used in this study. Outliers and mice that did not meet our inclusion criteria were removed from the study as described in the supplementary materials.

tion of stable measures in long-term testing of naïve mice, n = 8 [6 males, 2 females]; therapeutic evaluation of drNPC transplants, n = 53 [30 males, 23 females]) were used in this study. Outliers and mice that did not meet our inclusion criteria were removed from the study as described in the supplementary materials. Stroke Injury ET-1 stroke was performed in SCID/Beige mice as previously described [17, 28, 29]. Briefly, the skull was exposed, a small burr hole was drilled at the site of the right sensorimotor cortex at + 0.6 mm anterior and − 2.25 mm lateral to bregma. Mice received a 1-μL injection of ET-1 (Calbiochem, 800 picomolars) 1 mm deep from the surface of the brain at a rate of 0.1 μL/min using a 2.5 μL Hamilton Syringe with a 26 gage, 0.375″ long needle (Hamilton, Reno, NV). The needle was kept in place for 10 min following the injection and then slowly withdrawn. drNPC Reprogramming and Preparation Human bone marrow cells were reprogrammed with transient expression of transcription factors musashi-1 (Msi1), neurogenin-2 (Ngn2), and methyl-CpG binding domain protein 2 (MBD2), as described in detail in the supplementary material. The cells were cultured until they reached ~ 80% confluence in each passage and collected for transplantation after 4–9 passages. Sister cultures were prepared for in vitro analysis to characterize the cells using immunocytochemistry and PCR.

omain protein 2 (MBD2), as described in detail in the supplementary material. The cells were cultured until they reached ~ 80% confluence in each passage and collected for transplantation after 4–9 passages. Sister cultures were prepared for in vitro analysis to characterize the cells using immunocytochemistry and PCR. Cell Transplantation For transplantation, drNPCs were suspended in artificial cerebrospinal fluid (aCSF) or Hyaluronan Methylcellulose (HAMC; supplementary methods) hydrogel (100,000 cells/μL). Cells were transplanted into the stroke site 4 days following stroke with the same surgical procedures used for ET-1-induced ischemia in the sensorimotor cortex. Control animals received 1 μL injections of aCSF or HAMC only. Live/Dead Assay A live/dead assay was performed to determine the percent of surviving cells post-injection through the syringe [30]. Using the same protocol as was used for transplantation, 100,000 cells in 1 μL of vehicle were injected into a well containing 14 μL of warm (37 °C) aCSF at a rate of 0.1 μL/min. Following injection, 15 μL of the live/dead stain solution (0.2% Ethidium Homodimer and 0.05% calcein AM) (L3224, ThermoFisher) was added, followed by a 5-min incubation period after which 220 μL aCSF was added for a final volume of 250 μL. The solution was imaged using an AxioVision Zeiss UV microscope (5× magnification) and images were visualized in FIJI [31]. The percent of live/dead cells was calculated using the “analyze particles” feature on the FIJI software.

a 5-min incubation period after which 220 μL aCSF was added for a final volume of 250 μL. The solution was imaged using an AxioVision Zeiss UV microscope (5× magnification) and images were visualized in FIJI [31]. The percent of live/dead cells was calculated using the “analyze particles” feature on the FIJI software. Immunostaining Fixed tissue and cells were rinsed with 1× phosphate buffered saline (PBS), permeabilized with 0.3% Triton-X100 in 0.01 M PBS for 20 min and blocked in 10% Normal Goat Serum (NGS) with 0.3 M glycine for 1 h at room temperature. Samples were then treated with primary antibodies (Table 1) in 0.01 M PBS and left overnight at 4 °C in a humid chamber. Samples were washed with 1× PBS and exposed to secondary antibodies for 1 h at room temperature (Table 2). The samples were then washed, cover-slipped with mowiol® 4-88 (Sigma-Aldrich), and imaged using an AxioVision Zeiss UV microscope, an Olympus FV1000 confocal point-scanning microscope, or a ZEN Zeiss spinning disk confocal microscope.Table 1 Primary antibodies used in this study

s for 1 h at room temperature (Table 2). The samples were then washed, cover-slipped with mowiol® 4-88 (Sigma-Aldrich), and imaged using an AxioVision Zeiss UV microscope, an Olympus FV1000 confocal point-scanning microscope, or a ZEN Zeiss spinning disk confocal microscope.Table 1 Primary antibodies used in this study Antibody Concentration Product code Company Species Type Oct4 1:500 sc-5279 Santa Cruz Mouse IgG2b Sox2 1:200 ab97959 abcam Rabbit Polyclonal Human nestin 1:500 ABD69 Milipore Rabbit Polyclonal HuNu 1:200 MAB1281 Milipore Mouse IgG STEM121 1:1000 Y40410 Takara Mouse IgG1 Ki67 1:200 ab16667 Abcam Rabbit Monoclonal Ki67 1:500 ab15580 Abcam Rabbit Polyclonal TUJ1 1:1000 802,001 Biolegend Rabbit Polyclonal TUJ1 1:1000 T8660 Sigma Mouse IgG2b GFAP 1:600 Z0334 Dako Rabbit Polyclonal Olig2 1:200 AB9610 Milipore Rabbit Polyclonal NeuN 1:500 ABN78 Milipore Rabbit Polyclonal Synaptophysin 1:400 AB32127 Abcam Rabbit Polyclonal DAPI 1:10000 D1306 Invitrogen N/A N/A Hoechst 33342 1:1000 H3750 ThermoFisher N/A N/A Table 2 Secondary antibodies used in this study Antibody Concentration Wavelength Goat anti-mouse 1:400 488 Goat anti-rabbit 1:400 488 Goat anti-rabbit 1:400 568 Goat anti-mouse 1:400 555 Goat anti-mouse 1:400 568

-stroke. Similar to what we observed in the foot fault task, only mice that received drNPCs recovered back to baseline levels, whereas mice that received vehicle-only injections did not recover (Supplementary Fig. 2b). Taken together, these data support the conclusion that drNPC transplants promote functional recovery. Long-Term Surviving drNPCs Are Found in a Subpopulation of Stroke-Injured Mice To assess drNPC survival post-transplant, we sacrificed mice at 8 and 32 days post-stroke and stained for human cell markers HuNu and/or STEM121. At 8 days post-stroke (i.e., 4 days post-transplant), all mice (100%) had drNPCs at the site of injection. By 32 days post-stroke (i.e., 28 days post-transplant), the time when functional recovery was observed, drNPCs were only observed in 71% (HAMC, 8/14; aCSF, 9/10) of transplanted brains. A Fisher’s exact test and chi-square test revealed no significant association between vehicle and cell survival (p > 0.05). Interestingly, in all brains that had drNPCs present, irrespective of the time of sacrifice, transplanted cells were confined within the lesion boundary demarcated by NeuN expression, and did not penetrate deep into the uninjured tissue (Fig. 4d).Fig. 4 Transplanted drNPCs can survive and proliferate in the stroke-injured cortex. a HuNU+ (red) drNPCs are seen within the stroke-injured cortex of SCID/Beige mice at 8 and 32 days post-stroke. A subpopulation of HuNu+ cells are Ki67+ (green) at both survival times. b The number of HuNu+ drNPCs found within the stroke-injured cortex at 8 or 32 days post-stroke was not significantly different between cells transplanted in aCSF or HAMC (8 days, p = 0.34; 32 days, p = 0.99). Significantly fewer HuNu+ cells were observed between 8 and 32 days post-stroke when transplanted in HAMC (p = 0.019) but not aCSF (p = 0.34). c There was no significant difference between transplant vehicles in the percentage of Ki67+/HuNu+ drNPCs at 8 or 32 days post-stroke (day 8, p = 0.49; day 32, p = 0.79). There was also no significant difference in Ki67+/HuNu+ drNPCs between 8 and 32 days within each vehicle group (aCSF, p = 0.66; HAMC, p = 0.46). Percent of ki67+/HuNu+ drNPCs at Day 8, 8.5% ± 1.5 (aCSF) and 11.0% ± 2.0 (HAMC); and Day 32, 6.8% ± 1.3 (aCSF) and 8.2% ± 1.0 (HAMC). d IHC reveals that drNPCs (HuNu+) remain at the boundary of the stroke injury (demarcated by NeuN+ cells) following transplantation into the stroke-injured cortex 4 days post-stroke.

Antibody Concentration Product code Company Species Type Oct4 1:500 sc-5279 Santa Cruz Mouse IgG2b Sox2 1:200 ab97959 abcam Rabbit Polyclonal Human nestin 1:500 ABD69 Milipore Rabbit Polyclonal HuNu 1:200 MAB1281 Milipore Mouse IgG STEM121 1:1000 Y40410 Takara Mouse IgG1 Ki67 1:200 ab16667 Abcam Rabbit Monoclonal Ki67 1:500 ab15580 Abcam Rabbit Polyclonal TUJ1 1:1000 802,001 Biolegend Rabbit Polyclonal TUJ1 1:1000 T8660 Sigma Mouse IgG2b GFAP 1:600 Z0334 Dako Rabbit Polyclonal Olig2 1:200 AB9610 Milipore Rabbit Polyclonal NeuN 1:500 ABN78 Milipore Rabbit Polyclonal Synaptophysin 1:400 AB32127 Abcam Rabbit Polyclonal DAPI 1:10000 D1306 Invitrogen N/A N/A Hoechst 33342 1:1000 H3750 ThermoFisher N/A N/A Table 2 Secondary antibodies used in this study Antibody Concentration Wavelength Goat anti-mouse 1:400 488 Goat anti-rabbit 1:400 488 Goat anti-rabbit 1:400 568 Goat anti-mouse 1:400 555 Goat anti-mouse 1:400 568 Lesion Volume Analysis Following cresyl violet staining (see Supplementary methods), serial 20- μm thick coronal sections (200 μm apart) spanning a total of 3–4 mm surrounding the injury site were imaged at 5× magnification using an AxioCam ICc1 camera. The cortical lesion was measured on FIJI using the lasso and polygon tools to outline and quantify the total cortical lesion infarct area, as defined by the area with atypical tissue morphology including pale areas with lost Nissl staining and areas filled with dark pyknotic stained debris [32]. The total volume of the injury was estimated by averaging the area measured in each coronal section and multiplying by the total length of the scar, which was calculated from the number of sections in which the lesion was present.

e areas with lost Nissl staining and areas filled with dark pyknotic stained debris [32]. The total volume of the injury was estimated by averaging the area measured in each coronal section and multiplying by the total length of the scar, which was calculated from the number of sections in which the lesion was present. Gliosis Measurement and Analysis A set of serial coronal sections (20 μm thick) immunostained for GFAP+ expression were visualized at 5× magnification at 200 μm intervals using FIJI [31]. The total area of cortical GFAP+ expression was measured in each section. Measurements were taken from anterior to posterior through the scar and the maximal GFAP expression, as well as the total gliosis volume, was calculated per brain. Synaptophysin Imaging and Quantification All of the images were taken with identical parameters using confocal microscopy on a ZEN Zeiss spinning disk confocal microscope to generate z-stacks comprised of eight optical sections at 0.49 μm per section. The channel exposure was fixed at 1000 ms throughout the imaging of the entire set. Quantification of total synaptophysin-positive pixels per analyzed brain section was conducted by using FIJI [31] to measure the number of positive pixels in the perilesional areas. The mean pixel intensity in two perilesional regions of interest (ROIs) was used to measure the amount of staining (Fig. 7a), as this measure represents the sum of all detected bright pixels (gray values) divided by the total number of pixels within the channel. Imaging, ROI selection, and analysis were conducted by a blinded observer.

tensity in two perilesional regions of interest (ROIs) was used to measure the amount of staining (Fig. 7a), as this measure represents the sum of all detected bright pixels (gray values) divided by the total number of pixels within the channel. Imaging, ROI selection, and analysis were conducted by a blinded observer. Cellular Characterization and Quantification Three cell culture wells per biological replicate were stained for each specific antibody and were counted within the field of view in five areas within each well at 20× magnification. The percentage of each cell type was calculated as a percent of all DAPI or Hoechst labeled cells. For in vivo analysis, coronal sections 20 μm thick at 200 μm intervals were immunostained for HuNu or STEM121 and antibodies found in Table 1. Total numbers of surviving transplanted cells were calculated by extrapolating the average number of surviving drNPCs per section over the total number of sections that contained drNPCs (ranging from 15 to 20 sections). To analyze proliferation, the numbers of Ki67+/HuNu+ cells were counted in the same representative sections and calculated as a percent of all HuNu+ cells. Cell differentiation in vivo post-transplantation was analyzed by immunohistochemistry in brains that had surviving drNPCs.

drNPCs (ranging from 15 to 20 sections). To analyze proliferation, the numbers of Ki67+/HuNu+ cells were counted in the same representative sections and calculated as a percent of all HuNu+ cells. Cell differentiation in vivo post-transplantation was analyzed by immunohistochemistry in brains that had surviving drNPCs. Reverse Transcription Polymerase Chain Reaction Cultured drNPCs were collected into Buffer RL (Norgen Biotek) with β-mercapthenol and then processed according to the manufacturer’s directions using Total RNA Purification Kit (Norgen Biotek — Cat#17200). Cycling conditions consisted of polymerase activation and DNA denaturation (3 min at 98 °C), followed by 35 cycles of 10 s at 95 °C and 30 s at 60 °C. Primer sequences used are listed in Table 3.Table 3 Polymerase chain reaction primer sequences used in this study Target Sequence Expected product size Sox2 Fwd GGAGCTTTGCAGGAAGTTTG Rev. GGAAAGTTGGGATCGAACAA 460 Oct4 Fwd CTGAGGGTGAAGCAGGAGTC Rev. CTTGGCAAATTGCTCGAGTT 170 Nanog Fwd AAGGCCTCAGCACCTACCTA Rev. GAGACGGCAGCCAAGGTTAT 979 Nestin Fwd GCGTTGGAACAGAGGTTGGA Rev. TGGGAGCAAAGATCCAAGAC 327 Pax6 Fwd CAATCAAAACGTGTCCAACG Rev. TGGTATTCTCTCCCCCTCCT 431 Ascl1 Fwd GTCGAGTACATCCGCGCGCTG Rev. AGAACCAGTTGGTGAAGTCGA 220 CD133 Fwd CAGTCTGACCAGCGTGAAAA Rev. GGCCATCCAAATCTGTCCTA 200 Map2 Fwd TCAGAGGCAATGACCTTACC Rev. GTGGTAGGCTCTTGGTCTTT 320 Actb Fwd TCACCCACACTGTGCCCATCTACGA Rev. CAGCGGAACCGCTCATTGCCAATGG 295 GAPDH Fwd CTCTGCTCCTCCTGTTCGAC Rev. GCGCCCAATACGACCAAATC 121

327 Pax6 Fwd CAATCAAAACGTGTCCAACG Rev. TGGTATTCTCTCCCCCTCCT 431 Ascl1 Fwd GTCGAGTACATCCGCGCGCTG Rev. AGAACCAGTTGGTGAAGTCGA 220 CD133 Fwd CAGTCTGACCAGCGTGAAAA Rev. GGCCATCCAAATCTGTCCTA 200 Map2 Fwd TCAGAGGCAATGACCTTACC Rev. GTGGTAGGCTCTTGGTCTTT 320 Actb Fwd TCACCCACACTGTGCCCATCTACGA Rev. CAGCGGAACCGCTCATTGCCAATGG 295 GAPDH Fwd CTCTGCTCCTCCTGTTCGAC Rev. GCGCCCAATACGACCAAATC 121 Quantitative Reverse Transcription Polymerase Chain Reaction (RT-qPCR) Samples were collected into Buffer RL (Norgen Biotek) and processed according to the manufacturer’s directions using Total RNA Purification Kit (Norgen Biotek — Cat#17200). cDNA synthesis was carried out with iScript gDNA Clear cDNA Synthesis Kit (Bio-Rad — Cat# 1725034). RT-qPCR reactions were prepared according to the manufacturer’s directions using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad — Cat# 172-5270). RT-qPCR was carried out on Bio-Rad CFX384 Touch Real-Time PCR System (Bio-Rad). Cycling conditions consisted of polymerase activation and DNA denaturation (3 min at 98 °C), followed by 40 cycles of 10 s at 95 °C and 30 s at 60 °C. All reactions were concluded by incubation at 65 °C and increasing the temperature (at 0.5 °C increments) to 95 °C for melting-curve analysis. Prior to performing relative expression analyses, standard curves were generated for targets (see below) via the serial dilutions of pooled cDNA. In accordance with MIQE (Minimum Information for Publication of Quantitative Real-Time PCR Experiments) guidelines, the amplification efficiencies (E) of reported runs were between 97% and 113% and R2 > 0.9 with minimum two technical replicates per reaction. The Bio-Rad SYBR Green Assays used were Nestin (qHsaCED0044457), Tuj1 (qHsaCED0005794), Olig2 (qHsaCED0007834), Gfap (qHsaCID0022307), BDNF (qHsaCED0047199), and Gapdh (qHsaCED0038674). Relative expression data were normalized to the reference gene Gapdh to control for variability in expression levels and were analyzed using the Livak and Schmittgen (i.e., 2−ΔΔCT) and Pfaffl methods. The relative expression of each target was assessed by unpaired two-tailed t test. A p value of less than 0.05 was considered significant.

data were normalized to the reference gene Gapdh to control for variability in expression levels and were analyzed using the Livak and Schmittgen (i.e., 2−ΔΔCT) and Pfaffl methods. The relative expression of each target was assessed by unpaired two-tailed t test. A p value of less than 0.05 was considered significant. Enzyme-Linked Immunosorbent Assay (ELISA) Brain-derived neurotrophic factor (BDNF) released into the conditioned medium of drNPC cultures that were differentiated towards a neural lineage was measured by antigen-capture ELISA at different time points and compared to the release of BDNF in the conditioned medium of mature neurons (cat #1520, ScienCell). Conditioned medium from each group was collected, centrifuged, and then stored at − 80 °C until assaying. BDNF concentrations were measured by ELISA kit (BDNF Emax Immunoassay System, Promega Corporation, USA), according to the manufacturer’s instructions. Briefly, 96-well ELISA immunoplates were coated with Anti-BDNF (CatNb#G700B) diluted 1/1000 in carbonate buffer (pH 9.7), and incubated at 4 °C overnight. The following day, all wells were washed with TBS-Tween 0.5% before incubation with Block/Sample buffer 1× at room temperature for 1 h without shaking. After blocking, standards and samples were added to the plates and incubated and shaken (450 ± 100 rpm) for 2 h at room temperature. Subsequently, after washing with TBS-Tween wash buffer, plates were incubated for 2 h with Anti-Human BDNF (1:500 dilution in Block & Sample 1× Buffer) at 4 °C. After incubation, plates were washed five times with TBS-Tween 0.5% wash buffer and 100 μl of diluted Anti-IgYHRP Conjugate was added to each well (1:200 dilution in Block & Sample 1X Buffer) and incubated for 1 h at room temperature with shaking (450 ± 100 rpm). Then, plates were washed five times with TBS-Tween 0.5% wash buffer and 100 μl of TMB One Solution was added to each well. Following 10 min incubation at room temperature with shaking (450 ± 100 rpm) for the BDNF plate, a blue color formed in the wells. After stopping the reaction by adding 100 μl of 1 N hydrochloric acid, the absorbance was read at 450 nm on a microplate reader (Synergy 4) within 30 min of stopping the reactions. Concentration of released BDNF in the supernatants was determined according to the standard curves. BDNF concentrations were compared using an unpaired, two-tailed t test for each time point.

hydrochloric acid, the absorbance was read at 450 nm on a microplate reader (Synergy 4) within 30 min of stopping the reactions. Concentration of released BDNF in the supernatants was determined according to the standard curves. BDNF concentrations were compared using an unpaired, two-tailed t test for each time point. Behavioral Tests Behavioral analysis was performed using the foot fault task, measuring gross motor functions such as coordination and balance, as well as fine sensorimotor function like reaching and stepping [33], 3 days prior to injury (baseline) and at 3, 8, 18, and 32 days post-stroke; and the cylinder test at 3 days prior to injury and at 3 and 32 days post-stroke. Detailed methods of behavioral tests can be found in the supplementary material. All behavioral tests were recorded with a digital camera (SX 60 HS, Canon) and viewed on VLC Media Player (Version 2.2.1, VideoLAN Organizarion). Videos were scored by a blinded observer. Statistical Analysis Statistical analysis was performed using Prism 6 (GraphPad Software, San Diego, CA) and IBM SPSS Statistics (International Business Machines Corp., Armonk, NY). Data was analyzed using a variety of statistical methods which can be found in the supplementary material. All data is reported as mean ± SEM.

istical Analysis Statistical analysis was performed using Prism 6 (GraphPad Software, San Diego, CA) and IBM SPSS Statistics (International Business Machines Corp., Armonk, NY). Data was analyzed using a variety of statistical methods which can be found in the supplementary material. All data is reported as mean ± SEM. Results drNPCs Are Neurally Committed at the Time of Transplant To confirm the identity of drNPCs, we characterized their expression of pluripotency and neural markers using immunostaining and polymerase chain reaction (PCR). We examined Oct4 expression as a measure of pluripotency; Sox2 and Nestin for NPCs; Ki67 to assess proliferation; GFAP for astrocytes; TUJ1 for neurons; and Olig2 for oligodendrocytes. We found that drNPCs do not express Oct4, and ubiquitously express NPC markers Sox2 and Nestin, indicating a precursor phenotype (Fig. 1a, Supplementary Fig. 1). In addition, we determined that 71.8 ± 4.0% of drNPCs were Ki67+ in vitro (Fig. 1a). We also confirmed that drNPCs have the ability to differentiate into the three neural subtypes and express GFAP, TUJ1, and Olig2 (Fig. 1b). PCR analysis confirmed the expression of neural lineage markers including Nestin, Sox2, Ascl1, Pax6, MAP2, and CD133 and the absence of pluripotency markers Nanog and Oct4 (Fig. 1c). RT-qPCR analysis comparing the expression levels of neural markers Nestin, Tuj1, Olig2, and Gfap in drNPCs prior to culturing (from frozen vials) and drNPCs that were cultured prior to transplantation (cultured drNPCs) revealed similar expression of Nestin and Gfap between the two cell populations and increased expression of Tuj1 and Olig2 in cultured drNPCs (Fig. 1d). Collectively, these results confirm that drNPCs are neurally committed, remain in a precursor state at time of transplant, and can give rise to all three neural cell types.Fig. 1 drNPCs remain in a neural precursor state in vitro at time of transplantation. a Immunocytochemistry reveals that drNPCs do not express the pluripotency marker Oct4, but do ubiquitously express neural precursor cell markers Sox2 and Nestin, while 71.8 ± 4.0% of cells express the proliferation marker Ki67. b Cultured drNPCs can also express differentiated neural cell markers for astrocytes (GFAP), neurons (TUJ1), and oligodendrocytes (Olig2). c Polymerase chain reaction confirms that drNPCs express neural markers Nestin, Sox2, Ascl1, Pax6, MAP2, CD311, and do not express pluripotency markers Nanog and Oct4.

marker Ki67. b Cultured drNPCs can also express differentiated neural cell markers for astrocytes (GFAP), neurons (TUJ1), and oligodendrocytes (Olig2). c Polymerase chain reaction confirms that drNPCs express neural markers Nestin, Sox2, Ascl1, Pax6, MAP2, CD311, and do not express pluripotency markers Nanog and Oct4. d Cultured drNPCs express equivalent levels of Nestin and Gfap, but increased Tuj1 and Olig2 mRNA compared to the frozen drNPCs. Data are shown as mean ± SEM. Gene expression levels are relative to frozen drNPCs and normalized to the reference gene Gapdh. n = 3/cohort. a, bn = 3, with three technical replicates per stain. Scale bar = 25 μm, blue = nuclear stain

of Nestin and Gfap, but increased Tuj1 and Olig2 mRNA compared to the frozen drNPCs. Data are shown as mean ± SEM. Gene expression levels are relative to frozen drNPCs and normalized to the reference gene Gapdh. n = 3/cohort. a, bn = 3, with three technical replicates per stain. Scale bar = 25 μm, blue = nuclear stain Cell Viability Is Not Dependent on the Transplant Vehicle With the goal of enhancing cell viability at the time of transplant and within the host, we tested two transplant vehicles; (1) artificial cerebrospinal fluid (aCSF), a buffer to mimic circulating CSF, and (2) a hyaluronan methylcellulose hydrogel (HAMC), which has been shown to improve xenograft survival in the CNS [17, 22]. To determine the effect of vehicle on the number of live cells at the time of transplantation, we performed a live/dead assay on cells prior to and following injection through the syringe. Cell viability was determined at 0 h and 2 h (aCSF) or 6 h (HAMC) after cell preparation (Fig. 2a), which was reflective of the longest elapsed time from preparation to transplanting in vivo. Cell viability was always > 88% of the original cell population (aCSF 0 h = 95.5 ± 0.2%, 2 h = 94.5 ± 1.5%; HAMC 0 h = 95.2 ± 0.8%, 6 h = 88.2 ± 0.3%). Cell viability following injection through the syringe, relative to the numbers of viable cells placed in the syringe, was > 95% in all conditions and was not significantly different between vehicles at any time point (Fig. 2b). Thus, all mice received a minimum of ~ 85,000 viable drNPCs at the time of transplantation.Fig. 2 drNPC transplant injection paradigm results in minimal cell death. a Live/dead assay images before and after injection at the first and last injection time points within each vehicle. All cells were maintained on ice within their respective vehicle until transplanted. b There is no significant difference in % survival between groups at any time point. Cell survival as a result of injection through the syringe was greater than 95% in all instances (aCSF 0 h = 99.60 ± 0.53%, 2 h = 98.52 ± 1.29%; HAMC 0 h = 96.00 ± 1.32%, 6 h = 98.49 ± 1.32%). Data presented as mean ± SEM; n = 3 per vehicle within each time point, with three technical replicates each; green cells = live cells, red cells (white arrowheads) = dead cells; scale bar = 100 μm

than 95% in all instances (aCSF 0 h = 99.60 ± 0.53%, 2 h = 98.52 ± 1.29%; HAMC 0 h = 96.00 ± 1.32%, 6 h = 98.49 ± 1.32%). Data presented as mean ± SEM; n = 3 per vehicle within each time point, with three technical replicates each; green cells = live cells, red cells (white arrowheads) = dead cells; scale bar = 100 μm drNPC Transplantation Promotes Functional Recovery To determine the efficacy of drNPCs for stroke recovery, we used a clinically relevant model of ET-1 stroke in the sensorimotor cortex. This resulted in consistent tissue damage, gliosis, and functional impairment in the foot fault task as early as 4 days post-stroke, which was maintained up to 32 days post-stroke, the longest time point examined (Supplementary Fig. 2). Cultured drNPCs were injected directly into the stroke lesion at 4 days post-stroke based on previous work [17]. Mice were tested in the foot fault task at 3 days prior to stroke to establish a baseline measure, and at 3, 8, 18, and 32 days post-stroke (Fig. 3a). Only stroke-injured mice that had significant motor impairments on the foot fault task by 8 days post-stroke (deficits present at day 3 or day 8) were included in our analysis (aCSF = 11/12 mice; HAMC = 11/13 mice; drNPCs+aCSF = 10/13 mice; drNPCs+HAMC = 14/15 mice). Animals not showing any deficits on either day 3 or 8 post-stroke were excluded from our analysis and completely removed from the study. These exclusion criteria were used to prevent animals that did not actually have a deficit following stroke from skewing the outcomes of the study to falsely showing improved performance.Fig. 3 drNPC transplantation promotes functional recovery. a Mice were tested for functional performance on the foot fault test at 3 days prior to stroke and 3, 8, 18, and 32 days post-stroke, and sacrificed 32 days post-stroke, when tissue analysis was performed. b–d All stroke-injured mice have significant functional deficits by 3 days following stroke. b drNPC transplants promote functional recovery back to uninjured control levels by 32 days post-stroke, whereas vehicle-only injections did not. c The transplant vehicle had no impact on functional recovery, as only those mice that received drNPCs recovered to naïve levels and those that received vehicle injections did not, irrespective of transplant vehicle (HAMC or aCSF). d Mice that received drNPCs had significantly better performance on the foot fault test than those that received vehicle-only injections at 32 days post-stroke.

hose mice that received drNPCs recovered to naïve levels and those that received vehicle injections did not, irrespective of transplant vehicle (HAMC or aCSF). d Mice that received drNPCs had significantly better performance on the foot fault test than those that received vehicle-only injections at 32 days post-stroke. Data is presented as mean ± SEM; * = b significantly different from naïve d significant difference between groups; c a (aCSF alone), b (HAMC alone), c (drNPCs + aCSF), d (drNPCs + HAMC) = significantly different from naïve; n.s. = not significant, b) p < 0.0001, c) p < 0.003, and d) p < 0.05

hose mice that received drNPCs recovered to naïve levels and those that received vehicle injections did not, irrespective of transplant vehicle (HAMC or aCSF). d Mice that received drNPCs had significantly better performance on the foot fault test than those that received vehicle-only injections at 32 days post-stroke. Data is presented as mean ± SEM; * = b significantly different from naïve d significant difference between groups; c a (aCSF alone), b (HAMC alone), c (drNPCs + aCSF), d (drNPCs + HAMC) = significantly different from naïve; n.s. = not significant, b) p < 0.0001, c) p < 0.003, and d) p < 0.05 Stroke-injured mice that received vehicle-only (aCSF or HAMC) injections displayed functional impairments at all times examined, relative to their own baseline performance as well as compared to naïve (uninjured) controls (Fig. 3b). Interestingly, mice that received drNPC transplants in either HAMC or aCSF recovered to their baseline performance by 32 days post-stroke and were not significantly different from naïve performance (Fig. 3b). Our analysis revealed that the observed functional recovery was due to drNPC transplants, irrespective of vehicle (Fig. 3c). We investigated this relationship further by comparing the performance of mice that received drNPCs vs vehicle-alone injections and found that mice that received drNPCs performed significantly better than those that received vehicle-only injections at day 32 post-stroke (Fig. 3d). Functional recovery following drNPC transplantation was seen in both male and female mice (Supplementary Fig. 3a), indicating that drNPC transplants promote recovery regardless of recipient sex.

eived drNPCs performed significantly better than those that received vehicle-only injections at day 32 post-stroke (Fig. 3d). Functional recovery following drNPC transplantation was seen in both male and female mice (Supplementary Fig. 3a), indicating that drNPC transplants promote recovery regardless of recipient sex. To further assess functional outcomes, we performed the cylinder test in a subset of mice that demonstrated impairments in the foot fault task post-stroke. Similar to what we observed in the foot fault task, only mice that received drNPCs recovered back to baseline levels, whereas mice that received vehicle-only injections did not recover (Supplementary Fig. 2b). Taken together, these data support the conclusion that drNPC transplants promote functional recovery.

localization between markers. a, d Scale bars = 100 μm (b), n = 4 for both vehicles at day 8 and n = 7 for both vehicles at day 32. cn = 4 for both vehicles at day 8, and n = 6 for aCSF and n = 3 for HAMC at day 32 (e) Scale bar = 50 μm, at least three brains were analyzed per marker and representative images were used The total number of viable HuNu+ drNPCs within the transplanted brains 8 days post-stroke was 12,780 ± 2963 when delivered in aCSF and 21,633 ± 9880 when delivered with HAMC (Fig. 4a, b). A decline in the numbers of surviving HuNu+ cells was observed by day 28 post-transplantation in both vehicles; by 32 days post-stroke, the total numbers of HuNu+ cells were 4961 ± 1266 for cells transplanted in aCSF and 5130 ± 1815 for cells transplanted in HAMC (Fig. 4b). However, this decrease was only significant for brains that received cells in HAMC. No significant difference in drNPC survival was observed between vehicles at 8 or 32 days following stroke (Fig. 4b). Proliferation of drNPCs at 8 and 32 days post-stroke was measured by counting the number of Ki67+/HuNu+ cells as a percent of all HuNu+ cells in brains (Fig. 4a). We found no significant difference between the two times examined in either vehicle and no significant effect of transplant vehicle (Fig. 4c).

The total number of viable HuNu+ drNPCs within the transplanted brains 8 days post-stroke was 12,780 ± 2963 when delivered in aCSF and 21,633 ± 9880 when delivered with HAMC (Fig. 4a, b). A decline in the numbers of surviving HuNu+ cells was observed by day 28 post-transplantation in both vehicles; by 32 days post-stroke, the total numbers of HuNu+ cells were 4961 ± 1266 for cells transplanted in aCSF and 5130 ± 1815 for cells transplanted in HAMC (Fig. 4b). However, this decrease was only significant for brains that received cells in HAMC. No significant difference in drNPC survival was observed between vehicles at 8 or 32 days following stroke (Fig. 4b). Proliferation of drNPCs at 8 and 32 days post-stroke was measured by counting the number of Ki67+/HuNu+ cells as a percent of all HuNu+ cells in brains (Fig. 4a). We found no significant difference between the two times examined in either vehicle and no significant effect of transplant vehicle (Fig. 4c). We further characterized the in vivo profile of surviving drNPCs (HuNu+ or STEM121+ cells) at 32 days post-stroke using immunostaining for Sox2 and Nestin (undifferentiated NPCs), GFAP (astrocytes), TUJ1 (immature neurons), NeuN (mature neurons), and Olig2 (oligodendrocytes). Irrespective of the transplant vehicle, the vast majority of surviving drNPCs remain undifferentiated (Sox2+, Nestin+). A subpopulation of drNPCs expressed GFAP and a small minority expressed the immature neuronal marker TUJ1, but no mature neurons (NeuN+) or oligodendrocytes (Olig2+) were observed at 28 days post-transplant (Fig. 4e).

he transplant vehicle, the vast majority of surviving drNPCs remain undifferentiated (Sox2+, Nestin+). A subpopulation of drNPCs expressed GFAP and a small minority expressed the immature neuronal marker TUJ1, but no mature neurons (NeuN+) or oligodendrocytes (Olig2+) were observed at 28 days post-transplant (Fig. 4e). drNPC Transplant Survival Is Not Necessary for Recovery Approximately 30% of brains that received drNPC transplants did not have drNPCs present at 28 days post-transplant when functional recovery was observed. We asked whether cell survival was necessary for maintaining functional recovery by comparing behavioral performance in mice with (n = 17) and without (n = 7) drNPCs present at day 32 post-stroke. Our findings reveal that functional recovery occurs irrespective of the presence of cells at day 32 post-stroke (Fig. 5a). Interestingly, mice without cells at 32 days post-stroke had already recovered by 18 days post-stroke. There was no significant difference in performance between the two groups at any time-point tested. Moreover, Pearson’s and Point-Biseral analyses revealed a significant correlation between functional performance at day 32 (compared to baseline) and the injection of drNPCs into the cortex (r = 0.37, n = 46, p = 0.011), but no correlation between whether the drNPCs were present on day 32 post-stroke (r = − 0.17, n = 24, p = 0.43) or with the absolute number of surviving drNPCs (r = 0.21, n = 21, p = 0.36) (Fig. 5b). These findings reveal that long-term transplanted cell survival is not necessary for maintaining functional recovery.Fig. 5 Xenograft survival is not necessary for functional recovery. a Performance in the foot fault task revealed no significant difference in functional recovery between mice that had drNPCs present at day 32 versus mice with no drNPCs present. b There is no significant correlation (r = 0.21, n = 21, p = 0.36) between the absolute drNPC survival numbers in vivo and functional recovery by 32 days post-stroke. a* = brains with surviving drNPCs different from baseline, # = brains with no drNPC survival different from baseline; p < 0.05

ce with no drNPCs present. b There is no significant correlation (r = 0.21, n = 21, p = 0.36) between the absolute drNPC survival numbers in vivo and functional recovery by 32 days post-stroke. a* = brains with surviving drNPCs different from baseline, # = brains with no drNPC survival different from baseline; p < 0.05 Lesion Volume and Glial Scarring Are Unaffected by Transplant and Not Correlated to Functional Recovery We asked whether drNPC transplants affected the extent of gliosis and the size of the lesion following ET-1 stroke, and whether these outcomes were related to the observed motor recovery. The extent of gliosis was determined using GFAP staining in vehicle and drNPC transplanted mice at 32 days post-stroke (Fig. 6a) by measuring the maximal cortical GFAP+ area, which was strongly correlated (r = 0.89, n = 35, p < 0.001) to total gliosis volume per brain (Supplementary Fig. 4). A comparison between the gliotic response in mice that received drNPCs versus vehicle-only injections (n ≥ 16 per group) revealed no significant difference between groups (Fig. 6b). There was also no significant difference between vehicle type in drNPC or vehicle-only treated groups (Supplementary Fig. 5a). Furthermore, a Pearson’s correlation revealed no significant correlation (r = − 0.102, n = 36, p = 0.554) between the size of the glial scar and functional performance at 32 days post-stroke (Fig. 6c).Fig. 6 Glial scarring and lesion volume are unaffected by treatment and do not correlate with functional recovery. a Gliosis was present in brains from all groups at 32 days post-stroke. b There was no significant difference in terms of maximal GFAP+ area between drNPC transplanted brains and vehicle-only brains (n = 20, drNPC transplant; n = 16, vehicle-only) c There is no significant correlation (r = − 0.10, n = 36, p = 0.55) between the extent of gliosis and the functional performance at 32 days post-stroke compared to baseline performance. d Cresyl violet staining reveals stroke lesions in all groups at 32 days post-stroke. e No significant difference in lesion volume is seen between drNPC transplant and vehicle-only treated brains (n = 8 per group).f There is no significant correlation (r = − 0.38, n = 16, p = 0.15) between the lesion volume and the functional performance at 32 days post-stroke compared to baseline performance. Dashed line = a GFAP+ expression boundary or d cresyl violet lesion boundary, dashed line within graphs.

treated brains (n = 8 per group).f There is no significant correlation (r = − 0.38, n = 16, p = 0.15) between the lesion volume and the functional performance at 32 days post-stroke compared to baseline performance. Dashed line = a GFAP+ expression boundary or d cresyl violet lesion boundary, dashed line within graphs. c, f = 95% confidence bands, Cx = cortex, CC = corpus callosum, n.s. = not significant, r = Pearson’s coefficient, significance based on p < 0.05. Data are represented as mean ± SEM. Scale bars = 500 μm A similar observation was made comparing the ischemic lesion volume between vehicle-only and drNPC-transplanted mice. Cresyl violet staining revealed no difference in the size of the lesion between mice that received vehicle only versus drNPCs (Fig. 6d, e) or between the vehicle type within the groups (Supplementary Fig. 5b). A Pearson’s correlation also revealed no significant correlation (r = − 0.38, n = 16, p = 0.15) between behavioral recovery and lesion volumes (Fig. 6f). Hence, drNPC transplants do not affect the extent of gliosis or size of the lesion volume post-ET-1 stroke and these tissue outcomes are not correlated with functional improvement.

rrelation also revealed no significant correlation (r = − 0.38, n = 16, p = 0.15) between behavioral recovery and lesion volumes (Fig. 6f). Hence, drNPC transplants do not affect the extent of gliosis or size of the lesion volume post-ET-1 stroke and these tissue outcomes are not correlated with functional improvement. drNPC Transplants Increase Perilesional Expression of Synaptophysin To investigate the possibility that drNPC transplants influenced recovery by promoting synaptic plasticity, we examined the expression of synaptophysin, a presynaptic vesicle protein, in treated stroke-injured brains. Using immunohistochemistry and confocal imaging (Fig. 7a), we compared the mean pixel intensity (MPI) of synaptophysin expression in the perilesional tissue of stroke-injured mice that received drNPCs (n = 6) and those that received vehicle-only injections (n = 9). Mice that received drNPC transplants had a significant increase (p = 0.012) in synaptophysin expression compared to vehicle-treated mice (Fig. 7b). A Pearson’s correlation also revealed that decreased functional impairment is strongly correlated (r = − 0.80, n = 15, p = 0.0003) with increased synaptophysin MPI (Fig. 7c). This supports the hypothesis that drNPC transplants lead to increased synaptic plasticity which may underlie the observed functional recovery.Fig. 7 drNPC transplants result in increased synaptophysin expression in the perilesional area. a Two ROIs within the perilesional tissue (medial and lateral) were selected in one coronal section per brain analyzed. The ROIs were imaged through eight optical planes and the settings were all kept identical for each section. ai–ii: higher magnification images of perilesional areas. b Mice that received drNPCs had significantly greater MPI for Alexa Fluor 488 Staining (Synaptophysin) within the ROIs than mice that received vehicle alone injections. c Pearson correlation analysis reveals that synaptophysin MPI is strongly correlated (r = − 0.80, n = 15, p = 0.0003) with improved functional outcomes. d RT-qPCR analysis of drNPCs. BDNF expression levels are relative to frozen drNPCs and normalized to the reference gene Gapdh. Both frozen and cultured drNPCs express BDNF. Data are shown as mean ± SEM. n = 3/cohort. e Differentiated drNPCs release BDNF at levels comparable to that of mature neurons in vitro, as determined by quantification of BDNF (ng/mL) release using antigen-capture.

frozen drNPCs and normalized to the reference gene Gapdh. Both frozen and cultured drNPCs express BDNF. Data are shown as mean ± SEM. n = 3/cohort. e Differentiated drNPCs release BDNF at levels comparable to that of mature neurons in vitro, as determined by quantification of BDNF (ng/mL) release using antigen-capture. The populations of cells tested were mature neurons (positive control) and drNPCs in differentiation conditions. Data are shown as mean ± SEM. n = 3 independent samples per timepoint. Dashed lines = lesion boundary, white boxes = ROIs ROIm = Medial ROI, ROIL = Lateral ROI, a scale bar = 200 μm, ai–ii scale bar = 50 μm, bn = 6 for drNPC group and n = 9 for vehicle group, * = significant difference; p = 0.012 To elucidate a potential mechanism by which drNPC transplants exert their beneficial effects, we asked whether drNPCs express and secrete brain-derived neurotrophic factor (BDNF), which is known to promote neuroplasticity [34]. RT-qPCR analysis revealed that BDNF is expressed in both frozen and cultured drNPCs (Fig. 7d). An ELISA analysis of culture medium derived from drNPC cultures that were differentiated towards a neural lineage over time reveals that BDNF is released at levels comparable to mature neurons (Fig. 7e). Hence, BDNF-mediated plasticity may play a role in the functional recovery observed following drNPC transplantation.

An ELISA analysis of culture medium derived from drNPC cultures that were differentiated towards a neural lineage over time reveals that BDNF is released at levels comparable to mature neurons (Fig. 7e). Hence, BDNF-mediated plasticity may play a role in the functional recovery observed following drNPC transplantation. Discussion We have shown that drNPC transplantation during the subacute phase in a pre-clinical mouse model of stroke is able to promote functional recovery, regardless of the transplant vehicle or the sex of the recipient. Furthermore, we found that functional recovery does not require the long-term survival of transplanted cells, and that recovery is maintained beyond transplant survival. In the brains of mice that did have surviving drNPCs at late survival times, the majority of the transplanted drNPCs remained undifferentiated and non-proliferative. Most interesting, brains that received drNPC transplants had higher levels of synaptophysin in the perilesional stroke-injured cortex, supporting the idea that synaptogenesis may underlie the drNPC-mediated recovery.

urvival times, the majority of the transplanted drNPCs remained undifferentiated and non-proliferative. Most interesting, brains that received drNPC transplants had higher levels of synaptophysin in the perilesional stroke-injured cortex, supporting the idea that synaptogenesis may underlie the drNPC-mediated recovery. Cell survival is a challenge common to transplant therapies in general. Herein, we used two transplant vehicles with the goal of establishing the best parameters to enhance cell survival and promote recovery. Interestingly, the frequency and absolute number of viable drNPCs observed in vivo were not different between HAMC and aCSF. Previous studies report that HAMC has pro-survival properties and improves cell transplant survival outcomes using murine cells [17, 35–37], which has been attributed in part to the immunomodulatory effects of HAMC [38–41]. Accordingly, the lack of pro-survival effects of HAMC in this study may be due to the immunomodulatory advantage of HAMC being negated in the immunodeficient mouse strain (lacking adaptive immune cells). Of note, drNPC proliferation was also not affected by the vehicle. In both vehicles, the proliferative ability of drNPCs decreased following injection into the stroke-injured brain; dropping from 71.8 ± 4.0% at time of transplantation to approximately 10% by 4 days post-transplant. Importantly, we also found no evidence of tumor formation in any of the animals, similar to previous work with drNPCs [42].

he proliferative ability of drNPCs decreased following injection into the stroke-injured brain; dropping from 71.8 ± 4.0% at time of transplantation to approximately 10% by 4 days post-transplant. Importantly, we also found no evidence of tumor formation in any of the animals, similar to previous work with drNPCs [42]. Our results indicate that the long-term survival of transplanted cells is not necessary for maintaining functional recovery, although their presence at early times is important, as vehicle-only treated mice did not recover. We found no correlation between functional recovery and the extent of gliosis or lesion volumes, consistent with observations in other models of stroke where interventions lead to recovery but had no effect on tissue outcomes [43–45]. The mechanism by which transplanted cells mediate recovery is still unknown but there is evidence that suggests transplanted cells can promote recovery through trophic support, by promoting plasticity and synaptogenesis, inducing angiogenesis, immunomodulation, reducing excitotoxicity, and even activating endogenous cells to proliferate and migrate to the site of the lesion [15, 16, 18, 20, 43, 45–55]. Notably, the short-term survival of the transplanted cells is consistent with the hypothesis that the presence of drNPCs promotes recovery through an indirect mechanism.

ation, reducing excitotoxicity, and even activating endogenous cells to proliferate and migrate to the site of the lesion [15, 16, 18, 20, 43, 45–55]. Notably, the short-term survival of the transplanted cells is consistent with the hypothesis that the presence of drNPCs promotes recovery through an indirect mechanism. Our observation that drNPC transplants lead to functional recovery and increased synaptophysin expression in the perilesional stroked hemisphere suggests that one underlying mechanism for drNPC-mediated recovery for stroke is enhancing host brain plasticity; through increased synaptogenesis via the development of new synaptic junctions, potentially resulting from axonal sprouting and endogenous cortical remapping [43, 56]. Exploring the secretome of transplanted drNPCs may provide further insight into the mechanisms and pathways that result in functional recovery. Supplementary to our findings, recent studies transplanting drNPCs that were pre-differentiated towards an oligodendrogenic fate prior to transplantation in a rat model of spinal cord injury resulted in improved functional outcomes via migration and integration within the injured tissue, where they participated in tissue sparing and axonal remyelination [42]. Thus, it is possible that the observed mechanism of recovery depends on a variety of factors, such as injury, host, and status of drNPCs, which is an important consideration for drNPCs as autologous transplants since they could have additional mechanisms of action related to cell replacement in humans.

yelination [42]. Thus, it is possible that the observed mechanism of recovery depends on a variety of factors, such as injury, host, and status of drNPCs, which is an important consideration for drNPCs as autologous transplants since they could have additional mechanisms of action related to cell replacement in humans. drNPCs have the potential to provide a safe, autologous, and plentiful source of cells for clinical neural repair strategies and our findings support the conclusion that drNPCs are a promising candidate to treat stroke. The potential of translating the results from our study to the clinic raises important questions with regard to the optimal timing of transplantation and the associated mechanism that induces recovery. Transplantation of drNPCs in a different model that produces a larger lesion or a chronic model of stroke, in addition to selective ablation of transplanted cells at various times post-stroke, may provide additional insight into the optimal therapeutic window for transplantation and further our understanding of the underlying cell-based mechanisms that promote recovery. Further understanding of these mechanisms will support the development of novel therapeutics for neural repair. Electronic Supplementary Material ESM 1 (DOCX 4560 kb) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ilan Vonderwalde and Ashkan Azimi contributed equally to this work.

drNPCs have the potential to provide a safe, autologous, and plentiful source of cells for clinical neural repair strategies and our findings support the conclusion that drNPCs are a promising candidate to treat stroke. The potential of translating the results from our study to the clinic raises important questions with regard to the optimal timing of transplantation and the associated mechanism that induces recovery. Transplantation of drNPCs in a different model that produces a larger lesion or a chronic model of stroke, in addition to selective ablation of transplanted cells at various times post-stroke, may provide additional insight into the optimal therapeutic window for transplantation and further our understanding of the underlying cell-based mechanisms that promote recovery. Further understanding of these mechanisms will support the development of novel therapeutics for neural repair. Electronic Supplementary Material ESM 1 (DOCX 4560 kb) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ilan Vonderwalde and Ashkan Azimi contributed equally to this work. We would like to thank Nadia Sachewsky for helping with drNPC culture work, Monoleena Khan and Ritika Kompella for helping with tissue processing and preparation, Ricky Siu for assisting with animal surgery, and Priya Anandakumaran, Ana Fokina, and Tobias Fuehrmann for helping with HAMC preparation and drNPC suspension. Thank you to Emily Gilbert and Jessica Livingston for their discussions and assistance.

n and Ritika Kompella for helping with tissue processing and preparation, Ricky Siu for assisting with animal surgery, and Priya Anandakumaran, Ana Fokina, and Tobias Fuehrmann for helping with HAMC preparation and drNPC suspension. Thank you to Emily Gilbert and Jessica Livingston for their discussions and assistance. Authors’ Contributions IV: designed and performed experiments, contributed to study design, collected and analyzed data, prepared figures and wrote manuscript AA: designed and performed experiments, contributed to study design, collected and analyzed data, contributed to manuscript writing GR: collected and analyzed data JEA: contributed to study design and provided financial support MS: contributed to study design CM: designed and supervised studies, analyzed data, prepared final manuscript, provided financial support for all studies All authors approved the final manuscript. Funding This work was funded by a collaborative industry–sponsored research grant between the Canadian Institutes of Health Research (CIHR) and New World Laboratories Inc., and the Canada First Research Excellent Fund (CFREF, Medicine by Design). The directly reprogrammed human neural precursor cells are being commercialized by Fortuna Fix Ltd., who provided the cells for the current study. Compliance with Ethical Standards Conflict of Interest JEA is a shareholder of New World Laboratories Inc. and Fortuna Fix Ltd. All other authors declare that they have no conflict of interest.

Funding This work was funded by a collaborative industry–sponsored research grant between the Canadian Institutes of Health Research (CIHR) and New World Laboratories Inc., and the Canada First Research Excellent Fund (CFREF, Medicine by Design). The directly reprogrammed human neural precursor cells are being commercialized by Fortuna Fix Ltd., who provided the cells for the current study. Compliance with Ethical Standards Conflict of Interest JEA is a shareholder of New World Laboratories Inc. and Fortuna Fix Ltd. All other authors declare that they have no conflict of interest. Ethical Approval All experimental protocols were in accordance with the policies established in the Guide to the Care and Use of Experimental Animals prepared by the Canadian Council of Animal Care and approved by the animal care committee at the University of Toronto.

Introduction Glutamate, which is the principal excitatory amino acid, mediates the physiological excitatory synaptic transmission in the brain [1]. Excessive glutamate released from presynaptic neurons induces excitotoxicity, which contributes to the neuropathology observed in acute traumatic and ischemic brain injury, chronic neurodegenerative diseases, and other neurological diseases, such as seizures [2–5]. Rapid glutamate clearance in the synaptic cleft is required for maintaining synaptic homeostasis and preventing excitotoxicity [6–8]. The presynaptic membrane, postsynaptic membrane, and neighboring astrocytes comprise a physical “tripartite synapse” that limits the diffusion of neurotransmitters [9, 10]. Glutamate is predominantly taken up by sodium-dependent transporters known as excitatory amino acid transporters (EAATs), which are mainly localized in tripartite synaptic astrocytes [11, 12]. The glutamate taken up by EAATs is cotransported with Na+ into astrocytes [13]. Therefore, the transmembrane Na+ gradient in the synaptic space, produced by the export of Na+ by Na+/K+-ATPase, is the essential driving force for glutamate uptake [14, 15].

are mainly localized in tripartite synaptic astrocytes [11, 12]. The glutamate taken up by EAATs is cotransported with Na+ into astrocytes [13]. Therefore, the transmembrane Na+ gradient in the synaptic space, produced by the export of Na+ by Na+/K+-ATPase, is the essential driving force for glutamate uptake [14, 15]. N-myc downstream regulated gene 2 (NDRG2), a known tumor suppressor protein [16], is primarily expressed in astrocytes rather than in neurons or other glial cells in various brain areas [17–19]. NDRG1, NDRG2, NDRG3, and NDRG4 constitute the NDRG family and are required in cell proliferation and differentiation [20, 21]. NDRG2 was reported as an early-stage stress response gene, and its expression is upregulated under excitotoxic conditions in the brain, including ischemia [22], hemorrhage [23], trauma [24], and Alzheimer’s disease. To evaluate the potential physiological or pathological roles of NDRG2 in the brain, we generated NDRG2 knockout (Ndrg2−/−) mice [25]. These mice are susceptible to cerebral ischemia and exhibit increased interstitial glutamate levels in the brain, suggesting that NDRG2 plays an essential role in controlling glutamate excitotoxicity. Here, we examined this hypothesis and provided evidence that strongly suggests that NDRG2 is required for sodium-dependent glutamate uptake into astrocytes. The NDRG2-mediated astroglial glutamate uptake from the cerebral interstitial fluid is essential for protecting the brain from glutamate excitotoxicity following ischemia.

examined this hypothesis and provided evidence that strongly suggests that NDRG2 is required for sodium-dependent glutamate uptake into astrocytes. The NDRG2-mediated astroglial glutamate uptake from the cerebral interstitial fluid is essential for protecting the brain from glutamate excitotoxicity following ischemia. Materials and Methods Mice The experimental protocols were reviewed and approved by the Ethics Committee of the Fourth Military Medical University. We made all efforts to minimize the number of mice used and their suffering. The animals were group-housed under a regular 12-h light/dark cycle with access to food and water ad libitum. The Ndrg2flox/flox mice were crossed with B6.C-Tg(CMV-cre)1Cgn/J mice (Jackson Labs, USA) to generate the Ndrg2−/− mice. The line was backcrossed to C57BL/6 J more than 20 times. Young adult male (8–10 weeks old) mice were used in this study. All animal procedures were performed in accordance with the guidelines established by the Fourth Military Medical University and with the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines.

e was backcrossed to C57BL/6 J more than 20 times. Young adult male (8–10 weeks old) mice were used in this study. All animal procedures were performed in accordance with the guidelines established by the Fourth Military Medical University and with the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines. Transient Middle Cerebral Artery Occlusion Mice were anesthetized with 1.5% isoflurane. Body temperature was monitored and maintained at 36.5 to 37.5 °C using a thermostatic pad. The right MCAO surgery was performed as described previously [26], and the cortical blood perfusion was monitored using laser-Doppler flowmetry. A sudden blood flow drop below 15–20% of the baseline value was considered to indicate sufficient occlusion (Fig. S1). After 60 min of occlusion, the mice were reanesthetized to facilitate the removal of blood vessel occlusion for recanalization and reperfusion.

od perfusion was monitored using laser-Doppler flowmetry. A sudden blood flow drop below 15–20% of the baseline value was considered to indicate sufficient occlusion (Fig. S1). After 60 min of occlusion, the mice were reanesthetized to facilitate the removal of blood vessel occlusion for recanalization and reperfusion. Immunohistochemistry (IHC) and Immunofluorescence (IF) Whole mice were fixed with 4% paraformaldehyde via cardiac perfusion. The brain tissues were flash-frozen on dry ice and cut into 12-μm-thick sections using a freezing microtome. First, the brain sections were incubated with the primary antibodies at 4 °C for 12 h. For the immunohistochemistry studies, the sections were incubated with biotinylated secondary antibodies at room temperature for 2 h. The immunostaining was visualized using the streptavidin/peroxidase complex and diaminobenzidine. For the immunofluorescence staining studies, the brain sections were incubated with fluorophore-conjugated secondary antibodies for 2 h at room temperature. The sections then were dyed with 4′,6-diamidino-2-phenylindole (DAPI) for nuclear staining. The staining results were observed and photographed under a laser-scanning confocal microscope. Additional information regarding the antibodies used in this study is provided in the Supplementary material (Table S1).

emperature. The sections then were dyed with 4′,6-diamidino-2-phenylindole (DAPI) for nuclear staining. The staining results were observed and photographed under a laser-scanning confocal microscope. Additional information regarding the antibodies used in this study is provided in the Supplementary material (Table S1). Neurological Score Assessment and Cerebral Infarct Volume Measurement After 60 min of MCAO and 24 h of reperfusion, a neurological score assessment was performed by two observers who were blinded to the study. The following rating scale was used: 0 = no deficit; 1 = failure to extend the right forepaw; 2 = decreased grip strength in the right forepaw; 3 = circling to the right after the tail was pulled; and 4 = spontaneous circling. Then, the mouse brains were removed and cut into coronal slices, and infarct volume was evaluated by 2,3,5-triphenyltetrazolium chloride (TTC) staining as described previously [26]. The edema index was calculated by dividing the right (ipsilateral to the transient middle cerebral artery occlusion (tMCAO)) hemisphere volume by the left (contralateral to the tMCAO) hemisphere volume. The infarct volume was corrected by dividing the infarct volume by the edema index. The infarct volume without the correction for edema was also provided.

iding the right (ipsilateral to the transient middle cerebral artery occlusion (tMCAO)) hemisphere volume by the left (contralateral to the tMCAO) hemisphere volume. The infarct volume was corrected by dividing the infarct volume by the edema index. The infarct volume without the correction for edema was also provided. Microdialysis and High-Performance Liquid Chromatography A microdialysis probe (4-mm length guide cannula, 0.22-mm membrane outer diameter, 1-mm membrane length, MW cutoff 50 kD; Eicom Corp, Tokyo, Japan) was stereotaxically inserted into the right striatum through the cannula guide (2 mm right lateral from the bregma, 0.5 mm anterior, and 4 mm ventral from the dura). Artificial cerebrospinal fluid (ACSF, in mM: NaCl, 124; KCl, 4.4; CaCl2, 2; NaHCO3, 25; MgSO4, 2; KH2PO4, 1; glucose, 10; pH 7.4) was perfused at a rate of 1 μl/min. The microdialysis samples were continuously collected for 4 h into microvials after 1 h at equilibrium, and these samples were subsequently lyophilized and redissolved in 20 μl of ACSF. The concentrations of glutamate in the microdialysis samples were analyzed by high-performance liquid chromatography (HPLC) as described previously [27]. The concentrations were calculated using LC solution software (Waters, USA) based on the standard samples (Sigma-Aldrich, USA).

ized and redissolved in 20 μl of ACSF. The concentrations of glutamate in the microdialysis samples were analyzed by high-performance liquid chromatography (HPLC) as described previously [27]. The concentrations were calculated using LC solution software (Waters, USA) based on the standard samples (Sigma-Aldrich, USA). Primary Astrocytic and Neuronal Cultures Primary astrocytes were obtained from the cerebral cortices of 1- to 3-day-old mouse pups. The brains were minced and trypsinized (0.25% trypsin-EDTA) to produce cell suspensions, which were then plated in poly-L-lysine-coated flasks in Dulbecco’s Modified Eagle’s Medium with 10% fetal bovine serum and maintained at 37 °C and 5% CO2. Every 3 days, half of the medium was removed and replaced. When the cells reached confluence after 10 to 14 days, the flasks were shaken at 200 to 220 rpm for 14 to 16 h to remove the microglia and oligodendrocytes. After shaking, the cultures included more than 95% astrocytes as determined by immunofluorescence staining for GFAP. After isolation, the cells were subcultured in different dishes according to distinct protocols.

ays, the flasks were shaken at 200 to 220 rpm for 14 to 16 h to remove the microglia and oligodendrocytes. After shaking, the cultures included more than 95% astrocytes as determined by immunofluorescence staining for GFAP. After isolation, the cells were subcultured in different dishes according to distinct protocols. The neuronal cultures were prepared from the cortices of 18-day-old mouse embryos obtained from pregnant mice by cesarean section. The cerebral cortices were extracted from the embryos and incubated for 30 min in 0.25% trypsin-EDTA. Digested tissues were dissociated by trituration and seeded on poly-L-lysine-coated plates. The culture medium consisted of Neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen). After 2 days, two-thirds of the medium was replaced with fresh medium. The cultures were maintained at 37 °C in a humidified 5% CO2 incubator. On the fifth day, the mature neurons were used for the experiments.

ed plates. The culture medium consisted of Neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen). After 2 days, two-thirds of the medium was replaced with fresh medium. The cultures were maintained at 37 °C in a humidified 5% CO2 incubator. On the fifth day, the mature neurons were used for the experiments. Adeno-Associated Virus Construction and Stereotaxic Injection NDRG2 adeno-associated virus (AAV) was constructed with the glial fibrillary acidic protein (GfaABC1D) promoter that selectively expresses proteins in astrocytes by Genechem Company (Shanghai, China). Intracerebroventricular injection of AAV was carried out using a stereotaxic instrument. Mice were anesthetized with sodium pentobarbital (50 mg/kg, i.p.). A 26-gauge cannula was stereotaxically inserted into the right lateral ventricle (coordinates: 0.22 mm anterior, 1.2 mm right lateral from the bregma, and 2.5 mm ventral from the dura), and 3 μl of AAV (1 × 1012 vector genomes/ml) was injected at a constant rate of 1.0 μl/min over 5 min using a micro-syringe pump. The needle was withdrawn slowly over 5 min. Three weeks after AAV injection, mice were subjected to microdialysis or tMCAO. Evaluation of Glutamate Uptake Fluorescein isothiocyanate (FITC) was linked to glutamate. Then, 200 μM FITC-coupled glutamate was added to the primary mice astrocyte cultures. At each time point, the fluorescence intensity in the cultured astrocytes was analyzed. Glutamate uptake was detected using a Live Cell Imaging System (Olympus, Japan).

Uptake Fluorescein isothiocyanate (FITC) was linked to glutamate. Then, 200 μM FITC-coupled glutamate was added to the primary mice astrocyte cultures. At each time point, the fluorescence intensity in the cultured astrocytes was analyzed. Glutamate uptake was detected using a Live Cell Imaging System (Olympus, Japan). Measurement of Intracellular Na+ The astrocytes were washed twice with PBS, and 1 μM CoroNa Green indicator (Invitrogen, USA) was then added by dilution from a concentrated stock solution in DMSO. The cells were incubated for 30 min at 37 °C. The loaded cells were washed twice with PBS before the fluorescence was measured. Subsequently, the CoroNa Green indicator was excited at 492 nm, and emission was collected above 516 nm. The images were observed and captured under a laser-scanning confocal microscope (Olympus, Japan).

ere incubated for 30 min at 37 °C. The loaded cells were washed twice with PBS before the fluorescence was measured. Subsequently, the CoroNa Green indicator was excited at 492 nm, and emission was collected above 516 nm. The images were observed and captured under a laser-scanning confocal microscope (Olympus, Japan). Na+/K+-ATPase Activity Assay The cellular Na+/K+-ATPase activity was measured via colorimetric determination using the ATPase Assay Kit (Innova Biosciences). The cells were lysed in deionized water with ultrasonic decomposition. The solubilized protein mixture was centrifuged to remove the cellular debris. The cell suspension (100 μl) was transferred to 2 × 96-well microplates with 100 μl of the substrate/buffer mix and incubated at 37 °C for 15 min. Then, 50 μl of Gold Mix was added to stop the reaction. After 2 min, 20 μl of the stabilizer was added, and the mixture was incubated at room temperature for 30 min. The absorbance was measured using an automated microplate reader at a wavelength of 590 nm. The Na+/K+-ATPase activity was calculated as the difference between the tested samples (total ATPase activity) and samples assayed in the presence of 2 mM ouabain.

d, and the mixture was incubated at room temperature for 30 min. The absorbance was measured using an automated microplate reader at a wavelength of 590 nm. The Na+/K+-ATPase activity was calculated as the difference between the tested samples (total ATPase activity) and samples assayed in the presence of 2 mM ouabain. Immunoblotting Analysis Proteins were extracted from tissues or cells using a lysis buffer composed of 150 mM NaCl, 50 mM Tris, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 1 mM EDTA, 1 mM Na3VO4 and a proteinase inhibitor mixture. The samples were then separated on 10% SDS-PAGE gels, transferred to polyvinylidene difluoride membranes and immunoblotted with the indicated antibodies. The blots were enhanced using a chemiluminescence detection reagent kit (Pierce, USA) and visualized using a Bio-Rad imager and Image Lab software (Bio-Rad, USA). See Supplementary material, Table S1, for more information regarding the antibodies. Plasmid Construction and Cell Transfection The plasmids of Flag-tagged full-length NDRG2, Myc-tagged full-length Na+/K+-ATPase β1 and its different truncation mutants were constructed by GeneChem (Shanghai, China) by inserting NDRG2 or different Na+/K+-ATPase β1 fragments into GV140 using XhoI/EcoRI restriction sites. The cells were grown to 80% confluence in 10-cm cell culture dishes before being transiently transfected using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s protocol. The cells were transfected for 48 h at 37 °C before harvesting for the biochemical analyses.

0 using XhoI/EcoRI restriction sites. The cells were grown to 80% confluence in 10-cm cell culture dishes before being transiently transfected using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s protocol. The cells were transfected for 48 h at 37 °C before harvesting for the biochemical analyses. Coimmunoprecipitation (Co-IP) The cells were incubated with 1 ml of lysis buffer containing 50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1% Lubrol, and a protease inhibitor mixture for 20 min on ice. The insoluble fraction was discarded after centrifugation at 10,000×g for 20 min at 4 °C. After centrifugation, the lysates were incubated with different antibodies and Sepharose beads conjugated with protein A/G for 10 h at 4 °C. The Sepharose beads were washed 4 times with the lysis buffer, and the proteins were eluted from the beads in sample buffer at 95 °C for 5 min. Subsequently, the proteins were detected by an immunoblotting (IB) analysis. Peptides All peptides used in this study were synthesized by Bankpeptide Biological Technology Company (Hefei, China). The peptides were HPLC-purified to reach a purity of more than 95%. All peptides were stored in a powder form and freshly diluted to 20 μM or 10 mg/ml for use in the cell- or animal-based experiments.

Coimmunoprecipitation (Co-IP) The cells were incubated with 1 ml of lysis buffer containing 50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1% Lubrol, and a protease inhibitor mixture for 20 min on ice. The insoluble fraction was discarded after centrifugation at 10,000×g for 20 min at 4 °C. After centrifugation, the lysates were incubated with different antibodies and Sepharose beads conjugated with protein A/G for 10 h at 4 °C. The Sepharose beads were washed 4 times with the lysis buffer, and the proteins were eluted from the beads in sample buffer at 95 °C for 5 min. Subsequently, the proteins were detected by an immunoblotting (IB) analysis. Peptides All peptides used in this study were synthesized by Bankpeptide Biological Technology Company (Hefei, China). The peptides were HPLC-purified to reach a purity of more than 95%. All peptides were stored in a powder form and freshly diluted to 20 μM or 10 mg/ml for use in the cell- or animal-based experiments. Cell Viability Assessment Cell death was quantitatively assessed using the LDH-Cytotoxicity Colorimetric Assay Kit (BioVision, USA), and the proportion of necrotic cells was detected (propidium iodide and Hoechst 33342 staining). LDH release was defined based on the ratio of LDH in the media to the total LDH and was normalized to the control. The quantification of the necrotic/healthy cells was performed by costaining the samples with 5 μM Hoechst 33342 and 2 μM propidium iodide, followed by blinded counts. The percentage of cell death was determined as the ratio of the number of neurons stained with propidium iodide to the number stained with Hoechst 33342.

he quantification of the necrotic/healthy cells was performed by costaining the samples with 5 μM Hoechst 33342 and 2 μM propidium iodide, followed by blinded counts. The percentage of cell death was determined as the ratio of the number of neurons stained with propidium iodide to the number stained with Hoechst 33342. Statistical Analysis The data are reported as the means ± SD or means ± SEM, and the analysis was performed using GraphPad Prism 6.0 software. The significance of the differences was determined by Student’s t test, one-way analysis of variance (ANOVA) followed by Tukey-Kramer’s post hoc test, repeated-measures one-way ANOVA, the Wilcoxon rank sum test, or the Kruskal-Wallis test unless otherwise specified. Statistical significance was defined as p < 0.05.

re. The significance of the differences was determined by Student’s t test, one-way analysis of variance (ANOVA) followed by Tukey-Kramer’s post hoc test, repeated-measures one-way ANOVA, the Wilcoxon rank sum test, or the Kruskal-Wallis test unless otherwise specified. Statistical significance was defined as p < 0.05. Results NDRG2 Knockout Exacerbates Cerebral Ischemic Damage and Glutamate Excitotoxicity NDRG2 was reported as an early-stage stress-responsive gene, and its expression can be induced by several types of stimulation, including cerebral ischemia [22]. Here, we also found that the expression of NDRG2 increased after cerebral ischemia in vivo or glutamate stimulation in vitro (Fig. S2). To explore the potential role of NDRG2 in cerebral ischemia, mice carrying an NDRG2 deletion that targeted exons 2–6 (Fig. 1a, top panel) were generated. These Ndrg2−/− mice were fully fertile and grew at a normal rate; no detectable brain structural differences were observed between the 8-week-old Ndrg2−/− mice and their WT littermates (Fig. 1a, bottom panel). We first examined whether NDRG2 is associated with the changed neurological outcomes in ischemia following a timeline (Fig. 1b). The infarct volume after MCAO and 24-h reperfusion was markedly larger in the Ndrg2−/− brains than that in the WT brains (Fig. 1c, d). We also observed that the Ndrg2−/− mice exhibited significantly increased neurological deficits after ischemia and 24-h reperfusion (Fig. 1e). Consistent with the 24-h reperfusion data, the KO mice exhibited an increased infarct volume compared with WT mice after 3 and 7 days of reperfusion (Fig. S3a and b). In addition, although the neurological outcome 7 days after reperfusion was better than that at 24 h or 3 days, the KO mice exhibited increased neurological deficits compared with the WT mice (Fig. S3c and d). No lesions were found in the Ndrg2−/− mice or WT mice after the sham operations (Fig. S3e).Fig. 1 NDRG2 deficiency exacerbates focal cerebral ischemia and promotes interstitial glutamate accumulation. a Schematic representation of the generation of the Ndrg2−/− (KO) mice and immunohistochemical identification using an NDRG2-specific antibody in WT and KO mouse brains. b The flow diagram of tMCAO, microdialysis, behavioral and infarct volume analysis. c–f WT and KO mice were subjected to MCAO and 24 h of reperfusion (n = 8).

tic representation of the generation of the Ndrg2−/− (KO) mice and immunohistochemical identification using an NDRG2-specific antibody in WT and KO mouse brains. b The flow diagram of tMCAO, microdialysis, behavioral and infarct volume analysis. c–f WT and KO mice were subjected to MCAO and 24 h of reperfusion (n = 8). Representative coronal brain sections stained with TTC (c), quantification of the infarct volume (d), and neurological scoring (e) after tMCAO. (f) Microdialysis samples were collected from extracellular fluid in the striatum after tMCAO as indicated. The glutamate concentrations in the microdialysis samples were measured by HPLC. *p < 0.05, **p < 0.01, Wilcoxon rank sum test (d, e). **p < 0.01, Student’s t test (f). Error bars, mean ± SD (d, f); Horizontal bars, medians (e) The sudden cessation of cerebral blood flow is accompanied by a marked increase in glutamate release [28]. Glutamate excitotoxicity plays a prominent role in ischemic cerebral injury [29]. We previously found increased glutamate in the brains of Ndrg2−/− mice, which exhibit attention deficit/hyperactivity-like behavior [25]. Therefore, we examined the interstitial glutamate levels in the Ndrg2−/− mice brains after ischemia by microdialysis. The levels of interstitial glutamate were significantly higher in the striatum of the Ndrg2−/− mice than those in the striatum of the WT mice both at the baseline level and after tMCAO treatment (Fig. 1f). These data suggest that NDRG2 plays a neuroprotective role in cerebral ischemia by regulating glutamate transport.

The levels of interstitial glutamate were significantly higher in the striatum of the Ndrg2−/− mice than those in the striatum of the WT mice both at the baseline level and after tMCAO treatment (Fig. 1f). These data suggest that NDRG2 plays a neuroprotective role in cerebral ischemia by regulating glutamate transport. Impaired Glutamate Uptake in Ndrg2−/− Astrocytes Because NDRG2 is mainly expressed in astrocytes [17–19], we investigated the glutamate uptake ability in cultured Ndrg2−/− astrocytes. FITC-coupled glutamate was used to visualize dynamic glutamate trafficking. The speed and amount of the FITC-coupled glutamate uptake were significantly lower in the Ndrg2−/− astrocytes than those in the WT astrocytes at various time points (Fig. 2a, b). We constructed an adeno-associated virus (AAV) that specifically expresses NDRG2 in astrocytes. Strikingly, the NDRG2 rescue treatment reversed the decreased FITC-coupled glutamate uptake into Ndrg2−/− astrocytes (Fig. 2a, b). Here, we observed that NDRG2 levels are related to glutamate uptake in real time, although we previously also found decreased glutamate clearance in cultured Ndrg2−/− astrocytes by indirectly detecting the remnant glutamate in the culture medium [25].Fig. 2 NDRG2 is required for the sodium-dependent glutamate uptake into astrocytes. WT astrocytes, KO astrocytes, and KO astrocytes infected with a control virus (AAV-Ctrl) or the NDRG2 virus (AAV-NDRG2). a Representative photomicrographs of FITC-coupled glutamate uptake at various time points. Scale bar = 100 μm. b Green fluorescence intensity of FITC-coupled glutamate in each group shown in a. Data are presented as the means ± SEM of three independent experiments, each of which was performed in 5 fields of view and evaluated using repeated-measures one-way ANOVA. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl. c Representative fluorescence images showing CoroNa Green (sodium indicator) trapped inside astroglial cytoplasm and counterstained with Hoechst 33342. Scale bar = 100 μm. d Green fluorescence intensity of CoroNa in each group shown in d. Data are presented as the means ± SEM of three independent experiments, each performed in 5 fields of view and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl. e Na+/K+-ATPase activity was determined following the experimental procedure in different astrocytes.

as the means ± SEM of three independent experiments, each performed in 5 fields of view and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl. e Na+/K+-ATPase activity was determined following the experimental procedure in different astrocytes. Data are expressed as the means ± SEM from three independent determinations, each of which was performed in quadruplicate and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl

as the means ± SEM of three independent experiments, each performed in 5 fields of view and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl. e Na+/K+-ATPase activity was determined following the experimental procedure in different astrocytes. Data are expressed as the means ± SEM from three independent determinations, each of which was performed in quadruplicate and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl Sodium-dependent and sodium-independent glutamate uptake in astrocytes are important for preventing excitotoxicity [6]. We previously found that NDRG2 can interact with Na+/K+-ATPase β1 [30], which is required for Na+ transport. To determine whether the NDRG2-related glutamate uptake is sodium-dependent, we examined the intracellular Na+ concentration and the activity of Na+/K+-ATPase in cultured astrocytes. The Na+ fluorescence imaging revealed a substantially higher amount of intracellular Na+ in the Ndrg2−/− astrocytes than in the WT astrocytes (Fig. 2c, d), which is suggestive of a decreased transmembrane Na+ gradient in the Ndrg2−/− astrocytes. The NDRG2-silenced astrocytes showed a reduction in the activity of Na+/K+-ATPase (Fig. 2e), which is consistent with our previous study in salivary epithelial cells [30]. In addition, attenuated intracellular Na+ and increased Na+/K+-ATPase activity were detected following the rescue of NDRG2 expression via the AAV treatment in the Ndrg2−/− astrocytes. Together, these results indicate that NDRG2 is required for the maintenance of sodium-dependent astroglial glutamate uptake.

cells [30]. In addition, attenuated intracellular Na+ and increased Na+/K+-ATPase activity were detected following the rescue of NDRG2 expression via the AAV treatment in the Ndrg2−/− astrocytes. Together, these results indicate that NDRG2 is required for the maintenance of sodium-dependent astroglial glutamate uptake. Because stroke results from a combination of excitotoxicity and energy pump failure, we next tested astrocytic Na+/K+-ATPase activity and glutamate uptake under oxygen-glucose deprivation/reoxygenation (OGD/OGR) conditions. Na+/K+-ATPase activity and glutamate uptake were significantly reduced in both KO and WT astrocytes after 2 h of OGD and 6 h of OGR. Moreover, KO astrocytes exhibited a more severe impairment in Na+/K+-ATPase activity and glutamate uptake upon OGD compared with WT astrocytes (Fig. S4a and b). Therefore, these data suggest that NDRG2 deficiency aggravates the disruption of Na+/K+-ATPase function and Na+/K+-ATPase-associated glutamate uptake under energy blocking.

astrocytes exhibited a more severe impairment in Na+/K+-ATPase activity and glutamate uptake upon OGD compared with WT astrocytes (Fig. S4a and b). Therefore, these data suggest that NDRG2 deficiency aggravates the disruption of Na+/K+-ATPase function and Na+/K+-ATPase-associated glutamate uptake under energy blocking. Na+/K+-ATPase β1 Is Involved in NDRG2-Regulated Glutamate Uptake To determine the mechanism by which an NDRG2 deficiency leads to impaired glutamate uptake, we examined the expression levels of the molecules implicated in this process. The IB analysis revealed that the levels of EAAT1, EAAT2, and Na+/K+-ATPase β1, but not the level of Na+/K+-ATPase α1, were markedly decreased in the cortex and striatum of the Ndrg2−/− mice (Fig. 3a, b). These results suggest that NDRG2 stabilizes the Na+/K+-ATPase β1 protein in astrocytes, consistent with our previous report indicating that NDRG2 inhibited the ubiquitination and degradation of Na+/K+-ATPase β1 in salivary epithelial cells [30]. Next, we injected the NDRG2 AAV into Ndrg2−/− mice via the lateral ventricles, which specifically rescued the astrocytic NDRG2 expression in the brain, including striatal ischemic penumbra (Fig. S5). The expression levels of Na+/K+-ATPase β1, EAAT1, and EAAT2 were remarkably recovered by restoring the expression of NDRG2 in the cortex and striatum of the Ndrg2−/− mice (Fig. 3a, b).Fig. 3 NDRG2 promotes glutamate uptake by regulating Na+/K+-ATPase β1. a Representative immunoblots of NDRG2, Na+/K+-ATPase α1 (α1), Na+/K+-ATPase β1 (β1), EAAT1, and EAAT2 proteins from the cortex and striatum of WT, KO, and KO mice injected with a control virus (AAV-Ctrl) or the NDRG2 virus (AAV-NDRG2). b The data shown in a were quantified and normalized to β-tubulin. Data are presented as the means ± SEM of three independent experiments and were evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, ##p < 0.01 versus KO + AAV-Ctrl. c Representative photomicrographs of FITC-coupled glutamate uptake in KO astrocytes infected with a control virus (AAV-Ctrl) or the Na+/K+-ATPase β1 virus (AAV-β1) at various time points. Scale bar = 100 μm. d Green fluorescence intensity of FITC-coupled glutamate in each group shown in c. Data are presented as the means ± SEM of three independent experiments, each of which was performed in 5 fields of view and evaluated using repeated-measures one-way ANOVA. **p < 0.01 versus KO + AAV-Ctrl.

ime points. Scale bar = 100 μm. d Green fluorescence intensity of FITC-coupled glutamate in each group shown in c. Data are presented as the means ± SEM of three independent experiments, each of which was performed in 5 fields of view and evaluated using repeated-measures one-way ANOVA. **p < 0.01 versus KO + AAV-Ctrl. e Representative fluorescence images showing CoroNa Green (sodium indicator) trapped inside the KO astrocytes infected with AAV-Ctrl or AAV-β1 and counterstained with Hoechst 33342. Scale bar = 100 μm. f Green fluorescence intensity of CoroNa in each group shown in e. Data are presented as the means ± SEM of three independent experiments, each of which was performed in 5 fields of view and evaluated using Student’s t test. *p < 0.05 versus KO + AAV-Ctrl. g Na+/K+-ATPase activity was determined following the experimental procedure in different astrocytes. Data are expressed as the means ± SEM of three independent determinations, each of which was performed in quadruplicate and evaluated using Student’s t test. **p < 0.01 versus KO + AAV-Ctrl

< 0.05 versus KO + AAV-Ctrl. g Na+/K+-ATPase activity was determined following the experimental procedure in different astrocytes. Data are expressed as the means ± SEM of three independent determinations, each of which was performed in quadruplicate and evaluated using Student’s t test. **p < 0.01 versus KO + AAV-Ctrl We previously screened 12 potential interaction partners of NDRG2 using the yeast two-hybrid system [30]; of these candidate partners, three were identified as Na+/K+-ATPase β1 (Table S2). We hypothesized that Na+/K+-ATPase β1 is involved in NDRG2-mediated glutamate uptake because Na+/K+-ATPase β1 plays a critical role in the membrane translocation of Na+/K+-ATPase and is required for enzymatic activity [31, 32]. We confirmed that exogenous NDRG2 interacted with exogenous Na+/K+-ATPase β1 in HEK293 cells (Fig. S6a). We further detected endogenous NDRG2 coprecipitated with endogenous Na+/K+-ATPase β1 in cultured astrocytes (Fig. S6b). In addition, we constructed an AAV that specifically expresses Na+/K+-ATPase β1 in astrocytes. The decreased uptake of FITC-coupled glutamate into Ndrg2−/− astrocytes was significantly rescued by the Na+/K+-ATPase β1 AAV treatment (Fig. 3c, d). Attenuated intracellular Na+ and increased activity of Na+/K+-ATPase were detected following the rescue of NDRG2 expression in the Ndrg2−/− astrocytes (Fig. 3e–g). Together, these results suggest that Na+/K+-ATPase β1 is critical for the expression and function of EAAT1/2 and underlies the NDRG2-regulated glutamate uptake.

nuated intracellular Na+ and increased activity of Na+/K+-ATPase were detected following the rescue of NDRG2 expression in the Ndrg2−/− astrocytes (Fig. 3e–g). Together, these results suggest that Na+/K+-ATPase β1 is critical for the expression and function of EAAT1/2 and underlies the NDRG2-regulated glutamate uptake. A Na+/K+-ATPase β1 Peptide Attenuates NDRG2-Mediated Glutamate Uptake To identify the key segment of Na+/K+-ATPase β1 binding with NDRG2, we truncated Na+/K+-ATPase β1 into peptides of different lengths. Deletion mapping using these truncated peptides of Na+/K+-ATPase β1 to conduct the coimmunoprecipitation showed that the region encompassing amino acid residues 244–304 is critical for the interaction between Na+/K+-ATPase β1 and NDRG2 (Fig. 4a, b). However, a peptide containing 60 amino acids is still too large. Therefore, the 60 amino acid residues were subdivided into three peptides, each containing 20 amino acid residues. We further combined each peptide with an HIV trans-activator of transcription (TAT: YGRKKRRQRRR) sequence, which allows the peptide to cross the blood-brain barrier and cell membrane. Therefore, we synthesized three peptides containing a TAT sequence, which allows the peptide to cross the blood-brain barrier and cell membrane [33, 34], and subdivided regions of the 244–304 peptide sequence (Fig. 4c). These peptides were individually added to cultured WT astrocytes. TAT-β1244–263, but not TAT-β1264–283 or TAT-β1284–304, competitively blocked the binding of NDRG2 to endogenous Na+/K+-ATPase β1 in astrocytes (Fig. 4d).Fig. 4 Dissociation of the NDRG2-Na+/K+-ATPase β1 interaction abolishes the NDRG2-mediated cerebral protection. a Co-IP analysis of NDRG2 and Na+/K+-ATPase β1 (β1). Flag-tagged full-length NDRG2 was cotransfected with Myc-tagged full-length Na+/K+-ATPase β1 or its truncated mutants into HEK293 cells for 48 h. IP was performed using an anti-Myc antibody. Precipitated proteins were analyzed by IB with an anti-Flag antibody or anti-Myc antibody (upper panel). IP was also performed using an anti-Flag antibody and analyzed by IB with anti-Myc and anti-Flag antibodies (lower panel). Similar results were observed in three independent experiments. b Schematic of the interactions between Myc-tagged full-length Na+/K+-ATPase β1 and its different truncation mutants with Flag-tagged full-length NDRG2. c Synthetic peptides TAT-β1244–263, TAT-β1264–283, and TAT-β1284–304 and the control peptide TAT-Ctrl with a scrambled sequence.

rved in three independent experiments. b Schematic of the interactions between Myc-tagged full-length Na+/K+-ATPase β1 and its different truncation mutants with Flag-tagged full-length NDRG2. c Synthetic peptides TAT-β1244–263, TAT-β1264–283, and TAT-β1284–304 and the control peptide TAT-Ctrl with a scrambled sequence. d Co-IP analysis of endogenous NDRG2 and Na+/K+-ATPase β1 was performed following different peptide treatments in cultured astrocytes for 2 h. IP and IB analyses were performed as indicated. Similar results were obtained from three independent experiments. e Representative immunoblots of Na+/K+-ATPase β1 following TAT-Ctrl or TAT-β1244–263 administration with 100 μM emetine treatment. β-tubulin served as a loading control. f Relative levels of Na+/K+-ATPase β1 to β-tubulin shown in e were quantified by densitometry, and the percentage of the value at 0 min was calculated at various time points (from 0 to 150 min). Data are shown as the means ± SEM of three independent experiments. g Representative photomicrographs of FITC-coupled glutamate uptake after the TAT-Ctrl or TAT-β1244–263 treatment in cultured astrocytes. Scale bar = 100 μm. Similar results were obtained in each of three independent experiments. h Green fluorescence intensity of FITC-coupled glutamate in each group shown in g. Data are presented as the means ± SEM of three independent experiments, each of which was performed in 5 fields of view and evaluated using repeated-measures one-way ANOVA. *p < 0.05 versus TAT-Ctrl. i–l Mice were intravenously injected with TAT-Ctrl or TAT-β1244–263 2 h before tMCAO (n = 8). i The concentrations of glutamate in the microdialysis samples were measured using HPLC. Student’s t test. *p < 0.05 versus TAT-Ctrl. Representative coronal brain sections stained with TTC (j), quantification of infarct volume (k), and neurological score (l) after tMCAO. i, k, Student’s t test; l, Wilcoxon test. *p < 0.05, **p < 0.01 versus TAT-Ctrl. Error bars, mean ± SD (i, k); horizontal bars, medians (l)

udent’s t test. *p < 0.05 versus TAT-Ctrl. Representative coronal brain sections stained with TTC (j), quantification of infarct volume (k), and neurological score (l) after tMCAO. i, k, Student’s t test; l, Wilcoxon test. *p < 0.05, **p < 0.01 versus TAT-Ctrl. Error bars, mean ± SD (i, k); horizontal bars, medians (l) Therefore, we hypothesize that the TAT-β1244–263 peptide can competitively combine with endogenous NDRG2 and disturb the binding of NDRG2-Na+/K+-ATPase β1, leading to faster degradation of endogenous Na+/K+-ATPase β1. Next, we examined the protein stability of Na+/K+-ATPase β1 and glutamate uptake after TAT-β1244–263 treatment in the astrocytes. The half-life of endogenous Na+/K+-ATPase β1 following the TAT-β1244–263 treatment (65 min) was much shorter than that following the TAT-Ctrl treatment (110 min) (Fig. 4e, f), indicating that endogenous Na+/K+-ATPase β1 is less stable when unable to interact with NDRG2. In addition, the speed and amount of the astroglial glutamate uptake were significantly reduced after the TAT-β1244–263 administration (Fig. 4g, h). Increased intracellular Na+ and reduced Na+/K+-ATPase activity were also detected in the cultured astrocytes after the TAT-β1244–263 treatment (Fig. S7). Altogether, these findings suggest that the NDRG2-Na+/K+-ATPase β1 interaction is required not only for the stability of Na+/K+-ATPase β1 but also for sodium-dependent glutamate uptake into astrocytes.

uced Na+/K+-ATPase activity were also detected in the cultured astrocytes after the TAT-β1244–263 treatment (Fig. S7). Altogether, these findings suggest that the NDRG2-Na+/K+-ATPase β1 interaction is required not only for the stability of Na+/K+-ATPase β1 but also for sodium-dependent glutamate uptake into astrocytes. We further investigated whether the NDRG2-Na+/K+-ATPase β1 interaction is required for the NDRG2-mediated neuronal protection against glutamate excitotoxicity in vitro. The astrocytes were pretreated with TAT-β1244–263 or TAT-Ctrl for 2 h and then indirectly cocultured with neurons. The glutamate clearance was markedly decreased in the TAT-β1244–263 pretreated group compared with that in the TAT-Ctrl pretreated group after 2 h of glutamate exposure (Fig. S8a). We also observed significant increases in the LDH release and neuronal death in the TAT-β1244–263 pretreated group after glutamate administration (Fig. S8b and c).

was markedly decreased in the TAT-β1244–263 pretreated group compared with that in the TAT-Ctrl pretreated group after 2 h of glutamate exposure (Fig. S8a). We also observed significant increases in the LDH release and neuronal death in the TAT-β1244–263 pretreated group after glutamate administration (Fig. S8b and c). To determine whether the NDRG2-Na+/K+-ATPase β1 association is required for the NDRG2-mediated neuroprotection in vivo, we performed tMCAO experiments. The mice were intravenously injected with either TAT-Ctrl or TAT-β1244–263 2 h before tMCAO (Fig. S9a). Consistent with the results of the in vitro experiments, TAT-β1244–263 triggered a dissociation of the NDRG2-Na+/K+-ATPase β1 interaction in the cortex and striatum of the mouse brain (Fig. S9b). Compared with TAT-Ctrl, TAT-β1244–263 induced a significant increase in interstitial glutamate after ischemia (Fig. 4i). The brain infarct volume was notably larger in the TAT-β1244–263-injected mice compared with the TAT-Ctrl-injected mice after tMCAO (Fig. 4j, k). Consistently, we observed that the ischemia-associated neurological deficits were significantly worsened in the TAT-β1244–263-injected mice compared with those in the TAT-Ctrl-injected mice (Fig. 4l). Altogether, these results indicate that the NDRG2-Na+/K+-ATPase β1 interaction is functionally necessary for NDRG2-mediated neuroprotection against glutamate excitotoxicity.

neurological deficits were significantly worsened in the TAT-β1244–263-injected mice compared with those in the TAT-Ctrl-injected mice (Fig. 4l). Altogether, these results indicate that the NDRG2-Na+/K+-ATPase β1 interaction is functionally necessary for NDRG2-mediated neuroprotection against glutamate excitotoxicity. Astroglial NDRG2 Is Required for Neuronal Survival under Glutamate Excitotoxicity Next, we used mixed neuron-astrocyte cultures to detect whether astroglial NDRG2 is associated with neuronal survival following a glutamate challenge in vitro. The mixed cultures were obtained from indirect neuron and astrocyte cocultures (WT or Ndrg2−/− astrocytes were plated in the upper chamber, and WT neurons were seeded in the lower chamber), allowing the two cell types to share the same diffusible materials but remain divided by a physical filter (Fig. 5a). These indirect neuron-astrocyte cocultures were treated with 200 μM glutamate for 2 h. The remaining glutamate in the medium of the Ndrg2−/− astrocyte-neuron cocultures was more than two times higher than that in the medium of the WT astrocyte-neuron cocultures (Fig. 5b). This finding is consistent with the FITC-coupled glutamate uptake results in monocultures of Ndrg2−/ astrocytes (Fig. 2a, b).Fig. 5 NDRG2 deletion increases neuronal death following a glutamate challenge in vitro. Neurons from WT mice were cocultured with WT astrocytes, KO astrocytes, and KO astrocytes infected with a control virus (AAV-Ctrl) or the NDRG2 virus (AAV-NDRG2). a Schematic diagram of the neuron-astrocyte indirect cocultures. Astrocytes were seeded on the top filter, and WT neurons were planted in the bottom chamber. b Cocultures were treated with 200 μM glutamate, and the glutamate levels in the culture medium were measured at various time points. Data represent the means ± SEM of three independent determinations, each of which was performed in quadruplicate and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl. Quantification of LDH release from neurons (c) and neuronal death percentage (d) after 2 h of glutamate treatment and another 24 h of further culture without glutamate. Data represent the means ± SEM of three independent determinations, each of which was performed in quadruplicate and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl

of glutamate treatment and another 24 h of further culture without glutamate. Data represent the means ± SEM of three independent determinations, each of which was performed in quadruplicate and evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT, #p < 0.05 versus KO + AAV-Ctrl We also examined the cell viability of the neurons. The Ndrg2−/− astrocyte-neuron cocultures showed higher lactate dehydrogenase (LDH) release from neurons (Fig. 5c) and more neuronal death (Fig. 5d) than the WT astrocyte-neuron cocultures after glutamate treatment. The increased NDRG2 expression in the Ndrg2−/− astrocytes effectively rescued the glutamate clearance (Fig. 5b) and attenuated the neuronal LDH release and neuronal death (Fig. 5c, d). These results indicate that NDRG2 protects neurons from glutamate excitotoxicity by promoting astroglial glutamate clearance.

reatment. The increased NDRG2 expression in the Ndrg2−/− astrocytes effectively rescued the glutamate clearance (Fig. 5b) and attenuated the neuronal LDH release and neuronal death (Fig. 5c, d). These results indicate that NDRG2 protects neurons from glutamate excitotoxicity by promoting astroglial glutamate clearance. NDRG2 Confers Neuroprotection Against Ischemia Glutamate-induced excitotoxicity plays a prominent role in various neurological disorders, such as ischemic cerebral injury [29]. We therefore explored whether an NDRG2 treatment could improve neurological function after brain ischemia. An intracerebroventricular injection of the NDRG2 AAV effectively decreased the infarct volume and neurological deficits in both WT and Ndrg2−/− mice after tMCAO (Fig. 6a–c).Fig. 6 NDRG2 protects the brain from ischemic stroke. WT and KO mice injected with a control virus (AAV-Ctrl) or the NDRG2 virus (AAV-NDRG2) were subjected to tMCAO. Representative coronal brain sections stained with TTC (a), quantification of infarct volume (b), and neurological score (c) after tMCAO (n = 8). b One-way ANOVA followed by Tukey-Kramer’s post hoc test; c Kruskal-Wallis test. **p < 0.05, **p < 0.01 versus WT + AAV-Ctrl or KO + AAV-Ctrl. d Microdialysis samples were collected from striatal extracellular fluid after tMCAO as indicated. The concentrations of glutamate in the microdialysis samples were measured by HPLC. Data were evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT + AAV-Ctrl or KO + AAV-Ctrl. Error bars, mean ± SD (b, d); horizontal bars, medians (c). e A working model of astroglial NDRG2-mediated neuroprotection following brain ischemia

odialysis samples were measured by HPLC. Data were evaluated using one-way ANOVA followed by Tukey-Kramer’s post hoc test. **p < 0.01 versus WT + AAV-Ctrl or KO + AAV-Ctrl. Error bars, mean ± SD (b, d); horizontal bars, medians (c). e A working model of astroglial NDRG2-mediated neuroprotection following brain ischemia Thus, we investigated the interstitial glutamate levels in the mouse brain after ischemia. The increase in interstitial glutamate was significantly attenuated after injection of the NDRG2 AAV into WT and Ndrg2−/− brains after tMCAO (Fig. 6d). In addition, the NDRG2 AAV injection effectively decreased the interstitial glutamate levels in the Ndrg2−/− brains after the sham operation (Fig. S10). Collectively, these results indicate that astroglial NDRG2 plays a neuroprotective role in cerebral ischemia and is a potential therapeutic target in brain stroke. Discussion The results of the present study revealed that NDRG2 plays an important role in maintaining the synaptic glutamate balance and alleviating glutamate excitotoxicity. NDRG2 is required for astroglial glutamate uptake from cerebral interstitial fluid, as it interacts with Na+/K+-ATPase β1, which protects the brain from ischemic stroke.

the present study revealed that NDRG2 plays an important role in maintaining the synaptic glutamate balance and alleviating glutamate excitotoxicity. NDRG2 is required for astroglial glutamate uptake from cerebral interstitial fluid, as it interacts with Na+/K+-ATPase β1, which protects the brain from ischemic stroke. In mammalian brains, NDRG2 is primarily expressed in astrocytes and has been considered a new marker of mature astrocytes [17–19]. However, the functions of NDRG2 in astrocytes remain largely unknown. Astrocytes are the predominant cells contributing to the removal of excitatory glutamate from the synaptic space [35]. Here, we found that Ndrg2−/− mice exhibited increased interstitial glutamate in the brain and that impaired astroglial glutamate uptake underlies the glutamate accumulation. Many studies have focused on the modulation of glutamate uptake by astrocytic EAATs, which are driven by the transmembrane Na+ gradient [11, 36, 37]. In Xenopus laevis oocytes, NDRG2 stimulates an amiloride-sensitive Na+ current by regulating the epithelial sodium channel [38]. We determined that NDRG2 is required for the Na+/K+-ATPase-mediated Na+ current in salivary cells [39]. Here, we demonstrated that NDRG2 silencing resulted in a build-up of intracellular Na+ and a decrease in Na+/K+-ATPase activity in astrocytes. Therefore, we speculate that NDRG2 regulates the function of EAATs through Na+/K+-ATPase. Although most mammalian cells maintain a very stable intracellular Na+ concentration under physiological conditions, in astrocytes, the Na+ concentrations widely fluctuate to buffer neighboring extracellular Na+ derived from neurons and other cells in the brain [40–42], which may explain why the cultured Ndrg2−/− astrocytes could survive a high intracellular Na+ concentration.

ellular Na+ concentration under physiological conditions, in astrocytes, the Na+ concentrations widely fluctuate to buffer neighboring extracellular Na+ derived from neurons and other cells in the brain [40–42], which may explain why the cultured Ndrg2−/− astrocytes could survive a high intracellular Na+ concentration. We previously found that NDRG2 interacts with and stabilizes Na+/K+-ATPase β1 [31, 43]. In the present study, we developed a Na+/K+-ATPase β1 peptide that could dissociate the NDRG2- Na+/K+-ATPase β1 interaction, sequentially leading to a disruption in sodium-dependent glutamate uptake and glutamate excitotoxicity. Na+/K+-ATPase β1 plays important roles in the membrane translocation and enzyme function of the alpha subunit of Na+/K+-ATPase [31, 43]. The Na+/K+-ATPase α subunit and EAATs are physically associated within a macromolecular complex in the plasma membrane [15]. Thus, a dynamic NDRG2-Na+/K+-ATPase β1-EAAT complex may modulate sodium-dependent glutamate uptake in astrocytes [25].

location and enzyme function of the alpha subunit of Na+/K+-ATPase [31, 43]. The Na+/K+-ATPase α subunit and EAATs are physically associated within a macromolecular complex in the plasma membrane [15]. Thus, a dynamic NDRG2-Na+/K+-ATPase β1-EAAT complex may modulate sodium-dependent glutamate uptake in astrocytes [25]. The NDRG2-mediated glutamate uptake and glutamate homeostasis are not only required for the physiological function of the brain but also enhance the tolerance of the brain to glutamate excitotoxicity under neuropathological conditions. While NDRG2 is normally expressed at a moderate level in the brain, its expression was notably increased under excitotoxic conditions, including cerebral ischemia and trauma [22, 24, 44]. The upregulated NDRG2 expression may reflect the brain’s attempt to maintain the glutamate balance and protect itself from excitotoxicity under restricted blood supply and other neuropathological conditions. Increasing or rescuing the expression of NDRG2 in WT and Ndrg2−/− mice significantly reduces the infarct volume, neurological outcome, and interstitial glutamate levels, suggesting a potentially therapeutic target for brain ischemia. A recent study reported that the loss of NDRG2 increases astrocytic MMP activity and BBB permeability, which exacerbate subsequent brain damage after permanent focal cerebral ischemia [45]. This is consistent with the function of NDRG2 as a neuroprotective protein observed in the present study, although we focused on astrocyte-mediated glutamate uptake but not astrocyte-associated blood cell infiltration and BBB permeability.

exacerbate subsequent brain damage after permanent focal cerebral ischemia [45]. This is consistent with the function of NDRG2 as a neuroprotective protein observed in the present study, although we focused on astrocyte-mediated glutamate uptake but not astrocyte-associated blood cell infiltration and BBB permeability. Here, we used the Ndrg2−/− mice to explore the underlying function of NDRG2 in cerebral ischemia and concluded that NDRG2 plays a neuroprotective role. However, the results of another report on the role of NDRG2 in stroke [46] were contradictory to our conclusion. In that study, the OGD model in vitro was used to mimic the situation in which astrocytes suffered from ischemic-reperfusion injury, and NDRG2 silencing significantly reduced astrocytic reactive oxygen species production and apoptosis after exposure to OGR. This discrepancy could be attributed to the difference between in vivo and in vitro pathological models. Thus, more studies are needed to clarify this issue in the future.

mic-reperfusion injury, and NDRG2 silencing significantly reduced astrocytic reactive oxygen species production and apoptosis after exposure to OGR. This discrepancy could be attributed to the difference between in vivo and in vitro pathological models. Thus, more studies are needed to clarify this issue in the future. In addition, in a mouse model of cortical stab injury, the NDRG2 expression was elevated in astrocytes surrounding the wounded area and deletion of NDRG2 resulted in a lower induction of reactive astrogliosis and inflammatory response in the injured cortex [44]. NDRG2 also participated in other signaling pathways relating to neuronal death. In a previous study [47], the authors used phospho-peptide library screening and found that phosphorylation of NDRG2 on Ser350 by DAPK1 could be a novel mechanism of NDRG2 activation and could be involved in neuronal cell death. This result indicates that NDRG2 may play multiple roles in the central nervous system.

th. In a previous study [47], the authors used phospho-peptide library screening and found that phosphorylation of NDRG2 on Ser350 by DAPK1 could be a novel mechanism of NDRG2 activation and could be involved in neuronal cell death. This result indicates that NDRG2 may play multiple roles in the central nervous system. In summary, our results support a scenario (Fig. 6e) of NDRG2-mediated neuroprotection. In response to cerebral ischemia, excessive glutamate is released from the presynaptic membrane, and astroglial NDRG2 is rapidly upregulated. Increased NDRG2 interacts with Na+/K+-ATPase β1 and facilitates the Na+/K+-ATPase-triggered transmembrane Na+ gradient, leading to the cotransportation of both glutamate and Na+ into astrocytes via the EAATs. Subsequently, glutamate receptors are appropriately activated, but not overloaded, to maintain excitatory synaptic transmission. Therefore, the accelerated glutamate clearance protects the neurons from further glutamate excitotoxicity. Our findings provide a potential target for future interventions for glutamate excitotoxicity-associated diseases such as ischemia. Electronic Supplementary Material ESM 1 (DOCX 4136 kb) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Anqi Yin and Hang Guo contributed equally to this work.

In summary, our results support a scenario (Fig. 6e) of NDRG2-mediated neuroprotection. In response to cerebral ischemia, excessive glutamate is released from the presynaptic membrane, and astroglial NDRG2 is rapidly upregulated. Increased NDRG2 interacts with Na+/K+-ATPase β1 and facilitates the Na+/K+-ATPase-triggered transmembrane Na+ gradient, leading to the cotransportation of both glutamate and Na+ into astrocytes via the EAATs. Subsequently, glutamate receptors are appropriately activated, but not overloaded, to maintain excitatory synaptic transmission. Therefore, the accelerated glutamate clearance protects the neurons from further glutamate excitotoxicity. Our findings provide a potential target for future interventions for glutamate excitotoxicity-associated diseases such as ischemia. Electronic Supplementary Material ESM 1 (DOCX 4136 kb) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Anqi Yin and Hang Guo contributed equally to this work. We thank He Zhang at Loma Linda University for the critical evaluation of the manuscript. We also thank many colleagues at the Fourth Military Medical University: Xiangyang Qin in the Department of Pharmacy for the synthesis of the FITC-coupled glutamate, Qi Yang in the Department of Pharmacy for providing technical help with the HPLC analysis, Kailong Zhu in the Department of Anesthesiology for assisting with the animal behavioral scoring after tMCAO, and Ting Gu in the Department of Anesthesiology for her kind assistance in isolating the primary neurons and astrocytes.

mRNA expression [1–6]. Though these studies demonstrated proof-of-principle, most of the stroke transcriptome which is comprised of all alternatively spliced isoforms remains unstudied in stroke. The importance of alternative splicing is supported by evidence implicating it in the pathogenesis of many diseases [7, 8] . Alternative splicing is the process whereby exons from a single gene are included or excluded in the final mRNA transcript (Supplementary Figure 1). A single gene can produce several alternatively spliced isoforms which have specific functions in different cells, tissues, developmental stages, and disease states. Thus, the ~20,000 known genes code for >250,000 different mRNAs and proteins. Differential alternative splicing (DAS) is alternative splicing that differs between groups. We hypothesized that DAS would vary for different causes of IS (cardioembolic, large vessel, and lacunar) and for ICH when compared to each other and to controls. RNA-seq is a new technology that allows for estimation of expression of each splice variant (Supplementary Figure 1), a significant advance over previous technologies. Because there have been no studies of alternative splicing related to stroke etiology, or for IS versus ICH either in humans or in animal models, we performed this pilot RNA-seq study to examine DAS in whole blood following IS and ICH in humans.

Supplementary Figure 1), a significant advance over previous technologies. Because there have been no studies of alternative splicing related to stroke etiology, or for IS versus ICH either in humans or in animal models, we performed this pilot RNA-seq study to examine DAS in whole blood following IS and ICH in humans. Methods Stroke patients and control subjects were randomly selected from those recruited at the University of California Davis Medical Center between 2008 and 2012. Stroke patients were chosen to represent the major IS etiologies (cardioembolic, large vessel atherosclerotic, lacunar) or had ICH. IS diagnosis and causes were assessed as described previously [5, 9]. ICH patients had deep ICH (basal ganglia, deep white matter, or thalamus) confirmed by CT and/or MRI brain scans and were associated with hypertension without evidence of vascular malformation, tumor, or aneurysm. Control subjects were selected to match stroke subjects for age, race, sex, and vascular risk factors and had no history of previous stroke or cardiovascular events. Blood from all subjects was drawn into PAXgene tubes between 5.8 and 101.2 h following IS or ICH. RNA from whole blood was isolated as previously described [3]. The UCD Institutional Review Board approved this study and all subjects provided informed consent.

had no history of previous stroke or cardiovascular events. Blood from all subjects was drawn into PAXgene tubes between 5.8 and 101.2 h following IS or ICH. RNA from whole blood was isolated as previously described [3]. The UCD Institutional Review Board approved this study and all subjects provided informed consent. Whole blood RNA was used to prepare mRNA libraries using the TruSeq RNA Sample Prep v2 kit and protocol (Illumina). Two hundred million PE 100-bp RNA-seq reads were obtained from each mRNA library using Illumina Solexa sequencing by synthesis on the Illumina HiSeq 2000. TopHat v2.0.7 (Bowtie v2.0.6) was used with default parameters to map reads to a reference genome (Hg19) and generate bam files for analysis [10]. RNA transcript quantification was performed using Hg19 AceView transcripts in the Partek Genomics Suite 6.6 RNA-seq workflow.

encing by synthesis on the Illumina HiSeq 2000. TopHat v2.0.7 (Bowtie v2.0.6) was used with default parameters to map reads to a reference genome (Hg19) and generate bam files for analysis [10]. RNA transcript quantification was performed using Hg19 AceView transcripts in the Partek Genomics Suite 6.6 RNA-seq workflow. The raw reads for genes displaying DAS are shown in Supplementary Table 2 and the raw reads for genes displaying differential exon usage are shown in Supplementary Table 6. They were generated from aligned bam files using featureCounts against AceView (NCBI 37) [11] with options allowing for any and multiple overlaps [12]. However, they were not used directly for the statistical analysis. Instead, raw aligned reads were normalized, and differential alternatively spliced transcript expression and exon expression quantification were performed using the expectation/maximization (E/M) algorithm (briefly described below) as implemented in Partek Genomics Suite [13]. DAS was determined with one-way ANOVA on Group (Benjamini-Hochberg false discovery rate, FDR; p < 0.05), and differential exon usage was assessed between each two groups (p < 0.0005, fold change > |1.2|). Principal components analysis (PCA) and hierarchical clustering were performed in Partek Genomics Suite. Ingenuity Pathway Analysis (IPA®) and DAVID identified regulated pathways and processes as described previously [6].

The raw reads for genes displaying DAS are shown in Supplementary Table 2 and the raw reads for genes displaying differential exon usage are shown in Supplementary Table 6. They were generated from aligned bam files using featureCounts against AceView (NCBI 37) [11] with options allowing for any and multiple overlaps [12]. However, they were not used directly for the statistical analysis. Instead, raw aligned reads were normalized, and differential alternatively spliced transcript expression and exon expression quantification were performed using the expectation/maximization (E/M) algorithm (briefly described below) as implemented in Partek Genomics Suite [13]. DAS was determined with one-way ANOVA on Group (Benjamini-Hochberg false discovery rate, FDR; p < 0.05), and differential exon usage was assessed between each two groups (p < 0.0005, fold change > |1.2|). Principal components analysis (PCA) and hierarchical clustering were performed in Partek Genomics Suite. Ingenuity Pathway Analysis (IPA®) and DAVID identified regulated pathways and processes as described previously [6]. Results Subject Demographics Subject demographics and clinical characteristics are presented in Table 1. Only Caucasian males were studied because of the small group sizes. Age, time since event for IS or ICH, and vascular risk factors were not significantly different between groups. Coverage of a wide range of post-stroke biology was obtained by selecting patients with early (5.8 h) through late (101.2 h) blood draw times after IS and ICH. However, the means were similar between the stroke groups. Cardioembolic IS post-event blood draw times were, on average, 33.7 h; large vessel averaged 47.4 h; lacunar averaged 34.6 h; and ICH averaged 29.4 h (Table 1).Table 1 Subject demographics and clinical characteristics

101.2 h) blood draw times after IS and ICH. However, the means were similar between the stroke groups. Cardioembolic IS post-event blood draw times were, on average, 33.7 h; large vessel averaged 47.4 h; lacunar averaged 34.6 h; and ICH averaged 29.4 h (Table 1).Table 1 Subject demographics and clinical characteristics Ischemic stroke Intracerebral hemorrhage Controls Cardioembolic Large vessel Lacunar Subjects (total n = 20) 4 4 4 4 4 Age, years (mean ± SD) 62.3 ± 9.6 61.0 ± 8.2 58.9 ± 9.0 60.1 ± 2.3 60.8 ± 9.2 Time since event, h (mean ± SD) 33.7 ± 18.9 47.4 ± 47.8 34.6 ± 23.7 29.4 ± 15.5 N/A Hypertension 4 3 2 3 3 Diabetes 2 2 0 0 1 Hyperlipidemia 3 2 2 0 2 RNA Sequencing Alignments RNA sequencing alignment statistics for all samples among the five groups are presented in Supplementary Table 1. Cardioembolic stroke samples had on average, 1.60E+08 alignments; large vessel had 1.65E+08 alignments; lacunar stroke had 1.64E+08 alignments; and ICH and control groups each averaged 1.59E+08 alignments. These data show that there is no bias in the numbers of alignments for any of the five groups.

plementary Table 1. Cardioembolic stroke samples had on average, 1.60E+08 alignments; large vessel had 1.65E+08 alignments; lacunar stroke had 1.64E+08 alignments; and ICH and control groups each averaged 1.59E+08 alignments. These data show that there is no bias in the numbers of alignments for any of the five groups. Distinct Alternative Splicing of Genes in Whole Blood of Stroke Patients and Controls A total of 412 genes displayed differential alternative splicing (DAS) in the whole blood transcriptomes of the five groups of patients with ischemic stroke (cardioembolic, large vessel, and lacunar), intracerebral hemorrhage (ICH) and controls (FDR p < 0.05; Supplementary Table 2, raw reads; Supplementary Table 3, ANOVA results). These 412 genes are those predicted using the E/M algorithm [13] as implemented in Partek, to have DAS for IS (cardioembolic, large vessel, lacunar) versus ICH versus controls. The E/M algorithm probabilistically assigns reads to known isoforms/exons of a gene [13]. Partek then uses a log-likelihood ratio test to identify genes with DAS across samples [13, 14]. The 412 significant genes displaying DAS across IS, ICH, and controls are involved in cellular immunity, cytokine signaling, and cell death and survival pathways (Supplementary Table 4).

known isoforms/exons of a gene [13]. Partek then uses a log-likelihood ratio test to identify genes with DAS across samples [13, 14]. The 412 significant genes displaying DAS across IS, ICH, and controls are involved in cellular immunity, cytokine signaling, and cell death and survival pathways (Supplementary Table 4). Pathways highly over-represented with differentially alternative spliced genes between the five groups included CD28 signaling in T helper cells, CDC42 signaling, Nur77 signaling in T lymphocytes, fMLP signaling in neutrophils, and interferon signaling (Supplementary Table 4). Molecular and cellular functions most highly associated with the differentially alternatively spliced genes were cell death and survival of immune cells, cell-cell signaling, activation and recruitment of leukocytes, antigen-presenting cells, activation of T lymphocytes, adhesion of vascular endothelial cells, and immune response of neutrophils (Supplementary Table 5).

ated with the differentially alternatively spliced genes were cell death and survival of immune cells, cell-cell signaling, activation and recruitment of leukocytes, antigen-presenting cells, activation of T lymphocytes, adhesion of vascular endothelial cells, and immune response of neutrophils (Supplementary Table 5). Specific Exon-Usage Profiles for ICH and Different Ischemic Stroke Etiologies A total of 308 exons from 292 genes were differentially expressed for the three causes of IS (cardioembolic, large vessel, lacunar), ICH, and controls (p < 0.0005, fold change >|1.2|; Supplementary Table 6, raw reads; Supplementary Table 7, ANOVA results). These exons separated the five groups, including the three causes of IS (cardioembolic, large vessel, lacunar), ICH, and controls on principal components analysis (PCA) plots (Fig. 1a) and using unsupervised hierarchical clustering (Fig. 1b). Given that the E/M algorithm uses counts of the numbers of reads on each exon [13, 14], the differential expression of exons across the five groups represents differential exon usage across the five groups. These results are relevant to DAS because DAS results from differential exon usage.Fig. 1 Principal components analysis (PCA) (Fig. 1a) and unsupervised hierarchical clustering (Fig. 1b) of the 308 exons (292 genes) with differential exon usage among intracerebral hemorrhage (n = 4), ischemic strokes (IS) (cardioembolic, large vessel, and lacunar) (n = 12), and control subjects (n = 4). In Fig. 1a, the expression of the 308 exons is compressed on to three axes in the PCA plot. The three principal components on the PCA plot account for 64.1 % of the variance. In Fig. 1b, exon expression is shown on the X-axis and subjects are shown on the Y-axis. Each row on the Y-axis represents a single individual, with five individuals per group. The dendrograms were removed from this figure. Red indicates increased expression. Green indicates decreased expression

64.1 % of the variance. In Fig. 1b, exon expression is shown on the X-axis and subjects are shown on the Y-axis. Each row on the Y-axis represents a single individual, with five individuals per group. The dendrograms were removed from this figure. Red indicates increased expression. Green indicates decreased expression Biological functions and networks represented by genes with differentially expressed exons in each group (Fig. 1b) are summarized in Supplementary Table 8. Cardioembolic stroke genes with differential exon usage were involved in ion binding/transport and cellular assembly/organization. Large-vessel stroke genes were associated with cell death, transcription, and chromatin remodeling. Lacunar stroke genes were associated with cellular compromise, cell cycle, cell death and survival. ICH genes were involved with protein transport and localization (Supplementary Table 8).

cellular assembly/organization. Large-vessel stroke genes were associated with cell death, transcription, and chromatin remodeling. Lacunar stroke genes were associated with cellular compromise, cell cycle, cell death and survival. ICH genes were involved with protein transport and localization (Supplementary Table 8). Discussion Although differential alternative splicing (DAS) is implicated in many human diseases, this is the first study to show that DAS differs between intracerebral hemorrhage (ICH), ischemic stroke, and control subjects. In addition, it is the first study to show that DAS differs between different etiologies of ischemic stroke including cardioembolic, large vessel, and lacunar causes. Identification of DAS in RNA from whole blood for specific stroke etiologies and ICH suggests the immune response varies for each condition. This will be important for understanding the pathogenesis of each condition and will be important for developing biomarkers to differentiate ischemic stroke from ICH and for developing biomarkers to differentiate the different causes of ischemic stroke.

es and ICH suggests the immune response varies for each condition. This will be important for understanding the pathogenesis of each condition and will be important for developing biomarkers to differentiate ischemic stroke from ICH and for developing biomarkers to differentiate the different causes of ischemic stroke. This study identified several pathways, molecular functions, and genes previously reported in human ischemic stroke using 3′-biased microarrays [6, 15]. These included actin cytoskeleton signaling, CCR5 signaling in macrophages, NF-κB activation, α-adrenergic signaling, cellular growth and proliferation, cell death and survival, cell morphology, hematopoiesis, hematological system development, and inflammatory response [4, 5, 16, 17]. Moreover, a number of the pathways implicated in different etiologies of ischemic stroke in our previous microarray studies were confirmed in these RNA-seq studies [4, 5, 16, 17]. This study is the first to describe genes with DAS and pathways unique for ICH. Among the genes that differentiated ICH from IS were INPP5D (inositol polyphosphate-5-phosphatase) and ITA4 (integrin alpha 4). The INPP5D enzyme regulates myeloid cell proliferation and programming, and its expression correlates with hemorrhagic transformation of ischemic stroke [18]. ITA4 is involved in leukocyte recruitment after intracerebral hemorrhage [19], and leukocytes are intimately associated with ICH. For example, leukocytes are involved in clotting and interact with injured vessels and brain following ICH [15].

xpression correlates with hemorrhagic transformation of ischemic stroke [18]. ITA4 is involved in leukocyte recruitment after intracerebral hemorrhage [19], and leukocytes are intimately associated with ICH. For example, leukocytes are involved in clotting and interact with injured vessels and brain following ICH [15]. Other genes with DAS associated with ICH in this study included NAV1 (neuron navigator 1), PDGFC (platelet derived growth factor C), and CCM2 (cerebral cavernous malformation 2) which participate in vascular endothelial growth factor (VEGF) signaling, which predisposes the brain to hemorrhage because of new vessel formation [20]. Of interest, mutations of CCM2 cause cerebral cavernous malformations which can lead to intracerebral hemorrhage [21]. Other genes with DAS associated with ICH included EXOSC1 (exosome component 1) and EXOSC9 (exosome component 9) which code for core components of the exosome complex [22]. Although exosomes have been implicated in neuroinflammation, neurodegeneration, and cancer, they have not previously been associated with ICH [23, 24]. Lastly, another gene with DAS associated with ICH included DGCR8 (DiGeorge syndrome critical region 8, a microprocessor complex subunit) which is involved in the biogenesis of microRNAs [25], which could suggest that miRNAs are involved with differential alternative splicing following ICH.

associated with ICH [23, 24]. Lastly, another gene with DAS associated with ICH included DGCR8 (DiGeorge syndrome critical region 8, a microprocessor complex subunit) which is involved in the biogenesis of microRNAs [25], which could suggest that miRNAs are involved with differential alternative splicing following ICH. Study Limitations Sample sizes in this pilot study were small. Thus, we cannot rule out splicing changes due to vascular risk factors. However, hierarchical clustering of the differentially expressed exons demonstrates separation on diagnosis and not on vascular risk factors (Supplementary Figure 2). Validation of these findings in a separate cohort is needed to confirm the present results. These results are important because they provide evidence for differential alternative splicing in the pathophysiology of the immune response to ischemic stroke and intracerebral hemorrhage and also might provide novel biomarkers for ICH and different causes of IS. Electronic Supplementary Material Supplementary Figure 1 Schematic of Alternative Splicing. The 5′ and 3′ untranslated regions in mRNA are not depicted. In this example the primary mRNA transcript is transcribed into three different mRNA (mRNA1, mRNA2, mRNA3) which are translated into three different proteins (protein1, protein 2, protein 3) which are all derived from a single gene. (GIF 42 kb) High Resolution (TIFF 317 kb)

Electronic Supplementary Material Supplementary Figure 1 Schematic of Alternative Splicing. The 5′ and 3′ untranslated regions in mRNA are not depicted. In this example the primary mRNA transcript is transcribed into three different mRNA (mRNA1, mRNA2, mRNA3) which are translated into three different proteins (protein1, protein 2, protein 3) which are all derived from a single gene. (GIF 42 kb) High Resolution (TIFF 317 kb) Supplementary Figure 2 Unsupervised Hierarchical Clustering of 308 exons (292 genes) with differential exon usage among Intracerebral Hemorrhage (n = 4), Ischemic Strokes (Cardioembolic, Large Vessel, and Lacunar) (n = 12) and Control Subjects (n = 4). Dendrograms are not displayed. This is similar to Fig. 1, with the addition of information on age, time since event, diabetes, hypertension and hyperlipidemia. Exon expression is on the X-axis. Subjects are on the Y axis. Red indicates increased expression and green indicates decreased expression. (GIF 145 kb) High Resolution (TIFF 2305 kb) Supplementary Table 1 Over all details of the RNA sequencing reads and quality. (PDF 36 kb) Supplementary Table 2 Raw sequencing count for the 412 genes displaying Differential Alternative Splicing (DAS). IS, Ischemic Stroke; CE, Cardioembolic IS; LV, Large Vessel IS; ICH, Intracerebral Hemorrhage. (PDF 60 kb) Supplementary Table 3 The 412 genes with DAS among Large Vessel Ischemic Stroke (IS), Cardioembolic IS, Lacunar IS and ICH and Controls (ANOVA, FDR corrected p < 0.05). (PDF 72 kb)

Supplementary Table 2 Raw sequencing count for the 412 genes displaying Differential Alternative Splicing (DAS). IS, Ischemic Stroke; CE, Cardioembolic IS; LV, Large Vessel IS; ICH, Intracerebral Hemorrhage. (PDF 60 kb) Supplementary Table 3 The 412 genes with DAS among Large Vessel Ischemic Stroke (IS), Cardioembolic IS, Lacunar IS and ICH and Controls (ANOVA, FDR corrected p < 0.05). (PDF 72 kb) Supplementary Table 4 Canonical pathways of the 412 genes with DAS among Large Vessel Ischemic Stroke (IS), Cardioembolic IS, Lacunar IS, Intracerebral Hemorrhage (ICH) and Control subjects. Benjamini-Hochberg corrected P values for multiple comparison corrections. (PDF 51 kb) Supplementary Table 5 Functions associated with the top 2 molecular and cellular functions over-represented in the 412 differentially alternatively spliced genes. (PDF 113 kb) Supplementary Table 6 Raw sequencing count for the 308 exons displaying differential exon usage for Ischemic Stroke (Cardioembolic, Large Vessel, Lacunar), ICH and Controls. (PDF 53 kb) Supplementary Table 7 Differential exon usage for the 308 exons for Large Vessel IS, Cardioembolic IS, Lacunar IS, ICH and Controls (p < 0.0005, FC > |1.2|). (PDF 84 kb) Supplementary Table 8 Over-represented pathways and gene ontology for the 308 exons that showed differential exon usage between the five groups (Large Vessel IS, Cardioembolic IS, Lacunar IS, ICH, Controls). (PDF 65 kb)

Supplementary Table 7 Differential exon usage for the 308 exons for Large Vessel IS, Cardioembolic IS, Lacunar IS, ICH and Controls (p < 0.0005, FC > |1.2|). (PDF 84 kb) Supplementary Table 8 Over-represented pathways and gene ontology for the 308 exons that showed differential exon usage between the five groups (Large Vessel IS, Cardioembolic IS, Lacunar IS, ICH, Controls). (PDF 65 kb) We appreciate the patients and families who participated in the study, as well as the support of the MIND Institute and the UCD Department of Neurology. We thank Dr. Frank Sharp for his advice and support during the course of these studies. Compliance with Ethics Guidelines ᅟ Ethical Approval All procedures and protocols were performed in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and in accordance with the Helsinki Declaration of 1975, as revised in 2008 (5). This study was approved by the IRB at the University of California at Davis. Informed Consent Informed consent was obtained from all individuals included in the study. Funding This study was funded by NINDS, a division of the National Institutes of Health (RO1NS075035, RO1NS075035) and by a Fellow to Faculty Award from the American Heart Association (GCJ). Conflict of Interest All authors declare they have no competing interests.

Introduction Atherosclerosis is an inflammatory disease with both lifestyle and genetic risk factors. Receptors for the Fc portion of immunoglobulin G (IgG), known as FcγRs, are the binding link between humoral and cellular immunologic reactions. Endothelial cells express FcγRs and pro-inflammatory mediators, such as immune complexes and C-reactive protein, that can activate FcγR-dependent pathways that lead to oxidative burst, degranulation, phagocytosis, cytokine production and antibody-dependent cell-mediated cytotoxicity (ADCC) [1]. The FcγRs are surface glycoproteins, encoded by eight genes located on chromosome 1q21–23. Based on structural homology and differences in affinity for IgG, this family is divided into three subfamilies: FcγRI (CD64), FcγRII (CD32) and FcγRIII (CD16). FcγR isoforms can be activating receptors (FcRI, FcγRIIA and FcγRIII) or inhibitory receptors (FcγRIIB) [1]. Two studies have demonstrated associations between certain FcγRIIA polymorphisms and coronary artery disease (CAD) and ischemic stroke [2, 3]. However, a study of patients with myocardial infarction and patients with CAD, documented by angiography, did not confirm this association [4]. Protective effects have been reported for the FcγRIIA polymorphism and peripheral atherosclerosis [5], and an inverse relation has been found for FcγRIIIA and CAD [6].

[2, 3]. However, a study of patients with myocardial infarction and patients with CAD, documented by angiography, did not confirm this association [4]. Protective effects have been reported for the FcγRIIA polymorphism and peripheral atherosclerosis [5], and an inverse relation has been found for FcγRIIIA and CAD [6]. Interleukin (IL)-10 is another important immune regulator that may be important in the pathogenesis of atherosclerosis [7]. Low levels of expression of IL-10 are associated with atherosclerosis and cardiac and vascular dysfunction in mouse models [8, 9], and in humans, reduced levels of IL-10 are associated with carotid atherosclerosis and coronary disease [10, 11]. IL-10 production varies due to single-nucleotide polymorphisms (SNPs) in the promoter region of the IL-10 gene [12]. IL-10 SNPs have been identified at positions −1082 (A/G), −819 (C/T) and −592 (A/C) in Caucasians and form the haplotypes ATA, ACC and GCC in six possible combinations [13]. The genotype GCC/GCC is associated with high production of IL-10; GCC/ACC and GCC/ATA with medium production; and ATA/ATA, ATA/ACC and ACC/ACC with low production of IL-10 [14]. Although the IL-10 gene is also localised on chromosome 1, the polymorphisms of IL-10 and FcγR are not related. However, it has been shown that cross-linking of FcγRs on macrophages influence the IL-10 production [15].

d GCC/ATA with medium production; and ATA/ATA, ATA/ACC and ACC/ACC with low production of IL-10 [14]. Although the IL-10 gene is also localised on chromosome 1, the polymorphisms of IL-10 and FcγR are not related. However, it has been shown that cross-linking of FcγRs on macrophages influence the IL-10 production [15]. We have previously reported data from 232 patients who had suffered acute arterial ischemic stroke at ages between 15 and 49 years and found a 10-fold higher mortality rate and a 5-fold increased vascular morbidity rate among long-term survivors compared with controls [16]. After a mean observation time of 12 years from the index stroke, individual maximum carotid intima-media thickness (cIMT) values were ≥1.0 mm in 76 % of 140 examined patients, based on 5944 wall-segmental measurements [17]. In the present study, we analysed these retrospectively selected young stroke patients with the following biomarkers: C-reactive protein (CRP), homocysteine, cholesterol and triglycerides, sedimentation rate (SR), haemoglobin (Hb), glycolysed haemoglobin (HbA1c), creatinine, leukocytes, thrombocytes, and FcγR polymorphisms, IL-10 polymorphisms and Chlamydia pneumoniae antibodies. For multivariate analysis, we related these parameters and other relevant clinical information, such as age, gender, blood pressure, number of smoking years and additional arterial events other than the index stroke to cIMT.

thrombocytes, and FcγR polymorphisms, IL-10 polymorphisms and Chlamydia pneumoniae antibodies. For multivariate analysis, we related these parameters and other relevant clinical information, such as age, gender, blood pressure, number of smoking years and additional arterial events other than the index stroke to cIMT. Methods Patients and Study Design We identified 232 patients of less than 50 years of age with first-ever ischemic stroke between 1988 and 1997 from five acute care hospitals in Hordaland County in western Norway [16]. The mean age (standard deviation, SD) of the subjects was 41.1 (±7.5) years at the time of the index stroke. Ischemic stroke was defined according to the Baltimore Washington Cooperative Young Stroke Study Criteria as neurological deficits lasting longer than 24 h or clinical transient ischemic attacks (TIAs) in which cerebral computed tomography (CT) or magnetic resonance imaging (MRI) showed acute arterial cerebral infarction related to clinical findings [18]. About two thirds of the patients had anterior circulation infarction, about one third had posterior circulation infarction, and despite more serious affection of patients with anterior circulation infarction, the short-term outcome at discharge was favourable with modified Rankin score ≤2 in 80 % [19]. From 1988 to 1997, patients were mainly treated by platelet inhibition or anticoagulation. Thrombolysis and treatment with statins were not usual at that time. Figure 1 shows the flow chart of the study. The first follow-up was done from 1998 to 2001 after a mean time of 5.7 years from the index stroke. Blood samples were taken from 198 participating patients and frozen at −80 °C. Four patients refused blood sampling. A second follow-up was done from 2004 to 2005 after a mean time of 11.9 years from the index stroke. Until the second follow-up, 45 patients had died, mainly from cardiovascular disease [20]. Of the 187 survivors, 144 (77 %) participated in clinical examinations, and 140 patients were examined with cIMT measurements by B-mode ultrasound [17]. Two patients refused blood sampling but permitted the use of samples taken for other control purposes within 1 month. The mean age at this follow-up was 52.9 years (SD ± 8.1, range 28–65 years).Fig.

) participated in clinical examinations, and 140 patients were examined with cIMT measurements by B-mode ultrasound [17]. Two patients refused blood sampling but permitted the use of samples taken for other control purposes within 1 month. The mean age at this follow-up was 52.9 years (SD ± 8.1, range 28–65 years).Fig. 1 Flow chart of 232 retrospectively included ischemic stroke patients after the index stroke at age 17 to 49 years; n = number of patients who were identified and participated in follow-ups, nd = numbers of patients who died between inclusion and first follow-up and between first and second follow-up, nbs = number of collected blood samples. Four patients at first follow-up and two patients at second follow-up refused blood sampling, but both patients from second follow-up permitted the use of blood samples from other health controls, done within 1 month; nu = number of patients with carotid ultrasound measurements FcγR Polymorphisms FcγR polymorphisms were analysed as previously described [6]. In brief, leukocytes were isolated from peripheral venous blood samples (EDTA blood). DNA was extracted with QIAampTM Blood Kit (Qiagen GmbH, Hilden, Germany). FcγR genotypes were determined by polymerase chain reaction (PCR), and all PCR products were analysed by electrophoresis [6].

lysed as previously described [6]. In brief, leukocytes were isolated from peripheral venous blood samples (EDTA blood). DNA was extracted with QIAampTM Blood Kit (Qiagen GmbH, Hilden, Germany). FcγR genotypes were determined by polymerase chain reaction (PCR), and all PCR products were analysed by electrophoresis [6]. IL-10 Polymorphisms DNA was extracted from whole-blood samples and examined for IL-10 polymorphisms at positions −1082 (G/A), −819 (T/C) and −592 (A/C) using PCR and gene sequencing as previously described [21]. The IL-10 haplotypes were determined and grouped into combinations with high IL-10 expression, medium IL-10 expression and low IL-10 expression.

d from whole-blood samples and examined for IL-10 polymorphisms at positions −1082 (G/A), −819 (T/C) and −592 (A/C) using PCR and gene sequencing as previously described [21]. The IL-10 haplotypes were determined and grouped into combinations with high IL-10 expression, medium IL-10 expression and low IL-10 expression. Carotid Artery Imaging Right- and left-sided cIMT examination of near and far walls of the distal common carotid artery (CCA), bifurcation (BIF) and the proximal internal carotid artery (ICA) were performed at four predefined angles. CCA, BIF and ICA vessel segments were identified by using the tip of the flow divider as an internal landmark. Segments had a length of 10 mm, and the mean value was measured. A total of 5944 segmental mean IMT or plaque values were obtained. The maximum value of 48 possible individual cIMT measurements was used as outcome variable. cIMT results were used to divide subjects into two cIMT categories: <1.0 and ≥1.0 mm. Our 140 patients had a mean age of 53 years, the oldest patients were 65 years old at that final clinical follow-up, and total maximum IMT of 1.0 mm was chosen as measure of dichotomisation due to literature [22, 23]. Methods, results and patients’ characteristics were previously described in detail [17].

and ≥1.0 mm. Our 140 patients had a mean age of 53 years, the oldest patients were 65 years old at that final clinical follow-up, and total maximum IMT of 1.0 mm was chosen as measure of dichotomisation due to literature [22, 23]. Methods, results and patients’ characteristics were previously described in detail [17]. Data Analysis Biomarkers were compared between the groups with cIMT <1 mm and cIMT ≥1 mm. Age, additional arterial events, systolic and diastolic BP, number of smoking years, SR, HbA1c, creatinine, Hb, leukocytes, thrombocytes and triglycerides were used as results from the second follow-up. CRP, homocysteine, total cholesterol, LDL cholesterol and HDL cholesterol were used as results from the first and second follow-up. Categorical variables were analysed with 3×2 chi-square test and Fisher’s exact test, and Wilcoxon–Mann–Whitney test was used for continuous variables. Unadjusted logistic regression analysis of the dichotomised cIMT was used to evaluate the potential relationships with FcγR polymorphisms, IL-10 haplotypes with ATA/ATA as reference variable and with C. pneumoniae antibodies. p values were calculated from the likelihood ratio test. Multiple logistic regression analysis was done using the dichotomised cIMT results as outcome in a backward stepwise selection process retaining variables significant at the 5 % level in the final model. Multiple linear regression analysis was done using continuous maximum cIMT results as outcome in a backward stepwise selection process retaining variables significant at the 5 % level in the final model. The analysis was performed using the Statistical Package for Social Sciences (SPSS) 20. p values <0.05 were considered significant.

regression analysis was done using continuous maximum cIMT results as outcome in a backward stepwise selection process retaining variables significant at the 5 % level in the final model. The analysis was performed using the Statistical Package for Social Sciences (SPSS) 20. p values <0.05 were considered significant. Results Immunological Parameters and cIMT Table 1 shows FcγRIIIB polymorphisms associated with cIMT ≥1 mm. Table 2 shows total cholesterol, LDL cholesterol and homocysteine from the first follow-up associated with cIMT ≥1 mm, and further association with Hb, HbA1c and creatinine from the second follow-up. Age, gender, systolic and diastolic BP and additional arterial events were also associated with cIMT ≥1 mm (Table 2). We found a significant association between the FcγRIIIB NAII/NAII polymorphism and cIMT ≥1 mm by using unadjusted logistic regression and adjusted backward stepwise logistic regression analysis (Table 3). However, linear regression analysis showed non-significant results for the FcγRIIIB NAII/NAII polymorphism relation with continuous cIMT. There was also a significant association between the IL-10 GCC/ACC polymorphisms and cIMT ≥1 mm, but this was not correlated with other polymorphisms associated with the degree of IL-10 production (Table 1). Therefore, we did not follow up the IL-10 results by performing final backward stepwise logistic and linear regression analysis. C. pneumoniae antibodies, sedimentation rate and C-reactive protein were not associated with cIMT ≥1 mm.Table 1 Fc gamma receptor (FcγR) polymorphisms and interleukin (IL)-10 polymorphisms related to carotid intima-media thickness (cIMT) with ≥1 mm as outcome variable

d stepwise logistic and linear regression analysis. C. pneumoniae antibodies, sedimentation rate and C-reactive protein were not associated with cIMT ≥1 mm.Table 1 Fc gamma receptor (FcγR) polymorphisms and interleukin (IL)-10 polymorphisms related to carotid intima-media thickness (cIMT) with ≥1 mm as outcome variable Immunology Variables Number % cIMT <1 mm n = 34 cIMT ≥1 mm n = 104 LR test p value Fc gamma receptor (FcγR) polymorphisms FcγRIIA n = 138 0.313 H/H 15.9 5 17 H/R 42.0 18 40 R/R 42.0 11 47 FcγRIIIB n = 138 0.036 NaI/NaI 12.0 6 10 NaI/NaII 43.5 19 41 NaII/NaII 44.9 9 53* FcγRIIIA n = 137 0.362 V/V 12.4 2 15 V/F 40.9 15 41 F/F 46.7 17 47 Interleukin (IL)-10 polymorphisms n = 138 0.102 ATA/ATA 7.2 4 6 ATA/ACC 9.4 5 8 ACC/ACC 7.2 1 9 GCC/ACC 26.8 4 33* GCC/ATA 23.2 10 22 GCC/GCC 26.1 10 26 IL-10 producing genotypes [14] Low IL-10 production: ATA/ATA, ATA/ACC and ACC/ACC 10 23 0.353 Medium IL-10 production: GCC/ACC and GCC/ATA 14 55 0.238 High IL-10 production: GCC/GCC 10 26 0.655 n number, LR likelihood ratio test from logistic regression *Significant p value, p ≤ 0.05 Table 2 Other clinical chemistry and clinical biomarkers related to carotid intima-media thickness (cIMT) ≥1 mm as outcome variable n of samples cIMT <1 mm mean value (±SD) cIMT ≥1 mm mean value (±SD) p value Data from first follow-up, 6 years after the index stroke

n number, LR likelihood ratio test from logistic regression *Significant p value, p ≤ 0.05 Table 2 Other clinical chemistry and clinical biomarkers related to carotid intima-media thickness (cIMT) ≥1 mm as outcome variable n of samples cIMT <1 mm mean value (±SD) cIMT ≥1 mm mean value (±SD) p value Data from first follow-up, 6 years after the index stroke Chlamydia pneumoniae antibodies 140 3 8 0.728 Homocysteine (μM) 139 12.8 (16.7) 11.4* (7.7) 0.011 CRP (mg/L) 134 2.1 (3.1) 2.5 (6.5) 0.344 Total cholesterol (mM) 139 5.5 (1.1) 6.1*(1.1) 0.007 LDL cholesterol (mM) 139 3.48 (0.9) 3.90* (1.0) 0.030 HDL cholesterol (mM) 139 1.33 (0.4) 1.29 (0.4) 0.778 Data from second follow-up, 12 years after the index stroke CRP (mg/L) 139 2.6 (3.0) 3.7 (5.7) 0.344 Sedimentation rate (mm) 135 8.8 (6.3) 12.9 (10.1) 0.055 Homocysteine (μM) 132 11.8 (10.4) 12.4 (7.5) 0.098 Total cholesterol (mM) 139 5.37 (1.0) 5.47 (1.2) 0.590 LDL-cholesterol (mM) 138 3.41 (0.9) 3.54 (1.1) 0.543 HDL-cholesterol (mM) 140 1.64 (0.5) 1.51 (0.4) 0.331 Triglycerides (mM) 137 1.43 (0.6) 1.75 (0.8) 0.065 Haemoglobin (g/dL) 139 14.0 (1.1) 14.6* (1.3) 0.024 Glycolysed haemoglobin (%) 135 5.4 (0.7) 5.8* (0.9) <0.001 Creatinine (μM) 138 64.3 (13.6) 83.1* (70.7) <0.001 Leukocytes (109/L) 140 7.3 (2.2) 7.2 (1.9) 0.948 Thrombocytes (109/L) 138 268 (55.4) 262 (74.6) 0.221 Other variables at second follow-up Age (years) 140 45.6 (9.9) 55.2* (5.8) <0.001 Gender * 0.001 Males (N) 10 68 Females (N) 24 38 Blood pressure (mmHg) 137 Systolic 131 (18.4) 143* (16.8) 0.001 Diastolic 84 (11.1) 90* (9.9) 0.002

109/L) 140 7.3 (2.2) 7.2 (1.9) 0.948 Thrombocytes (109/L) 138 268 (55.4) 262 (74.6) 0.221 Other variables at second follow-up Age (years) 140 45.6 (9.9) 55.2* (5.8) <0.001 Gender * 0.001 Males (N) 10 68 Females (N) 24 38 Blood pressure (mmHg) 137 Systolic 131 (18.4) 143* (16.8) 0.001 Diastolic 84 (11.1) 90* (9.9) 0.002 n of smoking years 53 active and 54 ex-smokers 22.3 (10.0) 26.9 (10.2) 0.078 AAE 6 46* 0.008 3×2 chi-square test and Fisher’s exact test for categorical variables and Wilcoxon–Mann–Whitney test for continuous variables were used for comparison of data. Mean value (±SD) n number, SD standard deviation, μM micromole, CRP C-reactive protein, LDL low density lipoprotein, HDL high density lipoprotein, AAE additional arterial event other than the index stroke (recurrent ischemic stroke, angina pectoris, myocardial infarction, peripheral artery disease) Table 3 Significant results of unadjusted logistic regression and adjusted backward stepwise logistic regression analysis, related to maximum carotid intima-media thickness, categorised as <1 and ≥1 mm in 140 patients, measured 12 years after acute ischemic stroke

n number, SD standard deviation, μM micromole, CRP C-reactive protein, LDL low density lipoprotein, HDL high density lipoprotein, AAE additional arterial event other than the index stroke (recurrent ischemic stroke, angina pectoris, myocardial infarction, peripheral artery disease) Table 3 Significant results of unadjusted logistic regression and adjusted backward stepwise logistic regression analysis, related to maximum carotid intima-media thickness, categorised as <1 and ≥1 mm in 140 patients, measured 12 years after acute ischemic stroke Unadjusted logistic regression Adjusted logistic regression Variables OR (95 % CI) p values OR (95 % CI) p values FcγRIIIB NaI/NaI 1.00 FcγRIIIB NaII/NaII 2.78 1.18, 6.52 0.019 3.94 1.08, 14.3 0.037 Cholesterol 1.62 1.12, 2.34 0.010 – – – SR 1.06 1.01, 1.12 0.031 – – – Hb 1.41 1.02, 1.95 0.036 – – – HbA1c 3.35 1.27, 8.86 0.015 6.65 1.21, 36.5 0.029 Triglycerides 1.75 1.00, 3.07 0.049 – – – Creatinine 1.05 1.02, 1.08 0.001 – – – Age 1.17 1.10, 1.25 <0.001 1.13 1.04, 1.23 0.003 Male gender 4.30 1.86, 9.93 0.001 4.07 1.15, 14.5 0.030 Systolic BP 1.05 1.02, 1.08 0.002 – – – Diastolic BP 1.06 1.01, 1.10 0.008 – – – AAE 3.64 1.39, 9.53 0.009 – – – Unadjusted univariate logistic regression analysis was performed for all blood samples from the first and second follow-up: FcγR polymorphisms, IL-10 polymorphisms, Chlamydia pneumoniae antibodies, homocysteine, C-reactive protein, total cholesterol, LDL cholesterol, HDL cholesterol, SR, triglycerides, Hb, HbA1c, creatinine, leukocytes and thrombocytes and for clinical variables, such as age, gender, systolic BP, diastolic BP and AAE. Only significant found variables are listed and used in the adjusted backward stepwise logistic regression analysis

tal cholesterol, LDL cholesterol, HDL cholesterol, SR, triglycerides, Hb, HbA1c, creatinine, leukocytes and thrombocytes and for clinical variables, such as age, gender, systolic BP, diastolic BP and AAE. Only significant found variables are listed and used in the adjusted backward stepwise logistic regression analysis OR odds ratio, CI confidence interval, FcγR Fc gamma receptor polymorphism, SR sedimentation rate, Hb haemoglobin, HbA1c glycolysed haemoglobin, BP blood pressure, AAE additional arterial events (recurrent stroke and/or angina pectoris and/or myocardial infarction and/or peripheral artery disease) other than the index stroke Non-immunological Parameters and cIMT cIMT ≥1 mm was associated with total cholesterol, LDL cholesterol and homocysteine at the first, but not at the second, follow-up (Table 2). Haemoglobin, HbA1c and creatinine were also associated with cIMT ≥1 mm as were age, male gender, systolic and diastolic BP and additional arterial events other than the index stroke. HbA1c was associated with cIMT ≥1 mm in addition to age and male gender after unadjusted logistic regression and results after adjusted backward stepwise logistic regression analysis (Table 3). Final linear regression analysis for continuous cIMT showed associations with age, additional arterial events, systolic BP and HbA1c (Table 4). Male gender was no longer associated with cIMT in that analysis.Table 4 Final results of linear regression analysis, related to maximum carotid intima-media thickness (cIMT), measured in 140 patients 12 years after acute ischemic stroke

associations with age, additional arterial events, systolic BP and HbA1c (Table 4). Male gender was no longer associated with cIMT in that analysis.Table 4 Final results of linear regression analysis, related to maximum carotid intima-media thickness (cIMT), measured in 140 patients 12 years after acute ischemic stroke b Standard error (SE) p values Age 0.25 0.007 0.002 Systolic BP 0.18 0.003 0.029 AAE 0.20 0.111 0.019 HbA1c 0.18 0.061 0.028 Linear regression analysis was performed with Fc gamma receptor IIIB NaII/NaII polymorphism, total cholesterol, sedimentation rate, haemoglobin, HbA1c, triglycerides, creatinine, age, gender, systolic BP and AAE. Gender was non-significant in the final results (b = 0.13, SE = 0.1, p = 0.80) b estimated regression coefficient, BP blood pressure, AAE additional arterial event other than the index stroke, such as recurrent ischemic stroke, angina pectoris, myocardial infarction and peripheral artery disease, HbA1c glycolysed haemoglobin Discussion This is the first clinical study of long-term survivors of ischemic stroke at a young age to evaluate follow-up blood samples and carotid IMT measurements. The mean observation time was nearly 12 years, and the cIMT measurements also included plaques, as we measured at standardised sites in order to perform a staging of the carotid artery wall disease.

f long-term survivors of ischemic stroke at a young age to evaluate follow-up blood samples and carotid IMT measurements. The mean observation time was nearly 12 years, and the cIMT measurements also included plaques, as we measured at standardised sites in order to perform a staging of the carotid artery wall disease. We have studied immunological and non-immunological risk factors for cIMT as parameters for carotid atherosclerosis. Our results showed no association with C. pneumoniae antibodies and cIMT. The association between C. pneumoniae antibodies and cardiovascular disease is unclear [24]. Furthermore, we did not find any association with cIMT and sedimentation rate or C-reactive protein. However, these parameters are crude markers for the evaluation of specific immune responses in the vascular wall. We also studied the FcγR and IL10 genotypes as they have been linked to atherosclerosis. The frequencies of the FcγR and IL-10 genotypes among our patients were comparable to 272 healthy controls previously reported from western Norway [25, 26]. We found a significant association between FcγRIIIB NaII/NaII and cIMT ≥1 mm. However, linear regression analysis showed non-significant results for continuous maximum cIMT values. Others have reported that soluble FcγRIIIA is associated with maximum cIMT [27], that the FcγRIIIA VV-genotype is inversely related to the extent of coronary artery disease [6] and that the FcγRIIA R allele is associated with impaired endothelial vasodilation in patients with hypercholesterolemia [28].

mum cIMT values. Others have reported that soluble FcγRIIIA is associated with maximum cIMT [27], that the FcγRIIIA VV-genotype is inversely related to the extent of coronary artery disease [6] and that the FcγRIIA R allele is associated with impaired endothelial vasodilation in patients with hypercholesterolemia [28]. The present study is the first to show a possible association with the low-affinity FcγRIIIB receptor and atherosclerosis as measured by cIMT ≥1 mm. FcγRIIIB, as well as FcγRIIA and FcγRIIIA, activates immune functions such as phagocytosis, ADCC and release of inflammatory mediators and superoxide radicals [29]. FcγRIIIB is only present on granulocytes, and the FcγRIIIB-NA2 allotype shows lower levels of phagocytosis of IgG1 and IgG3 opsonised particles than the NA1 allotype [30]. Our results therefore indicate that reductions in effector functions such as phagocytosis and ADCC may contribute to the development of atherosclerosis. Such findings are supported by knockout studies of the inhibitory FcγRIIB which shows an increased plaque formation [31]. In addition to myeloid cells, FcγRIIB is expressed on B cells where it suppresses B cell signaling and regulates IgG production [32]. Defects of FcγRIIB on B cell will therefore increase the IgG production, which was showed in this mouse model [31]. Our results indicate that similar mechanisms may take place in humans, as the NA2 allotype is associated with reduced phagocytosis of IgG1 and IgG3 [30].

cell signaling and regulates IgG production [32]. Defects of FcγRIIB on B cell will therefore increase the IgG production, which was showed in this mouse model [31]. Our results indicate that similar mechanisms may take place in humans, as the NA2 allotype is associated with reduced phagocytosis of IgG1 and IgG3 [30]. Stimulation of activating FcγRs on macrophages has been found to induce an anti-inflammatory immune response with increased IL-10 production [15]. It has also been proposed that elevated IgG with IL-10 acting downstream is driving the plaque formation [31]. Others have reported that reduced levels of IL-10 and low-producing IL-10 genotypes are associated with increased cIMT [10, 33] and that low-producing IL-10 genotypes are associated with coronary disease [11] and ischemic stroke [34].

that elevated IgG with IL-10 acting downstream is driving the plaque formation [31]. Others have reported that reduced levels of IL-10 and low-producing IL-10 genotypes are associated with increased cIMT [10, 33] and that low-producing IL-10 genotypes are associated with coronary disease [11] and ischemic stroke [34]. We have also studied the polymorphisms of IL-10 since 50–70 % of the variance in IL-10 levels is genetically determined [35]. However, we found no association between high-, medium- or low-producing IL-10 genotypes and cIMT ≥1 mm. Unfortunately, we had no access to serum IL-10 levels in our patients. Although we could not find any genetic risk factor with IL-10 and cIMT, this does not exclude the importance of this anti-inflammatory cytokine in the pathogenesis of atherosclerosis. The major roles of IL-10 in terms of atherosclerosis include inhibition of macrophage activation, as well as inhibition of matrix metalloproteinase, pro-inflammatory cytokines and cyclooxygenase-2 expression in lipid-loaded and activated macrophage foam cells. Furthermore, another important role of IL-10 in atherosclerosis is its ability to alter lipid metabolism in macrophages [36]. Of the non-immunological parameters, we found that cholesterol, triglycerides, blood pressure and cardiovascular events were risk factors in the unadjusted logistic regression analysis but not in the adjusted logistic regression analysis. However, these are known risk factors for subclinical carotid atherosclerosis [37, 38] At the first follow-up, statins were only used by 30 (15.5 %) of 194 patients with registered medication, and at the second follow-up, statins were used by 55 (38.2 %) of 144 patients [16]. The increasing use of statins probably explains why total cholesterol/homocysteine/and LDL-cholesterol levels were associated with cIMT at the first follow-up but not at the second follow-up.

of 194 patients with registered medication, and at the second follow-up, statins were used by 55 (38.2 %) of 144 patients [16]. The increasing use of statins probably explains why total cholesterol/homocysteine/and LDL-cholesterol levels were associated with cIMT at the first follow-up but not at the second follow-up. We found that HbA1c, SR, haemoglobin and creatinine were risk factors in the unadjusted logistic regression analysis, but only HbA1c showed significance in the final adjusted logistic and linear regression analysis. Elevated blood glucose is a known risk factor for endothelial dysfunction and early atherosclerosis [39], and the high percentage of carotid artery atherosclerosis observed in our patients is in agreement with this. Furthermore, we found that increased age and male gender are correlated with increased cIMT as has been shown in large population studies [37, 40]. Surprisingly, male gender disappeared as significant result in the linear regression analysis, and other risk factors for increased cIMT and atherosclerosis, such as systolic BP and additional arterial events, became significant. Probably, the number of patients played a role for the varying results by the different methods of regression analysis.

appeared as significant result in the linear regression analysis, and other risk factors for increased cIMT and atherosclerosis, such as systolic BP and additional arterial events, became significant. Probably, the number of patients played a role for the varying results by the different methods of regression analysis. The strength of our study is that we have a very well characterised patient population and clinically related immunological and non-immunological parameters were available for analysis. The frequencies of FcγR and IL-10 polymorphisms in our patients are similar to those observed in population studies performed from the same area [25, 26]. We have measurements of the various blood samples both at first and at second follow-up. In addition, we did meticulous studies to measure cIMT and plaques. The weakness of our study is that patients were included retrospectively and selection was affected by death and a 23 % dropout rate of long-term survivors. The 23 % dropout rate at our last clinical follow-up, without any missing data concerning the dead-alive status, is, however, comparable to that of other clinical studies with follow-ups after 10 years [41].

ere included retrospectively and selection was affected by death and a 23 % dropout rate of long-term survivors. The 23 % dropout rate at our last clinical follow-up, without any missing data concerning the dead-alive status, is, however, comparable to that of other clinical studies with follow-ups after 10 years [41]. Differences occurred after logistic and linear regression analysis concerning significancies related to cIMT in this study. However, age and HbA1c were in both analyses related to cIMT, and these parameters are well-known risk factors for increased cIMT and atherosclerosis. Male gender and increased BP are also well-known risk factors for atherosclerosis and additional arterial events, even if the results by the two types of regression analysis were slightly different. A new finding in this study is that FcγRIIIB-NAII/NAII appeared as a significant factor for increased cIMT ≥1 mm. However, further prospective studies with higher sample sizes are necessary to confirm these results. In conclusion, we found that in patients with ischemic stroke at a young age, FcγRIIIB-NAII/NAII may be a contributing factor for increased cIMT together with known risk factors, such as age, male gender, systolic BP and increased HbA1c. We thank Terje Ertkjern and co-workers from the laboratory of clinical biochemical medicine for the analysis of routine blood samples.

In conclusion, we found that in patients with ischemic stroke at a young age, FcγRIIIB-NAII/NAII may be a contributing factor for increased cIMT together with known risk factors, such as age, male gender, systolic BP and increased HbA1c. We thank Terje Ertkjern and co-workers from the laboratory of clinical biochemical medicine for the analysis of routine blood samples. Conflict of interest Ulrike Waje-Andreassen, Halvor Naess, Lars Thomassen, Tove Helene Maroy, Kibret Yimer Mazengia, Geir Egil Eide and Christian Alexander Vedeler declare that they have no conflicts of interest concerning this study. Ethical approval The study procedures were in accordance with the Helsinki Declaration of 1975. The Western Norway Regional Committee for Medical Research Ethics approved the study (project ID: 82.04); The Norwegian Royal Department of Health approved the research biobank, connected to the project (ID number: 200403628). Informed consent was obtained from all patients for being included in the study. Investigators were not allowed to ask patients about reasons for non-participation during the first and second follow-up.

Collateral circulation refers to pre-existing vascular redundancies that provide a route for blood to reach a target tissue when a primary channel is blocked [1–5]. Primary cerebral arterial collaterals refer to short arterial segments in the circle of Willis that allow blood flow between the territories of the internal carotid arteries and the vertebrobasilar system or between cerebral hemispheres in the event of proximal occlusion. The secondary cerebral collaterals include the pial (or leptomeningeal) collaterals, which are anastomotic connections located on the surface of the cortex that connect distal branches of the anterior, middle, and posterior cerebral arteries (ACA, MCA, PCA). These collateral channels permit blood flow from the territory of an unobstructed artery into the territory of an occluded artery (e.g., retrograde filling of the MCA territory via anastomoses with the ACA after middle cerebral artery occlusion (MCAo)) [2–4, 6]. Collateral extent is crucial, as in both animal models and human stroke patients the degree of collateral perfusion during cerebral ischemia is a predictor of stroke severity, prognosis, and response to reperfusion therapy [7–15].

ory via anastomoses with the ACA after middle cerebral artery occlusion (MCAo)) [2–4, 6]. Collateral extent is crucial, as in both animal models and human stroke patients the degree of collateral perfusion during cerebral ischemia is a predictor of stroke severity, prognosis, and response to reperfusion therapy [7–15]. “Collateral therapeutics” augment blood flow through collaterals to improve perfusion of penumbral tissue during acute ischemic stroke. Recently, promising pre-clinical and clinical data support the use of transient aortic occlusion (TAO) to increase global cerebral perfusion and reduce damage due to stroke. Pre-clinical work in TAO-treated rats has demonstrated reduced infarct size [16] and augmentation of flow through pial collaterals due to TAO after thromboembolic MCAo [17]. Data from the Safety and Efficacy of NeuroFlo Technology in Ischemic Stroke (SENTIS) trial suggests that TAO is safe and improves stroke outcome in subsets of stroke patients [4, 16, 18–21]. Patients with cortical ischemic stroke presenting within 5 h of onset, with National Institutes of Health Stroke Scale (NIHSS) between 8 and 14 (moderate severity), and patients older than 70 years of age showed the greatest benefit from TAO [19]. Additional analysis of the SENTIS trial reports a significant reduction in stroke-related mortality and severe disability with TAO [22]. Notably, the mortality difference was concentrated in the patients at highest risk (NIHSS scores >14 and those older than age 70) [19, 20]. This reduction in stroke-related mortality in patients with severe stroke suggests that TAO may be beneficial for patients with proximal occlusions resistant to intravenous thrombolysis, such as internal carotid occlusions or proximal occlusion of the MCA [23, 24]. Here, we evaluated the efficacy of TAO to improve collateral blood flow and reduce mortality in a model of proximal MCA/ACA occlusion.

ts that TAO may be beneficial for patients with proximal occlusions resistant to intravenous thrombolysis, such as internal carotid occlusions or proximal occlusion of the MCA [23, 24]. Here, we evaluated the efficacy of TAO to improve collateral blood flow and reduce mortality in a model of proximal MCA/ACA occlusion. Methods Experimental protocols conform to the guidelines established by the Canadian Council on Animal Care and were approved by the Health Sciences Animal Care and Use Committee at University of Alberta. Urethane-anaesthetized (i.p., 1.5 g/kg in sterile saline) male Sprague Dawley rats (400–450 g, 3–4 months of age) divided into two treatment groups (MCAo + TAO, n = 10; MCAo + Sham TAO, n = 11) underwent laser speckle contrast imaging (LSCI) through a thinned-skull imaging window. LSCI maps of blood flow were acquired at baseline, post-MCAo, 15 min after cessation of TAO (or sham, 2 h post-MCAo), and 75 min after TAO (or sham, 3 h post-MCAo). Immediately after the final imaging session, animals were euthanized and their brains were removed. Early indices of infarct size and location were assessed on cryostat-sectioned 20-μm coronal brain sections stained with hematoxylin and eosin (H&E). Animals that did not survive through all imaging sessions (4, all from the sham-TAO group) or with poor image quality (1, due to degradation of the optical window) were excluded from LSCI analyses.

ze and location were assessed on cryostat-sectioned 20-μm coronal brain sections stained with hematoxylin and eosin (H&E). Animals that did not survive through all imaging sessions (4, all from the sham-TAO group) or with poor image quality (1, due to degradation of the optical window) were excluded from LSCI analyses. LSCI LSCI was performed via a ~5 × 5-mm thin skull cranial window over the distal regions of the MCA territory [6, 17, 25]. Back-scattered light was collected by a Dalsa 1M60 Pantera CCD camera (5 ms exposure time) during illumination with a 784-nm (32 mW) laser (StockerYale, Inc.). Speckle contrast (K = σs/I) was determined using ImageJ software (NIH) [6, 26]. Maps of K show the pattern of blood flow on the cortical surface during imaging, with darker vessels demonstrating relatively faster blood flow than lighter vessels. K values were converted to correlation times (τc) that are approximately inversely proportional to blood flow velocity [17, 26, 27]. Data are expressed as τBaseline/τc or τMCAo/τc values that illustrate changes in blood flow from baseline or post-MCAo values, respectively. Vessel diameters were determined using an ImageJ plug-in that uses a full width at half-maximum algorithm [17, 28].

ely inversely proportional to blood flow velocity [17, 26, 27]. Data are expressed as τBaseline/τc or τMCAo/τc values that illustrate changes in blood flow from baseline or post-MCAo values, respectively. Vessel diameters were determined using an ImageJ plug-in that uses a full width at half-maximum algorithm [17, 28]. Filament Occlusion To induce proximal occlusions of the MCA and ACA, the right common, internal, and external carotid arteries (CCA, ICA, ECA) were exposed. A silicon rubber-coated monofilament (Doccol Corporation, filament size 3-0, diameter 0.20 mm, length 30 mm; diameter with coating 0.54+/− 0.02 mm; coating length 5–6 mm) was advanced through the ECA into the ICA until lodging proximal to the origin of MCA and ACA and then sutured in place. Transient Aortic Occlusion To transiently occlude the descending aorta [16, 17], a dilation catheter (2.0 mm diameter, Cordis Fire Star RX PTCA) was advanced past the origin of the renal artery from the femoral artery. Aortic flow diversion was initiated 60 min after ischemic onset and maintained for 45 min.

Filament Occlusion To induce proximal occlusions of the MCA and ACA, the right common, internal, and external carotid arteries (CCA, ICA, ECA) were exposed. A silicon rubber-coated monofilament (Doccol Corporation, filament size 3-0, diameter 0.20 mm, length 30 mm; diameter with coating 0.54+/− 0.02 mm; coating length 5–6 mm) was advanced through the ECA into the ICA until lodging proximal to the origin of MCA and ACA and then sutured in place. Transient Aortic Occlusion To transiently occlude the descending aorta [16, 17], a dilation catheter (2.0 mm diameter, Cordis Fire Star RX PTCA) was advanced past the origin of the renal artery from the femoral artery. Aortic flow diversion was initiated 60 min after ischemic onset and maintained for 45 min. Results LSCI was performed to monitor the blood flow in cortical surface vessels during proximal MCA and ACA occlusion followed by TAO (n = 10) or Sham-TAO (n = 11). Consistent with previous studies, anastomoses between distal ACA and MCA segments were apparent after MCAo but not during baseline imaging (Fig. 1a, b, white arrows). Qualitative analysis of LSCI data showed clear decreases in blood flow by MCAo and increased flow that persists after TAO in some but not all TAO-treated rats (Fig. 1a, c; note an increase in speckle contrast after MCAo that reflects a reduction in flow and a darkening of the image as blood flow increases due to TAO). Arrows in Fig. 1c show regions of drastic increases in flow in MCA segments downstream of ACA-MCA anastomoses in a TAO-treated animal. In other animals (from both treatment groups), reductions in flow after MCAo were less drastic and distinct changes due to TAO or sham were not apparent (Fig. 1b). Notably, early mortality (between 1 and 3 h after ischemic onset, prior to completion of the final imaging session 3 h after ischemic onset) was significantly reduced in TAO-treated rats relative to Sham-TAO (Fig. 1d, χ2(2) = 4.49, P = 0.034). Analysis of blood flow (τBaseline/τc) in distal MCA segments downstream of ACA anastomoses after MCAo but prior to treatment (or sham) suggests that early mortality occurred in Sham-TAO rats with severe ischemia (Fig. 1e). Multivariate analyses of physiological parameters monitored via pulse oximetry (Table 1) did not reveal any differences in oxygen saturation, heart rate, or breath rate that could account for differences in early mortality.Fig. 1 a Laser speckle contrast imaging (LSCI) data showing changes in blood flow associated with proximal middle cerebral artery occlusion (MCAo) and treatment with transient aortic occlusion (TAO). Images show blood flow recorded at baseline (pre-MCAo), post-MCAo, 15 min after TAO, and 75 min after TAO. Anastomoses between the MCA and anterior cerebral artery (ACA) are apparent after MCAo.

od flow associated with proximal middle cerebral artery occlusion (MCAo) and treatment with transient aortic occlusion (TAO). Images show blood flow recorded at baseline (pre-MCAo), post-MCAo, 15 min after TAO, and 75 min after TAO. Anastomoses between the MCA and anterior cerebral artery (ACA) are apparent after MCAo. An increase in speckle contrast intensity post-MCAo reflects a decrease in blood flow (ischemia), while the darker images after TAO demonstrate increased blood flow after treatment. b Representative LSCI data from a sham-TAO animal showing a more moderate ischemia (less intense speckle contrast, but clear formation of anastomoses indicating MCAo) and little change in blood flow after sham. c Pseudo-colored LSCI maps of blood flow (with cooler colors reflecting greater blood flow, warmer colors showing lower flow) showing increases in flow through MCA segments downstream of ACA connections (see white arrows) after TAO (note increased diameter and darker colors of identified MCA segments). d Contigency table illustrating reduced mortality prior to imaging completion (within 3 h of MCAo onset) in animals treated with TAO as opposed to sham-TAO. e τ Baseline /τ c values reflecting changes in blood flow velocity relative to baseline show that TAO animals that survived through imaging had a trend towards lower blood flow than did sham-TAO animals that survived. Sham-TAO animals that died within 3 h of MCAo onset had more severe ischemia (reduced τ Baseline /τ c), suggesting that lower mean τ Baseline /τ c in TAO animals results from reduced mortality in severely ischemic animals. Scale bars, 1 mm

ng had a trend towards lower blood flow than did sham-TAO animals that survived. Sham-TAO animals that died within 3 h of MCAo onset had more severe ischemia (reduced τ Baseline /τ c), suggesting that lower mean τ Baseline /τ c in TAO animals results from reduced mortality in severely ischemic animals. Scale bars, 1 mm Table 1 Physiological parameters Oxygen saturation Heart rate Breath rate Pre-MCAo TAO Mean 98.83 366.75 100.43 S.E.M 0.38 10.69 5.92 Sham—survived Mean 99.09 386.76 97.55 S.E.M 0.34 20.66 6.80 Sham—died Mean 99.45 365.26 85.22 S.E.M 0.05 22.84 7.61 Post-MCAo TAO Mean 98.84 389.54 96.53 S.E.M 0.14 11.91 6.56 Sham—survived Mean 99.18 414.15 89.58 S.E.M 0.22 20.28 3.22 Sham—died Mean 99.56 394.81 86.03 S.E.M 0.05 22.84 7.61 Post-TAO or sham TAO Mean 95.98 404.40 96.99 S.E.M 0.93 13.15 5.68 Sham—survived Mean 98.20 381.01 94.99 S.E.M 1.08 31.18 8.85 Multivariate analyses of physiological parameters monitored via pulse oximetry did not reveal any differences in oxygen saturation, heart rate, or breath rate that could account for differences in early mortality

O Mean 95.98 404.40 96.99 S.E.M 0.93 13.15 5.68 Sham—survived Mean 98.20 381.01 94.99 S.E.M 1.08 31.18 8.85 Multivariate analyses of physiological parameters monitored via pulse oximetry did not reveal any differences in oxygen saturation, heart rate, or breath rate that could account for differences in early mortality Because early mortality resulted in group differences in post-MCAo blood flow, τc were normalized to post-MCAo values within each animal to examine treatment-induced changes in blood flow in distal MCA segments and in surface veins draining the ischemic territories (Fig. 2a, b). Multivariate ANOVA did not reveal a significant main effect of treatment on τMCAo/τc (MCA segments, F(1, 14) = 3.020, P = 0.1042; veins, F(1, 14) = 2.306, P = 0.1511). However, a clear trend towards augmented collateral flow in MCA segments was apparent in a subset of TAO-treated rats. Notably, drastic increases in flow after TAO did not occur in any Sham-TAO animals. TAO “responders” were therefore defined as animals in which τMCAo/τc was greater than the mean τMCAo/τc from sham animals plus two standard deviations [5/9 (56 %) of rats in the TAO group met this criteria; see Fig. 1c for representative speckle contrast images showing increased flow]. Vessel diameter (normalized to baseline diameter, measured 75 min after cessation of TAO or Sham-TAO) was not significantly different between treatment groups in MCA segments or surface veins (Fig. 2c, d) (unpaired t test, P > .05). However, single sample t tests (hypothetical baseline of 1.0) suggest that MCA segments were significantly dilated relative to baseline only in the TAO group (P = .032).Fig. 2 a, b Scatter plots showing τ MCAo /τ c values measured in distal MCA segments or surface veins, respectively, for individual rats after middle cerebral artery occlusion (MCAo) and transient aortic occlusion (TAO) or sham-TAO. The horizontal line denotes the mean. While multivariate ANOVA did not identify a significant main effect, significant increases in blood flow from post-MCAo values were observed in some (5/9) TAO-treated rats. No comparable changes in flow were observed in sham-TAO rats. c, d Diameter of distal MCA segments and surface veins, respectively, measured 75 min after cessation of TAO or Sham-TAO. # denotes a significant difference from a hypothetical (normalized) baseline value of 1.0 on a single sample t test. e, f A variance-stabilizing transformation was achieved by performing a log transform of τ MCAo /τ c values.

segments and surface veins, respectively, measured 75 min after cessation of TAO or Sham-TAO. # denotes a significant difference from a hypothetical (normalized) baseline value of 1.0 on a single sample t test. e, f A variance-stabilizing transformation was achieved by performing a log transform of τ MCAo /τ c values. e shows a clear separation among rats that responded to TAO with an increase in τ MCAo /τ c in distal MCA segments (+, denoted by dashed oval) and rats that did not respond to TAO. f Multivariate ANOVA detected a significant effect of group on log(τ MCAo /τ c) when rats were categorized as TAO “responding” or “non-responding” (i.e., rats that exhibited a clear increase in MCA flow after TAO vs. rats with no significant change in MCA flow) or sham-TAO. **** denotes significant (P < .0001) Holm-Sidak post hoc comparisons. g, h τ Baseline /τ c values from TAO responders and non-responders, respectively, showing that TAO responders had severe ischemia prior to treatment that was normalized near baseline values by TAO (whereas TAO non-responders had less severe ischemia and no change due to treatment). i Representative H&E-stained tissue sections showing demarcation of early infarct from healthy tissue (black arrows, distances from Bregma shown below images). j Pooling τ Baseline /τ c values across treatment groups, a clear inverse relationship between amount of blood flow (relative to baseline) in distal MCA segments (measured 75 min after TAO or sham-TAO, 2 h after MCAo onset) and the volume of early infarct detected 3 h after MCAo onset

s from Bregma shown below images). j Pooling τ Baseline /τ c values across treatment groups, a clear inverse relationship between amount of blood flow (relative to baseline) in distal MCA segments (measured 75 min after TAO or sham-TAO, 2 h after MCAo onset) and the volume of early infarct detected 3 h after MCAo onset TAO Responders Versus Non-responders A variance-stabilizing log transformation of MCA flow ([log(τMCAo/τc)], Fig. 2a) was performed to better visualize response patterns. Figure 2e shows log(τMCAo/τc) data from MCA segments in TAO-treated rats, indicating a clear separation of rats who responded to TAO with an increase in collateral flow (+, denoted by dashed line) or did not respond to treatment (−). Multivariate ANOVA of log(τMCAo/τc) for TAO “responders,” TAO “non-responders,” and Sham-TAO rats (Fig. 2f) revealed a significant main effect of group (F(2, 13) = 26.59, P < 0.0001), with Holm-Sidak post hoc comparisons demonstrating that TAO responders had significantly greater increases in log(τMCAo/τc) relative to TAO non-responders or sham animals (P < .0001). Notably, TAO responders had very low blood flow after MCAo (τBaseline/τc = 0.06 ± 0.016 or ~6 % of baseline) and showed a drastic increase after TAO (Fig. 2g). Non-responders had more moderate ischemia (τBaseline/τc = 0.38 ± 0.14 or ~38 % of baseline) and no change after TAO (Fig. 2h). Friedman’s repeated measures tests on τBaseline/τc in responders and non-responding animals confirmed a statistically significant relationship between TAO and blood flow in TAO responders (χ2(3) = 7.60, P = .0239) but not non-responders (P > .05).

c = 0.38 ± 0.14 or ~38 % of baseline) and no change after TAO (Fig. 2h). Friedman’s repeated measures tests on τBaseline/τc in responders and non-responding animals confirmed a statistically significant relationship between TAO and blood flow in TAO responders (χ2(3) = 7.60, P = .0239) but not non-responders (P > .05). Histology Early infarct was visualized using H&E staining (Fig. 2i). Infarct volume was 93.84 ± 21.92 mm3 in TAO-treated rats and 107.1 ± 36.89 mm3 in Sham-TAO rats (unpaired t test, P > .05). While difference in pre-treatment but post-MCAo flow in animals that survived to 3-h post-MCAo makes comparisons between treatment groups difficult, plotting infarct volume as a function of blood flow (τBaseline/τc) in MCA segments 75 min after TAO or Sham-TAO illustrates a clear inverse relationship between collateral flow and early infarct volume (Fig. 2j).

ppocampus. In each region, the degree of neuronal cell injury was assessed according to an incremental 5-point scale: 0 is normal, 1 denotes 10 % selective neuronal injury, 2 denotes 10 to 50 % neuronal injury, 3 denotes 50 % neuronal injury, and 4 indicates confluent areas of pan-necrosis as previously described [42]. Tissue damage was not found in either sham-operated rats or the hemisphere contralateral to the ischemic side. Ischemic tissue damage was clearly seen in the vascular territory supplied by the MCA, i.e., the sensorimotor cortex and striatum. The pallid area was measured by utilizing an image analysis system (Image-Pro Plus; Image & Computer, Milan, Italy). Striatal and cortical damage was calculated in cubic millimeters. All histological studies were performed in a blinded fashion.

t-MCAo flow in animals that survived to 3-h post-MCAo makes comparisons between treatment groups difficult, plotting infarct volume as a function of blood flow (τBaseline/τc) in MCA segments 75 min after TAO or Sham-TAO illustrates a clear inverse relationship between collateral flow and early infarct volume (Fig. 2j). Discussion In the SENTIS trial, 515 patients with cortical ischemic stroke received 45 min of TAO or standard stroke treatment. SENTIS data suggest that TAO is safe and may be an effective therapy for subsets of stroke patients [18–20, 22, 29]. These trials in cortical stroke patients suggested maximal treatment efficacy for those with moderate stroke severity [19]. By redistributing blood towards the head from the peripheral circulation, TAO has the potential to augment both primary (circle of Willis) and secondary (pial) collateral flows [1, 2]. Here, using a large filament-based occlusion to block the origins of the MCA and ACA, we evaluated whether TAO could improve collateral blood flow and reduce mortality during severe stroke. Because this model would occlude the origins of the ACA and MCA, enhanced collateral flow through the circle of Willis (i.e., through the anterior and/or posterior communicating arteries) would be required to reduce ischemia and mortality. Our data show a significant reduction in early mortality with TAO. Moreover, LSCI data suggest that in animals with severe ischemia, TAO can improve flow in MCA segments downstream of ACA anastomoses and reduce ischemia. In animals with more moderate ischemia, TAO did not induce significant increases in blood flow. This may suggest that in animals or patients with effective primary collaterals, TAO does not further enhance flow. However, in severe ischemia associated with poor primary collateral flow and high mortality [30], TAO may be an effective acute therapy to maintain blood flow to ischemic regions at a level sufficient to prevent early mortality.

animals or patients with effective primary collaterals, TAO does not further enhance flow. However, in severe ischemia associated with poor primary collateral flow and high mortality [30], TAO may be an effective acute therapy to maintain blood flow to ischemic regions at a level sufficient to prevent early mortality. While promising, it is important to note that our measure of early mortality may not reflect total mortality outside the acute period. Previous reports suggest that mortality after permanent MCAo in Sprague Dawley rats results in the days following ischemic onset rather than the first few hours [31]. As such, our early mortality was surprising. However, our occlusion model used a large calibre (~0.54 mm including coating) filament to block carotid flow to the ACA and MCA, and this severe ischemia may account for the early mortality. Consistent with this interpretation, mortality of 13 % in the first 24 h has been reported after MCAo with 0.193-mm diameter filaments, whereas smaller filaments (0.180 mm) produced smaller infarcts and did not result in mortality in this period [32]. We observed increased TAO efficacy (greater increase in flow relative to post-stroke levels) in rats with severe ischemia prior to treatment, and early mortality may result from this severe ischemia in untreated rats. However, while we postulate a mechanism involving augmentation of primary collateral flow (via the circle of Willis), this was not verified by our imaging (which measured flow in surface arteries downstream of any increases in primary collateral flow). Nonetheless, the current data suggests that TAO warrants further investigation as stand-alone neuroprotective strategy to reduce acute mortality after severe stroke and as an adjunct therapy to maintain tissue viability during treatment of severe ischemic stroke.

downstream of any increases in primary collateral flow). Nonetheless, the current data suggests that TAO warrants further investigation as stand-alone neuroprotective strategy to reduce acute mortality after severe stroke and as an adjunct therapy to maintain tissue viability during treatment of severe ischemic stroke. Compliance with Ethical Standards Funding This work was supported by the Canadian Institutes of Health Research (IRW, 110967) and Alberta Innovates Health Solutions (Scholarship Award, IRW, 200900509). Conflict of Interest The authors have no conflicts of interest. Ethical Approval All applicable international, national, and institutional guidelines for the care and use of animals were followed.

Introduction Stroke is the second leading cause of death and the preeminent cause of neurological disability worldwide [1–3]. Ischemic stroke results from the sudden decrease or loss of blood circulation to an area of the brain, resulting in a corresponding loss of neurological function. Deficits can include partial paralysis, difficulties with memory, thinking, language, and movement. While the prompt restoration of blood flow to the ischemic tissue is the current strategy of choice in clinical stroke treatment, this can cause additional damage and exacerbate neurocognitive deficits. Inflammation is a key feature of cerebral ischemia [4, 5] with major immune system players, namely microglia [6, 7] and mast cells [8, 9] acting as early responders. This response leads to the production of pro-inflammatory mediators and infiltration of other inflammatory cell populations (e.g., neutrophils, T cell subsets, monocyte/macrophages) into the ischemic brain tissue. Additional, late-phase responders include reactive astrocytes and the resulting glial scar which forms in the boundary zone of the ischemic core and may play a critical role, both in detrimental and beneficial terms [10].

ulations (e.g., neutrophils, T cell subsets, monocyte/macrophages) into the ischemic brain tissue. Additional, late-phase responders include reactive astrocytes and the resulting glial scar which forms in the boundary zone of the ischemic core and may play a critical role, both in detrimental and beneficial terms [10]. Experimental animal models of stroke enabled the identification of a wide array of anti-inflammatory/neuroprotective compounds, including anti-epileptics, inhibitors of inducible nitric oxide synthase and kinases, minocycline, antioxidants, and polyphenols [11–15]. Some of these putative neuroprotectants have been tested in human clinical trials, but they have yielded little positive outcome [3, 16]. Nevertheless, neuroprotection studies utilizing animal models still provide useful insights into strategies for limiting stroke severity as they continue to offer translational potential for improving future stroke outcome [3, 17]. The failure of therapies targeted only to neuronal cell protection may be attributable, in part, to a lack of concomitant protection of cerebral blood vessels from secondary injury by inflammation and reactive oxygen species/reactive nitrogen species. Although anti-inflammatory approaches have proven successful in animal models of stroke [18, 19], attempts to translate this into clinical application have fallen short of expectations [20, 21].

ection of cerebral blood vessels from secondary injury by inflammation and reactive oxygen species/reactive nitrogen species. Although anti-inflammatory approaches have proven successful in animal models of stroke [18, 19], attempts to translate this into clinical application have fallen short of expectations [20, 21]. In response to tissue injury and stress, the body is known to respond by producing molecules “on demand” which function to restore homeostatic balance and prevent further damage [22]. Among these is a class of lipid signaling molecules, the N-acylethanolamines (NAEs) [23, 24]. One NAE, in particular N-palmitoylethanolamine (PEA or palmitoylethanolamide), is surrounded by a large number of observations supporting its role in maintaining cellular homeostasis by acting as a mediator of resolution of inflammatory processes [25, 26]. These past years have witnessed a continually growing number of studies confirming the anti-neuroinflammatory and neuroprotective actions of PEA [27–30]. Interestingly, several recent studies have shown that a co-ultramicronized PEA/luteolin composite (co-ultraPEALut, 10:1 mass ratio) is more efficacious than PEA alone [31, 32], including animal models of spinal cord injury [33] and traumatic brain injury [34].

uroinflammatory and neuroprotective actions of PEA [27–30]. Interestingly, several recent studies have shown that a co-ultramicronized PEA/luteolin composite (co-ultraPEALut, 10:1 mass ratio) is more efficacious than PEA alone [31, 32], including animal models of spinal cord injury [33] and traumatic brain injury [34]. Based on the above observations, we carried out a two-part study with co-ultraPEALut in cerebral ischemia. In the first part of the study, we analyzed the neuroprotective and anti-inflammatory properties of co-ultraPEALut in a rat model of middle cerebral artery occlusion (MCAo), while in the second part, the effects of co-ultraPEALut (Glialia®) were assessed in stroke patients undergoing rehabilitation therapy. We chose to conduct the study in subacute-phase stroke patients, given that even at this stage, certain characteristic pathological features, in particular neuro-inflammatory processes, are still active and able to cause continued neuronal cell damage. Moreover, many patients with ischemic stroke, despite optimal medical treatment received during the acute phase, often fail to recover (or only partially), leading to persistent disability requiring rehabilitation. As Glialia® is already a marketed product, we investigated whether treatment with Glialia®, carried out simultaneously with rehabilitation therapy, can bring about a better functional recovery in stroke patients in the subacute phase over a prolonged time frame.

eading to persistent disability requiring rehabilitation. As Glialia® is already a marketed product, we investigated whether treatment with Glialia®, carried out simultaneously with rehabilitation therapy, can bring about a better functional recovery in stroke patients in the subacute phase over a prolonged time frame. Materials and Methods Middle Cerebral Artery Occlusion Male Wistar rats (Harlan, Italy) weighing 270–290 g were used. Animals were housed in groups of three and kept on a 12-h light/dark cycle, under standardized temperature, humidity, and light conditions with free access to food and water. Animal care and use followed directives of the Council of the European Community (86/609/EC). All efforts were made to minimize animal suffering and to reduce the number of animals used. Focal cerebral ischemia was induced by transient MCAo in the right hemisphere. The rats were anesthetized with 5.0 % isoflurane (Baxter International) and spontaneously inhaled 1.0–2.0 % isoflurane in air by the use of a mask. Body core temperature was maintained at 37 °C with a heating pad and was monitored via an intrarectal type T thermocouple (Harvard, Kent, UK). The rats were placed in a stereotaxic system (Kopf). Middle cerebral carotid artery (MCA) occlusion was performed by inserting a 4-0 nylon monofilament (Ethilon; Johnson & Johnson, Somerville, NJ, USA), pre-coated with silicone (Xantopren; Heraeus Kulzer, Germany) and mixed with a hardener (Omnident, Germany), via the external carotid artery into the internal carotid artery to block the MCA, as originally described by Longa et al. [35] and modified by Melani et al. [36]. Sham rats were subjected to the same surgical procedure, and the filament was inserted into the internal carotid artery and immediately withdrawn. At the end of the surgical procedure, anesthesia was discontinued and the animals were returned to a prone position. Recovery from anesthesia took 15 min; animals were then allowed free access to food and water.

urgical procedure, and the filament was inserted into the internal carotid artery and immediately withdrawn. At the end of the surgical procedure, anesthesia was discontinued and the animals were returned to a prone position. Recovery from anesthesia took 15 min; animals were then allowed free access to food and water. Regional cerebral blood flow (rCBF) was monitored by laser Doppler flowmetry (PeriFlux System 5000; Perimed AB, Stockholm, Sweden) with the use of a flexible probe over the skull as described earlier [37]. The rats that did not show a CBF reduction of at least 70 % were excluded from subsequent experiments. Cranial temperature was maintained at 36.8–37.5 °C with a heating pad. Some physiologic parameters including cranial temperature, arterial pH, PaCO2, PaO2, and glucose were measured in six additional rats. Arterial blood samples were taken before ischemia (baseline), during ischemia, and after ischemia for gases and plasma glucose measurements (Table 1).Table 1 Physiological parameters, mean ± SD (n = 8) Time point Temperature (°C) pH PaCO2 (mmHg) PaO2 (mmHg) Glucose (mmol/l) Baseline 37.2 ± 0.3 37.32 ± 0.2 38.2 ± 3.9 104.5 ± 10.9 4.31 ± 0.31 During ischemia 37.0 ± 0.2 37.36 ± 0.3 38.5 ± 5.1 103.2 ± 8.1 4.25 ± 0.22 After ischemia 36.8 ± 0.3 37.33 ± 0.1 38.3 ± 4.2 102.9 ± 9.9 4.27 ± 0.19

Regional cerebral blood flow (rCBF) was monitored by laser Doppler flowmetry (PeriFlux System 5000; Perimed AB, Stockholm, Sweden) with the use of a flexible probe over the skull as described earlier [37]. The rats that did not show a CBF reduction of at least 70 % were excluded from subsequent experiments. Cranial temperature was maintained at 36.8–37.5 °C with a heating pad. Some physiologic parameters including cranial temperature, arterial pH, PaCO2, PaO2, and glucose were measured in six additional rats. Arterial blood samples were taken before ischemia (baseline), during ischemia, and after ischemia for gases and plasma glucose measurements (Table 1).Table 1 Physiological parameters, mean ± SD (n = 8) Time point Temperature (°C) pH PaCO2 (mmHg) PaO2 (mmHg) Glucose (mmol/l) Baseline 37.2 ± 0.3 37.32 ± 0.2 38.2 ± 3.9 104.5 ± 10.9 4.31 ± 0.31 During ischemia 37.0 ± 0.2 37.36 ± 0.3 38.5 ± 5.1 103.2 ± 8.1 4.25 ± 0.22 After ischemia 36.8 ± 0.3 37.33 ± 0.1 38.3 ± 4.2 102.9 ± 9.9 4.27 ± 0.19 Animal Groups and Treatments The rats were randomly allocated to the following groups:Ischemia/reperfusion + vehicle: the rats (n = 20) were subjected to MCAo (2 h) followed by 24 h reperfusion [38]. Carboxymethyl cellulose in saline (1.5 %, w/v) was administered (os) 1 h after ischemia and 6 h after reperfusion, and the rats were sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10);

vehicle: the rats (n = 20) were subjected to MCAo (2 h) followed by 24 h reperfusion [38]. Carboxymethyl cellulose in saline (1.5 %, w/v) was administered (os) 1 h after ischemia and 6 h after reperfusion, and the rats were sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10); Ischemia/reperfusion + co-ultraPEALut: the rats (n = 20) were subjected to surgical procedures as described above. Co-ultraPEALut (1 mg/kg in 1.5 % carboxymethyl cellulose, os) was administered (os) 1 h after ischemia and 6 h after reperfusion, and the rats were sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10); Sham + vehicle: the rats (n = 20) were subjected to identical surgical procedures except for MCA occlusion and were kept under anesthesia for the duration of the experiment. The animals were treated (os) with 1.5 % (w/v) carboxymethyl cellulose in saline at the same time point as the MCAo group and sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10); Sham + co-ultraPEALut: the rats (n = 20) identical to sham-operated rats except for administration of co-ultraPEALut (1 mg/kg in 15 % carboxymethyl cellulose, os) 1 h after ischemia and 6 h after reperfusion were sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10).

Sham + vehicle: the rats (n = 20) were subjected to identical surgical procedures except for MCA occlusion and were kept under anesthesia for the duration of the experiment. The animals were treated (os) with 1.5 % (w/v) carboxymethyl cellulose in saline at the same time point as the MCAo group and sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10); Sham + co-ultraPEALut: the rats (n = 20) identical to sham-operated rats except for administration of co-ultraPEALut (1 mg/kg in 15 % carboxymethyl cellulose, os) 1 h after ischemia and 6 h after reperfusion were sacrificed 24 h later for evaluation of histological damage (n = 10) and Western blot (n = 10). In a separate set of experiments, 10 animals from each group were observed until 24 h after MCAo to evaluate motor behavior, perform neurological testing, and evaluate infarct damage (Table 2). The dose (1 mg/kg) of co-ultraPEALut, administration route (os), and vehicle used were based on our previous study [39].Table 2 Effect of co-ultraPEALut (1 mg/kg, os, 1 h after ischemia and 6 h after reperfusion) on neurological score and ischemic brain damage evaluated 24 h after transient MCAo Treatment Neurological score Cortical damage (mm3) Striatal damage (mm3) Vehicle 7.57 ± 0.21* 57.26 ± 2.33§§ 24.28 ± 2.37§

In a separate set of experiments, 10 animals from each group were observed until 24 h after MCAo to evaluate motor behavior, perform neurological testing, and evaluate infarct damage (Table 2). The dose (1 mg/kg) of co-ultraPEALut, administration route (os), and vehicle used were based on our previous study [39].Table 2 Effect of co-ultraPEALut (1 mg/kg, os, 1 h after ischemia and 6 h after reperfusion) on neurological score and ischemic brain damage evaluated 24 h after transient MCAo Treatment Neurological score Cortical damage (mm3) Striatal damage (mm3) Vehicle 7.57 ± 0.21* 57.26 ± 2.33§§ 24.28 ± 2.37§ Co-ultraPEALut 9.13 ± 0.23# 37.72 ± 1.26 12.23 ± 2.1 Sham operated 17.5 ± 0.15 0 0 Data are the mean ± SD of n of rats (neurological score and brain damage, n = 10). The volume of the ipsilateral hemisphere was as follows (mean ± SD): 127.75 ± 3.59 mm3 in co-ultraPEALut-treated rats. One-way ANOVA followed by Bonferroni post hoc test was performed for neurological score *p < 0.001, versus sham-operated rats; # p < 0.05, versus MCAo + vehicle (cortical and striatal damage was analyzed using unpaired Student’s t test); § p < 0.05; §§ p < 0.01, versus co-ultraPEALut-treated rats

Co-ultraPEALut 9.13 ± 0.23# 37.72 ± 1.26 12.23 ± 2.1 Sham operated 17.5 ± 0.15 0 0 Data are the mean ± SD of n of rats (neurological score and brain damage, n = 10). The volume of the ipsilateral hemisphere was as follows (mean ± SD): 127.75 ± 3.59 mm3 in co-ultraPEALut-treated rats. One-way ANOVA followed by Bonferroni post hoc test was performed for neurological score *p < 0.001, versus sham-operated rats; # p < 0.05, versus MCAo + vehicle (cortical and striatal damage was analyzed using unpaired Student’s t test); § p < 0.05; §§ p < 0.01, versus co-ultraPEALut-treated rats Motor Behavior About 1 h after MCAo, the rats showed spontaneous turning behavior that was evaluated every 15 min for 3–4 h post-MCAo, as described by Melani et al. [40]. Motor activity was assessed in sham-operated and vehicle- and drug-treated rats. After being placed in a round cage, the rat began rapid unidirectional walking along the perimeter of the cage and to chase its tail. The number of complete rotations was recorded manually. Five separate counting periods of 3 min each, separated by 15-min intervals, were made. Values are reported as mean rotation number/h during the five counting periods. The same rats were also evaluated after awakening and 24 h later for failure to extend fully the left forepaw and for contralateral turning when pulled by the tail. Under this condition, all the rats showed a clear circling to the left side.

alues are reported as mean rotation number/h during the five counting periods. The same rats were also evaluated after awakening and 24 h later for failure to extend fully the left forepaw and for contralateral turning when pulled by the tail. Under this condition, all the rats showed a clear circling to the left side. Neurological Scoring Neurological evaluation of motor sensory functions was carried out prior to and 24 h after MCAo by an examiner blinded to the procedure, always between 10:00 a.m. and 11:00 a.m. to exclude behavioral changes based on circadian rhythm. The neurological examination consisted of six tests [41]: (i) spontaneous activity, (ii) symmetry in limb movements, (iii) forepaw outstretching, (iv) climbing, (v) body proprioception, and (vi) response to vibrissae touch. The six test scores were summed to generate an assigned final score, which ranged from a minimum of 3 to a maximum of 18.

ion consisted of six tests [41]: (i) spontaneous activity, (ii) symmetry in limb movements, (iii) forepaw outstretching, (iv) climbing, (v) body proprioception, and (vi) response to vibrissae touch. The six test scores were summed to generate an assigned final score, which ranged from a minimum of 3 to a maximum of 18. Histological Evaluation of Infarct Damage After evaluation of motor behavior and neurological deficit 24 h post-MCAo, the vehicle-treated (n = 10), co-ultraPEALut-treated (n = 10), and sham-operated rats (n = 10) were randomly selected and analyzed for ischemic damage using published histological methods [36]. The rats were anesthetized and sacrificed by decapitation. The brains were rapidly removed and fixed with Carnoy’s solution (6:3:1 absolute ethanol, chloroform, and glacial acetic acid) and then embedded in paraffin after dehydration in graded series of ethanol and xylol. Coronal sections (7–8 μm) were collected at 1-mm intervals at eight different levels through the striatum (from +2.2 Bregma to −4.8 Bregma, corresponding to the ischemic area) and were stained with acetate cresyl violet (1 %) or hematoxylin and eosin (H&E). The lesioned area was viewed under light microscopy (Axiostar Plus equipped with AxioCam MRc, Zeiss) as a pale zone lacking acetate cresyl violet staining, a refection of the extent of unlabeled necrotic neurons 24 h after MCAo [41]. The specific regions included the cortical areas and hippocampus. In each region, the degree of neuronal cell injury was assessed according to an incremental 5-point scale: 0 is normal, 1 denotes 10 % selective neuronal injury, 2 denotes 10 to 50 % neuronal injury, 3 denotes 50 % neuronal injury, and 4 indicates confluent areas of pan-necrosis as previously described [42].

territory supplied by the MCA, i.e., the sensorimotor cortex and striatum. The pallid area was measured by utilizing an image analysis system (Image-Pro Plus; Image & Computer, Milan, Italy). Striatal and cortical damage was calculated in cubic millimeters. All histological studies were performed in a blinded fashion. Quantification of Infarct Volume The rats were anesthetized with ketamine and decapitated, and the brains were carefully removed and cut into five coronal slices of 2 mm in thickness. Slices were incubated in a 2 % solution of 2,3,5-triphenyltetrazolium chloride (TTC) at 37 °C for 30 min and immersion fixed in 10 % buffered formalin solution. TTC stains viable brain tissue red, while infracted tissue remains unstained [43]. For quantification of infracted area and volumes, the brain slices were photographed using a digital camera (Canon 4×; Canon Inc., China) and image analysis was performed on a personal computer with ImageJ for Windows [44]. To compensate for the effect of brain edema, the corrected infarct area equals the left hemisphere area minus (right hemisphere area minus infarct area) [45]. The corrected total infarct volume was calculated by summing the infarct area in each slice and multiplying by slice thickness (2 mm).

h ImageJ for Windows [44]. To compensate for the effect of brain edema, the corrected infarct area equals the left hemisphere area minus (right hemisphere area minus infarct area) [45]. The corrected total infarct volume was calculated by summing the infarct area in each slice and multiplying by slice thickness (2 mm). Immunohistochemical Localization of Chymase, Tryptase, and Glial Fibrillary Acidic Protein Twenty-four hours after ischemia, tissues were fixed in 10 % (w/v) phosphate-buffered formaldehyde and 7-μm sections were prepared from paraffin-embedded tissues. After deparaffinization, endogenous peroxidase was quenched with 0.3 % (v/v) hydrogen peroxide in 60 % (v/v) methanol for 30 min. The sections were permeabilized with 0.1 % (w/v) Triton X-100 in phosphate-buffered saline (PBS) for 20 min. Nonspecific adsorption was minimized by incubation in 2 % (v/v) normal goat serum in PBS for 20 min. Endogenous biotin or avidin binding sites were blocked by sequential incubation for 15 min with biotin and avidin, respectively. Sections were then incubated overnight at room temperature with one of the following primary antibodies: anti-chymase rabbit polyclonal antibody (1:1000 in PBS), anti-tryptase rabbit polyclonal antibody (1:1000 in PBS), or anti-glial fibrillary acidic protein (GFAP) rabbit polyclonal antibody (1:100 in PBS) (Santa Cruz, DBA, Milan, Italy). Sections were then washed with PBS and incubated with a biotin-conjugated goat anti-rabbit IgG and avidin–biotin–peroxidase complex (Vector Laboratories, DBA). No immunostaining was observed when either the primary or secondary antibody was omitted. Images captured (n = 5 photos from each sample collected from all rats in each experimental group) were quantitatively assessed for a difference in immunoreactivity by computer-assisted color image analysis (Leica QWin V3, Cambridge, UK). All analyses were carried out by an investigator blinded to the treatment. The number of GFAP-positive cells was acquired in three defined brain regions: (i) a cortical infarct border zone (IBZ), defined by the middle of the tangent to the cortical lesion; (ii) an ipsilateral cortical remote zone (RZ), adjacent to the interhemispheric gap; and (iii) a control region within the contralateral hemisphere (CH) cortex equivalent to the location of the IBZ. The number of GFAP-positive cells was counted in three sections per animal and presented as the number of positive cells per high-power field.

ortical remote zone (RZ), adjacent to the interhemispheric gap; and (iii) a control region within the contralateral hemisphere (CH) cortex equivalent to the location of the IBZ. The number of GFAP-positive cells was counted in three sections per animal and presented as the number of positive cells per high-power field. Western Blot Analysis Cytosolic extracts were prepared with slight modifications. Briefly, brain tissues from each rat were suspended in extraction buffer A containing 10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 3 μg/ml pepstatin A, 2 μg/ml leupeptin, 15 μg/ml trypsin inhibitor, and 40 μM benzamidine; homogenized at the highest setting for 2 min, and centrifuged at 13,000g for 3 min at 4 °C. Supernatants that represented the cytosolic fraction were retained and stored at −80 °C until being used. Equivalent amounts of protein for each sample were loaded per lane and electrophoretically separated using 10 % denaturing polyacrylamide gel electrophoresis (PAGE) (SDS-PAGE). The filters were blocked with 1× Tris-buffered saline (pH 7.5) and 5 % (w/v) nonfat dried milk for 1 h at room temperature and subsequently probed with the following primary antibodies: rabbit polyclonal anti-GFAP (1:2000), anti-brain-derived neurotrophic factor (BDNF) (1:1000; Santa Cruz Biotechnology), or anti-glial cell line-derived neurotrophic factor (GDNF) (1:1000; Santa Cruz Biotechnology), in 1× PBS, 5 % (w/v) nonfat dried milk, and 0.1 % Tween-2 at 4 °C overnight. Membranes were incubated with peroxidase-conjugated bovine anti-mouse IgG secondary antibody or peroxidase-conjugated goat anti-rabbit IgG (1:2000; Jackson ImmunoResearch, West Grove, PA, USA) for 1 h at room temperature. To control for protein loading, the blots were stripped by agitation with 3 % glycine (pH 2) and blocked in 1× PBS/5 % (w/v) nonfat dried milk for 1 h at room temperature and then probed for 2 h at room temperature with anti-β-actin antibody (Sigma-Aldrich). Relative expression of the protein bands was quantified by densitometric scanning of the X-ray films with a GS-700 Imaging Densitometer (Bio-Rad Laboratories) and a computer program (Molecular Analyst, IBM) and standardized for densitometric analysis to β-actin levels.

rature with anti-β-actin antibody (Sigma-Aldrich). Relative expression of the protein bands was quantified by densitometric scanning of the X-ray films with a GS-700 Imaging Densitometer (Bio-Rad Laboratories) and a computer program (Molecular Analyst, IBM) and standardized for densitometric analysis to β-actin levels. Materials Unless otherwise stated, all compounds were obtained from Sigma-Aldrich Company Ltd. (Milan, Italy). All other chemicals were of the highest commercial grade available. All stock solutions were prepared in non-pyrogenic saline (0.9 % NaCl; Baxter, Italy). Co-ultraPEALut was kindly supplied by the Epitech Group (Saccolongo (PD), Italy). The methodology underlying its preparation has been described previously [33].

ly). All other chemicals were of the highest commercial grade available. All stock solutions were prepared in non-pyrogenic saline (0.9 % NaCl; Baxter, Italy). Co-ultraPEALut was kindly supplied by the Epitech Group (Saccolongo (PD), Italy). The methodology underlying its preparation has been described previously [33]. Assessment of Co-ultraPEALut in Stroke Patients An observational study to evaluate the effects of co-ultraPEALut (Glialia®) administration in stroke patients was carried out between April 2013 and June 2014, involving 37 neuromotor rehabilitation facilities throughout Italy. The patients selected had experienced a first ischemic stroke, were clinically stabilized, and underwent rehabilitative therapy. Individuals with previous hospitalizations for stroke and with hemorrhagic stroke were excluded. Of the 267 patients initially enrolled, 250 completed the study. Seventeen patients were excluded from statistical evaluation for the following reasons: bilateral stroke (4), no stroke at first event (3), ischemic event occurred ≥18 months before the start of the study (7), and inadequate information about ischemic event (3). Eighty-four percent of patients were undergoing rehabilitative treatment in inpatient centers and the remaining 16 % on an outpatient basis. Table 3 summarizes the patient demographics, while the location and type of stroke injury are collated in Table 4.Table 3 Patient demographics Total 250 Male 132 Female 118 Age (years) Mean ± SD 71.4 ± 12.4 Range 31–100 Inpatient, % (n) 84 (210) Outpatient, % (n) 16 (40) Table 4 Location and type of stroke injury

Assessment of Co-ultraPEALut in Stroke Patients An observational study to evaluate the effects of co-ultraPEALut (Glialia®) administration in stroke patients was carried out between April 2013 and June 2014, involving 37 neuromotor rehabilitation facilities throughout Italy. The patients selected had experienced a first ischemic stroke, were clinically stabilized, and underwent rehabilitative therapy. Individuals with previous hospitalizations for stroke and with hemorrhagic stroke were excluded. Of the 267 patients initially enrolled, 250 completed the study. Seventeen patients were excluded from statistical evaluation for the following reasons: bilateral stroke (4), no stroke at first event (3), ischemic event occurred ≥18 months before the start of the study (7), and inadequate information about ischemic event (3). Eighty-four percent of patients were undergoing rehabilitative treatment in inpatient centers and the remaining 16 % on an outpatient basis. Table 3 summarizes the patient demographics, while the location and type of stroke injury are collated in Table 4.Table 3 Patient demographics Total 250 Male 132 Female 118 Age (years) Mean ± SD 71.4 ± 12.4 Range 31–100 Inpatient, % (n) 84 (210) Outpatient, % (n) 16 (40) Table 4 Location and type of stroke injury % of patients Location Supratentorial 35.9 Subtentorial 10.5 Anterior circle 22.7 Posterior circle 6.4 Lacunar 8.6 Multiple location 15.9 Type Diffuse of injury 62.0 Internal capsule 26.6 External capsule 11.4

Total 250 Male 132 Female 118 Age (years) Mean ± SD 71.4 ± 12.4 Range 31–100 Inpatient, % (n) 84 (210) Outpatient, % (n) 16 (40) Table 4 Location and type of stroke injury % of patients Location Supratentorial 35.9 Subtentorial 10.5 Anterior circle 22.7 Posterior circle 6.4 Lacunar 8.6 Multiple location 15.9 Type Diffuse of injury 62.0 Internal capsule 26.6 External capsule 11.4 Therapy with Glialia® commenced at a median time of 18 days after the onset of the acute phase of stroke for inpatients and after a median time of 104 days for patients undergoing rehabilitation on an outpatient basis. Written informed consent was obtained from each patient or responsible family member prior to the beginning of the study. All patients were administered Glialia® (composed of co-ultramicronized 700 mg PEA and 70 mg luteolin, in microgranular form) sublingually, twice daily (every 12 h) for 60 days in association with the specific therapy (e.g., thrombolytics, aspirin, and anticoagulants) normally administered and/or with drugs prescribed for comorbidities, where present in the patient.

d of co-ultramicronized 700 mg PEA and 70 mg luteolin, in microgranular form) sublingually, twice daily (every 12 h) for 60 days in association with the specific therapy (e.g., thrombolytics, aspirin, and anticoagulants) normally administered and/or with drugs prescribed for comorbidities, where present in the patient. All patients underwent the following evaluations at baseline (T0) and after 30 days (T30) of treatment: (i) stroke severity by the Canadian Neurological Scale (CNS) [46], a tool used to assess and monitor the neurological status of patients with stroke; (ii) impairment of cognitive abilities by the Mini Mental State Examination (MMSE) [47] adjusted for age and educational level; (iii) degree of spasticity, by means of the Ashworth Scale; (iv) pain, by the Numeric Rating Scale (NRS), a scale of 0 to 10 points with 0 being no pain at all and 10 being the worst pain imaginable; and (v) independence in activities of daily living by the Barthel Index, which scales functional capacity in terms of patient self-care [48]. Evaluation of the degree of autonomy and/or dependence of the patient from help, both physical and verbal, in carrying out daily living activities was performed also at the 60th day, concurrently with treatment end. During the period of rehabilitation therapy, patients were subjected to routine blood chemistry and hematology analyses and monitored for the possible occurrence of adverse events.

elp, both physical and verbal, in carrying out daily living activities was performed also at the 60th day, concurrently with treatment end. During the period of rehabilitation therapy, patients were subjected to routine blood chemistry and hematology analyses and monitored for the possible occurrence of adverse events. Statistical Analysis Preclinical Data Statistically significant differences in neurological score were evaluated by one-way ANOVA, followed by the Newman–Keuls multiple comparison test. Statistically significant differences in the volume of brain ischemic damage were evaluated using unpaired Student’s t test. The results were analyzed by one-way ANOVA followed by the Bonferroni post hoc test for multiple comparisons in neurological score. All values are expressed as mean ± SD. A P value of less than 0.05 was considered significant. Clinical Data To evaluate changes of means in time, the generalized linear mixed model (GLMM) with SAS 9.2 was used. Variables such as gender, age, and hospitalization type were included in the model as covariates. The Bonferroni–Holm correction for multiple comparisons was adopted for the CNS evaluation.

Statistical Analysis Preclinical Data Statistically significant differences in neurological score were evaluated by one-way ANOVA, followed by the Newman–Keuls multiple comparison test. Statistically significant differences in the volume of brain ischemic damage were evaluated using unpaired Student’s t test. The results were analyzed by one-way ANOVA followed by the Bonferroni post hoc test for multiple comparisons in neurological score. All values are expressed as mean ± SD. A P value of less than 0.05 was considered significant. Clinical Data To evaluate changes of means in time, the generalized linear mixed model (GLMM) with SAS 9.2 was used. Variables such as gender, age, and hospitalization type were included in the model as covariates. The Bonferroni–Holm correction for multiple comparisons was adopted for the CNS evaluation. Results Physiologic Parameters and rCBF The changes of physiological parameters are shown in Table 1. No statistical significance among different time points for any of these parameters was noted. All physiological parameters were within the normal range for the experimental process. There was no significant difference in rCBF between the control and treated group at the corresponding time points (Fig. 1). Monitoring of rCBF showed successful MCAo.Fig. 1 Regional cerebral blood flow analysis (rCBF). rCBF was monitored 10 min before ischemia (baseline); at 10, 60, and 110 min during ischemia; and 5 min after ischemia. rCBF monitoring was used to confirm successful induction of MCAo

the corresponding time points (Fig. 1). Monitoring of rCBF showed successful MCAo.Fig. 1 Regional cerebral blood flow analysis (rCBF). rCBF was monitored 10 min before ischemia (baseline); at 10, 60, and 110 min during ischemia; and 5 min after ischemia. rCBF monitoring was used to confirm successful induction of MCAo Neuroprotective Effects of Co-ultraPEALut on Turning Behavior, Neurological Deficit, and Ischemic Brain Damage in MCAo Rats Turning behavior after permanent intraluminal MCAo is a precocious index of neurological deficit and neuronal cell damage [40, 41]. One hour after MCAo, the rats engaged in turning behavior towards the side contralateral to the ischemic hemisphere. This acute behavioral response lasted several hours and was no longer evident 24 h after ischemia. Co-ultraPEALut, administered orally, blocked this acute behavioral response. Turning behavior, expressed as mean rotations per hour, was 788 ± 133 in MCAo vehicle-treated rats (n = 10) and 138 ± 42 (n = 10) in co-ultraPEALut-treated rats (mean ± SD). This effect of co-ultraPEALut was statistically significant (unpaired Student’s t test: p < 0.001 vs. vehicle-treated rats). Sham-operated rats did not show any turning behavior.

ean rotations per hour, was 788 ± 133 in MCAo vehicle-treated rats (n = 10) and 138 ± 42 (n = 10) in co-ultraPEALut-treated rats (mean ± SD). This effect of co-ultraPEALut was statistically significant (unpaired Student’s t test: p < 0.001 vs. vehicle-treated rats). Sham-operated rats did not show any turning behavior. Neurological deficit and ischemic striatal and cortical damage of MCAo vehicle-treated, MCAo co-ultraPEALut-treated, and sham-operated rats were evaluated after 24 h (Table 2). Co-ultraPEALut significantly improved MCAo-induced neurological deficit in comparison to vehicle-treated rats and reduced both cortical and striatal damages in comparison to vehicle-treated rats. No damage was found in the sham-operated rats or in the hemisphere contralateral to the ischemic side.

after 24 h (Table 2). Co-ultraPEALut significantly improved MCAo-induced neurological deficit in comparison to vehicle-treated rats and reduced both cortical and striatal damages in comparison to vehicle-treated rats. No damage was found in the sham-operated rats or in the hemisphere contralateral to the ischemic side. Effect of Co-ultraPEALut on Infarct Area and Morphological Changes Twenty-two hours after ischemia/reperfusion, the rats developed infarcts affecting the cortex and striatum (Fig. 2a). The co-ultraPEALut treatment group (n = 10) (Fig. 2b) had a significantly smaller infarct area (n = 10) (Fig. 2c) 24 h post-MCAo compared with the control (n = 10) (Fig. 2a). H&E staining revealed morphologically healthy cerebral neurons in the sham group (data not shown), whereas 24 h after reperfusion, the brain sections from ischemic rats showed a paucity of intact neurons in those areas (Fig. 2d (D1); see histological score H) and the presence of multiple vacuolated interspaces (Fig. 2d (D1); see histological score H). In contrast, the corresponding areas in the co-ultraPEALut group displayed partial neuronal cell loss only and the presence of intact neurons between the vacuolated spaces (Fig. 2e (E1); see histological score H). Hippocampal CA1 pyramidal cells 24 h after reperfusion were layered and arranged uniformly with large, round, transparent, and intact nuclei (data not shown). In the ischemia/reperfusion group, CA1 neurons were significantly reduced in numbers and characterized by pyknotic and indistinct nuclei (Fig. 2f (F1); see histological score H). Co-ultraPEALut treatment significantly decreased neuronal cell death in CA1 (Fig. 2g (G1); see histological score H).Fig. 2 Co-ultraPEALut reduces infarct area and histological damage in a rat MCAo model. Co-ultraPEALut (1 mg/kg) was administered orally 1 h after ischemia and 6 h post-reperfusion. TTC staining of coronal brain sections 24 h after MCAo. Infarcted brain tissue appears unstained (a), while co-ultraPEALut treatment (b) significantly reduced infarct area (c). H&E staining of MCAo tissue shows a loss of neurons and the presence of multiple vacuolated interspaces (d, D1), which were significantly ameliorated in co-ultraPEALut-treated animals (e, E1). Moreover, damage to CA1 hippocampal neurons following ischemia/reperfusion (f, F1) was significantly attenuated by co-ultraPEALut (g, G1). This figure is representative of at least three independent experiments. Quantification of histological score (h).

gnificantly ameliorated in co-ultraPEALut-treated animals (e, E1). Moreover, damage to CA1 hippocampal neurons following ischemia/reperfusion (f, F1) was significantly attenuated by co-ultraPEALut (g, G1). This figure is representative of at least three independent experiments. Quantification of histological score (h). *p < 0.01, versus sham group; °p < 0.01, versus MCAo + vehicle group. ND: not detectable

gnificantly ameliorated in co-ultraPEALut-treated animals (e, E1). Moreover, damage to CA1 hippocampal neurons following ischemia/reperfusion (f, F1) was significantly attenuated by co-ultraPEALut (g, G1). This figure is representative of at least three independent experiments. Quantification of histological score (h). *p < 0.01, versus sham group; °p < 0.01, versus MCAo + vehicle group. ND: not detectable Effect of Co-ultraPEALut on Astrocyte Activation After MCAo Astrocytes were revealed by GFAP immunohistochemistry; active astrocytes were defined as GFAP-stained cells with an increased GFAP immunoreactivity. A basal number of astrocytes were detected in the sham controls (Fig. 3b). Twenty-two hours of ischemia/reperfusion resulted in significant increases in the number of astrocytes and GFAP immunoreactivity in the cortical IBZ and RZ when compared to the sham control or contralateral hemisphere (Fig. 3b, c). GFAP-positive astrocytes were more numerous in the border zone compared to the remote zone (Fig. 3b). The number of GFAP-stained cells in the contralateral hemisphere was similar to the sham control (Fig. 3c). Brain sections from sham-operated rats did not exhibit appreciable immunostaining for GFAP (data not shown), unlike infarct border zone sections from MCAo rats (Fig. 3c). Co-ultraPEALut reduced GFAP immunostaining in the infarct border zone of rats subjected to MCAo (Fig. 3d). Western blot analysis confirmed the immunohistochemical findings, with a low level of GFAP expression in extracts from the occluded hemisphere of sham-operated rats and a significant increase in the MCAo vehicle-treated group (Fig. 3e (E1)). Co-ultraPEALut decreased the damage-induced rise in GFAP expression to a significant extent (Fig. 3d (D1)).Fig. 3 Co-ultraPEALut moderates MCAo-induced activation of astrocytes. A representative brain section showing the regions of interest used for analysis: cortical infarct border zone (IBZ), ipsilateral cortical remote zone (RZ), and contralateral hemisphere (CH) (a). The number of glial fibrillary acidic protein (GFAP)-positive cells was counted in three sections per animal in each experimental region and presented as the number of positive cells per high-power field (b). Immunohistochemical localization of GFAP shows positive staining in the IBZ of vehicle-treated MCAo animals (c) in contrast to the co-ultraPEALut-treated ischemic rats (d). Western blot analysis demonstrated a marked increase in GFAP-immunoreactive protein in ischemic tissue and its moderation by co-ultraPEALut treatment (e, E1).

chemical localization of GFAP shows positive staining in the IBZ of vehicle-treated MCAo animals (c) in contrast to the co-ultraPEALut-treated ischemic rats (d). Western blot analysis demonstrated a marked increase in GFAP-immunoreactive protein in ischemic tissue and its moderation by co-ultraPEALut treatment (e, E1). Values in E1 are mean ± SD for five to six rats from each group. *p < 0.01, versus sham group; # p < 0.01, versus MCAo group. Data are representative of three independent experiments Co-ultraPEALut Limits Ischemia-Induced Loss of BDNF and GDNF Expression Brain tissues from MCAo rats showed a marked reduction in the levels of the neurotrophic factors BDNF and GDNF, as determined by Western blot 24 h post-ischemia (Fig. 4a (A1), b (B1)). Treatment with co-ultraPEALut increased the expression levels of both BDNF and GDNF in comparison to with non-treated injured rats.Fig. 4 Co-ultraPEALut restores the MCAo-induced loss of BDNF and GDNF expression. Co-ultraPEALut (1 mg/kg) was administered orally 1 h after ischemia and 6 h post-reperfusion. Animals were sacrificed 24 h after MCAo, and brain tissue lysates were analyzed by Western blot for BDNF and GDNF. Ischemic rats exhibited a marked reduction in the levels of both BDNF (a, A1) and GDNF (b, B1) which were restored by co-ultraPEALut treatment. Data are mean ± SD for five to six rats from each group. *p < 0.01, versus sham group; **p < 0.001, versus sham group; # p < 0.01, versus MCAo group. Representative blots are shown in a, b, with quantification of all animals given in A1, B1

BDNF (a, A1) and GDNF (b, B1) which were restored by co-ultraPEALut treatment. Data are mean ± SD for five to six rats from each group. *p < 0.01, versus sham group; **p < 0.001, versus sham group; # p < 0.01, versus MCAo group. Representative blots are shown in a, b, with quantification of all animals given in A1, B1 Co-ultraPEALut Reduces Ischemia-Induced Mast Cell Infiltration and Chymase and Tryptase Expression Toluidine blue staining of the brain tissue from MCAo vehicle-treated rats identified cells with metachromatic granules, a characteristic of mast cells (Fig. 5a, c). In contrast, significantly fewer cells of this type were seen in the ischemic tissue of rats treated with co-ultraPEALut (Fig. 5b, c) and none in sham-operated rats (data not shown). While immunoreactivity for chymase and tryptase, two serine peptidases characteristic of mast cells, was absent in brain tissues from sham-operated rats (data not shown), there were substantial increases 24 h after ischemia (Fig. 5d, g, respectively, see densitometries F and I). Chymase and tryptase expression was attenuated in tissue sections from rats receiving co-ultraPEALut (Fig. 5e, h, respectively, see densitometries F and I).Fig. 5 Effect of co-ultraPEALut treatment on mast cells in MCAo rats. Co-ultraPEALut (1 mg/kg) was administered orally 1 h after ischemia and 6 h post-reperfusion. Animals were sacrificed 24 h after MCAo, and tissues were prepared for immunohistochemical analysis. (a) Toluidine blue-stained mast cells with their characteristic metachromatic granular inclusions were evident in MCAo-vehicle treated rats. (b) Co-ultraPEALut-treated ischemic animals displayed only occasional cells of this type. (c) Mast cell density per unit area. A marked increase in immunoreactivity for the mast cell enzymes chymase (d) and tryptase (g) was seen in the brain sections from MCAo vehicle-treated rats. This expression was significantly attenuated in the MCAo group treated with co-ultraPEALut (e, h). Densitometric analysis for chymase (f) and tryptase (i) immunoreactivities is given (five photomicrographs from each sample collected from all rats in each experimental group) and was carried out using Optilab Graftek software on a Macintosh personal computer (CPU G3-266). Data, expressed as percentage of total tissue area, are mean ± SD. *p < 0.01, versus sham group; °p < 0.01, versus MCAo + vehicle group. ND not detectable

rom each sample collected from all rats in each experimental group) and was carried out using Optilab Graftek software on a Macintosh personal computer (CPU G3-266). Data, expressed as percentage of total tissue area, are mean ± SD. *p < 0.01, versus sham group; °p < 0.01, versus MCAo + vehicle group. ND not detectable Co-ultraPEALut Effect on Bax and Bcl-2 Expression Western blot analysis was employed to determine the levels of Bax and Bcl-2 expression. As shown in Fig. 6, co-ultraPEALut administration effectively normalized the Bax level in the brain of rats subjected to MCAo (Fig. 6b). At the same time, co-ultraPEALut significantly restored the reduction in Bcl-2 expression observed in MCAo rats (Fig. 6c). As such, co-ultraPEALut administration normalized the Bax/Bcl-2 expression level (Fig. 6d).Fig. 6 Effects of co-ultraPEALut on Bax and Bcl-2. Co-ultraPEALut (1 mg/kg) was administered orally 1 h after ischemia and 6 h post-reperfusion. Animals were sacrificed 24 h after MCAo, and brain tissue lysates were analyzed by Western blot for Bax and Bcl-2. (a, b) Co-ultraPEALut treatment reduced the levels of Bax in the brain of ischemic rats. β-Actin was used as internal control. (a, c) Co-ultraPEALut administration increased the Bcl-2 levels in the brain tissue of ischemic rats. Co-ultraPEALut administration normalized Bax/Bcl-2 expression levels (d). Data are mean ± SD for five to six rats from each group. *p < 0.01, versus sham group; **p < 0.001, versus sham group; ### p < 0.001; # p < 0.01, versus MCAo group. Representative blots are shown in a, with quantification of all animals given in b–d

ultraPEALut administration normalized Bax/Bcl-2 expression levels (d). Data are mean ± SD for five to six rats from each group. *p < 0.01, versus sham group; **p < 0.001, versus sham group; ### p < 0.001; # p < 0.01, versus MCAo group. Representative blots are shown in a, with quantification of all animals given in b–d Clinical Assessment of Stroke Patients Treated with Glialia® Nine of the 250 enrolled patients dropped out because of death owing to severity of concomitant diseases or seriousness of ischemic state (2), transfer to another rehabilitation center (3), onset of diarrhea (2), gastric discomfort (1), and agitation (1). None of these reasons were considered by the attending physician to be treatment related. The CNS showed a significant improvement over time (p < 0.0001), with the mean total score increasing by 1.7 ± 0.10 between T0 and T30 (Fig. 7). Furthermore, GLMM analysis showed that while baseline average score in women was indicative of greater severity than males (p = 0.0049), improvement over time was independent of both gender and recovery type.Fig. 7 Neurological scoring for stroke patients treated with Glialia®. Patients were administered Glialia®, as described in the “Materials and Methods” section, for a period of 60 days and underwent the following evaluations at baseline (T0) and after 30 days (T30). Canadian Neurological Score values were 6.4 ± 0.16 and 8.1 ± 0.15 at T0 (242 patients) and T30 (237 patients), respectively. Delta: T30–T0 was calculated on the basis of 237 patients. There was a significant difference in time (***p < 0.0001) between T0 and T30. Mini Mental State Exam values were 20.2 ± 0.57 and 22.7 ± 0.47 at T0 (169 patients) and T30 (196 patients), respectively. Delta: T30–T0 was calculated on the basis of 167 patients. There were significant differences in time (***p < 0.0001) between T0 and T30. Seventy-two patients were unable to carry out the exam at T0; 28 of these were able to take the test at T30. Ashworth Scale values were 4.1 ± 0.32 and 3.7 ± 0.27 at T0 (246 patients) and T30 (244 patients), respectively. Delta: T30–T0 was calculated on the basis of 244 patients. There was a significant difference in time (**p < 0.0015) between T0 and T30. Numerical Rating Scale values were 2.1 ± 0.17 and 1.1 ± 0.11 at T0 (242 patients) and T30 (241 patients), respectively. Delta: T30–T0 was calculated on the basis of 241 patients. There was a significant difference in time (***p < 0.0001)

nts. There was a significant difference in time (**p < 0.0015) between T0 and T30. Numerical Rating Scale values were 2.1 ± 0.17 and 1.1 ± 0.11 at T0 (242 patients) and T30 (241 patients), respectively. Delta: T30–T0 was calculated on the basis of 241 patients. There was a significant difference in time (***p < 0.0001) Cognitive function evaluation performed by MMSE showed a significant improvement (p < 0.0001) compared to baseline performance after 30 days of Glialia® treatment (Fig. 7). The average score over these 30 days of treatment increased from 20.2 ± 0.57 (T0) to 22.7 ± 0.47 (T30). Out of the 78 patients who were unable to perform the MMSE at baseline, 28 were capable of carrying out the test at follow-up after 30 days of treatment. The MMSE analysis showed also that at baseline, female patients had a significantly (p = 0.0450) greater cognitive impairment compared to males. However, there was no time × recovery and gender effect. Muscle spasticity evaluated by the Ashworth Scale showed that the global spasticity significantly improved over time (p < 0.0015) (Fig. 7). The severity of spasticity at baseline was greater in outpatients (p < 0.0003) than in inpatients, and outpatient improvement was statistically significantly greater (p < 0.0075) than that of inpatients. These observations apply to the assessment of spasticity both overall and at the level of each limb.

0015) (Fig. 7). The severity of spasticity at baseline was greater in outpatients (p < 0.0003) than in inpatients, and outpatient improvement was statistically significantly greater (p < 0.0075) than that of inpatients. These observations apply to the assessment of spasticity both overall and at the level of each limb. The average pain intensity detected by NRS fell within a range of values considered clinically very mild. The average value for NRS at baseline was 2.1 ± 0.17; after 30 days of Glialia® treatment, this had decreased one point to a mean value of 1.1 ± 0.11 (p < 0.0001) (Fig. 7). The baseline mean score of pain intensity appeared higher in outpatients (3.1 ± 0.49) compared to inpatients (1.9 ± 0.18) and reached a statistical significance (p = 0.0014).

t baseline was 2.1 ± 0.17; after 30 days of Glialia® treatment, this had decreased one point to a mean value of 1.1 ± 0.11 (p < 0.0001) (Fig. 7). The baseline mean score of pain intensity appeared higher in outpatients (3.1 ± 0.49) compared to inpatients (1.9 ± 0.18) and reached a statistical significance (p = 0.0014). Patients’ independence and mobility in daily living activities (Barthel Index) showed a significant improvement (p < 0.0001) after 30 and 60 days of treatment (mean scores of 26.6 ± 1.69, 48.3 ± 1.91, and 60.5 ± 1.95 at T0, T30, and T60, respectively) (Fig. 8). Moreover, there was a highly significant difference between T30 and T60, indicative of a continuing improvement with time. The baseline degree of disability in daily living activity was significantly more severe in inpatients (p < 0.0003) than outpatients and in females (p < 0.0047) compared to males; improvement over time, however, appeared better in inpatients and was not influenced by gender.Fig. 8 Barthel Index score for stroke patients treated with Glialia®. Patients were administered Glialia®, as described in the “Materials and Methods” section, for a period of 60 days. Barthel Index values were 26.6 ± 1.69, 48.3 ± 1.91, and 60.5 ± 1.95 at T0 (242 patients), T30 (229 patients), and T60 (218 patients), respectively. There was a significant difference in the improvement between T0 and T30 (***p < 0.0001) and between T0 and T60 (### p < 0.0001). Moreover, there was a highly significant difference also between T30 and T60 (p < 0.0001). Female patients exhibited lower scores than males, and disability was worse in inpatients

ectively. There was a significant difference in the improvement between T0 and T30 (***p < 0.0001) and between T0 and T60 (### p < 0.0001). Moreover, there was a highly significant difference also between T30 and T60 (p < 0.0001). Female patients exhibited lower scores than males, and disability was worse in inpatients Tolerability to Glialia® treatment was excellent, with no adverse events ever having been observed over the course of this study. Further, routine blood chemistry and hematology analyses did not reveal any deviations from their normal ranges in relation to Glialia® treatment.

ectively. There was a significant difference in the improvement between T0 and T30 (***p < 0.0001) and between T0 and T60 (### p < 0.0001). Moreover, there was a highly significant difference also between T30 and T60 (p < 0.0001). Female patients exhibited lower scores than males, and disability was worse in inpatients Tolerability to Glialia® treatment was excellent, with no adverse events ever having been observed over the course of this study. Further, routine blood chemistry and hematology analyses did not reveal any deviations from their normal ranges in relation to Glialia® treatment. Discussion Cerebral ischemia continues to represent one of the principal unmet medical needs in today’s society. Stroke is especially devastating, given that it constitutes the most frequent cause of neurological disability worldwide. The underlying cellular mechanisms of stroke neuropathology are complex. An initial episode of focal hypoperfusion subsequently leads to excitotoxicity, oxidative damage, microvascular injury, blood–brain barrier dysfunction, and post-ischemic inflammation [4, 49–51]. Despite almost four decades of experimental animal investigations on stroke and the identification of a spectrum of anti-inflammatory/neuroprotective compounds [11–15], translatability of these findings to human clinical trials until now have been proven uniformly disappointing [3, 16]. In the present study, we describe the neuroprotective effects of co-ultramicronized palmitoylethanolamide/luteolin in a rat model of focal cerebral ischemia and, more importantly, the ability of co-ultraPEALut (Glialia®) to improve the neurological status of stroke patients undergoing neurorehabilitation.

ting [3, 16]. In the present study, we describe the neuroprotective effects of co-ultramicronized palmitoylethanolamide/luteolin in a rat model of focal cerebral ischemia and, more importantly, the ability of co-ultraPEALut (Glialia®) to improve the neurological status of stroke patients undergoing neurorehabilitation. Focal cerebral ischemia is accompanied by reactive astrogliosis [52] and activation of microglial cells in the hippocampal area [53]. Reactive gliosis can lead to the production of excessive amounts of cytokines as well as inflammatory products that exacerbate ischemic damage [54]. In our study, an increased in GFAP immunoreactivity was observed in the MCAo group in comparison to sham-operated animals while there was a marked reduction in the co-ultraPEALut-treated group. As previously reported [55], co-ultraPEALut is able to significantly reduce expression levels of GFAP after MCAo. In the current MCAo model, oral administration of co-ultraPEALut 1 h after ischemia and 6 h after reperfusion improved neurological score, reduced lesion size and histological damage, inhibited mast cell infiltration/degranulation and astrocyte activation (as measured by GFAP accumulation, a characteristic neuropathologic feature of ischemic brain injury [56]), and restored expression of BDNF and GDNF. Experimental and clinical studies have shown that BDNF and GDNF are upregulated at very early stages during brain ischemia [56]. Furthermore, exogenous administration of GDNF and BDNF reduced the toxic effects of excitatory amino acids, attenuated nitric oxide production, and lowered apoptosis/cell death in stroke animal models [56, 57]. In our preclinical findings, protein analysis showed that both BDNF and GFAP levels were downregulated by cerebral ischemia while a local and sustained increase in their expression in the perilesioned tissue followed oral administration of co-ultraPEALut.

d lowered apoptosis/cell death in stroke animal models [56, 57]. In our preclinical findings, protein analysis showed that both BDNF and GFAP levels were downregulated by cerebral ischemia while a local and sustained increase in their expression in the perilesioned tissue followed oral administration of co-ultraPEALut. Mast cell activation and degranulation is known to contribute to blood–brain barrier disruption in cerebral ischemia [58]. Other elements play a role in ischemic brain damage, such as activation of the transcription factor nuclear factor-κB [59, 60], inducible nitric oxide synthase [61, 62], and poly(ADP-ribose)synthetase [63]; increased expression of the pro-apoptotic protein Bax [64]; and decreased expression of the anti-apoptotic molecule Bcl-2 [65]. In the present study, co-ultraPEALut significantly reduced nuclear factor-κB translocation, attenuated poly(ADP-ribose)synthetase activation, and normalized Bax/Bcl-2 expression levels. These results are qualitatively similar to our earlier investigations with PEA alone, although in the latter, 10-fold higher doses of PEA (compared to co-ultraPEALut) were required for efficacy [55, 66].

ctor-κB translocation, attenuated poly(ADP-ribose)synthetase activation, and normalized Bax/Bcl-2 expression levels. These results are qualitatively similar to our earlier investigations with PEA alone, although in the latter, 10-fold higher doses of PEA (compared to co-ultraPEALut) were required for efficacy [55, 66]. PEA actions are mediated, at least in part, by activation of peroxisome proliferator-activated receptors, accompanied by a decrease in neutrophil influx and expression of pro-inflammatory proteins, such as inducible nitric oxide synthase and cyclooxygenase-2 [39, 67]. Luteolin displays specific anti-inflammatory effects, which are only partly explained by its antioxidant capacities. The anti-inflammatory activity of luteolin includes activation of antioxidative enzymes, suppression of the nuclear factor-κB pathway, and inhibition of pro-inflammatory substances [68, 69]. The molecular basis behind the superior pharmacological efficacy of co-ultraPEALut compared to comparable concentrations of the single chemical components is currently under investigation.

ion of antioxidative enzymes, suppression of the nuclear factor-κB pathway, and inhibition of pro-inflammatory substances [68, 69]. The molecular basis behind the superior pharmacological efficacy of co-ultraPEALut compared to comparable concentrations of the single chemical components is currently under investigation. Although inflammatory signaling is considered to participate in the early post-ischemic period, as the ischemic cascade progresses, cell death leads to a new phase of the inflammatory response, whereby the immune system becomes activated. There is evidence that both microglia and mast cells have a role at later times following the ischemic episode. Mast cells possess a potent armamentarium to target the components of the blood–brain barrier and basal lamina shortly after their activation in ischemia, whereas de novo production of mediators reactivates and maintains the process over the longer term [70–73]. Further, a newly published study demonstrates that mast cells can undergo a delayed and long-term activation following traumatic brain injury [74]. While the latter condition is not stroke, this report demonstrates that mast cells are not only early responders but also long-term players in brain injury. Further, it bears keeping in mind that in terms of acute versus late-stage treatment regimens in rats and man, one cannot necessarily make a direct comparison between time scales. In other words, a patient treated starting 60 days post-stroke does not have a rat equivalent.

s but also long-term players in brain injury. Further, it bears keeping in mind that in terms of acute versus late-stage treatment regimens in rats and man, one cannot necessarily make a direct comparison between time scales. In other words, a patient treated starting 60 days post-stroke does not have a rat equivalent. Despite literally hundreds of compounds and interventions that provide benefit in experimental models of cerebral ischemia, efficacy in humans remains to be demonstrated [75]. Many patients with ischemic stroke, despite optimal medical treatment received during the acute phase, often fail to recover (or only partially), leading to persistent disability requiring rehabilitation. As Glialia® is already a marketed product, we investigated whether treatment with Glialia®, carried out simultaneously with rehabilitation therapy, can bring about a better functional recovery in stroke patients in the subacute phase. The observations reported here demonstrate that a positive outcome in an animal stroke model can be translated into stroke patients. Open studies, however, necessarily impose certain limitations as, for example, the lack of a randomized controlled trial’s robustness and the absence of a control group. Further, outcome of an observational study may be biased by patient attributes that affect treatment selection. Formal clinical trials typically employ randomization to address some of these issues by balancing baseline characteristics among the treatment groups. This open study was intended to consider patients who are observed in the context of neurological disability resulting from cerebral ischemia and undergoing rehabilitation—independent of their gravity. The large patient base provided a range in times from initial ischemic episode until the beginning of treatment with Glialia® and can serve to gauge the evolution of a patient’s disability with treatment in a long-term rehabilitation setting with the same rehabilitation team. We believe that such information may allow one to derive considerations for clinical practice on choice of therapy to combine with rehabilitation for patients with persistent stroke-related neurological disabilities.

nt’s disability with treatment in a long-term rehabilitation setting with the same rehabilitation team. We believe that such information may allow one to derive considerations for clinical practice on choice of therapy to combine with rehabilitation for patients with persistent stroke-related neurological disabilities. Systematic reviews comparing the results of randomized controlled trials and observational studies of the same agents have failed to demonstrate significant differences in outcomes across multiple study designs [76]. In order to address the absence of a control patient group, we compared the results obtained by the CNS, Ashworth Scale, and Barthel Index in patients treated with Glialia® with historical literature [77] data observed in patients having similar pathologic conditions but never receiving Glialia® (Table 5). A z test comparing the mean value reported (standard error was not given) and that of the present study was performed; this comparison is limited to the observed mean value and a fixed value. As such, the test tends to overestimate the probability value and therefore was fixed (alpha = 0.01).Table 5 Comparison of current multicenter study results with reported values for stroke patients not receiving Glialia® Multicenter study Study Aa Study Bb Δc Δc Δc Canadian Neurological Scale 1.80 0.01 (p < 0.0001) 1.02 (p < 0.0001) Ashworth Scale −0.20 – 0.58 (p < 0.0009) Barthel Index 36.2 0.62 (p < 0.0001) 34.8 n.s. aFornasari et al. [77] bCurrent data (C. Cisari) cThe mean value achieved between basal and study end

Systematic reviews comparing the results of randomized controlled trials and observational studies of the same agents have failed to demonstrate significant differences in outcomes across multiple study designs [76]. In order to address the absence of a control patient group, we compared the results obtained by the CNS, Ashworth Scale, and Barthel Index in patients treated with Glialia® with historical literature [77] data observed in patients having similar pathologic conditions but never receiving Glialia® (Table 5). A z test comparing the mean value reported (standard error was not given) and that of the present study was performed; this comparison is limited to the observed mean value and a fixed value. As such, the test tends to overestimate the probability value and therefore was fixed (alpha = 0.01).Table 5 Comparison of current multicenter study results with reported values for stroke patients not receiving Glialia® Multicenter study Study Aa Study Bb Δc Δc Δc Canadian Neurological Scale 1.80 0.01 (p < 0.0001) 1.02 (p < 0.0001) Ashworth Scale −0.20 – 0.58 (p < 0.0009) Barthel Index 36.2 0.62 (p < 0.0001) 34.8 n.s. aFornasari et al. [77] bCurrent data (C. Cisari) cThe mean value achieved between basal and study end These caveats aside, this represents the first description of co-ultraPEALut administration to human stroke patients and clinical improvement not otherwise expected from spontaneous recovery. Based on these observations, we believe that controlled trials are warranted to confirm the utility of co-ultraPEALut to improve clinical outcome in human stroke, also in consideration of the excellent tolerability associated with Glialia® treatment. A double-blind, randomized, and placebo-controlled trial of Glialia® in stroke patients within 12 h of the initial ischemic episode has been initiated.

nfirm the utility of co-ultraPEALut to improve clinical outcome in human stroke, also in consideration of the excellent tolerability associated with Glialia® treatment. A double-blind, randomized, and placebo-controlled trial of Glialia® in stroke patients within 12 h of the initial ischemic episode has been initiated. Functional Recovery and Rehabilitation Centres

nfirm the utility of co-ultraPEALut to improve clinical outcome in human stroke, also in consideration of the excellent tolerability associated with Glialia® treatment. A double-blind, randomized, and placebo-controlled trial of Glialia® in stroke patients within 12 h of the initial ischemic episode has been initiated. Functional Recovery and Rehabilitation Centres F Ventura, IRCSS Az. Osp. Univ. San Martino, Genova; M Casaleggio, ASL3 Polo Levante, Genova; V Leoni, ASL4 Chiavarese, Genova; T Tassinari, ASL2, Pietra Ligure (SV); P Cipolli, ASL1 Imperiese, Bordighera; G Deinite, Casa di Cura Papa Giovanni XXIII, Torino; A Porcella, CTR Ambulatori Riab., Cagliari; S Clemente, Centro di Riabilitazione di Macomer, Nuoro; A Cornaggia, Presidio di Bellano (LC); G Ferlini, CdC Eremo Arco (TN); M Ballotta, ASL2 Lamon-Feltre (BL); G Zara, Azienda Ospedaliera, Padova; E Vincenti, ASL13, Dolo (VE); F Mayer, CdC Ulivella e Glicini, Firenze; R Pecori, CdC Nomentana Hospital, Fonte Nuova (RM); R Quadrini, CdC Villa Sandra, Roma; R Proietti, CdC Città Bianca, Veroli (FR); RM Colognola, ASL VT3, Viterbo; D Topini, CdC Villa Immacolata, San Martino al Cimino (VT); D Tomaccini, Villa S Margherita, Montefiascone (VT); P Matteucci, CdC Villa Immacolata, San Martino al Cimino (VT); P Milia, Ist. Prosperius Tiberino, Umbertide (PG); MM Terracciano, CdC Santa Maria del Pozzo, Somma Vesuviana (NA); R Senatore, Campolongo Hospital, Eboli (SA); S Serra, CdC Istituto S Anna, Crotone; A Boccadamo, CdC Riabilitativa Euroitalia, Casarano (LE); C Lanzillotti, Fond. S Raffaele, Ceglie Messapica (BR); G Di Quarto, CdC Villa Verde, Taranto; G Russo, Fond. S Raffaele Cittadella della Carità, Taranto; P Del Priore, Osp. Santa Maria Bambina, Foggia; F Colonna, CdC Villa Verde, Lecce; G Vastola, Centro Riabilitazione Don M Gala-Fondazione Don Gnocchi, Acerenza (PZ); M Gesualdi, Polo Specialistico Riabilitativo-Fondazione Don Gnocchi, Tricarico (MT); M Spinnato, Villa Margherita, Palermo; P Marano, CdC Villa dei Gerani, Catania; S Russo, CdC Riabilitativa Villa Sofia, Acireale (CT); R Mantia, Ist. Ricerca e Cura Sergio Mantia, Palermo.

M Gala-Fondazione Don Gnocchi, Acerenza (PZ); M Gesualdi, Polo Specialistico Riabilitativo-Fondazione Don Gnocchi, Tricarico (MT); M Spinnato, Villa Margherita, Palermo; P Marano, CdC Villa dei Gerani, Catania; S Russo, CdC Riabilitativa Villa Sofia, Acireale (CT); R Mantia, Ist. Ricerca e Cura Sergio Mantia, Palermo. The authors would like to thank Maria Antonietta Medici for her excellent technical assistance, Mr. Francesco Soraci for his secretarial and administrative assistance, and Miss Valentina Malvagni for her editorial assistance with the manuscript. Compliance with Ethical Standard Ethical Approval All procedures performed in those studies involving human participants were in accordance with the ethical standards of the relevant institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. Informed Consent Informed consent was obtained from all individual participants included in the study. Funding The experimental part of this study was supported in part by MIUR, PON “Ricerca e Competitivita” 2007–2013 project PON01_02512.

Compliance with Ethical Standard Ethical Approval All procedures performed in those studies involving human participants were in accordance with the ethical standards of the relevant institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. Informed Consent Informed consent was obtained from all individual participants included in the study. Funding The experimental part of this study was supported in part by MIUR, PON “Ricerca e Competitivita” 2007–2013 project PON01_02512. Conflict of Interest Carlo Caltagirone declares that he has no conflict of interest. Carlo Cisari declares that he has no conflict of interest. Carlo Schievano declares that he has no conflict of interest. Rosanna Di Paola declares that she has no conflict of interest. Marika Cordaro declares that she has no conflict of interest. Giuseppe Bruschetta declares that he has no conflict of interest. Emanuela Esposito declares that she has no conflict of interest. Salvatore Cuzzocrea is a co-inventor on patent WO2013121449 A8 (Epitech Group Srl) which deals with compositions and methods for the modulation of amidases capable of hydrolyzing N-acylethanolamines useable in the therapy of inflammatory diseases.

“There are many aspects to P3PT that are important including quality standards, assessment criteria, funding, data management, and choice of participating centres. All of these will have to be rigorously thought through, and dealt with before P3PT goes ahead.” – selected statement from an individual commenting on the P3PT concept Introduction The concept of multicenter ‘phase III’ preclinical trials (P3PT) for the evaluation of neuroprotective strategies is suggested to overcome the translational roadblock in stroke research [1]. Importantly, it is not meant to replace exploratory scientific work as usually performed by individual laboratories (‘phases I and II’), but represents a type of confirmative research conducted by collaborating centers [2]. P3PT should be performed prior to early stage clinical investigations and shall contribute to the strongly desired predictive value increase in stroke research. Expert consortia are currently defining potential P3PT frameworks and guidelines, including ways for their future implementation. In parallel, a number of editorials, white papers, and commentaries have reviewed the concept. However, such publications exclusively represent statements by groups of selected experts or renowned individuals. While this provides well-thought through and highly relevant impulses expediting and refining the P3PT idea, it omits the chance to include ideas and feedback from a broader audience.

reviewed the concept. However, such publications exclusively represent statements by groups of selected experts or renowned individuals. While this provides well-thought through and highly relevant impulses expediting and refining the P3PT idea, it omits the chance to include ideas and feedback from a broader audience. The first P3PT has recently been completed [3], but wider adoption and sustained utilization of the P3PT concept requires acceptance throughout the community down to its grassroots, i.e., junior investigators in smaller stroke research laboratories, as well as technicians and students conducting experiments everyday. Hence, the currently ongoing, expert consortia-based design of the P3PT framework might benefit from sensory input from a diverse audience representing the wider stroke research community without balancing towards a specific subgroup.

ch laboratories, as well as technicians and students conducting experiments everyday. Hence, the currently ongoing, expert consortia-based design of the P3PT framework might benefit from sensory input from a diverse audience representing the wider stroke research community without balancing towards a specific subgroup. Collecting a Community Feedback on the P3PT Concept We sought community feedback on the P3PT concept by a public call for an online questionnaire which was announced in a previous publication [4]. The roster contained single (SA) and multiple answer (MA) questions plus five free text answers (see supplementary information). It was hosted by SoSci Survey (Munich, Germany) for 6 months. Answers were assessed for repeated access to avoid bias from counting multiple, but similar and potentially extreme statements. Received information was further subjected to a plausibility check to ensure information consistency (see supplementary material for details). Only those contributions addressing an a priori defined minimum of one survey section were included in the final analysis. An exception was a negative statement regarding general acceptance of the P3PT concept (first question), which was recorded even in case no further question was answered to prevent missing any potential negative statements on the concept. All survey questions, methodological details on the feedback acquisition strategy and data analysis are given in the supplementary material, which also contains complete collection of all free text answers.

ven in case no further question was answered to prevent missing any potential negative statements on the concept. All survey questions, methodological details on the feedback acquisition strategy and data analysis are given in the supplementary material, which also contains complete collection of all free text answers. Of note, the survey was designed as a completely anonymous platform and did not weight individual contributions by the responder’s level of responsibility, experience, or visibility in the field. Nevertheless, the feedback received on the free text answers suggests that a significant proportion of individuals responding to our call are experienced scientists, having profound experience with clinical research, and/or oversee a wide spectrum of research activities. A total of 93 contributions were considered for analysis based on aforementioned plausibility checks, with 81 individuals completing the entire survey.

individuals responding to our call are experienced scientists, having profound experience with clinical research, and/or oversee a wide spectrum of research activities. A total of 93 contributions were considered for analysis based on aforementioned plausibility checks, with 81 individuals completing the entire survey. The Community View on P3PT Organization and Quality Assurance Overall acceptance of the P3PT concept was very high (Fig. 1a). Only 10 % questioned the concept while 90 % acknowledged at least a theoretical benefit. Respondents found a clear and significant overall benefit, recommending to test the concept at a limited scale initially, or assumed it hard to implement but acknowledged its theoretical value (16 % each). The majority (42 %, p < 0.05) requested a careful implementation to ensure maximum benefit. We support this position because careful and potentially stepwise implementation of the concept is warranted in order to investigate which organizational items provide the best balance between practicability and study design complexity, and mitigate the risk of larger failures.Fig. 1 Feedback on P3PT organization and quality assurance as provided by survey participants. a, b provide answer frequency on questions regarding organization while c shows community statements regarding quality standards and their assurance in P3PTs. *p < 0.05

ign complexity, and mitigate the risk of larger failures.Fig. 1 Feedback on P3PT organization and quality assurance as provided by survey participants. a, b provide answer frequency on questions regarding organization while c shows community statements regarding quality standards and their assurance in P3PTs. *p < 0.05 The application of high quality standards seems mandatory to ensure a maximum benefit from P3PT. Accordingly, most participants recommended restricting P3PT projects to centers evidently applying quality standards (40 %, see below for implementation-specific standards), or even to pre-selected labs (33 %, Fig. 1a). This would, however, exclude smaller and less experienced laboratories from P3PT projects at least in initial stages.

participants recommended restricting P3PT projects to centers evidently applying quality standards (40 %, see below for implementation-specific standards), or even to pre-selected labs (33 %, Fig. 1a). This would, however, exclude smaller and less experienced laboratories from P3PT projects at least in initial stages. A centrally determined study protocol for each lab (33 %) was not recommended statistically more often over decentralized monitoring approaches or strict centralized surveillance (Fig. 1b). Most participants preferred a common study design with pre-set endpoints for all participating centers. This indicates a high awareness for standardization within P3PT. However, a majority of participants (44 %) also advocated for a design which, next to addressing centrally determined endpoints according to a common plan, allows additional endpoints to be investigated by individual labs (p < 0.05; Fig. 1b). This suggestion is remarkable since it offers an option to capitalize on benefits from a centralized study organization (e.g., reliable study results on primary endpoints) without omitting the possibility to receive valuable secondary endpoint data by utilizing individual lab competencies. Although the assessment of such secondary endpoints is statistically less powerful due to the smaller number of subjects/cases investigated, overall information content and translational relevance are likely to be increased. Importantly, this also underpins the importance of academic centers as the main stakeholders in P3PTs because other well organized and equipped entities such as commercial contract research organizations often cannot offer a similar diversity of available methods and readout assays. No clear preference was given for data analysis, but centralized data analysis (29 %) or at least analysis surveillance (27 %) were selected most often (Fig. 1b).

ized and equipped entities such as commercial contract research organizations often cannot offer a similar diversity of available methods and readout assays. No clear preference was given for data analysis, but centralized data analysis (29 %) or at least analysis surveillance (27 %) were selected most often (Fig. 1b). Application of minimum quality assurance criteria plus a pre-defined experimental plan applying to each lab (39 %), or even restricting participating labs to those applying ARRIVE (http://www.nc3rs.org/ARRIVE) or STAIR [5] criteria recommendations (31 %; p < 0.05 each), were considered superior to intra-lab quality assurance (Fig. 1c). Official agreement on quality assurance criteria by all partners (43 %) and standardized experimenter training (38 %) were superior to all other options, including pre-study round robin trials for quality check-up (13 %; p < 0.05 each; Fig. 1c).

each), were considered superior to intra-lab quality assurance (Fig. 1c). Official agreement on quality assurance criteria by all partners (43 %) and standardized experimenter training (38 %) were superior to all other options, including pre-study round robin trials for quality check-up (13 %; p < 0.05 each; Fig. 1c). P3PT Methodology Each aspect suggested for intra-experimental quality assurance was at least recommended by 60 %, with blinding/randomization (69 %/72 %) and definition of exclusion/inclusion criteria (80 %) being selected most frequently. All were considered superior to omitting standardization (p < 0.05; Fig. 2a). Appropriate positive/negative controls were recommended most frequently for the experimental design (76 %; Fig. 2a), corroborating a clear community affirmation to quality assurance principles.Fig. 2 Feedback P3PT methodology, financing and result publication. a to c provide answer frequencies on questions regarding numerous central aspects of P3PT methodology. Abbreviations are as follows: HTN hypertension, DM diabetes mellitus, HL hyperlipidemia, d distal, e embolic, f filament middle cerebral artery occlusion (MCAO). d shows preferred options regarding P3PT funding and results publication. *p < 0.05

ncies on questions regarding numerous central aspects of P3PT methodology. Abbreviations are as follows: HTN hypertension, DM diabetes mellitus, HL hyperlipidemia, d distal, e embolic, f filament middle cerebral artery occlusion (MCAO). d shows preferred options regarding P3PT funding and results publication. *p < 0.05 However, 31 % did not put special emphasis on blinding and/or randomization. This is a striking finding since (i) neglecting these aspects has been discussed to contribute to the translational failure in stroke research and (ii) blinding to avoid bias is an essential, uniformly adopted approach in clinical trials. The picture was also less consistent when it came to monitoring of important physiological parameters. Recording of body weight and temperature during surgery was found recommendable by most responders (80 %) while cerebral blood flow (CBF) monitoring to ensure stroke induction was not considered as important by 38 % of the participants. Although stroke induction can be monitored by alternative methods such as magnetic resonance imaging, those are technically more complex and less widely available. From our perspective, this indicates a need for a broader awareness of the necessity to thoroughly control stroke induction. Even fewer participants (53 %) considered arterial blood gases/pH monitoring during surgery to be necessary despite recommendations that thorough monitoring of blood chemistry parameters is critical to ensure result reliability in stroke research [5]. In summary, considering these key methodological aspects will be critical to ensure maximum predictability of P3PT studies. This likely requires increased awareness throughout the community or even mandatory.

rough monitoring of blood chemistry parameters is critical to ensure result reliability in stroke research [5]. In summary, considering these key methodological aspects will be critical to ensure maximum predictability of P3PT studies. This likely requires increased awareness throughout the community or even mandatory. Testing of therapeutic efficacy in aged versus young (76 %), male versus female (75 %), and in comorbid (73 %) rodent species were recommended most often to enhance the predictive value of P3PT (p < 0.05 each; Fig. 2c). Surprisingly, only 55 % considered investigation of at least two species, a central recommendation of the STAIR expert consortium [5]. Conducting large animal experiments (39 %) and considering polypharmacy (37 %) were recommended least often. This is understandable given the complexity of appropriate model systems and their limited availability. Nevertheless, studies utilizing large animal and polypharmacy models are important for a number of reasons. Large animal models may provide an additional benefit in the assessment of novel stroke therapies with respect to brain anatomy (gyrencephalic species) and potential distribution aspects of a pharmacological treatment (larger brain) [6] with primate models being the main expert recommendation [7]. Polypharmacy is a frequent observation in human stroke patients and an interaction between an experimental therapeutic and the patient’s medication is relatively likely [2]. Being hard to investigate by most single centers, P3PT may offer a practicable framework to address these aspects. The same applies to the heterogeneity of strokes often seen in clinical trials but rarely represented in experimental studies. Consequently, the use of multiple stroke models to reflect patient heterogeneity was found more useful than the reliance on any specific stroke model (67 %; p < 0.05; Fig. 2c).

s these aspects. The same applies to the heterogeneity of strokes often seen in clinical trials but rarely represented in experimental studies. Consequently, the use of multiple stroke models to reflect patient heterogeneity was found more useful than the reliance on any specific stroke model (67 %; p < 0.05; Fig. 2c). Special emphasis was given on post-stroke care and preset exclusion/inclusion criteria with half of the participants also pointing at factors such as nutrition and fluid supply (Fig. 2c). Accordingly, participating colleagues also called for mimicking clinically realistic scenarios such as combination with tPA, body weight/surface-based dosing, and i.v. administration of the therapeutic agent when testing a therapeutic paradigm. Using same dose in all animals (3 %) or selecting administration protocols promising best effect size (28 %) were considered inferior approaches (p < 0.05; Fig. 2c).

rios such as combination with tPA, body weight/surface-based dosing, and i.v. administration of the therapeutic agent when testing a therapeutic paradigm. Using same dose in all animals (3 %) or selecting administration protocols promising best effect size (28 %) were considered inferior approaches (p < 0.05; Fig. 2c). P3PT Financing and Publication Financing and publication represent challenges in large scale preclinical studies. Although we expected a heterogeneous opinion spectrum, clear statements were provided by the community. The most popular option for publication was that study initiators should cover first and/or senior author positions and/or invite members of the writing committee, with all other experimenters listed as co-authors/contributors (43 %; p < 0.05; Fig. 2d). This supports the idea of forming writing committees as proven useful for large-scale clinical trials. Importantly, balancing transparency versus justified background interests (patents, technological knowhow) will require individual solutions, which should be informed by good practice in clinical trials.

05; Fig. 2d). This supports the idea of forming writing committees as proven useful for large-scale clinical trials. Importantly, balancing transparency versus justified background interests (patents, technological knowhow) will require individual solutions, which should be informed by good practice in clinical trials. Using (still non-existing) global funding schemes was found to be the most appropriate approach for P3PT financing (86 %; p < 0.05; Fig. 2d). This is consistent with the fact that stroke is a global burden requiring mobilization of global resources to counter it. Requesting industrial support was recommended by 63 % and is warranted since the pharmaceutical industry is expected to benefit, e.g., from concept falsification by P3PTs prior to significant reputational and financial losses in failed clinical trials [8]. Nevertheless, realization of global funding mechanisms may be hard to achieve while potential conflicts and intellectual property issues arising within academic-industrial partnerships demand careful consideration. Collaboration between leaders from basic research, industry, and regulatory authorities as previously proven beneficial [5] may be required to orchestrate P3PT realization.

to achieve while potential conflicts and intellectual property issues arising within academic-industrial partnerships demand careful consideration. Collaboration between leaders from basic research, industry, and regulatory authorities as previously proven beneficial [5] may be required to orchestrate P3PT realization. Information Derived from Free Text Answers When categorizing answers regarding content and counting for answers referring to the particular aspect, three major points of interested within the participant community became evident: (i) implementation of very high quality standards in P3PT (n = 7), (ii) careful selection of endpoints, models, and participating labs while at the same time ensuring high interaction among P3PT participants (n = 10) as well as (iii) organizing P3PT close to the design of late stage clinical trials (n = 11) was particularly stressed.

ion of very high quality standards in P3PT (n = 7), (ii) careful selection of endpoints, models, and participating labs while at the same time ensuring high interaction among P3PT participants (n = 10) as well as (iii) organizing P3PT close to the design of late stage clinical trials (n = 11) was particularly stressed. Limitations of the Survey As with any voluntary participation in opinion surveys, scientists who are less supportive of the P3PT concept may not have taken the time to contribute their thoughts to our or other initiatives and thus potentially bias our results. However, it is equally possible that those who are not supportive of the P3PT idea could have taken this anonymous opportunity to express their opinions by answering the survey. We did not collect information regarding respondents’ positions (e.g., senior vs. junior) or their actual participation in preclinical stroke research and drug testing. This impedes weighting of feedbacks and positions provided with respect to experience and level of expertise of the respective respondent. On the other hand, the strictly anonymous nature may have helped to receive a broader feedback from the stroke community, which is critical for the acceptance and large-scale implementation of the P3PT concept.

dbacks and positions provided with respect to experience and level of expertise of the respective respondent. On the other hand, the strictly anonymous nature may have helped to receive a broader feedback from the stroke community, which is critical for the acceptance and large-scale implementation of the P3PT concept. Conclusions Despite its limitations, our survey provides profound feedback from a considerable number of individual respondents. Their feedback encourages further steps implementing P3PT studies into translational research strategies, but also highlights controversies around specific aspects of its implementation. Based on the analysis of received answers, we suggest drawing the following conclusions and recommendations with respect to five core areas (Table 1).Table 1 Conclusions and recommendations for the implementation of P3PT based on community feedback analysis

s controversies around specific aspects of its implementation. Based on the analysis of received answers, we suggest drawing the following conclusions and recommendations with respect to five core areas (Table 1).Table 1 Conclusions and recommendations for the implementation of P3PT based on community feedback analysis Area Conclusion/recommendation Benefit P3PT implementation Careful and stepwise implementation recommended Widespread acceptance more likely Initial preclinical multicenter studies should be performed by experienced centers, ideally having a long-standing history of collaboration Swift and exact estimation of P3PT benefit under practical conditions P3PT organization and governance Centralized study governance and central study protocol Clinical-trial like design, enhanced result comparability Core endpoints addressed by all centers according to P3PT protocol Enhanced statistical power and higher predictability for primary endpoints Additional: individual endpoints addressed by single centers with outstanding competencies Broad spectrum of translationally endpoints addressable (but no benefit for study power) P3PT animal models Use of multiple models, if applicable Better representation of patient population (polypharmacy, age, sex, comorbidities) Use of large animal models, if available Reflecting gyrencephalic brain structure, closer similarity to human situation P3PT quality assurance Further enhancing awareness for those (might require institutional support [9]) Increasing scientific rigor and result comparability, reducing divergences in relevance acknowledgment throughout the community Strict application of quality assurance criteria Enhanced result comparability and relevance P3PT financing and result publication Establishing and recruiting of global funds International and -continental collaboration facilitated, reduced financial burden for national public funding authorities Early enrolment of high-quality academic-industry collaborations Timely involvement of key stakeholders, preventing failure of clinical trials, bolstering financial resources for P3PT Establishing centralized writing committees More efficient workflow, comparability to large scale clinical trials

ic funding authorities Early enrolment of high-quality academic-industry collaborations Timely involvement of key stakeholders, preventing failure of clinical trials, bolstering financial resources for P3PT Establishing centralized writing committees More efficient workflow, comparability to large scale clinical trials Since the P3PT idea can only be successful if supported by the entire stroke research community, input from other initiatives and expert committees such as the MULTIPART project (http://www.dcn.ed.ac.uk/multipart/default.htm) responding to our results are essential and helpful to shape P3PT implementation and to ensure its maximum benefit and impact. Electronic Supplementary Material Below is the link to the electronic supplementary material.ESM 1 (DOCX 34.2 kb) Nikolaus Plesnila and Cenk Ayata contributed equally to this work. We thank Drs. Antonia Weingart and Jürgen Peters, University of Munich Medical School for online survey maintenance and expert statistical counseling, respectively. Author Contribution Statement JB, CA, and NP designed the survey. DCW and NH provided valuable input to improve its design and content. JB, CA, and NP conducted the statistical analysis of the obtained information and all authors contributed to their interpretation. JB, CA, and NP wrote the initial manuscript draft. All authors corrected and improved the draft, and approved the final manuscript version. Compliance with Ethical Standards Conflict of Interest The authors declare that they have no competing interests. Funding Only institutional funds were used for this project.

Introduction Stroke is the second leading cause of death globally and composed of hemorrhagic and ischemic stroke, with the latter being more common and resulting from disruption of blood supply to the brain [1, 2]. Thrombolysis with tissue plasminogen activator (tPA) and mechanical thrombectomy [3, 4] is effective at recanalization of an occluded artery, thus restoring cerebral circulation. However, reperfusion upon spontaneous or medically induced recanalization after cerebral ischemia can cause tissue injury which in turn contributes to worsening neurological status as well as increased morbidity and mortality in patients with acute ischemic stroke [5]. Reperfusion injury consists of a multi-step cascade with a wide range of mechanisms, including disturbance of protein synthesis, oxidative stress, platelet activation, inflammatory immune responses, disruption of the blood–brain barrier (BBB), glial activation, and neuronal apoptosis [6–8]. Some pharmacological agents have been studied to address ischemia–reperfusion injury by blocking reactive oxygen species (ROS) and neuronal excitotoxicity [9, 10]. However, these agents have thus far failed to demonstrate efficacy in clinical trials [11, 12]. The complexity of the ischemia–reperfusion biological cascade, inadequate dosing of antioxidants, inappropriate targeting by antioxidants, and most importantly the narrow time window (from minutes to approximately 3 h) for targeting oxidative stress in clinical settings are among the possible reasons that have led to the failure of these clinical trials [13–15]. It remains an urgent need to develop neuroprotective strategies targeting multiple key steps in the biochemical cascade of ischemia–reperfusion injury.

to approximately 3 h) for targeting oxidative stress in clinical settings are among the possible reasons that have led to the failure of these clinical trials [13–15]. It remains an urgent need to develop neuroprotective strategies targeting multiple key steps in the biochemical cascade of ischemia–reperfusion injury. Dimethyl fumarate (DMF), derived from fumaric acid esters (FAE), represents a class of molecules exhibiting a multitude of biological effects including anti-oxidative stress and anti-apoptotic and immunomodulatory properties as well as providing protection from microvascular dysfunction in a variety tissues [16]. DMF exerts immunomodulatory effects on T cell subsets, glial cells, via the reduction of proinflammatory cytokines such as IL-2, TNF-α, and ICAM in inflammatory cascades. DMF stabilizes and activates the transcription factor nuclear factor (erythroid-derived 2)-like 2 (Nrf2), regulating many target genes such as HO-1, quinone oxidoreductase 1 (NADPH), and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) [17, 18]. DMF is approved for the treatment of relapsing multiple sclerosis in the US and European countries.

factor nuclear factor (erythroid-derived 2)-like 2 (Nrf2), regulating many target genes such as HO-1, quinone oxidoreductase 1 (NADPH), and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) [17, 18]. DMF is approved for the treatment of relapsing multiple sclerosis in the US and European countries. In a Parkinson’s disease model, DMF has been shown to ameliorate dopaminergic neurotoxicity [19]. In intracerebral hemorrhage animal models, DMF induced Nrf2 target genes, reduced cerebral edema and inflammation, improved hematoma resolution, and enhanced neurological recovery [20, 21]. Using a preconditioning, acute ischemic stroke model, Kunze et al. demonstrated that prophylactic treatment with DMF did not change the size of cerebral infarct, but was able to attenuate edema formation [22]. The unpredictable onset of stroke extremely limits the prophylactic use of DMF in clinical practice. The immediate impact of DMF on acute ischemic stroke and its post-conditioning role in treating acute ischemic stroke have not been assessed.

the size of cerebral infarct, but was able to attenuate edema formation [22]. The unpredictable onset of stroke extremely limits the prophylactic use of DMF in clinical practice. The immediate impact of DMF on acute ischemic stroke and its post-conditioning role in treating acute ischemic stroke have not been assessed. In the present study, we examined a potential therapeutic role for DMF in the acute and subacute stages following middle cerebral artery occlusion (MCAO) and reperfusion injury. Given the significant biological difference between DMF and its major metabolite, monomethyl fumarate (MMF) [23, 24], comparisons between DMF and MMF were carried out. Our data demonstrated that DMF and MMF exert their protective role by reducing infarct size during the subacute stage of stroke, but not immediately following MCAO. This protection is likely due to increased Nrf2 activity.

metabolite, monomethyl fumarate (MMF) [23, 24], comparisons between DMF and MMF were carried out. Our data demonstrated that DMF and MMF exert their protective role by reducing infarct size during the subacute stage of stroke, but not immediately following MCAO. This protection is likely due to increased Nrf2 activity. Materials and Methods Animals Nrf2 knockout mouse line was generously provided by Dr. Thomas W. Kensler of the University of Pittsburgh [25]. Mice were backcrossed to the C57BL/6 background for more than 10 generations. Heterozygous (Nrf2+/−) mice were used to produce homozygous (Nrf2–/–) and wild-type (WT) littermates. Animals (20 to 25 g, 8 to 10 weeks old) had access to food and water ad libitum and were housed under controlled conditions (23 ± 2 °C, 12-h light/dark periods). Adequate measures were taken to minimize the number of experiment animals used and to ensure minimal pain or discomfort in animals. All mice were randomly assigned to the different experimental groups. Animal exclusion criteria were as follows: mice died within the observation period and subarachnoid or intracerebral hemorrhage macroscopically or by magnetic resonance imaging were excluded. All animal experiments and procedures were approved by the Animal Experiments Ethical Committee of Tianjin Medical University General Hospital.

n criteria were as follows: mice died within the observation period and subarachnoid or intracerebral hemorrhage macroscopically or by magnetic resonance imaging were excluded. All animal experiments and procedures were approved by the Animal Experiments Ethical Committee of Tianjin Medical University General Hospital. Middle Cerebral Artery Occlusion Model Focal cerebral ischemia was modeled by occluding the left middle cerebral artery (MCAO), based on the methods described by Longa et al. [26]. Briefly, the mice were anesthetized with chloral hydrate (30 mg/kg, intraperitoneal injection). A midline neck incision was then made to expose the left common carotid artery, the external carotid artery, and the internal carotid artery, which were all then isolated and ligated. A monofilament coated with silicone rubber (Xinong, 1418A, Beijing, China) was inserted into the internal carotid artery (9–10 mm) through the common carotid artery, to the beginning of the middle cerebral artery (MCA). A laser Doppler approach was used to monitor MCA occlusion and reperfusion as we previously described [27]. For this procedure, a small incision was made in the skin overlying the temporalis muscle and a 0.7-mm flexible laser Doppler probe (model P10) was positioned on the superior portion of the temporal bone (6 mm lateral and 2 mm posterior from the bregma). One hour after the induction of ischemia, the monofilament was removed to restore blood flow. Relative cerebral blood flow had to rise to at least 50 % of preischemic levels for the mice to be included in the study and subjected to further analyses. The body temperature of the mice was maintained at 37.0 ± 0.5 °C during surgery, and the mice were kept in a well-ventilated room at 25 ± 3 °C in individual cages, with the provision of food and water, until they regained full consciousness.

ls for the mice to be included in the study and subjected to further analyses. The body temperature of the mice was maintained at 37.0 ± 0.5 °C during surgery, and the mice were kept in a well-ventilated room at 25 ± 3 °C in individual cages, with the provision of food and water, until they regained full consciousness. DMF and MMF Administration DMF (Sigma-Aldrich, Steinheim, Germany) was dissolved in 10 % dimethyl sulfoxide (DMSO). Given the evidence of a dose-dependent effect in antioxidant strategy in general [13] and with DMF [28–30], we performed dose-finding experiments with oral gavage administration of DMF in a mouse MCAO model (Supplemental Fig. 1). Subsequently, DMF was given at 30 or 45 mg/kg body weight, twice a day, for seven consecutive days with the first dose given 15 min before reperfusion (Fig. 1a). MMF (Sigma-Aldrich, Steinheim, Germany) was dissolved in 0.01 M phosphate-buffered saline (PBS) and administered intraperitoneally (i.p.) at a dosage of 30 or 45 mg/kg body weight, twice a day, for seven consecutive days with the first dose given 15 min before reperfusion (Fig. 1a).Fig. 1 DMF and MMF improve neurological deficits and reduce brain infarct volume and cerebral edema in mice with MCAO. a Schematic experimental design of DMF and MMF treatment for acute ischemic cerebral stroke. b Administration of DMF or MMF reduces infarct volume determined by 2,3,5-tripenyltetrazolium chloride (TTC) staining. Brain sections from a representative mouse are shown from each group (vehicle control, DMF 30, 45 mg and MMF 30, 45 mg). Each column shows four TTC-stained coronal brain slices arranged in cranial to caudal order. c DMF and MMF treatments reduce the infarct volume due to MCAO during the subacute phase. Mice in the DMF- and MMF-treated groups had smaller volume of infarct on days 3 and 7 post-ischemia–reperfusion injury. Data shown are mean ± SEM at day 3; *p < 0.05, **p < 0.01 as compared to the vehicle-treated group; ##p < 0.01, as compared between the DMF 45 mg group and DMF 30 mg group, or MMF 45 mg and MMF 30 mg group, n = 10 per group. d Neurobehavioral improvement in mice with MCAO upon DMF or MMF treatment is time- and dose-dependent. Significant improvement in mNSS scores is evident in mice treated with 45 mg/kg DMF or MMF (p < 0.01) as compared to mice in the vehicle control group on day 3 and day 7, but not on day 1 post-ischemia–reperfusion injury.

urobehavioral improvement in mice with MCAO upon DMF or MMF treatment is time- and dose-dependent. Significant improvement in mNSS scores is evident in mice treated with 45 mg/kg DMF or MMF (p < 0.01) as compared to mice in the vehicle control group on day 3 and day 7, but not on day 1 post-ischemia–reperfusion injury. A lesser degree of improvement is observed in mice treated with DMF or MMF at a dose of 30 mg/kg body weight. Mean ± SEM; **p<0.01, as compared to those in the vehicle control group. e DMF and MMF attenuate cerebral edema associated with transient MCAO in a dose-dependent manner. Data depicted here are mean ± SEM; *p < 0.05 and **p < 0.01 as compared to the vehicle-treated group; ##p < 0.01 as compared between DMF 45 mg/kg and DMF 30 mg/kg group or MMF 45 mg/kg and MMF 30 mg/kg group; n = 10 per group To discern a potential post-ischemic role for both DMF and MMF, WT mice were divided into five groups: vehicle (MCAO + PBS, n = 50), DMF 30 mg/kg (MCAO + DMF 30 mg/kg, n = 50), DMF 45 mg (MCAO + DMF 45 mg/kg, n = 50), MMF 30 mg/kg (MCAO + MMF 30 mg/kg, n = 50), and MMF 45 mg/kg (MCAO + MMF 45 mg/kg, n = 50). Assessments included neurobehavioral testing and infarct volume assessment using MRI in vivo and 2,3,5-tripenyltetrazolium chloride (TTC) staining in vitro at days 1, 3, and 7 post-ischemia as well as brain edema and tissue pathology (Fig. 1a)

CAO + MMF 30 mg/kg, n = 50), and MMF 45 mg/kg (MCAO + MMF 45 mg/kg, n = 50). Assessments included neurobehavioral testing and infarct volume assessment using MRI in vivo and 2,3,5-tripenyltetrazolium chloride (TTC) staining in vitro at days 1, 3, and 7 post-ischemia as well as brain edema and tissue pathology (Fig. 1a) To test whether the Nrf2/HO-1 pathway mediates the effects provided by DMF and MMF in cerebral ischemia reperfusion injury, Nrf2−/− mice with MCAO were randomly divided into three groups: control (Nrf2−/− + PBS, n = 18), DMF 45 mg/kg (Nrf2−/− + DMF 45 mg/kg, n = 18), and MMF 45 mg/kg (Nrf2−/− + DMF 45 mg/kg, n = 18). Assessments included neurobehavioral testing as well as brain edema and tissue pathology assessment (Fig. 1a). The arbitrary time points after MCAO in this study represent the following stages of stroke [27, 31]: <1 day, acute stage; 2–7 days, subacute stage. Neurobehavioral Monitoring A battery of neurobehavioral tests was performed before MCAO and on days 1, 3, and 7 after MCAO induction by two investigators who were blinded to the experimental group assignment. These tests were summarized and expressed as the modified neurological severity score (mNSS), a composite of motor, sensory, reflex, and balance tests. Neurological function was graded on a scale of 0–18 as previously described [32], with the higher score, the more severe impairment from the ischemia–reperfusion injury.

hese tests were summarized and expressed as the modified neurological severity score (mNSS), a composite of motor, sensory, reflex, and balance tests. Neurological function was graded on a scale of 0–18 as previously described [32], with the higher score, the more severe impairment from the ischemia–reperfusion injury. 2,3,5-Tripenyltetrazolium Chloride Assessment of Infarct Size Brain tissues of mice were obtained at days 1, 3, and 7 post-ischemia and immediately sliced into coronal sections (2 mm thick) from the rostral to the caudal frontal tip using scalpels. The sections were stained with 1.5 % TTC (Sigma-Aldrich, USA), followed by immersion in normal saline at 37 °C for 20 min. Brain sections were then fixed in 4 % paraformaldehyde at 4 °C overnight before being photographed. With this staining method, viable tissues stain deep red based on intact mitochondrial function, while infarcts remain white. The infarcted regions in each section were evaluated using Image-Pro® Plus v 4.0 image analysis software (Media Cybernetics, Washington, DC, USA). The total infarct volume was calculated as the sum of the infarct volume of each section. The infarct volume percentage was calculated as follows: ([total contralateral hemispheric volume] − [total ipsilateral hemispheric stained volume])/(total contralateral hemispheric volume) × 100 % [33].

ics, Washington, DC, USA). The total infarct volume was calculated as the sum of the infarct volume of each section. The infarct volume percentage was calculated as follows: ([total contralateral hemispheric volume] − [total ipsilateral hemispheric stained volume])/(total contralateral hemispheric volume) × 100 % [33]. Magnetic Resonance Imaging Magnetic resonance imaging (MRI) data were acquired using a 3.0-Tesla clinical MR system (Discovery MR750, General Electric, Milwaukee, WI, USA), using an eight-channel phased-array head coil. Coronal T2-weighted turbo spin echo (10 continuous slices, repetition time (TR) = 2980 ms, echo time (TE) = 78 ms, section thickness = 1 mm, in-plane resolution 1 mm2) was used for lesion detection. The lesion borders on each slice were traced manually by using Image-Pro Plus (Media Cybernetics, Rockville, MD, USA) to measure the lesion area, then summed and multiplied by the slice thickness to determine the lesion volume. Brain Edema Measurement At day 3 after MCAO, mice were sacrificed and brains were harvested. Each brain was carefully divided into three parts as left hemisphere, right hemisphere, and cerebellum. Tissues were quickly weighed on an electronic analytical balance to obtain the wet weight. The brain tissues were then dried for 72 h at 100 °C to obtain the dry weight. Brain water content calculation was achieved using the following formula: (wet weight − dry weight) / wet weight × 100 %.

hemisphere, and cerebellum. Tissues were quickly weighed on an electronic analytical balance to obtain the wet weight. The brain tissues were then dried for 72 h at 100 °C to obtain the dry weight. Brain water content calculation was achieved using the following formula: (wet weight − dry weight) / wet weight × 100 %. Immunostaining The extent of cell death was assessed using a terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) kit (Roche, USA). Once stained, the specimens were analyzed under a florescence microscope (Nikon C-HGFI, Japan). The total number of nuclei and TUNEL-positive cells were counted in four random fields of a ×20 view of the edge of the infarct, and the ratio of apoptotic cells to nuclei was calculated as apoptotic cells in percent. Paraffin sections from mouse brains sacrificed 24 h after MCAO, cut to a thickness of 5 μm, were first deparaffinized in xylene and subsequently rehydrated with various grades of ethanol. After rehydrating the sections, nonspecific binding was blocked by incubating the sections in 10 % bovine serum albumin for 20 min. The sections were then incubated overnight at room temperature with anti-glial fibrillary acidic protein (GFAP) or anti-ionized calcium-binding adapter molecule (Iba-1) antibodies to identify astrocytes and microglia, respectively. Finally, mounting media containing DAPI was applied and a coverslip was placed over the sections. The stained sections were examined and analyzed with a fluorescence microscope (Olympus, Tokyo, Japan).

r anti-ionized calcium-binding adapter molecule (Iba-1) antibodies to identify astrocytes and microglia, respectively. Finally, mounting media containing DAPI was applied and a coverslip was placed over the sections. The stained sections were examined and analyzed with a fluorescence microscope (Olympus, Tokyo, Japan). Determination of Indicators of Oxidative Stress Brain tissues were collected at day 3 after MCAO. Superoxide dismutase (SOD) activity as well as the levels of malondialdehyde (MDA) and glutathione (GSH) were measured as indicators of oxidative stress [34]. The brains were washed, weighed, and then homogenized in ice-cold saline (nine volumes) for 20 min to prepare a 10 % (w/v) homogenate. The homogenate was then centrifuged at 4000 rpm/min for 10 min at 4 °C. SOD activity and levels of MDA and GSH were measured as described previously [34, 35]. Data were calculated in reference to the protein concentration in each sample.

d in ice-cold saline (nine volumes) for 20 min to prepare a 10 % (w/v) homogenate. The homogenate was then centrifuged at 4000 rpm/min for 10 min at 4 °C. SOD activity and levels of MDA and GSH were measured as described previously [34, 35]. Data were calculated in reference to the protein concentration in each sample. Western Blot Analysis On day 3 post-MCAO, ipsilateral hemispheres were homogenized in RIPA lysis buffer (Sigma, USA) and 1 mmol/L phenylmethanesulfonyl fluoride (PMSF) (Sigma, USA). After centrifugation, the supernatants were collected as total proteins. Proteins were loaded and transferred to a PVDF membrane (Millipore, USA). After being blocked, membranes were incubated overnight at 4 °C with anti-Nrf2 (1:1000, Millipore), anti-HO-1 (1:1000, Millipore), or rabbit polyclonal antibodies. Membranes were then incubated for 1 h at room temperature with horseradish peroxidase (HRP)-labeled goat anti-rabbit secondary antibody (1:4000, Vector, Burlingame, USA). The membranes were placed into a gel imaging system (Bio-Rad, ChemiDoc XRS, USA) and then exposed. The intensity of blots was quantified using the Quantity One Analysis software (Bio-Rad, USA). β-Actin was used as an internal control.

(HRP)-labeled goat anti-rabbit secondary antibody (1:4000, Vector, Burlingame, USA). The membranes were placed into a gel imaging system (Bio-Rad, ChemiDoc XRS, USA) and then exposed. The intensity of blots was quantified using the Quantity One Analysis software (Bio-Rad, USA). β-Actin was used as an internal control. RNA Isolation and Real-Time PCR Total RNA was extracted from the ischemic hemisphere using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions at day 3 after MCAO. The concentration of RNA was quantified by ultraviolet spectrophotometry at 260/280 nm. cDNA was transcribed using TransScript First-Strand cDNA Synthesis SuperMix Kit (TransGen; catalog no. AT301) in accordance with the manufacturer’s instructions. PCR was performed on the Opticon 2 Real-Time PCR Detection System (Bio-Rad) using the following primers: Nrf2 forward GGTTGCCCACATTCCCAAAC, Nrf2 reverse TCCTGCCAAACTTGCTCCAT; HO-1 forward CGACAGCATGTCCCAGGATT, HO-1 reverse CTGGGTTCTGCTTGTTTCGC; and β-actin forward AAATCGTGCGTGACATCAAAGA, β-actin reverse GGCCATCTCCTGCTCGAA; SYBR Green PCR Master Mix (Roche) was also used. Samples were run in duplicate and normalized to β-actin using the 2–ΔΔCt method. The expression levels of the messenger RNAs (mRNAs) were then reported as fold changes vs control.

CTTGTTTCGC; and β-actin forward AAATCGTGCGTGACATCAAAGA, β-actin reverse GGCCATCTCCTGCTCGAA; SYBR Green PCR Master Mix (Roche) was also used. Samples were run in duplicate and normalized to β-actin using the 2–ΔΔCt method. The expression levels of the messenger RNAs (mRNAs) were then reported as fold changes vs control. Statistical Analysis All data were analyzed by SPSS 18.0 software and expressed as mean ± SEM. Sample size per group was determined using a priori sample size calculation (G*Power version 3.1). To achieve α = 0.05 at β = 0.2 (power 80 %) with a mean 20 % standard deviation, results from sample size calculation show that n = 6–10 mice per group was appropriate. Statistical differences were measured by unpaired two-tailed Student t test for comparison of two groups or ANOVA followed by Bonferroni post hoc test for multiple group comparisons. Values of p < 0.05 were considered significant.

, results from sample size calculation show that n = 6–10 mice per group was appropriate. Statistical differences were measured by unpaired two-tailed Student t test for comparison of two groups or ANOVA followed by Bonferroni post hoc test for multiple group comparisons. Values of p < 0.05 were considered significant. Results Improvement in Neurological Deficits at Subacute Stage in MCAO Mice Treated with DMF or MMF On day 1 post-MCAO ischemia–reperfusion injury, there was no statistical difference in mNSS between groups of MCAO mice treated with vehicle control, DMF (30, 45 mg/kg), or MMF (30, 45 mg/kg) groups. Surprisingly, on day 3 post-MCAO ischemia–reperfusion injury, mice in both of the DMF (30 and 45 mg/kg)-treated groups exhibited significantly reduced mNSS. A similar impact was evident in mice receiving MMF (30 and 45 mg/kg) treatment. The improvement of neurobehavioral function with DMF or MMF treatment was dose-dependent with a better outcome in mice treated with the relatively higher dose of DMF (45 mg/kg) or MMF (45 mg/kg) as compared with those treated with the lower dose of DMF (30 mg/kg) or MMF (30 mg/kg) body weight, respectively (Fig. 1d). Further, the reduction of mNSS was sustained on day 7 post-MCAO ischemia–reperfusion injury.

th a better outcome in mice treated with the relatively higher dose of DMF (45 mg/kg) or MMF (45 mg/kg) as compared with those treated with the lower dose of DMF (30 mg/kg) or MMF (30 mg/kg) body weight, respectively (Fig. 1d). Further, the reduction of mNSS was sustained on day 7 post-MCAO ischemia–reperfusion injury. DMF and MMF Reduce Infarct Volume The infarct volume of brain tissue was measured on days 1, 3, and 7 post-MCAO ischemia–reperfusion injury with TTC staining (Fig. 1b). On day 1 post-MCAO, there was no statistical difference in the volume of the infarct in mice treated with vehicle and DMF (30, 45 mg per kg body weight) (Supplemental Fig. 2). However, a significant reduction of infarct volume was observed in MCAO mice treated with DMF on day 3 post-MCAO. The reduction of infarct volume with DMF treatment was sustained in mice on day 7 post-MCAO (Fig. 1b, c). A dose effect was observed with the higher dose of DMF at 45 mg/kg body weight having a greater effect on the infarct volume reduction. To further confirm the protective role of DMF, we examined its primary metabolite MMF. A similar impact on acute ischemic infarct volume was observed with MMF in a dose-dependent manner (Fig. 1b, c). Administration of DMF (30, 45 mg/kg) or MMF (30, 45 mg/kg) significantly decreased the percentage of the infarct volume from 49.36 ± 5.33 % in the vehicle group to 39.69 ± 6.65 % (p < 0.05), 30.64 ± 2.85 % (p < 0.01), 38.05 ± 3.60 % (p < 0.01), or 28.21 ± 2.58 % (p < 0.01) on day 7 post-ischemia–reperfusion injury, respectively.

tion of DMF (30, 45 mg/kg) or MMF (30, 45 mg/kg) significantly decreased the percentage of the infarct volume from 49.36 ± 5.33 % in the vehicle group to 39.69 ± 6.65 % (p < 0.05), 30.64 ± 2.85 % (p < 0.01), 38.05 ± 3.60 % (p < 0.01), or 28.21 ± 2.58 % (p < 0.01) on day 7 post-ischemia–reperfusion injury, respectively. Brain Edema The brain water content (BWC) in mice of the vehicle group was 19.6 ± 2.5 %, while the DMF-treated (30, 45 mg/kg) and MMF-treated (30, 45 mg/kg) groups had decreased BWC of 11.47 ± 2.3 % (p < 0.05, as compared with vehicle group), 8.1 ± 1.9 % (p < 0.01), 11.42 ± 3.52 % (p < 0.05), or 6.42 ± 2.51 % (p < 0.01), respectively (Fig. 1e).

r content (BWC) in mice of the vehicle group was 19.6 ± 2.5 %, while the DMF-treated (30, 45 mg/kg) and MMF-treated (30, 45 mg/kg) groups had decreased BWC of 11.47 ± 2.3 % (p < 0.05, as compared with vehicle group), 8.1 ± 1.9 % (p < 0.01), 11.42 ± 3.52 % (p < 0.05), or 6.42 ± 2.51 % (p < 0.01), respectively (Fig. 1e). Longitudinal Examination of Infarct Volume with MRI To further monitor the evolution of ischemic brain lesions following MCAO ischemia–reperfusion injury, MRI was employed to image the lesions in vivo on days 1, 3, and 7 post-MCAO ischemia–reperfusion injury. T2-weighted images were used to calculate the volume of ischemic lesions. On day 1 post-MCAO ischemia–reperfusion injury, there was no difference in the volume of ischemic lesion in the vehicle-treated control group as compared to those in the DMF (30 mg or 45 mg/kg)-treated groups. However, on day 3 post-ischemia, the volume of ischemic lesion was significantly reduced in mice treated with DMF at a dose of 30 or 45 mg per kg body weight. The DMF-induced ischemic volume reduction on T2-weighted images was sustained on following MRI examinations at day 7 post-MCAO ischemia–reperfusion injury (Fig. 2).Fig. 2 Effects of DMF on infarct volume were sustained. Coronal MRI sections show T2 lesions in vivo on days 1, 3, and 7 after MCAO. The volume reduction in ischemic lesions with DMF was sustained. Data shown are mean ± SEM; *p < 0.05, **p < 0.01 as compared to the vehicle group; #p < 0.05, ##p < 0.01, as compared between DMF 45 mg/kg and DMF 30 mg/kg group, n = 10 per group

oronal MRI sections show T2 lesions in vivo on days 1, 3, and 7 after MCAO. The volume reduction in ischemic lesions with DMF was sustained. Data shown are mean ± SEM; *p < 0.05, **p < 0.01 as compared to the vehicle group; #p < 0.05, ##p < 0.01, as compared between DMF 45 mg/kg and DMF 30 mg/kg group, n = 10 per group DMF and MMF Inhibit Neural Apoptosis A TUNEL assay was used to assess neuronal apoptosis. TUNEL staining was performed in brain tissue sections obtained at day 3 after MCAO. In the sham group, no discernable neural apoptosis was seen. In the vehicle group, TUNEL-positive cells were densely distributed in the ischemic cortex (53.4 ± 2.61 %). The number of apoptotic cells was decreased in mice treated with DMF 30 mg (44.1 ± 2.64 %, p < 0.05 as compared to that in the vehicle-treated group), DMF 45 mg (29.2 ± 1.3 %, p < 0.01), MMF 30 mg (40.7 ± 0.77 %, p < 0.05), or MMF 45 mg/kg body weight (22.2 ± 2.12 %, p < 0.01). These results suggest a protective role of DMF or MMF against apoptosis of cells in the central nervous system (CNS) arising from ischemia–reperfusion injury (Fig. 3).Fig. 3 DMF and MMF reduce cell death caused by cerebral ischemic–reperfusion injury. TUNEL staining shows a large number of apoptotic cells in brain sections at 72 h post-ischemia–reperfusion injury. a In the vehicle-treated group, numerous TUNEL-positive cells are present in the ischemic cortex. b–e DMF or MMF treatment reduces the number of apoptotic cells. f Quantitative analysis of TUNEL-positive cells shows DMF- and MMF-induced protection from cell death and its dose-dependent pattern. Data are presented as mean ± SEM; *p < 0.05, **p < 0.01, compared to the vehicle group; ##p < 0.01, as compared between DMF 45 mg and DMF 30 mg group, or between MMF 45 mg and MMF 30 mg group; n = 8 per group

s of TUNEL-positive cells shows DMF- and MMF-induced protection from cell death and its dose-dependent pattern. Data are presented as mean ± SEM; *p < 0.05, **p < 0.01, compared to the vehicle group; ##p < 0.01, as compared between DMF 45 mg and DMF 30 mg group, or between MMF 45 mg and MMF 30 mg group; n = 8 per group DMF and MMF Treatment Leads to Suppressed Glial Activation Neural cell death in cerebral ischemia–reperfusion injury is associated with astrocytosis and microgliosis [36]. To test the hypothesis that cytoprotective properties of DMF and MMF might diminish the glial activation elicited by transient ischemia and reperfusion, immunohistochemistry studies were performed. As shown in Fig. 4, the numbers of GFAP-positive and Iba-1-positive cells were decreased in the ischemic region of the MCAO mice as compared to the sham group. Compared with the vehicle group (27.4 ± 2.90 %), the number of GFAP-positive cells was decreased in the DMF 30 mg (20.3 ± 2.47 %, p < 0.05), DMF 45 mg (16 ± 3.08 %, p < 0.01), MMF 30 mg (20.5 ± 2.23 %, p < 0.05), and MMF 45 mg/kg body weight (17 ± 2.71 %, p < 0.01) treated groups. The number of Iba-1-bearing cells was also decreased in the DMF 30 mg (27.6 ± 2.11 %, p < 0.05), DMF 45 mg (22.8 ± 2.7 %, p < 0.01), MMF 30 mg (29.8 ± 1.57 %, p < 0.05), and MMF 45 mg/kg body weight (19.7 ± 3.12 %, p < 0.01) treated groups as compared with the vehicle control group (40.0 ± 2.98 %) (Fig. 4).Fig. 4 DMF or MMF suppresses glial activation associated with transient MCAO. a Brain slices from mice with MCAO stained with anti-glial fibrillary acidic protein (GFAP) antibody. b Brain slices stained with anti-ionized calcium-binding adapter molecule (Iba-1) antibody. Brain tissue sections obtained from mice 72 h after transient MCAO show increased astrocytosis and microgliosis (vehicle control in a, b, respectively). Treatment with DMF 30, 45 mg/kg or MMF 30, 45 mg/kg body weight reduces the expression of GFAP and Iba-1. c Quantitative analysis of GFAP-bearing cells and d that of Iba-1-bearing cells in cerebral tissues from mice with transient MCAO treated with DMF or MMF. The percentages of GFAP- and Iba-1-expressing cells are decreased by DMF or MMF treatment in a dose-dependent manner. Data are presented as mean ± SEM; *p < 0.05, **p < 0.01 as compared to the vehicle group; ##p < 0.01 as compared between DMF 45 mg vs DMF 30 mg or MMF 45 mg vs MMF 30 mg groups; n = 8 per group

or MMF. The percentages of GFAP- and Iba-1-expressing cells are decreased by DMF or MMF treatment in a dose-dependent manner. Data are presented as mean ± SEM; *p < 0.05, **p < 0.01 as compared to the vehicle group; ##p < 0.01 as compared between DMF 45 mg vs DMF 30 mg or MMF 45 mg vs MMF 30 mg groups; n = 8 per group DMF and MMF Attenuate Ischemia–Reperfusion Injury Evoked Intracellular Oxidative Stress SOD is often regarded as the first line of defense against an upswing of ROS and is responsible for the conversion of superoxide to H2O2 in the cytoplasm and mitochondria [37]. GSH plays a major role in the detoxification of peroxides [38]. Brain tissues with MCAO ischemia–reperfusion injury have decreased SOD activity and GSH levels. MDA, one of the products of membrane lipid peroxidation, reflects the degree of ROS-inflicted damage via membrane lipid peroxidation. Compared with the vehicle-treated group, there were significantly increased levels of SOD activity and GSH and decreased levels of MDA in the groups treated with DMF and MMF (30 and 45 mg/kg) (Table 1). These data suggest that DMF and MMF reduced MCAO-induced oxidative stress.Table 1 Effects of DMF and MMF on transient MCAO-induced oxidative stress Group SOD (U/mg protein) MDA (nmol/mg protein) GSH (μmol/g protein) Vehicle 92.04 ± 4.20 3.98 ± 0.65 8.54 ± 2.96 DMF 30 mg/kg 114.46 ± 6.57* 2.96 ± 0.58* 9.99 ± 2.91* DMF 45 mg/kg 131.98 ± 6.13**# 2.18 ± 0.39**# 11.96 ± 2.16**# MMF 30 mg/kg 109.98 ± 5.96* 2.99 ± 0.57* 10.07 ± 1.87* MMF 45 mg/kg 125.46 ± 5.91**# 2.21 ± 0.51**# 11.64 ± 1.22**#

Group SOD (U/mg protein) MDA (nmol/mg protein) GSH (μmol/g protein) Vehicle 92.04 ± 4.20 3.98 ± 0.65 8.54 ± 2.96 DMF 30 mg/kg 114.46 ± 6.57* 2.96 ± 0.58* 9.99 ± 2.91* DMF 45 mg/kg 131.98 ± 6.13**# 2.18 ± 0.39**# 11.96 ± 2.16**# MMF 30 mg/kg 109.98 ± 5.96* 2.99 ± 0.57* 10.07 ± 1.87* MMF 45 mg/kg 125.46 ± 5.91**# 2.21 ± 0.51**# 11.64 ± 1.22**# Data are presented as mean ± SEM. n = 8 per group DMF dimethyl fumarate, MMF monomethyl fumarate, MCAO middle cerebral artery occlusion, SOD superoxide dismutase, MDA malondialdehyde, GSH glutathione *p < 0.05; **p < 0.01, as compared to the vehicle group; #p < 0.01, comparisons between DMF 45 mg and DMF 30 mg group or MMF 45 mg and MMF 30 mg group