Browse the corpus

Caffeine does not affect [Ca2+]i levels Possible involvement of altered [Ca2+]i levels after caffeine exposure to parathyroid adenoma cells was investigated using the Ca2+ indicator Fura-2 and calculation of resulting F340/F380 ratios as representing the [Ca2+]i level. Caffeine was administered in the perfusion system at 1.25 mm Ca2+. No significant changes were recorded for F340/F380 ratios after the addition of caffeine concentrations ranging from 10 μm to 5 mm (Figure 2). Figure 2. Measurement of [Ca2+]i using Fura-2 after application of 50, 200, or 500 μm caffeine (A) and 10 μm and 5 mm caffeine (B). Expression of ADORA1 and ADORA2A in parathyroid tissues Expression of the ADORA1 and ADORA2A proteins in parathyroid tissues was determined by Western blot analysis. For both antibodies, a product of expected size was detected in the positive control and in parathyroid tissues (Figure 3A). Our results showed that both ADORA1 and ADORA2A are present in normal parathyroid tissue. ADORA1 showed relatively even expression between the normal parathyroid and the 7 parathyroid adenomas investigated. Protein expression of ADORA1 was also demonstrated by immunohistochemistry. As illustrated in Figure 3B, similar expression intensity was observed in adenoma and normal parathyroid rim present on the same slides. However, for ADORA2A, varying levels of expression were observed between adenomas (Figure 3A).

Caffeine is the most commonly consumed substance with stimulatory effects on the central nervous system. The main sources are dietary beverages such as coffee, tea, and caffeinated soft drinks. The daily caffeine intake varies largely between countries and individuals, with an average daily consumption above or well above 150 mg in most Western countries (1). Multiple epidemiological and experimental studies have investigated side effects of caffeine and associations to disease phenotypes, such as low bone mineral density (BMD). In vitro studies of osteoblast cultures and in vivo animal studies have shown negative effects of caffeine on osteoblast function, bone matrix formation, and BMD (2–4). Although the results from epidemiological studies are not fully conclusive, many reports found an association between high coffee consumption of more than 3 cups per day and low BMD in women and men, especially when calcium intake was low (5–10).

of caffeine on osteoblast function, bone matrix formation, and BMD (2–4). Although the results from epidemiological studies are not fully conclusive, many reports found an association between high coffee consumption of more than 3 cups per day and low BMD in women and men, especially when calcium intake was low (5–10). PTH, vitamin D, and calcitonin are main regulators of calcium homeostasis and bone remodeling involving mainly the kidneys, gut, parathyroid glands, and skeleton. The effects of caffeine on human body functions are not yet fully conclusive. In the kidneys, caffeine intake led to reduced renal reabsorption of calcium (Ca2+) without significant change of glomerular filtration (11), which was not fully compensated during nighttime conservation (12), hence giving a net loss of urinary calcium. This calcium loss can be simply overcome by the addition of 1–2 tablespoons milk per cup of coffee (13). In the gut, caffeine intake leads to inhibition of Ca2+ absorption (14). PTH, which is secreted from the parathyroid glands, has a direct and positive effect on osteoblast survival and differentiation (15). PTH has been successfully used in the treatment of osteoporosis, and the positive effects on bone formation have been mainly attributed to the increased population of osteoblasts (15). In a Swedish study, a correlation was seen between high coffee consumption and low serum levels of intact PTH in men (16), suggesting a relationship between caffeine and parathyroid gland function, although similar results were not reported in another study of young women (17). However, possible associations between caffeine and PTH secretion from the parathyroid cell have so far not been experimentally explored.

of intact PTH in men (16), suggesting a relationship between caffeine and parathyroid gland function, although similar results were not reported in another study of young women (17). However, possible associations between caffeine and PTH secretion from the parathyroid cell have so far not been experimentally explored. Binding to adenosine receptors (ADORAs) is 1 important mechanism for caffeine action at the cellular level. Four different ADORA types encoded by separate genes are presently known. ADORA1 and especially ADORA2A are already significantly activated at a low coffee consumption of 1 cup and are increasingly activated at higher caffeine doses (1). By contrast, ADORA2B and ADORA3 are only weakly responsive to caffeine (18). Receptors ADORA1 and ADORA2A are G protein-coupled with activating or inactivating effects on adenylyl cyclase, which in turn affects cAMP, protein kinase A (PKA), subsequently regulates cellular functions (18). Given the observational associations between caffeine, BMD, and PTH, we aimed to explore possible effects of caffeine on PTH secretion from the parathyroid cell. For this purpose, we have applied an established system for studies of human parathyroid cell function in short-term cultures of cells (19–22) from patients with hyperparathyroidism. The effects of caffeine on PTH secretion and PTH gene expression, intracellular Ca2+ ([Ca2+]i), cAMP, and PKA activity were assessed as well as expression of ADORA1 and ADORA2A.

ablished system for studies of human parathyroid cell function in short-term cultures of cells (19–22) from patients with hyperparathyroidism. The effects of caffeine on PTH secretion and PTH gene expression, intracellular Ca2+ ([Ca2+]i), cAMP, and PKA activity were assessed as well as expression of ADORA1 and ADORA2A. Patients and Methods Patients, parathyroid tissue samples, and cell preparation Human parathyroid tissue samples were collected with ethical approval at the Karolinska University Hospital, Stockholm, Sweden. Informed consent was obtained from all patients, as documented in the patients' files. Diagnoses were according to the World Health Organization classification (23). Fresh pathological parathyroid cells used in studies of PTH secretion; PTH, ADORA1, and ADORA2A gene expression; intracellular cAMP; and PKA activity assays were from 25 patients (19 females and 6 males) with a median age of 57 years (range, 42–87 y), median serum PTH levels of 361 ng/L (range, 69–1163 ng/L), and median serum ionized calcium levels of 1.46 mm (range, 1.38–1.59 mm). The histopathological diagnoses were primary adenoma (n = 22), lithium-induced hyperplasia or adenoma (n = 2), and secondary parathyroid hyperplasia (n = 1). For physiological studies, fresh parathyroid tissues were randomly collected directly after tissue dissection at surgery, transported in MEM, and isolated into cell clusters by collagenase digestion using previously described methodology (19). To obtain single cells, subsequent digestion in 1× Accutase solution (catalog no. AT104; Innovative Cell Technologies, San Diego, California) was carried out for 6 minutes. Single cells were treated with caffeine within 72 hours after isolation, and samples were collected for analysis of PTH secretion; PTH, ADORA1, and ADORA2A expression; intracellular cAMP level; and PKA activity. In those experiments, omission of caffeine was used as control in each experiment, and its value was set as 1, to which values from caffeine-treated samples were then normalized. Fresh-frozen samples and slides of paraffin-embedded tissues from reference nontumorous parathyroid and parathyroid adenomas (20) were used for Western blot analyses and immunohistochemistry, respectively.

n each experiment, and its value was set as 1, to which values from caffeine-treated samples were then normalized. Fresh-frozen samples and slides of paraffin-embedded tissues from reference nontumorous parathyroid and parathyroid adenomas (20) were used for Western blot analyses and immunohistochemistry, respectively. Measurement of PTH secretion by batch incubation Cells were isolated and plated (2–5 × 105) in 24-well plates and cultured overnight to allow cells to recover from collagenase digestion and attach to the plate bottom. Before each experiment, cells were preincubated with extracellular solution (EC) containing 1.5 mm Ca2+ and 1 mg/mL BSA for 1 hour at 37°C. The medium was then changed to EC containing 1.25 mm Ca2+ and 0, 1, 10, or 50 μm caffeine. After 30-minute incubation, medium was collected and directly put on ice. Medium was then centrifuged at 3000 rpm for 5 minutes at 4°C to precipitate cells possibly present in the medium. After removal of the medium, 100 μL Mammalian Protein Extraction Reagent (Thermo Scientific, Hudson, New Hampshire) was added to each well to lyse cells. Protein concentrations were measured using Bio-Rad protein assay. Supernatants were stored at −20°C and quantified for intact PTH using electrochemiluminescence immunoassay (catalog no. 11972103; Roche Diagnostics, Indianapolis, Indiana) at the routine clinical chemistry laboratory in the Karolinska University Hospital, Stockholm, Sweden. For each sample, the PTH value was normalized to the corresponding total protein. Each treatment was performed in duplicate or triplicate. The effects of 50 and 10 μm caffeine were investigated in parallel using cells from the same adenoma. EC with 0.5 and 1.8 mm Ca2+ was applied as positive controls for cell responsiveness. For each concentration, 13–26 independent experiments were performed on 5 to 10 patient samples.

rformed in duplicate or triplicate. The effects of 50 and 10 μm caffeine were investigated in parallel using cells from the same adenoma. EC with 0.5 and 1.8 mm Ca2+ was applied as positive controls for cell responsiveness. For each concentration, 13–26 independent experiments were performed on 5 to 10 patient samples. Measurement of [Ca2+]i by Fura-2 Measurements of [Ca2+]i were carried out using Fura-2 as an indicator, following procedures previously described in detail (19). Isolated cells were grown on glass coverslips overnight until they attached and loaded with 2.5 μm Fura-2 AM in EC solution with 1.25 mm Ca2+. Coverslips were placed in a 37°C perfusion chamber exposed to an inverted fluorescence microscope with a ×40 oil objective and connected to a cooled charged-coupled device camera with an imaging system. Cells were stepwise stimulated with 0.5 mm Ca2+ and 1.25 mm Ca2+, followed by the addition of caffeine at concentrations of 10, 50, 200, 500 μm or 5 mm for 5–10 minutes. Fluorescence was provided to cells with excitation at 340 and 380 nm, and emission was monitored at 505 nm. Three to 4 independent experiments were performed on at least 2 patient samples.

5 mm Ca2+ and 1.25 mm Ca2+, followed by the addition of caffeine at concentrations of 10, 50, 200, 500 μm or 5 mm for 5–10 minutes. Fluorescence was provided to cells with excitation at 340 and 380 nm, and emission was monitored at 505 nm. Three to 4 independent experiments were performed on at least 2 patient samples. Protein expression analysis Western blot analysis was carried out according to previously described methodology for previously published parathyroid tissue samples (20). Total protein extracts from 7 parathyroid chief cell adenomas and 1 biopsy from a normal parathyroid gland were analyzed together with U251 cells used as positive control. The human glioblastoma cell line U251 was kindly provided by Anna-Maria Marino, Department of Clinical Neurosciences, Karolinska Institutet, Stockholm, Sweden. In short, the analysis involved SDS-PAGE, blotting onto nitrocellulose membranes, incubation overnight at 4°C with primary antibodies, followed by incubation with appropriate secondary antibodies. Three primary antibodies were used: rabbit polyclonal anti-ADORA1 (catalog no. ab82477; Abcam Inc., Cambridge, Massachusetts) at 1:1000 dilution; rabbit polyclonal anti-ADORA2A (catalog no. ab101678; Abcam Inc.) at 1:800; and anti-GAPDH (catalog no. sc-32233; Santa Cruz Biotechnology, Santa Cruz, California) at 1:5000 as a protein-loading control. Detection was carried out by enhanced chemiluminescence and by exposure to hyperfilm. Immunohistochemistry was carried out on tissue sections from 4 parathyroid adenomas with a normal rim using previously described methods and the anti-ADORA1 antibody at a dilution of 1:1000. Incubation without antibody served as negative control.

s carried out by enhanced chemiluminescence and by exposure to hyperfilm. Immunohistochemistry was carried out on tissue sections from 4 parathyroid adenomas with a normal rim using previously described methods and the anti-ADORA1 antibody at a dilution of 1:1000. Incubation without antibody served as negative control. Quantitative real-time PCR (qRT-PCR) Expression of ADORA1, ADORA2A, and PTH was determined in 7 cell culture samples from 5 parathyroid adenomas treated with 0, 1, 10, or 50 μm caffeine at 1.25 mm Ca2+ for 30 minutes. Total RNA was extracted using RNeasy Mini Kit (QIAGEN, Valencia, California), quantified by NanoDrop, and reverse transcribed (600 ng) using the High Capacity RNA-to-cDNA Kit (Life Technologies, Gaithersburg, Maryland). cDNA samples (2 μL) were applied for gene expression quantification using TaqMan Universal Master Mix II (Life Technologies) with gene-specific primers and TaqMan probes for ADORA1 (Hs00379752_m1), ADORA2A (Hs00169123_m1), and PTH (Hs00757710_g1). Samples were amplified following a standard protocol in an ABI Real-Time PCR 7900HF Fast System (Life Technologies). The housekeeping gene RPLP0 (Hs00420895_gH) was used as an endogenous control, and a no-template sample was used as a negative control. All samples were loaded in triplicate and run together in a 384-well plate.

were amplified following a standard protocol in an ABI Real-Time PCR 7900HF Fast System (Life Technologies). The housekeeping gene RPLP0 (Hs00420895_gH) was used as an endogenous control, and a no-template sample was used as a negative control. All samples were loaded in triplicate and run together in a 384-well plate. Measurement of intracellular cAMP Isolated cells were cultured overnight in 24-well plates with 1–2 × 105 cells per well, followed by incubation with caffeine dissolved in 1.25 mm Ca2+ EC at concentrations of 0, 1, 10, and 50 μm and 5 mm at 37°C for 30 minutes. Cells incubated with medium without caffeine were used as control. After caffeine treatment, medium was aspirated, and 100 μL 0.1 m HCl was added to the wells, followed by incubation at room temperature for 20 minutes. Cells were collected with a cell scraper and centrifuged at 1000 × g for 10 minutes. Supernatant was then acetylated with KOH and acetic anhydride, and cAMP was detected using the cAMP ELA kit (catalog no. 581001; Cayman Chemical Company, Ann Arbor, Michigan) according to the protocol recommended by the manufacturer. Plates were read at 405 nm in a Thermomax microplate reader (Molecular Devices Corp., Sunnyvale, California), and results were analyzed with the 4-Parameter Logistic model in the Masterplex 2010 software (MiraiBio Group, Hitachi Solutions America, Ltd, San Francisco, California). Protein concentrations were measured as described above, and cAMP levels were then normalized to total protein. Independent experiments were repeated 9 times in 5 adenomas.

ed with the 4-Parameter Logistic model in the Masterplex 2010 software (MiraiBio Group, Hitachi Solutions America, Ltd, San Francisco, California). Protein concentrations were measured as described above, and cAMP levels were then normalized to total protein. Independent experiments were repeated 9 times in 5 adenomas. PKA activity assay PKA activity was evaluated using the antibody-based DetectX PKA activity kit (Arbor Assays, Ann Arbor, Michigan) in primary cultured cells treated with 0, 1, 10, or 50 μm caffeine at 1.25 mm Ca2+ for 30 minutes. Cells were lysed with 1% Nonidet P-40 lysis buffer containing 0.1% protease inhibitor cocktail, 1 mm phenylmethylsulfonylfluoride, and 10 mm activated Na3VO4 and assayed according to the protocol of the manufacturer. Horseradish peroxidase conjugate activity was detected by TMB ELISA Substrates system read at 450 nm. PKA activity of each sample was calculated using the 4-Parameter Logistic model in the Masterplex 2010 software and normalized by total protein. Independent experiments were repeated 10 times in 5 adenomas. Statistics Results are given as median (25th, 75th percentiles). Box-plots are used in figures. Statistical significance was analyzed using nonparametric Wilcoxon signed-rank test between 2 groups. A P value of less than .05 was considered significant.

PKA activity assay PKA activity was evaluated using the antibody-based DetectX PKA activity kit (Arbor Assays, Ann Arbor, Michigan) in primary cultured cells treated with 0, 1, 10, or 50 μm caffeine at 1.25 mm Ca2+ for 30 minutes. Cells were lysed with 1% Nonidet P-40 lysis buffer containing 0.1% protease inhibitor cocktail, 1 mm phenylmethylsulfonylfluoride, and 10 mm activated Na3VO4 and assayed according to the protocol of the manufacturer. Horseradish peroxidase conjugate activity was detected by TMB ELISA Substrates system read at 450 nm. PKA activity of each sample was calculated using the 4-Parameter Logistic model in the Masterplex 2010 software and normalized by total protein. Independent experiments were repeated 10 times in 5 adenomas. Statistics Results are given as median (25th, 75th percentiles). Box-plots are used in figures. Statistical significance was analyzed using nonparametric Wilcoxon signed-rank test between 2 groups. A P value of less than .05 was considered significant. Results Inhibition of PTH secretion by caffeine Possible effects of caffeine on PTH secretion were determined by batch incubation studies of human parathyroid adenoma cells. Three physiologically relevant concentrations of caffeine were evaluated: 1, 10, and 50 μm. Caffeine treatment was applied for periods of 30 minutes in each plate. The results are summarized in Figure 1. Treatment with 50 μm caffeine led to a significant inhibition by 10.4% (25th, 75th percentiles: 7.6, 21) of PTH secretion compared with control (P < .05). For comparison, 0.5 mm Ca2+ caused a 12.4% (0.2, 47.8) increase, whereas 1.8 mm Ca2+ led to a 30.5% (9.8, 41.5) decrease of PTH secretion.

re summarized in Figure 1. Treatment with 50 μm caffeine led to a significant inhibition by 10.4% (25th, 75th percentiles: 7.6, 21) of PTH secretion compared with control (P < .05). For comparison, 0.5 mm Ca2+ caused a 12.4% (0.2, 47.8) increase, whereas 1.8 mm Ca2+ led to a 30.5% (9.8, 41.5) decrease of PTH secretion. Figure 1. Summary of PTH secretion from human parathyroid cells after caffeine treatment at concentrations of 0, 1, 10, and 50 μm at extracellular calcium ([Ca2+]e) of 1.25 mm. The PTH level without caffeine was set at 1.0 for each experiment; 0.5 and 1.8 mm Ca2+ were applied as controls for cell responsiveness. n, number of experiments. Box-plots show median levels together with 10th, 25th, 75th, and 90th percentiles, outline indicated as open circle. *, P < .05 as compared to control cells. Caffeine does not affect [Ca2+]i levels Possible involvement of altered [Ca2+]i levels after caffeine exposure to parathyroid adenoma cells was investigated using the Ca2+ indicator Fura-2 and calculation of resulting F340/F380 ratios as representing the [Ca2+]i level. Caffeine was administered in the perfusion system at 1.25 mm Ca2+. No significant changes were recorded for F340/F380 ratios after the addition of caffeine concentrations ranging from 10 μm to 5 mm (Figure 2). Figure 2. Measurement of [Ca2+]i using Fura-2 after application of 50, 200, or 500 μm caffeine (A) and 10 μm and 5 mm caffeine (B).

nvestigated. Protein expression of ADORA1 was also demonstrated by immunohistochemistry. As illustrated in Figure 3B, similar expression intensity was observed in adenoma and normal parathyroid rim present on the same slides. However, for ADORA2A, varying levels of expression were observed between adenomas (Figure 3A). Figure 3. Protein expression of ADORA1 and ADORA2A in parathyroid tissues. A, Autoradiograms of Western blot analyses for protein expression of ADORA1 and ADORA2A in 7 human parathyroid adenomas, normal parathyroid tissue, and human glioblastoma U251 cells used as positive control. Subsequent incubation with anti-GAPDH was used as protein-loading control. B, Photomicrographs showing ADORA1 expression in parathyroid adenoma and in the adjacent rim of normal parathyroid (magnification, ×36). ADORA1, ADORA2A, and PTH gene expressions after caffeine treatment The expression of ADORA1, ADORA2A, and PTH after 30-minute caffeine treatment of cultured parathyroid adenoma cells was evaluated by qRT-PCR. A significant decrease of PTH mRNA by 15.3% (4.6, 32.4) was found after addition of 50 μm caffeine, but not at 1 or 10 μm (Figure 4). Decrease of ADORA1 expression was detected at 10 μm caffeine treatment by 11.1% (10.6, 29.7) and after 50 μm by 11% (7.8, 19.4) (Figure 4), but not after 1 μm. However, caffeine did not cause any significant change of ADORA2A mRNA level at any tested concentration (Figure 4).

caffeine, but not at 1 or 10 μm (Figure 4). Decrease of ADORA1 expression was detected at 10 μm caffeine treatment by 11.1% (10.6, 29.7) and after 50 μm by 11% (7.8, 19.4) (Figure 4), but not after 1 μm. However, caffeine did not cause any significant change of ADORA2A mRNA level at any tested concentration (Figure 4). Figure 4. Relative mRNA expression levels of PTH, ADORA1, and ADORA2A determined by qRT-PCR after incubation with 0, 1, 10, or 50 μm caffeine. The expression level in the noncaffeine control was set at the arbitrary value of 1.0 for each experiment. Box-plots show median, 10th, 25th, 75th, and 90th percentiles. *, P < .05 as compared to control cells. Effects of caffeine on intracellular cAMP levels and PKA activity We subsequently determined possible alterations in the levels of cAMP and PKA activity, critical second messengers mediating caffeine cellular function. For this purpose, the content of acetylated cAMP and PKA activity was quantitated after exposure to caffeine in the culturing media for 30 minutes. The results are summarized in Figure 5. The cAMP levels were not significantly affected at 10 μm. A significant decrease of cAMP by 9.8% (2.0, 26.6) was found after caffeine treatment at 50 μm. However, a very high concentration of caffeine (5 mm) was found to significantly elevate the cAMP level by 43.8% (19.8, 90) above control cells (P < .05). A significant inhibition of PKA was observed at 1 and 50 μm caffeine by 18.8% (11.6, 18.8) and 31.3% (19.6, 43.3), respectively.

d after caffeine treatment at 50 μm. However, a very high concentration of caffeine (5 mm) was found to significantly elevate the cAMP level by 43.8% (19.8, 90) above control cells (P < .05). A significant inhibition of PKA was observed at 1 and 50 μm caffeine by 18.8% (11.6, 18.8) and 31.3% (19.6, 43.3), respectively. Figure 5. Results from measurements of intracellular cAMP and PKA in human parathyroid adenoma cells exposed to caffeine for 30 minutes. A, Intracellular cAMP measured after administration of caffeine at concentrations of 1, 10, 50 μm and 5 mm, and for control cells without caffeine treatment. B, PKA activities after incubation with 0, 1, 10, or 50 μm caffeine. Intracellular cAMP and PKA activity of the noncaffeine control was set at 1.0 for each experiment. Box-plots show median levels together with 10th, 25th, 75th, and 90th percentiles, outline indicated as open circle. *, P < .05 as compared to control cells.

A activities after incubation with 0, 1, 10, or 50 μm caffeine. Intracellular cAMP and PKA activity of the noncaffeine control was set at 1.0 for each experiment. Box-plots show median levels together with 10th, 25th, 75th, and 90th percentiles, outline indicated as open circle. *, P < .05 as compared to control cells. Discussion Osteoporosis is a common disease in the elderly population. Bone fracture, which is the main consequence of osteoporosis, leads to decreased quality of life and increased morbidity and mortality. In view of the lack of effective therapies, prevention of osteoporosis is of great interest. Many epidemiology studies have been carried out with the aim to identify risk factors for bone loss (24, 25). Factors like increased age, low body weight, and weight loss have consistently been found to be associated with bone loss. However, possible associations to dietary components, such as caffeine intake, are more obscure. Although some studies did not find an association, other studies reported associations between high coffee intake (more than 3 cups per day that contain 300 mg caffeine) and low BMD (5–9), especially when calcium intake was low.

er, possible associations to dietary components, such as caffeine intake, are more obscure. Although some studies did not find an association, other studies reported associations between high coffee intake (more than 3 cups per day that contain 300 mg caffeine) and low BMD (5–9), especially when calcium intake was low. It is presently unknown how caffeine may influence the bone quality in our body. Some early studies have shown that caffeine intake caused acute increase of urinary calcium excretion (12, 13) and reduced calcium absorption in the gut (14, 15). Cell and animal studies have found that caffeine directly inhibits osteoblast cell function (4) or induces calcium release from bone (26). In the last decade, PTH has been found to be an important factor for bone remodeling, and injection of PTH has been approved for treatment of osteoporosis. Given that both hyper- and hypoparathyroidism are associated with increased risk of bone fracture (27), we speculated that caffeine might affect bone quality through modulation of PTH secretion. Initial data obtained with an alternative technique to study PTH secretion, ie, cell perfusion, revealed a significant inhibition of PTH secretion at nonphysiological high concentrations (200 and 500 μm and 5 mm) of caffeine (data not shown). The present experiments were carried out under physiological conditions; ie, the effect of caffeine was investigated at 1.25 mm Ca2+, which is close to the physiological set point for half inhibition of maximal PTH secretion in humans. Our observations included caffeine concentrations of 1, 10, and 50 μm. One to 10 μm corresponds to the serum concentration after the intake of 1 cup (177 mL) of coffee. No significant inhibition of PTH secretion or PTH mRNA expression was seen at 1 and 10 μm. Raising the caffeine concentration to 50 μm resulted in significantly decreased PTH secretion by 10.4% and significant inhibition of PTH gene expression. Caffeine is quickly absorbed after digestion and easily distributed to the whole body with a half-life of 2.5 to 4.5 hours (1). Therefore, high doses of caffeine intake may inhibit pulsatile secretion of PTH.

m resulted in significantly decreased PTH secretion by 10.4% and significant inhibition of PTH gene expression. Caffeine is quickly absorbed after digestion and easily distributed to the whole body with a half-life of 2.5 to 4.5 hours (1). Therefore, high doses of caffeine intake may inhibit pulsatile secretion of PTH. In the parathyroid cell, [Ca2+]i is a central player for PTH secretion, in which high levels of [Ca2+]i result in reduced secretion of PTH. However, even caffeine levels as high as 5 mm did not result in significant changes in [Ca2+]i levels. These observations indicate that the inhibitory effect of caffeine on PTH secretion is not due to the elevation of [Ca2+]i level. Caffeine at high concentrations is a well-known activator of the ryanodine receptor (RYR), leading to calcium release from sacro/endoplasmic reticulum in many cell types (28, 29). However, it has also been found to reduce inositol trisphosphate (IP3)-stimulated calcium release and the following store operated calcium entry (30), which exists in human parathyroid cells (20). The effect of caffeine on [Ca2+]i (intracellular calcium) levels varies between cells depending on the distribution of RYR receptors and IP3 receptors (26–28). Expression of RYR receptors and IP3 receptors in human parathyroid has not been addressed. The lack of altered [Ca2+]i after caffeine stimulation may be due to the balance between the two receptors.

racellular calcium) levels varies between cells depending on the distribution of RYR receptors and IP3 receptors (26–28). Expression of RYR receptors and IP3 receptors in human parathyroid has not been addressed. The lack of altered [Ca2+]i after caffeine stimulation may be due to the balance between the two receptors. cAMP is another important second messenger that mediates caffeine function. Caffeine is known to modulate the intracellular cAMP level through inactivation of the G protein-coupled receptors ADORA1 and ADORA2A and inhibition of phosphodiesterase. In parathyroid cells, cAMP has been found to have a positive association with PTH secretion. Several compounds, such as the β-adrenergic agonists isoproterenol and epinephrine and forskolin (31, 32), have been shown to enhance cAMP accumulation and PTH release. In our study, we observed a decrease of cAMP at 50 μm caffeine in parallel with a decrease of PTH secretion. The expected increase of cAMP at a very high concentration of caffeine is most probably a result from inhibition of phosphodiesterase (1). Simultaneously, we observed decreased ADORA1 gene expression at 10 and 50 μm, whereas no obvious alterations of ADORA2A expression were recorded. We detected both ADORA1 and ADORA2A at the protein level in normal parathyroid and parathyroid adenomas. ADORA1 showed an equal expression in normal and parathyroid adenomas, whereas ADORA2A was found to be weak or absent in some parathyroid adenoma samples. Increased or unaltered ADORA1 mRNA expression has been reported in brain after long-term caffeine treatment (33, 34). Other studies also found that short-term caffeine treatment reduces adenylyl cyclase activity, whereas chronic caffeine treatment results in increased adenylyl cyclase activity due to adaptation after caffeine intake (35). The decrease of ADORA1 expression observed in this study may be due to the short-term caffeine incubation. Once cAMP is produced, it activates PKA and mediates multiple cellular functions. In our study, we found that PKA activity was decreased in parallel with a decrease of cAMP after administration of 50 μm caffeine.

The decrease of ADORA1 expression observed in this study may be due to the short-term caffeine incubation. Once cAMP is produced, it activates PKA and mediates multiple cellular functions. In our study, we found that PKA activity was decreased in parallel with a decrease of cAMP after administration of 50 μm caffeine. In conclusion, our results show that a high concentration of caffeine causes inhibition of PTH secretion in human parathyroid cells. This effect may be due to inhibition of intracellular cAMP and the PKA pathway, but it does not involve alterations of [Ca2+]i. Our study is the first to investigate the effect of caffeine on PTH secretion in human parathyroid cells. The observation demonstrates a functional link between caffeine and parathyroid cell function. Abbreviations: ADORAadenosine receptor BMDbone mineral density [Ca2+]iintracellular calcium ECextracellular solution IP3inositol trisphosphate PKAprotein kinase A qRT-PCRquantitative real-time PCR RYRryanodine receptor. Acknowledgments The authors thank Drs Martin Bäckdahl, Catharina Ihre Lundgren, Inga-Lena Nilsson, Jörgen Nordenström, Eva Reihnér, and Jan Zedenius at the Department of Breast and Endocrine Surgery, Karolinska University Hospital, and Dr Anders Höög and Lisa Ånfalk at the Department of Pathology for excellent support in collection of parathyroid tissues. This study was supported by grants from the Fredrik and Ingrid Thuring Foundation, the AFA Insurance Foundation, the Tornspiran Foundation, the Swedish Research Council, the Stockholm County Council, Karolinska Institutet, and the Swedish Cancer Society.

Acknowledgments The authors thank Drs Martin Bäckdahl, Catharina Ihre Lundgren, Inga-Lena Nilsson, Jörgen Nordenström, Eva Reihnér, and Jan Zedenius at the Department of Breast and Endocrine Surgery, Karolinska University Hospital, and Dr Anders Höög and Lisa Ånfalk at the Department of Pathology for excellent support in collection of parathyroid tissues. This study was supported by grants from the Fredrik and Ingrid Thuring Foundation, the AFA Insurance Foundation, the Tornspiran Foundation, the Swedish Research Council, the Stockholm County Council, Karolinska Institutet, and the Swedish Cancer Society. Author Contributions: All authors participated in the study design, data interpretation, and writing of the manuscript. M.L. was mainly responsible for the study conduct, data collection, and data analysis. Disclosure Summary: All authors state that they have no conflict of interest.

The results of biochemical assays provide critical, often determinative components of medical decision making. Hence, normal reference intervals are critical to accurate interpretation of patient values. Reference intervals are conventionally determined using data sets comprised of test results obtained for a specific apparently healthy population and generating 95% confidence interval to define the normal range (1). The assumption underlying this approach is that only a small proportion of a normal population will consist of subjects with an abnormal test result, and thus, the effect of these outliers will not influence the final reference interval. This assumption has recently been challenged by the recognition that reference intervals for TSH were skewed because of the inclusion of subjects with biochemical hypothyroidism and mildly elevated TSH levels (2–4). These insights led to important revisions in the normal reference intervals for TSH in the adult population.

assumption has recently been challenged by the recognition that reference intervals for TSH were skewed because of the inclusion of subjects with biochemical hypothyroidism and mildly elevated TSH levels (2–4). These insights led to important revisions in the normal reference intervals for TSH in the adult population. Because vitamin D [25-hydroxyvitamin D (25[OH]D)] deficiency (25[OH]D < 20 ng/dL (50 nmol/L)] and insufficiency [25(OH)D 30 ng/dL (75 nmol/L)] can reduce calcium absorption and cause hypocalcemia, we hypothesized that the high prevalence of vitamin D deficiency in the pediatric population (5) might affect serum calcium reference intervals. Here we report an innovative approach to determination of age-adjusted reference intervals for serum calcium excluding subjects with vitamin D deficiency. Our results provide refined reference intervals for calcium and show that many children and adolescents with serum concentrations of 25(OH)D between 20 and 30 ng/dL (50–75 nmol/L) have mildly depressed serum total calcium concentrations that skew calculated reference intervals.

uding subjects with vitamin D deficiency. Our results provide refined reference intervals for calcium and show that many children and adolescents with serum concentrations of 25(OH)D between 20 and 30 ng/dL (50–75 nmol/L) have mildly depressed serum total calcium concentrations that skew calculated reference intervals. Materials and Methods We measured total serum calcium by a colorimetric assay with the Ortho VITROS 5, 1 FS automated chemistry system (Ortho Clinical Diagnostics). This assay uses calibrators traceable to the certified National Institute of Standards and Technology reference material. The reportable range is from 1.0 to 14.0 mg/dL (0.3–3.5 mmol/L). The between-day coefficient of variation (22 d) is 1.4% and 1.6% at concentrations of 8.9 and 12.6 mg/dL (2.2–3.2 mmol/L), respectively. HPLC coupled with tandem mass spectrometry was used to measure serum total 25(OH)D based on the procedure of Maunsell et al (7) (2005) with modifications (6). The assay gave a linear response from 1.3 to 135 ng/mL (3.2–337 nmol/L) for both 25-hydroxyvitamin D2 [25(OH)D2] and 25-hydroxyvitamin D3[25(OH)D3]. The limit of quantitation was 1.3 ng/mL (3.2 nmol/L) for both compounds. The interassay variation was measured for both compounds by measuring the metabolite concentrations of three spiked serum specimens on each of 38 different days. The coefficients of variation for 25(OH)D2 were 10.0%, 9.0%, and 7.3% at 18, 35, and 100 ng/mL (45, 87, and 250 nmol/L), respectively, and for 25(OH)D3, 4.2%, 4.9%, and 4.8% at 21, 43, and 60 ng/mL (52, 107, and 149.8 nmol/L), respectively (7). Certified reference material and external quality control samples were analyzed to meet the standards outlined by the National Institute of Standards and Technology. Validation steps included recovery and both precision and accuracy under inter- and intraday variation limit of detection and analysis of each analyte over a linear range as described in Clinical and Laboratory Standards Institute guidelines. This assay detects both 25(OH)D2 and 25(OH)D3, and serum concentrations of 25(OH)D refer to the total concentrations of both metabolites.

curacy under inter- and intraday variation limit of detection and analysis of each analyte over a linear range as described in Clinical and Laboratory Standards Institute guidelines. This assay detects both 25(OH)D2 and 25(OH)D3, and serum concentrations of 25(OH)D refer to the total concentrations of both metabolites. We reviewed serum concentrations of total calcium and 25(OH)D that had been determined during the calendar year July 1, 2011, through June 30, 2012, from all patients at The Children's Hospital of Philadelphia (CHOP, n = 6868 unique patients). This population was drawn from all visits to CHOP during this time period, consisting of approximately 28 000 inpatient admissions and 1 140 000 outpatient visits. We included patients who had a serum 25(OH)D concentration between 20 ng/mL and 80 ng/mL measured within 30 days of their serum calcium measurement (n = 5449). We excluded patients admitted to the renal unit, the endocrine unit, or any critical care unit, cardiac intensive care unit (ICU), pediatric intensive care unit, and neonatal intensive care unit. These criteria yielded 4629 unique calcium values for 4629 patients. For comparison, values for all unique calcium measurements without any selection of patients were generated simultaneously (n = 106 220 unique values). We used EPEvaluator version 9 (Data Innovations, Inc) in accordance with Clinical and Laboratory Standards Institute guidelines to generate age-specific normal reference intervals. The CHOP Institutional Review Board determined that this study was exempt from institutional review board approval because it included only deidentified data.

version 9 (Data Innovations, Inc) in accordance with Clinical and Laboratory Standards Institute guidelines to generate age-specific normal reference intervals. The CHOP Institutional Review Board determined that this study was exempt from institutional review board approval because it included only deidentified data. Results Using these data sets, we determined reference intervals for serum calcium in the patients selected for 0- to 90-day-old infants [7.8–11.3 mg/dL (1.9–2.8 mmol/L)]; 91- to 180-day-old infants [8.8–11.2 mg/dL (2.2–2.8 mmol/L)], 181- to 365-day-old children [8.8–11.4 mg/dL (2.2–2.9 mmol/L)], 1- to 3-year-old children [8.8–11.1 mg/dL (2.2–2.8 mmol/L)], 4- to 11-year-old children [8.8–10.7 mg/dL (2.2–2.68 mmol/L)], 12- to 19-year-old children [8.5–10.6 mg/dL (2.1–2.7 mmol/L)], patients younger than 19 years of age [8.6–10.9 mg/dL (2.2–2.7 mmol/L)] and patients older than 19 years old [8.6–10.9 mg/dL (2.2–2.7 mmol/L)] (Table 1 and Figure 1). For comparison, we generated calcium reference intervals for unselected subjects: 0- to 90-day-old infants [7.1–11.1 mg/dL (1.8–2.8 mmol/L)]; 91- to 180-day-old infants [7.6–11.0 mg/dL (1.9–2.8 mmol/L)], 181- to 365-day-old children [7.3–10.9 mg/dL (1.8–2.7 mmol/L)], 1- to 3-year-old children [7.6–10.5 mg/dL (1.9–2.6 mmol/L)], 4- to 11-year-old children [7.5–10.4 mg/dL (1.9–2.6 mmol/L)], 12- to 19-year-old children [7.3–10.2 mg/dL (1.8–2.6 mmol/L)], and patients older than 19 years old [7.2–10.2 mg/dL (1.8–2.6 mmol/L)]. A two-way ANOVA with Tukey's correction showed significant differences between the lower limits of normal (P < .001) and the normal range (P < .001), but not for the upper limit of normal, between these two groups.

7.3–10.2 mg/dL (1.8–2.6 mmol/L)], and patients older than 19 years old [7.2–10.2 mg/dL (1.8–2.6 mmol/L)]. A two-way ANOVA with Tukey's correction showed significant differences between the lower limits of normal (P < .001) and the normal range (P < .001), but not for the upper limit of normal, between these two groups. Table 1. Age-Specific Reference Intervals of Serum Calcium Concentrations in Milligrams Per Deciliter

7.3–10.2 mg/dL (1.8–2.6 mmol/L)], and patients older than 19 years old [7.2–10.2 mg/dL (1.8–2.6 mmol/L)]. A two-way ANOVA with Tukey's correction showed significant differences between the lower limits of normal (P < .001) and the normal range (P < .001), but not for the upper limit of normal, between these two groups. Table 1. Age-Specific Reference Intervals of Serum Calcium Concentrations in Milligrams Per Deciliter Age Range Values Eliminating Patients on the Renal Unit or Endocrine Unit as Well as ICU (NICU, PICU and CICU) Patients Values for Subjects From All Units With All 25(OH)D Values for All Subjects (With Or Without 25(OH)D Concentrations) 25(OH)D >30 and <80 (75–200 nmol/L ) 25(OH)D Between 20 and 80 (50–200 nmol/L) 25(OH)D Between 20 and 30 (50–75 nmol/L) Birth to 90 d 8.0–11.3, n = 78 (2.0–2.8 mmol/L) 7.8–11.3, n = 110 (2.0–2.8 mmol/L) 7.5–11.1, n = 32 (1.9–2.8 mmol/L) 7.5–11.6, n = 140 (1.9–2.9 mmol/L) 7.1–11.1, n = 14,659 (1.8–2.8 mmol/L) 91–180 d 8.9–11.2, n = 96 (2.2–2.8 mmol/L) 8.8–11.2, n = 124 (2.2–2.8 mmol/L) 8.6–10.9, n = 28 (2.2–2.7 mmol/L) 8.5–11.2, n = 164 (2.1–2.8 mmol/L) 7.6–11.0, n = 5657 (1.9–2.8 mmol/L) 181–364 d 9.0–11.3, n = 163 (2.3–2.8 mmol/L) 8.8–11.4, n = 192 (2.2–2.9 mmol/L) 8.1–11.3, n = 29 (2.0–2.8 mmol/L) 8.2–11.6, n = 231 (2.1–2.9 mmol/L) 7.3–10.9, n = 5952 (1.8–2.7 mmol/L) 1–3 y 8.9–11.1, n = 568 (2.2–2.8 mmol/L) 8.8–11.1, n = 756 (2.2–2.8 mmol/L) 8.5–11.0, n = 188 (2.1–2.8 mmol/L) 8.6–10.8, n = 832 (2.2–2.7 mmol/L) 7.6–10.5, n = 16708 (1.9–2.6 mmol/L) 4–11 y 8.7–10.7, n = 962 (2.2–2.7 mmol/L) 8.8–10.7, n = 1440 (2.2–2.7 mmol/L) 8.6–10.7, n = 478 (2.2–2.7 mmol/L) 8.4–10.6, n = 2186 (2.1–2.7 mmol/L) 7.5–10.4, n = 27102 (1.9–2.6 mmol/L) 12–19 y 8.5–10.7, n = 943 (2.1–2.7 mmol/L) 8.5–10.6, n = 1716 (2.1–2.7 mmol/L) 8.5–10.5, n = 773 (2.1–2.6 mmol/L) 8.2–10.5, n = 2788 (2.1–2.6 mmol/L) 7.3–10.2, n = 30909 (1.8–2.6 mmol/L) >19 y 8.5–10.5, n = 195 (2.2–2.6 mmol/L) 8.5–10.5, n = 27 (2.1–2.6 mmol/L) 8.4–10.4, n = 92 (2.1–2.6 mmol/L) 8.2–10.3, n = 526 (2.1–2.6 mmol/L) 7.2–10.2, n = 5223 (1.8–2.6 mmol/L) Birth to 19 y 8.6–10.9, n = 2810 (2.2–2.7 mmol/L) 8.6–10.9, n = 4338 (2.2–2.7 mmol/L) 8.5–10.7, n = 1528 (2.1–2.7 mmol/L) 8.3–10.7, n = 6341 (2.1–2.7 mmol/L) All ages 8.6–10.9, n = 3008 (2.2–2.7 mmol/L) 8.8–10.8, n = 4629 (2.2–2.7 mmol/L) 8.5–10.7, n = 1621 (2.1–2.7 mmol/L) 8.2–10.7, n = 6867 (2.1–2.7 mmol/L) 7.4–10.6, n = 106220 (1.9–2.7 mmol/L) Abbreviation: NICU, neonatal intensive care unit; PICU, pediatric intensive care unit. Values include interval, n, below in parentheses (in nanomoles per liter).

–2.7 mmol/L) 8.8–10.8, n = 4629 (2.2–2.7 mmol/L) 8.5–10.7, n = 1621 (2.1–2.7 mmol/L) 8.2–10.7, n = 6867 (2.1–2.7 mmol/L) 7.4–10.6, n = 106220 (1.9–2.7 mmol/L) Abbreviation: NICU, neonatal intensive care unit; PICU, pediatric intensive care unit. Values include interval, n, below in parentheses (in nanomoles per liter). Figure 1. Calcium concentrations by age and 25(OH)D. A, Calcium concentrations (nanograms per deciliter) in unselected subjects compared with subjects not on the endocrine unit, not on the renal unit, and not in an ICU with 25(OH)D between 20 and 80 ng/mL. B, Calcium concentrations (nanograms per deciliter) eliminating patients with endocrine or renal diagnoses, and in ICUs stratified by 25(OH)D. C, Unselected subjects with fitted dose-response curve; this curve revealed an inflection point at 28.4 with 95% confidence intervals ranging from 27.7 to 29.0, and an R2 value of 0.022.

m concentrations (nanograms per deciliter) eliminating patients with endocrine or renal diagnoses, and in ICUs stratified by 25(OH)D. C, Unselected subjects with fitted dose-response curve; this curve revealed an inflection point at 28.4 with 95% confidence intervals ranging from 27.7 to 29.0, and an R2 value of 0.022. Within the group of selected patients, we attempted to determine a physiological lower limit of normal for serum 25(OH)D based on observation of an inflection point in serum calcium concentration. To do this, we attempted to fit a dose-response curve to all calcium values by 25(OH)D concentrations. This curve revealed an inflection point at 28.4 ng/mL (70.9 nmol/L) with 95% confidence intervals ranging from 27.7 to 29.0 (69.1–72.3 nmol/L) and an R2 value of 0.022 (Figure 1). Additionally, based on the controversy of whether 20 or 30 ng/mL (50 or 75 nmol/L) is the appropriate lower limit of normal for 25(OH)D, we generated new reference intervals based on two groups: those with a 25(OH)D from 20 to 30 ng/dL (50–75 nmol/L) comprised one group, and those with a 25(OH)D from 30 to 80 ng/dL (75–200 nmol/L) comprised the second group. Within this cohort, serum 25(OH)D levels were between 20 and 30 ng/mL (50–75 nmol/L) in 31% of subjects; t tests revealed significant differences at all pediatric ages aside from greater than 19 between calcium concentrations in these two groups (Figure 1B). Furthermore, paired t tests showed a significant [two tailed P < .03, difference 0.24 mg/dL (0.06 mmol/L)] between the lower limits of normal for these subjects as well as the upper limits [two tailed P < .002, difference 0.12 mg/dL (0.03 mmol/L)] but no difference in range compared with subjects with 25(OH)D levels between 30 and 80 ng/mL (75 and 200 nmol/L). These results support the proposition, shown in prior studies based on serum PTH levels (8) and calcium absorption efficiency (9, 10), that 30 ng/mL (75 nmol/L) is a physiological lower limit of normal for serum 25(OH)D.

ce in range compared with subjects with 25(OH)D levels between 30 and 80 ng/mL (75 and 200 nmol/L). These results support the proposition, shown in prior studies based on serum PTH levels (8) and calcium absorption efficiency (9, 10), that 30 ng/mL (75 nmol/L) is a physiological lower limit of normal for serum 25(OH)D. Discussion Here we report age-adjusted reference intervals for total serum calcium concentration based on apparently normal subjects with vitamin D sufficiency. Similar to previous studies, we found that serum calcium levels were greatest in younger subjects, and that the upper limits of the reference ranges declined with age (11, 12). We found that at all ages, the lower limit of the reference intervals for normal subjects with vitamin D sufficiency [ie, 25(OH)D levels greater than 20 ng/ml (50 nmol/L)] were significantly greater (P < .001) than those for unselected subjects. By contrast, the upper limit of these reference ranges did not vary between groups defined by vitamin D status. Hence, the reference intervals for serum calcium that we have determined likely reflect a more physiologically appropriate reference interval, rather than a reference range based on values from an unselected population that includes many subjects with vitamin D insufficiency. That is, the increase in precision in the normal range of serum calcium actually reflects an increase in accuracy. Our observations suggest that many subjects with 25(OH)D levels less than 30 ng/mL have mild hypocalcemia. We did not observe a relationship between 25(OH)D and higher levels of total serum calcium, consistent with the notion that within the physiological range, levels of 25(OH)D do not lead to higher concentrations of serum calcium in patients with intact parathyroid gland function. Most importantly, our results indicate that inclusion of patients with low serum levels of 25(OH)D has a significant impact on the determination of reference ranges for serum calcium.

gical range, levels of 25(OH)D do not lead to higher concentrations of serum calcium in patients with intact parathyroid gland function. Most importantly, our results indicate that inclusion of patients with low serum levels of 25(OH)D has a significant impact on the determination of reference ranges for serum calcium. Our studies also addressed the ongoing controversy in the field of vitamin D biology concerning the definition of vitamin D deficiency. Several attempts to define the lower limit of the normal range for 25(OH)D have relied on population data showing an increase in serum levels of PTH when serum levels of 25(OH)D are less than 30 ng/mL (75 nmol/L) (13, 14). Mortality data, as well as the difficulty in determining the inflection point of the PTH curve, have limited acceptance of these results, however (15–17). Based on other studies, some authorities (15, 18) have proposed that a serum concentration of 25(OH)D less than 20 ng/dL represents a state of vitamin D deficiency. Our data indicate that many patients with serum 25(OH)D concentrations between 20 and 30 ng/mL have mildly depressed serum calcium levels and thereby support the premise that 30 ng/mL represents a more valid lower physiological limit.

ncentration of 25(OH)D less than 20 ng/dL represents a state of vitamin D deficiency. Our data indicate that many patients with serum 25(OH)D concentrations between 20 and 30 ng/mL have mildly depressed serum calcium levels and thereby support the premise that 30 ng/mL represents a more valid lower physiological limit. There are several weaknesses to this study. We cannot be certain that our patients are really apparently healthy because many hospitalized children were acutely ill, which may have influenced serum calcium or 25(OH)D levels. Moreover, we did not use albumin levels to adjust calcium, nor we did not measure serum levels of vitamin D binding protein or PTH. Although we did not adjust 25(OH)D concentrations for season, we included measurements obtained throughout the year. Finally, because of the retrospective nature of our study, we had no knowledge of calcium or vitamin D intake for any of the patients. However, the strengths of this study include the large number of patients and an extensive data set, a broad representation of age spectrum, and inclusion of a diverse and free-living patient population comprised of all ethnic groups.

ur study, we had no knowledge of calcium or vitamin D intake for any of the patients. However, the strengths of this study include the large number of patients and an extensive data set, a broad representation of age spectrum, and inclusion of a diverse and free-living patient population comprised of all ethnic groups. Our reference intervals are considerably broader than other recently published pediatric calcium reference intervals (11, 12). This broadened dispersion, in the context of a significantly greater number of values, may reflect seasonal heterogeneity or ethnic heterogeneity because the previously reported values were reported for ethnically homogenous groups (11, 12). Nevertheless, our study represents the first attempt to develop calcium reference ranges for a pediatric population that excludes subjects who are vitamin D deficient. These new reference intervals refine previous normal ranges that likely included many subjects who had abnormal vitamin D status and suggest that future studies to define the reference intervals for normal serum calcium levels should consider vitamin D status in the selected cohort. Abbreviations: CHOPChildren's Hospital of Philadelphia ICUintensive care unit 25(OH)D25-hydroxyvitamin D 25(OH)D225-hydroxyvitamin D2 25(OH)D325-hydroxyvitamin D3. Acknowledgments The content of this work is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Abbreviations: CHOPChildren's Hospital of Philadelphia ICUintensive care unit 25(OH)D25-hydroxyvitamin D 25(OH)D225-hydroxyvitamin D2 25(OH)D325-hydroxyvitamin D3. Acknowledgments The content of this work is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The project described in this work was supported by the Center of Excellence in Environmental Toxicology at the University of Pennsylvania, the National Institutes of Health through Grants T32-HD043021 and R01DK079970, and the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant UL1TR000003. Disclosure Summary: The authors have nothing to disclose.

In the absence of sufficient nonskeletal outcomes data (1–3), the 2010 Institute of Medicine (IOM) committee considered only bone studies to estimate pediatric vitamin D requirements (4). Although the committee concluded that there were no skeletal benefits of 25-hydroxyvitamin D (25(OH)D) >50 nmol/L and that intakes of 600 IU vitamin D per day were satisfactory, knowledge gaps persist. For example, 25(OH)D responds to oral vitamin D in a dose-dependent manner between 200 and 2000 IU/d in children (5–8); however, 25(OH)D responses to higher intakes are unknown. Most childhood intervention trials lack 1,25-dihydroxyvitamin D (1,25(OH)2D) measures, but those that report 1,25(OH)2D show dose-dependent increases (9). It remains unknown whether 1,25(OH)2D responds to vitamin D intakes >2000 IU, whether increases in 1,25(OH)2D are related to increased calcium absorption, or whether the 1,25(OH)2D response is similar in white and nonwhite populations.

OH)2D) measures, but those that report 1,25(OH)2D show dose-dependent increases (9). It remains unknown whether 1,25(OH)2D responds to vitamin D intakes >2000 IU, whether increases in 1,25(OH)2D are related to increased calcium absorption, or whether the 1,25(OH)2D response is similar in white and nonwhite populations. Controversy exists regarding a 25(OH)D inflection point in children at which maximal suppression of serum intact PTH (iPTH) occurs and calcium absorption is maximized. Although childhood trials identified iPTH inflection points of ∼75 nmol/L 25(OH)D (5, 10), we were unable to reproduce similar results (11). Child intervention studies suggest that iPTH suppression occurs with higher vitamin D intake (8), although doses <2000 IU/d do not suppress iPTH in blacks (6). Based on these data, considering iPTH suppression for defining optimal 25(OH)D in children remains questionable. We have shown that blacks vs whites have greater calcium retention (12), but relationships between calcium absorption and 25(OH)D have not been detected in cross-sectional child studies (13). Likewise, supplementation with ≤1000 IU vitamin D per day does not alter calcium absorption (8, 14). To clarify the iPTH and fractional calcium absorption response to vitamin D in children, studies using a wide range of inputs and powered to examine race differences are needed.

ed in cross-sectional child studies (13). Likewise, supplementation with ≤1000 IU vitamin D per day does not alter calcium absorption (8, 14). To clarify the iPTH and fractional calcium absorption response to vitamin D in children, studies using a wide range of inputs and powered to examine race differences are needed. The IOM report (4) called for dose-response studies in children using >2,000 IU vitamin D per day to help establish an optimal level of 25(OH)D based on functional outcomes. The aims of this study were to determine in children 1) the dose response in 25(OH)D to oral vitamin D3; 2) the degree to which vitamin D3 supplementation alters fractional calcium absorption, iPTH, and 1,25(OH)2D; and 3) whether race and sex modify these responses.

ptimal level of 25(OH)D based on functional outcomes. The aims of this study were to determine in children 1) the dose response in 25(OH)D to oral vitamin D3; 2) the degree to which vitamin D3 supplementation alters fractional calcium absorption, iPTH, and 1,25(OH)2D; and 3) whether race and sex modify these responses. Subjects and Methods Subjects Males and females aged 9 to 13 years (n = 323) participated in the 12-week randomized, double blind, placebo-controlled GAPI (University of Georgia [UGA], Purdue University [PU], and Indiana University [IU]) trial. Children were enrolled in 2009 to 2010 and 2010 to 2011 during the winter (October through December) when serum 25(OH)D is at its nadir. Study sites included U.S. latitudes 34° N (Athens, Georgia [UGA]) and 40° N (West Lafayette [PU] and Indianapolis [IU], Indiana). Within each of 4 strata defined by race (black/white) and latitude, children were assigned to 1 of 5 vitamin D3 doses (Figure 1), with block randomization (blocks were of size 5; the number of treatments) and stratified by sex, race, and latitude. Project statistician (D.B.Hal.) produced the randomization scheme using the blockrand function in R (available at: http://cran.r-project.org/). The supplement manufacturer (Douglas Laboratories) labeled the doses arbitrarily as A, B, C, D, and E. Children were enrolled by 3 study site coordinators, who assigned participants to the intervention based on chronological order of enrollment. Once enrolled, children attended 5 study visits (baseline and weeks 3, 6, 9, and 12). All participants and investigators, including biostatisticians, were blinded to dose until all data were analyzed.

by 3 study site coordinators, who assigned participants to the intervention based on chronological order of enrollment. Once enrolled, children attended 5 study visits (baseline and weeks 3, 6, 9, and 12). All participants and investigators, including biostatisticians, were blinded to dose until all data were analyzed. Figure 1. A, Study design. B, Participants recruited and retained. Children were recruited at sexual maturity stages 2 and 3, estimated using self-administered questionnaires for genitalia or breast development (15–18). Both parents and grandparents were the same race as the child and considered themselves non-Hispanic (19). Children taking nutritional supplements were enrolled after a 4-week washout. Children agreed to not alter dietary or physical activity patterns while enrolled. Exclusion criteria included menarche, growth disorders, diseases (eg, cerebral palsy), and medications (eg, corticosteroids) known to influence bone metabolism. Each university's Institutional Review Board for Human Subjects approved the procedures.

eed to not alter dietary or physical activity patterns while enrolled. Exclusion criteria included menarche, growth disorders, diseases (eg, cerebral palsy), and medications (eg, corticosteroids) known to influence bone metabolism. Each university's Institutional Review Board for Human Subjects approved the procedures. Supplements Vitamin D3 tablets (Douglas Laboratories) contained 0 (placebo), 400, 1000, 2000, or 4000 IU vitamin D3. Supplements were confirmed independently (Covance, Inc) as 0.184, 486, 1140, 1880, and 4710 IU vitamin D3, respectively. Compliance was estimated by pill counts. A subject was considered compliant if pill bottles were returned at all 5 time points and ≥80% of pills were consumed. Children who returned without pill bottles at ≥1 visit were not included in estimating compliance. A questionnaire was interviewer-administered to seek adverse events information. Anthropometric measures Anthropometric measures included weight (nearest 0.1 kg) using an electronic scale and height (nearest 0.1 cm) using a wall-mounted stadiometer (20). With a one-way random-effects model, single-measure intraclass correlation coefficients (ICCs) were computed among 6- to 10-year-old females (n = 10), measured by the same individual twice over 2 weeks. The ICCs and test-retest coefficients of variation (CV) (percent) were height (0.99% and 0.4%) and weight (0.99% and 1.4%), respectively. Body mass index (BMI)-for-age percentiles were also calculated.

nts (ICCs) were computed among 6- to 10-year-old females (n = 10), measured by the same individual twice over 2 weeks. The ICCs and test-retest coefficients of variation (CV) (percent) were height (0.99% and 0.4%) and weight (0.99% and 1.4%), respectively. Body mass index (BMI)-for-age percentiles were also calculated. Biochemical analyses Fasting blood and second-void urine samples were collected at each visit and stored at <−70°C until analysis. Reference controls (kits) and internal controls (in-house pooled samples) were included with each assay run for quality control. Repeat analyses were conducted when duplicate samples differed by ≥10%. Serum 25(OH)D was assessed using a 2-step RIA (Diasorin) (21, 22). The inter- and intra-assay CV were 5.6% to 8.4% and 5.5% to 7.0%, respectively. Analytical reliability of 25(OH)D assays was further monitored through DEQAS (the Vitamin D External Quality Assessment Scheme). Serum iPTH was measured using an immunoradiometric assay (Diasorin). The inter- and intra-assay CV were 4.8% to 6.9% and 2.3% to 5.7%, respectively. Serum 1,25(OH)2D was quantitated using a 2-step RIA (Diasorin). The inter- and intra-assay CV were 12.4% to 17.6% and 11.6%, respectively. Serum and urine calcium (CV = 2.1%) and creatinine (Cr) (CV = 3.5%) were measured using a clinical analyzer (Cobas Mira). Hypercalcemia was defined as serum calcium >10.6 mg/dL and hypercalciuria as urine calcium corrected for Cr >0.22 mg. Serum 25(OH)D values >200 nmol/L characterized hypervitaminosis D.

respectively. Serum and urine calcium (CV = 2.1%) and creatinine (Cr) (CV = 3.5%) were measured using a clinical analyzer (Cobas Mira). Hypercalcemia was defined as serum calcium >10.6 mg/dL and hypercalciuria as urine calcium corrected for Cr >0.22 mg. Serum 25(OH)D values >200 nmol/L characterized hypervitaminosis D. Fractional calcium absorption Fractional calcium absorption was measured at baseline and 12 weeks using a single oral stable isotope and calculated as follows: 1.9458 × (3-hour SSA)0.8597 × BSA1.8608 × e(−0.1918×Tanner), where SSA is serum specific activity expressed as fraction of administered tracer dose per gram of Ca in a 3-hour blood sample; BSA is body surface area, calculated as 0.20247 × (weight [kilograms]0.425 × height [meters]0.725 (23); e is the base of the natural logarithm (∼2.71828), Tanner is pubertal stage. This simplified method, using a single oral isotope and 2 blood samples (3 hours apart), predicts calcium absorption by double-isotope (R2 = 0.90, P < .01) (24). Five milligrams of 44Ca as CaCl2 in 1 mL saline was consumed as part of a standardized breakfast containing 150 mg 40Ca. Blood was drawn before and 3 hours after isotope administration. Sample preparation was performed by calcium oxalate precipitation, and isotope ratios were measured by High Resolution Inductively Coupled Plasma mass spectrometer.

l2 in 1 mL saline was consumed as part of a standardized breakfast containing 150 mg 40Ca. Blood was drawn before and 3 hours after isotope administration. Sample preparation was performed by calcium oxalate precipitation, and isotope ratios were measured by High Resolution Inductively Coupled Plasma mass spectrometer. Body composition measurement Fat mass, percent fat, and fat-free soft tissue were assessed at baseline using dual-energy x-ray absorptiometry (DXA) (Delphi-A, Hologic Inc [UGA]; Lunar iDXA, GE Medical Instruments [PU]; and Hologic Discovery-W [IU]). The same technician at each site conducted scans and performed analyses using instrument-specific software and protocols. ICCs were calculated in females aged 5 to 8 years (n = 10) scanned twice at UGA over 7 days for body composition (all ≥0.98). Short- and long-term precision of DXA at IU was <2%. The UGA/PU sites were cross-calibrated by scanning 26 children on the Delphi-A and an iDXA, whereas the IU and PU sites were cross-calibrated by scanning 10 children on the Discovery-W and iDXA. Regression formulae between UGA/PU and IU/PU were derived and used to adjust data from UGA/IU to PU values.

XA at IU was <2%. The UGA/PU sites were cross-calibrated by scanning 26 children on the Delphi-A and an iDXA, whereas the IU and PU sites were cross-calibrated by scanning 10 children on the Discovery-W and iDXA. Regression formulae between UGA/PU and IU/PU were derived and used to adjust data from UGA/IU to PU values. Demographic, dietary, and physical activity assessment Parents assisted children in answering interviewer-administered demographic questions and completing 3-day diet records and 3-day physical activity recalls at home on 2 weekdays and 1 weekend day (25–28). Records were analyzed by Food Processor SQL version 9.7.3 (ESHA Research) (29), entered by 2 researchers and statistically compared for agreement. Average measure (3-day) ICCs were calculated in girls aged 6 to 10 years (n = 10), whose 3-day diet records were completed twice over 2 weeks and calculated for vitamin D, calcium, and energy (≥0.86). For the physical activity recalls, a metabolic equivalent (MET) value was assigned to each time block based on the type and intensity of activity described, and average METs per day were calculated.

0), whose 3-day diet records were completed twice over 2 weeks and calculated for vitamin D, calcium, and energy (≥0.86). For the physical activity recalls, a metabolic equivalent (MET) value was assigned to each time block based on the type and intensity of activity described, and average METs per day were calculated. Statistical analyses and sample size determination Statistical analyses were performed using SAS System for Windows version 9.2 (SAS Institute) and R: A Language and Environment for Statistical Computing, version 2.14.1 (R Foundation for Statistical Computing). Descriptive statistics, range, and normality checks and three-way ANOVA with race, sex, latitude, and all interactions at baseline were performed. To investigate the effect of vitamin D supplementation on biochemical outcomes, 25(OH)D and 1,25(OH)2D were each modeled over time via a nonlinear mixed-effects model (30) of the following form: yijkl = αjkl + βjkl[1 − exp(−eγjkltijkl)] + ϵijkl, where yijkl and tijkl denote the measured response (eg, 25[OH]D) and corresponding time (in days from baseline) at measurement on occasion i for subject j within treatment k (k = 1, …, 5) of race l (l = 1, 2). This model is commonly known as the asymptotic regression model or monomolecular growth model and has been used to characterize 25(OH)D levels over time in response to supplementation in previous literature (31). It describes a pattern of change over time in which the response increases smoothly to a long-run or asymptotic level from an initial baseline value. In this model, for the subject corresponding to j, k, and l, αjkl denotes the baseline or initial concentration; αjkl + βjkl denotes the asymptotic concentration, that is, βjkl represents the asymptotic or long-run gain due to supplementation; and γjkl characterizes the rate of increase from baseline to the asymptote. More specifically, it denotes the natural logarithm of the elimination rate.

e baseline or initial concentration; αjkl + βjkl denotes the asymptotic concentration, that is, βjkl represents the asymptotic or long-run gain due to supplementation; and γjkl characterizes the rate of increase from baseline to the asymptote. More specifically, it denotes the natural logarithm of the elimination rate. Of central interest is the asymptotic gain, β, and how it differs across the supplementation treatments and other experimental factors. Of secondary interest is whether the rate parameter, γ, and, to a lesser extent, baseline value α depend upon experimental factors. To address these questions, each mixed effect was modeled with ANOVA-type specifications involving two-way interactions and main effects among dose and race with additive site and gender effects. For example, the subject-specific long-run gain parameter was specified as βjkl = μ + λk + τl + (λτ)lk + δ1Mjkl + δ2Gjkl + bjkl, where λk, τl and (λτ)lk are main effects and interactions for treatment and race; Mjkl and Gjkl are indicators for male gender and being from Georgia, respectively, with corresponding regression coefficients δ1 and δ2; and bjkl is the normally distributed random subject effect with mean zero. Similar specifications were used for α and γ. These specifications reflect the experimental design and allow a structured analysis of whether the characteristic features of the pattern of change over time in the presence of supplementation depend upon dose (main effect of treatment), whether that dependence has a dose-response form (linear trend in treatment effects), and whether the dose effects differ by race (treatment by race interaction), while controlling for expected differences across sites and gender. These tests were conducted using F tests suitable for parametric hypotheses on fixed effects in nonlinear mixed-effects models (30).

form (linear trend in treatment effects), and whether the dose effects differ by race (treatment by race interaction), while controlling for expected differences across sites and gender. These tests were conducted using F tests suitable for parametric hypotheses on fixed effects in nonlinear mixed-effects models (30). As is typical in this class of models, the subject-specific random effects in α, β, and γ were assumed jointly normal and independent across subjects. In addition, the error terms, ϵijkl, were assumed to be independent, normal random variables with variance proportional to a power of the mean to account for nonconstant (increasing) variance that was observed in residual plots and model diagnostics. The model was fitted simultaneously to data from all children to yield greater power to determine dose effects and to eliminate problems of nonconvergence encountered by previous researchers who used a 2-stage estimation procedure for a very similar model and application (31). Generally speaking, nonsignificant terms were not removed from the model with the exception of the gender effect in γ, which was found to be nonsignificant in initial model fitting and eliminated to facilitate convergence of subsequent models. This decision was made on practical grounds but is supported by the results of other authors (31) who found the rate parameter to exhibit little variation across subjects.

the gender effect in γ, which was found to be nonsignificant in initial model fitting and eliminated to facilitate convergence of subsequent models. This decision was made on practical grounds but is supported by the results of other authors (31) who found the rate parameter to exhibit little variation across subjects. A test of linear association, on a log-transformed scale as necessary, was conducted to elucidate dependence of iPTH, 1,25(OH)2D, and fractional calcium absorption on 25(OH)D. Inverse regression methodology was used to estimate the 25(OH)D value associated with a given targeted mean of 1/iPTH. Although the primary analysis was intent-to-treat, secondary analyses adjusting for compliance were conducted where β and γ were modeled using analysis of covariance (32). Covariates included significant baseline predictors of subsequent compliance as determined by a separate regression model built for this purpose.

A test of linear association, on a log-transformed scale as necessary, was conducted to elucidate dependence of iPTH, 1,25(OH)2D, and fractional calcium absorption on 25(OH)D. Inverse regression methodology was used to estimate the 25(OH)D value associated with a given targeted mean of 1/iPTH. Although the primary analysis was intent-to-treat, secondary analyses adjusting for compliance were conducted where β and γ were modeled using analysis of covariance (32). Covariates included significant baseline predictors of subsequent compliance as determined by a separate regression model built for this purpose. Statistical power was calculated based on fractional calcium absorption, 25(OH)D, 1,25(OH)2D, and iPTH with the analysis of 25(OH)D regarded as the primary analysis driving the sample size determination. The sample size necessary to detect a linear change in the asymptotic gain parameter β proportional to dose with a power of 80% was calculated based upon assumptions regarding differences between the 0- and 2000-IU groups within each race via standard sample size and power routines designed for contrasts in multi-way ANOVA models (33). Our primary method of statistical analysis, inference on nonlinear mixed-effects models, can handle arbitrary patterns of missing data and yields valid statistical inferences provided that data are missing at random.

race via standard sample size and power routines designed for contrasts in multi-way ANOVA models (33). Our primary method of statistical analysis, inference on nonlinear mixed-effects models, can handle arbitrary patterns of missing data and yields valid statistical inferences provided that data are missing at random. Results Subject characteristics and supplement compliance Baseline characteristics are shown in Tables 1 and 2. Both blacks vs whites and females vs males were younger and had greater BMI-for-age percentiles. Blacks vs whites had greater lean mass. Males vs females were taller and leaner and had lower fat mass and percent fat. Of those enrolled, 93% were retained (Figure 1B). Approximately 12% (40 of 323) returned without bottles for pill counts at all time points and were not included in the compliance-based subanalyses. Of these, 8 white and 32 black children returned without pill bottles at ≥1 of the time points. Overall compliance was 52.3%, differed across races, with blacks less compliant than whites (P = .01), but did not differ across treatments. There was no treatment by race interaction affecting compliance. Table 1. Baseline Characteristics by Race, Sex, and Locationa

Results Subject characteristics and supplement compliance Baseline characteristics are shown in Tables 1 and 2. Both blacks vs whites and females vs males were younger and had greater BMI-for-age percentiles. Blacks vs whites had greater lean mass. Males vs females were taller and leaner and had lower fat mass and percent fat. Of those enrolled, 93% were retained (Figure 1B). Approximately 12% (40 of 323) returned without bottles for pill counts at all time points and were not included in the compliance-based subanalyses. Of these, 8 white and 32 black children returned without pill bottles at ≥1 of the time points. Overall compliance was 52.3%, differed across races, with blacks less compliant than whites (P = .01), but did not differ across treatments. There was no treatment by race interaction affecting compliance. Table 1. Baseline Characteristics by Race, Sex, and Locationa Variable n Overall Characteristics Differences,b P < .05 Overall (n = 323) White Males (n = 80) Black Males (n = 82) White Females (n = 78) Black Females (n = 83) Georgia (n = 160) Indiana (n = 163) Age, y 323 11.3 (1.2) 12.1 (1.0) 11.8 (1.2) 11.0 (1.0) 10.5 (1.0) 11.3 (1.2) 11.4 (1.2) B < W; F < M Anthropometry Weight, kg 323 47.4 (12.2) 47.8 (13.8) 49.4 (13.0) 44.3 (9.3) 47.8 (11.8) 48.0 (11.8) 46.7 (12.5) Height, cm 323 151 (9.3) 154 (9.7) 152 (9.6) 149 (8.8) 148 (8.2) 151 (8.9) 151 (9.6) F < M BMI-for-age, % 323 68.0 (29.2) 56.6 (32.2) 72.1 (26.1) 66.2 (27.4) 76.7 (27.3) 70.3 (28.7) 65.8 (29.6) W < B; M < F Body compositionc Fat mass, kg 320 14.9 (7.4) 13.6 (8.0) 14.0 (7.5) 14.4 (6.1) 17.0 (7.6) 14.2 (7.5) 15.4 (7.3) M < F % body fat 320 31.1 (9.4) 28.2 (9.4) 28.4 (9.5) 32.6 (7.9) 35.2 (8.7) 29.7 (9.8) 32.6 (8.7) M < F Lean mass, kg 320 30.1 (6.9) 32.2 (7.8) 32.5 (7.6) 28.2 (5.0) 28.6 (5.4) 31.6 (6.9) 29.2 (6.6) W < B; F < M Biochemical 25(OH)D, nmol/L 318 70.0 (18.6) 80.6 (13.4) 61.2 (14.6) 79.6 (16.6) 59.3 (18.0) 72.0 (17.1) 68.0 (19.8) B < W; IN B < GA B 1,25(OH)2D, pmol/L 318 144 (42.8) 130 (34.6) 146 (46.4) 147 (47.9) 152 (38.6) 146 (45.6) 142 (39.9) W < B; M < F iPTH (pg/mL) 318 27.4 (10.8) 24.2 (9.0) 28.4 (11.3) 25.7 (9.2) 31.1 (12.1) 28.7 (11.4) 26.1 (10.0) W < B Fractional Ca absorption 311 0.44 (0.14) 0.46 (0.14) 0.39 (0.14) 0.47 (0.13) 0.45 (0.13) 0.43 (0.14) 0.45 (0.13) B < W; M < F Urine Ca (Ca/Cr) 305 0.06 (0.08) 0.07 (0.06) 0.05 (0.05) 0.08 (0.08) 0.05 (0.12) 0.06 (0.09) 0.07 (0.07) B < W Serum Ca, mg/dL 305 9.8 (0.3) 9.7 (0.3) 9.9 (0.3) 9.9 (0.4) 9.9 (0.3) 9.8 (0.3) 9.9 (0.4) W < B; WM < WF; GA < IN Dietary intake (per day) Energy, kcal 307 2001 (556) 2143 (541) 2025 (602) 1975 (543) 1853 (505) 1944 (528) 2053 (577) F < M Vitamin D, IU 307 169 (124) 188 (163) 168 (95) 157 (121) 162 (105) 153 (101) 184 (142) GA < IN Calcium (mg) 307 901 (395) 964 (417) 849 (401) 964 (419) 825 (321) 888 (373) 912 (416) B < W; GA M < IN M Energy expenditure, METs/d 304 62.2 (10.0) 63.4 (9.0) 61.3 (11.3) 64.8 (10.2) 59.3 (8.7) 63.4 (10.5) 61.1 (9.4) B < W; IN B < GA B Abbreviations: B, black; F, female; GA, Georgia; IN, Indiana; M, male; W, white.

IN Calcium (mg) 307 901 (395) 964 (417) 849 (401) 964 (419) 825 (321) 888 (373) 912 (416) B < W; GA M < IN M Energy expenditure, METs/d 304 62.2 (10.0) 63.4 (9.0) 61.3 (11.3) 64.8 (10.2) 59.3 (8.7) 63.4 (10.5) 61.1 (9.4) B < W; IN B < GA B Abbreviations: B, black; F, female; GA, Georgia; IN, Indiana; M, male; W, white. a Values are presented as means (SD). Overall characteristics represent data collapsed across the 5 treatment groups. b Results of three-way ANOVA for race, gender, latitude, and all interactions investigated. Differences shown are significant at α = .05. c Body composition measures assessed using DXA. Table 2. Baseline Characteristics by Treatmenta

a Values are presented as means (SD). Overall characteristics represent data collapsed across the 5 treatment groups. b Results of three-way ANOVA for race, gender, latitude, and all interactions investigated. Differences shown are significant at α = .05. c Body composition measures assessed using DXA. Table 2. Baseline Characteristics by Treatmenta Variable n Characteristics by Vitamin D Dose Placebo (n = 66) 400 IU (n = 64) 1000 IU (n = 65) 2000 IU (n = 64) 4000 IU (n = 64) Age, y 323 11.5 (1.2) 11.3 (1.2) 11.1 (1.1) 11.4 (1.4) 11.5 (1.2) Anthropometry Weight, kg 323 45.5 (11.3) 46.6 (10.4) 46.1 (11.1) 52.0 (14.8) 46.7 (12.2) Height, cm 323 151 (8.9) 151 (8.1) 149 (9.2) 153 (10.4) 150 (9.4) BMI-for-age, % 323 63.3 (29.5) 67.6 (27.8) 70.4 (28.4) 71.5 (30.4) 67.4 (30.1) Body compositionb Fat mass, kg 320 13.7 (7.0) 14.3 (7.1) 14.6 (7.3) 16.3 (8.2) 14.9 (7.5) % body fat 320 29.8 (8.9) 30.8 (9.4) 31.5 (9.9) 31.6 (9.7) 31.9 (9.0) Lean mass, kg 320 30.0 (6.3) 30.1 (5.6) 29.2 (5.8) 32.9 (9.1) 29.7 (6.6) Biochemical3 25(OH)D, nmol/L 318 71.5 (18.6) 71.4 (19.5) 71.1 (19.7) 65.8 (7.3) 70.0 (17.5) 1,25(OH)2D, pmol/L 318 146 (45) 143 (40) 140 (39) 149 (44) 142 (46) iPTH, pg/mL 318 26.6 (10.8) 27.6 (9.1) 26.2 (10.7) 30.2 (12.8) 26.4 (9.9) Serum Ca, mg/dL 305 9.8 (0.3) 9.9 (0.3) 9.9 (0.3) 9.9 (0.3) 9.8 (0.3) Urine Ca (Ca/Cr) 305 0.07 (0.06) 0.07 (0.13) 0.08 (0.10) 0.05 (0.05) 0.05 (0.05) Fractional Ca absorption 310 0.46 (0.16) 0.45 (0.15) 0.44 (0.20) 0.47 (0.17) 0.46 (0.17) Dietary intake (per day) Energy, kcal 307 1978 (513) 1996 (577) 1986 (523) 2000 (642) 2048 (536) Vitamin D, IU 307 151 (96) 198 (140) 143 (111) 184 (160) 175 (101) Calcium, mg 307 837 (321) 1000 (467) 822 (375) 914 (411) 945 (378) Energy expenditure, METs/d 304 60.9 (8.2) 63.8 (12.0) 61.0 (10.0) 63.0 (9.2) 62.4 (10.3) a Values are presented as means (SD).

307 1978 (513) 1996 (577) 1986 (523) 2000 (642) 2048 (536) Vitamin D, IU 307 151 (96) 198 (140) 143 (111) 184 (160) 175 (101) Calcium, mg 307 837 (321) 1000 (467) 822 (375) 914 (411) 945 (378) Energy expenditure, METs/d 304 60.9 (8.2) 63.8 (12.0) 61.0 (10.0) 63.0 (9.2) 62.4 (10.3) a Values are presented as means (SD). b Body composition measures assessed using DXA.

307 1978 (513) 1996 (577) 1986 (523) 2000 (642) 2048 (536) Vitamin D, IU 307 151 (96) 198 (140) 143 (111) 184 (160) 175 (101) Calcium, mg 307 837 (321) 1000 (467) 822 (375) 914 (411) 945 (378) Energy expenditure, METs/d 304 60.9 (8.2) 63.8 (12.0) 61.0 (10.0) 63.0 (9.2) 62.4 (10.3) a Values are presented as means (SD). b Body composition measures assessed using DXA. Serum 25(OH)D Serum 25(OH)D ranged from 25.3 to 114.7 nmol/L at baseline where 15% (47 of 318) of children had values <50 nmol/L (insufficiency), 6% (18 of 318) had values <40 nmol/L (estimated average requirement), and <1% (3 of 323) of children had values <30 nmol/L (deficiency). After 12 weeks, 25(OH)D ranged from 24.2 to 237.4 nmol/L with 9% (26 of 302) having serum 25(OH)D <50 nmol/L. At baseline, blacks vs whites had lower 25(OH)D and blacks living in Indiana vs Georgia had lower 25(OH)D. Changes in serum 25(OH)D in response to supplementation are illustrated in Figure 2. Serum 25(OH)D increased in a dose-dependent manner, and higher doses resulted in higher long-run concentrations (Figure 3A; P < .01). The rate of increase (γ) in 25(OH)D was not different by dose or race. The mean values for the equation parameters predicting serum 25(OH)D by treatment are shown in Table 3. The mean increment in 25(OH)D increased with dose and changes over 12 weeks ranged from −10 nmol/L for placebo to 76 nmol/L for the 4000-IU dose. The increment with each dose was significantly different from placebo (P < .05), except for the 400-IU dose in whites. To obtain prediction equations for β, a reduced model was fit in which qualitative effects of dose were replaced by the statistically significant quantitative dose effects (linear in the case of whites, linear and quadratic for blacks). This model was validated, found to fit similarly to the original model (likelihood ratio test statistic = 5.99 on 5 degrees of freedom, P = .307) and yielded the following population-level prediction equations for the asymptotic gain: blacks, ai = −8.01 + 5.22latitudei − 3.99sexi = 3.27dosei − 0.417dose2; whites, ai = −9.73 + 5.22latitudei − 3.99sexi = 2.55dosei, where latitudei = 1 if subject was from 34° N, 0 if 40° N; sexi = 1 if subject was male, 0 if female; and dosei = dose in hundreds of international units.

equations for the asymptotic gain: blacks, ai = −8.01 + 5.22latitudei − 3.99sexi = 3.27dosei − 0.417dose2; whites, ai = −9.73 + 5.22latitudei − 3.99sexi = 2.55dosei, where latitudei = 1 if subject was from 34° N, 0 if 40° N; sexi = 1 if subject was male, 0 if female; and dosei = dose in hundreds of international units. Figure 2. Change in serum 25(OH)D after 12 weeks of D3 supplementation (n = 323). In the schematic box plots, diamonds indicate means, horizontal lines indicate medians, shaded boxes indicate interquartile ranges (IQR), whiskers indicate highest value below the upper fence (1.5 × IQR above the 75th percentile) and the lowest value above the lower fence (1.5 × IQR below the 25th percentile), and circles indicate values outside the upper and lower fences. For vitamin D dose, P < .0001 for trend. Figure 3. A, Fitted 25(OH)D curves over time for the overall sample (n = 323). B, Fitted 1,25(OH)2D curves over time for the overall sample (n = 323). The majority of subjects completed the study within 12 weeks; however, data were included from several subjects who were enrolled up to 65 days after the 12-week intervention. Table 3. Model Variables to Predict 25(OH)D by Daily Vitamin D3 Dosea

Figure 3. A, Fitted 25(OH)D curves over time for the overall sample (n = 323). B, Fitted 1,25(OH)2D curves over time for the overall sample (n = 323). The majority of subjects completed the study within 12 weeks; however, data were included from several subjects who were enrolled up to 65 days after the 12-week intervention. Table 3. Model Variables to Predict 25(OH)D by Daily Vitamin D3 Dosea Dose, IU/d Parameter C(0), nmol/L a, nmol/L k Estimate SE Estimateb SE Estimate SE 0 70.89 1.90 −10.12 2.86 −3.67 0.32 400 70.82 1.93 5.54 2.59 −2.90 0.52 1,000 70.22 1.92 20.29 2.61 −2.91 0.17 2,000 65.35 1.92 37.57 2.66 −3.08 0.10 4,000 69.54 1.94 76.07 2.95 −3.05 0.07 a Means are ± SE in nanomoles per liter by daily vitamin D3 dose, averaged over race, sex, and latitude. C(t) = C(0) + a[1 − exp(−exp(k)t)] + e(t), where, C(t) is the concentration of 25(OH)D at time t for each subject, C(0) is the corresponding initial concentration of 25(OH)D at baseline, C(0) + a is the asymptotic or equilibrium concentration as t→∞ for a given constant supplementation level, k is the natural log elimination rate constant, and e(t) is a mean zero error term. b The increment with each dose was significantly different from placebo (P < .05); the 400-IU dose increment for whites was not different from placebo.

Dose, IU/d Parameter C(0), nmol/L a, nmol/L k Estimate SE Estimateb SE Estimate SE 0 70.89 1.90 −10.12 2.86 −3.67 0.32 400 70.82 1.93 5.54 2.59 −2.90 0.52 1,000 70.22 1.92 20.29 2.61 −2.91 0.17 2,000 65.35 1.92 37.57 2.66 −3.08 0.10 4,000 69.54 1.94 76.07 2.95 −3.05 0.07 a Means are ± SE in nanomoles per liter by daily vitamin D3 dose, averaged over race, sex, and latitude. C(t) = C(0) + a[1 − exp(−exp(k)t)] + e(t), where, C(t) is the concentration of 25(OH)D at time t for each subject, C(0) is the corresponding initial concentration of 25(OH)D at baseline, C(0) + a is the asymptotic or equilibrium concentration as t→∞ for a given constant supplementation level, k is the natural log elimination rate constant, and e(t) is a mean zero error term. b The increment with each dose was significantly different from placebo (P < .05); the 400-IU dose increment for whites was not different from placebo. Pairwise comparisons using Bonferroni adjustments between each of the 4 treatments and placebo showed that 25(OH)D gains in each dose were different from placebo (P < .01), except for 400 IU in whites. Table 4 shows that larger 25(OH)D gains were observed for whites vs blacks at the highest dose (P < .01). To determine whether adiposity modified the treatment effect on the asymptotic gain in 25(OH)D, the original model was refit with baseline fat mass as a covariate in the submodels for parameters α, β, and γ. Results from this secondary analysis showed that the dose effects on 25(OH)D were not modified while taking fat mass into consideration.

posity modified the treatment effect on the asymptotic gain in 25(OH)D, the original model was refit with baseline fat mass as a covariate in the submodels for parameters α, β, and γ. Results from this secondary analysis showed that the dose effects on 25(OH)D were not modified while taking fat mass into consideration. Table 4. Model Variables to Predict 25(OH)D by Race in the 4000-IU Groupa Race Parameter: C(0) + a, nmol/L Estimate SE Black 117.21 4.43 White 174.03 5.15 Averaged over race 145.62 3.40 Means are ± SE in nanomoles per liter for the 4000-IU group, averaged over sex and latitude. Larger 25(OH)D gains were observed for whites vs blacks at the 4000-IU dose (P < .01). See Table 3 for explanation of terms. Serum 1,25(OH)2D and iPTH Both blacks vs whites and females vs males had higher 1,25(OH)2D at baseline. No differences in 1,25(OH)2D were observed by latitude. Figure 3B illustrates the 12-week time course of 1,25(OH)2D by dose for the overall sample. Asymptotic increases for 1,25(OH)2D were not statistically significant (P = .07). There was also no evidence of race differences in the increase from baseline to asymptote. At baseline, blacks vs whites had higher iPTH. Despite the fact that 25(OH)D and iPTH were inversely correlated (r = −0.37, P < .01), iPTH did not change in a dose-dependent manner with supplementation, nor were there differences between races from baseline to asymptote.

nces in the increase from baseline to asymptote. At baseline, blacks vs whites had higher iPTH. Despite the fact that 25(OH)D and iPTH were inversely correlated (r = −0.37, P < .01), iPTH did not change in a dose-dependent manner with supplementation, nor were there differences between races from baseline to asymptote. Fractional calcium absorption Baseline fractional calcium absorption was lower in blacks vs whites and in males vs females. At baseline, 25(OH)D and fractional calcium absorption were not related; however, when adjusted for race, a negative relationship existed between 25(OH)D with higher absorption for whites than blacks (Figure 4A). The relationship between 25(OH)D and fractional calcium absorption did not differ by sex or latitude. There was no relationship between 12-week 25(OH)D and fractional calcium absorption before or after adjusting for race (Figure 4B). There was also no effect of vitamin D on fractional calcium absorption (change from baseline to 12 weeks) (Figure 4D). Change in 25(OH)D was not related to change in fractional calcium absorption before or after adjusting for race (Figure 4C) and was not affected when adjusted for sex, latitude, baseline 25(OH)D, or compliance. Furthermore, 1,25(OH)2D was not related to fractional calcium absorption at baseline or 12 weeks, and taking race, sex, and latitude into account did not affect these relationships.

tion before or after adjusting for race (Figure 4C) and was not affected when adjusted for sex, latitude, baseline 25(OH)D, or compliance. Furthermore, 1,25(OH)2D was not related to fractional calcium absorption at baseline or 12 weeks, and taking race, sex, and latitude into account did not affect these relationships. Figure 4. A–C, Relationship between serum 25(OH)D and fractional calcium absorption at baseline (A) (serum 25(OH)D, P = .001; race [white vs black], P < .0001; slope = −0.002, R2 = 0.073, n = 297); 12 weeks (B) (25(OH)D, P = .13; race (white vs black), P < .0001; R2 = 0.071, n = 297) and (C) as change from baseline to 12 weeks (C) (25(OH)D, P = .66; race, P = .12; R2 = 0.013, n = 297). D, Relationship between vitamin D dose and change in fractional calcium absorption after 12 weeks of supplementation (vitamin D dose, P = .54; race, P = .12). Filled circles and solid lines indicate blacks, and open circles and dotted lines indicate whites. Supplement safety Black vs white children had lower baseline urinary calcium and higher serum calcium. No differences in baseline urinary calcium were observed by sex or latitude. Among whites only, females vs males had higher serum calcium. Children living in Indiana vs Georgia had higher serum calcium. Over 12 weeks, 3 children met the criteria defining hypercalciuria, 3 for hypercalcemia, and 7 for hypervitaminosis D.

o differences in baseline urinary calcium were observed by sex or latitude. Among whites only, females vs males had higher serum calcium. Children living in Indiana vs Georgia had higher serum calcium. Over 12 weeks, 3 children met the criteria defining hypercalciuria, 3 for hypercalcemia, and 7 for hypervitaminosis D. Diet and physical activity Males had greater energy intake vs females at baseline. Children living in Indiana vs Georgia had greater intake of vitamin D. Both whites vs blacks and males living in Indiana vs Georgia consumed more calcium. Whites vs blacks and blacks living in Georgia vs Indiana had greater energy expenditure. Discussion The IOM described difficulties in establishing pediatric Dietary Reference Intakes for vitamin D partially because of the lack of dose-response studies (4). Many of the studies used to simulate these dose-response data included 1 lower dose or placebo and 1 single, larger dose. The GAPI trial was the first multisite, randomized dose-response trial conducted in children with vitamin D3 doses ranging from 400 to 4000 IU/d. Higher doses resulted in higher asymptotic 25(OH)D responses in both white and black children, but there was no effect on iPTH or calcium absorption. Although not statistically significant, 1,25(OH)2D exhibited a roughly asymptotic increase over time.

ted in children with vitamin D3 doses ranging from 400 to 4000 IU/d. Higher doses resulted in higher asymptotic 25(OH)D responses in both white and black children, but there was no effect on iPTH or calcium absorption. Although not statistically significant, 1,25(OH)2D exhibited a roughly asymptotic increase over time. The dose-dependent increases in 25(OH)D in this study were lower than reported in other trials using a range of 200 to 2000 IU/d (6, 9) and higher than observed in a study using either 200 or 1000 IU/d (34). The response to supplementation with 2000 IU/d in the current study was approximately 38 nmol/L, lower than 60 nmol/L observed when ∼2000 IU/d was provided to U.S. black adolescents over 16 weeks (6) or to female Lebanese adolescents over 1 year (9). The difference in serum responses is likely due to higher baseline 25(OH)D. Even with constrained wintertime testing to minimize UVB exposure, baseline 25(OH)D was ∼70 nmol/L in this study, more than double baseline concentrations observed in previous trials. The IOM recommends intakes of 600 IU/d for children (4), but there is considerable debate as to whether this recommendation is sufficient. Although the present trial was conducted for only 12 weeks, we found among healthy children with 25(OH)D ∼70 nmol/L that vitamin D3 supplementation with 400 IU/d was sufficient in maintaining wintertime 25(OH)D in healthy black, but not white, children.

nsiderable debate as to whether this recommendation is sufficient. Although the present trial was conducted for only 12 weeks, we found among healthy children with 25(OH)D ∼70 nmol/L that vitamin D3 supplementation with 400 IU/d was sufficient in maintaining wintertime 25(OH)D in healthy black, but not white, children. An important goal of this project was to ascertain whether race modified the biochemical responses of vitamin D metabolism with increasing vitamin D doses. There was a significant interaction between race and dose, such that at 400, 1000, and 2000 IU/d, there were equal 25(OH)D gains for whites and blacks, but at 4000 IU, whites had greater gains than blacks. Other trials that included black adolescents (5, 6) and one with adults (35, 36) demonstrated that 25(OH)D responses were similar among races. Rather than a biological explanation for the race interaction in this study, it is likely that compliance contributed to the discrepancy observed at 4000 IU. However, because our estimation of compliance included only children who returned pill bottles at all visits, we were not able to capture a significant dose by race interaction in compliance rates. It is noteworthy that 32 black vs 8 white children did not return pill bottles at ≥1 testing occasion. Of these, 6 of 32 black children and 2 of 8 white children were assigned 4000 IU/d.

dren who returned pill bottles at all visits, we were not able to capture a significant dose by race interaction in compliance rates. It is noteworthy that 32 black vs 8 white children did not return pill bottles at ≥1 testing occasion. Of these, 6 of 32 black children and 2 of 8 white children were assigned 4000 IU/d. This study was one of the few to assess 1,25(OH)2D responses to vitamin D supplementation. Unlike findings in children supplemented with ≤1000 IU vitamin D per day (7, 37), we report that there was a marginal, albeit nonsignificant, increase in 1,25(OH)2D with increasing dose, which was not different by race. These results are consistent with studies that used 1600 or 2000 IU/d regardless of baseline 25(OH)D. The marginal increases in serum 1,25(OH)2D observed in the present study should have been linked with increased calcium absorption, but that was not the case. Assessment of serum 1,25(OH)2D using liquid chromatography–mass spectrometry should be pursued to better characterize the 1,25(OH)2D increases observed in this study.

arginal increases in serum 1,25(OH)2D observed in the present study should have been linked with increased calcium absorption, but that was not the case. Assessment of serum 1,25(OH)2D using liquid chromatography–mass spectrometry should be pursued to better characterize the 1,25(OH)2D increases observed in this study. Based on adult data (35), it would be expected that increases in 25(OH)D and 1,25(OH)2D would be accompanied by iPTH suppression after supplementation. Cross-sectional studies in children have reported inverse or no association between iPTH and 25(OH)D, and in one trial in younger children, a significant reduction in iPTH occurred after 13 weeks of supplementation with 1000 IU/d (8). We report that 25(OH)D and iPTH were inversely correlated at baseline. However, after 12 weeks of supplementation with doses up to 4000 IU/d, there was no effect of increased 25(OH)D on iPTH suppression. In adults, the inflection point at which maximal suppression occurs is used in defining vitamin D requirements. Based on the findings from the current trial, the use of iPTH suppression in children for identifying recommended 25(OH)D cutoffs is not warranted.

here was no effect of increased 25(OH)D on iPTH suppression. In adults, the inflection point at which maximal suppression occurs is used in defining vitamin D requirements. Based on the findings from the current trial, the use of iPTH suppression in children for identifying recommended 25(OH)D cutoffs is not warranted. The most relevant functional outcome in this trial, fractional calcium absorption, did not improve over 12 weeks with increased 25(OH)D. This confirms cross-sectional (38) and intervention data in children (8, 14) who received 1000 IU/d. The null effect of supplementation on fractional calcium absorption implies that baseline 25(OH)D, ranging from 25.3 to 114.7 nmol/L, can be considered adequate for fractional calcium absorption in this age during winter. The response of fractional calcium absorption to vitamin D supplementation was not affected by baseline 25(OH)D; however, 85% of children had concentrations >50 nmol/L throughout the study. Therefore, our findings may not be generalizable to children with lower 25(OH)D. This study also showed that, despite lower baseline calcium absorption and 25(OH)D in blacks vs whites, concomitant increases in 25(OH)D and 1,25(OH)2D with supplementation did not result in calcium absorption increases. Therefore, 25(OH)D in healthy black children at baseline was adequate for calcium absorption.

OH)D. This study also showed that, despite lower baseline calcium absorption and 25(OH)D in blacks vs whites, concomitant increases in 25(OH)D and 1,25(OH)2D with supplementation did not result in calcium absorption increases. Therefore, 25(OH)D in healthy black children at baseline was adequate for calcium absorption. This pediatric trial was the first to employ a daily dose >2000 IU/d; therefore, it was important to monitor safety particularly in the group receiving 4000 IU. Serum calcium indicative of hypercalcemia occurred in 3 children, and another 3 distinct children had urinary calcium levels defined as hypercalcuria. None of these children were assigned to 4000 IU/d, and it did not appear that these elevated levels were associated with dose. In contrast, 7 children, or 10% of those receiving 4000 IU/d met the criteria for hypervitaminosis D by the trial's end. The consequences of these high serum levels are unknown, but because no subject reported adverse events over the 12-week study, we conclude that short-term supplementation with 4000 IU/d appears safe. Longer-term studies are needed to ascertain the safety of 25(OH)D exceeding 200 nmol/L. In conclusion, vitamin D3 doses ≤4000 IU/d appear to be safe over the short term and indicate that healthy U.S. children do not require vitamin D supplementation to improve calcium absorption. Abbreviations: BMIbody mass index Crcreatinine CVcoefficient of variation DXAdual-energy x-ray absorptiometry ICCintraclass correlation coefficient iPTHintact PTH METmetabolic equivalent 25(OH)D25-hydroxyvitamin D 1,25(OH)2D1,25-dihydroxyvitamin D.

This pediatric trial was the first to employ a daily dose >2000 IU/d; therefore, it was important to monitor safety particularly in the group receiving 4000 IU. Serum calcium indicative of hypercalcemia occurred in 3 children, and another 3 distinct children had urinary calcium levels defined as hypercalcuria. None of these children were assigned to 4000 IU/d, and it did not appear that these elevated levels were associated with dose. In contrast, 7 children, or 10% of those receiving 4000 IU/d met the criteria for hypervitaminosis D by the trial's end. The consequences of these high serum levels are unknown, but because no subject reported adverse events over the 12-week study, we conclude that short-term supplementation with 4000 IU/d appears safe. Longer-term studies are needed to ascertain the safety of 25(OH)D exceeding 200 nmol/L. In conclusion, vitamin D3 doses ≤4000 IU/d appear to be safe over the short term and indicate that healthy U.S. children do not require vitamin D supplementation to improve calcium absorption. Abbreviations: BMIbody mass index Crcreatinine CVcoefficient of variation DXAdual-energy x-ray absorptiometry ICCintraclass correlation coefficient iPTHintact PTH METmetabolic equivalent 25(OH)D25-hydroxyvitamin D 1,25(OH)2D1,25-dihydroxyvitamin D. Acknowledgments We thank Ms Jessica Smith for the overall coordination of this project as well as the participants and their families for their commitment to this research. This work was supported by Eunice Kennedy Shriver National Institute of Child Health and Human Development Grant R01HD057126.

1,25(OH)2D1,25-dihydroxyvitamin D. Acknowledgments We thank Ms Jessica Smith for the overall coordination of this project as well as the participants and their families for their commitment to this research. This work was supported by Eunice Kennedy Shriver National Institute of Child Health and Human Development Grant R01HD057126. Disclosure Summary: The authors have nothing to disclose.

Adults with a history of childhood-onset GH deficiency or with hypothalamic/pituitary disease, surgery, or irradiation to these areas, head trauma, or evidence of other pituitary hormone deficiencies are at risk for adult GH deficiency (AGHD). Because symptoms are usually nonspecific, in the absence of panhypopituitarism and low serum IGF-I levels, the diagnosis of AGHD requires biochemical confirmation with at least 1 GH stimulation test (1). The insulin tolerance test (ITT) is considered the gold standard test for AGHD, having a sensitivity of 96% and a specificity of 92% (1). However, because it induces hypoglycemia, the test is contraindicated in patients with coronary artery disease, seizures, and in the elderly (1). GHRH combined with arginine (arginine+GHRH) has been endorsed by several consensus guidelines (2–4) as the main alternative when the ITT is contraindicated, having a sensitivity of 95% and a specificity of 91% (1); but when GHRH analog (Geref Diagnostic; Serono Laboratories, Rockland, Massachusetts) was withdrawn in the United States in 2008, the need for an alternative to the ITT increased (2). The diagnosis of AGHD is important, given that the treatment of this condition, although expensive, has consistently shown improvements in body composition, exercise capacity, endothelial function, inflammatory biomarkers, bone mineral density, lipoprotein metabolism, and self-reported quality of life measures (5–10).

The diagnosis of AGHD is important, given that the treatment of this condition, although expensive, has consistently shown improvements in body composition, exercise capacity, endothelial function, inflammatory biomarkers, bone mineral density, lipoprotein metabolism, and self-reported quality of life measures (5–10). Ghrelin is the endogenous ligand for the GH secretagogue receptor [also called the ghrelin receptor (GHS-R1a)] (11, 12). Pharmacological treatment of rats and mice with ghrelin increases GH secretion, and Ghsr−/− mice are refractory to the stimulatory effects of ghrelin on GH release, confirming that the GHS-R1a is the physiologically relevant ghrelin receptor mediating GH secretion (13). Synthetic agonists of this receptor, known as ghrelin mimetics or GH secretagogues, are molecules that evoke dose-dependent increases in GH levels (14, 15). Macimorelin (formerly known as AEZS-130, ARD-07, and prior to that, EP-01572) is a novel GH secretagogue with good stability and oral bioavailability, which binds the GHS-R1a receptor and to pituitary and hypothalamic extracts with an affinity similar to ghrelin (16). In phase I clinical studies in healthy volunteers, macimorelin stimulated GH release in a dose-dependent manner, achieving maximum blood levels within 1 hour with good tolerability (17). The objective of this clinical trial was to determine the diagnostic efficacy and safety of macimorelin in the diagnosis of AGHD.

6). In phase I clinical studies in healthy volunteers, macimorelin stimulated GH release in a dose-dependent manner, achieving maximum blood levels within 1 hour with good tolerability (17). The objective of this clinical trial was to determine the diagnostic efficacy and safety of macimorelin in the diagnosis of AGHD. Materials and Methods Study design This study was conducted at 11 centers across the United States. The protocol was approved by the institutional review board at each institution. Patients were recruited between July 2007 and July 2011. The study was conducted in compliance with ethical principles that have their origin in the Declarations of Helsinki and its amendments and the International Conference on Harmonization Guideline for Good Clinical Practices. This multicenter, open-label study was originally designed as a crossover trial of oral macimorelin (0.5 mg/kg) vs GHRH (Geref Diagnostic; Serono) iv bolus of 1 μg/kg + arginine (Ar-Gine; Pfizer, New York, New York) iv infusion of 30 g over 30 minutes in AGHD patients and in controls, matched for body mass index (BMI), estrogen status, gender, and age. After 43 AGHD patients and 10 controls had been tested, Geref Diagnostic became unavailable in the United States. The study was completed by testing 10 more AGHD patients and 38 controls with macimorelin alone.

inutes in AGHD patients and in controls, matched for body mass index (BMI), estrogen status, gender, and age. After 43 AGHD patients and 10 controls had been tested, Geref Diagnostic became unavailable in the United States. The study was completed by testing 10 more AGHD patients and 38 controls with macimorelin alone. Study drug Macimorelin was provided by Ardana Biosciences Ltd (Edinburgh United Kingdom) and later by Æterna Zentaris Inc (Basking Ridge, New Jersey). Macimorelin is a ghrelin receptor agonist with a molecular weight of 474.5 g/mol. The solubility in water is 0.3 mg/mL. The structural formula of macimorelin is shown in Supplemental Figure 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org. GHRH (Geref Diagnostic; Serono) and arginine (Ar-Gine; Pfizer) were purchased from their manufacturers.

confirmed AGHD. Their data were not included in this analysis. The safety population (n = 53 AGHD patients and 48 controls) included all of the subjects who received at least 1 dose of study medication and for whom any safety information was available. The analysis of all safety variables was based on this population. The primary end point was the area under the receiver-operating characteristic (ROC) curve for the peak GH after macimorelin. The basis for the ROC analysis was the patient or control status of the subject. Secondary efficacy end points included the calculation of Youden's Index for the cutoff with maximum accuracy and the peak IGF-I concentration after treatment. The primary efficacy analysis was the ROC analysis performed on peak GH concentrations when the subjects were treated with macimorelin (see Supplemental Data).

4.5 g/mol. The solubility in water is 0.3 mg/mL. The structural formula of macimorelin is shown in Supplemental Figure 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org. GHRH (Geref Diagnostic; Serono) and arginine (Ar-Gine; Pfizer) were purchased from their manufacturers. Eligibility criteria Subjects were 18 years old or older and provided written informed consent. GH deficiency was confirmed by low age- and sex-adjusted IGF-I levels and pituitary hormone deficiencies in 3 or more hormones (TSH, ACTH, LH/FSH, and/or arginine vasopressin) or, for those patients with fewer deficiencies, by one of the following stimulation tests with these cut points for peak GH levels: 1) arginine+GHRH (cut point 4.1 μg/L), 2) ITT (cut point 5.0 μg/L), 3) glucagon stimulation test (cut point 3.0 μg/L), or 4) arginine (cut point 0.4 μg/L) as recommended previously (1, 2, 4, 18). Subjects who required replacement therapy for hormone deficiencies other than AGHD had been on stable treatment for 3 months or longer. Subjects with hypogonadism were on sex steroid replacement therapy, excluding women older than 50 years of age. Controls were matched to AGHD patients already enrolled in the study based on sex; age (±5 years); BMI (±2 kg/m2); and estrogen status (women only). They were required to have undergone normal growth and development and have normal prolactin and free T4 levels. Females had a history of regular, age-appropriate menses, and males were required to have had normal serum testosterone levels.

based on sex; age (±5 years); BMI (±2 kg/m2); and estrogen status (women only). They were required to have undergone normal growth and development and have normal prolactin and free T4 levels. Females had a history of regular, age-appropriate menses, and males were required to have had normal serum testosterone levels. Subjects with AGHD were excluded from the study if they had untreated hypothyroidism, intracranial lesions not documented to be stable for 12 months or longer, GH therapy within 1 month of study entry (subjects may have been washed out from previous GH therapy and then screened for other entry criteria), significant cardiovascular or cerebrovascular disease, current active malignancy other than nonmelanoma skin cancer, renal or hepatic dysfunction (≥3 times the upper limit of normal aspartate aminotransferase, alanine aminotransferase, and gamma glutamyl transpeptidase, creatinine > 2 times the upper limit of normal), pregnancy or lactation, active Cushing's disease, or clinically relevant electrocardiogram (ECG) abnormalities (including QT/QTc interval > 450 milliseconds) at any time prior to dosing. Exclusion criteria for the controls included current pregnancy or lactation or clinically relevant ECG abnormalities (including QT/QTc interval > 450 milliseconds).

Cushing's disease, or clinically relevant electrocardiogram (ECG) abnormalities (including QT/QTc interval > 450 milliseconds) at any time prior to dosing. Exclusion criteria for the controls included current pregnancy or lactation or clinically relevant ECG abnormalities (including QT/QTc interval > 450 milliseconds). Study procedures Investigators and subjects were not blinded during the study; however, GH assays were performed on blinded samples. In the original protocol, subjects were randomized upon entry into the study to group 1 or group 2, which determined the order of the 2 tests. Half of the patients received macimorelin first, whereas the remaining half received arginine+GHRH first. Each patient's matched control was treated in the same order. In the second phase of the study, subjects received only macimorelin due to the unavailability of Geref Diagnostic (Serono); therefore, the crossover aspect of the design and the randomization no longer applied. All subjects were administered a single oral dose of macimorelin (0.5 mg/kg) in the morning after an overnight fast. (See Supplemental Material for more details.) The GH levels were measured 15–30 minutes before dosing, immediately prior to dosing, and 30, 45, 60, 75, 90, 120, and 150 minutes after dosing. Taste perception was assessed at the time of dosing, and IGF-I was measured before and 150 minutes after dosing. Complete blood count and complete metabolic panel were measured at screening and upon the completion of the study (poststudy visit). In addition, ECGs were also performed prior to dosing and 60 minutes after dosing. For the first portion of the study, the subjects were also asked to rate their preference between the 2 tests.

te blood count and complete metabolic panel were measured at screening and upon the completion of the study (poststudy visit). In addition, ECGs were also performed prior to dosing and 60 minutes after dosing. For the first portion of the study, the subjects were also asked to rate their preference between the 2 tests. Study measures A central laboratory, Esoterix Clinical Trials (LabCorp) (Cranford, New Jersey), was used to analyze the blood samples. GH levels were measured by a validated immunochemiluminometric assay with an intraassay coefficient of variation of 7.4%, an interassay coefficient of variation of 7%–14%, and a lower limit of quantitation of 0.013 ng/mL. The serum IGF-I levels were measured by a RIA. (See Supplemental Data for more details.)

samples. GH levels were measured by a validated immunochemiluminometric assay with an intraassay coefficient of variation of 7.4%, an interassay coefficient of variation of 7%–14%, and a lower limit of quantitation of 0.013 ng/mL. The serum IGF-I levels were measured by a RIA. (See Supplemental Data for more details.) Statistical analysis All statistical analyses were performed using SAS for Windows (version 9.0; SAS Institute, Inc, Cary, North Carolina), MedCalc (version 11.3.1.0; Ostend, Belgium), or Classification and Regression Tree (CART; version 6.0; Statsoft, Tulsa, Oklahoma). Analysis sets reported in the study included per-protocol and safety populations. The per-protocol set (n = 50 cases and 48 controls) included all intent-to-treat subjects who were treated with macimorelin according to the protocol without any major deviations. A protocol deviation was the enrollment of 2 patients who did not meet the entry criteria for confirmed AGHD. Their data were not included in this analysis. The safety population (n = 53 AGHD patients and 48 controls) included all of the subjects who received at least 1 dose of study medication and for whom any safety information was available. The analysis of all safety variables was based on this population.

Secondary efficacy end points included the calculation of Youden's Index for the cutoff with maximum accuracy and the peak IGF-I concentration after treatment. The primary efficacy analysis was the ROC analysis performed on peak GH concentrations when the subjects were treated with macimorelin (see Supplemental Data). Results Fifty-three AGHD subjects were enrolled at 11 centers across the United States and 52 received macimorelin. The remaining subject discontinued the study prior to dosing due to collapsed veins after arginine+GHRH. Fifty subjects had confirmed AGHD prior to study entry and were included in this analysis along with 48 controls. Two subjects were excluded from this per-protocol analysis: 1 subject had only 2 pituitary deficiencies instead of the 3 required, and another subject did not have 3 hormone deficiencies and a low IGF-I or confirmed AGHD from an arginine+GHRH test and was randomized in error. Two AGHD subjects could not be matched with a control due to the combination of young age, high BMI, and estrogen use (Figure 1). The 2 groups had similar sex, age, and BMI distribution as designed by the matching criteria, and most subjects were obese. There were more African Americans in the control group. In the AGHD group, 4 subjects were categorized by the investigators as having hypothalamic AGHD, 36 as pituitary AGHD, and 10 as unknown cause. Pituitary adenomas were the most common cause of hormone deficiencies, and within that group nonfunctioning adenomas were noted most often (Table 1). Gonadal, thyroid, and adrenal insufficiency were the 3 most common absent pituitary hormones (83%, 85%, and 59% of AGHD patients, respectively), and sex steroids (87%), T4 (85%), and glucocorticoids (62%) were the most commonly used replacement therapies. Surgery (transsphenoidal in 28 of 53 patients, 52.8%; transcranial, 10 patients, 18.9%), medication (7 patients, 13.2%), and radiation (conventional, 8 patients, 15.1%; stereotactic, 5 patients, 9.4%) were the treatments patients had received for their pituitary disorders.

the most commonly used replacement therapies. Surgery (transsphenoidal in 28 of 53 patients, 52.8%; transcranial, 10 patients, 18.9%), medication (7 patients, 13.2%), and radiation (conventional, 8 patients, 15.1%; stereotactic, 5 patients, 9.4%) were the treatments patients had received for their pituitary disorders. Figure 1. Study design and patient disposition. *, After 43 AGHD patients and 10 controls had been tested, Geref Diagnostic (Serono) became unavailable. The study was completed by testing 10 more AGHD patients and 38 controls with macimorelin alone; **, two AGHD subjects could not be matched due to the combination of young age, high BMI, and estrogen use. Table 1. Demographics and Baseline Characteristics (All Enrolled Subjects) AGHD Patients (n = 53) Matched Controls (n = 48) Sex, n, % Male 31 (58.5) 30 (62.5) Female 22 (41.5) 18 (37.5) Race, n, %a White 49 (92.5) 29 (60.4) Black or African American 2 (3.8) 18 (37.5) Asian 2 (3.8) 0 (0.0) Other 0 (0.0) 1 (2.1) Ethnicity, n, % Hispanic or Latino 11 (20.8) 9 (18.8) Not Hispanic or Latino 42 (79.2) 39 (81.3) Age, y, n, % Mean (SD) 52 (13.4) 53 (12.9) Estrogen status, n, % None 12 (22.6) 12 (25.0) Oral 9 (17.0) 6 (12.5) Patch 1 (1.9) 0 (0.0) BMI, kg/m2, n, % Lean (<25) 7 (13.2) 7 (14.6) Overweight (≥25 and < 30) 15 (28.3) 13 (27.1) Obese (≥30) 31 (58.5) 28 (58.3) Mean 32.1 (7.4) 31.8 (7.1) Etiology, n, %b NA Pituitary adenoma 34 (64.2) CNS tumors 8 (15.1) Othersc 18 (34.0) Abbreviation: CNS, central nervous system. a P < .05 between groups. b More than 1 cause may be applicable per patient.

AGHD Patients (n = 53) Matched Controls (n = 48) Sex, n, % Male 31 (58.5) 30 (62.5) Female 22 (41.5) 18 (37.5) Race, n, %a White 49 (92.5) 29 (60.4) Black or African American 2 (3.8) 18 (37.5) Asian 2 (3.8) 0 (0.0) Other 0 (0.0) 1 (2.1) Ethnicity, n, % Hispanic or Latino 11 (20.8) 9 (18.8) Not Hispanic or Latino 42 (79.2) 39 (81.3) Age, y, n, % Mean (SD) 52 (13.4) 53 (12.9) Estrogen status, n, % None 12 (22.6) 12 (25.0) Oral 9 (17.0) 6 (12.5) Patch 1 (1.9) 0 (0.0) BMI, kg/m2, n, % Lean (<25) 7 (13.2) 7 (14.6) Overweight (≥25 and < 30) 15 (28.3) 13 (27.1) Obese (≥30) 31 (58.5) 28 (58.3) Mean 32.1 (7.4) 31.8 (7.1) Etiology, n, %b NA Pituitary adenoma 34 (64.2) CNS tumors 8 (15.1) Othersc 18 (34.0) Abbreviation: CNS, central nervous system. a P < .05 between groups. b More than 1 cause may be applicable per patient. c Includes 11% head trauma, 4% empty sella, 6% childhood-onset GHD.

AGHD Patients (n = 53) Matched Controls (n = 48) Sex, n, % Male 31 (58.5) 30 (62.5) Female 22 (41.5) 18 (37.5) Race, n, %a White 49 (92.5) 29 (60.4) Black or African American 2 (3.8) 18 (37.5) Asian 2 (3.8) 0 (0.0) Other 0 (0.0) 1 (2.1) Ethnicity, n, % Hispanic or Latino 11 (20.8) 9 (18.8) Not Hispanic or Latino 42 (79.2) 39 (81.3) Age, y, n, % Mean (SD) 52 (13.4) 53 (12.9) Estrogen status, n, % None 12 (22.6) 12 (25.0) Oral 9 (17.0) 6 (12.5) Patch 1 (1.9) 0 (0.0) BMI, kg/m2, n, % Lean (<25) 7 (13.2) 7 (14.6) Overweight (≥25 and < 30) 15 (28.3) 13 (27.1) Obese (≥30) 31 (58.5) 28 (58.3) Mean 32.1 (7.4) 31.8 (7.1) Etiology, n, %b NA Pituitary adenoma 34 (64.2) CNS tumors 8 (15.1) Othersc 18 (34.0) Abbreviation: CNS, central nervous system. a P < .05 between groups. b More than 1 cause may be applicable per patient. c Includes 11% head trauma, 4% empty sella, 6% childhood-onset GHD. Effect of macimorelin on GH and IGF-I levels Mean ± SD peak GH levels in AGHD patients and controls after macimorelin administration were 2.36 ± 5.69 ng/mL and 17.71 ± 19.11 ng/mL, respectively (P < .0001, Figure 2, A and B). Obesity (BMI > 30 kg/m2) was present in 58% of AGHD patients and controls, and peak GH levels were inversely associated with BMI in controls (Figure 2C). As expected, IGF-I levels were lower in the AGHD patients (58.1 ± 37.29 ng/mL) than in the controls (128.1 ± 52.47 ng/mL, P < .0001) but did not rise significantly after a single dose of macimorelin (53 ± 34.57 ng/mL and 125.9 ± 54.26 ng/mL respectively). For more details, please see Supplemental Data, Supplemental Tables 1 and 2, and Supplemental Figure 2.

ated an unpleasant taste and 41.7% found a neutral taste). Test preference could be evaluated only for the first portion of the trial when subjects received both treatments. In that portion, 72.0% of subjects, significantly more (P = .0019) than random choice, preferred the macimorelin test over the arginine+GHRH test. Safety and tolerability Adverse events (AEs) were generally mild or moderate in severity and occurred in 19 of 52 AGHD patients (37%) and in 10 of 48 (21%) controls after macimorelin treatment. In contrast, 26 of 43 AGHD subjects (61%) and 3 of 10 controls (30%) experienced AEs with arginine+GHRH. Only 1 drug-related serious AE was reported in a control subject receiving macimorelin (2.1%), whose postdose ECG showed asymptomatic QT prolongation and nonspecific T wave abnormalities that resolved spontaneously within 24 hours. The AE was rated as serious due to the choice to hospitalize the subject as a precautionary measure. The subject had been taking citalopram, a drug that was later reported by the Food and Drug Administration (FDA) to be associated with QT prolongation (19), although the patient had stopped this medication 7 days prior to dosing (∼ 5 half-lives before the AE).

the AGHD patients (58.1 ± 37.29 ng/mL) than in the controls (128.1 ± 52.47 ng/mL, P < .0001) but did not rise significantly after a single dose of macimorelin (53 ± 34.57 ng/mL and 125.9 ± 54.26 ng/mL respectively). For more details, please see Supplemental Data, Supplemental Tables 1 and 2, and Supplemental Figure 2. Figure 2. Mean ± SD (A), scatter plot (B) of peak GH concentrations in response to macimorelin and correlation analysis of BMI and peak GH response to macimorelin (C; controls: n = 48, r2 = −0.368, P = .01; cases: n = 50, r2 = −0.14, P = .33). ROC and CART analysis The ROC plot analysis yielded an optimal GH cut point of 2.7 ng/mL, with 82% sensitivity, 92% specificity, and a 13% misclassification rate. A cut point of 4.5 ng/mL yielded a higher sensitivity (90%) but lower specificity (79%) with a misclassification rate of 15% (Figure 3). The CART analysis of peak GH after macimorelin showed that misclassifications of patient and control subjects are slightly decreased when the covariate predictors of BMI and age are included in the analysis. The ROC area under the curve (AUC) and cut point for peak GH alone are slightly different from those found using the logistic regression modeling approach (Supplemental Table 3). When BMI-specific cut points were used on subgroup analyses, the ROC analysis improved, yielding a sensitivity of 86% and a specificity of 92%, with a misclassification rate of 11% (for BMI < 30 kg/m2, Youden's cut point 6.8 ng/mL, ROC AUC for BMI < 30 kg/m2 = 0.913; for BMI ≥ 30 kg/m2, Youden's cut point 2.7 ng/mL, ROC AUC 0.937).

ific cut points were used on subgroup analyses, the ROC analysis improved, yielding a sensitivity of 86% and a specificity of 92%, with a misclassification rate of 11% (for BMI < 30 kg/m2, Youden's cut point 6.8 ng/mL, ROC AUC for BMI < 30 kg/m2 = 0.913; for BMI ≥ 30 kg/m2, Youden's cut point 2.7 ng/mL, ROC AUC 0.937). Figure 3. ROC curve and CART for analysis of peak GH in response to macimorelin. Post hoc analyses were performed to identify the minimum number of blood draws needed to discriminate AGHD patients from controls. Using a GH threshold of 6.8 ng/mL for lean/overweight subjects (BMI < 30 kg/m2) and a threshold of 2.7 ng/mL for obese subjects (BMI ≥ 30 kg/m2), a sensitivity of 90%, a specificity of 85%, and a 12.2% misclassification rate were attained with a single postdose blood draw at the 45-minute time point. Using time points at 45 and 60 minutes, the sensitivity, specificity, and misclassification rate were 86%, 90%, and 12%, respectively. Using time points at 30, 45, and 60 minutes, the sensitivity, specificity, and misclassification rate were 86%, 92%, and 11%, respectively. The AGHD cases with fewer than 3 deficiencies had a slightly higher mean GH peak value but not significantly different from the cases with at least 3 deficiencies (Supplemental Table 4).

time points at 30, 45, and 60 minutes, the sensitivity, specificity, and misclassification rate were 86%, 92%, and 11%, respectively. The AGHD cases with fewer than 3 deficiencies had a slightly higher mean GH peak value but not significantly different from the cases with at least 3 deficiencies (Supplemental Table 4). Macimorelin vs arginine+GHRH test Subgroup analyses of those individuals who underwent both interventions showed that responses were blunted in AGHD patients after arginine+GHRH and macimorelin in comparison with a brisk GH response in normal controls. Receiver operating characteristic curve analysis showed a numerically better discrimination with macimorelin than with arginine+GHRH, but the difference was not statistically significant (P = .29). In this subgroup of subjects, macimorelin at peak GH responses of 4.3 μg/L provided a sensitivity of 93%, a specificity of 100%, and a misclassification rate of 6% (ROC AUC 0.99), whereas arginine+GHRH at peak GH responses of 7.4 μg/L showed a sensitivity of 88% with a specificity of 90% and a misclassification rate of 12% (ROC AUC 0.94). As shown in Figure 4, the etiology of AGHD (hypothalamic vs pituitary) did not affect the performance of these tests. Figure 4. Correlation analysis of peak GH response to macimorelin and arginine+GHRH. The GH detection limit was 0.05 μg/L.

Macimorelin vs arginine+GHRH test Subgroup analyses of those individuals who underwent both interventions showed that responses were blunted in AGHD patients after arginine+GHRH and macimorelin in comparison with a brisk GH response in normal controls. Receiver operating characteristic curve analysis showed a numerically better discrimination with macimorelin than with arginine+GHRH, but the difference was not statistically significant (P = .29). In this subgroup of subjects, macimorelin at peak GH responses of 4.3 μg/L provided a sensitivity of 93%, a specificity of 100%, and a misclassification rate of 6% (ROC AUC 0.99), whereas arginine+GHRH at peak GH responses of 7.4 μg/L showed a sensitivity of 88% with a specificity of 90% and a misclassification rate of 12% (ROC AUC 0.94). As shown in Figure 4, the etiology of AGHD (hypothalamic vs pituitary) did not affect the performance of these tests. Figure 4. Correlation analysis of peak GH response to macimorelin and arginine+GHRH. The GH detection limit was 0.05 μg/L. Taste perception and test preference Most subjects found the unflavored macimorelin solution to have an unpleasant taste. This was true for both AGHD patients (66.7%) and controls (57.4%). Subjects who did not receive arginine+GHRH (the second portion of the trial) had a better perception of the macimorelin taste (52.1% indicated an unpleasant taste and 41.7% found a neutral taste). Test preference could be evaluated only for the first portion of the trial when subjects received both treatments. In that portion, 72.0% of subjects, significantly more (P = .0019) than random choice, preferred the macimorelin test over the arginine+GHRH test.

oice to hospitalize the subject as a precautionary measure. The subject had been taking citalopram, a drug that was later reported by the Food and Drug Administration (FDA) to be associated with QT prolongation (19), although the patient had stopped this medication 7 days prior to dosing (∼ 5 half-lives before the AE). Among AGHD patients treated with macimorelin, AEs included the unpleasant taste (10 subjects, 19.2%). All other adverse events in the AGHD group treated with macimorelin were reported by a single patient (1.9%) with the exception of diarrhea, which was reported by 2 patients (3.8%). In the matched control group treated with macimorelin, only diarrhea and unpleasant taste were reported by 1 or more subjects (2 subjects each, 4.2%). (See Supplemental Data and Supplemental Table 5 for details.) Discussion Our study demonstrates that a novel oral ghrelin mimetic is both safe and accurate in diagnosing AGHD. The peak GH response after macimorelin treatment allows the establishment of the diagnosis of AGHD with good sensitivity and specificity. Furthermore, this test was well tolerated by patients and does not require parenteral administration of the agent.

oral ghrelin mimetic is both safe and accurate in diagnosing AGHD. The peak GH response after macimorelin treatment allows the establishment of the diagnosis of AGHD with good sensitivity and specificity. Furthermore, this test was well tolerated by patients and does not require parenteral administration of the agent. Although the ITT is considered the standard reference test for diagnosing AGHD, alternative tests are needed because this test is often contraindicated due to the risks associated with hypoglycemia. In addition, performing an ITT may be challenging in some settings because it requires trained personnel, monitored facilities, and other resources that may not be available to every clinician (2). The arginine+GHRH test had emerged as the best alternative, but unfortunately, Geref Diagnostic (Serono) was removed from the US market in 2008 (1). Other currently available tests such as arginine, clonidine, levodopa, and arginine in combination with levodopa have much lower specificity and sensitivity in adults (1).

The arginine+GHRH test had emerged as the best alternative, but unfortunately, Geref Diagnostic (Serono) was removed from the US market in 2008 (1). Other currently available tests such as arginine, clonidine, levodopa, and arginine in combination with levodopa have much lower specificity and sensitivity in adults (1). Ghrelin and its mimetics are potent GH stimulants, and ghrelin has been used to diagnose GH deficiency in animals and very recently in humans (20, 21). GHRH in combination with the ghrelin mimetic GH-releasing hexapeptide-6 was tested as a method for diagnosing AGHD in one study, and it was found to have excellent sensitivity and specificity with good tolerability (22). Administration of GH-releasing hexapeptide-6 alone was not found useful in the diagnosis of GH deficiency in a different report (23). However, none of these drugs are currently available in the United States.

AGHD in one study, and it was found to have excellent sensitivity and specificity with good tolerability (22). Administration of GH-releasing hexapeptide-6 alone was not found useful in the diagnosis of GH deficiency in a different report (23). However, none of these drugs are currently available in the United States. The current multicenter study of macimorelin was conducted to determine its efficacy and safety in the diagnosis of AGHD. Because other GH stimulation tests require iv or im administration of the GH stimulant, macimorelin offers the advantage of being administered orally. In addition, the drug is rapidly absorbed and has good bioavailability, reaching peak serum concentrations 45 minutes after administration (17). Although the unavailability of Geref Diagnostic (Serono) to complete the protocol as initially designed greatly decreased the statistical power of the study to detect a significant difference in the accuracy of the macimorelin vs the arginine+GHRH test in diagnosing AGHD, the results on the subgroup of subjects whom we could test with both agents show that macimorelin is at least as accurate in diagnosing this condition as arginine+GHRH, and it is preferred by patients.

dy to detect a significant difference in the accuracy of the macimorelin vs the arginine+GHRH test in diagnosing AGHD, the results on the subgroup of subjects whom we could test with both agents show that macimorelin is at least as accurate in diagnosing this condition as arginine+GHRH, and it is preferred by patients. A single blood draw at 45 minutes after the dose showed very good sensitivity, specificity, and misclassification rate that was only marginally improved by addition of the 30- and 60-minute blood draws. However, in the clinical setting, providers may choose to perform these 3 draws (30, 45, and 60 minutes) to maximize the chance of a correct diagnosis because missed specimens are not uncommon. This test would also present several advantages over other current alternatives such as the glucagon stimulation test. Only 3 draws are needed after drug administration without the need to follow up the patients for extended periods of time, it does not require parenteral administration of drugs (although it has to be diluted in water before administration), and it is not associated with nausea or symptomatic hypoglycemia.

lation test. Only 3 draws are needed after drug administration without the need to follow up the patients for extended periods of time, it does not require parenteral administration of drugs (although it has to be diluted in water before administration), and it is not associated with nausea or symptomatic hypoglycemia. Peak GH levels in response to macimorelin were inversely related to BMI in control subjects. This is important, given that obesity affects more than 30% of the adult US population and the specificity of GH provocative tests may be decreased in this setting (24, 25). Others have suggested that BMI-specific cut points should be used for GH-provocative tests (26). Our data show that the sensitivity of the macimorelin-stimulated GH test can be improved while maintaining specificity by using different cut points based on BMI level.

ative tests may be decreased in this setting (24, 25). Others have suggested that BMI-specific cut points should be used for GH-provocative tests (26). Our data show that the sensitivity of the macimorelin-stimulated GH test can be improved while maintaining specificity by using different cut points based on BMI level. Macimorelin was well tolerated in our study. The only adverse event reported frequently was dysgeusia (unpleasant taste). Nevertheless, most subjects preferred this test over the arginine+GHRH test. Only 1 subject experienced a serious adverse event of asymptomatic QT prolongation and T wave abnormality on the ECG that resolved overnight. To our knowledge, there are no prior published reports of this or other drugs in this class causing ECG abnormalities. It is worth mentioning that the individual that experienced this AE had been taking citalopram, a drug now known to cause QT prolongation (19). Other known effects of activating the ghrelin receptor including increased appetite and body weight were not seen and are unlikely to happen with a single dose.

abnormalities. It is worth mentioning that the individual that experienced this AE had been taking citalopram, a drug now known to cause QT prolongation (19). Other known effects of activating the ghrelin receptor including increased appetite and body weight were not seen and are unlikely to happen with a single dose. There are several strengths to our study. The use of BMI-, estrogen status-, and age-matched controls decreased the chances of misclassifying individuals due to variability in these variables. Also, a central reference laboratory was used to measure the GH levels. Limitations of our study include the fact that we were not able to perform the arginine+GHRH test in all subjects as originally planned because Geref Diagnostic (Serono) became unavailable during the study and that there was some overlap between AGHD patients and controls. Nevertheless, the sensitivity and specificity thresholds determined by this study were similar to those reported previously for other GH provocative tests (1). Also, the limited number of patients did not allow for a cause-specific (hypothalamic, n = 4, vs pituitary, n = 36) analysis of the data or for further analyses based on other patient features (ie, 3 confirmed hormone deficiencies vs less than 3 hormone deficiencies and a positive confirmatory test). Subjects with diabetes, renal, or hepatic dysfunction also were excluded from this trial. Further studies including a larger number of these patients will be needed to determine the sensitivity and specificity of this test in these scenarios. It is possible that the combination of arginine and macimorelin will cause a greater release of GH than macimorelin alone, given that arginine is thought to potentiate the response to GHRH via inhibition of hypothalamic somatostatin release (27). Future studies also will be needed to address this question.

e scenarios. It is possible that the combination of arginine and macimorelin will cause a greater release of GH than macimorelin alone, given that arginine is thought to potentiate the response to GHRH via inhibition of hypothalamic somatostatin release (27). Future studies also will be needed to address this question. There was a difference in cut points for macimorelin-stimulated peak GH levels for those controls receiving both tests vs the entire cohort. Several factors could have contributed to this difference, including chance differences due to the small sample size of the first cohort (n = 10). The use of this agent to discriminate those patients with hypothalamic disease from pituitary disease including patients with recent irradiation or tumors in the hypothalamus such as hamartomas remains to be seen. Arginine+GHRH is associated with false-normal responses in these circumstances. It may turn out that patients with hypothalamic disease patients will not respond to this agent and provide an important discriminating tool. Further studies using a larger sample size would be needed to definitively answer this question, although this would be possible only if Geref Diagnostic (Serono) becomes available again. Unpleasant taste was the only adverse event reported frequently with macimorelin. Alternative strategies to improve its palatability including using a diluent other than water are being tested in ongoing trials.

er this question, although this would be possible only if Geref Diagnostic (Serono) becomes available again. Unpleasant taste was the only adverse event reported frequently with macimorelin. Alternative strategies to improve its palatability including using a diluent other than water are being tested in ongoing trials. In conclusion, this study shows that oral macimorelin is safe and effective in diagnosing GH deficiency in adults, with sensitivity and specificity comparable with other provocative tests. This novel oral test agent could be a simple, rapid, and convenient alternative, especially for patients in whom ITT is contraindicated in establishing the diagnosis of AGHD that could be performed in most outpatient settings. Abbreviations: AEadverse event AGHDadult GH deficiency AUCarea under the curve BMIbody mass index CARTClassification and Regression Tree ECGelectrocardiogram GHS-R1aGH secretagogue receptor 1a ITTinsulin tolerance test ROCreceiver-operating characteristic. Acknowledgments We thank the patients for their participation and the nursing and research staff for their expert professional support in the conducting of these studies. We also thank Paul Blake and Ines Altemose from Æterna Zentaris, Inc for their support and Patty Ruppel for her statistical help. This study was registered with the trial registration, ClinicalTrials.gov NCT00448747.

he nursing and research staff for their expert professional support in the conducting of these studies. We also thank Paul Blake and Ines Altemose from Æterna Zentaris, Inc for their support and Patty Ruppel for her statistical help. This study was registered with the trial registration, ClinicalTrials.gov NCT00448747. This work was supported by Ardana Biosciences Ltd and later by Æterna Zentaris Inc and by the facilities and services of the Michael E. DeBakey Veterans Affairs Medical Center and the Veterans Affairs Puget Sound Health Care System. All investigators received research support from Ardana Biosciences and/or Æterna Zentaris Inc. Disclosure Summary: J.M.G. is a consultant for Æterna Zentaris Inc and Helsinn Therapeutics Inc and receives research support from Helsinn Therapeutics, Inc. B.M.K.B. has served as a consultant to Æterna Zentaris Inc and received research support from and consulted for Novo Nordisk and Pfizer. K.C.J.Y. has received research support from and consulted for Novo Nordisk and Pfizer. G.R.M. has received research support from Eli Lilly, Genentech, and Pfizer and consulted for Novo Nordisk and Teva. M.E.M. received research support from Ipsen, Eli Lilly & Co, Novo Nordisk, Novartis, and Corcept and has consulted for Novo Nordisk, Corcept, and Novartis. The other authors have nothing to declare.

GH is approved in various countries for treatment of adult GH deficiency (GHD) due to hypothalamic or pituitary disease, pediatric short stature due to several causes, and certain catabolic states (HIV-associated muscle wasting and short bowel syndrome). The increasing use of GH for these indications, for unapproved indications, and for sports doping (1) underscores the importance of safety information from rigorous large-scale studies. This information is especially important for adults with GHD who may receive lifelong GH replacement. Regulatory approval of GH treatment for adult GHD was based on placebo-controlled clinical trials of 6 to 12 months' duration, each with <200 patients (2). Because uncommon adverse drug reactions cannot be detected reliably in studies of this size, postmarketing research programs with larger sample sizes have been conducted to expand the safety data for adult GH replacement (3–6). Although such studies have been generally reassuring, they have not compared the outcomes of GH-treated and GH-untreated patients in a prospective observational cohort. Such comparison is needed because hypopituitarism itself may increase rates of myocardial infarction, cerebrovascular events, malignancies, and overall mortality (7). Scientific societies have also recommended additional surveillance for diabetes, tumor recurrence, de novo tumors, and potential unforeseen adverse effects (8).

mparison is needed because hypopituitarism itself may increase rates of myocardial infarction, cerebrovascular events, malignancies, and overall mortality (7). Scientific societies have also recommended additional surveillance for diabetes, tumor recurrence, de novo tumors, and potential unforeseen adverse effects (8). Unlike GH surveillance programs that lack a control group (9), the US Hypopituitary Control and Complication Study (HypoCCS) was designed to compare the incidence rate of events between GH-treated and untreated GH-deficient adult patients who had similar hypothalamic-pituitary disorders over a prospective follow-up period of 5 years. Here, we report the safety profile of GH treatment in patients with adult GHD compared with similar patients not receiving GH treatment. The mortality rates for both treatment groups were compared with that of the US general population. Subjects and Methods Study design The US HypoCCS was a prospective observational study sponsored by Eli Lilly and Company to examine long-term safety of GH (Humatrope; Eli Lilly and Company, Indianapolis, Indiana) treatment in adults with GHD. Investigators at 157 US centers participated between 1996 and 2002 (10). Data were verified against source documents by monitors reviewing patient records at the sites (data source verification). The 2430 subjects (1988 GH-treated and 442 untreated) who enrolled in US HypoCCS and had follow-up data comprise the focus of this report.

at 157 US centers participated between 1996 and 2002 (10). Data were verified against source documents by monitors reviewing patient records at the sites (data source verification). The 2430 subjects (1988 GH-treated and 442 untreated) who enrolled in US HypoCCS and had follow-up data comprise the focus of this report. In 2002, a new global HypoCCS study was launched, merging the US and European HypoCCS studies, with some differences from the US study in data collection; most notably, data source verification was not performed. Patients who participated in US HypoCCS were allowed to enroll in the new global HypoCCS, although only a subset were subsequently enrolled. To evaluate a selected number of new safety signals identified during analysis of US HypoCCS, an interim analysis of the US patients enrolled in the global study was performed in 2008, with a focus on new patients who had not previously participated in US HypoCCS (1034 GH-treated and 233 untreated). Local institutional review boards approved both protocols; patients provided written informed consent.

S HypoCCS, an interim analysis of the US patients enrolled in the global study was performed in 2008, with a focus on new patients who had not previously participated in US HypoCCS (1034 GH-treated and 233 untreated). Local institutional review boards approved both protocols; patients provided written informed consent. The study design, inclusion/exclusion criteria, and hormone assay methods for US HypoCCS have been described previously (10, 11). Enrolled patients met the criteria for adult GHD as specified in the US package insert for Humatrope (10, 11). According to the observational study design, the choice of whether to receive GH replacement therapy or remain untreated was made by each patient in consultation with his or her endocrinologist. Investigators were allowed to individualize GH treatment based on the clinical and biochemical (serum IGF-I and IGF binding protein-3 [IGFBP-3]) responses of each patient; the protocol recommended a starting dose of not more than 6 μg/kg/d and a maximum dose of 12.5 μg/kg/d based on the US package insert for Humatrope as worded at the beginning of the study. Follow-up visits were at 6-month intervals (±1 month) for both treatment groups. All patients who met the diagnostic criteria and had at least 1 follow-up visit were analyzed.

not more than 6 μg/kg/d and a maximum dose of 12.5 μg/kg/d based on the US package insert for Humatrope as worded at the beginning of the study. Follow-up visits were at 6-month intervals (±1 month) for both treatment groups. All patients who met the diagnostic criteria and had at least 1 follow-up visit were analyzed. Adverse event (AE) reporting A treatment-emergent adverse event (TEAE) was defined as a condition that developed or was present but worsened in severity after enrollment in the study. Serious adverse events (SAEs) were reported as defined by regulatory criteria (life-threatening, hospitalization, severe disability, congenital anomaly, cancer, drug overdose, death, or investigator-designated as serious for other reason). To ensure accuracy of event descriptions, investigators were contacted for follow-up information on all SAEs. AEs were analyzed based on individual preferred terms defined by the Medical Dictionary for Regulatory Activities (MedDRA, version 7.0), as recommended by the International Conference for Harmonization and adopted by most regulatory authorities worldwide. TEAEs were classified as expected if they were previously reported side effects of GH treatment (2, 12).

eferred terms defined by the Medical Dictionary for Regulatory Activities (MedDRA, version 7.0), as recommended by the International Conference for Harmonization and adopted by most regulatory authorities worldwide. TEAEs were classified as expected if they were previously reported side effects of GH treatment (2, 12). Baseline patient characteristics Baseline patient characteristics are summarized in Table 1. Untreated patients were older, had more preexisting medical problems, and were more likely to be male and to have an intracranial tumor (including pituitary adenoma) as the cause of GHD. The proportions of patients with specific pituitary hormone deficiencies and hormonal replacements were similar, except that diabetes insipidus and estrogen replacement (among women) were more commonly encountered in GH-treated patients. Table 1. Patient Characteristics

Baseline patient characteristics Baseline patient characteristics are summarized in Table 1. Untreated patients were older, had more preexisting medical problems, and were more likely to be male and to have an intracranial tumor (including pituitary adenoma) as the cause of GHD. The proportions of patients with specific pituitary hormone deficiencies and hormonal replacements were similar, except that diabetes insipidus and estrogen replacement (among women) were more commonly encountered in GH-treated patients. Table 1. Patient Characteristics Characteristic GH-Treated (n = 1988a) Untreated (n = 442a) P Valueb Follow-up, y 2.3 ± 1.4 2.3 ± 1.6 .555 Age at entry, y 46 ± 15 55 ± 16 <.001 Sex (male/female), % 56/44 62/38 .023 Body mass index, kg/m2 31 ± 7 30 ± 6 .002 History of smoking, y 7 ± 12 9 ± 14 .006 Intracranial tumor as cause of GHD, % 63 76 <.001 Radiotherapy associated with GHD, % 29 32 .257 Isolated GH deficiency, % 13 10 .032 Onset of GHD (adult/childhood), % 84/16 88/12 .023 GH therapy before study entry, % Adult-onset GHD 12 5 <.001 Childhood-onset GHD 72 59 .051 Preexisting medical problem, % Visual impairment 27 34 .002 Hypertension 25 33 .001 Hyperlipidemia 42 50 .003 Diabetes mellitus 8 15 <.001 Coronary artery disease 6 12 <.001 Pituitary hormone deficiency, %c Secondary hypothyroidism 72 75 .163 Thyroid hormone replacement 97 95 .013 Secondary hypoadrenalism 52 56 .083 Glucocorticoid replacement 95 95 .783 Secondary hypogonadism (men) 82 85 .302 Androgen replacement 88 84 .117 Secondary hypogonadism (women) 52 51 .783 Estrogen replacement 64 48 .004 Diabetes insipidus 22 13 <.001 Vasopressin replacement 88 84 .439 Serum IGF-I, μg/L 108 ± 61 90 ± 51 <.001 Serum IGF-I SD score −2.6 ± 1.9 −2.5 ± 1.7 .695 Serum IGFBP-3, μg/L 2.4 ± 0.9 2.1 ± 1.0 <.001 a The number of subjects is smaller for some variables.

hypogonadism (women) 52 51 .783 Estrogen replacement 64 48 .004 Diabetes insipidus 22 13 <.001 Vasopressin replacement 88 84 .439 Serum IGF-I, μg/L 108 ± 61 90 ± 51 <.001 Serum IGF-I SD score −2.6 ± 1.9 −2.5 ± 1.7 .695 Serum IGFBP-3, μg/L 2.4 ± 0.9 2.1 ± 1.0 <.001 a The number of subjects is smaller for some variables. b Comparisons unadjusted for propensity score. χ2 test for categorical variables; t test for continuous variables (mean ± SD shown). c Percentage of patients with specific pituitary hormone deficiencies and of those the percentage receiving hormonal replacement.

hypogonadism (women) 52 51 .783 Estrogen replacement 64 48 .004 Diabetes insipidus 22 13 <.001 Vasopressin replacement 88 84 .439 Serum IGF-I, μg/L 108 ± 61 90 ± 51 <.001 Serum IGF-I SD score −2.6 ± 1.9 −2.5 ± 1.7 .695 Serum IGFBP-3, μg/L 2.4 ± 0.9 2.1 ± 1.0 <.001 a The number of subjects is smaller for some variables. b Comparisons unadjusted for propensity score. χ2 test for categorical variables; t test for continuous variables (mean ± SD shown). c Percentage of patients with specific pituitary hormone deficiencies and of those the percentage receiving hormonal replacement. Statistical analysis To adjust for imbalances in patient characteristics between GH-treated and untreated groups due to the observational study design, stratified propensity score analysis was used for group comparisons of TEAEs and SAEs (13). Propensity scores (conditional probability of being treated) were derived from a logistic regression model that included 37 covariates, selected because of a baseline imbalance between treatment groups or their perceived impact on the occurrence of AEs. (Details of the 37 covariates used in the logistic regression model are found in Supplemental Methods published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org.) Missing quantitative baseline data were imputed by treatment mean imputation (14). Patients were stratified into quintiles based on the propensity scores (15). Consequently, 97% of baseline covariate comparisons within propensity score quintiles showed no statistically significant treatment group differences. TEAE and SAE rates were compared using the Cochran-Mantel-Haenszel test, controlling for the propensity score quintiles (thus controlling for baseline differences). A P value of ≤.05 was considered significant.

omparisons within propensity score quintiles showed no statistically significant treatment group differences. TEAE and SAE rates were compared using the Cochran-Mantel-Haenszel test, controlling for the propensity score quintiles (thus controlling for baseline differences). A P value of ≤.05 was considered significant. Because hundreds of treatment group comparisons were performed, multiplicity adjustment was necessary. The false discovery rate (FDR) method was used to distinguish between probable false positives and probable true findings from the propensity score analysis (16). Different from traditional approaches to multiplicity adjustment (eg, the Bonferroni procedure), which control the probability of making 1 or more false discoveries (type I error), the FDR controls the expected proportion of false positives while maintaining power to uncover real differences. We eliminated from the analysis TEAEs with incidence rates so low (<0.5% in both groups) that achieving a statistically significant treatment difference was unlikely (16). Within each MedDRA system organ class, a P value adjusted to control the FDR was calculated for each of the remaining TEAEs and compared with the prespecified cutoff of 10% (16). For this analysis, a 10% false-positive rate among the significant findings was considered acceptable. We report all TEAEs with an incidence of >1.5% in either treatment group that were significantly more common (P ≤ .05) in GH-treated than in untreated patients and provide the FDR-adjusted P value for interpretation of these events as “probable true positives” or “probable false positives.” Associations between unexpected TEAEs and other potentially related conditions were examined via logistic regression of the unexpected event on the condition of interest, adjusting for therapy and propensity score quintile.

e for interpretation of these events as “probable true positives” or “probable false positives.” Associations between unexpected TEAEs and other potentially related conditions were examined via logistic regression of the unexpected event on the condition of interest, adjusting for therapy and propensity score quintile. Standardized mortality ratios (SMRs) were used to quantify the risk of death among the GH-treated and untreated groups. To compute the SMR, an expected number of deaths was calculated using age- and sex-specific US mortality rates, reported by the National Center for Health Statistics. Person-years of observation for the patient population were applied to age-specific rates for 5-year age intervals to calculate all-cause SMRs and the corresponding 95% confidence interval (CI) (17). Results Follow-up time in study Mean follow-up time did not differ between GH-treated and untreated patients (Table 1). Median (25th percentile, 75th percentile, maximum) follow-up was 2.2 (1.1, 3.5, 5.5) years for GH-treated patients and 2.1 (1.0, 3.8, 5.2) years for untreated patients. Serum IGF-I and IGFBP-3 concentrations Mean baseline serum concentrations of IGF-I and IGFBP-3 are shown in Table 1; serum IGF-I standard deviation scores (based on age-adjusted normative data, measured in a central laboratory) were <−2 in 56% and 54% of GH-treated and untreated patients, respectively (P = .406). At follow-up visits, the percentage of GH-treated patients with elevated (SD score ≥2) and low (SD score <−2) IGF-I concentrations ranged from 0% to 2.0% and 4.8% to 18.5%, respectively.

mative data, measured in a central laboratory) were <−2 in 56% and 54% of GH-treated and untreated patients, respectively (P = .406). At follow-up visits, the percentage of GH-treated patients with elevated (SD score ≥2) and low (SD score <−2) IGF-I concentrations ranged from 0% to 2.0% and 4.8% to 18.5%, respectively. TEAEs Overall, the incidence of TEAEs was higher in GH-treated than in untreated patients (84.1% vs 69.2%, P < .001). Table 2 lists TEAEs that were more common (P ≤ .05) in GH-treated than in untreated patients, after adjusting for baseline differences. After application of the FDR criteria (P ≤ .05 and FDR-adjusted P ≤ .1), most expected TEAEs were retained as probable true positives, with the exception of joint swelling, joint stiffness, and acne. Conversely, several unexpected events were identified as probable false positives. Table 2. TEAEs That Were Significantly More Common (P ≤ .05) in the GH-Treated Group Than in the Untreated Group After Adjusting for Baseline Differencesa

TEAEs Overall, the incidence of TEAEs was higher in GH-treated than in untreated patients (84.1% vs 69.2%, P < .001). Table 2 lists TEAEs that were more common (P ≤ .05) in GH-treated than in untreated patients, after adjusting for baseline differences. After application of the FDR criteria (P ≤ .05 and FDR-adjusted P ≤ .1), most expected TEAEs were retained as probable true positives, with the exception of joint swelling, joint stiffness, and acne. Conversely, several unexpected events were identified as probable false positives. Table 2. TEAEs That Were Significantly More Common (P ≤ .05) in the GH-Treated Group Than in the Untreated Group After Adjusting for Baseline Differencesa AEs GH-Treated (n = 1988) Untreated (n = 442) P Valueb FDR-Adjusted P Value Expected events Probable true positive, n (%) Arthralgia 397 (20.0) 32 (7.2) <.001 <.001 Edema peripheral 307 (15.4) 26 (5.9) <.001 <.001 Back pain 175 (8.8) 18 (4.1) .004 .056 Hypoesthesia 103 (5.2) 10 (2.3) .009 .098 Myalgia 100 (5.0) 9 (2.0) .006 .059 Energy increased 90 (4.5) 0 (0.0) <.001 <.001 Paresthesia 85 (4.3) 7 (1.6) .009 .098 Carpal tunnel syndrome 79 (4.0) 5 (1.1) .014 .103 Hormone level abnormalc 43 (2.2) 1 (0.2) .003 .037 IGF-I increasedd 37 (1.9) 1 (0.2) .008 .055 Probable false positive, n (%) Joint swelling 51 (2.6) 5 (1.1) .037 .213 Acne 40 (2.0) 1 (0.2) .046 .228 Joint stiffness 35 (1.8) 1 (0.2) .021 .154 Unexpected events Probable true positive, n (%) Insomnia 128 (6.4) 12 (2.7) .014 .090 Dyspnea 83 (4.2) 9 (2.0) .006 .090 Anxiety 67 (3.4) 4 (0.9) .021 .090 Sleep apnea syndrome 66 (3.3) 4 (0.9) .010 .098 Libido decreased 42 (2.1) 1 (0.2) .016 .090 Probable false positive, n (%) Headache 276 (13.9) 43 (9.7) .038 .182 Depression 195 (9.8) 23 (5.2) .041 .127 Nausea 163 (8.2) 24 (5.4) .040 .417 Hypertension 144 (7.2) 23 (5.2) .050 .132 Nasopharyngitis 92 (4.6) 11 (2.5) .042 .503 Abdominal distension 33 (1.7) 2 (0.5) .045 .417 Asthma 33 (1.7) 1 (0.2) .044 .350 a Events (MedDRA preferred terms) were defined as “expected” or “unexpected” based on previous studies. An event was considered a “probable true positive” if FDR-adjusted P ≤ .1. AEs with an incidence >0.5% in either group were included in the statistical analysis; events with an incidence of >1.5% are shown in the table.

vents (MedDRA preferred terms) were defined as “expected” or “unexpected” based on previous studies. An event was considered a “probable true positive” if FDR-adjusted P ≤ .1. AEs with an incidence >0.5% in either group were included in the statistical analysis; events with an incidence of >1.5% are shown in the table. b Cochran-Mantel-Haenszel general association. c Includes abnormal serum concentrations of IGF-I, IGFBP-3, and dehydroepiandrosterone. d Indicates an increase in serum IGF-I that the investigator considered to be an AE. The incidence of TEAEs related to benign or malignant neoplasms did not differ between groups (GH-treated, 8.1%; untreated, 10.0%; P = .77). For MedDRA preferred terms related to glucose metabolism and diabetes mellitus, frequencies of TEAEs were <2%, and there were no significant group differences. Cardiac or vascular disorders were not identified as probable true-positive findings. Unexpected events in GH-treated patients Five TEAEs in the GH-treated group were identified as probable true positives but were unexpected based on previous studies (Table 2). For GH-treated patients, reported events of insomnia, dyspnea, anxiety, sleep apnea, and decreased libido were mild to moderate in severity (as reported by the investigator) in 93%, 92%, 82%, 85%, and 90% of patients, respectively, and were reported during the first year of GH treatment in 58%, 61%, 52%, 44%, and 55% of patients, respectively.

reported events of insomnia, dyspnea, anxiety, sleep apnea, and decreased libido were mild to moderate in severity (as reported by the investigator) in 93%, 92%, 82%, 85%, and 90% of patients, respectively, and were reported during the first year of GH treatment in 58%, 61%, 52%, 44%, and 55% of patients, respectively. To further clarify the nature of these unexpected TEAEs, possible clinical associations were examined. Sleep apnea was associated with both obesity (73% had a body mass index ≥30 kg/m2; P = .003) and fluid retention (P < .001). Dyspnea was associated with fluid retention (P < .001). Insomnia was associated with anxiety (P < .001); both insomnia and anxiety were associated with painful AEs (such as musculoskeletal disorders and others) (P < .001). Decreased libido was not associated with anxiety, depression, or painful AEs. Among patients reporting dyspnea, 78% had a related disorder or symptom (cardiac or respiratory disorder, infection, or edema). Among those reporting anxiety, 33% either had anxiety listed as a preexisting condition or were taking anxiolytic medication at baseline. For those with decreased libido, most were hypogonadal and/or menopausal at baseline (87% women and 85% men); a minority of these hypogonadal patients were not receiving hormonal replacement when decreased libido was reported (15% women and 26% men).

eexisting condition or were taking anxiolytic medication at baseline. For those with decreased libido, most were hypogonadal and/or menopausal at baseline (87% women and 85% men); a minority of these hypogonadal patients were not receiving hormonal replacement when decreased libido was reported (15% women and 26% men). Compared with the overall study population followed for 2 years, GH-treated patients reporting sleep apnea more commonly had higher (≥0.8 mg/d) GH doses (26% vs 14% for all GH-treated patients) and elevated serum IGF-I concentrations (35% vs 14% for all GH-treated patients), but this association was not observed for the other 4 unexpected events. During follow-up, reduction in clinical severity was reported in 28%, 49%, 25%, 18%, and 19% of GH-treated patients experiencing insomnia, dyspnea, anxiety, sleep apnea, and decreased libido, respectively. Of these, a reduction of the GH dose was associated with clinical severity decrease in 25%, 17%, 35%, 33%, and 50% of those subjects with insomnia, dyspnea, anxiety, sleep apnea, and decreased libido, respectively.

reated patients experiencing insomnia, dyspnea, anxiety, sleep apnea, and decreased libido, respectively. Of these, a reduction of the GH dose was associated with clinical severity decrease in 25%, 17%, 35%, 33%, and 50% of those subjects with insomnia, dyspnea, anxiety, sleep apnea, and decreased libido, respectively. To evaluate these unexpected findings further, an independent interim analysis of US patients subsequently enrolled in the global HypoCCS was performed in 2008 (mean follow-up period, 2.0 years; range, 0.1–5.4 years). This analysis examined (using the same methodology) 4 events of particular interest (dyspnea, sleep apnea, hypertension, and decreased libido) based on the findings reported herein, reviews of published literature, and the Lilly spontaneous AE database. Among US patients who had not previously participated in US HypoCCS (1034 GH-treated and 233 untreated), there were no statistically significant differences in the proportion of GH-treated and untreated patients reporting dyspnea (1.5% vs 0.9%, respectively), sleep apnea (1.6% vs 2.6%), hypertension (3.8% vs 2.1%), or decreased libido (0.2% vs 0%).

previously participated in US HypoCCS (1034 GH-treated and 233 untreated), there were no statistically significant differences in the proportion of GH-treated and untreated patients reporting dyspnea (1.5% vs 0.9%, respectively), sleep apnea (1.6% vs 2.6%), hypertension (3.8% vs 2.1%), or decreased libido (0.2% vs 0%). SAEs After controlling for baseline group differences, the proportion of GH-treated and untreated patients experiencing death, cancer, or benign extracranial tumors or cysts (de novo or recurrent) did not differ (Tables 3, 4, and 5). Among patients with a previous intracranial tumor and evaluable baseline data, no difference in growth or recurrence rates of pituitary adenoma, craniopharyngioma, or other intracranial tumors was observed between GH-treated and untreated patients (Table 6). Table 3. Summary of Number of SAEs Reported as Deathsa

SAEs After controlling for baseline group differences, the proportion of GH-treated and untreated patients experiencing death, cancer, or benign extracranial tumors or cysts (de novo or recurrent) did not differ (Tables 3, 4, and 5). Among patients with a previous intracranial tumor and evaluable baseline data, no difference in growth or recurrence rates of pituitary adenoma, craniopharyngioma, or other intracranial tumors was observed between GH-treated and untreated patients (Table 6). Table 3. Summary of Number of SAEs Reported as Deathsa GH-Treated (n = 1988) Untreated (n = 442) Cause of death, n (%) Cause unknown or not specified 9 (0.45) 3 (0.68) Cardiac arrhythmia or arrest 6 (0.30) 1 (0.23) Sepsis, pneumonia, or other infection 3 (0.15) 1 (0.23) Myocardial infarction 2 (0.10) 0 Cerebrovascular accident or cerebral hemorrhage 1 (0.05) 2 (0.45) Suicide 3 (0.15) 0 Respiratory failure or arrest 3 (0.15) 0 Cardiac failure 1 (0.05) 2 (0.45) Motor vehicle accident 1 (0.05) 2 (0.45) Acute myeloid leukemia 1 (0.05) 0 Astrocytoma (malignant) 1 (0.05) 0 Ruptured aortic aneurysm 1 (0.05) 0 Carbon monoxide poisoning 1 (0.05) 0 Total deaths, n (%) 33 (1.66) 11 (2.49) a There was no significant difference in the total proportions of patients dying in the 2 groups after controlling for baseline group differences (P = .73). Table 4. Summary of Number of SAEs Related to Cancer (Either New Cancers or Recurrence of Previous Disease)a

GH-Treated (n = 1988) Untreated (n = 442) Cause of death, n (%) Cause unknown or not specified 9 (0.45) 3 (0.68) Cardiac arrhythmia or arrest 6 (0.30) 1 (0.23) Sepsis, pneumonia, or other infection 3 (0.15) 1 (0.23) Myocardial infarction 2 (0.10) 0 Cerebrovascular accident or cerebral hemorrhage 1 (0.05) 2 (0.45) Suicide 3 (0.15) 0 Respiratory failure or arrest 3 (0.15) 0 Cardiac failure 1 (0.05) 2 (0.45) Motor vehicle accident 1 (0.05) 2 (0.45) Acute myeloid leukemia 1 (0.05) 0 Astrocytoma (malignant) 1 (0.05) 0 Ruptured aortic aneurysm 1 (0.05) 0 Carbon monoxide poisoning 1 (0.05) 0 Total deaths, n (%) 33 (1.66) 11 (2.49) a There was no significant difference in the total proportions of patients dying in the 2 groups after controlling for baseline group differences (P = .73). Table 4. Summary of Number of SAEs Related to Cancer (Either New Cancers or Recurrence of Previous Disease)a GH-Treated (n = 1988) Untreated (n = 442) Cancer type, n (%) Skin cancerb 10 (0.50) 5 (1.13) Prostate cancer 3 (0.15) 3 (0.68) Breast cancer 4 (0.20) 1 (0.23) Lung cancerc 4 (0.20) 0 Colorectal cancer 2 (0.10) 1 (0.23) Acute leukemia 2 (0.10) 0 Carcinoid tumor 0 1 (0.23) Lymphoma 0 1 (0.23) Ovarian cancer 1 (0.05) 0 Ewing's sarcoma 1 (0.05) 0 Pancreatic islet cell tumor 1 (0.05) 0 Bladder/urethral cancer 1 (0.05) 0 Fibrosarcoma 1 (0.05) 0 Laryngeal cancer 1 (0.05) 0 Polycythemia vera 1 (0.05) 0 Total cancers, n (%) 32 (1.61) 12 (2.71) a For total cancer events, there was no significant difference between the 2 groups after controlling for baseline differences (P = .57).

let cell tumor 1 (0.05) 0 Bladder/urethral cancer 1 (0.05) 0 Fibrosarcoma 1 (0.05) 0 Laryngeal cancer 1 (0.05) 0 Polycythemia vera 1 (0.05) 0 Total cancers, n (%) 32 (1.61) 12 (2.71) a For total cancer events, there was no significant difference between the 2 groups after controlling for baseline differences (P = .57). b Skin cancers comprised 10 basal cell carcinomas (6 GH-treated patients and 4 untreated patients), 3 melanomas (3 GH-treated patients), and 2 squamous cell carcinomas (1 GH-treated patient and 1 untreated patient); 1 of the patients with a melanoma also had a basal cell carcinoma that is not included in the table. c Includes 1 patient with a pulmonary mass that was presumed to be a lung cancer by the oncologist, but a tissue diagnosis was not obtained. Table 5. Summary of Number of SAEs Related to Benign Extracranial Tumors or Cysts (Either New Tumors or Progression of Preexisting Tumors)a GH-Treated (n = 1988) Untreated (n = 442) Tumor type, n (%) Uterine leiomyoma 3 (0.15) 0 Ovarian cyst or adenoma 3 (0.15) 0 Hemangioma 2 (0.10) 0 Histiocytosis 0 1 (0.23) Colon adenoma 1 (0.05) 0 Lipoma 1 (0.05) 0 Synovial cyst 1 (0.05) 0 Total benign tumors and cysts, n (%) 11 (0.55) 1 (0.23) a For total events, there was no significant difference between the 2 groups after controlling for baseline group differences (P = .46). Table 6. SAEs Related to Growth or Recurrence of Preexisting Intracranial Tumors Including Pituitary Adenomasa

GH-Treated (n = 1988) Untreated (n = 442) Tumor type, n (%) Uterine leiomyoma 3 (0.15) 0 Ovarian cyst or adenoma 3 (0.15) 0 Hemangioma 2 (0.10) 0 Histiocytosis 0 1 (0.23) Colon adenoma 1 (0.05) 0 Lipoma 1 (0.05) 0 Synovial cyst 1 (0.05) 0 Total benign tumors and cysts, n (%) 11 (0.55) 1 (0.23) a For total events, there was no significant difference between the 2 groups after controlling for baseline group differences (P = .46). Table 6. SAEs Related to Growth or Recurrence of Preexisting Intracranial Tumors Including Pituitary Adenomasa Intracranial Tumor GH-Treated Untreated P Value Pituitary adenomas 24/965 (2.5%) 12/273 (4.4%) .21 Craniopharyngiomas 8/211 (3.8%) 2/36 (5.6%) .82 Other intracranial tumorsb 4/133 (3.0%) 1/30 (3.3%) .90 a For each category, the denominator is patients with a history of the tumor before study entry. P values have been adjusted for baseline group differences. b The other intracranial tumor events included 2 meningiomas (both GH-treated patients), 1 astrocytoma (GH-treated patient), 1 medulloblastoma (GH-treated patient), and 1 Rathke cleft cyst (untreated patient). Comparison with US general population The all-cause SMR was not increased in either treatment group (GH-treated, 0.86 [95% CI, 0.59–1.21]; untreated, 0.58 [95% CI, 0.29–1.04]) and did not differ significantly between GH-treated and untreated patients after 4655 and 1019 person-years of observation, respectively (P = .26).

Comparison with US general population The all-cause SMR was not increased in either treatment group (GH-treated, 0.86 [95% CI, 0.59–1.21]; untreated, 0.58 [95% CI, 0.29–1.04]) and did not differ significantly between GH-treated and untreated patients after 4655 and 1019 person-years of observation, respectively (P = .26). Discussion We compared the safety profile of GH treatment vs no treatment in patients with adult GHD in the setting of routine clinical practice and identified unexpected AEs not previously associated with GH treatment. Notably, sleep apnea and dyspnea were identified as new risks of GH treatment. Although the mean follow-up period (2.3 years) was of insufficient duration to be conclusive, there was no increased rate of death, new cancer, intracranial tumor recurrence, diabetes mellitus, or cardiovascular events in GH-treated patients compared with untreated patients. In addition, the SMR compared with that of the general US population was not increased in either group.

ficient duration to be conclusive, there was no increased rate of death, new cancer, intracranial tumor recurrence, diabetes mellitus, or cardiovascular events in GH-treated patients compared with untreated patients. In addition, the SMR compared with that of the general US population was not increased in either group. A prerequisite to these findings was use of an analytical approach that would reduce biases inherent in the observational study design. Untreated patients were older, sicker, and more likely to have an intracranial tumor as the cause of GHD than GH-treated patients. Thus, the decision whether to replace GH was influenced by medical history. This selection bias was reduced by stratifying analyses within patient subgroups, balanced for baseline covariates, using the propensity score method (13, 15). In addition, we used a statistical method to control the proportion of false-positive findings while maintaining the ability to detect new safety concerns (16). These methods provided robust analysis of observational data.

patient subgroups, balanced for baseline covariates, using the propensity score method (13, 15). In addition, we used a statistical method to control the proportion of false-positive findings while maintaining the ability to detect new safety concerns (16). These methods provided robust analysis of observational data. Because an association has been reported between serum IGF-I levels and prostate, breast, and colon cancer risk, increases in GH/IGF-I levels during medically indicated GH replacement have been considered a potential safety concern (18, 19). Although specific cancers in the current study were too few for analysis by tumor type, no concerning trends were noted, considering the greater than 4-fold difference in treatment group size. Similarly, the absence of a GH effect on overall de novo cancer occurrence and intracranial tumor recurrence is reassuring and consistent with previous reports in adults with GHD (3, 4, 6, 20, 21), with the caveat that longer-term follow-up is needed. In addition, a recent analysis of global HypoCCS data revealed a standardized incidence ratio for all cancers of 0.88 (95% CI, 0.74–1.04) in GH-treated patients globally and, among US patients, the standardized incidence ratio was 0.94 (95% CI, 0.73–1.18) for GH-treated patients and 1.16 (95% CI, 0.76–1.69) for untreated patients (6). In long-term follow-up studies of adults previously treated during childhood with human pituitary GH, the risk of dying from cancer (specifically, colorectal cancer and Hodgkin disease) was increased in 1 UK study of 1848 patients (22) but not in a US study of 6107 patients (23). All-type cancer-related mortality was not increased in recent reports from the Safety and Appropriateness of Growth Hormone Treatments in Europe study (24, 25). Childhood cancer survivors subsequently treated with GH had an increased risk of a second neoplasm but not of recurrence of the first neoplasm (26, 27). Patients with childhood-onset GHD represented a small minority of subjects enrolled in US HypoCCS, so they were not analyzed separately.

ents in Europe study (24, 25). Childhood cancer survivors subsequently treated with GH had an increased risk of a second neoplasm but not of recurrence of the first neoplasm (26, 27). Patients with childhood-onset GHD represented a small minority of subjects enrolled in US HypoCCS, so they were not analyzed separately. Most AEs identified as GH-related in the current study were expected based on previous short-term controlled trials (2, 12), supporting the validity of the current statistical approach. Although GH replacement in adults with GHD may result in increased fasting glucose concentrations (10), the incidence of diabetes mellitus was not significantly increased, in agreement with a recent report (5). However, 5 unexpected AEs were encountered as significantly more frequent in GH-treated than in untreated patients: sleep apnea, dyspnea, insomnia, anxiety, and decreased libido (occurrence rates, 2%–6%). These events often occurred in patients with predisposing factors, such as high body mass index (sleep apnea) or cardiopulmonary disease (dyspnea), or in whom the condition was either preexisting and already treated (anxiety) or chronologically related to known AEs likely to explain the symptoms (sleep apnea and dyspnea associated with fluid retention or insomnia associated with musculoskeletal pain). Events were observed commonly during the first year of GH treatment, and a substantial proportion decreased in severity after GH dose reduction. Although our analysis sought to constrain the false-positive rate to 10%, it is possible that 1 or 2 of these 5 events may be a false-positive finding.

musculoskeletal pain). Events were observed commonly during the first year of GH treatment, and a substantial proportion decreased in severity after GH dose reduction. Although our analysis sought to constrain the false-positive rate to 10%, it is possible that 1 or 2 of these 5 events may be a false-positive finding. The increased incidences of sleep apnea and dyspnea in GH-treated than in untreated patients are important unexpected findings. Our results suggest that sleep apnea may be unmasked or precipitated by GH replacement in patients already at risk for sleep apnea (eg, obese patients), particularly those receiving higher GH doses and those exhibiting higher serum IGF-I levels. This finding is consistent with observations in untreated acromegaly (28) and in GH-treated patients with Prader-Willi syndrome (29). Results of polysomnographic studies in adults with GHD have been conflicting. One series of 5 cases described improvement in sleep apnea after discontinuation of GH (30). However, a prospective, uncontrolled study using lower doses of GH reported a high prevalence of sleep apnea in untreated adults with GHD (12 of 19 patients) but no induction or aggravation of sleep apnea with GH treatment (31). Our results also suggest that GH therapy, which commonly causes dose-related edema, may increase the risk of dyspnea in patients with cardiac and respiratory disorders. These findings emphasize the importance of careful GH dose titration to achieve IGF-I levels within the age-adjusted normal range and to avoid edema (8), particularly in patients with preexisting obesity or cardiopulmonary disorders.

, may increase the risk of dyspnea in patients with cardiac and respiratory disorders. These findings emphasize the importance of careful GH dose titration to achieve IGF-I levels within the age-adjusted normal range and to avoid edema (8), particularly in patients with preexisting obesity or cardiopulmonary disorders. After US HypoCCS closed in 2002, individualized GH dosing, which decreases the occurrence of edema, has become more common in clinical practice (8). Our 2008 interim analysis of US patients enrolled in global HypoCCS did not identify a significant difference in reported rates for dyspnea and sleep apnea between GH-treated and untreated patients. Factors other than individualized dosing that may explain this difference include the smaller number of patients in the interim analysis (1267 vs 2430), differences in study design (no source data verification in global HypoCCS), and changes in patient selection over time. The diagnostic characteristics of GH-deficient patients included in the 3 HypoCCS protocols over the period 1996–2005 have changed significantly over this decade; for example, the proportion of patients harboring pituitary adenomas before entry has decreased from 50.2% to 38.6% (32).

ges in patient selection over time. The diagnostic characteristics of GH-deficient patients included in the 3 HypoCCS protocols over the period 1996–2005 have changed significantly over this decade; for example, the proportion of patients harboring pituitary adenomas before entry has decreased from 50.2% to 38.6% (32). Mild to moderate anxiety and insomnia were associated with known GH AEs such as musculoskeletal pain. Anxiety associated with initiation of GH replacement may also result in insomnia. The finding of decreased libido may reflect reporting bias; it was not associated with anxiety, depression, or painful AEs, and only a minority of patients had untreated hypogonadism. In addition, in the 2008 interim analysis of US patients who enrolled in the global HypoCCS after 2002, only 0.2% of GH-treated patients reported decreased libido. Quality-of-life data were not collected in the current study, but results from the European arm of HypoCCS showed improved quality of life during GH replacement in routine clinical practice (33). In a separate study, the ability to become sexually aroused was significantly decreased at baseline in GH-deficient adults compared with age- and sex-matched control subjects and improved after 6 months of GH treatment (34).

ypoCCS showed improved quality of life during GH replacement in routine clinical practice (33). In a separate study, the ability to become sexually aroused was significantly decreased at baseline in GH-deficient adults compared with age- and sex-matched control subjects and improved after 6 months of GH treatment (34). This observational study has several inherent limitations. Importantly, the mean follow-up period is short in relation to the longer latency period for cancers to appear. Thus, these results cannot exclude the possibility that longer-term GH treatment might be associated with a higher risk for cancer. Nonetheless, our study was of sufficient size and duration to identify TEAEs not previously identified. The cause of death was not ascertained for 12 of 44 reported deaths, despite investigator efforts to obtain these data from the primary care physicians. The propensity score methodology cannot control for unmeasured baseline imbalances, and it is not known whether there were socioeconomic differences between GH-treated and untreated patients. Because the study was not blinded, reporting bias probably occurred, consisting potentially of closer monitoring of GH-treated patients by physicians and increased reporting of unrelated AEs by patients. The latter, termed the nocebo phenomenon, occurs commonly when patients starting new medications have had AEs with other drugs or have preexisting symptoms that predispose them to attribute new or worsening events to the medication (35). No statistical methods are available to control for reporting bias.

s by patients. The latter, termed the nocebo phenomenon, occurs commonly when patients starting new medications have had AEs with other drugs or have preexisting symptoms that predispose them to attribute new or worsening events to the medication (35). No statistical methods are available to control for reporting bias. In summary, this observational US study of GH-treated and GH-untreated patients with adult GHD demonstrates a safety profile for adult GH replacement therapy similar to that observed in clinical trials. At a mean follow-up of 2.3 years, there is no evidence for an effect of GH therapy on deaths, cancers, diabetes mellitus, cardiovascular events, and intracranial tumor growth or recurrence. Careful GH dose titration is recommended for patients who may be at risk for sleep apnea or cardiopulmonary disorders. As treatment paradigms evolve and long-term surveillance continues, changes in the safety profile of GH replacement therapy may be expected. Notably, GH side effects reported here should not be extrapolated to nonapproved GH uses in pituitary-replete adults (1). Abbreviations: AEadverse event CIconfidence interval FDRfalse discovery rate GHDgrowth hormone deficiency HypoCCSHypopituitary Control and Complication Study IGFBP-3IGF binding protein-3 MedDRAMedical Dictionary for Regulatory Activities SAEserious adverse event SMRstandardized mortality ratio TEAEtreatment-emergent adverse event.

In summary, this observational US study of GH-treated and GH-untreated patients with adult GHD demonstrates a safety profile for adult GH replacement therapy similar to that observed in clinical trials. At a mean follow-up of 2.3 years, there is no evidence for an effect of GH therapy on deaths, cancers, diabetes mellitus, cardiovascular events, and intracranial tumor growth or recurrence. Careful GH dose titration is recommended for patients who may be at risk for sleep apnea or cardiopulmonary disorders. As treatment paradigms evolve and long-term surveillance continues, changes in the safety profile of GH replacement therapy may be expected. Notably, GH side effects reported here should not be extrapolated to nonapproved GH uses in pituitary-replete adults (1). Abbreviations: AEadverse event CIconfidence interval FDRfalse discovery rate GHDgrowth hormone deficiency HypoCCSHypopituitary Control and Complication Study IGFBP-3IGF binding protein-3 MedDRAMedical Dictionary for Regulatory Activities SAEserious adverse event SMRstandardized mortality ratio TEAEtreatment-emergent adverse event. Acknowledgments We thank the US HypoCCS investigators, study coordinators, and numerous Lilly employees for their dedication to this study and Peter Bates (Cambridge Medical Writing Services, United Kingdom) for assistance with preparation of the first draft of the manuscript. The US investigators and members of the International HypoCCS Advisory Board have been reported previously (10). This work was supported by Eli Lilly and Company.

Acknowledgments We thank the US HypoCCS investigators, study coordinators, and numerous Lilly employees for their dedication to this study and Peter Bates (Cambridge Medical Writing Services, United Kingdom) for assistance with preparation of the first draft of the manuscript. The US investigators and members of the International HypoCCS Advisory Board have been reported previously (10). This work was supported by Eli Lilly and Company. Disclosure Summary: M.L.H., B.J.C., and A.G.Z. are employees of Lilly and have equity interests in Lilly. R.X., W.W.W., G.B.C., and J.J.C. were previously employed by Lilly and have equity interests in Lilly. D.L.K. and S.M. have received research grants from Lilly. L.L.R., E.M.E., D.L.K., W.W.W., G.B.C., and S.M. have received consulting fees from Lilly.

Uterine natural killer (uNK) cells, an important component of the innate immune system, are the most abundant immune cells in midluteal (periimplantation) endometrium and in the decidua of early pregnancy (1). They represent a unique subset of natural killer cells, staining intensely for CD56 but not for CD16 antigens. uNK cells play a significant role in the establishment and maintenance of early pregnancy by promoting decidual angiogenesis, spiral artery remodeling, and trophoblast invasion (2, 3). In contrast to their circulating (CD56+/CD16+) counterparts, there is little evidence for a cytotoxic role of uNK cells at the fetomaternal interface. However, uNK cells express killer cell immunoglobulin-like receptors that preferentially bind to human leukocyte antigen-C molecules expressed on placental cells, suggesting a role in maternal allorecognition of fetal trophoblast (4). They are abundant around the spiral arteries, near endometrial glands, and adjacent to extravillous trophoblast in early pregnancy. Thus, uNK cells are unique in terms of their tissue distribution, phenotype, and function.

essed on placental cells, suggesting a role in maternal allorecognition of fetal trophoblast (4). They are abundant around the spiral arteries, near endometrial glands, and adjacent to extravillous trophoblast in early pregnancy. Thus, uNK cells are unique in terms of their tissue distribution, phenotype, and function. Both the maternal killer cell immunoglobulin-like receptor and fetal human leukocyte antigen-C gene systems are highly polymorphic and certain genotypic combinations are associated with a modest increase or decrease in pregnancy complications, including miscarriage, fetal growth restriction, and preeclampsia (4). In addition, several studies reported an association between elevated uNK cell levels in midluteal endometrium and reproductive failure (4–6). In particular, there is compelling evidence to link increased uNK density to recurrent pregnancy loss (RPL), defined here as three or more consecutive miscarriages. RPL is a prevalent disorder that affects 1%–2% of couples and a cause of considerable physical and psychological morbidity (7). Furthermore, RPL is associated with an increased likelihood of obstetric complications and adverse perinatal outcome in a subsequent ongoing pregnancy (8). Whether midluteal uNK cell testing in a nonconception cycle predicts subsequent pregnancy complications remains unresolved (9).

cal and psychological morbidity (7). Furthermore, RPL is associated with an increased likelihood of obstetric complications and adverse perinatal outcome in a subsequent ongoing pregnancy (8). Whether midluteal uNK cell testing in a nonconception cycle predicts subsequent pregnancy complications remains unresolved (9). Resident human endometrial stromal cells (HESCs) are thought to serve as gatekeepers for the recruitment and distribution of immune cells in the periimplantation endometrium (10). For example, decidualizing (differentiating) HESCs secrete IL-11 and IL-15, two multifaceted cytokines implicated in trafficking and differentiation of uNK cells (11–13). uNK cells express the glucocorticoid receptor (GR) but lack progesterone receptor, rendering them directly responsive to cortisol but not progesterone (14, 15). Consistent with this notion, preconceptual glucocorticoid (prednisolone) treatment significantly reduces uNK cell density in RPL subjects as well as inhibiting endometrial angiogenesis (16, 17). We recently demonstrated that progesterone massively enhances the expression and activity of 11β-hydroxysteroid dehydrogenase type 1 (11βHSD1) in decidualizing HESCs (18), suggesting that local cortisol biosynthesis plays an integral role in the preparation of the endometrium for implantation. Decidualization is further associated with a decline in GR expression and reciprocal induction of the mineralocorticoid receptor (MR), which in turn drives the expression of several key enzymes involved in lipid and retinoid metabolism, including retinol saturase (RETSAT), various short-chain dehydrogenases/reductases (such as dehydrogenase/reductase (DHRS) member 3, DHRS4, and DHRS4L2), and steroidogenic acute regulatory protein-related lipid transfer protein 5 (18).

ves the expression of several key enzymes involved in lipid and retinoid metabolism, including retinol saturase (RETSAT), various short-chain dehydrogenases/reductases (such as dehydrogenase/reductase (DHRS) member 3, DHRS4, and DHRS4L2), and steroidogenic acute regulatory protein-related lipid transfer protein 5 (18). Emerging evidence suggests that aberrant differentiation of resident HESCs into specialized decidual cells is the hallmark of RPL (19, 20). Taken together, these observations raise the possibility that excessive uNK cell levels in midluteal endometrial samples reflect relative local corticosteroid deficiency, caused by inadequate induction of decidual 11βHSD1 and result in impaired local metabolic function.

d decidual cells is the hallmark of RPL (19, 20). Taken together, these observations raise the possibility that excessive uNK cell levels in midluteal endometrial samples reflect relative local corticosteroid deficiency, caused by inadequate induction of decidual 11βHSD1 and result in impaired local metabolic function. Materials and Methods Patient selection and endometrial sampling The study was approved by the local ethics committee (1997/5065). Subjects were recruited in the Implantation Clinic, a dedicated research clinic at University Hospitals Coventry and Warwickshire National Health Service Trust for patients suffering RPL or recurrent in vitro fertilization treatment failure. Written informed consent was obtained prior to tissue collection. Endometrial biopsies were timed between 7 and 10 days after the preovulatory LH surge. Samples were obtained using a Wallach Endocell sampler (Wallach) under ultrasound guidance, starting from the uterine fundus and moving downward to the internal cervical ostium. Each biopsy was divided, with one part fixed in formalin for immunohistochemistry and the other processed for primary cell culture. The demographic details of participating subjects are summarized in Supplemental Tables 1 and 2, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org.

h biopsy was divided, with one part fixed in formalin for immunohistochemistry and the other processed for primary cell culture. The demographic details of participating subjects are summarized in Supplemental Tables 1 and 2, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org. Primary cell culture HESCs were isolated, cultured, and maintained as described (20). Primary cultures were passaged once, allowed to grow to confluency, and then decidualized with 0.5 mM 8-bromoadenosine cAMP (8-bromo-cAMP; Sigma), 1 μM progesterone (P4; Sigma), and 0.1 μM cortisone (E; Sigma). Cortisone, which is inactive, was added to decidualizing HESC cultures as the substrate for endogenous conversion by 11βHSD11 to cortisol (18).

e, allowed to grow to confluency, and then decidualized with 0.5 mM 8-bromoadenosine cAMP (8-bromo-cAMP; Sigma), 1 μM progesterone (P4; Sigma), and 0.1 μM cortisone (E; Sigma). Cortisone, which is inactive, was added to decidualizing HESC cultures as the substrate for endogenous conversion by 11βHSD11 to cortisol (18). Immunohistochemistry Five-micrometer-thick formalin-fixed, paraffin-embedded tissue sections were labeled with antibody to CD56 (NCL-CD56-1B6; Novacastra) using standard methods and detection systems (3). The uNK cell density was determined as the percentage of uNK cells within the stromal cell population. Because uNK cell density varies with endometrial depth, counting of CD56+ cells was confined to the stroma underlying the luminal epithelium. Five randomly selected high-power magnification fields per biopsy were assessed using ImageJ software (Rasband, W. S., ImageJ, National Institutes of Health) to minimize interobserver variability (21, 22). Normal uNK cell density was defined as 5% or less CD56+ cells in the stroma underlying the luminal epithelium (17, 23). A Mirax Midi slide scanner was used to scan bright-field sections with a ×20 objective with a resolution of 0.23 μm/pixel. This produces images that can be dynamically manipulated within the viewer software, allowing optical magnifications up to ×20 and digital magnification to ×200.

ng the luminal epithelium (17, 23). A Mirax Midi slide scanner was used to scan bright-field sections with a ×20 objective with a resolution of 0.23 μm/pixel. This produces images that can be dynamically manipulated within the viewer software, allowing optical magnifications up to ×20 and digital magnification to ×200. Tissue microarray (TMA) Areas of interest, ie, subepithelial regions, were spotted and tissue microarrays comprising duplicate 0.6-mm cores from 18 cases in each group were constructed using Alphelys TMA Designer R 2 version 1.0.0.8. Sections (3 μm) were cut from completed array blocks and transferred to silanized glass slides. Sections from these arrays were then stained for CD56, 1:200 (NCL-CD56-1B6; Novocastra); 11βHSD1, 1:300 (AB83522; Abcam); MR, 1:400 (H-300: SC-11412; Santa Cruz Biotechnology); and GR, 1:200 (E-20: SC-1003; Santa Cruz Biotechnology). Semiquantative analysis was performed using a Panoramic viewer to capture images. High-power (×400) images were analyzed with Image J (http://rsbweb.nih.gov/ij/) using the color deconvolution plugin and thresholding to assess the percentage of strongly immunopositive endometrial cells. The observers were blind as to the origin of the samples.

was performed using a Panoramic viewer to capture images. High-power (×400) images were analyzed with Image J (http://rsbweb.nih.gov/ij/) using the color deconvolution plugin and thresholding to assess the percentage of strongly immunopositive endometrial cells. The observers were blind as to the origin of the samples. Real-time quantitative PCR Total RNA was extracted with RNA STAT-60 from primary HESC cultures. After treatment with amplification-grade deoxyribonuclease I (Invitrogen Ltd), cDNA was generated using the SuperScript II first-strand synthesis system for RT-PCR kit (Invitrogen). Template quantification was performed with an ABI Step One system (Applied Biosystems) using Power SYBR Green PCR master mix (Applied Biosystems). RNA input variances were normalized against the levels of the L19 housekeeping gene, which encodes a ribosomal protein. All measurements were performed in duplicate. Specific primer pairs were designed using Primer3 software (http://frodo.wi.mit.edu): L19 sense, 5′-GCG GAA GGG TAC AGC CAA T-3′, L19-R antisense, 5′-GCA GCC GGC GCA AA-3′; 11βHSD1 sense, 5′-AGC AAG TTT GCT TTG GAT GG-3′, 11βHSD1 antisense, 5′-AGA GCT CCC CCT TTG ATG AT-3′; decidual prolactin (PRL) sense, 5′-AAG CTG TAG AGA TTG AGG AGC AAA C-3′, decidual PRL antisense, 5′-TCA GGA TGA ACC TGG CTG ACT A-3′; IGF-binding protein-1 (IGFBP1) sense, 5′-CGA AGG CTC TCC ATG TCA CCA-3′, IGFBP1 antisense, 5′-TGT CTC CTG TGC CTT GGC TAA AC-3′; IL-11 sense, 5′-CTC GAG TTT CCC CAG ACC CTC GG-3′, IL-11 antisense, 5′-TGT CAG CAC ACC TGG GAG CTG TAG-3′; IL-15 sense, 5′-TGG CTG CTG GAA ACC CCT TGC-3′, IL-15 antisense, 5′-CCC TGC ACT GAA ACA GCC CAA AA-3′; DHRS3 sense, 5′-AGC GCG GCG CCA GAA AGA TT-3′, DHRS3 antisense, 5′-TCA CCC ACC TTC TCC CGG ACG-3′; and RETSAT sense, 5′-CGC TGC CTG CCA GGT GTG AAG-3′, RETSAT antisense, 5′-AGA CGT AGC GCT CCA TCG CC-3′.

G CTG TAG-3′; IL-15 sense, 5′-TGG CTG CTG GAA ACC CCT TGC-3′, IL-15 antisense, 5′-CCC TGC ACT GAA ACA GCC CAA AA-3′; DHRS3 sense, 5′-AGC GCG GCG CCA GAA AGA TT-3′, DHRS3 antisense, 5′-TCA CCC ACC TTC TCC CGG ACG-3′; and RETSAT sense, 5′-CGC TGC CTG CCA GGT GTG AAG-3′, RETSAT antisense, 5′-AGA CGT AGC GCT CCA TCG CC-3′. Western blot analysis Whole-cell protein extracts were obtained by direct lysis in Laemmli buffer heated to 100°C. Proteins resolved by SDS-PAGE were transferred to a polyvinyl difluoride membrane (GE Healthcare) and probed with antibodies raised against 11βHSD1, 1:1000 (AB83522; Abcam); DHRS3, 1:1000 (15393-1-AP; ProteinTech Group); RETSAT, 1:1000 (SAB1407586; Sigma); and β-actin, 1:100 000 (A1978; Sigma). After incubation with horseradish peroxidase-conjugated secondary antibodies diluted 1:2000 (DAKO), immunoreactivity was visualized using the ECL+ chemoluminescent detection kit (Amersham). Statistical analysis Data were analyzed with the statistical package GraphPad Prism (GraphPad Software Inc). A Student's t test and a Mann-Whitney U test were used when appropriate. Logarithmic transformations were used when data were not normally distributed. Pearson's correlation coefficient (r) was used to assess the correlation between uNK cell densities in vivo and the induction of various genes upon decidualization of corresponding primary HESC cultures. Statistical significance was assumed when P < .05.

ransformations were used when data were not normally distributed. Pearson's correlation coefficient (r) was used to assess the correlation between uNK cell densities in vivo and the induction of various genes upon decidualization of corresponding primary HESC cultures. Statistical significance was assumed when P < .05. Results Midluteal uNK cell density correlates inversely with endometrial 11βHSD1 expression We routinely assess uNK cell densities by CD56 immunostaining of timed (d LH+7 to LH+10) endometrial biopsies from women suffering reproductive failure. Based on previous studies, normal uNK cell density is defined as 5% or less CD56+ cells in the stroma underlying the luminal epithelium (17, 23). We speculated that an apparent excess of uNK cells in the periimplantation endometrium may reflect impaired 11βHSD1 expression and relative cortisol deficiency. To test this hypothesis, a TMA was constructed using biopsies with normal (n = 18) as well as elevated (n = 18) uNK cell levels. The TMA was stained with anti-CD56 and anti-11βHSD1 antibodies. As shown in Figure 1A, a strong inverse correlation was observed between uNK cell density and 11βHSD1 immunoreactivity. This negative correlation was confirmed by semiquantitative image analysis (Figure 1B).

8) as well as elevated (n = 18) uNK cell levels. The TMA was stained with anti-CD56 and anti-11βHSD1 antibodies. As shown in Figure 1A, a strong inverse correlation was observed between uNK cell density and 11βHSD1 immunoreactivity. This negative correlation was confirmed by semiquantitative image analysis (Figure 1B). Figure 1. Inverse correlation between uNK cell density in the subluminal epithelium and expression of 11βHSD1. A, Representative CD56 and 11βHSD1 immunostaining in midluteal endometrial biopsies deemed to have normal or elevated uNK cell count (≤5% or >5% CD56+ cells in the stroma, respectively). Original magnification, ×20. B, Relative 11βHSD1 immunostaining after image analysis of a TMA containing biopsies with normal (≤5%; n = 18) or increased (>5%, n = 18) uNK cell density in the subluminal stroma. ***, P < .001.

to have normal or elevated uNK cell count (≤5% or >5% CD56+ cells in the stroma, respectively). Original magnification, ×20. B, Relative 11βHSD1 immunostaining after image analysis of a TMA containing biopsies with normal (≤5%; n = 18) or increased (>5%, n = 18) uNK cell density in the subluminal stroma. ***, P < .001. Elevated uNK cell density in vivo is associated with impaired induction of key decidual markers in vitro Previous studies have shown that an aberrant decidual response, associated with reproductive disorders such as endometriosis and RPL, is maintained in culture (24, 25). This prompted us to investigate whether uNK cell density in vivo could reflect the induction of 11βHSD1 in response to decidual cues. To this end, we divided timed endometrial biopsies and processed one part for CD56 immunostaining and the other part for primary HESCs. These cultures were passaged once, grown to confluency, and treated with 8-bromo-cAMP, P4, and E for either 4 or 8 days. As shown in Figure 2, there was a striking inverse correlation between uNK cell density in vivo and the responsiveness of paired primary cultures to differentiation stimuli (Supplemental Table 1). This inverse correlation was apparent for the induction of not only HSD11B1, the gene that encodes 11βHSD1 (Figure 2A) but also for PRL and IGFBP1, two classical decidual marker genes (Figure 2, B and C).

ensity in vivo and the responsiveness of paired primary cultures to differentiation stimuli (Supplemental Table 1). This inverse correlation was apparent for the induction of not only HSD11B1, the gene that encodes 11βHSD1 (Figure 2A) but also for PRL and IGFBP1, two classical decidual marker genes (Figure 2, B and C). Figure 2. Inverse correlation between uNK cell densities in vivo and the induction of decidual markers in vitro. A, The uNK cell densities in midluteal biopsies correlated inversely to the induction of 11βHSD1 transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). Note the logarithmic Y-axis. The right panel shows the mean (±SEM) induction of 11βHSD1 transcripts in biopsies deemed to have normal or elevated uNK cell counts. B, The uNK cell densities in midluteal biopsies correlated inversely to the induction of PRL transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of PRL transcripts in biopsies deemed to have normal or elevated uNK cell counts. Note the logarithmic Y-axis. C, The uNK cell densities in midluteal biopsies correlated inversely to the induction of IGFBP1 transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of PRL transcripts in biopsies deemed to have normal or elevated uNK cell counts. Note the logarithmic Y-axis. *, P < .05; **, P < .01.

related inversely to the induction of IGFBP1 transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of PRL transcripts in biopsies deemed to have normal or elevated uNK cell counts. Note the logarithmic Y-axis. *, P < .05; **, P < .01. We also examined whether this association extended to IL-11 and IL-15, cytokines implicated in regulating uNK cells (11–13). Rather surprisingly, there was a significant trend toward higher levels of induction IL-15, but not IL-11, transcripts in decidualizing HESC cultures obtained from biopsies with normal uNK cell densities (Figure 3, A and B). Figure 3. Induction of IL11 and IL15 mRNA in decidualizing HESCs in culture and uNK cell densities in vivo. A and B, The uNK cell densities in midluteal biopsies were correlated to the induction of IL11 and IL15 transcripts, respectively, in corresponding primary HESCs decidualized for either 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction in biopsies deemed to have normal or elevated uNK cell counts.

densities in midluteal biopsies were correlated to the induction of IL11 and IL15 transcripts, respectively, in corresponding primary HESCs decidualized for either 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction in biopsies deemed to have normal or elevated uNK cell counts. Elevated uNK cell density in vivo is associated with impaired expression of MR-dependent metabolic enzymes Next, we used the TMA to examine the expression of the cortisol-responsive receptors, GR and MR. As reported by others (26), GR and MR are both expressed in the endometrial stroma (Figure 4A). Semiquantitative analysis of the TMA showed no difference in stromal GR immunoreactivity in biopsies characterized by elevated vs normal uNK cell levels. In contrast, an increased percentage of CD56+ cell density was associated with significantly lower MR expression levels (P < .05; Figure 4B). Figure 4. Increased uNK cell levels are associated with impaired MR expression. A and B, The expression of GR and MR, respectively, in the subluminal stroma was assessed by immunostaining of a TMA containing biopsies with normal (≤5%; n = 18) or increased (>5%, n = 18) uNK cell density in the subluminal stroma. The left panel shows representative immunostaining (original magnification, ×40), whereas the right panel depicts semiquantitative image analysis of the immune staining. ***, P < .001.

nostaining of a TMA containing biopsies with normal (≤5%; n = 18) or increased (>5%, n = 18) uNK cell density in the subluminal stroma. The left panel shows representative immunostaining (original magnification, ×40), whereas the right panel depicts semiquantitative image analysis of the immune staining. ***, P < .001. Previous knockdown experiments in decidualizing HESCs have shown that the 11βHSD1/MR axis regulates the expression of several enzymes involved in lipid metabolism and retinoid acid biosynthesis and storage (18). We measured the expression levels of two target genes, DHRS3 and RETSAT, to monitor the MR activity in decidualizing primary HESC cultures established from biopsies with normal or elevated uNK cell levels. Both genes were moderately induced upon treatment with 8-bromo-cAMP/P4/E in a time-dependent manner when the primary HESC cultures were established from samples with 5% or less CD56+ cells in the subepithelial stroma (Figure 5). By contrast, this induction was significantly impaired in cultures established from high uNK cell samples.

moderately induced upon treatment with 8-bromo-cAMP/P4/E in a time-dependent manner when the primary HESC cultures were established from samples with 5% or less CD56+ cells in the subepithelial stroma (Figure 5). By contrast, this induction was significantly impaired in cultures established from high uNK cell samples. Figure 5. Elevated uNK cell density is associated with blunted expression of the MR-dependent genes in decidualizing HESCs. A, The uNK cell densities in 21 biopsies correlated inversely to the induction of DHRS3 transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of DHRS3 transcripts in biopsies deemed to have normal or elevated uNK cell counts. B, The uNK cell densities in 21 biopsies correlated inversely to the induction of RETSAT transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of RETSAT transcripts in biopsies deemed to have normal or elevated uNK cell counts. *, P < .05; **, P < .01.

nsities in 21 biopsies correlated inversely to the induction of RETSAT transcripts in primary HESCs decidualized for 4 days (left panel) or 8 days (middle panel). The right panel shows the mean (±SEM) induction of RETSAT transcripts in biopsies deemed to have normal or elevated uNK cell counts. *, P < .05; **, P < .01. To validate these findings, we performed Western blot analysis of primary HESCs decidualized with 8-bromo-cAMP/P4/E for 4 days. As shown in Figure 6A, 11βHSD1 and DHRS3 were abundantly expressed in primary cultures decidualized for 4 days. Induction of RETSAT, however, requires prolonged decidualization. This enzyme was barely detectable after 4 days of differentiation and only in primary cultures established from samples with 5% or less CD56+ cells in the subepithelial stroma. Semiquantitative analysis of the blots showed that elevated uNK cell density in vivo is associated with significantly lower 11βHSD1 expression in corresponding decidualizing primary HESC cultures and a trend toward lower DHRS3 levels (P = .03 and P = .07, respectively; Figure 6B). Figure 6. A, Composite figure showing 11βHSD1, DHRS3, and RETSAT protein expression in primary HESC cultures decidualized for 4 days. A total of eight primary cultures were established from biopsies with normal or elevated uNK cell densities. β-Actin served as a loading control. B, Semiquantitative analysis of 11βHSD1 and DHRS3 expression relative to β-actin. Because of the low level of expression, RETSAT expression was not quantified. *, P < .05.

s. A total of eight primary cultures were established from biopsies with normal or elevated uNK cell densities. β-Actin served as a loading control. B, Semiquantitative analysis of 11βHSD1 and DHRS3 expression relative to β-actin. Because of the low level of expression, RETSAT expression was not quantified. *, P < .05. Discussion Increased uNK cell density in midluteal endometrium has been associated with reproductive failure, especially RPL (4–6) However, the mechanisms that account for cyclic recruitment of uNK cell precursors and subsequent proliferation and differentiation within the periimplantation environment are not well understood. In addition to IL-11 and IL-15, several other endometrial factors may be implicated in this process, including chemokine motif ligand 14, IL-12, and IL-33 (25, 27, 28). Different strands of evidence suggest that induction of a cortisol gradient upon decidualization of HESCs is also a key regulator of uNK cells in periimplantation endometrium. First, preconceptual prednisolone treatment markedly reduces uNK cell density in RPL patients (17). uNK cells express GR but not progesterone receptors (15), the inference being that glucocorticoids are likely to act directly on these cells. Second, we found a strong negative correlation between uNK cell densities and expression of 11βHSD1 in differentiating stromal cells in vivo. We have shown previously that 11βHSD1 expression and enzyme activity in decidualizing HESCs is driven by cAMP and P4 signaling (18). Furthermore, inhibition of 11βHSD1 activity with either carbenoxolone disodium salt or PF 915275 virtually abolishes the induction of HSD11B1, indicating that local cortisol signaling reinforces the expression of this enzyme in decidualizing cells through an autocrine mechanism (18). By contrast, the expression of the type 2 isoform (11βHSD2), the dehydrogenase that converts cortisol into inactive cortisone, is low in both undifferentiated and decidualizing HESCs (18). Notably, the decidual process in the human endometrium is under tight spatiotemporal control (29). It is initiated in the midluteal phase of the cycle first in stromal cells surrounding the terminal spiral arteries and underlying the luminal epithelium.

one, is low in both undifferentiated and decidualizing HESCs (18). Notably, the decidual process in the human endometrium is under tight spatiotemporal control (29). It is initiated in the midluteal phase of the cycle first in stromal cells surrounding the terminal spiral arteries and underlying the luminal epithelium. Thus, rather than the total tissue concentration of uNK cells, it is possible that excessive migration of uNK cells from their usual position in the basal and perivascular regions of the endometrium to the subluminal region is the hallmark of an abnormal decidual response that predisposes for early pregnancy loss. Finally, there is increasing evidence that the responsiveness of endometrial cells to differentiation signals is subject to epigenetic programming (30), which explains how an aberrant decidual response in vivo is maintained, at least partly, upon differentiation of purified HESCs in vitro (24, 25, 31). In agreement, we found that high uNK cell density in vivo is associated with blunted induction of HSD11B1 as well as decidual marker genes, such as PRL and IGFBP1, in primary cultures. This strong inverse correlation suggests that the 5% threshold of uNK cell density is somewhat arbitrary. Whether increased uNK cells densities correlate to increased risk of pregnancy failure warrants further investigation.

ction of HSD11B1 as well as decidual marker genes, such as PRL and IGFBP1, in primary cultures. This strong inverse correlation suggests that the 5% threshold of uNK cell density is somewhat arbitrary. Whether increased uNK cells densities correlate to increased risk of pregnancy failure warrants further investigation. In addition to decidual marker genes, induction of IL-11 and IL-15 also tended to be lower in decidualizing cells established from biopsies with elevated uNK cell levels. This observation does not exclude that expression levels of these cytokines in situ correlate with uNK cell levels as reported for IL-15 in a recent study (32).

to decidual marker genes, induction of IL-11 and IL-15 also tended to be lower in decidualizing cells established from biopsies with elevated uNK cell levels. This observation does not exclude that expression levels of these cytokines in situ correlate with uNK cell levels as reported for IL-15 in a recent study (32). Human uNK cells have been described as immature and inactive before pregnancy (1). How elevated levels of uNK cells prior to conception predispose to subsequent pregnancy failure is unclear. Ablation of these cells in mice has been shown to compromise spiral arteriole remodeling and maintenance of decidual integrity seen after midpregnancy (3, 33). Yet uNK cell-deficient IL-15−/− mice are fertile and have normal gestation lengths and litter sizes comparable with wild-type mice (34). Similarly, human uNK cells are implicated in spiral artery remodeling. They are a rich source of angiogenic growth factors, although paradoxically the endometrium of RPL patients is characterized by reduced expression of several key factors, including platelet-derived growth factor-BB, angiotensin-2, vascular endothelial growth factor-A, and vascular endothelial growth factor-C (35). Our data suggest that increased density of CD56+ cells in the subluminal endometrial stromal compartment may be an indirect marker of local corticosteroid deficiency. Furthermore, our data indicate that the 11βHSD1/MR axis in target cells may be particularly affected as exemplified by the impaired induction of DHRS3 and RETSAT transcripts in decidualizing cultures established from high uNK cell biopsies. These enzymes are involved in lipid metabolism and retinoid acid (RA) biosynthesis and storage. Both shortage and excess of RA contribute to fetal malformation, suggesting that retinoid metabolism must be regulated closely at the fetomaternal interface (36). Alcohol dehydrogenase and nicotinamide adenine dinucleotide phosphate oxidase-dependent short-chain dehydrogenases/reductases, including DHRS3, oxidize retinal to retinol and promote its storage as retinyl esters. Likewise, RETSAT is involved in the regulation of retinoid storage as lipid droplets (37). Interestingly, RA inhibits decidualization of HESCs, and excess levels of RA or retinal are cytotoxic (38).

t-chain dehydrogenases/reductases, including DHRS3, oxidize retinal to retinol and promote its storage as retinyl esters. Likewise, RETSAT is involved in the regulation of retinoid storage as lipid droplets (37). Interestingly, RA inhibits decidualization of HESCs, and excess levels of RA or retinal are cytotoxic (38). Thus, high uNK cell density in the periimplantation endometrium may be associated with perturbations in the retinoid metabolism pathway in the stromal compartment, which in turn predisposes for an impaired decidual response and compromise histiotrophic support of the early conceptus.

t-chain dehydrogenases/reductases, including DHRS3, oxidize retinal to retinol and promote its storage as retinyl esters. Likewise, RETSAT is involved in the regulation of retinoid storage as lipid droplets (37). Interestingly, RA inhibits decidualization of HESCs, and excess levels of RA or retinal are cytotoxic (38). Thus, high uNK cell density in the periimplantation endometrium may be associated with perturbations in the retinoid metabolism pathway in the stromal compartment, which in turn predisposes for an impaired decidual response and compromise histiotrophic support of the early conceptus. Thus, it seems likely that complex and dynamic gradients of chemoattractants and chemorepellents control the spatiotemporal distribution of uNK cells in the periimplantation endometrium. In addition to glands and other immune cells, decidualizing stromal cells play a major role in governing this process, at least partially, by inducing a cortisol gradient that establishes a nutritive environment essential for postimplantation embryo development and fetal growth. Our data suggest that excessive uNK cells in the subluminal stromal compartment prior to conception may serve as a potential biomarker for a suboptimal decidual response in pregnancy. A recent pilot randomized, double-blind controlled clinical trial suggested an improvement in live birth rate with prednisolone in women with RPL and high midluteal uNK cell density (23), although this finding needs validating in a larger trial. In addition, assessment of uNK cell density varies greatly from laboratory to laboratory and intercycle variation has been reported (22). To be clinically useful, international standardization of uNK cell assessment is urgently needed.

eal uNK cell density (23), although this finding needs validating in a larger trial. In addition, assessment of uNK cell density varies greatly from laboratory to laboratory and intercycle variation has been reported (22). To be clinically useful, international standardization of uNK cell assessment is urgently needed. Abbreviations: DHRSdehydrogenase/reductase 8-bromo-cAMP8-bromoadenosine-cAMP Ecortisone GRglucocorticoid receptor HESChuman endometrial stromal cell 11βHSD111β-hydroxysteroid dehydrogenase IGFBP1IGF-binding protein-1 MRmineralocorticoid receptor P4progesterone PRLprolactin RAretinoic acid RETSATretinol saturase RPLrecurrent pregnancy loss TMAtissue microarray uNKuterine natural killer. Acknowledgments We are grateful to all the women who participated in this study and for the support from Biomedical Research Unit in Reproductive Health, a joint initiative between the University Hospitals Coventry and Warwickshire National Health Service Trust and Warwick Medical School. K.K. was supported by the Uehara Memorial Foundation Research Fellowship and a Naito Foundation subsidy for interinstitutional researchers. R.V., S.Q., and J.J.B. were supported by the Biomedical Research Unit in Reproductive Health, a joint initiative between the University Hospitals Coventry and Warwickshire National Health Service Trust and Warwick Medical School. Disclosure Summary: The authors have nothing to disclose.

The 5α-reductases (5αRs) convert testosterone to its more potent metabolite 5α-dihydrotestosterone (DHT). Investigation of rare cases of 5αR deficiency, presenting with a 46XY disorder of sexual development, led to the discovery of 2 isozymes (1): 5αR type 1 (5αR1) is expressed in metabolic tissues including liver (2), adipose (3) and skeletal muscle (4), and 5αR type 2 (5αR2) is expressed predominantly in the reproductive tract, where deficiency accounts for disordered sexual development, and in human liver (2). 5αR inhibitors, which reduce circulating and prostatic DHT levels, are prescribed commonly in patients with benign prostatic hyperplasia (BPH). Finasteride inhibits 5αR2 selectively, whereas dutasteride inhibits both 5αR1 and 5αR2 (5, 6). In addition to testosterone, 5αRs also catalyze reduction of a range of steroid hormones, including glucocorticoids (2). Due to widespread enzyme expression, and lack of substrate specificity, 5αR inhibition may alter local steroid concentrations in extraprostatic tissues. Targeting of another enzyme, 11β-hydroxysteroid dehydrogenase type 1, which metabolizes glucocorticoids in liver and adipose tissue, alters local but not systemic glucocorticoid levels and affects body fat distribution and insulin sensitivity (7, 8). Increased liver fat and decreased insulin sensitivity are seen in mice with targeted disruption of 5αR1, but not 5αR2 (9).

rogenase type 1, which metabolizes glucocorticoids in liver and adipose tissue, alters local but not systemic glucocorticoid levels and affects body fat distribution and insulin sensitivity (7, 8). Increased liver fat and decreased insulin sensitivity are seen in mice with targeted disruption of 5αR1, but not 5αR2 (9). We hypothesized that inhibition of 5αR1 decreases insulin sensitivity in humans, as it does in rodents. Previous studies of the metabolic effects of 5αR inhibitors in humans have been limited to simple but insensitive measures such as fasting plasma glucose (10). To determine the influence of 5αR1, we compared the dual 5αR1 and 5αR2 inhibitor dutasteride with the 5αR2 selective inhibitor finasteride. Subjects and Methods Study design This was a double-blind, randomized controlled study. Approval from the Lothian Research Ethics Committee and informed written consent were obtained. Participants were studied before and after 3 months of dutasteride (0.5 mg daily; Glaxo Smith Kline Pharmaceuticals), finasteride (5 mg daily; Gedeon Richter), or tamsulosin modified release (MR) (0.4 mg daily; Synthon Hispania) as a control group with doses as used in treatment of BPH (Figure 1). Fixed-size block randomization (n = 18 per block), without stratification or minimization, was performed by Tayside Pharmaceuticals. Figure 1. Summary of study protocol.

Subjects and Methods Study design This was a double-blind, randomized controlled study. Approval from the Lothian Research Ethics Committee and informed written consent were obtained. Participants were studied before and after 3 months of dutasteride (0.5 mg daily; Glaxo Smith Kline Pharmaceuticals), finasteride (5 mg daily; Gedeon Richter), or tamsulosin modified release (MR) (0.4 mg daily; Synthon Hispania) as a control group with doses as used in treatment of BPH (Figure 1). Fixed-size block randomization (n = 18 per block), without stratification or minimization, was performed by Tayside Pharmaceuticals. Figure 1. Summary of study protocol. Participants Participants (age 20–85 years) were recruited from secondary-care urology clinics, primary-care practices, and by advertising. Initial inclusion criteria were men with BPH aged 50 to 80 years, later expanded to all men ≥20 years old to improve recruitment. Exclusion criteria were 5αR inhibitor or glucocorticoid use in previous 3 months; diabetes mellitus or impaired glucose tolerance; significant hepatic, renal, or thyroid disease; hypogonadism; warfarin therapy; body mass index (BMI) ≥40 kg/m2; or any suspicion of urological malignancy.

years old to improve recruitment. Exclusion criteria were 5αR inhibitor or glucocorticoid use in previous 3 months; diabetes mellitus or impaired glucose tolerance; significant hepatic, renal, or thyroid disease; hypogonadism; warfarin therapy; body mass index (BMI) ≥40 kg/m2; or any suspicion of urological malignancy. Outcomes The primary outcome was insulin sensitivity assessed as glucose disposal during a hyperinsulinemic-euglycemic clamp (11). Secondary endpoints included fasting glucose/insulin relationships, effects of insulin on glucose production and lipolysis, body fat distribution, and gene transcript abundance in sc adipose tissue biopsies. Steroids were measured in blood, urine, and saliva to aid with mechanistic interpretation. Clinical methods Participants collected a 24-hour urine sample and 5 saliva samples (waking, 30 minutes after waking, noon, 4:00 pm, and bedtime) using Salivette collection tubes (Sarstedt) and then attended the Clinical Research Facility at 7:30 am after an overnight fast. Height, weight, blood pressure (BP), pulse rate, and hip and waist circumference were measured using standard techniques. Body fat was measured by bioimpedance using an OMRON BF306 body fat monitor (OMRON Healthcare Ltd). Blood was taken for measurements including glucose, insulin, C-peptide, sex steroids, cortisol, corticosteroid binding globulin (CBG), and adipokines. Biopsies of peri-umbilical sc abdominal adipose were taken with a 14-gauge needle under local anesthesia, with samples snap-frozen on dry ice.

or (OMRON Healthcare Ltd). Blood was taken for measurements including glucose, insulin, C-peptide, sex steroids, cortisol, corticosteroid binding globulin (CBG), and adipokines. Biopsies of peri-umbilical sc abdominal adipose were taken with a 14-gauge needle under local anesthesia, with samples snap-frozen on dry ice. A 3-phase, 2-step hyperinsulinemic-euglycemic clamp was conducted with infusion rates of tracers adjusted for body weight and those for insulin by body surface area as indicated below. From 0 to 90 minutes, only stable isotope tracers (Cambridge Isotope Laboratories, Inc) were infused: 6,6-[2H]2-glucose (d2-glucose; 17 μmol/kg for 1 minute, then 0.22 μmol/kg/min) and 1,1,2,3,3-[2H]5-glycerol (d5-glycerol; 1.6 μmol/kg for 1 minute, then 0.11 μmol/kg/min). Tracer infusions were continued, and from 90 to 180 minutes, low-dose insulin was infused (Actrapid; Novo Nordisk; 10 mU/m2/min) to measure inhibition of lipolysis and endogenous glucose production. From 180 to 270 minutes, high-dose insulin was infused (40 mU/m2/min) with tracers to assess peripheral glucose uptake. During insulin infusion, 20% dextrose (Baxter) infusion was adjusted to maintain euglycemia (4.5 mM–5.5 mM), measured from arterialized samples by glucometer (Accu-Check Advantage; Roche). One sample was taken at baseline, and 4 steady-state samples were taken over 20 minutes at the end of each phase from a hand vein arterialized by external heating of a retrograde cannula (12).

djusted to maintain euglycemia (4.5 mM–5.5 mM), measured from arterialized samples by glucometer (Accu-Check Advantage; Roche). One sample was taken at baseline, and 4 steady-state samples were taken over 20 minutes at the end of each phase from a hand vein arterialized by external heating of a retrograde cannula (12). Participants with BPH were unblinded individually on completion of their participation to allow decisions regarding ongoing care. Healthy participants were unblinded at either the interim or final analysis. Adherence was deemed adequate when drug was detected in serum.

djusted to maintain euglycemia (4.5 mM–5.5 mM), measured from arterialized samples by glucometer (Accu-Check Advantage; Roche). One sample was taken at baseline, and 4 steady-state samples were taken over 20 minutes at the end of each phase from a hand vein arterialized by external heating of a retrograde cannula (12). Participants with BPH were unblinded individually on completion of their participation to allow decisions regarding ongoing care. Healthy participants were unblinded at either the interim or final analysis. Adherence was deemed adequate when drug was detected in serum. Magnetic resonance imaging and proton magnetic resonance spectroscopy Magnetic resonance imaging (MRI) and proton magnetic resonance spectroscopy (MRS) measurements of adipose distribution and liver fat, respectively, were undertaken only at the end of the study and not at baseline. Participants without contraindications underwent MRI on a GE Signa Horizon 1.5-T HDxt scanner (General Electric) equipped with a self-shielding gradient set (33 mT m-1 maximum gradient strength) and a manufacturer-supplied torso array coil. Intra-abdominal visceral and sc fat volumes (10-mm slice at L4/5, using iterative decomposition of water and fat with echo asymmetry and least-squares estimation sequence) were quantified using SliceOmatic version 4.3 (TomoVision) software, assuming adipose density of 0.92 g/mL. Single-voxel proton MRS was performed for assessment of hepatic fat, using a point-resolved spectroscopy sequence, with and without water suppression. The voxel (10 mm3) was positioned within the liver, avoiding the edge of the liver and major vessels. Spectra were acquired during free breathing, with an echo time of 40 milliseconds and relaxation time of 5000 milliseconds. Postprocessing and quantification of MRS data was performed in jMRUI (13) using a nonlinear least-squares algorithm (Advanced Method for Accurate, Robust and Efficient Spectral fitting, AMARES) (14) with Gaussian line shapes to model each spectral peak of interest (eg, water at 4.7 ppm, methylene fat at 1.3 ppm).

onds. Postprocessing and quantification of MRS data was performed in jMRUI (13) using a nonlinear least-squares algorithm (Advanced Method for Accurate, Robust and Efficient Spectral fitting, AMARES) (14) with Gaussian line shapes to model each spectral peak of interest (eg, water at 4.7 ppm, methylene fat at 1.3 ppm). Laboratory methods A full blood count was measured on an XE-5000 automated flow cytometer (Sysmex UK); hemoglobin A1c by reverse-phase HPLC (HA8160 analyzer; Menarini); glucose, C-peptide, renal, liver, and thyroid function tests, and lipids by autoanalyzer (Architect c16000 analyzer; Abbott Diagnostics Ltd); serum SHBG by automated chemiluminescent assay (Immulite 2000 system; Siemens); plasma insulin by ultrasensitive ELISA (DRG); salivary cortisol by high-sensitivity ELISA (Salimetrics); plasma cortisol by 125I RIA (MP Biomedicals); plasma CBG by (125I RIA; DIAsource ImmunoAssays SA); plasma nonesterified fatty acids (NEFAs) by a coupled enzyme reaction assay (Zen-Bio, Inc); plasma leptin, monocyte chemoattractant protein 1, IL-8, adiponectin, and resistin by Milliplex immunoassay (Merck Millipore); and plasma estradiol by chemiluminescent microparticle immunoassay (Abbott Diagnostics) using an Architect c16000 analyzer. Tamsulosin was quantified from serum by liquid chromatography tandem mass spectrometry (LC-MS/MS) (15). Urinary steroids were extracted (16) and analyzed (17) as described previously, with the inclusion of the following transitions (collision energy) for androgens (androsterone, etiocholanolone m/z 360→270, 5α-androstane-3α,17α-diol (internal standard) m/z 331→241 [15 V]). mRNA abundance in sc adipose tissue was determined by real-time quantitative PCR (18), as detailed in Supplemental Table 1, and presented as abundance of gene of interest normalized to the mean of a panel of reference genes (PPIA, TBP, and GAPDH), the abundance of which did not differ between groups.

z 331→241 [15 V]). mRNA abundance in sc adipose tissue was determined by real-time quantitative PCR (18), as detailed in Supplemental Table 1, and presented as abundance of gene of interest normalized to the mean of a panel of reference genes (PPIA, TBP, and GAPDH), the abundance of which did not differ between groups. Expression of 5αR1 and -2 mRNA in human metabolic tissues Expression of 5αR1 and 2 mRNA was assessed in human tissues (sc adipose and liver collected with local ethical approval) and in commercially available skeletal muscle cDNA (Primer Design). Total mRNA was extracted using the QIAGEN RNeasy system, and 500 ng was reverse transcribed using the Applied Biosystems high-capacity reverse transcription kit with random primers. cDNA (10 ng) was subjected to PCR with primers specific for 5αR1 or 5αR2 (Supplemental Table 1) using the QIAGEN HotStarTaq Plus system, and products were separated by electrophoresis on a 1.2% agarose gel in 0.5× TBE buffer (Tris base, boric acid, EDTA). Supplemental laboratory methods Serum testosterone, DHT, finasteride, and dutasteride were quantified by LC-MS/MS (Supplemental Table 2), and plasma (during the euglycemic clamp) glucose, d2-glucose, glycerol, and d5-glycerol were quantified by gas chromatography/mass spectrometry.

Expression of 5αR1 and -2 mRNA in human metabolic tissues Expression of 5αR1 and 2 mRNA was assessed in human tissues (sc adipose and liver collected with local ethical approval) and in commercially available skeletal muscle cDNA (Primer Design). Total mRNA was extracted using the QIAGEN RNeasy system, and 500 ng was reverse transcribed using the Applied Biosystems high-capacity reverse transcription kit with random primers. cDNA (10 ng) was subjected to PCR with primers specific for 5αR1 or 5αR2 (Supplemental Table 1) using the QIAGEN HotStarTaq Plus system, and products were separated by electrophoresis on a 1.2% agarose gel in 0.5× TBE buffer (Tris base, boric acid, EDTA). Supplemental laboratory methods Serum testosterone, DHT, finasteride, and dutasteride were quantified by LC-MS/MS (Supplemental Table 2), and plasma (during the euglycemic clamp) glucose, d2-glucose, glycerol, and d5-glycerol were quantified by gas chromatography/mass spectrometry. Tracer kinetic calculations Tracer kinetics during the hyperinsulinemic-euglycemic clamp were calculated from average values in steady state: M value = glucose infusion rate at steady state; rate of disposal (Rd) of glucose = d2-glucose infusion rate/tracer-to-tracee ratio; endogenous glucose production (EGP) = Rd glucose − glucose infusion rate; and rate of appearance (Ra) of glycerol = d5-glycerol infusion rate/tracer-to-tracee ratio. Corrections were applied to adjust the peaks areas of d2-glucose for naturally occurring mass+2 glucose. Infusion rates were calculated specifically for mass+0 glucose and also d2-glucose.

Tracer kinetic calculations Tracer kinetics during the hyperinsulinemic-euglycemic clamp were calculated from average values in steady state: M value = glucose infusion rate at steady state; rate of disposal (Rd) of glucose = d2-glucose infusion rate/tracer-to-tracee ratio; endogenous glucose production (EGP) = Rd glucose − glucose infusion rate; and rate of appearance (Ra) of glycerol = d5-glycerol infusion rate/tracer-to-tracee ratio. Corrections were applied to adjust the peaks areas of d2-glucose for naturally occurring mass+2 glucose. Infusion rates were calculated specifically for mass+0 glucose and also d2-glucose. Sample size and statistical analysis A power calculation using previously published data (19) predicted 90% power to detect a 15% difference in glucose disposal rates to P < .05 with a sample size of 26 per group. A target group size of 33 allowed for a >20% dropout rate. A single planned interim analysis was conducted when at least half the planned participants had completed the study (n = 38). M values (mean steady-state glucose infusion rate) during hyperinsulinemia were analyzed, with P < .016 (P < .05/3) deemed sufficient for stopping the study. Interim data demonstrated a decrease in insulin sensitivity with dutasteride compared with finasteride (P = .002) and tamsulosin (P = .003). Therefore, recruitment was stopped and measurements for current participants completed in a final analysis. Analyses were specified a priori; therefore, no statistical adjustment was made for repeated analysis of M values. Results are presented from the final analysis.

finasteride (P = .002) and tamsulosin (P = .003). Therefore, recruitment was stopped and measurements for current participants completed in a final analysis. Analyses were specified a priori; therefore, no statistical adjustment was made for repeated analysis of M values. Results are presented from the final analysis. Statistical analysis was performed using SPSS for Windows, version 19 (IBM). Areas under the curve were calculated with Kinetica version 5.0 (Thermo Fisher Scientific). Data are presented as mean (SEM) unless stated otherwise. Analysis of covariance was not suitable because the primary and many secondary endpoints did not meet necessary statistical assumptions. ANOVA was therefore conducted on absolute change in each variable from baseline, with least significant difference (LSD) post hoc testing if ANOVA was significant (P < .05). If nonnormally distributed data could not be normalized by transformation, then Kruskal-Wallis testing was used. MRI scans were after treatment only, with absolute data rather than change from baseline analyzed by ANOVA as above. Values below the detection limit were considered to be one-third of the limit of detection for each assay. Missing values are indicated and were not imputed. Correlations with age were tested by Pearson correlation.

re after treatment only, with absolute data rather than change from baseline analyzed by ANOVA as above. Values below the detection limit were considered to be one-third of the limit of detection for each assay. Missing values are indicated and were not imputed. Correlations with age were tested by Pearson correlation. Results Participant recruitment, characteristics, and withdrawals Recruitment is summarized in Supplemental Figure 1. Fifty-one men consented, 47 completed the study, and 46 deemed adherent were included in the final analysis. Reasons for withdrawal were subclinical hypothyroidism (n = 1), side effects from study medication (urinary retention and impotence, n = 1), and unrelated illness before commencing study medication (n = 2). One BPH patient developed intolerable urinary symptoms upon cessation of his usual tamsulosin; he was able to complete the study with the addition of rescue tamsulosin to his study medication. Study medications were detected in serum for all but 1 participant (from the dutasteride group) who was deemed nonadherent and excluded from the final analysis. Serum concentrations in others were from 3.0 to 28.5 ng/mL (dutasteride), 2.0 to 64.0 ng/mL (finasteride), and 1.7 to 15.2 ng/mL (tamsulosin). Characteristics of participants at baseline are summarized in Table 1. Eleven participants were BPH patients (7 were being treated with α-blockers when recruited). Twelve participants were receiving concomitant regular medications, including simvastatin, aspirin, bendroflumethiazide, losartan, lansoprazole, and levothyroxine.

eristics of participants at baseline are summarized in Table 1. Eleven participants were BPH patients (7 were being treated with α-blockers when recruited). Twelve participants were receiving concomitant regular medications, including simvastatin, aspirin, bendroflumethiazide, losartan, lansoprazole, and levothyroxine. Table 1. Characteristics of Study Participants at Baselinea Dutasteride Finasteride Tamsulosin n 16 16 14 Age, y 35.3 (14.6) 40.3 (19.2) 49.4 (18.4) Range 20–64 21–85 21–73 BPH patients, n 2 4 5 BMI, kg/m2 25.3 (4.4) 26.8 (3.8) 25.5 (2.8) WHR 0.90 (0.08) 0.89 (0.07) 0.93 (0.06) Systolic BP, mm Hg 131 (11) 136 (15) 139 (18) Diastolic BP, mm Hg 78 (10) 78 (11) 81 (9) Body fat, % 19.8 (8.5) (n = 15) 22.1 (6.6) 24.7 (6.0) Fasting plasma/serum Glucose, mM 5.0 (0.5) 5.0 (0.5) 5.1 (0.4) Insulin, pM 59 (24) 54 (18) 63 (33) C-peptide, pM 539 (157) (n = 15) 539 (173) 613 (296) HOMA-IR 1.89 (0.79) 1.73 (0.63) 2.19 (1.14) Total cholesterol, mM 4.4 (0.6) 4.9 (1.0) 5.2 (1.0) Triglycerides, mM 1.2 (0.6) 1.2 (0.7) 1.4 (0.7) a Data are mean (SD). Effects of 5αR inhibition on insulin sensitivity As shown in Tables 2 and 3, insulin infusion had the predicted effects to suppress EGP and lipolysis (glycerol turnover and NEFA levels) and to stimulate glucose uptake. Table 2. Effects of Drug Interventions on Indices of Insulin Sensitivity for Glucose Metabolisma

Dutasteride Finasteride Tamsulosin n 16 16 14 Age, y 35.3 (14.6) 40.3 (19.2) 49.4 (18.4) Range 20–64 21–85 21–73 BPH patients, n 2 4 5 BMI, kg/m2 25.3 (4.4) 26.8 (3.8) 25.5 (2.8) WHR 0.90 (0.08) 0.89 (0.07) 0.93 (0.06) Systolic BP, mm Hg 131 (11) 136 (15) 139 (18) Diastolic BP, mm Hg 78 (10) 78 (11) 81 (9) Body fat, % 19.8 (8.5) (n = 15) 22.1 (6.6) 24.7 (6.0) Fasting plasma/serum Glucose, mM 5.0 (0.5) 5.0 (0.5) 5.1 (0.4) Insulin, pM 59 (24) 54 (18) 63 (33) C-peptide, pM 539 (157) (n = 15) 539 (173) 613 (296) HOMA-IR 1.89 (0.79) 1.73 (0.63) 2.19 (1.14) Total cholesterol, mM 4.4 (0.6) 4.9 (1.0) 5.2 (1.0) Triglycerides, mM 1.2 (0.6) 1.2 (0.7) 1.4 (0.7) a Data are mean (SD). Effects of 5αR inhibition on insulin sensitivity As shown in Tables 2 and 3, insulin infusion had the predicted effects to suppress EGP and lipolysis (glycerol turnover and NEFA levels) and to stimulate glucose uptake. Table 2. Effects of Drug Interventions on Indices of Insulin Sensitivity for Glucose Metabolisma Dutasteride (n = 16) Finasteride (n = 16) Tamsulosin (n = 14) P, ANOVA Before After Change Before After Change Before After Change Fasting before infusion Glucose, mM 5.0 (0.1) 5.1 (0.1) 0.1 (0.1) 5.0 (0.1) 4.9 (0.1) −0.1 (0.1) 5.1 (0.1) 5.1 (0.1) −0.1 (0.1) .34 Insulin, pM 59 (6) 69 (8) 10 (4) 54 (4) 58 (5) 4 (3) 63 (9) 59 (9) −4 (5) .07 C-peptide, pM 539 (41) 615 (44) 76 (26)b,d 539 (43) 526 (41) −13 (29) 613 (79) 588 (72) −24 (34) .04 HOMA-IR 1.89 (0.20) 2.28 (0.27) 0.39 (0.15)c 1.73 (0.16) 1.83 (0.16) 0.10 (0.10) 2.09 (0.31) 1.95 (0.32) −0.14 (0.16) .03 During tracer infusion without insulin infusion Glucose, mM 5.4 (0.1) 5.3 (0.1) 0.0 (0.1) 5.4 (0.1) 5.3 (0.1) −0.1 (0.1) 5.6 (0.1) 5.5 (0.1) −0.1 (0.1) .89 Insulin, pM 32 (3) 37 (5) 6 (3)c 36 (3) 35 (3) −1 (2) 37 (5) 31 (3) −6 (3) .03 EGP, μmol/kg FFM/min 9.03 (0.51) 9.10 (0.55) 0.07 (0.20) 10.23 (0.44) 10.02 (0.46) −0.21 (0.28) 9.83 (0.58) 10.21 (0.47) 0.38 (0.23) .24 During low-dose insulin infusion Glucose, mM 5.2 (0.1) 5.1 (0.0) −0.1 (0.1) 5.1 (0.1) 5.0 (0.1) 0.0 (0.1) 5.2 (0.1) 5.1 (0.1) −0.2 (0.1) .96 Insulin, pM 86 (9) 91 (10) 5 (11) 84 (6) 83 (6) −1 (7) 82 (7) 82 (6) −1 (5) .84 M value, μmol/kg FFM/min 7.84 (1.72) 8.04 (1.57) −0.02 (2.32) 9.06 (2.29) 9.13 (1.70) 0.08 (2.01) 6.72 (2.20) 7.12 (1.80) 0.40 (1.16) .99 EGP, μmol/kg FFM/min 5.10 (0.99) 5.54 (0.77) 0.44 (0.71) 5.80 (0.77) 5.54 (0.86) −0.25 (0.65) 6.67 (0.64) 6.56 (0.79) −0.11 (0.52) .72 During high-dose insulin infusion Glucose, mM 4.9 (0.1) 4.9 (0.0) −0.1 (0.1) 4.9 (0.1) 4.9 (0.1) 0.0 (0.1) 5.0 (0.1) 4.8 (0.1) −0.2 (0.1) .53 Insulin, pM 307 (18) 326 (14) 20 (16) 295 (13) 303 (19) 8 (16) 259 (15) 292 (15) 33 (12) .51 M value, μmol/kg FFM/min 45.2 (4.0) 39.0 (4.8) −6.2 (3.4)c,e 40.0 (4.1) 47.8 (5.1) 7.8 (3.2) 30.7 (4.2) 38.3 (4.7) 7.6 (2.2) .002 Rd glucose, μmol/kg FFM/min 41.9 (3.54) 36.1 (4.41) −5.7 (3.2)c,e 37.0 (3.77) 44.2 (4.74) 7.2 (3.0) 28.4 (3.88) 35.4 (4.38) 7.0 (2.0) .002 Abbreviation: FFM, fat-free mass.

(15) 33 (12) .51 M value, μmol/kg FFM/min 45.2 (4.0) 39.0 (4.8) −6.2 (3.4)c,e 40.0 (4.1) 47.8 (5.1) 7.8 (3.2) 30.7 (4.2) 38.3 (4.7) 7.6 (2.2) .002 Rd glucose, μmol/kg FFM/min 41.9 (3.54) 36.1 (4.41) −5.7 (3.2)c,e 37.0 (3.77) 44.2 (4.74) 7.2 (3.0) 28.4 (3.88) 35.4 (4.38) 7.0 (2.0) .002 Abbreviation: FFM, fat-free mass. a Data are mean (SEM) of the values from each study day obtained at baseline after overnight fast and as an average of 4 measurements in 15-minute steady-state periods after low-dose or high-dose insulin infusions (n = 11–12 per group). Steady state was confirmed, with the relative SD of the tracer-to-tracee ratios of d2-glucose to glucose between 0.4% and 3.8%. M value is the glucose infusion rate at steady state. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P < .05 vs tamsulosin. c P ≤ .01 vs tamsulosin. d P < .05 vs finasteride. e P ≤ .01 vs finasteride. Table 3. Effects of Drug Interventions on Lipid Profile and Insulin Sensitivity for Lipolysisa

a Data are mean (SEM) of the values from each study day obtained at baseline after overnight fast and as an average of 4 measurements in 15-minute steady-state periods after low-dose or high-dose insulin infusions (n = 11–12 per group). Steady state was confirmed, with the relative SD of the tracer-to-tracee ratios of d2-glucose to glucose between 0.4% and 3.8%. M value is the glucose infusion rate at steady state. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P < .05 vs tamsulosin. c P ≤ .01 vs tamsulosin. d P < .05 vs finasteride. e P ≤ .01 vs finasteride. Table 3. Effects of Drug Interventions on Lipid Profile and Insulin Sensitivity for Lipolysisa Dutasteride (n = 16) Finasteride (n = 16) Tamsulosin (n = 14) P, ANOVA Before After Change Before After Change Before After Change Total cholesterol, mM 4.4 (0.2) 4.3 (0.2) −0.1 (0.1) 4.9 (0.3) 5.0 (0.3) +0.1 (0.1) 5.2 (0.3) 4.9 (0.3) −0.3 (0.1) .08 HDL-cholesterol, mM 1.3 (0.1) 1.3 (0.1) −0.0 (0.0) 1.3 (0.1) 1.3 (0.1) 0.0 (0.0) 1.4 (0.2) 1.3 (0.1) −0.1 (0.1) .47 LDL-cholesterol, mM 2.6 (0.2) 2.6 (0.2) +0.0 (0.1) 3.0 (0.3) 3.1 (0.3) +0.1 (0.1) 3.2 (0.2) 3.1 (0.3) −0.1 (0.2) .43 Triglycerides, mM 1.3 (0.1) 1.1 (0.2) −0.2 (0.1) 1.2 (0.2) 1.2 (0.2) +0.0 (0.1) 1.4 (0.2) 1.2 (0.2) −0.1 (0.1) .38 Fasting before infusion Glycerol, μM 48.4 (5.3) 67.9 (18.4) 19.5 (14.5) 50.4 (7.2) 47.7 (7.0) −2.7 (7.3) 53.6 (9.5) 42.3 (5.4) −11.3 (5.8) .10 NEFAs, μM 485.1 (53.2) 523.8 (69.7) 38.7 (73.7) 443.9 (54.8) 493.1 (43.8) 49.2 (51.2) 649.4 (71.4) 580.2 (102.3) −69.2 (61.6) .37 During tracer infusion without insulin infusion Glycerol, μM 41.2 (5.0) 45.7 (6.3) +4.5 (5.2) 42.8 (4.7) 38.4 (4.7) −4.4 (4.2) 42.3 (6.3) 38.2 (4.3) −4.0 (3.7) .28 Ra glycerol, μmol/kg FFM/min 2.54 (0.34) 2.78 (0.42) 0.24 (0.21) 2.32 (0.28) 2.27 (0.16) −0.04 (0.25) 3.14 (0.39) 3.04 (0.39) −0.09 (0.30) .60 NEFAs, μM 484.2 (52.5) 520.1 (51.7) 35.9 (49.8) 479.6 (58.4) 497.5 (63.3) 17.9 (51.5) 628.9 (59.8) 577.2 (53.3) −51.7 (51.2) .46 During low-dose insulin infusion Glycerol, μM 18.0 (2.6) 22.4 (4.0) 4.4 (2.6) 17.4 (3.1) 16.0 (3.0) −1.5 (3.1) 22.6 (6.7) 16.0 (2.6) −6.7 (6.1) .17 Ra glycerol, μmol/kg FFM/min 1.23 (0.15) 1.39 (0.17) 0.16 (0.15) 1.32 (0.18) 1.30 (0.17) −0.02 (0.10) 1.99 (0.36) 1.60 (0.12) −0.39 (0.32) .16 NEFAs, μM 184.9 (27.9) 245.2 (36.6) 60.3 (30.0)b 193.4 (32.9) 189.7 (33.6) −3.7 (14.9) 295.1 (56.2) 214.1 (22.1) −81.0 (60.4) .04 During high-dose insulin infusion Glycerol, μM 14.7 (2.67) 13.1 (2.87) −1.8 (1.3) 10.6 (2.39) 8.9 (2.14) −1.7 (2.0) 15.0 (5.56) 6.9 (1.98) −8.1 (4.5) .21 NEFAs, μM 36.3 (4.0) 39.0 (5.9) 2.3 (5.9) 37.1 (4.5) 37.8 (4.9) 0.7 (3.7) 53.3 (9.8) 32.2 (2.8) −21.1 (9.8) .12 Abbreviations: FFM, fat-free mass; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

l, μM 14.7 (2.67) 13.1 (2.87) −1.8 (1.3) 10.6 (2.39) 8.9 (2.14) −1.7 (2.0) 15.0 (5.56) 6.9 (1.98) −8.1 (4.5) .21 NEFAs, μM 36.3 (4.0) 39.0 (5.9) 2.3 (5.9) 37.1 (4.5) 37.8 (4.9) 0.7 (3.7) 53.3 (9.8) 32.2 (2.8) −21.1 (9.8) .12 Abbreviations: FFM, fat-free mass; HDL, high-density lipoprotein; LDL, low-density lipoprotein. a Data are mean (SEM) of the values from each study day obtained at baseline after overnight fast and as an average of 4 measurements in 15-minute steady-state periods after low-dose or high-dose insulin infusions (n = 11–12 per group). Steady state was confirmed, with the relative SD of the tracer to tracee ratios of d5-glycerol to glycerol between 1.2% and 11.7%. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P ≤ .01 vs tamsulosin.

a Data are mean (SEM) of the values from each study day obtained at baseline after overnight fast and as an average of 4 measurements in 15-minute steady-state periods after low-dose or high-dose insulin infusions (n = 11–12 per group). Steady state was confirmed, with the relative SD of the tracer to tracee ratios of d5-glycerol to glycerol between 1.2% and 11.7%. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P ≤ .01 vs tamsulosin. Dutasteride, but not finasteride or tamsulosin, markedly decreased the glucose Rd (M value, the primary endpoint) during high-dose insulin infusion (Figure 2, A and B, and Supplemental Figure 2 and Table 2), increased fasting plasma C-peptide and homeostatic model assessment of insulin resistance (HOMA-IR) (Table 2), and increased plasma insulin levels when tracers were infused alone (Table 2). EGP during low-dose insulin was unaffected by study drugs (Table 2). Given the wide age range of participants, we tested whether age influenced the primary endpoint; the change in M value after drug treatment, measured during high-dose insulin infusion, did not correlate with age (dutasteride r = −0.28, P = .31; finasteride r = 0.17, P = .53; tamsulosin r = −0.13, P = .66).

s (Table 2). Given the wide age range of participants, we tested whether age influenced the primary endpoint; the change in M value after drug treatment, measured during high-dose insulin infusion, did not correlate with age (dutasteride r = −0.28, P = .31; finasteride r = 0.17, P = .53; tamsulosin r = −0.13, P = .66). Figure 2. Effects of 5αR inhibition on insulin sensitivity and body fat. A, Change in glucose (20% dextrose) infusion rate (milliliters per hour) required to maintain euglycemia during low-dose (10 mU/m2/min) and high-dose (40 mU/m2/min) insulin infusion. Data are mean (SEM). B, Change in glucose Rd during high-dose insulin infusion after dutasteride (black), finasteride (gray), or tamsulosin (white) treatment. Data are mean (SEM). C, Change in percent body fat measured by electrical bioimpedance after dutasteride (black), finasteride (gray), or tamsulosin (white) treatment. Data are mean (SEM). D, Transcripts of 5αR1 and -2 in human liver, skeletal muscle, and sc adipose tissue. Lanes 1, 5, and 9, 100-bp ladder; lanes 2, 6, and 10, 5αR1; lanes 3, 7, and 11, 5αR2; lanes 4, 8, and 12, negative control. Abbreviation: FFM, fat-free mass. Dutasteride, but not finasteride or tamsulosin, impaired suppression of plasma NEFA levels during low-dose insulin infusion only, although glycerol turnover was unaffected by drug treatment (Table 3).

Figure 2. Effects of 5αR inhibition on insulin sensitivity and body fat. A, Change in glucose (20% dextrose) infusion rate (milliliters per hour) required to maintain euglycemia during low-dose (10 mU/m2/min) and high-dose (40 mU/m2/min) insulin infusion. Data are mean (SEM). B, Change in glucose Rd during high-dose insulin infusion after dutasteride (black), finasteride (gray), or tamsulosin (white) treatment. Data are mean (SEM). C, Change in percent body fat measured by electrical bioimpedance after dutasteride (black), finasteride (gray), or tamsulosin (white) treatment. Data are mean (SEM). D, Transcripts of 5αR1 and -2 in human liver, skeletal muscle, and sc adipose tissue. Lanes 1, 5, and 9, 100-bp ladder; lanes 2, 6, and 10, 5αR1; lanes 3, 7, and 11, 5αR2; lanes 4, 8, and 12, negative control. Abbreviation: FFM, fat-free mass. Dutasteride, but not finasteride or tamsulosin, impaired suppression of plasma NEFA levels during low-dose insulin infusion only, although glycerol turnover was unaffected by drug treatment (Table 3). Effect of 5αR inhibition on body composition and adipose tissue There were no effects of drug treatment on BP, heart rate, body weight, BMI, or waist-to-hip ratio (WHR) (Table 4). There was, however, an increase in body fat (measured in kg or %) with dutasteride, but not finasteride, compared with tamsulosin (Figure 2C, Table 4). The increase in body fat with dutasteride was not accompanied by measurable differences in visceral or subcutaneous abdominal adipose volume on MRI (Table 4). Liver fat fraction (by MRS) was not measured at baseline and was compared only at the end of the study, when it was not different between treatment groups, either with (P = .22) or without adjustment for potential confounders (body weight, BMI, body fat, and WHR): median (interquartile ranges), dutasteride 9.4% (3.6, 23.6; n = 13), finasteride 4.7% (1.3, 43.2; n = 15), and tamsulosin 3.4% (1.8, 9.2; n = 9).

of the study, when it was not different between treatment groups, either with (P = .22) or without adjustment for potential confounders (body weight, BMI, body fat, and WHR): median (interquartile ranges), dutasteride 9.4% (3.6, 23.6; n = 13), finasteride 4.7% (1.3, 43.2; n = 15), and tamsulosin 3.4% (1.8, 9.2; n = 9). Table 4. Effects of Drug Interventions on Body Fat and BPa

of the study, when it was not different between treatment groups, either with (P = .22) or without adjustment for potential confounders (body weight, BMI, body fat, and WHR): median (interquartile ranges), dutasteride 9.4% (3.6, 23.6; n = 13), finasteride 4.7% (1.3, 43.2; n = 15), and tamsulosin 3.4% (1.8, 9.2; n = 9). Table 4. Effects of Drug Interventions on Body Fat and BPa Dutasteride (n = 16; body fat n = 15; MRI n = 13) Finasteride (n = 16) Tamsulosin (n = 14; MRI n = 11) P, ANOVA Before After Change Before After Change Before After Change Weight, kg 77.4 (3.2) 78.3 (3.0) +1.0 (0.6) 83.8 (3.5) 83.2 (3.4) −0.6 (0.5) 80.5 (2.8) 80.5 (3.0) 0.0 (0.7) .17 BMI, kg/m2 25.3 (1.1) 25.6 (1.0) +0.3 (0.2) 26.8 (1.0) 26.6 (0.9) −0.2 (0.2) 25.5 (0.7) 25.6 (0.9) +0.1 (0.2) .14 WHR 0.90 (0.02) 0.90 (0.02) 0.00 (0.01) 0.89 (0.02) 0.90 (0.01) +0.00 (0.01) 0.93 (0.02) 0.93 (0.01) 0.00 (0.01) .96 Systolic BP, mm Hg 131 (3) 135 (4) +4 (4) 136 (4) 140 (2) +4 (3) 139 (5) 138 (4) −1 (4) .56 Diastolic BP, mm Hg 78 (3) 82 (2) +4 (2) 78 (3) 80 (2) +2 (2) 81 (2) 82 (2) +1 (2) .60 Body fat, kg 16.5 (2.1) 17.8 (2.1) +1.2 (0.4)b 18.9 (1.7) 18.7 (1.6) −0.2 (0.5) 20.1 (1.6) 19.6 (1.8) −0.5 (0.6) .048 Body fat, % 19.8 (2.1) 21.5 (2.0) +1.6 (0.6)c 22.1 (1.7) 22.3 (1.7) +0.2 (0.5) 24.7 (1.6) 24.0 (1.9) −0.8 (0.6) .02 Visceral fat, kg 0.09 (0.01) 0.07 (0.01) 0.10 (0.02) .16 sc fat, kg 0.22 (0.03) 0.21 (0.02) 0.20 (0.02) .85 a Data are mean (SEM). Visceral and sc fat was measured in a cross-section at L4/5. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05).

eral fat, kg 0.09 (0.01) 0.07 (0.01) 0.10 (0.02) .16 sc fat, kg 0.22 (0.03) 0.21 (0.02) 0.20 (0.02) .85 a Data are mean (SEM). Visceral and sc fat was measured in a cross-section at L4/5. ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P < .05 vs tamsulosin. c P ≤ .01 vs tamsulosin. There were no differences in serum lipid profile (Table 3) and no drug-induced changes in serum adipokines (leptin, adiponectin, or resistin) or cytokines (monocyte chemoattractant protein 1 or IL-8) (Supplemental Table 3). In sc adipose, androgen receptor mRNA decreased from baseline in both dutasteride- and finasteride-treated groups compared with tamsulosin (Supplemental Table 4), but no other transcripts tested were altered. Effects of 5αR inhibitors on steroid profile Both dutasteride and finasteride, but not tamsulosin, decreased serum DHT and decreased urinary excretion of the A-ring-reduced metabolites of both androgens and glucocorticoids to a similar extent (Table 5). Steroid binding globulins, and cortisol in plasma (Table 5) and saliva (Supplemental Figure 3) did not differ between groups. There was a trend for 5αR inhibitors to increase estradiol levels in blood. Table 5. Effects of Drug Interventions on Steroids in Plasma and Urinea

Effects of 5αR inhibitors on steroid profile Both dutasteride and finasteride, but not tamsulosin, decreased serum DHT and decreased urinary excretion of the A-ring-reduced metabolites of both androgens and glucocorticoids to a similar extent (Table 5). Steroid binding globulins, and cortisol in plasma (Table 5) and saliva (Supplemental Figure 3) did not differ between groups. There was a trend for 5αR inhibitors to increase estradiol levels in blood. Table 5. Effects of Drug Interventions on Steroids in Plasma and Urinea Dutasteride (n = 16) Finasteride (n = 16) Tamsulosin (n = 14) P, ANOVA Before After Change Before After Change Before After Change Circulating steroids and binding proteins Testosterone, nM 25 (2) 30 (3) +5 (2) 21 (2) 24 (2) +3 (1) 21 (2) 23 (3) +2 (1) .22 DHT, nM 2.9 (0.4) 1.8 (0.4) −1.1 (0.2)b 2.1 (0.3) 1.0 (0.2) −1.1 (0.2)b 2.0 (0.3) 1.7 (0.3) −0.3 (0.2) .02 Cortisol, nM 788 (59) 692 (47) −96 (40) 769 (51) 689 (43) −80 (54) 818 (54) 757 (42) −61 (55) .88 Estradiol, pM 81.8 (7.9) 126.7 (16.9) +44.9 (15.8) 69.3 (7.4) 94.0 (8.9) +24.8 (6.9) 73.4 (9.4) 80.8 (10.9) +7.4 (7.5) .07 SHBG, nM 28 (2) 27 (3) −1 (1) 24 (2) 25 (2) +1 (1) 31 (3) 33 (4) +2 (2) .25 CBG, nM 988 (43) 966 (42) −22 (36) 936 (32) 905 (53) −31 (50) 959 (34) 929 (42) −31 (23) .75d Albumin, g/L 42 (1) 39 (1) −3 (1) 43 (1) 40 (1) −3 (1) 41 (1) 40 (1) −1 (1) .26 Urinary steroids Androsterone (α), μg/d 1806 (175) 121 (18) −1684 (163)b 2373 (434) 397 (78) −1975 (397)b 2116 (347) 2036 (336) −79 (268) <.001 Etiocholanolone (β), μg/d 817 (101) 2461 (338) +1643 (283)b 805 (151) 1710 (251) +905 (198)b 861 (130) 1029 (176) +168 (99) <.001d 5α-THF, μg/d 1664 (252) 32 (9) −1633 (249)b 1858 (356) 51 (13) −1807 (346)b 1786 (300) 1773 (308) −13 (200) <.001 5β-THF, μg/d 1724 (128) 1718 (130) −6 (127) 1670 (162) 1683 (127) +13 (159) 1793 (136) 1742 (165) −52 (199) .96 β-THF/α-THF 1.39 (0.20) 96.20 (16.68) +94.81 (16.75)b 1.21 (0.18) 55.69 (8.34) +54.48 (8.32)b 1.30 (0.18) 1.37 (0.24) +0.07 (0.12) <.001d F/α-THF 0.10 (0.03) 6.12 (0.92) +6.01 (0.92)b 0.10 (0.01) 4.22 (0.58) +4.11 (0.58)b 0.12 (0.03) 0.13 (0.03) +0.01 (0.02) <.001d F/β-THF 0.08 (0.01) 0.08 (0.01) 0.00 (0.01) 0.09 (0.01) 0.08 (0.01) −0.01 (0.00) 0.09 (0.01) 0.09 (0.01) 0.00 (0.01) .67 Etiocholanolone/androsterone 0.45 (0.05) 21.82 (2.43) +21.37 (2.43)b,c 0.36 (0.05) 8.26 (3.89) +7.89 (3.88)b 0.43 (0.04) 0.85 (0.39) +0.43 (0.38) <.001d Abbreviations: F, cortisol; THF, tetrahydrocortisol.

F 0.08 (0.01) 0.08 (0.01) 0.00 (0.01) 0.09 (0.01) 0.08 (0.01) −0.01 (0.00) 0.09 (0.01) 0.09 (0.01) 0.00 (0.01) .67 Etiocholanolone/androsterone 0.45 (0.05) 21.82 (2.43) +21.37 (2.43)b,c 0.36 (0.05) 8.26 (3.89) +7.89 (3.88)b 0.43 (0.04) 0.85 (0.39) +0.43 (0.38) <.001d Abbreviations: F, cortisol; THF, tetrahydrocortisol. a Data are mean (SEM). ANOVA was conducted on absolute change in each variable from baseline, with LSD post hoc testing if ANOVA was significant (P < .05). b P ≤ .01 vs tamsulosin. c P ≤ .01 vs finasteride. d Kruskal-Wallis test with pairwise comparisons. Expression of 5αR isozymes in human tissues Transcripts of both 5αR1 and 5αR2 were detected in human liver and skeletal muscle, but only 5αR1 mRNA was detected in sc adipose tissue (Figure 2D). Discussion These data highlight a previously unrecognized role of 5αR1 in modulating metabolic signaling in humans and detail the metabolic sequelae of 5αR inhibition in men. We demonstrate an increase in body fat and decrease in insulin sensitivity induced by the dual 5αR1/5αR2 inhibitor dutasteride, but not by the selective 5αR2 inhibitor finasteride, despite similar effects on circulating and urinary steroids. The metabolic effects of dutasteride are mediated in peripheral tissues, most likely including adipose tissue where 5αR1 but not 5αR2 is expressed. We therefore attribute these effects principally to inhibition of 5αR1 and consequent altered tissue steroid concentrations; this is supported by a recent publication demonstrating an adverse metabolic phenotype in 5αR1-deficient mice (9).

sues, most likely including adipose tissue where 5αR1 but not 5αR2 is expressed. We therefore attribute these effects principally to inhibition of 5αR1 and consequent altered tissue steroid concentrations; this is supported by a recent publication demonstrating an adverse metabolic phenotype in 5αR1-deficient mice (9). Although 5αR inhibitors have been used extensively clinically, previous studies of metabolism with 5αR inhibition (10, 20, 21) have neither been randomized nor adequately controlled; nor have they incorporated sensitive measures of insulin sensitivity. A crossover study is not feasible due to the long half-life of dutasteride (5 weeks) (22). We therefore designed a parallel-group randomized study to conduct detailed metabolic investigations and included a control group treated with tamsulosin, which is not known to have metabolic effects but allowed for inclusion of patients with symptomatic BPH.

ble due to the long half-life of dutasteride (5 weeks) (22). We therefore designed a parallel-group randomized study to conduct detailed metabolic investigations and included a control group treated with tamsulosin, which is not known to have metabolic effects but allowed for inclusion of patients with symptomatic BPH. The principal site of 5αR1 expression outside of the skin is the liver (2). Mice with life-long deficiency in 5αR1 exhibit liver fat accumulation after metabolic challenge (9), and we anticipated that effects of dutasteride on whole-body insulin sensitivity may be accompanied by liver fat accumulation and impaired suppression of EGP by insulin in the liver, with corresponding changes in serum lipid profile. However, our data in healthy men after dutasteride treatment for 3 months suggest preservation of hepatic insulin sensitivity after 5αR1 inhibition. Although in rodents 5αR1 is the predominant isozyme in liver, in humans, both 5αR1 and 5αR2 are expressed in liver (2), and their relative roles have not previously been described. We found that finasteride and dutasteride have similar effects on excretion of urinary 5α-reduced androgens and glucocorticoids, which reflect the intrahepatic steroid milieu as they are excreted as conjugates formed in the liver. The only difference we observed in steroid profiles between finasteride and dutasteride was a modestly higher etiocholanolone/androsterone ratio with dutasteride. This suggests that 5αR1 makes only a limited contribution, over and above that of 5αR2, to liver steroid metabolism in humans.

onjugates formed in the liver. The only difference we observed in steroid profiles between finasteride and dutasteride was a modestly higher etiocholanolone/androsterone ratio with dutasteride. This suggests that 5αR1 makes only a limited contribution, over and above that of 5αR2, to liver steroid metabolism in humans. Whereas hepatic insulin sensitivity was preserved, dutasteride strikingly decreased glucose disposal during high-dose insulin infusion, consistent with impaired insulin sensitivity in peripheral organs, including skeletal muscle and/or adipose tissue. This contrasted with an improvement in peripheral insulin sensitivity after 3 months treatment in the finasteride and tamsulosin groups, potentially explained by the Hawthorne effect of improved health during participation in clinical studies (23). We confirmed previous reports that 5αR1 is expressed in human skeletal muscle (4), but we did not assess skeletal muscle metabolism further here.

r 3 months treatment in the finasteride and tamsulosin groups, potentially explained by the Hawthorne effect of improved health during participation in clinical studies (23). We confirmed previous reports that 5αR1 is expressed in human skeletal muscle (4), but we did not assess skeletal muscle metabolism further here. Dutasteride increased body fat and reduced insulin-mediated suppression of NEFAs, consistent with impaired insulin sensitivity in adipose tissue. We could not attribute the increase in body fat to a specific change in sc, visceral, or hepatic adiposity, but this may reflect lack of statistical power for these secondary endpoints, particularly because MRI and proton MRS were not performed in every participant or at baseline. We did not demonstrate altered whole-body lipolysis by d5-glycerol turnover, but this may reflect biological or analytical variability. Alternatively, there may be an effect on fatty acid esterification, but this could not be demonstrated without the use of a palmitate tracer. We showed, using PCR, that 5αR1 but not 5αR2 is expressed in human adipose tissue. No alterations were found in intra-adipose transcript abundance or circulating adipokines that are likely to account for impaired insulin sensitivity; the observed reduction in androgen receptor mRNA might be metabolically adverse (24) but was observed with both finasteride and dutasteride so is most likely a response to altered circulating androgen levels. However, only sc adipose tissue was biopsied, whereas steroid signaling may exert greater effects in visceral adipose tissue. Taken together, these observations are consistent with metabolic effects of dutasteride being mediated by inhibition of 5αR1 in adipose tissue but do not exclude either a contribution from other tissues including skeletal muscle or a contribution from more potent inhibition of 5αR2 by dutasteride than finasteride.

sue. Taken together, these observations are consistent with metabolic effects of dutasteride being mediated by inhibition of 5αR1 in adipose tissue but do not exclude either a contribution from other tissues including skeletal muscle or a contribution from more potent inhibition of 5αR2 by dutasteride than finasteride. A third isozyme of 5αR has been described and is expressed in relevant tissues (25, 26). Its role in steroid metabolism is as yet not clearly defined, and furthermore, effects of 5αR inhibitors on this isozyme are uncertain (27), and any relevance to our findings is unclear.

sue. Taken together, these observations are consistent with metabolic effects of dutasteride being mediated by inhibition of 5αR1 in adipose tissue but do not exclude either a contribution from other tissues including skeletal muscle or a contribution from more potent inhibition of 5αR2 by dutasteride than finasteride. A third isozyme of 5αR has been described and is expressed in relevant tissues (25, 26). Its role in steroid metabolism is as yet not clearly defined, and furthermore, effects of 5αR inhibitors on this isozyme are uncertain (27), and any relevance to our findings is unclear. Previous studies have shown more potent effects of dutasteride than finasteride to lower circulating DHT levels (10, 28). Here, despite a higher etiocholanolone/androsterone ratio in urine, suggesting somewhat more potent overall 5αR inhibition by dutasteride, we did not find any differences in circulating DHT. This may reflect our use of a highly specific LC-MS/MS assay, although we might have obtained different results after longer-term treatment given the long half-life and very slow time to steady state for dutasteride (29). Most importantly, this indicates that differences in effects of dutasteride and finasteride on insulin sensitivity are not mediated by differences in circulating DHT. More studies are now justified to assess tissue steroid hormone concentrations and identify the downstream signaling pathways affected, particularly in adipose tissue and skeletal muscle. Such studies could test the hypotheses that either decreased androgen action and/or increased glucocorticoid action mediates these effects.

udies are now justified to assess tissue steroid hormone concentrations and identify the downstream signaling pathways affected, particularly in adipose tissue and skeletal muscle. Such studies could test the hypotheses that either decreased androgen action and/or increased glucocorticoid action mediates these effects. These results highlight a novel role for 5αR1 on modulating human metabolism; however, their clinical relevance is uncertain. The decrease in insulin sensitivity after 3 months of dutasteride (∼14%) is of similar magnitude to the beneficial effects of antidiabetic agents such as metformin (30). Impaired insulin sensitivity measured by euglycemic clamps predicts future risk of type 2 diabetes mellitus (31). Our study sample consisted mostly of healthy men who are younger than those affected by BPH with declining β-cell function (32) and increased body fat (33). Importantly, age did not confound the effect of dutasteride on insulin sensitivity in the study. Nonetheless, older men with already impaired insulin sensitivity might be more susceptible to the metabolic consequences of 5αR inhibition; the effect of disruption of 5αR1 in murine models is revealed with a high-fat diet (9). The association of BPH with the metabolic syndrome (34, 35), and the likelihood of long-term exposure to 5αR inhibitors once treatment is initiated, suggests that further studies should now be conducted to establish whether inhibition of 5αR1 has clinically important effects on adiposity and metabolism in men with BPH. Abbreviations: BMIbody mass index BPblood pressure

These results highlight a novel role for 5αR1 on modulating human metabolism; however, their clinical relevance is uncertain. The decrease in insulin sensitivity after 3 months of dutasteride (∼14%) is of similar magnitude to the beneficial effects of antidiabetic agents such as metformin (30). Impaired insulin sensitivity measured by euglycemic clamps predicts future risk of type 2 diabetes mellitus (31). Our study sample consisted mostly of healthy men who are younger than those affected by BPH with declining β-cell function (32) and increased body fat (33). Importantly, age did not confound the effect of dutasteride on insulin sensitivity in the study. Nonetheless, older men with already impaired insulin sensitivity might be more susceptible to the metabolic consequences of 5αR inhibition; the effect of disruption of 5αR1 in murine models is revealed with a high-fat diet (9). The association of BPH with the metabolic syndrome (34, 35), and the likelihood of long-term exposure to 5αR inhibitors once treatment is initiated, suggests that further studies should now be conducted to establish whether inhibition of 5αR1 has clinically important effects on adiposity and metabolism in men with BPH. Abbreviations: BMIbody mass index BPblood pressure BPHbenign prostatic hyperplasia CBGcorticosteroid binding globulin DHT5α-dihydrotestosterone EGPendogenous glucose production HOMA-IRhomeostatic model assessment of insulin resistance LC-MS/MSliquid chromatography tandem mass spectrometry LSDleast significant difference MRImagnetic resonance imaging MRSmagnetic resonance spectroscopy NEFAnonesterified fatty acid

BPHbenign prostatic hyperplasia CBGcorticosteroid binding globulin DHT5α-dihydrotestosterone EGPendogenous glucose production HOMA-IRhomeostatic model assessment of insulin resistance LC-MS/MSliquid chromatography tandem mass spectrometry LSDleast significant difference MRImagnetic resonance imaging MRSmagnetic resonance spectroscopy NEFAnonesterified fatty acid 5αR5α-reductase Rarate of appearance Rdrate of disposal WHRwaist-to-hip ratio. Acknowledgments We are grateful to the Wellcome Trust Clinical Research Facility and its Clinical Research Imaging Centre and Mass Spectrometry Core for nursing, radiography, and analytical support, respectively; Jill Harrison, Sanjay Kothiya, and Karen French for excellent technical support; and the staff of the Scottish Primary Care Research Network for assistance with recruitment. We are grateful to Professor Jane Norman, University of Edinburgh, for advice regarding the interim analysis. This work was supported by the Scottish Government Chief Scientist Office, the Wellcome Trust, and the Graham Aitken Nuffield Postgraduate Traveling Scholarship (New Zealand). Disclosure Summary: R.A. and B.R.W. are inventors on a patent owned by the University of Edinburgh, which describes the use of 5α-reduced metabolites of glucocorticoids as anti-inflammatory therapy.

Telomeres shorten with each cell division in most somatic cells. When the telomere length is reduced below a critical value, the Hayflick limit is approached and cellular senescence is triggered (1, 2). Therefore, immortal cells and cancer cells must recruit a mechanism for telomere stabilization to prevent senescence. Indeed, activation of telomerase in the presence of short telomeres is one of the most common features in many human cancers (3, 4). Telomerase is a ribonucleoprotein with two key subunits, the telomerase reverse transcriptase and the RNA component. It elongates telomeric DNA in human cells and telomerase activity is detectable in 85–90% of human malignancies. Telomerase activation is therefore a characteristic feature and a potential therapeutic target for cancer treatment (5). An alternative lengthening of telomeres (ALT) mechanism has also been described, which is thought to be based on homologous recombination (6, 7). Medullary thyroid carcinomas (MTC), arising from calcitonin-producing parafollicular cells (C cells), account for 5–8% of human thyroid cancers (8). Approximately 75% of the cases are sporadic and 25% present as multiple endocrine neoplasia (MEN) type 2 (MEN 2) including syndromic MEN 2A and MEN 2B or familial MTC (9). The development of MTC is strongly linked to activating mutations of the RET (Rearranged during Transfection) proto-oncogene. Almost all MEN 2 cases carry a constitutional RET mutation, and approximately 40% of sporadic MTCs have a somatic RET mutation M918T with prognostic implications (9, 10).

2B or familial MTC (9). The development of MTC is strongly linked to activating mutations of the RET (Rearranged during Transfection) proto-oncogene. Almost all MEN 2 cases carry a constitutional RET mutation, and approximately 40% of sporadic MTCs have a somatic RET mutation M918T with prognostic implications (9, 10). Although the involvement of TERT and telomerase has been described in follicular cell-derived thyroid cancer, the extent and role in MTC is less studied. Recently, activating TERT promoter mutations were identified as a cause of telomerase activation and associated with poor prognosis in follicular, papillary, and anaplastic thyroid carcinomas; however, such mutations have not been observed in MTC (11–15). Alternative splicing of TERT has been reported in follicular cell-derived thyroid tumors (16). Several TERT transcripts have been found, including three deletions and four insertions, which may affect telomerase enzyme activity and biological functions (17–19). Four insertions and the β− and γ− deletion result in nonfunctional proteins whereas the α− deletion is a dominant negative inhibitor of telomerase activity (20, 21). In follicular and papillary thyroid carcinomas three transcripts were detected including full-length, α− deletion and β− deletion (16). To further elucidate telomere stabilization in MTC we characterized a panel of tumors for activation of telomerase or the ALT mechanism in relation to RET mutational status, clinical characteristics, and patient outcomes.

Although the involvement of TERT and telomerase has been described in follicular cell-derived thyroid cancer, the extent and role in MTC is less studied. Recently, activating TERT promoter mutations were identified as a cause of telomerase activation and associated with poor prognosis in follicular, papillary, and anaplastic thyroid carcinomas; however, such mutations have not been observed in MTC (11–15). Alternative splicing of TERT has been reported in follicular cell-derived thyroid tumors (16). Several TERT transcripts have been found, including three deletions and four insertions, which may affect telomerase enzyme activity and biological functions (17–19). Four insertions and the β− and γ− deletion result in nonfunctional proteins whereas the α− deletion is a dominant negative inhibitor of telomerase activity (20, 21). In follicular and papillary thyroid carcinomas three transcripts were detected including full-length, α− deletion and β− deletion (16). To further elucidate telomere stabilization in MTC we characterized a panel of tumors for activation of telomerase or the ALT mechanism in relation to RET mutational status, clinical characteristics, and patient outcomes. Materials and Methods Patients and tissue specimens The study includes all patients operated on for MTC between 1986 and 2010 in the Karolinska University Hospital, Stockholm, from whom a fresh frozen tissue sample was available. All 42 cases were operated without preceding or subsequent chemotherapy or irradiation therapy. The details concerning age, sex, tumor size, TNM classification, MIB-1 proliferation index, RET mutation status, follow-up, and clinical outcome are summarized in Supplemental Table 1. Patients were followed up regularly (at 3, 6, and 12 months postoperatively, thereafter every 6 months for 5 years and then yearly), with clinical examination and measurement of basal serum calcitonin levels and radiology when suggested. In addition, 24 histopathologically verified noncancerous thyroid tissue samples obtained from patients surgically treated for other thyroid tumors than MTC were included as references. All tissue specimens were obtained through the Karolinska University Hospital Biobank. Paraffin-embedded tissues were also obtained for immunohistochemical purposes from a subset of MTCs and normal thyroid tissues. Histopathological classification of specimens was performed according to the criteria of the World Health Orginization Committee (8). RET mutation status was based on sequencing of exons 10, 11, 15, and 16 of RET in all MTC tissues (Wang et al, in preparation). Data for proliferation analysis determined by Ki-67 immunohistochemistry using the MIB-1 antibody was available for 23 of the cases (for cases with multiple surgical samples the highest MIB-1 index was chosen as representative). Informed consent was obtained from all patients and the study of the tissue samples was approved by the local Ethics Committee.

rmined by Ki-67 immunohistochemistry using the MIB-1 antibody was available for 23 of the cases (for cases with multiple surgical samples the highest MIB-1 index was chosen as representative). Informed consent was obtained from all patients and the study of the tissue samples was approved by the local Ethics Committee. Quantitative real-time PCR Total TERT expression was quantified by Quantitative real-time PCR (qRT-PCR) using Taqman Gene Expression Assays (Applied Biosystems) for TERT (Hs00 972656_m1) and 18S rRNA (Hs99999901_s1). TERT splice variants were analyzed using the methodology from Wang et al (16). The experimental procedures and quantifications are described in the Supplemental Material. Assessment of telomerase activity Telomerase activity was assayed in protein extracts with TeloTAGGG Telomerase PCR ELISA kit (Roche Diagnostics GmbH) based on the telomeric repeat amplification protocol described by Kim and Wu (22). Lysis buffer and HEK-293 cells served as negative and positive controls. All samples were performed in triplicate. The level of telomerase activity was determined in arbitrary units based on absorbance at OD450-OD690. Assessment of telomere length Mean relative telomere lengths were determined in MTCs and thyroid samples by real-time PCR according to published methodology (23), and telomere lengths were assessed by Southern blot analysis. The analyses are further described in the Supplemental Material.

Assessment of telomerase activity Telomerase activity was assayed in protein extracts with TeloTAGGG Telomerase PCR ELISA kit (Roche Diagnostics GmbH) based on the telomeric repeat amplification protocol described by Kim and Wu (22). Lysis buffer and HEK-293 cells served as negative and positive controls. All samples were performed in triplicate. The level of telomerase activity was determined in arbitrary units based on absorbance at OD450-OD690. Assessment of telomere length Mean relative telomere lengths were determined in MTCs and thyroid samples by real-time PCR according to published methodology (23), and telomere lengths were assessed by Southern blot analysis. The analyses are further described in the Supplemental Material. Detection of ALT-associated promyelocytic leukemia bodies (APBs) APBs were detected by combined telomere fluorescence in situ hybridization (Tel-FISH) and promyelocytic leukemia (PML) immunofluorescence as described (24). Representative tissue sections from 19 MTCs (nine telomerase negative and ten telomerase positive), and five normal thyroids were included. Telomerase-positive HeLa cell line was used as negative control. A detailed description of the methodology and scoring is given in the Supplemental Material.

nce as described (24). Representative tissue sections from 19 MTCs (nine telomerase negative and ten telomerase positive), and five normal thyroids were included. Telomerase-positive HeLa cell line was used as negative control. A detailed description of the methodology and scoring is given in the Supplemental Material. Statistical analysis The statistical analyses were performed using SPSS software version 18.0 for Windows. Differences between groups were evaluated by χ2 test or Fisher's exact test (where appropriate) and Mann-Whitney U test. Spearman rank order correlation was performed to analyze correlation between telomerase expression and telomerase activity as well as MIB-1 proliferation index and telomerase activity. Multivariate logistic regression was used to calculate the odds ratios and their 95% confidence interval (CI). Survival curves were illustrated by Kaplan-Meier plots, and significance was calculated by log-rank test. P values < .05 were regarded as statistically significant.

proliferation index and telomerase activity. Multivariate logistic regression was used to calculate the odds ratios and their 95% confidence interval (CI). Survival curves were illustrated by Kaplan-Meier plots, and significance was calculated by log-rank test. P values < .05 were regarded as statistically significant. Results Clinical and genetic characterization of the study cohort The clinical details for the 39 sporadic MTC and three MEN 2 cases are summarized in Supplemental Table 1. Of the 39 sporadic patients, 25 had a poor outcome at the end of follow-up, of which 12 died of the disease and 13 were alive with persistent disease, 12 were alive and free of disease at the end of follow-up, and 2 died resulting of causes unrelated to MTC. The three MEN 2 cases included a man with familial MTC diagnosed at age 14 years who survived for more than 34 years and two women with MEN 2A diagnosed at 28 and 54 years of age who are still alive after 22 and 9 years, respectively. In total, 24 of the 42 cases (57%) were RET-mutation positive (Supplemental Table 1). All three cases with familial disease harbored a constitutional mutation that was also present in the corresponding tumor DNA. Among the 39 sporadic MTCs, 17 tumors exhibited the common mutation M918T and four cases had mutations in exons 10, 11, or 15.

42 cases (57%) were RET-mutation positive (Supplemental Table 1). All three cases with familial disease harbored a constitutional mutation that was also present in the corresponding tumor DNA. Among the 39 sporadic MTCs, 17 tumors exhibited the common mutation M918T and four cases had mutations in exons 10, 11, or 15. TERT mRNA expression and telomerase activity in a subset of MTCs Among the 39 sporadic cases, 21 (54%) displayed TERT mRNA expression whereas 18 (46%) were negative (Figure 1A). The three MEN 2 cases did not reveal any TERT expression. Relative telomerase activity was then detected in the same 21 sporadic MTCs (54%), whereas the remaining 18 sporadic cases as well as three MEN 2 cases, the thyroid tissues, and the negative control were all negative. Although the relative telomerase activity showed variation between individuals (mean OD value, 1.04; range, 0.02–3.25), it showed strong positive correlation with the TERT gene expression (r = 0.967, P = .01) (Figure 1B).

ses as well as three MEN 2 cases, the thyroid tissues, and the negative control were all negative. Although the relative telomerase activity showed variation between individuals (mean OD value, 1.04; range, 0.02–3.25), it showed strong positive correlation with the TERT gene expression (r = 0.967, P = .01) (Figure 1B). Figure 1. Correlation between TERT mRNA expression and telomerase activity. A) Comparison of relative telomerase activity (filled circle) and relative TERT gene expression (red triangle) for each of the 42 cases. The left y-axis suggests the telomerase activity and the right y-axis refers to the TERT expression values given in arbitrary units. B) Scatter diagram of relative TERT expression and relative telomerase activity in the 42 MTCs shows a strong positive correlation between two variables (P = .01, Spearman rank order correlation). C) Comparison of relative telomere length (RTL) given in arbitrary units in MTCs and thyroid tissues. Box plots show mean telomere lengths determined by real-time PCR. MTCs had significantly shorter telomeres as compared with thyroid tissue. The statistical analyses were performed using Mann-Whitney U test. Cases suggested with closed circle are outliers.

ngth (RTL) given in arbitrary units in MTCs and thyroid tissues. Box plots show mean telomere lengths determined by real-time PCR. MTCs had significantly shorter telomeres as compared with thyroid tissue. The statistical analyses were performed using Mann-Whitney U test. Cases suggested with closed circle are outliers. Reduced mean telomere length in MTC In the MTCs the mean telomere lengths were significantly shorter compared with thyroid tissue samples (P = .0001; Figure 1C). Moreover, the relative telomere length of tumors with telomerase activation was shorter compared with the rest, although the difference was not significant (data not shown). However, the mean telomere length was not associated with clinical parameters of sporadic MTC cases (data not shown).

ples (P = .0001; Figure 1C). Moreover, the relative telomere length of tumors with telomerase activation was shorter compared with the rest, although the difference was not significant (data not shown). However, the mean telomere length was not associated with clinical parameters of sporadic MTC cases (data not shown). Association between telomerase activation and clinical features of sporadic MTC Sporadic MTCs with and without telomerase activation were compared for clinical phenotypes and follow-up. As shown in Table 1, telomerase activation was significantly more frequent in male compared with female patients (P = .02) and correlated with a more advanced stage at diagnosis (P < .0001) as well as larger tumors (P = .027). All telomerase-positive cases had stage III or stage IV disease. In the telomerase-negative group, nine patients had stage I or II disease and nine patients had stage III or IV disease. The three patients with MEN 2 had stage I or II disease. Furthermore, there was a positive correlation between telomerase activity and MIB-1 proliferation index (r = 0.683, P = .01; Supplemental Figure 1). These findings suggest that telomerase activation is associated with more aggressive growth in sporadic MTC. Table 1. Comparison of Telomerase Activation with Characteristics of Sporadic MTC

Association between telomerase activation and clinical features of sporadic MTC Sporadic MTCs with and without telomerase activation were compared for clinical phenotypes and follow-up. As shown in Table 1, telomerase activation was significantly more frequent in male compared with female patients (P = .02) and correlated with a more advanced stage at diagnosis (P < .0001) as well as larger tumors (P = .027). All telomerase-positive cases had stage III or stage IV disease. In the telomerase-negative group, nine patients had stage I or II disease and nine patients had stage III or IV disease. The three patients with MEN 2 had stage I or II disease. Furthermore, there was a positive correlation between telomerase activity and MIB-1 proliferation index (r = 0.683, P = .01; Supplemental Figure 1). These findings suggest that telomerase activation is associated with more aggressive growth in sporadic MTC. Table 1. Comparison of Telomerase Activation with Characteristics of Sporadic MTC Parameter Telomerase Activation Telomerase Negative P Valuea Informative cases (n) 21 18 Age at diagnosis (n = 39) Mean (min.-max.), y 53 (13–87) 58.2 (21–75) .294 Sex (n = 39) .019 Female 9 15 Male 12 3 Tissue type (n = 38) .282 Primary tumor 13 14 Metastasis 8 3 Tumor sizeb (n = 38) .027 T1 4 7 T2 5 7 T3 3 4 T4 8 0 TNM stage (n = 39) <.0001 Stage I 0 3 Stage II 0 6 Stage III 1 5 Stage IV 20 4 MIB1 proliferation index (n = 22) .134 0–2% positive nuclei 5 7 3–10% positive nuclei 5 2 >10% positive nuclei 3 0 Follow-up (n = 39) .005 Mean (min.-max.), y 7.3 (1–28) 13.7 (2–35) Outcome (n = 39) <.0001 Alive and disease free 0 12 Alive with spread disease 8 5 Dead (resulting from disease) 13 (12) 1 (0) a Significant P-values are indicated in bold.

.134 0–2% positive nuclei 5 7 3–10% positive nuclei 5 2 >10% positive nuclei 3 0 Follow-up (n = 39) .005 Mean (min.-max.), y 7.3 (1–28) 13.7 (2–35) Outcome (n = 39) <.0001 Alive and disease free 0 12 Alive with spread disease 8 5 Dead (resulting from disease) 13 (12) 1 (0) a Significant P-values are indicated in bold. b T1, <20 mm; T2, 20–40 mm; T3, >40 mm but limited to thyroid; T4, tumor extends beyond the thyroid capsule.

.134 0–2% positive nuclei 5 7 3–10% positive nuclei 5 2 >10% positive nuclei 3 0 Follow-up (n = 39) .005 Mean (min.-max.), y 7.3 (1–28) 13.7 (2–35) Outcome (n = 39) <.0001 Alive and disease free 0 12 Alive with spread disease 8 5 Dead (resulting from disease) 13 (12) 1 (0) a Significant P-values are indicated in bold. b T1, <20 mm; T2, 20–40 mm; T3, >40 mm but limited to thyroid; T4, tumor extends beyond the thyroid capsule. Telomerase activation as a prognostic factor of MTC patient survival Sporadic MTC cases were divided in two groups according to the outcomes at follow-up, including cases free of disease (alive and free of disease or dead resulting from other cause) and cases with persistent disease (alive with persistent disease or dead resulting from disease), and compared concerning clinical phenotypes and molecular features. Telomerase expression and telomerase activity were significantly associated with poor outcome (Table 2) as well as late-stage disease. Among the 21 telomerase-positive sporadic MTC cases, 12 died as a result of MTC, one died as a result of renal failure, and 8 were alive with spread disease at the end of follow-up. In the telomerase-negative group, all patients were alive except one who died as a result of lymphoma, including 12 who were free of disease and 5 patients with spread disease. Telomerase activity had a positive association with different outcomes; cases free of disease showed the lowest telomerase activity (Figure 2A). The Kaplan-Meier analysis showed a significantly shorter survival in the group with telomerase activation (P < .0001) (Figure 2B). Hence, we found a strong correlation between clinical outcome and telomerase activation (P < .0001). The multivariate analysis performed by logistic regression test showed that only telomerase activation was independently correlated with a poor outcome in MTC patients (odds ratio (OR), 24; 95% CI, 1.8–339; P = .017). Stage did not reach statistical significance in the multivariate analysis (OR, 3; 95% CI, 0.35–27; P = .345).

tivariate analysis performed by logistic regression test showed that only telomerase activation was independently correlated with a poor outcome in MTC patients (odds ratio (OR), 24; 95% CI, 1.8–339; P = .017). Stage did not reach statistical significance in the multivariate analysis (OR, 3; 95% CI, 0.35–27; P = .345). Table 2. Comparison of Telomerase and TERT Findings of Sporadic MTC with Final Outcome Parameter Free of Disease Persistent Disease P Valuea Informative cases (n) 14 25 Telomerase activation (n = 39) .0001 Postive 1 20 Negative 13 5 TERT mRNA expression (n = 39) .0001 Postive 1 20 Negative 13 5 Telomerase activity (n = 39) .0001 Mean 0.23 0.73 Telomere length (n = 39) .46 Mean (min.-max.) 0.86 (0.5–1.6) 0.83 (0.3–2.1) TERT promoter mutation (n = 39)b 0 0 a Significant P values are indicated in bold. b TERT promoter mutation data have been published in (15). Figure 2. Telomerase is a prognostic marker for survival of sporadic MTC patients. A) Box plot shows relative telomerase activity in MTC tumors from patients with different clinical outcomes. Differences between groups were evaluated by Mann-Whitney U test. Cases labeled with circle are outliers. B) Kaplan-Meier plots showing significantly shorter survival for patients with telomerase activation as compared with cases without telomerase activation (P < .0001). The statistical analysis was performed using log-rank test.

between groups were evaluated by Mann-Whitney U test. Cases labeled with circle are outliers. B) Kaplan-Meier plots showing significantly shorter survival for patients with telomerase activation as compared with cases without telomerase activation (P < .0001). The statistical analysis was performed using log-rank test. We found that telomerase activation was associated with aggressive disease and poor clinical outcome. Given that RET M918T is reported to be a marker of poor prognostic in MTC, we compared these two parameters in 39 sporadic MTC cases. However, no significant association was detected between M918T mutation and telomerase activation (Supplemental Figure 2).

s associated with aggressive disease and poor clinical outcome. Given that RET M918T is reported to be a marker of poor prognostic in MTC, we compared these two parameters in 39 sporadic MTC cases. However, no significant association was detected between M918T mutation and telomerase activation (Supplemental Figure 2). TERT alternative splice variants in MTCs TERT expression was further characterized concerning alternative splice variants in the 21 sporadic cases with positive expression of total TERT. Four TERT alternative splice variants were determined. Three of these four were detected in subsets of the tumors, as illustrated for representative cases in Figure 3A. Fourteen of the 21 cases exhibited a combination of full-length transcript, inhibitory α− deletion (382 bp), nonfunctional β− deletion (236 bp), and one case only showed full-length transcript. The remaining six cases showed the nonfunctional β− deletion only. The γ− deletion (91 bp) was not detected in any of the 21 cases examined. Cases with full-length transcript tended to have higher telomerase activity and shorter relative telomere length; however, the differences were not statistically significant (data not shown). Survival analysis showed a significantly shorter survival in the group with full-length transcript compared with cases without full-length transcript (P = .04) (Figure 3B). Hence, we found a correlation between clinical outcome and TERT full-length transcript. However, the presence or absence of full-length or α− deletion transcript was neither associated with other clinical parameters nor with RET M918T mutation. Seven of the 15 cases expressing the full-length transcript showed high levels of full-length TERT (>35% of the total TERT); however, there was no difference concerning clinical parameters for cases with high or low expression levels of full-length TERT.

iated with other clinical parameters nor with RET M918T mutation. Seven of the 15 cases expressing the full-length transcript showed high levels of full-length TERT (>35% of the total TERT); however, there was no difference concerning clinical parameters for cases with high or low expression levels of full-length TERT. Figure 3. Detection of TERT alternative splice variants in MTC. A) Representative electrophoretic gels showing the TERT full-length variant (FL), the α deletion (α−) and the β deletion (β−) detected by nested PCR. As shown in the lower panel the γ deletion (γ−) was not detected in this study. B) Kaplan-Meier plots showing significantly shorter survival for patients with the telomerase full-length transcript as compared with cases without the full-length transcript (P = .04). The statistical analysis was performed using log-rank test.

own in the lower panel the γ deletion (γ−) was not detected in this study. B) Kaplan-Meier plots showing significantly shorter survival for patients with the telomerase full-length transcript as compared with cases without the full-length transcript (P = .04). The statistical analysis was performed using log-rank test. Detection of the ALT activation in telomerase-negative MTCs To explore whether the ALT mechanism is involved in MTC cases without telomerase activation, we performed Southern blot analysis of 17 telomerase-negative and 6 telomerase-positive MTCs to evaluate the length of telomeres. As exemplified in Figure 4A, 11 of 17 telomerase-negative MTC cases (65%) exhibited a heterogeneous distribution of telomere lengths that is the characteristic of ALT-positive cells whereas telomerase-positive cases displayed a homogeneous pattern. Tel-FISH was performed on embedded tissue from nine telomerase-negative samples, 10 telomerase-positive samples, and five thyroid tissues. The prevalence of APBs varied among the cases. Six of nine telomerase-negative slides (67%) showed APBs whereas the rest had undetectable APB (Figure 4). As expected, APBs were not found in five thyroid and 10 telomerase-positive MTC tissues. The observed telomere length heterogeneity and APBs suggest that the ALT mechanism is involved in a subset of sporadic MTC. Intriguingly, several tumors showed neither telomerase activation nor involvement of ALT. ALT-positive tumors showed lower MIB-1 proliferation index (median, 1%; range, 1–3%) compared with ALT-negative cases (median, 3%; range, 1–80%) (P = .024; Supplemental Figure 1).

mechanism is involved in a subset of sporadic MTC. Intriguingly, several tumors showed neither telomerase activation nor involvement of ALT. ALT-positive tumors showed lower MIB-1 proliferation index (median, 1%; range, 1–3%) compared with ALT-negative cases (median, 3%; range, 1–80%) (P = .024; Supplemental Figure 1). Figure 4. Detection of ALT in telomerase-negative MTC. A) Southern blot analysis of terminal restriction fragment (TRF) shows the distribution of telomere length in telomerase-negative MTC. The length of telomeres of sample 2, 3, 4, 6, and 7 ranged from <2 kb to >21.2 kb. The length of telomeres of sample 1, 5, and 8 ranged from 3–8 kb, which had a more homogeneous distribution and a shorter telomere length. The heterogeneous distribution was scored as TRF (+), and detection of APBs is suggested below. Molecular weight marker (MWM), and positive (P) and negative (N) controls are included. B and C) Combined PML immunofluorescence (green) and telomere FISH (red). B) Analysis of MTC cases at ×100 magnification. Arrows suggest PML colocalized with telomere DNA foci, representing APBs. C) Tel-FISH performed in normal noncancerous thyroid control tissue.

nd positive (P) and negative (N) controls are included. B and C) Combined PML immunofluorescence (green) and telomere FISH (red). B) Analysis of MTC cases at ×100 magnification. Arrows suggest PML colocalized with telomere DNA foci, representing APBs. C) Tel-FISH performed in normal noncancerous thyroid control tissue. Discussion In the present study we found activation of telomerase in approximately half the MTC cases. Three different splice isoforms of TERT were detected among the telomerase positive cases; cases with expression of full-length transcripts showed a shorter survival. Telomerase activation had strong prognostic influence on patient survival but was independent of stage and the common RET M918T mutation. Activation of the ALT mechanism was demonstrated while a subset of cases was negative for both telomerase and ALT, suggesting that additional mechanisms of telomere stabilization might be operative in MTC.

ad strong prognostic influence on patient survival but was independent of stage and the common RET M918T mutation. Activation of the ALT mechanism was demonstrated while a subset of cases was negative for both telomerase and ALT, suggesting that additional mechanisms of telomere stabilization might be operative in MTC. Telomerase is a multienzymatic complex, which can be regulated at different levels (25). Because TERT gene expression is not always correlated with the enzyme activity, many studies have found that either the telomerase activity or the TERT expression is a marker of tumor aggressiveness and poor prognosis. In our study, there was a significant concordance between telomerase activity and TERT expression (r = 0.9). Still, some MTCs showed high TERT expression together with low telomerase activity. This observation could be related to differences in post-transcriptional modifications, translational efficiency, expression of other telomerase components, or TERT phosphorylation. We found that half of the investigated MTCs showed telomerase activation. Taken together, previously published studies have also reported recurrent telomerase-positive and -negative MTCs. However, because most studies are based on one or a few MTC cases only, direct comparisons of frequencies is not meaningful (Supplemental Table 3). Furthermore, the result may vary depending on the cases included concerning sporadic or MEN 2–related MTC as well as metastatic or nonmetastatic cases. Interestingly, low levels of telomerase activity were reported in MTCs compared with HEK-293 cells (26), and even low telomerase activity of less than 5% of HEK-293 may prolong the replicative life span of MTC cell strains (27).

uded concerning sporadic or MEN 2–related MTC as well as metastatic or nonmetastatic cases. Interestingly, low levels of telomerase activity were reported in MTCs compared with HEK-293 cells (26), and even low telomerase activity of less than 5% of HEK-293 may prolong the replicative life span of MTC cell strains (27). Approximately two thirds of the patients in the telomerase-positive group died as a result of MTC and the remaining patients had spread disease. Telomerase activation was independently correlated with a worse outcome of MTC; cases with telomerase activation had a significantly shorter survival. In adults, telomerase is silent in most tissues whereas expressed in stem cells, male germ line cells, and activated immune cells only. Somatic cells with short telomeres enter a senescent or apoptosis state after several divisions without activating telomerase (28, 29). However, some cells could bypass the checkpoints and avoid entering crisis by activating a mechanism of telomere maintenance. The primary mechanism for this is activation of telomerase. As shown in this study, half of the MTC cases used the primary mechanism to stabilize telomeres, i.e. telomerase activation. The role of telomerase is generally regarded as maintaining telomere length. However, accumulating evidence suggests that telomerase have additional functions in regulation of cell growth and survival, which may contribute to tumorigenesis (30). Here, although telomerase was not universally activated in MTCs, the frequency of telomerase activation was significantly associated with poor outcome, with large and late-stage tumors, demonstrating that up-regulation of telomerase is a relatively important event in MTC development. Several clinical and histopathological parameters such as lymph node invasion and distant metastases are important prognostic factors. Therefore, the association between telomerase activation and aggressive behavior (poor clinical outcome and advanced tumor stage), could be used to predict the prognosis of MTC patients, aiding in decision making for follow-up and additional therapy.

ode invasion and distant metastases are important prognostic factors. Therefore, the association between telomerase activation and aggressive behavior (poor clinical outcome and advanced tumor stage), could be used to predict the prognosis of MTC patients, aiding in decision making for follow-up and additional therapy. The TERT gene is known to generate seven alternative splice forms and produce multiple transcripts; here we detected three TERT splice forms. The α− deletion transcript has a reverse transcriptase motif A and can negatively regulate telomerase activity, whereas the β− deletion and the γ− deletion transcripts are without any known function (17, 20, 21). In a previous study of other cancers, telomerase activity was shown to depend on the existence of full-length TERT gene expression (18, 19). We found that most of the MTCs with full-length transcript exhibited relatively high telomerase activity. However, a few other cases exhibited high telomerase activity with low expression of the full-length splice form. Hence, in MTC telomerase, activity was not fully dependent of expression of the full-length transcript. The cases with full-length transcripts showed a shorter survival, implying that full-length TERT play an important role in telomerase functions in MTC.

ase activity with low expression of the full-length splice form. Hence, in MTC telomerase, activity was not fully dependent of expression of the full-length transcript. The cases with full-length transcripts showed a shorter survival, implying that full-length TERT play an important role in telomerase functions in MTC. Telomerase activation is the predominant mechanism of telomere maintenance in most malignancies. However, the mechanism for activation of telomerase in vivo is not fully understood. Some type of genetic event is suggested such as gene amplification, duplication, and translocation of the TERT locus, or the activating TERT promoter mutation. In our previous study, TERT promoter mutation led to telomerase activation in papillary and follicular thyroid cancer (15). However mutations were not revealed in the same series of MTCs investigated here (15), demonstrating that the telomerase activation was not a result of TERT promoter mutation. We also investigated the possible connection between telomerase activation and the common M918T RET mutation, which is associated with a less favorable clinical outcome (10, 31). In the present study, this M918T mutation was not significantly associated with specific clinical characteristics. Telomerase activation was twice as frequently observed in the M918T mutation group than the group without M918T mutation. However, the difference did not reach statistical significance (P = .065). Telomerase may be proliferation regulated, and there is a positive correlation between RET mutations and Ki-67 expression in MTC (32). In the present study, MIB-1 proliferation index was positively correlated with telomerase activity and negatively associated with ALT involvement.

h statistical significance (P = .065). Telomerase may be proliferation regulated, and there is a positive correlation between RET mutations and Ki-67 expression in MTC (32). In the present study, MIB-1 proliferation index was positively correlated with telomerase activity and negatively associated with ALT involvement. Besides the primary mechanism of telomere stabilization by telomerase activation, ALT-based chromosome recombination is involved in a subset of cancers (7, 33). We found involvement of ALT in a subset of MTC cases without telomerase activation. Heterogeneity in telomere length was observed in 11 of 17 cases, and presence of APBs was observed in six telomerase-negative cases. This is in accordance with previous observations of features of ALT, and it is the first description of a telomerase-independent telomere mechanism in MTC. The ALT mechanism has been described for tumor cells lacking telomerase activation (34), which is common in sarcomas and glioblastoma multiforme. The ALT activation is thought to be based on homologous recombination and copy switching of telomeric repeats. Telomere shortening in the absence of telomerase followed by recombination results in heterogeneous telomere length and the APBs are storage places for DNA synthesis (35). A subset of telomerase-negative MTCs exhibited heterogeneous telomeres and APBs, which was in accordance with the results of chromosomal instability. All these findings suggest the existence of the ALT mechanism in MTCs.

on results in heterogeneous telomere length and the APBs are storage places for DNA synthesis (35). A subset of telomerase-negative MTCs exhibited heterogeneous telomeres and APBs, which was in accordance with the results of chromosomal instability. All these findings suggest the existence of the ALT mechanism in MTCs. Of interest is the observation that some MTC cases exhibited neither telomerase activation nor features of ALT, which may suggest the existence of additional effective mechanisms of telomere stabilization. Two groups described an SV40-immortalized human cell line with neither telomerase activation nor APBs (36, 37), and another group showed the presence of an efficient telomere stabilization mechanism different from telomerase activation and ALT in non–small-cell lung cancer cell lines (38). In addition, a significant number of cancer cell lines maintain telomeres without signs of telomerase activation or ALT (39). These studies support the existence of other telomere stabilization mechanisms besides telomerase and ALT activation, which is in agreement with the present study. The telomere length is the result of a dynamic balance between shortening and elongation. Short telomeres are associated with poor prognosis of human carcinomas including thyroid cancer (40). The presence of short telomeres has been reported in patients with sporadic cancers of for example bladder, lung, kidney, and head and neck (4). We found that MTCs presented shorter telomeres compared with thyroid tissues, which is in accordance with previous observations in other cancers.

ncluding thyroid cancer (40). The presence of short telomeres has been reported in patients with sporadic cancers of for example bladder, lung, kidney, and head and neck (4). We found that MTCs presented shorter telomeres compared with thyroid tissues, which is in accordance with previous observations in other cancers. In conclusion, at least two mechanisms are involved in telomere maintenance in MTC including telomerase and ALT activation. Alternative splicing of TERT partly accounts for activation of the telomerase protein. Telomerase activation is of prognostic importance in MTC, and could serve as a prognostic marker for MTC patients. In addition, our data imply the existence of other possible mechanisms for telomere stabilization in MTC. Hence, MTC might be a useful tool for investigating the molecular background of telomere stabilization. Abbreviations: ALTalternative lengthening of telomeres APBpromyelocytic leukemia body CIconfidence interval Tel-FISHtelomere fluorescence in situ hybridization MENmultiple endocrine neoplasia MEN 2multiple endocrine neoplasia type 2 MTCmedullary thyroid carcinoma ORodds ratio PMLpromyelocytic leukemia. Acknowledgments The authors thank Ms. Lisa Ånfalk for her excellent assistance in collecting tissue samples, and the medical genetics group for valuable discussions. This work was supported by grants from the Swedish Cancer Society, the Swedish Research Council, the Cancer Society in Stockholm, the Gustav V Jubilee Foundation, the Stockholm County Council, and Karolinska Institutet. Disclosure Summary: The authors have nothing to disclose.

Familial glucocorticoid deficiency (FGD; Mendelian Inheritance in Man 202200) is characterized by ACTH resistance and isolated glucocorticoid deficiency, with typical biochemical findings of low serum cortisol levels and high plasma ACTH (1). Patients commonly present with hyperpigmentation, a consequence of overstimulation of melanocortin 1 receptor by proopiomelanocortin products and hypoglycemia, which, if untreated, can lead to neurological sequelae. Clinical presentation includes failure to thrive in the neonatal or early childhood periods and an increased susceptibility to infections (1). This disorder can lead to significant morbidity and is potentially fatal if untreated. Classical ACTH resistance is primarily caused by mutations in the ACTH receptor [melanocortin 2 receptor (MC2R)] and the MC2R accessory protein (MRAP), required for trafficking MC2R to the cell surface and subsequent signaling (2, 3). Mutations in MC2R and MRAP thus disrupt ACTH signaling in the zona fasciculata, resulting in isolated glucocorticoid deficiency.

ed by mutations in the ACTH receptor [melanocortin 2 receptor (MC2R)] and the MC2R accessory protein (MRAP), required for trafficking MC2R to the cell surface and subsequent signaling (2, 3). Mutations in MC2R and MRAP thus disrupt ACTH signaling in the zona fasciculata, resulting in isolated glucocorticoid deficiency. Our group recently described two novel pathogenic mechanisms for FGD (4, 5). A mutation in the DNA helicase, minichromosome maintenance 4 (MCM4) is responsible for FGD in the Irish traveler population, incorporating a phenotype of short stature, natural killer cell deficiency, and chromosomal fragility (4). Additionally, mutations in NNT, encoding the mitochondrial antioxidant nicotinamide nucleotide transhydrogenase, account for 10% of cases in our FGD cohort (5). Nicotinamide nucleotide transhydrogenase (NNT), located in the inner mitochondrial membrane, provides the high concentrations of reduced nicotinamide adenine dinucleotide phosphate (NADPH) required by the thioredoxin and glutathione systems to detoxify mitochondrial H2O2. In this study, we describe the first homozygous mutation in the mitochondrial selenoprotein, thioredoxin reductase 2 (TXNRD2) associated with FGD in an extended consanguineous Kashmiri kindred (Figure 1A).

Our group recently described two novel pathogenic mechanisms for FGD (4, 5). A mutation in the DNA helicase, minichromosome maintenance 4 (MCM4) is responsible for FGD in the Irish traveler population, incorporating a phenotype of short stature, natural killer cell deficiency, and chromosomal fragility (4). Additionally, mutations in NNT, encoding the mitochondrial antioxidant nicotinamide nucleotide transhydrogenase, account for 10% of cases in our FGD cohort (5). Nicotinamide nucleotide transhydrogenase (NNT), located in the inner mitochondrial membrane, provides the high concentrations of reduced nicotinamide adenine dinucleotide phosphate (NADPH) required by the thioredoxin and glutathione systems to detoxify mitochondrial H2O2. In this study, we describe the first homozygous mutation in the mitochondrial selenoprotein, thioredoxin reductase 2 (TXNRD2) associated with FGD in an extended consanguineous Kashmiri kindred (Figure 1A). Figure 1. Pedigree of the affected kindred and identification of p.Y447X TXNRD2 mutation leading to the loss of TXNRD2 protein. A, Pedigree of affected patients. Black-filled symbols indicate individuals homozygous and half-filled indicate individuals heterozygous for the mutation. White-filled symbols indicate wild-type individuals. Gray-filled symbols indicate individuals not tested. The asterisks denote the three affected individuals who were subjected to whole-exome sequencing. B, Gene structure of TXNRD2; p.Y447X mutation prior to the selenocysteine active site (SEC). C, Partial sequence chromatograms of genomic DNA from a wild-type, heterozygote carrier and a patient, showing the base change from T to G in exon 15, resulting in a premature stop codon in the affected individual. D, Lysates from a homozygous patient, heterozygote carrier, and control human lymphocytes were immunoblotted with an anti-TXNRD2 antibody. Although the control and the heterozygote carriers expressed the 56-kDa protein, this is absent in the homozygote patient, with no evidence of a truncated protein. All individuals express cytoplasmic TXNRD1 normally. E, RT-PCR of cDNA from the patient, heterozygote carrier, and control suggested nonsense mediated decay of mRNA.

the control and the heterozygote carriers expressed the 56-kDa protein, this is absent in the homozygote patient, with no evidence of a truncated protein. All individuals express cytoplasmic TXNRD1 normally. E, RT-PCR of cDNA from the patient, heterozygote carrier, and control suggested nonsense mediated decay of mRNA. Materials and Methods Study approval This study was approved by the Outer North East London Research Ethics Committee, reference number 09/H0701/12. Candidate gene sequencing Genomic DNA was extracted from peripheral blood leukocytes of affected individuals and their family members after obtaining informed consent from them and/or their parents. Sequencing of coding exon/intron boundaries of MC2R, MRAP, STAR, and NNT genes had previously been undertaken by conventional Sanger sequencing and no mutations were found. The coding exons of MCM4 and TXNRD2 were sequenced in 50 patients with a clinical diagnosis of FGD by a combination of whole-exome and Sanger sequencing. Whole-exome sequencing data were analyzed in each individual and for any exonic region with a coverage of less than 20 reads. PCR amplification and Sanger sequencing was carried out. No mutations were found.

sequenced in 50 patients with a clinical diagnosis of FGD by a combination of whole-exome and Sanger sequencing. Whole-exome sequencing data were analyzed in each individual and for any exonic region with a coverage of less than 20 reads. PCR amplification and Sanger sequencing was carried out. No mutations were found. Whole-exome sequencing Whole-exome sequencing using the Illumina HiSeq 2000 sequencer was conducted on three affected individuals (samples processed by Otogenetics Corp). The whole-exome sequencing samples were prepared as an Illumina sequencing library, and in the second step, the sequencing libraries were enriched using the Agilent V4 enrichment kit. The captured libraries were sequenced and downstream analysis conducted via DNAnexus (see Acknowledgments). Single-nucleotide polymorphisms (SNPs), with a threshold coverage of at least 10 reads on the respective nucleotide, were assessed. The number of variants was reduced by the following strategy: 1) identifying variants that were common to all three individuals; 2) excluding variants that were heterozygous; 3) removing variants, annotated in SNP databases (Ensembl SNP database, release 54), with a minor allele frequency of greater than 0.01; and 4) evaluating nonsynonymous coding variants, splice variants, and indels only. Finally, four candidate variants in three genes (OR2T35, MUC4, and TXNRD2) were investigated for the segregation of disease within the kindred by Sanger sequencing.

tabase, release 54), with a minor allele frequency of greater than 0.01; and 4) evaluating nonsynonymous coding variants, splice variants, and indels only. Finally, four candidate variants in three genes (OR2T35, MUC4, and TXNRD2) were investigated for the segregation of disease within the kindred by Sanger sequencing. PCR and sequencing Each exon of the genes of interest including intronic boundaries was amplified by PCR using specific primers (primer sequences are listed in Supplemental Table 1). The reaction mixture contained 100 ng DNA template, 1× PCR buffer, 200 μM each deoxynucleotide triphosphate, 200 mM each primer, and 1 U Taq DNA polymerase (Sigma-Aldrich). Cycling conditions were as follows: 95°C for 5 minutes (one cycle); 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds (30 cycles); and 72°C for 5 minutes. PCR products were visualized on 1% agarose gel and sequenced using the ABI Prism Big Dye sequencing kit and an ABI 3700 automated DNA sequencer (Applied Biosystems), in accordance with the manufacturer's instructions. RNA extraction and cDNA sequencing Total RNA was isolated and purified using the PAXgene blood RNA (QIAGEN) system according to the manufacturer's instructions. The RNA was reverse transcribed and cDNA subsequently used as a template for PCR amplification and sequencing of TXNRD2 exons 14–16 (primer sequences listed in Supplemental Table 1).

sequencing Total RNA was isolated and purified using the PAXgene blood RNA (QIAGEN) system according to the manufacturer's instructions. The RNA was reverse transcribed and cDNA subsequently used as a template for PCR amplification and sequencing of TXNRD2 exons 14–16 (primer sequences listed in Supplemental Table 1). Leukocyte separation for immunoblotting Fresh whole blood was collected in EDTA-containing tubes. Mononuclear cells were extracted using a gradient density centrifugation method with Histopaque-1077 according to the manufacturer's protocol (Sigma-Aldrich). Cells were lysed with radioimmunoprecipitation assay buffer [50 mM Tris-HCL (pH 8.0), with 150 mM sodium chloride, 1% IGEPAL CA-630 (Nonidet P-40), 0.5% sodium deoxycholate, and 0. 1% sodium dodecyl sulfate], supplemented with complete, Mini, EDTA-free protease inhibitor cocktail tablets (Roche), placed on ice for 30 minutes, and centrifuged at 15 000 × g for 12 minutes at 4°C. The supernatant was subsequently added to an equal volume of Laemmli loading buffer. Protein expression of TXNRD2 in patient lysates was determined by immunoblotting with a rabbit polyclonal anti-TXNRD2 antibody (Sigma-Aldrich). In vitro adrenocortical model Cell culture H295R adrenocortical tumor cells were cultured in GIBCO DMEM/F12-Ham (1:1) + GlutaMAX-I, supplemented with 5% NuSerum, penicillin/streptomycin, and insulin-transferrin-selenium. HEK293T cells were maintained in DMEM/F12 (1:1) supplemented with 2% NuSerum and penicillin/streptomycin. All cells were incubated in a humidified incubator at 37°C and 5% CO2.

ere cultured in GIBCO DMEM/F12-Ham (1:1) + GlutaMAX-I, supplemented with 5% NuSerum, penicillin/streptomycin, and insulin-transferrin-selenium. HEK293T cells were maintained in DMEM/F12 (1:1) supplemented with 2% NuSerum and penicillin/streptomycin. All cells were incubated in a humidified incubator at 37°C and 5% CO2. Short hairpin RNA (shRNA) lentiviral transduction Stable knockdown of TXNRD2 was established in H295R human adrenocortical cells by lentiviral shRNA transduction. Lentiviral plasmids (V3LHS_354173) were obtained from OpenBiosystems in a p.GIPZ backbone and contained shRNA specific for human TXNRD2 (NM 10587) under the control of the cytomegalovirus promoter, plus the puromycin resistance and green fluorescence protein (GFP) genes. HEK293T cells (packaging cells) were transiently transfected with the shRNA plasmids together with packaging vectors, PMDG.2 plasmid and the cytomegalovirus plasmid 8.74 plasmid, using Lipofectamine 2000 (Invitrogen Life Technologies) as per the manufacturer's guidelines. Two days after transfection, virus-containing media were collected, filtered using a 0.22-μm filter and used to transduce H295R cells. Five days after infection, GFP-positive cells were selected in 5 μg/mL puromycin. Transduction efficiency was determined by fluorescence microscopy. A scrambled (control) cell line was generated using a lentiviral plasmid vector containing a shRNA insert that does not target human and mouse genes (Open Biosystems).

ls. Five days after infection, GFP-positive cells were selected in 5 μg/mL puromycin. Transduction efficiency was determined by fluorescence microscopy. A scrambled (control) cell line was generated using a lentiviral plasmid vector containing a shRNA insert that does not target human and mouse genes (Open Biosystems). Immunoblotting An anti-TXNRD2 antibody was used to assess patient TXNRD2 expression by immunoblotting and confirm protein knockdown in the in vitro studies. Whole-cell lysates were prepared by washing the cells three times in PBS, and cells were lysed using radioimmunoprecipitation assay buffer on ice for 30 minutes. These were centrifuged at 13 000 × g for 12 minutes at 4°C and the supernatant added to Laemmli buffer. To obtain lysates in nonreducing conditions, the supernatant was added to a nonreducing buffer [250 mM Tris HCl (pH 6.8), with 8% sodium dodecyl sulfate, 40% glycerol, and 0.02% bromophenol blue].

ice for 30 minutes. These were centrifuged at 13 000 × g for 12 minutes at 4°C and the supernatant added to Laemmli buffer. To obtain lysates in nonreducing conditions, the supernatant was added to a nonreducing buffer [250 mM Tris HCl (pH 6.8), with 8% sodium dodecyl sulfate, 40% glycerol, and 0.02% bromophenol blue]. Samples were heated at 95–100°C for 5 minutes and loaded on 4%–12% SDS-PAGE gels. Proteins were then transferred to nitrocellulose membrane (GE Healthcare Life Sciences) using a semidry transfer blot (Bio-Rad Laboratories) at 15 V for 1 hour. Blots were immunolabeled overnight with a rabbit polyclonal anti-TXNRD2 antibody (Sigma-Aldrich; immunogen incorporating amino acid sequence 203–353 of TXNRD2) at 1:500 dilution, mouse monoclonal glyceraldehyde-3-phosphate dehydrogenase (GAPDH) at 1:5000 dilution (Abcam) as a loading control, or rabbit polyclonal anti-TXNRD1 antibody (Abcam). Rabbit polyclonal anti-Peroxiredoxin III antibody (Proteintech) was used at a 1:400 dilution. Visualization of the proteins was performed using Alexa-fluor 680 and 800 secondary antibodies (Invitrogen) at a 1:5000 dilution and the Li-CoR Odyssey system.

oading control, or rabbit polyclonal anti-TXNRD1 antibody (Abcam). Rabbit polyclonal anti-Peroxiredoxin III antibody (Proteintech) was used at a 1:400 dilution. Visualization of the proteins was performed using Alexa-fluor 680 and 800 secondary antibodies (Invitrogen) at a 1:5000 dilution and the Li-CoR Odyssey system. Quantitative real-time PCR Total RNA was extracted from cultured H295R cells 10 days after puromycin selection using the RNeasy kit (QIAGEN) according to the manufacturer's protocol. RNA samples were quantified with a spectrophotometer and 1.0 μg of RNA from each sample was reverse transcribed after deoxyribonuclease treatment. Quantitative RT-PCR was set up in triplicate (per sample) on a Stratagene Mx3000P thermocycler using KAPA SYBR Fast quantitative PCR master mix with 200 nM forward and reverse primers targeted to TXNRD2 or GAPDH (primer sequences listed in Supplemental Table 1); giving a total volume of 10 μL. After an initial denaturation step of 3 minutes at 95°C, PCR cycling was performed for 40 cycles of 95°C for 3 seconds, 55°C for 20 seconds, and 72°C for 1 second, followed by one cycle of 1 minute at 95°C, 55°C for 30 seconds, and 95°C for 30 seconds. The expression of TXNRD2 and GAPDH mRNA was investigated using a panel of cDNAs derived from 11 adult tissues (adrenal cortex, heart, liver, testes, thyroid, lung, kidney, spleen, ovary, brain, and skeletal muscle). Quantitative real-time PCR was performed for each tissue as described above.

Quantitative real-time PCR Total RNA was extracted from cultured H295R cells 10 days after puromycin selection using the RNeasy kit (QIAGEN) according to the manufacturer's protocol. RNA samples were quantified with a spectrophotometer and 1.0 μg of RNA from each sample was reverse transcribed after deoxyribonuclease treatment. Quantitative RT-PCR was set up in triplicate (per sample) on a Stratagene Mx3000P thermocycler using KAPA SYBR Fast quantitative PCR master mix with 200 nM forward and reverse primers targeted to TXNRD2 or GAPDH (primer sequences listed in Supplemental Table 1); giving a total volume of 10 μL. After an initial denaturation step of 3 minutes at 95°C, PCR cycling was performed for 40 cycles of 95°C for 3 seconds, 55°C for 20 seconds, and 72°C for 1 second, followed by one cycle of 1 minute at 95°C, 55°C for 30 seconds, and 95°C for 30 seconds. The expression of TXNRD2 and GAPDH mRNA was investigated using a panel of cDNAs derived from 11 adult tissues (adrenal cortex, heart, liver, testes, thyroid, lung, kidney, spleen, ovary, brain, and skeletal muscle). Quantitative real-time PCR was performed for each tissue as described above. MTS (3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium), inner salt assay The CellTiter96 Aqueous nonradioactive cell proliferation assay (Promega) was used to assess cell viability. Cells were plated at a density of 10 000 cells per well in a 96-well plate on day 0. Forty-eight hours later, standards were established by plating H295R cells at densities ranging between 1000 and 35 000 cells per well, in triplicate. Cells were then incubated for 2 hours in a humidified incubator at 37°C. The CellTiter96 kit was used according to the manufacturer's protocol, and the absorbance of the wells at 490 nm was read at 2.5 hours using an ELISA plate reader. There was no significant difference in absorbance readings, between control and TXNRD2-knockdown cells, 0.43A ± 0.03 vs 0.42A ± 0.03 (mean ± SD, n = 6).

°C. The CellTiter96 kit was used according to the manufacturer's protocol, and the absorbance of the wells at 490 nm was read at 2.5 hours using an ELISA plate reader. There was no significant difference in absorbance readings, between control and TXNRD2-knockdown cells, 0.43A ± 0.03 vs 0.42A ± 0.03 (mean ± SD, n = 6). Reduced glutathione (GSH)/oxidized glutathione (GSSG) measurements Measurement of total glutathione (GSH + GSSG) or GSSG was performed using GSH/GSSG-Glo Assay (Promega), a luminescence-based system according to the manufacturer's instructions. Briefly, control and TXNRD2-KD cells were plated on a 96-well plate at a density of 20 000 cells/well, and duplicate samples were assayed for total glutathione or GSSG. GSH to GSSG ratios were calculated directly from Net RLU (relative light units) measurements using the equation GSH to GSSG ratio = [net total glutathione RLU-Net GSSG RLU]/[Net GSSG RLU/2]. Flow cytometry with MitoSOX Cells were grown to 50%–70% confluency and fresh media added before each experiment. MitoSOX Red (Invitrogen) was added to a final concentration of 5 μM according to the manufacturer's recommendation. After 20 minutes of loading of MitoSOX, TXNRD2-knockdown and control cells were trypsinized for 4 minutes and neutralized with media. The cells were washed with PBS (with Ca/Mg) and resuspended in fresh media in a sterile fluorescence-activated cell sorting tube.

of 5 μM according to the manufacturer's recommendation. After 20 minutes of loading of MitoSOX, TXNRD2-knockdown and control cells were trypsinized for 4 minutes and neutralized with media. The cells were washed with PBS (with Ca/Mg) and resuspended in fresh media in a sterile fluorescence-activated cell sorting tube. For the determination of mitochondrial superoxide by flow cytometry, measurements were carried out using an LSR Fortessa (BD Bioscience), data obtained were recorded and subsequently analyzed using DIVA version 6.2 (BD Bioscience,). Ten thousand gated events were recorded. The following steps were carried out for gating: cell debris as represented by distinct low forward and side scatter were gated out for analysis (P1), only singlet events were gated (P2), GFP-positive cells (cells incorporating either scrambled control or TXNRD2 knockdown shRNA) were selected (excited by 488 nm blue laser, band pass filter 530/30 nm) (P3), and finally MitoSOX Red was excited by 561 nm yellow/green laser with a band pass filter of 670/30 nm (P4) (Supplemental Figure 1). For quantitative analysis the frequency of events in P4 was multiplied by the median fluorescence intensity (MFI) within P4 to give an integrated MFI (iMFI) (6), reflecting the total functional response of this population of cells to MitoSOX). A Student's t test was used for statistical analysis.

upplemental Figure 1). For quantitative analysis the frequency of events in P4 was multiplied by the median fluorescence intensity (MFI) within P4 to give an integrated MFI (iMFI) (6), reflecting the total functional response of this population of cells to MitoSOX). A Student's t test was used for statistical analysis. Results Case reports Affected individuals within the kindred exhibited a wide spectrum of severity in symptoms at diagnosis with late onset in several family members (Table 1). The index case (patient 1.1), whose parents were first cousins, was diagnosed with isolated glucocorticoid deficiency at the age of 10.8 years after hyperpigmentation during febrile illnesses. Biochemical results revealed ACTH resistance: raised 9:00 am plasma ACTH (160 ng/L, normal < 50) and low 9:00 am serum cortisol levels (<10 nmol/L, normal range 200–600). Her sister (patient 1.2) was subsequently diagnosed, aged 4.5 years, with a 2-year preceding history of hyperpigmentation. Table 1. Clinical Details of the Members of the Kindred

lele frequency of 30%, 17.5%, and 25%, respectively, in keeping with their frequencies in the European American cohort from GO-ESP. This would fit with the fact that most of our FGD patients are of European extraction. No other variants were discovered in 100 FGD alleles, making the TXNRD2 mutation a rare cause of FGD. TXNRD2 is a dimeric NADPH-dependent flavin adenine dinucleotide-containing enzyme that catalyzes the reduction of the active disulphide of thioredoxin 2 and other substrates (7). As a selenoprotein, it requires insertion of a selenocysteine residue for catalytic activity. The mutation was predicted to cause a protein truncation prior to the selenocysteine active site. However, Western blotting of patient lysates revealed absence of the protein in homozygous patients in comparison with a heterozygote carrier and control (Figure 1D). In contrast, patients express cytoplasmic TXNRD1 normally (Figure 1D). Absence of TXNRD2 cDNA on RT-PCR was consistent with nonsense-mediated decay (Figure 1E), with direct sequencing of the amplicon from a heterozygote carrier revealing amplification of the wild-type sequence alone.

Results Case reports Affected individuals within the kindred exhibited a wide spectrum of severity in symptoms at diagnosis with late onset in several family members (Table 1). The index case (patient 1.1), whose parents were first cousins, was diagnosed with isolated glucocorticoid deficiency at the age of 10.8 years after hyperpigmentation during febrile illnesses. Biochemical results revealed ACTH resistance: raised 9:00 am plasma ACTH (160 ng/L, normal < 50) and low 9:00 am serum cortisol levels (<10 nmol/L, normal range 200–600). Her sister (patient 1.2) was subsequently diagnosed, aged 4.5 years, with a 2-year preceding history of hyperpigmentation. Table 1. Clinical Details of the Members of the Kindred Patient Sex Age, y Age at Diagnosis, y Mode of Presentation Relevant Clinical History Degree of Pigmentation at Presentation 9:00 am Cortisol, nmol/L 9:00 am ACTH, ng/L Maximum Cortisol With ACTH Stimulation, nmol/L Echo 1.1 F 33.8 10.8 Hyperpigmentation Asymptomatic until diagnosis Moderate <10 160 Not done Normal 1.2 F 27.1 4.5 Hyperpigmentation Asymptomatic until diagnosis Severe <25 500 <25 Trivial TR and MR 2.1 M 13.9 2.9 Screening Mild neonatal jaundice None 65 8130 61 Normal 2.2 F 9.5 6.9 Screening Asymptomatic Mild 158 514 33 Normal 2.3 F 8.6 0.1 Screening Asymptomatic Severe 28 3249 147 Normal 2.4 F 7.4 Currently well Asymptomatic None 262a 23.2a 1052a Normal 2.5 M 2.1 0.1 Poor feeding Heart failure secondary to cardiac defect Mild 46 >1240 190 Truncus arteriosus and VSD Abbreviations: Echo, echocardiogram; F, female; M, male; MR, mitral regurgitation; TR, tricuspid regurgitation; VSD, ventricular septal defect. The 9:00 am cortisol normal range is 200–600 nmol/L, 9:00 am ACTH normal range is less than 50 ng/L. The maximum cortisol with ACTH stimulation normal is greater than 550 nmol/L. Biochemical data are at time of diagnosis.

e; M, male; MR, mitral regurgitation; TR, tricuspid regurgitation; VSD, ventricular septal defect. The 9:00 am cortisol normal range is 200–600 nmol/L, 9:00 am ACTH normal range is less than 50 ng/L. The maximum cortisol with ACTH stimulation normal is greater than 550 nmol/L. Biochemical data are at time of diagnosis. a Most recent levels.

e; M, male; MR, mitral regurgitation; TR, tricuspid regurgitation; VSD, ventricular septal defect. The 9:00 am cortisol normal range is 200–600 nmol/L, 9:00 am ACTH normal range is less than 50 ng/L. The maximum cortisol with ACTH stimulation normal is greater than 550 nmol/L. Biochemical data are at time of diagnosis. a Most recent levels. The children of the index case (patients 2.1–2.3) were screened from birth and diagnosed with glucocorticoid deficiency between the ages of 0.1 and 6.9 years. Patient 2.5 presented at 0.1 year with cardiac failure secondary to congenital truncus arteriosus and a ventricular septal defect. During his admission he was diagnosed with isolated glucocorticoid deficiency. He is the only affected member of the kindred known to have comorbidity. Echocardiograms and electrocardiograms were normal in all individuals homozygous for the mutation except patient 1.2, who has trivial tricuspid and mitral valve regurgitation. All clinically affected individuals demonstrated a poor cortisol response to ACTH stimulation [125 μg tetracosactide (Synacthen) im] requiring standard glucocorticoid replacement therapy. All have normal mineralocorticoid production. Patient 2.4, homozygous for the TXNRD2 mutation, is currently clinically well aged 7.4 years with normal biochemistry and is under close clinical surveillance. Interestingly, she has had raised ACTH levels in early infancy (9:00 am ACTH of 124 ng/L at 0.02 y of age; corresponding cortisol of 305 nmol/L), which subsequently normalized. Her older sister, individual 2.2, similarly had raised ACTH levels in infancy (9:00 am ACTH 171 ng/L at 0.98 y of age; corresponding cortisol of 448 nmol/L), which normalized, although she was later diagnosed with FGD at 6.9 years.

at 0.02 y of age; corresponding cortisol of 305 nmol/L), which subsequently normalized. Her older sister, individual 2.2, similarly had raised ACTH levels in infancy (9:00 am ACTH 171 ng/L at 0.98 y of age; corresponding cortisol of 448 nmol/L), which normalized, although she was later diagnosed with FGD at 6.9 years. Affected individuals were mutation negative for the known genetic causes of FGD. Whole-exome sequencing of three affected individuals (denoted by the asterisks in the pedigree, Figure 1A) identified four rare homozygous variants within coding sequences, common to all three individuals, after application of our filtration strategy (see Materials and Methods). Three variants in two candidate genes did not segregate with the disease. All three variants are in the Single Nucleotide Polymorphism Database (dbSNP); the OR2T35 variant (rs370874670) and the two MUC4 variants (rs374495657 and rs202060675), but there is no frequency or population data. The variants were discounted on the basis that disease affected patient 2.2 and was wild type for rs370874670 in OR2T35, and the unaffected husband of patient 1.1 (individual 1.3) was homozygous for both changes in MUC4. Only one variant, a stop gain mutation (c.1341T>G; p.Y447X) within exon 15 of TXNRD2 (RefSeq accession number NM_006440.3), encoding mitochondrial TXNRD2, segregated with the disease in this kindred (Figure 1, A–C). Individuals heterozygous for the change were clinically unaffected.

homozygous for both changes in MUC4. Only one variant, a stop gain mutation (c.1341T>G; p.Y447X) within exon 15 of TXNRD2 (RefSeq accession number NM_006440.3), encoding mitochondrial TXNRD2, segregated with the disease in this kindred (Figure 1, A–C). Individuals heterozygous for the change were clinically unaffected. Two SNPs at this position are annotated in dbSNP (identification rs202059967): an A>G change and the A>C change reported here. The A>G change is silent and seen in only 1 of 12 686 alleles in National Heart, Lung, and Blood Institute Gene Ontology Exome Sequencing Project, ESP6500 (see web site addresses); the A>C change is not recorded in this database but is listed in dbSNP submitted by EXOME_CHIP (submitter SNP identification ss491568437), with no frequency data. Sequencing of more than 1000 healthy adult British Pakistanis revealed a minor allele frequency of 1.04% for this variant, and the genotypes were A/A = 1080; A/C = 23; C/C = 0. Importantly, in this control population, the variant was never seen in homozygosity. One other stop gain mutation (R441*; rs200162480) is listed for TXNRD2 but is present in heterozygosity in only one individual (1 of 12 730 alleles). Sequencing of 50 FGD patients with unknown etiology identified variants at rs5748469 (p.A66S), rs5992495 (p.S299R), and rs1139793 (p.I370T) with minor allele frequency of 30%, 17.5%, and 25%, respectively, in keeping with their frequencies in the European American cohort from GO-ESP. This would fit with the fact that most of our FGD patients are of European extraction. No other variants were discovered in 100 FGD alleles, making the TXNRD2 mutation a rare cause of FGD.

te carrier and control (Figure 1D). In contrast, patients express cytoplasmic TXNRD1 normally (Figure 1D). Absence of TXNRD2 cDNA on RT-PCR was consistent with nonsense-mediated decay (Figure 1E), with direct sequencing of the amplicon from a heterozygote carrier revealing amplification of the wild-type sequence alone. The glutathione and thioredoxin systems maintain reduced peroxiredoxin 3 (PRDX3) (Figure 2), which is integral for redox regulation within the adrenal (8). NNT provides the thioredoxin and glutathione systems with high concentrations of NADPH required for this process. We found that although TXNRD2 mRNA is ubiquitously expressed in the human tissues tested, the highest levels are observed in the adrenal cortex (Figure 3A), consistent with mouse expression profiles (9). Figure 2. The thioredoxin and glutathione systems maintain mitochondrial redox homeostasis. TXNRD2 reduces TXN2 and GLRX2, both of which can reduce PRDX3, which in turn detoxifies H2O2 in mitochondria. Glutathione reductase (GSR) and GSH contribute to the process through the reduction of GLRX2. NNT provides the high concentrations of NADPH required by both the thioredoxin and glutathione systems.

homeostasis. TXNRD2 reduces TXN2 and GLRX2, both of which can reduce PRDX3, which in turn detoxifies H2O2 in mitochondria. Glutathione reductase (GSR) and GSH contribute to the process through the reduction of GLRX2. NNT provides the high concentrations of NADPH required by both the thioredoxin and glutathione systems. Figure 3. Knockdown of TXNRD2 impairs mitochondrial redox homeostasis in a human adrenocortical cell line. A, TXNRD2 is widely expressed in human tissues with highest the TXNRD2 mRNA levels in the adrenal cortex. B, Stable TXNRD2 knockdown (TXNRD2-KD) in H295R cells, confirmed by Western blot analysis with 95% knockdown of protein levels (n = 4). C, Increased pressure on the glutathione system is observed with a significant decrease in the reduced:oxidized glutathione (GSH:GSSG) ratio. Oxidative stress induced by 25 μM menadione in the control cells reduced this ratio to 0.4 ± 0.1 (n = 4). D, TXNRD2-KD leads to a decrease in the reduced to oxidized PRDX3 ratio [Western blot with densitometric analysis (n = 3)]. E, Quantitative analysis of superoxide production by MitoSOX, after fluorescence-activated cell sorting (FACS) shows a significant increase in superoxide production in KD cells relative to controls (n = 3). Error bars represent SD. *, P < .05.

to oxidized PRDX3 ratio [Western blot with densitometric analysis (n = 3)]. E, Quantitative analysis of superoxide production by MitoSOX, after fluorescence-activated cell sorting (FACS) shows a significant increase in superoxide production in KD cells relative to controls (n = 3). Error bars represent SD. *, P < .05. TXNRD2 knockdown in the H295R human adrenocortical cell line by shRNA (Figure 3B) had no effect on cell viability. However, a clear impact on mitochondrial redox homeostasis was demonstrated, with increased pressure on the glutathione system observed as a decrease in the GSH to GSSG ratio (10.3 ± 1.3 vs 7.7 ± 0.4 in scrambled vs knockdown cells, respectively; P = .02) (Figure 3C). The ability to maintain mitochondrial PRDX3 in its reduced form is impaired, with a significant decrease in the ratio of the monomeric reduced to oxidized dimeric form in the TXNRD2-deficient cells compared with controls (0.19 ± 0.06 vs 0.48 ± 0.09, respectively; P = .02; Figure 3D). Finally, as a consequence of TXNRD2 knockdown in the adrenocortical cells, an approximately 3-fold increase in levels of mitochondrial reactive oxygen species are seen, further demonstrating an impairment of redox regulation (Figure 3E and Supplemental Figure 1).

0.06 vs 0.48 ± 0.09, respectively; P = .02; Figure 3D). Finally, as a consequence of TXNRD2 knockdown in the adrenocortical cells, an approximately 3-fold increase in levels of mitochondrial reactive oxygen species are seen, further demonstrating an impairment of redox regulation (Figure 3E and Supplemental Figure 1). Discussion Absence of TXNRD2 is associated with adrenal insufficiency in this consanguineous kindred, with all affected members being homozygous for the identified genetic defect. TXNRD2 is 1 of 25 human selenoproteins, which require insertion of a highly reactive selenocysteine residue for enzymatic activity (10). Several selenoproteins, including the thioredoxin reductases and glutathione peroxidases, contribute significantly to redox regulation. TXNRD2, one of three thioredoxin reductases, is mitochondria specific and exists as an antiparallel homodimer (7). The N- and C-terminal redox active centers of the two subunits functionally interact and transfer electrons from NADPH/H+ to thioredoxin 2 (TXN2) and other substrates (7). Within the mitochondria the thioredoxin and glutathione systems, reliant on the provision of NADPH/H+ by NNT, contribute to the maintenance of redox homeostasis.

terminal redox active centers of the two subunits functionally interact and transfer electrons from NADPH/H+ to thioredoxin 2 (TXN2) and other substrates (7). Within the mitochondria the thioredoxin and glutathione systems, reliant on the provision of NADPH/H+ by NNT, contribute to the maintenance of redox homeostasis. Particularly high TXNRD2 mRNA levels were noted in the adrenal cortex compared with the other human tissues investigated, suggesting a critical role in this tissue. PRDX3 is reported to be the most important H2O2-eliminating enzyme in the mitochondria of the adrenal cortex with hyperoxidation of PRDX3 resulting in diminished steroidogenesis (8). During H2O2 elimination, two reduced PRDX3 subunits are converted to an oxidized disulphide-linked dimer that is reduced again by TXN2 (8). Thus, PRDX3, together with mitochondrial-specific TXN2 and TXNRD2, provide a primary line of defense against H2O2 produced by the mitochondrial respiratory chain in the adrenal gland. Additionally, glutaredoxin 2 has recently been identified as another electron donor for PRDX3 (11). Glutaredoxin 2 (GLRX2) itself is reduced by TXNRD2 as well as GSH, and the mitochondrial thioredoxin and GSH systems function in parallel to protect against oxidative stress (11, 12) (Figure 2). In our in vitro knockdown adrenocortical model, we demonstrate that the glutathione system is unable to fully compensate for the TXNRD2 deficiency leading to increased mitochondrial superoxide production.

chondrial thioredoxin and GSH systems function in parallel to protect against oxidative stress (11, 12) (Figure 2). In our in vitro knockdown adrenocortical model, we demonstrate that the glutathione system is unable to fully compensate for the TXNRD2 deficiency leading to increased mitochondrial superoxide production. Oxidative stress impedes steroidogenesis, and steroidogenesis itself induces oxidative stress as a result of electron leak throughout the steroidogenic pathway (13, 14). The final step of cortisol production, catalyzed by CYP11B1 within the mitochondria, accounts for approximately 40% of the total electron flow from NAPDH directed at reactive oxygen species production during steroidogenesis (14). This, together with the higher production of cortisol in comparison with aldosterone, may explain the particular susceptibility of the zona fasciculata to oxidative stress, and hence, individuals with TXNRD2 and NNT mutations primarily develop glucocorticoid deficiency. Oxidative stress has been implicated in other causes of adrenal insufficiency, including triple A syndrome, in which failure of nuclear import of DNA repair proteins and ferritin heavy chain are described (15, 16) and adrenoleukodystrophy, which is secondary to the accumulation of very-long chain fatty acids in the adrenal cortex and other tissues (17). The wide variability in age at diagnosis (0.1–10.8 y) within this family may suggest the existence of disease modifiers, which could be genetic, environmental, or both. Individual 2.4 currently shows no sign of adrenal disease, which may be due to incomplete penetrance or variable expressivity of the mutation. Because she is only 7.4 years old and because she has manifested abnormal ACTH readings in the past, similar to the clinical picture for her older sister (individual 2.2), she may go on to develop the disease.

o sign of adrenal disease, which may be due to incomplete penetrance or variable expressivity of the mutation. Because she is only 7.4 years old and because she has manifested abnormal ACTH readings in the past, similar to the clinical picture for her older sister (individual 2.2), she may go on to develop the disease. Because TXNRD2 is ubiquitously expressed in humans, individuals with this genetic defect are potentially at risk of developing extraadrenal manifestations. Txnrd2 deletion in mice is embryonically lethal at day 13 as a consequence of a combination of cardiac and hematopoietic defects, with cardiac-specific ablation resulting in a fatal dilated cardiomyopathy (18). Thus, in mice the TXNRD2 system is clearly indispensable for normal cardiac development and function. In humans, two novel heterozygous mutations in TXNRD2 were identified in 3 of 227 patients with a diagnosis of dilated cardiomyopathy (19). Interestingly, neither the heterozygote nor homozygote individual, with absence of TXNRD2, showed any evidence of cardiomyopathy or conduction disease. One affected family member presented with truncus arteriosus, an extremely rare congenital cardiac anomaly. Approximately 40% of truncus arteriosus cases are associated with Di George syndrome (DGS), secondary to haploinsufficiency of a region of varying length on chromosome 22q11.2 (20). Several genes within this region have been linked to defects in cardiac development including TBX1 and Crkl, essential for the survival, proliferation, and migration of neural crest cells (21, 22). Because TXNRD2 falls within this region on chromosome 22, this raises the possibility that TXNRD2 contributes to the cardiac phenotype of DGS. Because heterozygote carriers of the p.Y447X TXNRD2 mutation have normal adrenal function, we would predict that haploinsufficiency of TXNRD2 would not lead to an abnormal adrenal phenotype, and consistent with this, we were unable to identify any published reports of adrenal insufficiency in DGS.

iac phenotype of DGS. Because heterozygote carriers of the p.Y447X TXNRD2 mutation have normal adrenal function, we would predict that haploinsufficiency of TXNRD2 would not lead to an abnormal adrenal phenotype, and consistent with this, we were unable to identify any published reports of adrenal insufficiency in DGS. We report the first mutation in TXNRD2 associated with an adrenal phenotype in humans, signifying the importance of the thioredoxin system in maintaining redox homeostasis in the adrenocortical environment. The TXNRD2 mutation in this family and the other recently described FGD syndromes due to mutations in NNT and MCM4 highlight important pathogenic pathways in addition to defective ACTH signaling, causing glucocorticoid deficiency. A delicate balance of mitochondrial redox regulation controls steroidogenesis in the human adrenal gland (8, 23, 24), and therefore, other components of this complex network of mitochondrial antioxidants make good candidates for undiagnosed causes of adrenal insufficiency. Abbreviations: DGSDi George syndrome FGDfamilial glucocorticoid deficiency GAPDHglyceraldehyde-3-phosphate dehydrogenase GFPgreen fluorescence protein GLRX2glutaredoxin 2 GSHreduced glutathione GSSGoxidized glutathione MCM4minichromosome maintenance 4 MC2Rmelanocortin 2 receptor MRAPMC2R accessory protein NADPHreduced nicotinamide adenine dinucleotide phosphate NNTnicotinamide nucleotide transhydrogenase PRDX3peroxiredoxin 3 shRNAshort hairpin RNA SNPsingle-nucleotide polymorphism TXN2thioredoxin 2 TXNRD2thioredoxin reductase 2.

GSHreduced glutathione GSSGoxidized glutathione MCM4minichromosome maintenance 4 MC2Rmelanocortin 2 receptor MRAPMC2R accessory protein NADPHreduced nicotinamide adenine dinucleotide phosphate NNTnicotinamide nucleotide transhydrogenase PRDX3peroxiredoxin 3 shRNAshort hairpin RNA SNPsingle-nucleotide polymorphism TXN2thioredoxin 2 TXNRD2thioredoxin reductase 2. Acknowledgments We thank the patients and their families. We also thank R. Banerjee (Luton and Dunstable University Hospital) for the clinical information and G. Rosignoli and W. Day (Flow Cytometry Core Facility, William Harvey Research Institute) for their assistance in obtaining and analyzing the flow cytometry data. In addition, we thank Eamonn Maher, Richard Durbin, Shane McCarthy, Richard Trembath, and David van Heel for access to the exome sequence data on subjects supported by the Wellcome Trust Strategic Award WT102627 and sequenced at the Wellcome Trust Sanger Institute (Wellcome Trust Award WT098051). Web resources include the following: dbSNP (www.ncbi.nlm.nih.gov/snp/,); DNAnexus Classic (https://dnanexus.com/); and Exome Variant Server, National Heart, Lung, and Blood Institute Exome Sequencing Project, Seattle, Washington (http://evs.gs.washington.edu/evs/) (accessed October 2012). This work was supported by the Wellcome Trust (Clinical Research Training Fellowship Grant WT095984AIA, to R.P.) and the Medical Research Council United Kingdom (Project Grant MR/K020455/1, to L.A.M.; Medical Research Council/Academy of Medical Sciences Clinician Scientist Fellowship Grant G0802796, to L.F.C.).

Web resources include the following: dbSNP (www.ncbi.nlm.nih.gov/snp/,); DNAnexus Classic (https://dnanexus.com/); and Exome Variant Server, National Heart, Lung, and Blood Institute Exome Sequencing Project, Seattle, Washington (http://evs.gs.washington.edu/evs/) (accessed October 2012). This work was supported by the Wellcome Trust (Clinical Research Training Fellowship Grant WT095984AIA, to R.P.) and the Medical Research Council United Kingdom (Project Grant MR/K020455/1, to L.A.M.; Medical Research Council/Academy of Medical Sciences Clinician Scientist Fellowship Grant G0802796, to L.F.C.). Disclosure Summary: The authors have nothing to disclose.

Tumor-induced osteomalacia (TIO), one of the causes of hypophosphatemia, is commonly associated with benign mesenchymal tumors of the soft tissue and skeleton (1). Clinical characteristics include renal phosphate wasting, low or normal serum 1,25-dihydroxyvitamin D levels, bone pain, and elevated alkaline phosphatase levels (1). Fibroblast growth factor (FGF) 23, a phosphatonin secreted by these tumors, is responsible for the pathogenesis of TIO (1). Other phosphatonins such as matrix extracellular phosphoglycoprotein, secreted frizzled related protein-4, and FGF7 were also identified as contributing to the pathogenesis of TIO (2). TIO is also associated with malignancies such as prostate cancer, oat cell cancer, hematological malignancies, and colon cancer. In these cases, the primary disease is usually obvious, and treatment is focused on the underlying disease (3, 4). In this study, we report a case of ovarian cancer-related hypophosphatemic osteomalacia, which has not been previously reported in the literature to our knowledge.

tological malignancies, and colon cancer. In these cases, the primary disease is usually obvious, and treatment is focused on the underlying disease (3, 4). In this study, we report a case of ovarian cancer-related hypophosphatemic osteomalacia, which has not been previously reported in the literature to our knowledge. Patient and Methods Case description The 57-year-old woman examined in this study was otherwise healthy before presentation. The patient's menstruation was normal before the onset of menopause at age 52. She had experienced low back pain for 6 months before she visited the outpatient clinic at National Taiwan University Hospital. She also complained of night sweats and a weight loss of 14 kg during the previous 6 months. The low back pain developed while lying down and radiated to both lower limbs. The patient had no abdominal pain, diarrhea, or abnormal vaginal discharge. Methods This study was approved by the Institutional Review Board of the National Taiwan University Hospital (protocol no. 201105045RC) and is registered on Clinicaltrials.gov (protocol no. NCT01660308).

Patient and Methods Case description The 57-year-old woman examined in this study was otherwise healthy before presentation. The patient's menstruation was normal before the onset of menopause at age 52. She had experienced low back pain for 6 months before she visited the outpatient clinic at National Taiwan University Hospital. She also complained of night sweats and a weight loss of 14 kg during the previous 6 months. The low back pain developed while lying down and radiated to both lower limbs. The patient had no abdominal pain, diarrhea, or abnormal vaginal discharge. Methods This study was approved by the Institutional Review Board of the National Taiwan University Hospital (protocol no. 201105045RC) and is registered on Clinicaltrials.gov (protocol no. NCT01660308). Clinical, biochemical, and radiological assessments were undertaken. The patient's serum phosphate and FGF23 levels were evaluated at baseline and after treatment for ovarian cancer. FGF23 levels were measured using ELISA (Kainos Laboratories, Inc), according to manufacturer's instructions. Two specific murine monoclonal antibodies were bound to the full length of FGF23. One antibody was conjugated to horseradish peroxidase to allow for detection by a spectrophotometric reader. The other antibody was immobilized onto the microtiter well for capture. The normal range for serum FGF23 is 8.2–54.3 pg/mL (5).

Two specific murine monoclonal antibodies were bound to the full length of FGF23. One antibody was conjugated to horseradish peroxidase to allow for detection by a spectrophotometric reader. The other antibody was immobilized onto the microtiter well for capture. The normal range for serum FGF23 is 8.2–54.3 pg/mL (5). Results Physical examination The patient's height was 155 cm, and her weight was 40 kg (body mass index, 16.6 kg/m2). Her conjunctivae were pale. Two firm mass lesions, approximately 2 cm in diameter, were located in the parietal area on both sides. A thyroid nodule, approximately 1 cm in diameter, was noted on palpation. There was no abdominal tenderness or rebound tenderness. Neurological examination revealed normal muscle power and deep tendon reflex of the four limbs.

ass lesions, approximately 2 cm in diameter, were located in the parietal area on both sides. A thyroid nodule, approximately 1 cm in diameter, was noted on palpation. There was no abdominal tenderness or rebound tenderness. Neurological examination revealed normal muscle power and deep tendon reflex of the four limbs. Biochemical and imaging studies Spine radiography revealed a relatively radiolucent texture, scoliosis, and facet arthroses at the lower lumbar spine. With regard to bone mineral density, the lumbar spine had a T-score of −2.42 SD. Magnetic resonance imaging showed multiple spinal metastasis and stenosis over the lumbar spine. Radiography of the skull and extremities showed no obvious osteolytic or osteoblastic lesions. For evaluation of the abnormal bone mineral density and bony lesions, biochemical studies were performed and showed hypophosphatemia (1.6 mg/dL; normal range, 2.7–4.5 mg/dL), normocalcemia (2.2 mmol/L; normal range, 2.02–2.60 mmol/L), an elevated alkaline phosphatase level (597 U/L; normal range, 60–220 U/L), and a normal intact PTH level (17.1 pg/mL; normal range, 16–87 pg/mL). Renal function, serum uric acid level, and liver function were within normal limits (creatinine, 0.7 mg/dL; uric acid, 5.5 mg/dL; aspartate aminotransferase, 26 U/L; alanine aminotransferase, 29 U/L). Hemogram showed elevated white blood cell count (12.66 × 103/μL; normal range, 4–10 × 103/μL) and anemia (red blood cell count, 3.92 × 106/μL; normal range, 3.5–4.5 × 106/μL; hemoglobin, 8.4 g/dL; normal range, 12–15 g/dL). The phosphorus tubular reabsorption rate was lower than normal (75%; normal range, 80–90%). Hyperphosphaturic hypophosphatemia was pronounced, and TIO was highly suspected.

e, 4–10 × 103/μL) and anemia (red blood cell count, 3.92 × 106/μL; normal range, 3.5–4.5 × 106/μL; hemoglobin, 8.4 g/dL; normal range, 12–15 g/dL). The phosphorus tubular reabsorption rate was lower than normal (75%; normal range, 80–90%). Hyperphosphaturic hypophosphatemia was pronounced, and TIO was highly suspected. The tumor survey revealed a markedly elevated cancer antigen 125 level (598 U/mL; normal range, below 35 U/mL). Abdominal ultrasonography showed multiple tumors in the liver and left adnexa. Computed tomography (CT) showed multiple tumors in the thyroid, bone, lung, and liver, and a large tumor over the uterus and left adnexa (Figure 1). Fine-needle aspiration cytology of the thyroid nodule indicated a poorly differentiated carcinoma. CT-guided biopsy of the hepatic tumor revealed an undifferentiated carcinoma (Figure 2) with strongly positive immunohistochemical staining for cytokeratin 7, p16, Wilms' tumor 1, and β-catenin and negative staining for leukocyte common antigen. Under the diagnosis of stage IV ovarian carcinoma, neoadjuvant chemotherapy with carboplatin and paclitaxel was initiated, followed by debulking surgery including total hysterectomy, bilateral salpingo-oophorectomy, infracolic omentectomy, appendectomy, and pelvic lymph node dissection and adjuvant chemotherapy. The pathological diagnosis was also high-grade serous ovarian carcinoma. The serum FGF23 level decreased gradually after chemotherapy and debulking surgery (initial, 501.6 pg/mL; after neoadjuvant chemotherapy, 140.6 pg/mL; after surgery, 44.141 pg/mL; normal range, 8.2–54.3 pg/mL). The patient's requirement of a phosphate supplement decreased gradually until it was no longer needed after the treatment of ovarian cancer. The treatment course and the changes of laboratory data were presented in Figure 3.

adjuvant chemotherapy, 140.6 pg/mL; after surgery, 44.141 pg/mL; normal range, 8.2–54.3 pg/mL). The patient's requirement of a phosphate supplement decreased gradually until it was no longer needed after the treatment of ovarian cancer. The treatment course and the changes of laboratory data were presented in Figure 3. Figure 1. CT scan showing multilobulated cystic lesions over the left adnexa. Figure 2. Pathological specimen of the hepatic tumor showing an undifferentiated carcinoma with high cellularity and an infiltrative pattern (magnification, ×400; hematoxylin and eosin stain). Figure 3. The treatment course and laboratory data. P, Serum phosphate level; CA-125, cancer antigen 125.

Figure 1. CT scan showing multilobulated cystic lesions over the left adnexa. Figure 2. Pathological specimen of the hepatic tumor showing an undifferentiated carcinoma with high cellularity and an infiltrative pattern (magnification, ×400; hematoxylin and eosin stain). Figure 3. The treatment course and laboratory data. P, Serum phosphate level; CA-125, cancer antigen 125. Discussion We reported a case of high-grade ovarian carcinoma associated with an elevated serum FGF23 level and hyperphosphaturic hypophosphatemic osteomalacia, which has not been previously reported in the literature to our knowledge. TIO is a rare, paraneoplastic syndrome caused primarily by benign mesenchymal tumors. TIO has been associated with malignancies in rare cases (3). Phosphatonins, a group of endocrine hormones, decrease renal tubular reabsorption of phosphate and inhibit 1α-hydroxylation of vitamin D (2). Consequently, muscle weakness, osteomalacia, and even bone fracture may occur (2). The occult course often delays the diagnosis of TIO by an average of 5 years (6). FGF23 helps to diagnose TIO (6). Detailed physical examination, CT, magnetic resonance imaging, octreotide scintigraphy, and positron emission tomography may help to identify tumors (7). Venous sampling of FGF23 has been used to confirm the causative tumor preoperatively (8). Surgical resection of the tumor is the definitive treatment method. Serum FGF23 and phosphate levels typically return to normal within 5 days of the operation (3, 9). After searching the literature, we found neither a case of ovarian cancer-related osteomalacia nor detailed phosphate and FGF23 data of patients with ovarian cancer to draw a figure comparing serum phosphate and FGF23 levels purely in patients with ovarian cancer. Therefore, we compared serum FGF23 and phosphate levels between the present case and 19 other reported cases from three case series (1, 9, 10), in which the location of the responsible tumors included lower and upper extremities, the head, and the spine (Figure 4). We found that FGF23 levels were widely diverse and serum phosphate levels cannot predict FGF23 levels in TIO patients. This is compatible with the conclusion of another study that there were no correlations between FGF23 and the severity of X-link hypophosphatemia, including serum phosphate levels (11).

ine (Figure 4). We found that FGF23 levels were widely diverse and serum phosphate levels cannot predict FGF23 levels in TIO patients. This is compatible with the conclusion of another study that there were no correlations between FGF23 and the severity of X-link hypophosphatemia, including serum phosphate levels (11). Figure 4. Comparison of serum FGF23 and phosphorus (P) levels between the present case and 19 TIO cases reported in three case series (1, 9, 10). (The arrow points to the present case.) FGF23 levels > 2000 pg/mL are presented at the level of 2000, with the exact data shown above.

ine (Figure 4). We found that FGF23 levels were widely diverse and serum phosphate levels cannot predict FGF23 levels in TIO patients. This is compatible with the conclusion of another study that there were no correlations between FGF23 and the severity of X-link hypophosphatemia, including serum phosphate levels (11). Figure 4. Comparison of serum FGF23 and phosphorus (P) levels between the present case and 19 TIO cases reported in three case series (1, 9, 10). (The arrow points to the present case.) FGF23 levels > 2000 pg/mL are presented at the level of 2000, with the exact data shown above. Ovarian cancer is the leading cause of death due to gynecological malignancies (12). The prognosis is poor, with a 5-year survival rate of approximately 45.6% (13). The most common extra-abdominal metastasis is the pleural space. Other distant metastatic sites include liver, lung, pericardium, bone, and brain. In our patient, the metastatic site involved the thyroid gland, which has been rather uncommon and only occurred in 3–15% of the patients (14, 15). Angiogenesis is responsible for tumor spread and metastasis (16). Several growth factors have been identified to play key roles in driving angiogenesis, such as vascular endothelial growth factor, platelet-derived growth factor, FGF, and the angiopoietin/Tie2 receptor complex (16). FGF is expressed in epithelial ovarian cancer and has proangiogenic properties (13). It directly stimulates the proliferation and migration of endothelial cells, sensitizes epithelial cells to other angiogenic factors, facilitates tube formation, and stimulates the secretion of extracellular matrix remodeling proteases (17, 18). FGF is also secreted into malignant ascites together with vascular endothelial growth factor, which further contributes to ovarian cancer progression and angiogenesis (16). The FGF signaling pathway involves MAPK proteins, proteins of the phosphatidylinositol-3-kinase/Akt cascade, and proteins of the phospholipase-C and inositol triphosphate cascades and may cross talk with other pathways such as the Notch pathway (19, 20). There are 23 FGF isoforms and five receptor molecules identified so far (16). Several isoforms of FGF have been extensively studied in ovarian cancers. For example, serum FGF2 levels were higher in patients with ovarian cancer than in people with benign ovarian tumors or normal ovaries (21); the growth of ovarian cancer was regulated by FGF8 (22); overexpression of FGF18 was an independent predictive marker for poor outcomes in patients with high-grade serous ovarian cancer (23).

serum FGF2 levels were higher in patients with ovarian cancer than in people with benign ovarian tumors or normal ovaries (21); the growth of ovarian cancer was regulated by FGF8 (22); overexpression of FGF18 was an independent predictive marker for poor outcomes in patients with high-grade serous ovarian cancer (23). In a study of 13 women with advanced-stage ovarian cancer, 14 with early-stage ovarian cancer, 14 with benign ovarian tumors, and 39 healthy women, serum FGF23 levels were higher in patients with advanced-stage ovarian cancer, without decreased serum phosphate levels (24). Approximately half of the patients with advanced-stage ovarian cancer had elevated serum FGF23 levels before treatment. This suggests that elevated FGF23 levels in patients with ovarian tumors indicate advanced-stage disease (24). Many women still require surgery to differentiate benign ovarian tumor from cancer because current biochemical markers and imaging techniques are inadequate. Serum FGF23 levels may be a potential candidate for predicting the presence of malignant disease or clinical outcomes, and this is worthy of further study (24). As to the absence of hypophosphatemia in patients with advanced-stage ovarian cancer and high FGF23 in the study mentioned above, one possible explanation is that perhaps other factors are necessary to induce phosphate wasting, such as the overexpression of secreted frizzled related protein-4 and matrix extracellular phosphoglycoprotein (24). Besides, less than five patients with advanced-stage ovarian cancer had FGF23 levels above 500 pg/mL in the study mentioned above (24). The FGF23 level of the patient in our current study was 501.6 pg/mL. The severity of FGF23 level abnormality potentially could contribute to the development of hypophosphatemia to some extent. Other factors that may affect phosphate level are poor appetite and cachexia. Inadequate phosphate intake should also be taken into consideration as one of the reasons for hypophosphatemia in our patient. We reviewed the literature and found no data on the prevalence of hypophosphatemia in cases with ovarian cancer. Hypophosphatemia is associated with weakness, bone pain, and fracture, which may be underdiagnosed as purely cancer cachexia and bone metastasis. In such cases, medical therapy with phosphate supplementation and calcitriol or alfacalcidol may improve weakness and avoid fracture.

osphatemia in cases with ovarian cancer. Hypophosphatemia is associated with weakness, bone pain, and fracture, which may be underdiagnosed as purely cancer cachexia and bone metastasis. In such cases, medical therapy with phosphate supplementation and calcitriol or alfacalcidol may improve weakness and avoid fracture. It should be offered to these patients as in other TIO patients with tumor that cannot be localized or is not surgically resectable (3). Clinicians should have vigilance to measure serum phosphorus and/or FGF23 in patients with ovarian cancer, especially those with advanced-stage cancer or those with weakness, bone pain, and fracture. In conclusion, a subset of ovarian cancer cases may be associated with elevated FGF23 levels and should be taken into account during the differential diagnosis of TIO. In such cases, hypophosphatemia and bone pain can be relieved with the treatment of ovarian cancer. Besides, TIO should be considered in patients with ovarian cancer presenting with weakness, bone pain, and fractures. Investigation of TIO is appropriate when these patients present hypophosphatemia. Abbreviations: CTcomputed tomography FGFfibroblast growth factor TIOtumor-induced osteomalacia. Acknowledgments We thank the staff of the Eighth Core Lab in the Department of Medical Research of the National Taiwan University Hospital for technical support during the study. This work was supported by the Liver Disease Prevention and Treatment Research Foundation, Taiwan, the National Taiwan University College of Medicine, and the National Taiwan University Hospital.

Acknowledgments We thank the staff of the Eighth Core Lab in the Department of Medical Research of the National Taiwan University Hospital for technical support during the study. This work was supported by the Liver Disease Prevention and Treatment Research Foundation, Taiwan, the National Taiwan University College of Medicine, and the National Taiwan University Hospital. Clinical Trial Registration No. NCT01660308. Disclosure Summary: The authors have no conflict of interest.

Summary of Recommendations 1.1 The benefits of congenital hypothyroidism screening Early detection and treatment of congenital hypothyroidism (CH) through neonatal screening prevents neurodevelopmental disability and optimizes developmental outcomes (1|⊕⊕⊕). 1.2 Analytical methodology, effectiveness, and efficacy of CH screening strategies Screening for primary CH should be introduced worldwide. The initial priority of neonatal screening for CH should be the detection of all forms of primary CH: mild, moderate, and severe. The most sensitive test for detecting primary CH is TSH determination (1|⊕⊕⊕). 1.3 Screening in special categories of neonates at risk of CH A strategy of second screening should be considered for the following conditions: preterm neonates; low-birth weight (LBW) and very low-birth weight (VLBW) neonates; ill and preterm newborns admitted to neonatal intensive care units (NICU); specimen collection within the first 24 hours of life; and multiple births (particularly same-sex twins); (2|⊕⊕○). 2.1 Biochemical criteria used in the decision to initiate treatment If capillary TSH concentration from blood obtained on neonatal screening is ≥ 40 mU/L whole blood, we recommend starting treatment as soon as a good venous sample can be obtained, without waiting for the venous blood test result, unless venous thyroid function test (TFT) results are available on the same day (1|⊕⊕○). If capillary TSH concentration is < 40 mU/l of whole blood, the clinician may wait for the results of venous TFT, provided that these results are available on the following day (1|⊕⊕○).

2.1 Biochemical criteria used in the decision to initiate treatment If capillary TSH concentration from blood obtained on neonatal screening is ≥ 40 mU/L whole blood, we recommend starting treatment as soon as a good venous sample can be obtained, without waiting for the venous blood test result, unless venous thyroid function test (TFT) results are available on the same day (1|⊕⊕○). If capillary TSH concentration is < 40 mU/l of whole blood, the clinician may wait for the results of venous TFT, provided that these results are available on the following day (1|⊕⊕○). 2.2 Communication of elevated TSH result The detection of a high TSH concentration on screening should be communicated by an experienced person (eg screening laboratory staff or pediatric endocrine team) either by telephone or in person (2|⊕○○). When the child reaches nursery or school age, educators and teachers need not be informed about the child having CH to avoid stigmatization due to “labeling” (2|⊕○○). 2.3 Decision to start treatment on the basis of venous TFTs If venous free T4 (FT4) concentration is below norms for age, treatment should be started immediately (1|⊕⊕⊕). If venous TSH concentration is > 20 mU/L, treatment should be started, even if FT4 concentration is normal (2|⊕⊕○).

When the child reaches nursery or school age, educators and teachers need not be informed about the child having CH to avoid stigmatization due to “labeling” (2|⊕○○). 2.3 Decision to start treatment on the basis of venous TFTs If venous free T4 (FT4) concentration is below norms for age, treatment should be started immediately (1|⊕⊕⊕). If venous TSH concentration is > 20 mU/L, treatment should be started, even if FT4 concentration is normal (2|⊕⊕○). If venous TSH concentration is ≥ 6 to 20 mU/l beyond 21 days in a well baby with a FT4 concentration within the limits for age, we suggest a) investigation, which should include diagnostic imaging, to try to obtain a definitive diagnosis; b) consideration, in discussion with the family, of either initiating thyroxine supplementation immediately and retesting, off treatment, at a later stage; or withholding treatment but retesting two weeks later (2|⊕⊕○). 2.4 Use of imaging in assessing the severity and cause of CH X-ray of the knee may be carried out to assess the severity of intrauterine hypothyroidism by the presence or absence of femoral and tibial epiphyses (2|⊕⊕⊕). The thyroid gland should be imaged using either radioisotope scanning (scintigraphy) with or without the perchlorate discharge test; or ultrasonography; or both (1|⊕⊕○). Imaging should never be allowed to delay the initiation of treatment (1|⊕⊕○). 2.5 Associated malformations and syndromes All neonates with high TSH concentrations should be examined carefully for congenital malformations (particularly cardiac) and for dysmorphic features (1|⊕⊕⊕).

The thyroid gland should be imaged using either radioisotope scanning (scintigraphy) with or without the perchlorate discharge test; or ultrasonography; or both (1|⊕⊕○). Imaging should never be allowed to delay the initiation of treatment (1|⊕⊕○). 2.5 Associated malformations and syndromes All neonates with high TSH concentrations should be examined carefully for congenital malformations (particularly cardiac) and for dysmorphic features (1|⊕⊕⊕). 3.1 Treatment and monitoring of CH L-T4 alone is recommended as the medication of choice for treating CH (1|⊕⊕○). L-T4 treatment should be initiated as soon as possible and no later than 2 weeks after birth or immediately after confirmatory serum test results in infants in whom CH is detected by a second routine screening test (1|⊕⊕○). An initial L-T4 dose of 10–15 μg/kg per day should be given (1|⊕⊕○). Infants with severe disease, as defined by a very low pretreatment TT4 or FT4 concentration, should be treated with the highest initial dose (1|⊕⊕○). L-T4 should be administered orally; if intravenous treatment is necessary the dose should be no more than 80% of the oral dose, The dose should then be adjusted according to TSH and FT4 determinations (1|⊕⊕○). L-T4 tablets should be crushed and administered via a small spoon, in a few milliliters of water or breast milk (1|⊕⊕○). Brand rather than generic L-T4 tablets should be used, particularly during infancy and in severe cases (2|⊕⊕○). L-T4 liquid should only be used if pharmaceutically produced (1|⊕⊕○). Parents should be provided with written instructions on L-T4 treatment (1|⊕○○).

L-T4 tablets should be crushed and administered via a small spoon, in a few milliliters of water or breast milk (1|⊕⊕○). Brand rather than generic L-T4 tablets should be used, particularly during infancy and in severe cases (2|⊕⊕○). L-T4 liquid should only be used if pharmaceutically produced (1|⊕⊕○). Parents should be provided with written instructions on L-T4 treatment (1|⊕○○). 3.2 Monitoring of dose and follow-up Serum or plasma FT4 (or TT4) and TSH concentrations should be determined at least 4 hours after the last L-T4 administration (1|⊕⊕○). TSH concentration should be maintained in the age-specific reference range; TT4 or FT4 concentration should be maintained in the upper half of the age-specific reference range (1|⊕⊕○). Any reduction of L-T4 dose should not be based on a single increase in FT4 concentration during treatment (1|⊕⊕○). The first follow-up examination should take place 1–2 weeks after the start of L-T4 treatment (1|⊕○○). Subsequent evaluation should take place every 2 weeks until a complete normalization of TSH concentration is reached; then every 1 to 3 months thereafter until the age of 12 months (1|⊕○○). Between the ages of one and three years, children should undergo frequent clinical and laboratory evaluations (every 2 to 4 months) (1|⊕○○). Thereafter, evaluations should be carried out every 3 to 12 months until growth is completed (1|⊕○○). More frequent evaluations should be carried out if compliance is questioned or abnormal values are obtained (1|⊕○○). Additional evaluations should be carried out 4–6 weeks after any change in L-T4 dose or formulation (1| ⊕○○).

Subsequent evaluation should take place every 2 weeks until a complete normalization of TSH concentration is reached; then every 1 to 3 months thereafter until the age of 12 months (1|⊕○○). Between the ages of one and three years, children should undergo frequent clinical and laboratory evaluations (every 2 to 4 months) (1|⊕○○). Thereafter, evaluations should be carried out every 3 to 12 months until growth is completed (1|⊕○○). More frequent evaluations should be carried out if compliance is questioned or abnormal values are obtained (1|⊕○○). Additional evaluations should be carried out 4–6 weeks after any change in L-T4 dose or formulation (1| ⊕○○). Adequate treatment throughout childhood is essential and overtreatment should be avoided (1|⊕⊕⊕). 3.3 Thyroid re-evaluation Re-evaluation of the thyroid axis is indicated when no diagnostic assessment was carried out in infancy, and particularly when the infant was preterm/sick at the time of referral (1|⊕⊕⊕). For a precise diagnosis, L-T4 treatment should be phased out over a 4- to 6-week period, and a full reevaluation should be carried out, with both biochemical testing and thyroid imaging if hypothyroidism is confirmed (2|⊕⊕○). If the presence or absence of primary CH is being assessed, rather than an exact diagnosis being sought, re-evaluation may be carried out by decreasing the dose of L-T4 by 30% for 2–3 weeks and then rechecking thyroid function. If an increase in TSH concentrations to ≥ 10 mU/L is demonstrated, CH is confirmed. Otherwise the dose can be reduced further, with retesting after another 2–3 weeks. (2|⊕⊕○).

agnosis being sought, re-evaluation may be carried out by decreasing the dose of L-T4 by 30% for 2–3 weeks and then rechecking thyroid function. If an increase in TSH concentrations to ≥ 10 mU/L is demonstrated, CH is confirmed. Otherwise the dose can be reduced further, with retesting after another 2–3 weeks. (2|⊕⊕○). 3.4 Treatment and monitoring in pregnant women with CH We recommend an immediate increase in L-T4 dose by 25–30% following a missed menstrual cycle or positive home pregnancy test (1|⊕○○). TSH and FT4 (or total T4 [TT4]) levels should be monitored every 4–6 weeks during the pregnancy, aiming at TSH concentrations < 2.5 mU/L in the first trimester and < 3 mU/L later in pregnancy (1|⊕○○). 4.1 Outcomes in treated patients Psychomotor development and school progression should be monitored and recorded in all children with CH, and particularly in at-risk cases (absent knee epiphyses at term, very low TT4 or FT4, and very high TSH concentrations at diagnosis, athyreosis, delayed normalization of TSH, poor control during the first year, delayed milestones) (1|⊕⊕⊕). A personalized educational plan is required if school progress is affected in cases of severe CH (2|⊕⊕○). Concerns about behavior should be addressed from the time of diagnosis until school age (2|⊕⊕○). Memory deficits may be corrected by targeted training (2|⊕⊕○). Repeated (not just neonatal) hearing tests should be carried out before school age and, as required (2|⊕⊕○). We recommend assessment for evidence of visual processing problems (not just visual acuity) (2|⊕○○).

Concerns about behavior should be addressed from the time of diagnosis until school age (2|⊕⊕○). Memory deficits may be corrected by targeted training (2|⊕⊕○). Repeated (not just neonatal) hearing tests should be carried out before school age and, as required (2|⊕⊕○). We recommend assessment for evidence of visual processing problems (not just visual acuity) (2|⊕○○). We recommend screening for speech delay and referral for speech therapy by 3 years if required (2|⊕⊕○). 4.2 Health-related quality of life (HrQOL) Compliance with treatment should be promoted throughout life (1|⊕⊕⊕). There is a risk of subtle decrease in HrQOL in young adulthood, particularly if treatment is suboptimal (2|⊕○○). 4.3 Patient education and compliance/adherence Medical education about CH should be improved at all levels, with regular updates (1|⊕⊕⊕). The education of both parents and patients is essential particularly during transition to adult care and during pregnancy (1|⊕⊕⊕). 4.4 Growth, puberty, and fertility Adherence to treatment influences growth and should be promoted (1|⊕⊕⊕). Normal growth, puberty, and fertility can be anticipated, if adherence is reasonably good (1|⊕⊕⊕). 4.5 Bone health Patients with CH should be adequately treated with thyroxine, consume 800–1200 mg of calcium daily, and receive supplements added if intake is insufficient (2|⊕○○). 4.6 Metabolic and cardiovascular health We recommend lifestyle interventions, including diet and exercise, to optimize weight and health in individuals with CH (2|⊕○○).

4.5 Bone health Patients with CH should be adequately treated with thyroxine, consume 800–1200 mg of calcium daily, and receive supplements added if intake is insufficient (2|⊕○○). 4.6 Metabolic and cardiovascular health We recommend lifestyle interventions, including diet and exercise, to optimize weight and health in individuals with CH (2|⊕○○). 5.1 Criteria for genetic counselling Genetic counseling should involve explaining the risk of recurrence of CH in an affected family, based on family history and thyroid morphology (1|⊕⊕○). Each family with an affected child should have access to information about the two major forms of CH (dysgenesis and dyshormonogenesis) and should if possible receive an explanation of their inheritance and recurrence rate (1|⊕⊕○). Genetic counseling should be targeted rather than general (2|⊕⊕○). 5.2 Molecular biology in the diagnosis and management of CH Molecular genetic study should be preceded by a careful phenotypic description of CH patients (including morphology of the thyroid gland) (1|⊕⊕○). Any syndromic association should be studied genetically, to identify new CH genes and to make it possible to provide appropriate genetic counseling (1|⊕⊕○). The presence of familial cases of dysgenesis in siblings or parents should lead to a search for TSH receptor or PAX8 mutations, respectively (2|⊕⊕○).

5.2 Molecular biology in the diagnosis and management of CH Molecular genetic study should be preceded by a careful phenotypic description of CH patients (including morphology of the thyroid gland) (1|⊕⊕○). Any syndromic association should be studied genetically, to identify new CH genes and to make it possible to provide appropriate genetic counseling (1|⊕⊕○). The presence of familial cases of dysgenesis in siblings or parents should lead to a search for TSH receptor or PAX8 mutations, respectively (2|⊕⊕○). 5.3 Antenatal diagnosis, screening, and potential treatment of fetal CH We recommend antenatal diagnosis when goiter is fortuitously discovered during fetal ultrasound, with a family history of dyshormonogenesis and with known defects of genes involved in thyroid function or development (1|⊕⊕○). The therapeutic management of affected fetuses should comply with the laws of the country concerned (1|⊕⊕○). Cordocentesis, rather than amniocentesis, should be the reference method for assessing fetal thyroid function but only performed if prenatal intervention is considered (1|⊕⊕⊕). In a euthyroid pregnant woman, a large goiter in the fetus with progressive hydramnios and a risk of premature labor and delivery and/or concerns about tracheal occlusion are criteria in favor of fetal treatment in utero (1|⊕⊕○). Interventions such as intra-amniotic L-T4 injection should be performed only by multidisciplinary specialist teams (1|⊕⊕⊕). Conclusion Further research is required to improve our understanding of the pathophysiology and management of this heterogenous disorder.

In a euthyroid pregnant woman, a large goiter in the fetus with progressive hydramnios and a risk of premature labor and delivery and/or concerns about tracheal occlusion are criteria in favor of fetal treatment in utero (1|⊕⊕○). Interventions such as intra-amniotic L-T4 injection should be performed only by multidisciplinary specialist teams (1|⊕⊕⊕). Conclusion Further research is required to improve our understanding of the pathophysiology and management of this heterogenous disorder. Thyroid hormones play a crucial role in early neurodevelopment so that untreated severe CH results in neurological and psychiatric deficits, including intellectual disability, spasticity, and disturbances of gait and co-ordination. CH is one of the most common preventable causes of mental retardation. Screening programs, which have been in operation over the last 30 years in most industrialized countries, have led to the successful early detection and treatment of infants with CH and have eliminated the severe neurodevelopmental deficits resulting from late diagnosis. Studies on cognitive function in patients with CH treated soon after birth have shown that normal development can be achieved in most patients, although some may have subtle neurocognitive deficits (1).

reatment of infants with CH and have eliminated the severe neurodevelopmental deficits resulting from late diagnosis. Studies on cognitive function in patients with CH treated soon after birth have shown that normal development can be achieved in most patients, although some may have subtle neurocognitive deficits (1). Estimates of the prevalence of CH vary according to the method of ascertainment: about 1 in 2000 to 3000 live births in countries with neonatal screening vs about 1 in 6700 live births before the screening era (1). Recent reports have indicated that the incidence of primary CH may be increasing in some countries, particularly for cases with a normally located (eutopic) thyroid gland and milder dysfunction. The reasons for this remain unclear (2) but may relate to changes in screening thresholds (3, 4).

fore the screening era (1). Recent reports have indicated that the incidence of primary CH may be increasing in some countries, particularly for cases with a normally located (eutopic) thyroid gland and milder dysfunction. The reasons for this remain unclear (2) but may relate to changes in screening thresholds (3, 4). The results from neonatal screening programs have also helped to identify a broad spectrum of thyroid dysfunctions with different underlying etiologies. CH can be classified according to site: primary (thyroid) or secondary/central (pituitary and/or hypothalamic); to severity: FT4 levels within the normal range for age (compensated) or subnormal (decompensated); and to age of onset. The most common form of CH is primary hypothyroidism, with high TSH levels reflecting various types of abnormal thyroid gland development or dyshormonogenesis. Secondary hypothyroidism is much less frequent, either with isolated TSH deficiency due to mutations inactivating the TSH β-subunit, the TRH receptor, or IGSF1 (ImmunoGlobulin SuperFamily member 1), or more commonly with TSH deficiency associated with other pituitary hormone deficiencies.

elopment or dyshormonogenesis. Secondary hypothyroidism is much less frequent, either with isolated TSH deficiency due to mutations inactivating the TSH β-subunit, the TRH receptor, or IGSF1 (ImmunoGlobulin SuperFamily member 1), or more commonly with TSH deficiency associated with other pituitary hormone deficiencies. Impaired thyroid hormone production may also be temporary or permanent, the latter requiring lifelong treatment, and thyroid dysfunction may change in a given individual with growth and development stages (5). Transient primary CH can be defined as an increase in TSH levels during the neonatal period, with normal TFT results obtained off treatment at a later stage. The purely descriptive term “hyperthyrotropinemia” refers to a form of compensated CH in which there is a mild increase in TSH concentration (eg, 6–20 mU/L) with normal thyroid hormone concentrations. It may also be transient or permanent.

eonatal period, with normal TFT results obtained off treatment at a later stage. The purely descriptive term “hyperthyrotropinemia” refers to a form of compensated CH in which there is a mild increase in TSH concentration (eg, 6–20 mU/L) with normal thyroid hormone concentrations. It may also be transient or permanent. Method for Developing Evidence-Based Recommendations Given the importance of optimal screening, prompt diagnosis, and adequate treatment of CH, and in recognition of the considerable variations in its management worldwide, and the stated need for a consensus-building conference (6), the European Society for Paediatric Endocrinology (ESPE) decided to examine current best practice in CH and to formulate evidence-based recommendations. This was done by convening a panel of experts from the ESPE for a consensus conference on CH and also by inviting participation from members of the following societies: the Pediatric Endocrine Society [North America] (PES); the Asia Pacific Paediatric Endocrine Society (APPES); the Japanese Society for Pediatric Endocrinology (JSPE); the Sociedad Latino-Americana de Endocrinología Pediátrica (SLEP); the Australasian Paediatric Endocrine Group (APEG); and the Indian Society for Pediatric and Adolescent Endocrinology (ISPAE). The target audience for these guidelines includes general and specialist pediatricians, other professionals providing care for patients with CH, and policy makers, particularly in countries with developing economies currently in the process of initiating neonatal screening programs for CH.

Method for Developing Evidence-Based Recommendations Given the importance of optimal screening, prompt diagnosis, and adequate treatment of CH, and in recognition of the considerable variations in its management worldwide, and the stated need for a consensus-building conference (6), the European Society for Paediatric Endocrinology (ESPE) decided to examine current best practice in CH and to formulate evidence-based recommendations. This was done by convening a panel of experts from the ESPE for a consensus conference on CH and also by inviting participation from members of the following societies: the Pediatric Endocrine Society [North America] (PES); the Asia Pacific Paediatric Endocrine Society (APPES); the Japanese Society for Pediatric Endocrinology (JSPE); the Sociedad Latino-Americana de Endocrinología Pediátrica (SLEP); the Australasian Paediatric Endocrine Group (APEG); and the Indian Society for Pediatric and Adolescent Endocrinology (ISPAE). The target audience for these guidelines includes general and specialist pediatricians, other professionals providing care for patients with CH, and policy makers, particularly in countries with developing economies currently in the process of initiating neonatal screening programs for CH. Participants included individuals from Europe, North America (United States and Canada), Latin America, Asia, and Australia, with a balanced spectrum of professional seniority and expertise. In addition, an expert on the development of evidence-based guidelines was recruited to serve in an advisory capacity. Panel members declared whether they had any potential conflict of interest at the initial meeting of the group.

a, Asia, and Australia, with a balanced spectrum of professional seniority and expertise. In addition, an expert on the development of evidence-based guidelines was recruited to serve in an advisory capacity. Panel members declared whether they had any potential conflict of interest at the initial meeting of the group. Thirty-two participants were assigned to one of five groups to which topics 1–5 were allocated, and a chairperson was designated for each group. Each participant prepared a summary of the literature relating to a particular question distributed before the conference (which was held over 2 days in November 2011). Each group revised the summaries, which were then presented to the full conference. This report is based on the questions addressed. A detailed description of the grading scheme has been published elsewhere (7). Recommendations were based on published findings and on expert opinion when appropriate. The best available research evidence was used to develop recommendations. Preference was given to articles written in English, identified by PubMed searches with MeSH terms. For each point, recommendations and evidence are described, with a modification in the grading evidence as follows: 1 = strong recommendation (applies to most patients in most circumstances, benefits clearly outweigh the risk); and 2 = weak recommendation (consensus opinion of working group or should be considered; the best action may depend on circumstances or patient values, benefits, and risks closely balanced or uncertain).

ows: 1 = strong recommendation (applies to most patients in most circumstances, benefits clearly outweigh the risk); and 2 = weak recommendation (consensus opinion of working group or should be considered; the best action may depend on circumstances or patient values, benefits, and risks closely balanced or uncertain). Qualify of evidence is indicated as follows: ⊕⊕⊕, high quality (prospective cohort studies or randomized controlled trials at low risk of bias); ⊕⊕○, moderate quality (observational studies or trials with methodological flaws, inconsistent or indirect evidence); and ⊕○○, low quality (case series or nonsystematic clinical observations). 1.0 Neonatal Screening 1.1 The benefits of CH screening Recommendations 1.1.1 Neonatal screening programs for CH have been highly successful and economically beneficial over the last four decades. Affected children are detected very soon after birth, mostly before clinical symptoms and signs become evident. Early detection and treatment prevent morbidity, particularly neurodevelopmental disabilities (1|⊕⊕⊕). 1.1.1 Evidence Many studies have confirmed the early success of CH screening for normalizing the cognitive outcomes of children with severe primary CH (8, 9), and the timing of the normalization of thyroid function may influence the outcome (10). The avoided lifetime costs of care for children in whom intellectual disability is prevented as a result of screening for CH have been estimated to exceed the costs of screening and diagnosis by a large margin (11).

imary CH (8, 9), and the timing of the normalization of thyroid function may influence the outcome (10). The avoided lifetime costs of care for children in whom intellectual disability is prevented as a result of screening for CH have been estimated to exceed the costs of screening and diagnosis by a large margin (11). 1.2 Analytical methodology, effectiveness, and efficacy of CH screening strategies Recommendations 1.2.1 Screening for primary CH worldwide should be performed wherever possible on the basis of national resources. For new programs, there is a need to decide on the scope of screening to define the strategy for selecting neonatal screening tests. The goal of neonatal screening should be to detect all forms of primary CH—mild, moderate, and severe—but particularly those patients with severe CH in whom morbidity is high. The most sensitive test for detecting primary CH is the determination of TSH concentration (1|⊕⊕⊕).

tegy for selecting neonatal screening tests. The goal of neonatal screening should be to detect all forms of primary CH—mild, moderate, and severe—but particularly those patients with severe CH in whom morbidity is high. The most sensitive test for detecting primary CH is the determination of TSH concentration (1|⊕⊕⊕). 1.2.2 Primary CH screening has been shown to be effective for the testing of cord blood or blood collected after the age of 24 hours, although the best “window” for testing is 48 to 72 hours of age. Blood is spotted onto filter paper, allowed to dry, and eluted into a buffer for TSH analysis. This method detects primary CH more effectively than primary T4 screening. Primary T4 screening with confirmatory TSH testing entails a risk of missing some cases of mild forms of primary CH but can detect some cases of central CH (CCH). Screening strategies for the detection of CCH are based on two approaches: 1) a combination of primary T4 and primary TSH screening; and 2) a combination of primary T4 screening with secondary TSH testing followed by T4 binding protein determination. The inclusion of T4 binding protein determinations decreases the number of false positives. The criteria defining a positive result must be adapted to the target disease definition and the resources of the screening program (1|⊕⊕⊕).

y T4 screening with secondary TSH testing followed by T4 binding protein determination. The inclusion of T4 binding protein determinations decreases the number of false positives. The criteria defining a positive result must be adapted to the target disease definition and the resources of the screening program (1|⊕⊕⊕). 1.2.1–1.2.2 Evidence The most convincing justification for expanding neonatal screening for CH to every country in the world is that this approach is the most effective way of preventing mental retardation and ensuring normal IQ in this patient population (12, 13). Furthermore, because iodine deficiency is the most common preventable cause of mental retardation, developmental disabilities, and CH worldwide (8, 14, 15), neonatal screening for CH can be used as a sensitive indicator of neonatal and maternal iodine nutritional status (16). The strategy for selecting neonatal screening tests focuses on detecting the more severe forms of CH as early as possible because disability due to primary CH is greatest in patients not treated before the age of 3 months (8, 17). TSH screening is the most sensitive test for primary CH detection and should be the single most important test in any screening program (8, 17, 18). Predictably, an increase in the reported incidence of primary CH occurs when cutoff levels for TSH are lowered (3, 12, 17, 19, 20). Studies on long-term outcome are required to determine whether there is a risk of permanent disability in these milder cases with only moderate TSH elevation and normal T4 levels and whether these individuals have permanent or transient thyroid dysfunction (6).

hen cutoff levels for TSH are lowered (3, 12, 17, 19, 20). Studies on long-term outcome are required to determine whether there is a risk of permanent disability in these milder cases with only moderate TSH elevation and normal T4 levels and whether these individuals have permanent or transient thyroid dysfunction (6). There is some published evidence to suggest that neonatal screening for CCH may also fulfil criteria for disease screening (21–24): 1) CCH is a relatively frequent disease with an incidence similar to that of phenylketonuria in some populations; 2) screening tests are available and inexpensive; 3) treatment is available and effective; and 4) the risks of an unfavorable outcome in cases of delayed diagnosis are well known, although outcome studies showing that screening is superior to detection through clinical presentation are lacking (22).

n some populations; 2) screening tests are available and inexpensive; 3) treatment is available and effective; and 4) the risks of an unfavorable outcome in cases of delayed diagnosis are well known, although outcome studies showing that screening is superior to detection through clinical presentation are lacking (22). 1.3 Screening in special categories of neonates at risk of CH Recommendations 1.3.1 Specific biochemical criteria should be used for screening special categories of neonates at risk of transient and permanent CH and in whom initial screening tests may be inappropriate or provide normal results. A strategy of second screening may be required in the following conditions: preterm neonates with a gestational age (GA) of less than 37 weeks; LBW and VLBW neonates; ill and preterm neonates admitted to NICU; specimen collection within the first 24 hours of life; and multiple births, particularly in cases of same-sex twins. The repeat specimen should be collected at about 2 weeks of age, or 2 weeks after the first screening test was carried out. The interpretation of screening results should take into account the results of all specimens analyzed in a multiple sampling strategy. The criteria defining a positive screening test result should be adapted for the analytical parameters measured, the method used, and the age at sampling and maturity (GA/birth weight) of the infant (2|⊕⊕○).

creening results should take into account the results of all specimens analyzed in a multiple sampling strategy. The criteria defining a positive screening test result should be adapted for the analytical parameters measured, the method used, and the age at sampling and maturity (GA/birth weight) of the infant (2|⊕⊕○). 1.3.1 Evidence Several studies have generated data that argue for multiple sampling in preterm neonates with a GA of less than 37 weeks, LBW and very LBW neonates, ill and preterm neonates admitted to NICU, infants from whom a specimen is collected within the first 24 hours of life, and neonates from multiple births, particularly in case of monozygotic twins (8, 9, 12, 17, 19, 25–29). This approach reflects concern that primary CH may be masked in these situations due to the suppression of TSH caused by drug administration (30, 31), by hypothalamic-pituitary immaturity (32), by fetal blood mixing in multiple births (33), and by other effects of serious neonatal illnesses (34–36). Thus, it is the policy in many centers to remeasure dried blood spot (DBS) TSH in at-risk infants as they approach discharge from hospital. Repeat screening has not been adopted by all screening programs, some centers arguing that the limited data available suggest that, although it has identified neonates with delayed rise in TSH, this is mostly a transient problem (27). Further outcome data from this complex population of newborn infants are needed to better inform clinical practice.

adopted by all screening programs, some centers arguing that the limited data available suggest that, although it has identified neonates with delayed rise in TSH, this is mostly a transient problem (27). Further outcome data from this complex population of newborn infants are needed to better inform clinical practice. Some neonatal screening programs test for T4 alone in the initial screen, whereas TSH may also be assessed on the first specimen. However, most programs initially assessing T4 only subsequently evaluate TSH concentration for the neonates with the lowest T4 values (usually the lowest 10% of the T4 values for the day). If TSH concentration is high, the infant is recalled for evaluation and testing. Repeat DBS specimens are collected if the T4 value is below a defined cutoff value for GA (37, 38). If the TSH test result in the initial screening was normal, but repeat testing shows TSH concentration to be high, the evaluation and possible treatment of transient (in most cases) or permanent primary CH should be initiated promptly. Neonates with a persistently low T4 concentration in DBS tests should have serum FT4 and TSH determinations to confirm or exclude CCH (8, 9).

l, but repeat testing shows TSH concentration to be high, the evaluation and possible treatment of transient (in most cases) or permanent primary CH should be initiated promptly. Neonates with a persistently low T4 concentration in DBS tests should have serum FT4 and TSH determinations to confirm or exclude CCH (8, 9). 2.0 Criteria for Diagnosis 2.1 Biochemical criteria for use in the decision to initiate treatment in an infant with high TSH and/or low FT4 concentration Recommendations 2.1.1 We recommend starting treatment immediately after baseline serum TSH and FT4 determination if DBS TSH concentration is ≥ 40 mU/L of whole blood. If DBS TSH concentration is < 40 mU/L of whole blood, the clinician may postpone treatment, pending the serum results, for 1–2 days (1|⊕⊕○). 2.1.2 We recommend starting treatment immediately if serum FT4 concentration is below the norm for age, regardless of TSH concentration (1|⊕⊕⊕). 2.1.3 We suggest that treatment should be started if venous TSH concentration is persistently > 20 mU/L, even if serum FT4 concentration is normal (2|⊕○○). 2.1.4 When venous TSH concentration is between 6 and 20 mU/L in a well baby with an FT4 concentration within the normal limits for age, we suggest diagnostic imaging to try to establish a definitive diagnosis (2|⊕○○). If TSH concentration remains high for more than 3 to 4 weeks, we suggest (in discussion with the family) either starting L-T4 supplementation immediately and retesting, off treatment, at a later stage; or retesting 2 weeks later without treatment (2|⊕○○).

stic imaging to try to establish a definitive diagnosis (2|⊕○○). If TSH concentration remains high for more than 3 to 4 weeks, we suggest (in discussion with the family) either starting L-T4 supplementation immediately and retesting, off treatment, at a later stage; or retesting 2 weeks later without treatment (2|⊕○○). 2.1.1- 2.1.4 Evidence L-T4 treatment must be started immediately if venous FT4 or TT4 levels are low, given the known adverse effect of untreated decompensated CH on somatic growth and neurodevelopment (39). Previous work has shown that the likelihood of decompensated hypothyroidism is high if DBS TSH values are above 40 mU/L (40), justifying immediate treatment if DBS TSH concentration is above this value. Given that the period between birth and the age of 3 years is a critical time for neurocognitive development, most clinicians would advocate treatment when TSH concentration is > 20 mU/L, carefully monitoring thyroid function to avoid overtreatment, and retesting after 3 years if the thyroid is normally located (41). Management remains a matter of debate for cases in which TSH concentrations are high, but to a lesser extent (6–20 mU/L), and FT4 levels are normal (42). The family should be informed that this is a “gray area” but that many clinicians would advise “playing safe” and treating during early childhood in this situation (43).

agement remains a matter of debate for cases in which TSH concentrations are high, but to a lesser extent (6–20 mU/L), and FT4 levels are normal (42). The family should be informed that this is a “gray area” but that many clinicians would advise “playing safe” and treating during early childhood in this situation (43). 2.2 Communication of high TSH concentration results on neonatal screening to the family, the family doctor, and the local pediatrician Recommendations Communicating the initially high capillary TSH concentration result 2.2.1 The detection of a high TSH concentration on screening should be communicated by an experienced person (2|⊕○○), such as a member of the screening laboratory staff or one of the pediatric endocrine team. 2.2.2 The high TSH concentration should be communicated to the family either by telephone or in person as soon as possible (2|⊕○○). 2.2.3 Communication may be directly with the family or via the family doctor, health visitor, or midwife (2|⊕○○). Seeing the baby for clinical examination, investigation, and treatment 2.2.4 If possible, the baby should be seen for clinical assessment and venous thyroid function testing on the day of referral or on the next day at the latest (1|⊕○○). 2.2.5 Centers should have relevant information materials about the diagnosis and management of CH in the appropriate languages for their community (2|⊕○○). 2.2.6 The parents should be shown how to give the first dose of L-T4, by either the clinician or the pharmacist (2|⊕○○).

Seeing the baby for clinical examination, investigation, and treatment 2.2.4 If possible, the baby should be seen for clinical assessment and venous thyroid function testing on the day of referral or on the next day at the latest (1|⊕○○). 2.2.5 Centers should have relevant information materials about the diagnosis and management of CH in the appropriate languages for their community (2|⊕○○). 2.2.6 The parents should be shown how to give the first dose of L-T4, by either the clinician or the pharmacist (2|⊕○○). Communication with the family doctor, local pediatrician, and educators 2.2.7 The baby's family doctor and local pediatrician should be notified either by telephone or by letter to outline the provisional diagnosis and management (1|⊕○○). 2.2.8 Options regarding shared care should be discussed with the clinician at the local level (1|⊕○○). 2.2.9 When the child reaches nursery or school age, the consensus group advises against informing educators and teachers about the child having CH, to prevent “labeling” and stigmatization (2|⊕⊕○).

Communication with the family doctor, local pediatrician, and educators 2.2.7 The baby's family doctor and local pediatrician should be notified either by telephone or by letter to outline the provisional diagnosis and management (1|⊕○○). 2.2.8 Options regarding shared care should be discussed with the clinician at the local level (1|⊕○○). 2.2.9 When the child reaches nursery or school age, the consensus group advises against informing educators and teachers about the child having CH, to prevent “labeling” and stigmatization (2|⊕⊕○). 2.2.1–2.2.9 Evidence ESPE guidelines provide little information about contacting and counseling the families of neonates with high TSH concentrations (44). A survey of British pediatricians in 2006/2007 indicated that 54 of the 119 pediatricians questioned (43%) would always see the infant on the day of notification, whereas 65 (55%) would usually see the baby a day or so later (45). Age at treatment initiation during the first month of age was not shown to be a factor in educational attainment in a large French cohort study (46), but there are common sense grounds for starting L-T4 treatment as soon as possible after birth to prevent irreversible neurocognitive impairment.

a day or so later (45). Age at treatment initiation during the first month of age was not shown to be a factor in educational attainment in a large French cohort study (46), but there are common sense grounds for starting L-T4 treatment as soon as possible after birth to prevent irreversible neurocognitive impairment. 2.3 Criteria for assessing CH severity in terms of clinical, biochemical, and radiological features Recommendations 2.3.1 CH severity can be assessed clinically—on the basis of symptomatic hypothyroidism; biologically—as severe, moderate, or mild on the basis of serum FT4 levels of <5, 5 to <10, and 10 to 15 pmol/L, respectively; on the basis of delayed epiphyseal maturation on knee x-ray; and in terms of the etiology of CH (1|⊕⊕⊕). A serum thyroglobulin concentration below the detection threshold is highly suggestive of athyreosis or a complete thyroglobulin synthesis defect (2|⊕⊕⊕).

2.3 Criteria for assessing CH severity in terms of clinical, biochemical, and radiological features Recommendations 2.3.1 CH severity can be assessed clinically—on the basis of symptomatic hypothyroidism; biologically—as severe, moderate, or mild on the basis of serum FT4 levels of <5, 5 to <10, and 10 to 15 pmol/L, respectively; on the basis of delayed epiphyseal maturation on knee x-ray; and in terms of the etiology of CH (1|⊕⊕⊕). A serum thyroglobulin concentration below the detection threshold is highly suggestive of athyreosis or a complete thyroglobulin synthesis defect (2|⊕⊕⊕). 2.3.1 Evidence The clinical symptoms and signs of symptomatic CH include sleepiness, not waking for feeds, poor and slow feeding, cold extremities, prolonged neonatal jaundice, lethargy, hypotonia, macroglossia, umbilical hernia, and dry skin with or without a coarse/puffy face. Persistence of the posterior fontanelle, a large anterior fontanelle, and a wide sagittal suture all reflect delayed bone maturation, which can be further documented by knee x-ray. The absence of one or both knee epiphyses has been shown to be related to: 1) T4 concentration at diagnosis; and 2) IQ outcome, and is thus a reliable index of intrauterine hypothyroidism (47, 48). A threshold effect of subnormal TT4 concentration on IQ has also been claimed, with a 10-point IQ difference between children with initial TT4 concentrations below 40 nmol/L (equivalent to 5.5 pmol/L of FT4) and children with normal TT4 concentrations (49). Data for 82 10-day-old neonates yielded 2.5th percentile, median, and 97.5th percentile values of 1.18, 1.75, and 2.49 ng/dL—equivalent to 15.2, 22.5, and 32 pmol/L (50)—making it possible to construct a scale of biochemical severity on the basis of plasma FT4 concentrations of <5, 5 to <10, and 10 to 15 pmol/L. Imaging may reveal severe primary CH in cases of absence/severe hypoplasia of the gland or of complete organification defect with goiter. Alternatively, imaging may show various degrees of severity in cases of ectopic gland or normally shaped and located gland. A pragmatic conclusion as to the severity of CH can therefore be made when the clinical history, physical findings, initial venous blood biochemistry results, and knee x-ray and thyroid imaging results (if available) are considered together.

es of severity in cases of ectopic gland or normally shaped and located gland. A pragmatic conclusion as to the severity of CH can therefore be made when the clinical history, physical findings, initial venous blood biochemistry results, and knee x-ray and thyroid imaging results (if available) are considered together. 2.4 The place of imaging techniques—scintigraphy with or without perchlorate discharge test and ultrasonography—in the diagnosis of CH Recommendations 2.4.1 We recommend performing imaging studies to determine the specific etiology (1|⊕⊕○). 2.4.2 Both scintigraphy and ultrasound should be considered in neonates with high TSH concentrations (2|⊕⊕○). 2.4.3 Imaging should never be allowed to delay the initiation of treatment. Scintigraphy should be carried out within 7 days of starting L-T4 treatment (1|⊕⊕○). 2.4.4 Infants found to have a normal-sized gland in situ in the absence of a clear diagnosis should undergo further reassessment of the thyroid axis at a later age (1|⊕○○).

2.4.3 Imaging should never be allowed to delay the initiation of treatment. Scintigraphy should be carried out within 7 days of starting L-T4 treatment (1|⊕⊕○). 2.4.4 Infants found to have a normal-sized gland in situ in the absence of a clear diagnosis should undergo further reassessment of the thyroid axis at a later age (1|⊕○○). 2.4.1–2.4.4 Evidence Scintigraphy Scintigraphy may be carried out with either 10–20 MBq of technetium-99m (99mTc) or 1–2 MBq of iodine-123 (123I). 99mTc is more widely available, less expensive, and quicker to use than 123I. However, 123I is specifically taken up by the thyroid gland and gives a clearer scan than 99mTc (51). Scintigraphy can identify athyreosis (absence of uptake), hypoplasia of a gland in situ (with or without hemithyroid), a normal or large gland in situ with or without abnormally high levels of uptake, and an ectopic thyroid at any point along the pathway of the normal embryological descent from the foramen caecum at the base of the tongue to the thyroid cartilage. An enrichment of the tracer within the salivary glands can lead to misinterpretation, especially on lateral views, but this can be avoided by allowing the infant to feed before scintigraphy, thus emptying the salivary glands. When the thyroid is in the normal position, a discharge of >10% of the 123I dose when perchlorate is administered at 2 hours (the perchlorate discharge test) indicates an organification defect (51). Scintigraphy may show no uptake despite the presence of a eutopic thyroid gland with excess iodine intake through exposure (eg, from antiseptic preparations), maternal TSH receptor blocking antibodies, TSH suppression from L-T4 treatment, and inactivating mutations in the TSH receptor and the sodium/iodide symporter (NIS) (52, 53).

may show no uptake despite the presence of a eutopic thyroid gland with excess iodine intake through exposure (eg, from antiseptic preparations), maternal TSH receptor blocking antibodies, TSH suppression from L-T4 treatment, and inactivating mutations in the TSH receptor and the sodium/iodide symporter (NIS) (52, 53). Ultrasound The thyroid gland is a superficial structure that can be imaged by ultrasonography with a high-frequency linear array transducer (10–15 MHz) at a resolution of 0.7 to 1.0 mm. Ultrasound imaging, performed in the longitudinal and axial planes, can be used to investigate the absence or presence, size, echogenic texture, and structure of a thyroid gland in situ. However, it cannot always detect lingual and sublingual thyroid ectopy (54–56), although the use of color Doppler facilitates the identification of thyroid tissue by demonstrating marked increases in blood flow (57). Ultrasonography is highly observer-dependent, and investigators should be particularly wary of misdiagnosing nonthyroidal tissue in the thyroid fossa as a dysplastic thyroid gland in situ (54, 58). Thyroid tissue is more echogenic than muscle but less echogenic than fat. In the absence of thyroid tissue in the normal location, small hyperechoic structures of approximately the same echogenicity as fat are found laterally on both sides of the trachea, mimicking the appearance of a thyroid gland. Cysts have also been described within the empty thyroid area (59).

t less echogenic than fat. In the absence of thyroid tissue in the normal location, small hyperechoic structures of approximately the same echogenicity as fat are found laterally on both sides of the trachea, mimicking the appearance of a thyroid gland. Cysts have also been described within the empty thyroid area (59). Scintigraphy and ultrasound combined Combining scintigraphy and thyroid ultrasound in the individual patient helps to: 1) improve diagnostic accuracy (55, 60); 2) identify a eutopic gland, which may be normal, enlarged, or hypoplastic, thus guiding further diagnostic investigations, including molecular genetics studies; 3) prevent the incorrect diagnosis of athyreosis in the context of an absence of uptake on scintigraphy when ultrasound shows a normal gland in situ; and 4) detect thyroid ectopy reliably. Table 1 shows the diagnostic patterns to be found in thyroid dysgenesis, dyshormonogenesis, and some forms of transient CH when ultrasound, scintigraphy, and serum thyroglobulin measurement are combined. Table 1. Thyroid Ultrasound, Scintigraphy, and Serum Thyroglobulin Findings in Thyroid Dysgenesis, Dyshormonogenesis, and Some Forms of Transient CH

Scintigraphy and ultrasound combined Combining scintigraphy and thyroid ultrasound in the individual patient helps to: 1) improve diagnostic accuracy (55, 60); 2) identify a eutopic gland, which may be normal, enlarged, or hypoplastic, thus guiding further diagnostic investigations, including molecular genetics studies; 3) prevent the incorrect diagnosis of athyreosis in the context of an absence of uptake on scintigraphy when ultrasound shows a normal gland in situ; and 4) detect thyroid ectopy reliably. Table 1 shows the diagnostic patterns to be found in thyroid dysgenesis, dyshormonogenesis, and some forms of transient CH when ultrasound, scintigraphy, and serum thyroglobulin measurement are combined. Table 1. Thyroid Ultrasound, Scintigraphy, and Serum Thyroglobulin Findings in Thyroid Dysgenesis, Dyshormonogenesis, and Some Forms of Transient CH Defect Thyroid Ultrasound Thyroid Scintigraphy Serum Thyroglobulin Concentration Thyroid dysgenesis Apparent athyreosis No thyroid tissue seen No uptake Detectable (≥2 μg/L) True athyreosis No thyroid tissue seen No uptake Undetectable Ectopy Either no thyroid tissue seen or ectopic tissue seen (especially if in a sublingual or perihyoid location) Uptake into ectopic gland Usually ↑ but may be N or ↓ Hypoplasia in situ Small eutopic gland Low level of uptake in a normally sited gland N or ↓ Hemiagenesis Hemithyroid Hemithyroid N Dyshormonogenesis NIS/SCL5A5 Enlarged gland Uptake absent or ↓↓ ↑ Thyroid peroxidase, TPO Enlarged gland High level of uptake; positive perchlorate discharge test ↑↑ Dual oxidase 2, DUOX2/dual oxidase 2 maturation factor, DUOXA2 Enlarged gland High level of uptake; positive perchlorate discharge test ↑ Thyroglobulin, TG Enlarged gland Avid uptake; normal perchlorate discharge test ↓↓ or undetectable Pendred syndrome, pendrin PDS/SCL26A4 Normal/enlarged gland High level of uptake; positive perchlorate discharge test ↑ Dehalogenase, IYD/DEHAL1 Enlarged gland Avid uptake; normal perchlorate discharge test ↑ Transient CH Acute iodine excess Normal gland in situ No uptake N or ↓ Chronic iodine deficiency Large gland Avid uptake ↑ Maternal blocking antibodies N or small gland Uptake ↓ or absent N or ↓ TSH receptor, +/− N or small gland Uptake ↓ or absent N or ↓ Abbreviation: N, normal.

ke; normal perchlorate discharge test ↑ Transient CH Acute iodine excess Normal gland in situ No uptake N or ↓ Chronic iodine deficiency Large gland Avid uptake ↑ Maternal blocking antibodies N or small gland Uptake ↓ or absent N or ↓ TSH receptor, +/− N or small gland Uptake ↓ or absent N or ↓ Abbreviation: N, normal. 2.5. Congenital malformations and syndromes that should be systematically sought for in infants with CH Recommendations 2.5.1 A thorough physical examination should be carried out in all neonates with high TSH concentrations for the detection of congenital malformations, particularly those affecting the heart, and in children for the identification of any underlying dysmorphic syndrome or neurodevelopmental disorders (1|⊕⊕⊕). 2.5.1 Evidence The prevalence of congenital malformations, particularly cardiac malformations, including septal defects, renal abnormalities, and the risk of neurodevelopmental disorders is higher in subjects with CH than in the general population (61–65). However, care must be taken to distinguish between true CH and transient increases in TSH concentration in sick infants with and without extrathyroidal malformations, including heart and great vessel defects such as patent ductus arteriosus (66). There is no good evidence to suggest that additional screening measures, other than careful clinical examination, are required for detecting extrathyroidal malformations or comorbidities.

fants with and without extrathyroidal malformations, including heart and great vessel defects such as patent ductus arteriosus (66). There is no good evidence to suggest that additional screening measures, other than careful clinical examination, are required for detecting extrathyroidal malformations or comorbidities. Down syndrome is associated with a mild increase in TSH concentration from the neonatal period onward, although it is usually too small for detection by neonatal screening, as well as a shift to the left of the FT4 distribution so that mean values are lower when compared with the general population (67). Pendred syndrome, with or without goiter, and pseudohypoparathyroidism may both present with mild or moderate increases in TSH concentration during the neonatal period and should be included in the differential diagnosis of CH with gland in situ (60). 3.0 Treatment and Monitoring of CH 3.1 Initial treatment and therapeutic regimens Recommendations 3.1.1 L-T4 alone is recommended as the treatment of choice for CH (1|⊕⊕○). 3.1.2 Treatment with L-T4 should be started as soon as possible, and no later than the first 2 weeks of life or immediately after confirmatory serum test results in infants identified in a second routine screening test. We recommend an initial L-T4 dose of 10 to 15 μg/kg per day. Infants with severe disease, as defined by a very low pretreatment TT4 or FT4 concentration, should be treated with the highest initial dose, and those with a mild to moderate hypothyroidism with a lower dose (1|⊕⊕○).

ified in a second routine screening test. We recommend an initial L-T4 dose of 10 to 15 μg/kg per day. Infants with severe disease, as defined by a very low pretreatment TT4 or FT4 concentration, should be treated with the highest initial dose, and those with a mild to moderate hypothyroidism with a lower dose (1|⊕⊕○). 3.1.3 L-T4 should be administered orally. If oral administration is not possible, it can also be administered iv, in which case the iv dose should be no more than 80% the oral dose. The dose should then be adjusted according to TSH and FT4 determinations (1|⊕⊕○). 3.1.4 We recommend the administration of L-T4 in tablet form. In neonates and infants, the tablets can be crushed and administered via a small spoon, with suspension, if necessary, in a few milliliters of water or breast milk. L-T4 can also be administered in liquid form, but only if pharmaceutically produced and licensed L-T4 solutions are available (1|⊕⊕○). A brand rather than a generic formulation is recommended in CH, at least in infancy (2|⊕⊕○). 3.1.5 L-T4 can be administered or taken in the morning or evening, either before feeding or with food, but it should be administered in the same way every day. The dose should then be adjusted according to TSH and FT4 determinations to establish the appropriate dose in each setting. Caution should be taken with the administration of vitamin D during the first weeks of life, and the intake of soy, iron, and calcium at the time of L-T4 administration should be avoided (1|⊕⊕⊕).

The dose should then be adjusted according to TSH and FT4 determinations to establish the appropriate dose in each setting. Caution should be taken with the administration of vitamin D during the first weeks of life, and the intake of soy, iron, and calcium at the time of L-T4 administration should be avoided (1|⊕⊕⊕). 3.1.6 Parents should be provided with written instructions on L-T4 treatment to avoid uncertainties that might hinder compliance (1|⊕○○). 3.1.1–3.1.6 Evidence T3 is the biologically active hormone, but there is no evidence that combined therapy with L-T4 and L-T3 is more beneficial than treatment with L-T4 alone, probably due to the high degree of efficiency of endogenous deiodinases, which break T4 down into T3 (68, 69). L-T4 is available in tablet form (the most widely used form) or as a pharmaceutically produced and licensed liquid form. Unlike suspensions prepared by pharmacists, this licensed L-T4 solution allows reliable dosing and is a convenient way of administering L-T4, particularly to infants and young children (70–73). Recent evidence suggests that brand and generic L-T4 are not bioequivalent and that for CH, particularly in severe cases, it is prudent to use a brand preparation (74).

, this licensed L-T4 solution allows reliable dosing and is a convenient way of administering L-T4, particularly to infants and young children (70–73). Recent evidence suggests that brand and generic L-T4 are not bioequivalent and that for CH, particularly in severe cases, it is prudent to use a brand preparation (74). The orally administered hormone has a mean bioavailability of 50 to 80%, which may be influenced by the presence of food (soy) or minerals (calcium, iron). In infants with CH, hypersensitivity to vitamin D administration with hypercalcemia has been described during the first few weeks of L-T4 treatment, possibly due to the administration of prophylactic doses of vitamin D (75, 76). In adults, L-T4 administration at bedtime seems to be even more effective in terms of thyroid hormone levels than administration in the morning and is now considered to be as effective as morning administration in the fasting state (77).

t, possibly due to the administration of prophylactic doses of vitamin D (75, 76). In adults, L-T4 administration at bedtime seems to be even more effective in terms of thyroid hormone levels than administration in the morning and is now considered to be as effective as morning administration in the fasting state (77). The early initiation of L-T4 treatment, within the first 2 weeks of life, has been shown to be crucial for neurodevelopment and for the achievement of a normal intellectual outcome in affected children (78–81). Disease severity, as judged by very low initial levels of thyroid hormones due to the absence or loss of function of the thyroid gland and by severely delayed bone age, has also been shown to be an important predictive factor for neurodevelopment (49, 82–87). Severely affected children may benefit from a higher initial dose of L-T4, leading to the more rapid normalization of thyroid hormone levels and potentially resulting in a better intellectual outcome (88–91).

ly delayed bone age, has also been shown to be an important predictive factor for neurodevelopment (49, 82–87). Severely affected children may benefit from a higher initial dose of L-T4, leading to the more rapid normalization of thyroid hormone levels and potentially resulting in a better intellectual outcome (88–91). 3.2 Monitoring of treatment and adverse events Recommendations 3.2.1 The monitoring of L-T4 treatment should be based on periodic measurements of serum or plasma FT4 (or TT4) and TSH concentrations. The blood samples for laboratory evaluation should be collected at least 4 hours after the last L-T4 administration (1|⊕⊕○). We recommend maintaining TSH in the age-specific reference range (but to avoid undetectable TSH < 0.05 mU/L) and serum concentrations of TT4 or FT4 in the upper half of the age-specific reference range (1|⊕⊕○). If necessary, the treatment should be adjusted according to the hormone concentrations measured, but decreases in LT4 dose should not be based on a single high FT4 concentration during treatment (1|⊕⊕○). Clinicians should be familiar with the reference ranges for the methods used by the laboratory carrying out the tests (1|⊕○○).

ry, the treatment should be adjusted according to the hormone concentrations measured, but decreases in LT4 dose should not be based on a single high FT4 concentration during treatment (1|⊕⊕○). Clinicians should be familiar with the reference ranges for the methods used by the laboratory carrying out the tests (1|⊕○○). 3.2.2 We recommend performing the first follow-up examination 1 to 2 weeks after the start of L-T4 treatment, with intense follow-up over the first year of life (every 2 weeks until TSH levels are completely normalized and every 1 to 3 months thereafter, until the age of 12 months). Between the ages of 1 and 3 years, children should undergo frequent clinical and laboratory evaluations (every 2 to 4 months), with regular evaluations every 3 to 12 months thereafter until growth is completed. Measurements should be performed at more frequent intervals if compliance is questioned or abnormal values are obtained, and 4 to 6 weeks after any change in L-T4 dose or L-T4 formulation (eg, switch from brand to generic L-T4) (1|⊕○○). 3.2.3 The incidence of adverse events during L-T4 treatment is very low. The careful monitoring of thyroid hormone parameters, during both initial and maintenance treatment, is recommended to minimize the risk (1|⊕○○). 3.2.4 In children with pre-existing cardiac insufficiency, we suggest introducing L-T4 at 50% of the target replacement dose and increasing this in accordance with FT4 levels after 2 weeks (2|⊕○○).

3.2.3 The incidence of adverse events during L-T4 treatment is very low. The careful monitoring of thyroid hormone parameters, during both initial and maintenance treatment, is recommended to minimize the risk (1|⊕○○). 3.2.4 In children with pre-existing cardiac insufficiency, we suggest introducing L-T4 at 50% of the target replacement dose and increasing this in accordance with FT4 levels after 2 weeks (2|⊕○○). 3.2.1–3.2.4 Evidence The rapid normalization of thyroid hormone levels (within the first 2 weeks after treatment initiation) and the maintenance of relatively higher FT4 concentrations during the first year of life lead to a better intellectual outcome (90, 92, 93). The frequent monitoring of TSH and FT4 levels is required for this and also for preventing the occurrence of prolonged periods of supraphysiological thyroid hormone levels (94–96). After adjustment to L-T4 dosage, it is appropriate to recheck thyroid function, and the recommended interval of 4–6 weeks is in keeping with the American Academy of Pediatrics guidelines (8). The adequate treatment of CH minimizes the risk of treatment-related adverse effects (97–99). Based on neurological and cardiac complications that have occasionally been described (albeit primarily in patients with inadequate T4 treatment), patients with pre-existing health conditions or with a very late diagnosis may require special attention during treatment (100, 101).

atment-related adverse effects (97–99). Based on neurological and cardiac complications that have occasionally been described (albeit primarily in patients with inadequate T4 treatment), patients with pre-existing health conditions or with a very late diagnosis may require special attention during treatment (100, 101). 3.3. Criteria for re-evaluating the thyroid axis, to distinguish between permanent CH and transient increases in TSH concentration, and for treatment withdrawal in children with normally sited gland Criteria for re-evaluation of the thyroid axis Recommendations 3.3.1 We recommend re-evaluation of the thyroid axis in cases in which no etiological diagnostic assessment was carried out during early infancy and/or when treatment was started in the context of the infant being ill (eg, preterm). Re-evaluation is also mandatory when initial evaluation has shown a normally located gland, with or without goiter, in neonates with positive thyroid antibodies, in children who have required no increase in L-T4 dose since infancy, and in children in whom no enzyme defect has been identified, either because no molecular genetic investigations have been carried out or because investigations have proved negative for all mutations tested (1|⊕⊕○). 3.3.2 Retesting off treatment may be waived if venous TSH concentration has risen after the first year of life, due to either L-T4 underdosage or poor compliance with treatment (1|⊕⊕○).

3.3. Criteria for re-evaluating the thyroid axis, to distinguish between permanent CH and transient increases in TSH concentration, and for treatment withdrawal in children with normally sited gland Criteria for re-evaluation of the thyroid axis Recommendations 3.3.1 We recommend re-evaluation of the thyroid axis in cases in which no etiological diagnostic assessment was carried out during early infancy and/or when treatment was started in the context of the infant being ill (eg, preterm). Re-evaluation is also mandatory when initial evaluation has shown a normally located gland, with or without goiter, in neonates with positive thyroid antibodies, in children who have required no increase in L-T4 dose since infancy, and in children in whom no enzyme defect has been identified, either because no molecular genetic investigations have been carried out or because investigations have proved negative for all mutations tested (1|⊕⊕○). 3.3.2 Retesting off treatment may be waived if venous TSH concentration has risen after the first year of life, due to either L-T4 underdosage or poor compliance with treatment (1|⊕⊕○). 3.3.3 Re-evaluation of the thyroid axis is not indicated when thyroid dysgenesis has been conclusively shown on imaging or (with the exception of DUOX.2 mutations or Pendred syndrome) when dyshormonogenesis has been confirmed by molecular genetic testing (1|⊕⊕⊕).

3.3.2 Retesting off treatment may be waived if venous TSH concentration has risen after the first year of life, due to either L-T4 underdosage or poor compliance with treatment (1|⊕⊕○). 3.3.3 Re-evaluation of the thyroid axis is not indicated when thyroid dysgenesis has been conclusively shown on imaging or (with the exception of DUOX.2 mutations or Pendred syndrome) when dyshormonogenesis has been confirmed by molecular genetic testing (1|⊕⊕⊕). 3.3.1–3.3.3 Evidence Thyroid re-evaluation is important if no definitive diagnosis was made during the neonatal period; it has been shown that one-third of patients with CH and normally located glands may have transient thyroid dysfunction (41, 60). Re-evaluation is unnecessary if thyroid imaging at the time of neonatal screening showed thyroid ectopy, apparent athyreosis, or true athyreosis. However, caution is required when diagnosing athyreosis on the basis of isotope scanning alone because an absence of uptake in the context of a gland normally located on ultrasound scan may occur with excess iodine exposure, maternal antibodies blocking the TSH receptor, or NIS gene defects (52, 53).

r true athyreosis. However, caution is required when diagnosing athyreosis on the basis of isotope scanning alone because an absence of uptake in the context of a gland normally located on ultrasound scan may occur with excess iodine exposure, maternal antibodies blocking the TSH receptor, or NIS gene defects (52, 53). Re-evaluation is essential in subjects who were preterm or sick during the neonatal period (27). Other infants in whom initial evaluation showed a normal/slightly small gland on ultrasound with little or no uptake on scintigraphy should also be re-evaluated because this pattern is suggestive of either maternal antibodies blocking the TSH receptor or biallelic-TSH receptor mutations (102). Because iodine deficiency may mimic dyshormonogenesis (both conditions display thyroid enlargement and avid tracer uptake), retesting is indicated in children who appear to have mild dyshormonogenesis (103). Transient CH has also been linked to genetic defects such as heterozygous DUOX2 mutation (104, 105). Timing of thyroid re-evaluation Recommendations 3.3.4 Re-evaluation of the thyroid axis, off treatment, should normally take place after the age of 3 years (1|⊕⊕○). 3.3.5 Earlier re-evaluation may be indicated if the clinical context renders transient increases in TSH concentration highly probable; for example: 1) in the case of newborns in whom thyroid peroxidase or TSH receptor antibodies are detectable in the blood; and 2) when a eutopic, normally sized gland is found on ultrasound scans (2|⊕⊕○).

aluation may be indicated if the clinical context renders transient increases in TSH concentration highly probable; for example: 1) in the case of newborns in whom thyroid peroxidase or TSH receptor antibodies are detectable in the blood; and 2) when a eutopic, normally sized gland is found on ultrasound scans (2|⊕⊕○). 3.3.4–3.3.5 Evidence Magnetic resonance imaging studies have shown that the myelination of the central nervous system is completed by 36 to 40 months of age (106), at which point the child is more likely to be co-operative for thyroid imaging than at the age of 1 or 2 years. However, if transient increases in TSH concentration are likely, the clinician may consider the earlier withdrawal of treatment from 1 year of age. Method of thyroid re-evaluation Recommendations 3.3.6 If a precise diagnosis is sought, L-T4 should be phased out over a 4 to 6-week period (depending on the size of the maintenance dose) and a full re-evaluation carried out at the end of this period, with biochemical testing and thyroid imaging if hypothyroidism is confirmed (2|⊕○○). 3.3.7 If the clinician wishes to establish the presence or absence of primary hypothyroidism rather than to obtain an exact diagnosis, L-T4 dose may be decreased by 30% for 2 to 3 weeks. If an increase in TSH concentration to ≥ 10 mU/L is observed during this period, then continuing hypothyroidism can be assumed. By contrast, if thyroid function remains normal, the dose may be reduced further, followed by retesting (2|⊕⊕○).

obtain an exact diagnosis, L-T4 dose may be decreased by 30% for 2 to 3 weeks. If an increase in TSH concentration to ≥ 10 mU/L is observed during this period, then continuing hypothyroidism can be assumed. By contrast, if thyroid function remains normal, the dose may be reduced further, followed by retesting (2|⊕⊕○). 3.3.6–3.3.7 Evidence By the age of 2 to 3 years, the severity of thyroid impairment may be evident from a lower dose requirement (41, 103) or, if compliance has been imperfect, high TSH levels despite treatment. 3.4 Treatment and monitoring in pregnant women with CH Recommendations 3.4.1 For women with CH who are planning a pregnancy, even greater efforts should be made to ensure that maternal thyroid hormones are optimal. In newly pregnant patients, we recommend an immediate increase in L-T4 dose by 25 to 30% after a missed menstrual cycle or positive home pregnancy test (1|⊕○○). 3.4.2 TSH and FT4 (or TT4) levels should be evaluated as soon as pregnancy is confirmed, every 4 to 6 weeks during the pregnancy, and 4 weeks after every change in dose (1|⊕⊕⊕). 3.4.3 The goal of the treatment is to maintain TSH concentration below 2.5 mU/L in the first trimester and below 3 mU/L later in pregnancy (1|⊕○○). 3.4.1–3.4.3 Evidence There is currently no specific evidence for treatment schedules in pregnant women with CH. We therefore refer to the guidelines of the American Thyroid Association and The Endocrine Society regarding the management of thyroid disease during pregnancy (107–109).

3.4.3 The goal of the treatment is to maintain TSH concentration below 2.5 mU/L in the first trimester and below 3 mU/L later in pregnancy (1|⊕○○). 3.4.1–3.4.3 Evidence There is currently no specific evidence for treatment schedules in pregnant women with CH. We therefore refer to the guidelines of the American Thyroid Association and The Endocrine Society regarding the management of thyroid disease during pregnancy (107–109). 4.0 Outcomes of Treated Patients 4.1 Neurodevelopmental outcome Neurocognition, behavior, memory, psychomotor, school progression, language, hearing, and visuospatial skills Recommendations 4.1.1 Psychomotor and language development as well as school progression should be monitored and recorded in all children with CH (1|⊕⊕○). Clinicians should pay particular attention to developmental delays or learning difficulties and to attention problems in children with severe CH (athyreosis, absent knee epiphyses at term, very low T4 and very high TSH concentrations at diagnosis) or poor endocrine control, particularly during the first year, and in those from economically disadvantaged families (1|⊕⊕⊕). 4.1.2 We suggest specialized stimulation of motor development, if required, and a personalized educational plan if school progression is affected (2|⊕⊕○). 4.1.3 Memory deficits may be corrected by targeted training (2|⊕⊕○). 4.1.4 Concerns about behavior should be addressed from the time of diagnosis until school age (2|⊕⊕○). 4.1.5 Adequate treatment throughout childhood is essential, and overtreatment should be avoided (1|⊕⊕⊕).

4.1.2 We suggest specialized stimulation of motor development, if required, and a personalized educational plan if school progression is affected (2|⊕⊕○). 4.1.3 Memory deficits may be corrected by targeted training (2|⊕⊕○). 4.1.4 Concerns about behavior should be addressed from the time of diagnosis until school age (2|⊕⊕○). 4.1.5 Adequate treatment throughout childhood is essential, and overtreatment should be avoided (1|⊕⊕⊕). 4.1.6 The determinants of remaining deficits require further study. 4.1.1–4.1.6 Evidence With early and adequate treatment, intellectual disability (defined as an IQ < 70) has disappeared from cohorts screened for CH, and the mean global IQ is now 10 to 30 points higher in these patients than in the prescreening era (1). Some affected patients still have neurocognitive and behavioral sequelae of CH that persist into adolescence and adulthood and that are related to disease severity (78, 82, 110, 111). Cognitive outcome is related to age at treatment and L-T4 dose (90); school progression may be affected (112). Cognitive outcome is related to the parents' socioeducational status (49).

and behavioral sequelae of CH that persist into adolescence and adulthood and that are related to disease severity (78, 82, 110, 111). Cognitive outcome is related to age at treatment and L-T4 dose (90); school progression may be affected (112). Cognitive outcome is related to the parents' socioeducational status (49). Behavior scores on initial admission to school are within the normal range, (89) and the perception of the impact of CH on behavior varies with age and differs between children and their parents (113). The impact of informing teachers of the diagnosis of CH has not been investigated. Patients with CH have no increase in the risk of attention deficit-hyperactivity disorder but may have more sustained attention problems related to episodes of overtreatment (114, 115) and, in severe cases, slower information processing (116). Subtle and specific memory deficits and reduced hippocampal volumes may be observed (117). There is also a risk of fine motor impairment (118). However, most early-treated patients are well integrated into society with no impairment in educational level (46). Hearing, visual, verbal development Recommendations 4.1.7 Repeated (not just neonatal) hearing tests should be carried out before school age and as required (2|⊕⊕○). 4.1.8 Assessing patients for evidence of visual processing problems (not just visual acuity) is suggested (2|⊕○○). 4.1.9 We recommend screening for delays in speech acquisition by the age of 3 years and propose speech therapy as required (2|⊕⊕○).

Hearing, visual, verbal development Recommendations 4.1.7 Repeated (not just neonatal) hearing tests should be carried out before school age and as required (2|⊕⊕○). 4.1.8 Assessing patients for evidence of visual processing problems (not just visual acuity) is suggested (2|⊕○○). 4.1.9 We recommend screening for delays in speech acquisition by the age of 3 years and propose speech therapy as required (2|⊕⊕○). 4.1.7–4.1.9 Evidence Even after excluding patients with Pendred syndrome, a higher prevalence of hearing impairment has been observed in patients with CH than in the reference population, possibly necessitating the use of hearing aids in childhood. Substantial adverse effects on speech development, school performance, and social interactions may occur if hearing impairment is undiagnosed (119). This impairment may result from the role of thyroid hormone in cochlear development and auditory function (46, 120, 121). There is also a risk of visual processing problems (10, 46). 4.2 Health-related quality of life Recommendations 4.2.1 Compliance with treatment should be promoted throughout life (1|⊕⊕⊕). 4.2.2 In future studies of HrQOL, “focusing illusion” should be considered (a form of stigmatization similar to “labeling” at school) (2|⊕○○). 4.2.1–4.2.2 Evidence There is a risk of a subtle decrease in HrQOL, particularly in mental dimensions and if treatment is suboptimal (46, 122, 123). 4.3 Patient and professional education; compliance/adherence Recommendations 4.3.1 Medical education about CH should be improved at all levels, with regular updates (1|⊕⊕⊕).

4.2.1–4.2.2 Evidence There is a risk of a subtle decrease in HrQOL, particularly in mental dimensions and if treatment is suboptimal (46, 122, 123). 4.3 Patient and professional education; compliance/adherence Recommendations 4.3.1 Medical education about CH should be improved at all levels, with regular updates (1|⊕⊕⊕). 4.3.2 We suggest identifying resources in the community to which patients can be referred for continuing education about self-management of their condition and to update their knowledge of CH. The education of both parents and patients is essential, with particular attention paid to the transition to adult care and management during pregnancy (1|⊕⊕⊕). 4.3.1–4.3.2 Evidence The adherence of physicians to guidelines is low (45). Poor adherence with treatment and low treatment adequacy are prevalent at all ages (46, 124). 4.4 Growth, puberty, and fertility Recommendations 4.4.1 Adherence to treatment influences growth and should be promoted (1|⊕⊕⊕). 4.4.2 Provided treatment is adequate, it is appropriate to provide patients and their parents with reassurance about growth, puberty, and fertility (1|⊕⊕⊕). 4.4.1–4.4.2 Evidence Length and height increase within normal limits in patients with adequately treated CH (125). Patients may be overweight in early childhood and adulthood (46, 126). Head circumference may be greater than normal, but this reflects bone rather than brain development (127). The onset of puberty, age at menarche, and menstrual cycles are normal. Fecundity is generally normal, except in the most severely affected female patients (128, 129).

t in early childhood and adulthood (46, 126). Head circumference may be greater than normal, but this reflects bone rather than brain development (127). The onset of puberty, age at menarche, and menstrual cycles are normal. Fecundity is generally normal, except in the most severely affected female patients (128, 129). 4.5 Bone health 4.5.1 Recommendations We recommend that patients with CH should be adequately treated with L-T4 and consume 800–1200 mg of calcium daily, ideally from dietary sources, with calcium supplements added if dietary calcium intake is insufficient (2|⊕○○). 4.5.1 Evidence Thyroid hormones exert major effects on bone remodeling. Patients overtreated with L-T4 display higher levels of bone resorption than of bone formation, leading to progressive bone loss. The goal of therapy is to render the patient euthyroid, with a normal TSH concentration. Only a few studies have evaluated the impact of long-term L-T4 treatment on bone mineral density. Two studies reported that bone mineral density in children and young adults with CH is within the normal range (130, 131). More data for patients treated with the currently used doses are required. 4.6 Metabolic and cardiovascular health Recommendations 4.6.1 Weight should be closely monitored. Lifestyle interventions, including diet and exercise, should be encouraged in individuals with CH (2|⊕○○). 4.6.2 Optimal treatment of CH is essential for cardiovascular health (1|⊕○○).

4.5.1 Evidence Thyroid hormones exert major effects on bone remodeling. Patients overtreated with L-T4 display higher levels of bone resorption than of bone formation, leading to progressive bone loss. The goal of therapy is to render the patient euthyroid, with a normal TSH concentration. Only a few studies have evaluated the impact of long-term L-T4 treatment on bone mineral density. Two studies reported that bone mineral density in children and young adults with CH is within the normal range (130, 131). More data for patients treated with the currently used doses are required. 4.6 Metabolic and cardiovascular health Recommendations 4.6.1 Weight should be closely monitored. Lifestyle interventions, including diet and exercise, should be encouraged in individuals with CH (2|⊕○○). 4.6.2 Optimal treatment of CH is essential for cardiovascular health (1|⊕○○). 4.6.1–4.6.2 Evidence Patients with CH have a higher risk of being overweight and, thus, of metabolic complications (46). In addition to the higher risk of congenital heart malformations (61, 62), there is a subtle increase in cardiovascular risk factors in young adults with CH, related to treatment adequacy (132).

4.6.2 Optimal treatment of CH is essential for cardiovascular health (1|⊕○○). 4.6.1–4.6.2 Evidence Patients with CH have a higher risk of being overweight and, thus, of metabolic complications (46). In addition to the higher risk of congenital heart malformations (61, 62), there is a subtle increase in cardiovascular risk factors in young adults with CH, related to treatment adequacy (132). 5.0 Genetic Counseling and Antenatal Management 5.1 Criteria for genetic counseling Recommendations 5.1.1 Genetic counseling should provide information about the risk of recurrence of CH in a family with CH, based on family history and thyroid morphology. Each family with an affected child should have access to information on the two major forms of CH (dysgenesis and dyshormonogenesis) and should receive an explanation of inheritance and recurrence rate. A certified geneticist or a genetic counselor (depending upon the organization of healthcare in the country concerned) should be made available in some cases. In such cases, given the current state of knowledge, we propose targeted rather than general genetic counseling in the context of certain clinical situations, described in Table 2 (2|⊕⊕○). Table 2. Situations in Which Genetic Counseling Should Be Offered

5.0 Genetic Counseling and Antenatal Management 5.1 Criteria for genetic counseling Recommendations 5.1.1 Genetic counseling should provide information about the risk of recurrence of CH in a family with CH, based on family history and thyroid morphology. Each family with an affected child should have access to information on the two major forms of CH (dysgenesis and dyshormonogenesis) and should receive an explanation of inheritance and recurrence rate. A certified geneticist or a genetic counselor (depending upon the organization of healthcare in the country concerned) should be made available in some cases. In such cases, given the current state of knowledge, we propose targeted rather than general genetic counseling in the context of certain clinical situations, described in Table 2 (2|⊕⊕○). Table 2. Situations in Which Genetic Counseling Should Be Offered I. Pregnant women Positive family history for nonsyndromic CH Dyshormonogenesis (previously affected child) (1|⊕⊕⊕) Dysgenesis (at least 1 member of the family) (2|⊕⊕○) Positive family history of syndromic CH with Neurological disorders, including unexplained mental retardation Deafness Congenital heart disease, surfactant deficiency syndrome Cleft palate Kidney malformations Any sign of Albright hereditary osteodystrophy (GNAS mutation) (1|⊕⊕○) Unexplained abnormality of T4, T3, or TSH levels in family members (mild forms of CH) (2|⊕⊕○) II. Infant or child with CH (2|⊕⊕○) Subject with Deafness Neurological signs (hypotonia, choreoathetosis, intellectual disability) Lung disorders (surfactant deficiency syndrome, interstitial lung disease) Congenital heart disease Cleft palate Kidney malformations Any sign of Albright hereditary osteodystrophy (GNAS mutation) Family history Consanguinity Kidney malformations Deafness Specific malformations (as listed above) Unexplained mental retardation despite adequate treatment of CH in family members Any sign of Albright hereditary osteodystrophy (GNAS mutation) 5.1.1 Evidence In primary CH, about 80% of cases are due to thyroid dysgenesis and 20% are due to thyroid dyshormonogenesis. Thyroid dyshormonogenesis is caused by mutations in genes encoding proteins involved in thyroid hormone synthesis: the SCL5A5/NIS (iodide transport defect; OMIM No. 274400); pendrin, SCL26A4/PDS (Pendred syndrome; OMIM No. 274600); thyroglobulin, TG (OMIM No. 274700); thyroid peroxidase, TPO (OMIM No. 274500); dual oxidase 2, DUOX2 (OMIM No. 607200); dual oxidase maturation factor 2, DUOX2A (OMIM No. 274900); or iodotyrosine deiodinase, IYD/DEHAL1 (OMIM No. 274800). These mutations are inherited in an autosomal recessive manner and are associated with no other malformations, other than deafness in Pendred syndrome (133). Isolated thyroid dysgenesis (OMIM No. 218700) is generally a sporadic disease.

DUOX2A (OMIM No. 274900); or iodotyrosine deiodinase, IYD/DEHAL1 (OMIM No. 274800). These mutations are inherited in an autosomal recessive manner and are associated with no other malformations, other than deafness in Pendred syndrome (133). Isolated thyroid dysgenesis (OMIM No. 218700) is generally a sporadic disease. However, three observations suggest a possible, as yet unknown genetic basis: 1) a higher rate of familial cases (>15 times higher) than would be expected by chance alone; 2) minor morphological thyroid abnormalities in euthyroid first-degree relatives of patients with thyroid dysgenesis; and 3) a high incidence of associated extrathyroidal malformations (61–63, 134–137). Specific genetic forms of syndromic and nonsyndromic thyroid dysgenesis and TSH resistance may be associated with mutations in the NK2 homeobox 1 (NKX2–1, brain-lung-thyroid syndrome; OMIM No. 610978); Forkhead box E1 (FOXE-1, Bamforth-Lazarus syndrome; OMIM No. 241850), Paired box gene 8 (PAX8; OMIM No. 218700), NK2 homeobox 5 (NKX2–5; OMIM No. 225250), TSH receptor (TSHR; OMIM No. 275200); and Gs α (GNAS, pseudohypoparathyroidism type 1A; OMIM No. 103580) genes (133, 138) (Table 3). Table 3. Genetic Diagnosis to Detect the Individual Molecular Basis of CH

However, three observations suggest a possible, as yet unknown genetic basis: 1) a higher rate of familial cases (>15 times higher) than would be expected by chance alone; 2) minor morphological thyroid abnormalities in euthyroid first-degree relatives of patients with thyroid dysgenesis; and 3) a high incidence of associated extrathyroidal malformations (61–63, 134–137). Specific genetic forms of syndromic and nonsyndromic thyroid dysgenesis and TSH resistance may be associated with mutations in the NK2 homeobox 1 (NKX2–1, brain-lung-thyroid syndrome; OMIM No. 610978); Forkhead box E1 (FOXE-1, Bamforth-Lazarus syndrome; OMIM No. 241850), Paired box gene 8 (PAX8; OMIM No. 218700), NK2 homeobox 5 (NKX2–5; OMIM No. 225250), TSH receptor (TSHR; OMIM No. 275200); and Gs α (GNAS, pseudohypoparathyroidism type 1A; OMIM No. 103580) genes (133, 138) (Table 3). Table 3. Genetic Diagnosis to Detect the Individual Molecular Basis of CH Thyroid Morphology, as Assessed by Ultrasonography and/or Scintigraphy Family History Consanguinity or Siblings/Cousins With CH Parents With CH Isolated CH Normally located thyroid with normal perchlorate discharge test TSH-R (if hypoplasia), TG (if goiter, low TG level) PAX8 Normally located thyroid with abnormal perchlorate discharge test (ie, iodide organification defects) TPO, DUOX2/DUOXA2 +/− TG Normally located thyroid on ultrasonography, with no iodide uptake on scintigraphy SCL5A5/NIS, TSH-R (if hypoplasia) Syndromic CH Deafness Normally located thyroid SCL26A4/PDS Short stature, obesity, hypocalcemia Normally located thyroid GNAS Cleft palate, “spiky” hair Athyreosis (hypoplasia) FOXE1 (no mutations described in patients with ectopic or normally sized and sited thyroid gland to date) Kidney agenesis or any malformation of the genitourinary tract Athyreosis, ectopic thyroid gland, normally located thyroid +/− hypoplasia PAX8 PAX8 Choreoathetosis or neurological disease Normally located thyroid, hypoplasia (athyreosis) NKX2–1 (no mutations described in ectopic cases so far) NKX2–1 Lung disorders (surfactant deficiency syndrome at term, interstitial lung disease) Normally located thyroid, hypoplasia (athyreosis) NKX2–1 (no mutations described in ectopic cases so far) NKX2–1 Cardiac defects Ectopy (athyreosis) NKX2–5 NKX2–5 5.2 Molecular biology in the diagnosis and management of CH Recommendations 5.2.1 Careful phenotypic description of CH patients (including morphological analysis of the thyroid gland) is required, and we suggest that any syndromic association should be studied genetically to identify new genes involved in CH and to ensure that healthcare staff are in a position to offer appropriate genetic counseling. The presence of familial cases of dysgenesis should lead to a search for TSHR and PAX8 gene mutations (2|⊕⊕○).

, and we suggest that any syndromic association should be studied genetically to identify new genes involved in CH and to ensure that healthcare staff are in a position to offer appropriate genetic counseling. The presence of familial cases of dysgenesis should lead to a search for TSHR and PAX8 gene mutations (2|⊕⊕○). The genetic diagnosis of CH may facilitate the targeting of specific interdisciplinary follow-up and supportive care for patients (1|⊕⊕○). The relevant features to consider are listed in Table 3.

, and we suggest that any syndromic association should be studied genetically to identify new genes involved in CH and to ensure that healthcare staff are in a position to offer appropriate genetic counseling. The presence of familial cases of dysgenesis should lead to a search for TSHR and PAX8 gene mutations (2|⊕⊕○). The genetic diagnosis of CH may facilitate the targeting of specific interdisciplinary follow-up and supportive care for patients (1|⊕⊕○). The relevant features to consider are listed in Table 3. 5.2.1 Evidence Molecular biology techniques can identify the cause of CH on the basis of family history and thyroid morphology. The identification of a NKX2–1 mutation implies that special attention should be paid to neurological development and to lung disease in the follow-up of affected children (139, 140). The identification of a FOXE1 mutation implies that special attention should be paid to neurological development (141). The identification of a PAX8 mutation should lead to kidney and urinary tract ultrasound and, probably, to the monitoring of renal function if malformations are found (142). The identification of a SLC26A4/PDS mutation implies that special attention should be paid to the hearing of the child (143). The identification of a TG or TPO mutation implies a risk of thyroid cancer within the goiter in adulthood, as demonstrated by long-term follow-up studies in extremely rare published cases (133, 144). It is unclear whether the thyroid cancer is gene-specific or related to goiter development. The identification of a GNAS mutation should lead the clinician to focus on other potentially associated endocrine and nonendocrine disorders (145). Despite intensive and focused research, mutations in these genes have been found in fewer than 10% of CH patients to date, and the usual discordance between monozygotic twins (136) remains unexplained.

utation should lead the clinician to focus on other potentially associated endocrine and nonendocrine disorders (145). Despite intensive and focused research, mutations in these genes have been found in fewer than 10% of CH patients to date, and the usual discordance between monozygotic twins (136) remains unexplained. 5.3 Potential indications for antenatal diagnosis, screening methods for fetal hypothyroidism, management and criteria for fetal treatment in utero Recommendations 5.3.1 We recommend antenatal diagnosis in cases of goiter fortuitously discovered during systematic ultrasound examination of the fetus, in relation to thyroid dyshormonogenesis (1|⊕⊕⊕); a familial recurrence of CH due to dyshormonogenesis (25% recurrence rate) (1|⊕⊕⊕); and known defects of genes involved in thyroid function or development with potential germline transmission (1|⊕⊕○). Special issues should be considered for syndromic cases with potential mortality and possible germline mosaicism (as for NKX2–1 gene mutation/deletion and severe pulmonary dysfunction with possible transmission via germline mosaicism). In such circumstances, the discussion of the prenatal diagnosis should be open. The therapeutic management of affected fetuses should comply with the laws in force in the country concerned (1|⊕⊕○). The familial recurrence of CH due to dysgenesis (2% of familial occurrences) requires further study to determine the feasibility and clinical relevance for antenatal detection.

diagnosis should be open. The therapeutic management of affected fetuses should comply with the laws in force in the country concerned (1|⊕⊕○). The familial recurrence of CH due to dysgenesis (2% of familial occurrences) requires further study to determine the feasibility and clinical relevance for antenatal detection. 5.3.2 For the evaluation of fetal thyroid volume, we recommend ultrasound scans at 20 to 22 weeks gestation to detect fetal thyroid hypertrophy and potential thyroid dysfunction in the fetus. Goiter or an absence of thyroid tissue can also be documented by this technique. Measurements should be made as a function of GA, and thyroid perimeter and diameter should be measured to document goiter (1|⊕⊕⊕). 5.3.3 We recommend cordocentesis, rather than amniocentesis, as the reference method for assessing fetal thyroid function. Norms have been established as a function of GA. This examination should be carried out only if prenatal intervention is considered (see below) (1|⊕⊕⊕). 5.3.4 In most cases, fetal thyroid function can be inferred from context and ultrasound criteria, and fetal blood sampling is, therefore, only exceptionally required (2|⊕⊕○). 5.3.5 In a euthyroid pregnant woman, a large goiter in the fetus with progressive hydramnios and a risk of premature labor and delivery and/or concerns about tracheal occlusion are criteria in favor of fetal treatment in utero (1|⊕⊕○). 5.3.6 In a hypothyroid pregnant woman, the initial approach should be to treat the pregnant woman, rather than the fetus, with L-T4 (1|⊕⊕○).

5.3.5 In a euthyroid pregnant woman, a large goiter in the fetus with progressive hydramnios and a risk of premature labor and delivery and/or concerns about tracheal occlusion are criteria in favor of fetal treatment in utero (1|⊕⊕○). 5.3.6 In a hypothyroid pregnant woman, the initial approach should be to treat the pregnant woman, rather than the fetus, with L-T4 (1|⊕⊕○). 5.3.7 For goitrous nonimmune fetal hypothyroidism leading to hydramnios, intra-amniotic injections of L-T4 have been reported to decrease the size of the fetal thyroid gland. However, experience with this procedure is limited, and the risk of provoking premature labor or infections should be evaluated with care. Thus, follow-up studies are very important to determine the efficacy and possible adverse long-term consequences of medical interventions on the fetus. Such interventions should be performed only by multidisciplinary specialist teams (pediatric endocrinologists, adult endocrinologists for the pregnant mother, neonatologists, and obstetricians with experience in antenatal care and procedures) (1|⊕⊕⊕).

dverse long-term consequences of medical interventions on the fetus. Such interventions should be performed only by multidisciplinary specialist teams (pediatric endocrinologists, adult endocrinologists for the pregnant mother, neonatologists, and obstetricians with experience in antenatal care and procedures) (1|⊕⊕⊕). 5.3.8 Studies have confirmed the feasibility and safety of intra-amniotic L-T4 injection and strongly suggest that this treatment is effective for decreasing goiter size. However, none of the many L-T4 regimens used ensures euthyroidism at birth. It is therefore not possible to formulate guidelines from current data. The expert panel proposes the use of 10 μg/kg estimated fetal weight/15 days in the form of intra-amniotic injections. The risks to the fetus and the psychological burden on the parents should be factored into the risk/benefit evaluation (2|⊕○○). 5.3.9 Determination of the indications and optimal modes of prenatal treatment for nonimmune fetal goitrous hypothyroidism will require larger, well-designed studies that would be best conducted via international co-operation between multidisciplinary medical teams. Alternative ways of treating the fetus by administering drugs to the mother should also be investigated (2|⊕○○).

prenatal treatment for nonimmune fetal goitrous hypothyroidism will require larger, well-designed studies that would be best conducted via international co-operation between multidisciplinary medical teams. Alternative ways of treating the fetus by administering drugs to the mother should also be investigated (2|⊕○○). 5.3.1–5.3.9 Evidence Recent advances in fetal imaging techniques (ultrasonography) and fetal hormonology have made it possible to identify thyroid function disorders in the fetus that could potentially be treated in utero by administering drugs to the mother. Several interventions have also been proposed for improving the fetal outcomes of fetal hypothyroid disorders by considering the fetus as the patient to be treated and gaining direct access to the amniotic cavity. These approaches range from public health interventions with clear benefits and negligible risks, such as increasing the iodine intake of all pregnant women, to procedures with a much less clear benefit-to-risk ratio, such as cordocentesis for determining thyroid function in a fetus with goiter and repeated intra-amniotic injections of L-T4 (146–152) (Table 4). Table 4. Screening, Prevention, and Management of Fetal Hypothyroidism

5.3.1–5.3.9 Evidence Recent advances in fetal imaging techniques (ultrasonography) and fetal hormonology have made it possible to identify thyroid function disorders in the fetus that could potentially be treated in utero by administering drugs to the mother. Several interventions have also been proposed for improving the fetal outcomes of fetal hypothyroid disorders by considering the fetus as the patient to be treated and gaining direct access to the amniotic cavity. These approaches range from public health interventions with clear benefits and negligible risks, such as increasing the iodine intake of all pregnant women, to procedures with a much less clear benefit-to-risk ratio, such as cordocentesis for determining thyroid function in a fetus with goiter and repeated intra-amniotic injections of L-T4 (146–152) (Table 4). Table 4. Screening, Prevention, and Management of Fetal Hypothyroidism Adequate iodine intake should be ensured for all pregnant women (250 μg/d). For women with a personal or family history of thyroid disease, serum TSH and FT4 concentrations should be determined before pregnancy, at the start of pregnancy, and during pregnancy. On ultrasonography at about 22 and 32 wk gestation, fetal thyroid diameter and circumference should be measured; if above the 95th percentile for GA, a fetal thyroid disorder should be considered. If a pregnant woman is treated with L-T4, care should be taken to ensure an appropriate increase in dose during the course of the pregnancy. If fetal goiter is documented, cordocentesis and fetal serum FT4 and TSH determinations should be considered, and intra-amniotic L-T4 injections should be administered if severe hypothyroidism is diagnosed and progressive hydramnios develops. Conclusions Patients with CH benefit from neonatal screening, which makes it possible to initiate essential replacement therapy immediately. This consensus highlights the need to identify clear cutoff points for CH screening, without increasing the number of false-positive results. However, the results of long-term prospective studies in subjects with false-positive results are not yet available for the formulation of evidence-based recommendations for diagnosis and management. Based on current outcome data, an immediate treatment should be initiated, and it appears necessary to maintain adequate L-T4 treatment throughout the lifespan for most patients, with a particular emphasis on treatment in the first years of life and of treatment optimization in pregnant women with CH. Careful neurodevelopmental and neurosensory evaluations should be started early in life and repeated at important critical developmental phases, taking into account disease severity at diagnosis and providing appropriate interventions as required. Subsequent efforts should focus on educating both patients and caregivers to ensure that adequate treatment is continued into adulthood.

e started early in life and repeated at important critical developmental phases, taking into account disease severity at diagnosis and providing appropriate interventions as required. Subsequent efforts should focus on educating both patients and caregivers to ensure that adequate treatment is continued into adulthood. Future research should aim to improve our understanding of the pathophysiology of this heterogeneous disorder and to determine whether knowledge of the specific defect in thyroid development or function is likely to improve patient care and outcomes.

e started early in life and repeated at important critical developmental phases, taking into account disease severity at diagnosis and providing appropriate interventions as required. Subsequent efforts should focus on educating both patients and caregivers to ensure that adequate treatment is continued into adulthood. Future research should aim to improve our understanding of the pathophysiology of this heterogeneous disorder and to determine whether knowledge of the specific defect in thyroid development or function is likely to improve patient care and outcomes. Participants of the Congenital Hypothyroidism Consensus Conference Group Paulina Alexander, Department of Pediatric Endocrinology, and Diabetes, Charite Children's Hospital, Berlin, Germany; Miguel Alvarez, Institute of Neurology and Neurosurgery, Havana, Cuba; Oliver Blankenstein, Department of Pediatric Endocrinology, and Diabetes, Charite Children's Hospital, Berlin, Germany; Alessandra Cassio, Department of Pediatrics, University of Bologna, Bologna, Italy; Mireille Castanet, Department of Pediatrics, University of Rouen, Rouen, France; Graziano Cesaretti, Department of Pediatric Endocrinology, University of Pisa, Pisa, Italy; Tim Cheetham, Department of Paediatric Endocrinology, Newcastle University, Newcastle-upon-Tyne, United Kingdom; Anna Chiesa, Endocrinology Division, Ricardo Gutierrez Children's Hospital, Buenos Aires, Argentina; Carlo Corbetta, Neonatal Screening Laboratory, Ospedale dei Bambini “Vittore Buzzi,” Milan, Italy; Maria E. Craig, Institute of Endocrinology and Diabetes, Sydney Children's Hospital Network and University of Sydney, Sydney, Australia; Thomas P. Foley, Department of Pediatrics, University of Pittsburgh, Pittsburgh, Pennsylvania; Scott Grosse, Atlanta, Georgia; Marlies Kempers, Radboud University Nijmegen Medical Centre, Department of Genetics, Nijmegen, The Netherlands; Claire de Labriolle-Vaylet, UPMC Univ Paris IV and Nuclear Medicine, Hopital Trousseau, Paris, France; Stephen LaFranchi, Department of Pediatrics, Oregon Health and Sciences University, Portland, Oregon; Dominique Luton, Department of Obstetrics and Gynecology, University Diderot Paris 7, Clichy, France; Masanori Minagawa, Division of Endocrinology, Department of Pediatrics, Chiba Children's Hospital, Chiba, Japan; Palany Raghupathy, Department of Child Health, Christian Medical College, Vellore, India; Robert Rapaport, Department of Pediatric Endocrinology, Mount Sinai School of Medicine, New York, New York; Joanne Rovet, Departments of Psychology and Paediatrics, University of Toronto, Toronto, Canada; Diet S.

al, Chiba, Japan; Palany Raghupathy, Department of Child Health, Christian Medical College, Vellore, India; Robert Rapaport, Department of Pediatric Endocrinology, Mount Sinai School of Medicine, New York, New York; Joanne Rovet, Departments of Psychology and Paediatrics, University of Toronto, Toronto, Canada; Diet S. Rustama, Department of Child Health, Padjadjaran University School of Medicine, Bandung, Indonesia; Mariacarolina Salerno, Department of Pediatrics, University “Federico II,” Naples, Italy; Anna Simon, Department of Pediatrics, Christian Medical College, Vellore, India; Gabor Szinnai, Pediatric Endocrinology and Diabetology, University Children's Hospital Basel, University of Basel, Basel, Switzerland; Toshihiro Tajima, Department of Pediatrics, Hokkadio University School of Medecine, Sapporo, Japan; and Paul van Trotsenburg, Department of Paediatric Endocrinology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands. This article is simultaneously published in Hormone Research in Paediatrics (DOI: 10.1159/000358198). Abbreviations: CCHcentral CH CHcongenital hypothyroidism DBSdried blood spot FT4free T4 GAgestational age HrQOLhealth-related quality of life LBWlow birth weight L-T4levothyroxine NICUneonatal intensive care units NISsodium/iodide symporter T4thyroxine TFTthyroid function text TT4total T4 VLBWvery low birth weight.

This article is simultaneously published in Hormone Research in Paediatrics (DOI: 10.1159/000358198). Abbreviations: CCHcentral CH CHcongenital hypothyroidism DBSdried blood spot FT4free T4 GAgestational age HrQOLhealth-related quality of life LBWlow birth weight L-T4levothyroxine NICUneonatal intensive care units NISsodium/iodide symporter T4thyroxine TFTthyroid function text TT4total T4 VLBWvery low birth weight. Acknowledgments This work was supported by the European Society for Paediatric Endocrinology. The conference was hosted and partly funded by Istituto Superiore di Sanità of Italy. PerkinElmer Inc and Merck Serono S.A. also contributed educational grants but did not participate in the consensus and took no part in the preparation of this document. Data collection and analysis, data interpretation, and the decision to submit the paper for publication were the responsibility of the authors alone. Disclosure Summary: The authors have nothing to disclose.

Cardiovascular complications are a major cause of morbidity and mortality in patients with Cushing's syndrome (1–3). Patients with Cushing's syndrome are at increased risk of cardiovascular events, which does not fully normalize after remission (4–7). Increased blood pressure (BP), glucose intolerance or diabetes, central obesity, and metabolic syndrome (8) together with chronic hypokalemia (9) and a direct toxic effect of cortisol can all affect cardiac structure and function (2). Overt dilated cardiomyopathy with congestive heart failure is currently very rare, thanks to improved management of hypercortisolism and the use of antihypertensive drugs (eg, angiotensin converting enzyme inhibitors) with cardioprotective effects (10–12). However, subclinical structural and functional cardiac alterations are nearly always present but underdiagnosed.

t failure is currently very rare, thanks to improved management of hypercortisolism and the use of antihypertensive drugs (eg, angiotensin converting enzyme inhibitors) with cardioprotective effects (10–12). However, subclinical structural and functional cardiac alterations are nearly always present but underdiagnosed. To date, descriptions of cardiac structure and function in patients with Cushing's syndrome have been limited to two-dimensional (2D) echocardiography and restricted to the left ventricle (LV). A few studies have shown LV hypertrophy, concentric remodeling, and reduced systolic and diastolic performance (13–19). LV dysfunction was found to be associated with myocardial fibrosis in a single ultrasound (US) study (18) and was potentially reversed by normalization of the cortisol excess (17). One important limitation of these US studies is that the measurement of LV volumes and mass by 2D echocardiography requires assumptions to be made concerning LV geometry, introducing a source of inaccuracy and variability (20). In addition, patients with central obesity are especially prone to suboptimal image quality and off-axis images, which may limit the precision of echo-based calculations of ventricular mass and volumes.

aphy requires assumptions to be made concerning LV geometry, introducing a source of inaccuracy and variability (20). In addition, patients with central obesity are especially prone to suboptimal image quality and off-axis images, which may limit the precision of echo-based calculations of ventricular mass and volumes. In contrast, cardiac magnetic resonance imaging (CMR) provides a highly accurate and comprehensive assessment of cardiac geometry and function. CMR provides regional myocardial wall thickness, myocardial mass, and atrial and ventricular volumes based on precise delineation of myocardial borders without the need for geometric assumptions (21, 22). In addition, late gadolinium-enhanced CMR can depict dense myocardial fibrosis (23). The aim of this study was to characterize the consequences of cortisol excess on cardiac structure and function by means of CMR, comparing patients with Cushing's syndrome with healthy controls matched for age, gender, and body mass index (BMI), and evaluating the reversibility of cardiac abnormalities after the effective treatment of hypercortisolism.

cterize the consequences of cortisol excess on cardiac structure and function by means of CMR, comparing patients with Cushing's syndrome with healthy controls matched for age, gender, and body mass index (BMI), and evaluating the reversibility of cardiac abnormalities after the effective treatment of hypercortisolism. Materials and Methods Patients Patients were recruited in the Endocrinology Department of a tertiary referral center from September 2009 to March 2013. Patients aged 15–60 years were eligible if they had active endogenous Cushing's syndrome (newly diagnosed or uncontrolled after first surgery). Cushing's syndrome was diagnosed according to the usual clinical and biological criteria, including elevated urinary free cortisol (UFC) excretion, loss of the circadian plasma cortisol pattern, and lack of cortisol suppression during the overnight 1-mg dexamethasone suppression test (1). Healthy volunteers Healthy asymptomatic volunteers free of overt cardiovascular disease and risk factors for atherosclerosis, such as smoking, diabetes, dysplipidemia, and hypertension, were recruited by advertisement and matched with the patients for age, sex, and BMI. All the patients and volunteers gave their written informed consent, and the study protocol was approved by the Paris-Sud Ethics Committee (Le Kremlin-Bicêtre, France). All procedures conformed to the Declaration of Helsinki.

Healthy volunteers Healthy asymptomatic volunteers free of overt cardiovascular disease and risk factors for atherosclerosis, such as smoking, diabetes, dysplipidemia, and hypertension, were recruited by advertisement and matched with the patients for age, sex, and BMI. All the patients and volunteers gave their written informed consent, and the study protocol was approved by the Paris-Sud Ethics Committee (Le Kremlin-Bicêtre, France). All procedures conformed to the Declaration of Helsinki. Investigations Before cardiac investigations, the patients were admitted to the Endocrinology Department of our institution for clinical and biochemical investigations. BP (mean value of three measurements made at 2 min intervals after 30 min of rest in the sitting position) was determined with an automatic validated BP recorder (Press Mate BP 8800; Colin Co). The reported clinical BP is the mean value obtained in the 3 days preceding CMR. Plasma fasting glucose, triglycerides, total, low-density lipoprotein (LDL), and high-density lipoprotein (HDL) cholesterol, potassium and glycated hemoglobin (HBA1) were measured with standard procedures. In patients not known to have diabetes mellitus, glucose metabolism was assessed by a standard oral glucose tolerance test (OGTT). In the OGTT, the plasma glucose concentration was measured 2 hours after an oral intake of 75 g of glucose. Diabetes mellitus and impaired glucose tolerance were diagnosed according to the usual diagnostic criteria (24).

iabetes mellitus, glucose metabolism was assessed by a standard oral glucose tolerance test (OGTT). In the OGTT, the plasma glucose concentration was measured 2 hours after an oral intake of 75 g of glucose. Diabetes mellitus and impaired glucose tolerance were diagnosed according to the usual diagnostic criteria (24). The repeat CMR evaluation (second evaluation) was performed when successful treatment was achieved, with clinical and biological remission of hypercortisolism. Patients were considered in remission after treatment if they presented either with secondary adrenal insufficiency, with low morning serum cortisol concentrations and low UFC excretion rates or with eucortisolism, with normal UFC excretion rates (<65 μg per 24 hours or < 179 nmol per 24 h), normal overnight suppression of serum cortisol after the administration of 1 mg dexamethasone and normal sleeping midnight serum cortisol concentration (<2 μg/dL or 55 nmoL/L for both) (25). The period elapsed in the remission of hypercortisolism needed to be at least 2 months and no more than 12 months.

mol per 24 h), normal overnight suppression of serum cortisol after the administration of 1 mg dexamethasone and normal sleeping midnight serum cortisol concentration (<2 μg/dL or 55 nmoL/L for both) (25). The period elapsed in the remission of hypercortisolism needed to be at least 2 months and no more than 12 months. Assays Plasma glucose, potassium, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, and HBA1c were determined with standard analytical methods. UFC excretion was measured with a specific RIA using polyclonal antibodies as previously reported (26). Baseline UFC excretion is reported as the average of individual determinations in three consecutive daily urine samples. The normal range of UFC excretion is 10–65 μg per 24 hours (27–179 nmol per 24 hours). Serum cortisol concentrations were measured with a chemiluminescent competitive immunoassay using a polyclonal antibody as previously reported (26).

s the average of individual determinations in three consecutive daily urine samples. The normal range of UFC excretion is 10–65 μg per 24 hours (27–179 nmol per 24 hours). Serum cortisol concentrations were measured with a chemiluminescent competitive immunoassay using a polyclonal antibody as previously reported (26). Cardiac magnetic resonance imaging (MRI) All CMR examinations were performed in a tertiary academic referral hospital on a 1.5T (GEMS) MRI scanner using a dedicated torso 8-element phased array coil and electrocardiographic gating. All acquisitions were done with breath holding. Cine images were acquired using a steady-state free precession sequence and covered the heart in axial views (10–12 slice levels) and in cardiac short-axis views covering the entire right and left ventricles (10–12 slice levels) with the following parameters: echo time, 1.5 msec; repetition time, 3.5 msec; slice thickness, 8 mm; spacing, 10 mm; acquisition matrix, 224 × 192; spatial resolution, 0.78 mm; and number of phases, 40. Late gadolinium enhancement of the myocardium was studied with a T1-weighted 2D inversion recovery sequence on long- and short-axis end-diastolic views of the LV, 10–15 minutes after iv infusion of 0.2 mmol/kg of a gadolinated contrast agent (Gd-DTPA, Dotarem; Guerbet) with the following parameters: echo time, 1.5 msec; repetition time, 5.3 msec; slice thickness, 8 mm; spacing, 10 mm; acquisition matrix, 224 × 160; spatial resolution, 0.86 mm; and inversion time, 250 msec. These investigations were repeated in the same conditions during the patients' usual follow-up visits after effective treatment of Cushing's syndrome.

me, 1.5 msec; repetition time, 5.3 msec; slice thickness, 8 mm; spacing, 10 mm; acquisition matrix, 224 × 160; spatial resolution, 0.86 mm; and inversion time, 250 msec. These investigations were repeated in the same conditions during the patients' usual follow-up visits after effective treatment of Cushing's syndrome. Image analysis End-diastolic and end-systolic LV and right ventricular (RV) myocardial contours were generated on short-axis images by using semiautomated software (Qmass; MEDIS) by an experienced reader and were all validated by consensus after adjustment if necessary by an expert reader for all subjects. Endocardial contours were used to estimate ventricular end-diastolic and end-systolic volumes by the disk summation method. Similarly, end-diastolic LV myocardial contours were used to measure LV myocardial volume and, subsequently, LV mass (grams) as LV myocardial volume × myocardial density (1.06 g/mL). Left atrial end-diastolic and end-systolic endocardial contours were traced manually on axial cine images with the same method and used to measure left atrial (LA) volumes. LV and RV stroke volumes (milliliters) were then calculated as end-diastolic volume minus end-systolic volume, and ejection fractions were calculated (in percentages) as stroke volume/end-diastolic volume. Segmental wall thickness (millimeters) was measured as the mean segmental thickness on each American Heart Association segment (27) using epicardial and endocardial LV contours and the centerline method with 100 chords per slice level. Average wall thickness was calculated for the basal, mid-LV, and apical slices. LV concentricity was assessed by calculating the LV end-diastolic mass to volume ratio. Cardiac volumes and LV mass were indexed to BSA.

using epicardial and endocardial LV contours and the centerline method with 100 chords per slice level. Average wall thickness was calculated for the basal, mid-LV, and apical slices. LV concentricity was assessed by calculating the LV end-diastolic mass to volume ratio. Cardiac volumes and LV mass were indexed to BSA. Statistical analyses The LV and RV ejection fractions were used as the main measures of ventricular systolic performance. Continuous variables are expressed as the median and interquartile range. Differences between the controls and patients with active Cushing's syndrome were analyzed with the Mann-Whitney test. Treatment-induced changes in the patients are expressed as baseline-adjusted changes from baseline (pretreatment value minus posttreatment value/pretreatment value × 100%). The effects of Cushing's syndrome treatment were assessed with a Wilcoxon's rank-signed test. Univariate associations between clinical variables (age, gender, height, weight, BMI, systolic and diastolic BP) and biological variables (fasting glucose, HBA1c, UFC, morning plasma cortisol) and the LV, RV, and LA ejection fractions, the LV mass index and mean wall thickness were analyzed with the Spearman rank correlation, as were associations with treatment-induced changes in these parameters. To evaluate the influence of selected predictive variables on treatment-induced changes in the LV, RV, and LA ejection fractions, the LV mass index and mean wall thickness, we used multiple linear regression models with adjustment for age, gender, BMI, and systolic BP, testing each independent variable individually. Significance was assumed at P < .05. Stata 11SE (StataCorp) and SAS 9.1 software (SAS Institute) were used for statistical analyses.

the LV mass index and mean wall thickness, we used multiple linear regression models with adjustment for age, gender, BMI, and systolic BP, testing each independent variable individually. Significance was assumed at P < .05. Stata 11SE (StataCorp) and SAS 9.1 software (SAS Institute) were used for statistical analyses. Results Baseline clinical and biological characteristics Eighteen patients with active Cushing's syndrome (16 women and two men) with a median age of 35 years (range 15–59 y) were included. Fourteen patients had newly diagnosed Cushing's syndrome. Four patients had previously been treated by pituitary surgery and had developed recurrent hypercortisolism before entering the study. Fifteen patients had Cushing's disease, one had ectopic ACTH secretion, and two had cortisol-secreting adrenal adenomas. The patients' baseline clinical and hormonal characteristics are summarized in Supplemental Table 1. One patient (number 16) had New York Heart Association (NYHA) II-III dyspnea. The other patients had no symptoms of heart failure (NYHA I). Six patients were obese (BMI > 30 kg/m2), six were overweight (25 < BMI < 30 kg/m2), and six had a normal BMI (<25 kg/m2). Nine patients had arterial hypertension (World Health Organization/International Society of Hypertension definition), seven of whom were on antihypertensive treatment. Five patients were diabetic (World Health Organization/International Diabetes Foundation definition), of whom three required oral antidiabetic drugs. The OGTT revealed impaired glucose tolerance in another four patients. Five patients had hypertriglyceridemia and one had hypercholesterolemia. Three patients were on statin therapy and three patients required potassium supplementation.

al Diabetes Foundation definition), of whom three required oral antidiabetic drugs. The OGTT revealed impaired glucose tolerance in another four patients. Five patients had hypertriglyceridemia and one had hypercholesterolemia. Three patients were on statin therapy and three patients required potassium supplementation. As expected, sex, age, and BMI were similar in the patients and controls. Systolic and diastolic BPs were higher in the patients than in the controls (Table 1). Table 1. Baseline Clinical Parameters of the Control Subjects and Cushing's Syndrome Patients Controls (n = 18) Patients (n = 18) P Value Age, y 40.5 [24.2; 51.5] 35.0 [26.2; 47.2] NS Sex, female, % 13 (72%) 16 (89%) NS Body surface area, m2 1.73 [1.66; 1.96] 1.80 [1.68; 2.01] NS BMI, kg/m2 24.5 [21.8; 28.1] 27.5 [22.2; 33.0] NS Systolic BP, mm Hg 112 [108; 117] 120 [113; 130] .009 Diastolic BP, mm Hg 67 [64; 72] 77 [72; 88] <.001 Heart rate, beats/min 68 [64; 73] 72 [65; 83] NS Abbreviation: NS, not significant. Data are expressed as the median and interquartile range.

96] 1.80 [1.68; 2.01] NS BMI, kg/m2 24.5 [21.8; 28.1] 27.5 [22.2; 33.0] NS Systolic BP, mm Hg 112 [108; 117] 120 [113; 130] .009 Diastolic BP, mm Hg 67 [64; 72] 77 [72; 88] <.001 Heart rate, beats/min 68 [64; 73] 72 [65; 83] NS Abbreviation: NS, not significant. Data are expressed as the median and interquartile range. Treatment of Cushing's syndrome The second evaluation was performed a median of 6 months (range 2–12 mo) after correction of the glucocorticoid excess. Ten patients underwent transphenoidal surgery, four received steroidogenesis inhibitors, and one received both transphenoidal surgery and medical treatment. The patient with ectopic ACTH secretion was prepared with medical therapy before curative thoracic surgery. Both patients with cortisol-secreting adrenal adenomas were cured by unilateral adrenalectomy. Clinical and biological remission of hypercortisolism was obtained in all the patients (Table 2). Fourteen operated patients had secondary corticotroph deficiency after successful treatment of hypercortisolism and received effective hydrocortisone replacement therapy (20 mg daily). Table 2. Clinical, Biochemical, and Hormonal Parameters of Cushing's Syndrome Patients Before and After Successful Treatment

Treatment of Cushing's syndrome The second evaluation was performed a median of 6 months (range 2–12 mo) after correction of the glucocorticoid excess. Ten patients underwent transphenoidal surgery, four received steroidogenesis inhibitors, and one received both transphenoidal surgery and medical treatment. The patient with ectopic ACTH secretion was prepared with medical therapy before curative thoracic surgery. Both patients with cortisol-secreting adrenal adenomas were cured by unilateral adrenalectomy. Clinical and biological remission of hypercortisolism was obtained in all the patients (Table 2). Fourteen operated patients had secondary corticotroph deficiency after successful treatment of hypercortisolism and received effective hydrocortisone replacement therapy (20 mg daily). Table 2. Clinical, Biochemical, and Hormonal Parameters of Cushing's Syndrome Patients Before and After Successful Treatment Patients P Value Before T (n = 18) After T (n = 17) Clinical parameters Body surface area, m2 1.80 [1.68; 2.01] 1.77 [1.69; 1.98] NS BMI, kg/m2 27.5 [22.2; 33.0] 25.5 [21.0; 30.3] .020 Systolic BP, mm Hg 120 [113; 130] 114 [100; 123] .029 Diastolic BP, mm Hg 77 [72; 88] 72 [63; 79] .020 Heart rate, beats/min 72 [65; 83] 64 [61; 71] NS Biochemical parameters (plasma) Fasting glucose, mmol/L 4.80 [4.45; 5.50] 4.30 [4.05; 4.50] .002 Glucose 2 h after OGTT, mmol/L 8.75 [6.85; 10.5] 5.50 [4.60; 7.40] .027 HBA1c, % 6.00 [5.80; 6.40] 5.45 [5.20; 5.82] .004 Triglycerides, mmol/L 1.06 [0.67; 1.83] 1.00 [0.83; 1.99] NS Total cholesterol, mmol/L 5.39 [4.68; 5.75] 5.34 [4.66; 6.02] NS LDL cholesterol, mmol/L 3.02 [2.48; 3.43] 3.24 [2.54; 3.95] NS HDL cholesterol, mmol/L 1.67 [1.53; 1.95] 1.36 [1.23; 1.61] .029 K, mmol/L 4.00 [3.67; 4.30] 4.00 [3.82; 4.40] NS Hormonal parameters UFC, μg per 24 h 373 [178; 807] 6.0 [5.0; 19.0] <.001 Morning plasma cortisol, μg/dL 24.6 [19.4; 33.2] 4.8 [1.3; 8.3] .002 Midnight plasma cortisol, μg/dL 18.1 [14.0; 22.5] 1.1 [1.0; 1.9] .002 Abbreviations: K, potassium; T, treatment. Data are expressed as the median and interquartile range.

40] NS Hormonal parameters UFC, μg per 24 h 373 [178; 807] 6.0 [5.0; 19.0] <.001 Morning plasma cortisol, μg/dL 24.6 [19.4; 33.2] 4.8 [1.3; 8.3] .002 Midnight plasma cortisol, μg/dL 18.1 [14.0; 22.5] 1.1 [1.0; 1.9] .002 Abbreviations: K, potassium; T, treatment. Data are expressed as the median and interquartile range. Data for 17 patients were included in the analyses of the response to treatment. Patient 18 was excluded from further investigations because of pregnancy. In addition, data for the subgroup of 13 patients with surgical remission and secondary corticotroph deficiency also were analyzed separately. The impact of correction of the glucocorticoid excess on clinical and biological parameters is summarized in Table 2. BMI and systolic and diastolic BPs fell significantly after treatment, whereas the heart rate was unchanged. Antihypertensive therapy could be discontinued in four patients and reduced in another two. Fasting glucose levels, glucose levels 2 hours after the OGTT, and HBA1c levels also fell after treatment. Diabetes mellitus and impaired glucose tolerance reversed in all treated patients. Triglyceride, total cholesterol, and LDL cholesterol levels did not change, whereas HDL cholesterol levels fell after treatment. Statins could be discontinued in one patient.

rs after the OGTT, and HBA1c levels also fell after treatment. Diabetes mellitus and impaired glucose tolerance reversed in all treated patients. Triglyceride, total cholesterol, and LDL cholesterol levels did not change, whereas HDL cholesterol levels fell after treatment. Statins could be discontinued in one patient. Baseline cardiac parameters in patients and controls LV, RV, and LA parameters in the patients and controls are compared in Table 3. Patients and controls had similar LV end-diastolic volumes, whereas LV end-systolic volumes were higher in the patients. Consequently, the patients had lower LV ejection fractions (P < .001) and tended to have a smaller LV stroke volume index (P = .08) (Figure 1). In addition, compared with controls, patients had markedly increased end-diastolic LV segmental thickness in the basal, mid-LV and apical slices (all P < .001), and also a trend toward higher LV mass (Figure 2). Only one patient (number 16) had LV hypertrophy according to conventional CMR thresholds defining LV hypertrophy (28). At baseline, controls and patients had similar LV mass to end-diastolic volume ratio (Figure 2). Differences in RV parameters between controls and patients at baseline were similar to the differences in LV parameters. RV end-diastolic volumes were similar in the patients and controls, but the patients had higher RV end-systolic volumes (P = .012), lower RV stroke volumes (P = .033), and lower RV ejection fractions (P < .001) (Figure 1). Maximal LA volumes were similar in the patients and controls, but minimal LA volumes after atrial contraction were significantly larger in patients than controls at baseline (P = .042). Consequently, LA ejection fractions (Figure 1) were markedly lower in patients than controls (P < .001). In a univariate analysis, the patients' baseline CMR LV and RV function estimates did not correlate with their clinical or biological variables. The LA ejection fraction was associated only with elevated fasting glucose (r = 0.72, P = .043), and the LV mass index was associated only with systolic BP (r = 0.72, P = .045). No delayed myocardial enhancement was observed.

MR LV and RV function estimates did not correlate with their clinical or biological variables. The LA ejection fraction was associated only with elevated fasting glucose (r = 0.72, P = .043), and the LV mass index was associated only with systolic BP (r = 0.72, P = .045). No delayed myocardial enhancement was observed. Table 3. LA, LV, and RV Parameters Assessed by CMR in Control Subjects and in Cushing's Syndrome Patients Before and After Treatment

MR LV and RV function estimates did not correlate with their clinical or biological variables. The LA ejection fraction was associated only with elevated fasting glucose (r = 0.72, P = .043), and the LV mass index was associated only with systolic BP (r = 0.72, P = .045). No delayed myocardial enhancement was observed. Table 3. LA, LV, and RV Parameters Assessed by CMR in Control Subjects and in Cushing's Syndrome Patients Before and After Treatment Controls (n = 18) Patients Pre T (n = 18) Patients Post T (n = 17) P Value Controls vs Pre T Patients P Value Effect of T Left ventricular parameters End-diastolic volume index, mL/m2 67.9 [52.5; 77.6] 68.8 [62.5; 88.8] 71.7 [61.1; 87.9] NS NS End-systolic volume index, mL/m2 22.4 [16.7; 29.4] 31.8 [24.2; 37.4] 31.3 [21.4; 36.6] .002 NS Stroke volume index, mL/m2 44.8 [37.5; 49.7] 34.9 [31.4; 46.4] 42.3 [37.3; 52.6] .079 .042 Ejection fraction, % 64.3 [60.9; 69.6] 55.0 [51.0; 58.4] 57.8 [53.3; 66.7] <.001 .029 Mass index, g/m2 53.6 [50.1; 62.1] 59.1 [53.8; 69.0] 49.1 [39.8; 58.5] .085 <.001 Mass to end-diastolic volume, g/mL 0.81 [0.74; 0.87] 0.86 [0.80; 0.92] 0.62 [0.53; 0.88] NS .002 Basal wall thickness, mm 7.99 [6.78; 8.61] 11.2 [8.96; 12.2] 6.30 [5.52; 7.87] <.001 <.001 Midventricular wall thickness, mm 7.26 [6.33; 8.21] 10.3 [8.43; 11.3] 5.97 [5.54; 8.28] <.001 <.001 Apical wall thickness, mm 5.90 [5.16; 6.68] 8.45 [7.86; 9.52] 4.90 [4.61; 7.07] <.001 <.001 Right ventricular parameters End-diastolic volume index, mL/m2 69.6 [60.1; 80.0] 73.7 [62.8; 84.7] 83.5 [61.5; 91.1] NS NS End-systolic volume index, mL/m2 27.1 [21.6; 35.3] 36.6 [32.6; 42.5] 39.2 [28.6; 46.1] .012 NS Stroke volume index, mL/m2 41.9 [37.1; 48.3] 34.1 [26.9; 40.3] 42.2 [35.1; 51.5] .033 .029 Ejection fraction, % 60.0 [54.6; 64.3] 49.3 [42.5; 53.4] 51.1 [44.3; 55.6] <.001 NS Left atrial parameters Maximal volume index, mL/m2 37.7 [29.4; 48.5] 33.2 [27.6; 39.2] 33.8 [31.0; 42.4] NS NS Minimal volume index, mL/m2 15.6 [13.1; 20.7] 20.8 [15.5; 24.7] 15.8 [13.9; 21.2] .042 .004 Ejection fraction, % 56.1 [48.6; 61.2] 39.2 [34.8; 48.4] 52.0 [50.5; 55.2] <.001 <.001 Abbreviations: Pre T, before treatment; post T, after treatment; T, treatment. Data are expressed as the median and interquartile range.

S NS Minimal volume index, mL/m2 15.6 [13.1; 20.7] 20.8 [15.5; 24.7] 15.8 [13.9; 21.2] .042 .004 Ejection fraction, % 56.1 [48.6; 61.2] 39.2 [34.8; 48.4] 52.0 [50.5; 55.2] <.001 <.001 Abbreviations: Pre T, before treatment; post T, after treatment; T, treatment. Data are expressed as the median and interquartile range. Figure 1. Comparison of LV ejection fractions (A), LV stroke volume indexes (B), RV ejection fractions (C), RV stroke volume indexes (D), and LA ejection fractions between control subjects and Cushing's syndrome patients before and after treatment. Pre T, before treatment; post T, after treatment. Figure 2. Comparison of LV mass indexes (A), LV mass to end-diastolic volume ratios (B), basal wall thickness (C), midventricular wall thickness (D), and apical wall thickness (E) in control subjects and Cushing's syndrome patients before and after treatment. Pre T, before treatment; post T, after treatment.

Figure 1. Comparison of LV ejection fractions (A), LV stroke volume indexes (B), RV ejection fractions (C), RV stroke volume indexes (D), and LA ejection fractions between control subjects and Cushing's syndrome patients before and after treatment. Pre T, before treatment; post T, after treatment. Figure 2. Comparison of LV mass indexes (A), LV mass to end-diastolic volume ratios (B), basal wall thickness (C), midventricular wall thickness (D), and apical wall thickness (E) in control subjects and Cushing's syndrome patients before and after treatment. Pre T, before treatment; post T, after treatment. Changes in cardiac parameters after treatment of Cushing's syndrome Changes in LV, RV, and LA parameters after the treatment of Cushing's syndrome are presented in Table 3. Treatment did not significantly affect LV end-diastolic or end-systolic volumes. LV systolic function improved after treatment, as reflected by, respectively, 19% and 15% increases in LV stroke volume index (P = .042) and LV EF (P = .029) (Figure 1). Treatment of cortisol excess was also associated with a 17% reduction in the LV mass index (P < .001). Segmental myocardial wall thickness decreased by 37%, 34%, and 35% at the basal, mid-LV, and apical levels (all P < .001), respectively. LV mass/end-diastolic volume decreased by 10% (P < .001), pointing to a more eccentric LV geometry, because the decrease in LV mass was associated with a trend toward an increase in end-diastolic volume (Figure 2). Treatment-induced changes in RV parameters were similar to changes in LV parameters. RV end-diastolic volumes and end-systolic volumes were not affected by treatment, but RV stroke volumes and the RV EF increased by 29% (P = .033) and 11% (P = .16), respectively, reflecting the improvement in RV systolic function (Figure 1). Maximal LA volumes were unchanged by treatment, but minimal LA volumes fell by 16% (P = .004) and, consequently, the LA ejection fraction increased by 45% (P < .001) (Figure 1). The results for the subgroup of 13 patients with surgical remission and secondary corticotroph deficiency were similar to those found for all 17 patients, except for the increase in RV stroke volumes that did not reach statistical significance (P = .13) when the four nonsurgically treated patients were excluded from the analysis. The effects of Cushing's syndrome treatment on ventricular function and structure, as assessed by CMR in patient 16, are illustrated in Figure 3.

cept for the increase in RV stroke volumes that did not reach statistical significance (P = .13) when the four nonsurgically treated patients were excluded from the analysis. The effects of Cushing's syndrome treatment on ventricular function and structure, as assessed by CMR in patient 16, are illustrated in Figure 3. Figure 3. Steady-state free precession cine end-diastolic and end-systolic short- and long-axis images before and after treatment of Cushing's syndrome in patient 16 compared with a matched control subject. Note the global LV wall thickening without end-diastolic LV or RV dilatation and the decreased biventricular ejection in this symptomatic patient (NYHA II-III) prior to treatment and the clear reversal of this phenotype after treatment.

atment of Cushing's syndrome in patient 16 compared with a matched control subject. Note the global LV wall thickening without end-diastolic LV or RV dilatation and the decreased biventricular ejection in this symptomatic patient (NYHA II-III) prior to treatment and the clear reversal of this phenotype after treatment. In a univariate analysis, treatment-induced changes in the LV, RV, and LA ejection fraction did not correlate with changes in any of the clinical or biological parameters, except for an association between increased RV ejection fractions and decreased systolic BP (r = 0.68, P = .042). The treatment-induced decrease in the LV mass index correlated only with the decrease in blood glucose (r = −0.82, P = .007). In a multivariate analysis with adjustment for age, gender, BMI, and systolic BP, the only individual variable significantly associated with the increase in the LV ejection fraction was the decrease in mean wall thickness (r2 = 0.77, P = .030). Independent correlates of the decrease in the LV mass index included decreased fasting glucose (r2 = 0.82, P < .001), decreased BMI (r2 = 0.61, P = .017), and decreased mean wall thickness (r2 = 0.60, P = .048). None of the patients had late gadolinium enhancement of the myocardium on follow-up CMR.

= 0.77, P = .030). Independent correlates of the decrease in the LV mass index included decreased fasting glucose (r2 = 0.82, P < .001), decreased BMI (r2 = 0.61, P = .017), and decreased mean wall thickness (r2 = 0.60, P = .048). None of the patients had late gadolinium enhancement of the myocardium on follow-up CMR. Discussion This is the first CMR study of cardiac structure and function in Cushing's syndrome. We made three important findings. First, compared with controls, patients with Cushing's syndrome had a subclinical decrease in biventricular systolic performance, associated with increased LV mass and LA dysfunction. Second, effective treatment of hypercortisolism improved the systolic performance of both ventricles and LA, reduced LV mass, and LV wall thickness and led to the regression of the concentric LV remodeling pattern. The treatment-related decrease in LV mass was independently associated with changes in glucose metabolism and BMI. Finally, late gadolinium enhancement of the myocardium was absent in all patients, suggesting the absence of dense replacement myocardial fibrosis in uncomplicated Cushing's syndrome.

V remodeling pattern. The treatment-related decrease in LV mass was independently associated with changes in glucose metabolism and BMI. Finally, late gadolinium enhancement of the myocardium was absent in all patients, suggesting the absence of dense replacement myocardial fibrosis in uncomplicated Cushing's syndrome. This CMR study confirms and extends echocardiographic findings on LV systolic and diastolic dysfunction in patients with in Cushing's syndrome. Indeed, several echocardiographic studies have evaluated cardiac structure and function in patients with Cushing's syndrome, either in the active phase (13–16) or both before and after treatment in adult (17–19) and pediatric patients (29). However, echo-based evaluation of ventricular mass and volumes requires geometric assumptions that may limit its accuracy, especially in case of central obesity, which is frequent in Cushing's syndrome patients. Furthermore, US measurement of RV and LA volumes is challenging, and these studies almost exclusively focused on the LV (13–19). Here the cardiac consequences of cortisol excess were analyzed by CMR, allowing highly accurate, noninvasive assessment of cardiac geometry and function. In addition to LV parameters, we were able to analyze RV and LA structure and function. Patients with active Cushing's syndrome were compared with age-, sex-, and BMI-matched control subjects. The reversibility of cardiac abnormalities was evaluated after the effective treatment of hypercortisolism.

metry and function. In addition to LV parameters, we were able to analyze RV and LA structure and function. Patients with active Cushing's syndrome were compared with age-, sex-, and BMI-matched control subjects. The reversibility of cardiac abnormalities was evaluated after the effective treatment of hypercortisolism. One of our main findings is that LV and RV systolic performance is diminished in active Cushing's syndrome and that it improves after successful treatment of hypercortisolism. Compared with controls, patients had lower LV and RV ejection fractions, which increased by 15% and 11%, respectively, after successful treatment. End-diastolic volumes were similar in the patients (both before and after treatment) and controls, ruling out marked preload differences (30). Previous echo-based studies showed no changes in LV volumes or ejection fraction but revealed subclinical alteration of LV systolic performance based on calculated parameters requiring geometric assumptions, such as decreased LV endocardial and midwall fractional shortening (14, 19). Recently decreased LV circumferential and longitudinal strain was found by 2D speckle tracking imaging (17, 18). It is noteworthy that, except for patient 16, the patients in this study were all asymptomatic (NYHA I). Subclinical biventricular systolic dysfunction may therefore be largely underestimated in patients with Cushing's syndrome. Careful evaluation of cardiac structure and function by means of high-precision quantitative imaging in a specialized setting is thus important for appropriate cardiovascular management of Cushing's syndrome patients.

ular systolic dysfunction may therefore be largely underestimated in patients with Cushing's syndrome. Careful evaluation of cardiac structure and function by means of high-precision quantitative imaging in a specialized setting is thus important for appropriate cardiovascular management of Cushing's syndrome patients. In addition to biventricular systolic dysfunction, patients with Cushing's syndrome had a markedly lower LA ejection fraction than controls, pointing to diastolic LV dysfunction. This parameter showed a marked improvement after treatment. Diastolic LV dysfunction has previously been reported in terms of abnormal transmitral flow velocity patterns on Doppler US (14–16, 19) and as decreased diastolic myocardial strain rates on 2D speckle tracking (17, 18). This finding is important, given the adverse prognostic value of dilated LA, which could potentially promote atrial fibrillation, stroke, and heart failure in iatrogenic hypercortisolism (3).

velocity patterns on Doppler US (14–16, 19) and as decreased diastolic myocardial strain rates on 2D speckle tracking (17, 18). This finding is important, given the adverse prognostic value of dilated LA, which could potentially promote atrial fibrillation, stroke, and heart failure in iatrogenic hypercortisolism (3). This study further demonstrates that cortisol excess is associated with increased regional LV wall thickness. However, no significant increase in total LV mass was found, probably owing to the small number of patients we were able to study, this being a rare disease. The LV mass to end-diastolic volume ratio remained close to young healthy control values and lower than values associated with an adverse prognosis in the asymptomatic general population (30). However, treatment of Cushing's syndrome led to a 17% reduction in the LV mass index, associated with a noteworthy decrease in LV segmental myocardial wall thickness and a decrease in the LV mass to end diastolic volume ratio (28). Previous echo-based studies of Cushing's syndrome patients also showed an increased relative myocardial wall thickness of the LV (13) and an increased LV mass index (14, 17–19). LV hypertrophy was present in 24%–42% of patients in some of these studies, but echocardiographic assessment may overestimate these abnormalities. In our study, only one patient (number 16) had LV hypertrophy according to the conventional CMR threshold (28). Accurate evaluation of LV mass and remodeling are highly relevant for risk stratification and therapeutic management of patients with Cushing's syndrome because increased LV mass and LV concentric remodeling are independent risk factors for cardiovascular events, incident heart failure, and sudden death (30, 31).

(28). Accurate evaluation of LV mass and remodeling are highly relevant for risk stratification and therapeutic management of patients with Cushing's syndrome because increased LV mass and LV concentric remodeling are independent risk factors for cardiovascular events, incident heart failure, and sudden death (30, 31). The pathophysiology of cardiac hypertrophy and decreased systolic function in Cushing's syndrome remains unclear. The myocardium seems to be exempt from the generalized muscular atrophy seen in Cushing's syndrome, resulting from increased protein catabolism (1). Arterial hypertension, present in up to 75% of patients with Cushing's syndrome (1, 5), may partly explain the increased LV mass, but hypertrophy has also been described in patients with normal BP (13, 17, 19). In our study the LV mass index was related to baseline systolic BP, but no statistically significant correlation was found between the treatment-induced changes in LV mass index and BP. Other Cushing's syndrome-related cardiovascular risk factors such as visceral obesity, glucose intolerance, and dyslipidemia (1, 5) may also contribute to the elevated LV mass index. Indeed, we found that treatment-induced changes in plasma fasting glucose and BMI were independent correlates of the change in the LV mass index. However, patients with Cushing's syndrome have impaired systolic performance that is not observed in obese subjects, who have increased stroke volumes, an unchanged LV ejection fraction, and significantly higher LV mass values than Cushing patients (32). Lastly, the cortisol excess per se may exert a toxic effect on the heart, mediated directly through glucocorticoid and/or mineralcocorticoid receptors (33, 34), and indirectly by inducing the expression of adrenergic receptors (35).

ejection fraction, and significantly higher LV mass values than Cushing patients (32). Lastly, the cortisol excess per se may exert a toxic effect on the heart, mediated directly through glucocorticoid and/or mineralcocorticoid receptors (33, 34), and indirectly by inducing the expression of adrenergic receptors (35). Decreased systolic performance in Cushing's syndrome points to ultrastructural abnormalities in heart muscle, altering its contractility. Recently increased myocardial fibrosis, indirectly assessed by calibrated integrated backscatter, has been described in patients with Cushing's syndrome. Fibrosis was associated with systolic and diastolic dysfunction and regressed after treatment of hypercortisolism (18). In the present study, however, late gadolinium-enhanced CMR, the reference method for the noninvasive assessment of dense myocardial fibrosis (23), revealed no focal intramyocardial fibrosis or myocardial infarction scar either at baseline or during the follow-up in any of the patients. However, this does not preclude the presence of diffuse interstitial fibrosis. Cardiac steatosis, documented by magnetic resonance spectroscopy in patients with impaired glucose tolerance or diabetes mellitus (36), could also contribute to impairing the myocardial contractility in Cushing's syndrome patients, who frequently have visceral obesity and diabetes mellitus.

ffuse interstitial fibrosis. Cardiac steatosis, documented by magnetic resonance spectroscopy in patients with impaired glucose tolerance or diabetes mellitus (36), could also contribute to impairing the myocardial contractility in Cushing's syndrome patients, who frequently have visceral obesity and diabetes mellitus. The heterogeneity of the treatments used to control Cushing's syndrome is a potential limitation of this study. However, the results of reevaluation analysis did not chance substantially after exclusion of patients controlled only by medical therapy. Furthermore, we cannot rule out a potential impact of hydrocortisone replacement on the cardiac structure and function in patients with surgical remission and secondary corticotroph insufficiency. In conclusion, this CMR study shows that patients with Cushing's syndrome frequently have significant subclinical biventricular and LA systolic dysfunction and increased LV mass, in the absence of dense myocardial fibrosis. Successful treatment of the cortisol excess improves systolic ventricular and atrial performance and leads to regression of the LV mass and myocardial thickness, in parallel with an improvement in glucose metabolism and BMI. Structural and functional abnormalities of the myocardium occur early in Cushing's syndrome and should be actively investigated to guide strategies designed to prevent overt heart failure. Although endogenous Cushing's syndrome is a rare disease, our findings may also have implications for patients receiving long-term treatment with exogenous corticosteroids, a common clinical situation.

in Cushing's syndrome and should be actively investigated to guide strategies designed to prevent overt heart failure. Although endogenous Cushing's syndrome is a rare disease, our findings may also have implications for patients receiving long-term treatment with exogenous corticosteroids, a common clinical situation. Abbreviations: BMIbody mass index BPblood pressure CMRcardiac magnetic resonance 2Dtwo dimensional HBA1cglycated hemoglobin HDLhigh-density lipoprotein LAleft atrium LDLlow-density lipoprotein LVleft ventricle MRImagnetic resonance imaging NYHANew York Heart Association OGTToral glucose tolerance test RVright ventricle UFCurinary free cortisol USultrasound. Acknowledgments We thank the nursing staff of the Endocrinology Department of the Bicêtre Hospital and the imaging technicians of the Cardiovascular Radiology Department of the European George Pompidou Hospital who performed the experimental studies. P.K. is recipient of a Contrat d'Interface grant from INSERM. Disclosure Summary: The authors have nothing to declare.

Mutations that disrupt signaling through the melanocortin 4 receptor (MC4R) represent the commonest highly penetrant genetic form of obesity, being found in 2–5% of severely obese individuals (1). MC4Rs are widely expressed in the hypothalamus, brainstem, and other brain regions, where they mediate the anorectic response to the adipocyte-derived hormone leptin and the satiety response to gut hormones such as peptide YY and ghrelin (2). MC4Rs are also expressed in dopamine-rich regions of the striatum (3), and there is a growing body of evidence in rodents to suggest that melanocortin signaling modulates food reward (4). To investigate the impact of genetic disruption of central melanocortin signaling on the brain response to anticipatory food reward in humans, we studied obese individuals with heterozygous mutations in MC4R that completely disrupt melanocortin signaling in cells and are associated with increased food intake (1). We used functional magnetic resonance imaging (fMRI) to measure blood oxygen level-dependent responses to images of highly palatable, appetizing foods, bland foods, and non-food objects in eight MC4R-deficient individuals and in 10 equally obese and eight lean controls in whom MC4R mutations had been excluded.

ake (1). We used functional magnetic resonance imaging (fMRI) to measure blood oxygen level-dependent responses to images of highly palatable, appetizing foods, bland foods, and non-food objects in eight MC4R-deficient individuals and in 10 equally obese and eight lean controls in whom MC4R mutations had been excluded. Subjects and Methods Participants MC4R mutation carriers were identified by direct nucleotide sequencing (4). We identified eight MC4R-deficient patients with heterozygous mutations that had previously been shown to result in a complete loss of function in vitro by measuring cAMP production (1). MC4R mutations were excluded in the two controls groups—10 overweight/obese controls and eight lean controls. Details on the study participants are given in Tables 1 and 2. All subjects were weight stable for at least 3 months, were not taking any medication (healthy volunteers), were right-handed, and had no history of psychiatric disease. The study was approved by the Local Regional Ethics Committee in Suffolk. All subjects provided written informed consent. Motivational state (eg, hunger) modulates responses to food-related stimuli; we therefore studied all subjects in the satiated state. All participants received a standardized breakfast and lunch before the study at fixed times. Meals were calculated to provide 20 and 35% of daily energy requirements, respectively. Energy requirements were calculated with the Schofield formula (5), which allows for the different energy requirements of lean and obese individuals. The macronutrient content was 50% carbohydrate, 30% fat, and 20% protein. The fMRI scan was performed exactly 1 hour after consumption of the lunch in all instances. Hunger or fullness visual analog scores taken just before participants went into the fMRI scan were not different between the three groups (lean controls—hunger, 2.0 ± 0.4; fullness, 6.5 ± 0.7; obese controls—hunger, 2.7 ± 0.6; fullness, 5.4 ± 0.8; MC4R—hunger, 2.4 ± 0.5; fullness, 5.6 ± 0.7; hunger, P = .624; fullness, P = .551).

llness visual analog scores taken just before participants went into the fMRI scan were not different between the three groups (lean controls—hunger, 2.0 ± 0.4; fullness, 6.5 ± 0.7; obese controls—hunger, 2.7 ± 0.6; fullness, 5.4 ± 0.8; MC4R—hunger, 2.4 ± 0.5; fullness, 5.6 ± 0.7; hunger, P = .624; fullness, P = .551). Table 1. Characteristics of Groups Group Gender, M/F Age Range, y BMI (mean ± SEM), kg/m2 MC4R-deficient participants 5/3 18–47 34.5 ± 7.3; range 26–44 Obese participants 4/6 31–50 33.2 ± 5.4; range 28–41 Lean participants 2/6 19–54 22.4 ± 1.9; range 21–25 Table 2. Mutations of Subjects

llness visual analog scores taken just before participants went into the fMRI scan were not different between the three groups (lean controls—hunger, 2.0 ± 0.4; fullness, 6.5 ± 0.7; obese controls—hunger, 2.7 ± 0.6; fullness, 5.4 ± 0.8; MC4R—hunger, 2.4 ± 0.5; fullness, 5.6 ± 0.7; hunger, P = .624; fullness, P = .551). Table 1. Characteristics of Groups Group Gender, M/F Age Range, y BMI (mean ± SEM), kg/m2 MC4R-deficient participants 5/3 18–47 34.5 ± 7.3; range 26–44 Obese participants 4/6 31–50 33.2 ± 5.4; range 28–41 Lean participants 2/6 19–54 22.4 ± 1.9; range 21–25 Table 2. Mutations of Subjects Subject Mutation 1 I125K, related to subjects 5 and 8 2 Y35X;D37V, related to subject 3 3 Y35X;D37V, related to subject 2 4 G252S 5 I125K, related to subjects 1 and 8 6 R236C 7 R165W 8 I125K, related to subjects 1 and 5 Experimental design While lying in the scanner, participants viewed images of appetizing (eg, chocolate cake), disgusting (eg, rotten meat), and bland foods (eg, uncooked rice), plus non-food household objects (30 unique items in each category). To control for the possibility that areas engaged by appetizing foods reflect increased emotional arousal, we included a disgusting food category matched to the appetizing foods on rated arousal (6). Images of appetizing and disgusting food images were preselected to be equal in rated arousal (6). Color images were selected from the International Affective Picture Series, supplemented by copyright-unrestricted images obtained from the internet. All images were cropped to be of the same size. During the fMRI experiment, stimuli were presented in 18-second blocks, with each block containing 10 images from the same category. Each image was displayed for 1300 milliseconds, followed by a 500-millisecond fixation screen. The four block types were presented in random order, with nine blocks for each category. Stimuli were viewed via an angled mirror above the participants' eyes, which reflected images back-projected from a translucent screen positioned in the bore of the magnet to the rear of the participants' head. Images were shifted randomly, slightly to the left or right of the center of the screen (1.7° visual angle). Participants were asked to respond by button press whether the image was presented to the left or right of the center of the screen.

en positioned in the bore of the magnet to the rear of the participants' head. Images were shifted randomly, slightly to the left or right of the center of the screen (1.7° visual angle). Participants were asked to respond by button press whether the image was presented to the left or right of the center of the screen. fMRI data acquisition Functional imaging data were acquired using a 3T Tim Trio (Siemens) scanner. Whole-brain T2*-weighted echo planar images (EPIs) were acquired with a repetition time of 2000 milliseconds, echo time of 30 milliseconds, flip angle of 78º, and 32 axial oblique slices with 3-mm isotropic resolution. A total of 336 volumes were acquired, for a total imaging time of 11 minutes 12 seconds. A high-resolution structural MP-RAGE scan for normalization purposes was also acquired (voxel size, 1 × 1×1 mm; repetition time, 2250 ms; echo time, 2.98 ms; inversion time, 900 ms; flip angle, 9°; total scan time, 4 min 16 s).

on. A total of 336 volumes were acquired, for a total imaging time of 11 minutes 12 seconds. A high-resolution structural MP-RAGE scan for normalization purposes was also acquired (voxel size, 1 × 1×1 mm; repetition time, 2250 ms; echo time, 2.98 ms; inversion time, 900 ms; flip angle, 9°; total scan time, 4 min 16 s). fMRI analysis Data were analyzed using SPM5 (Wellcome Trust Centre for Neuroimaging, www.fil.ion.ucl.ac.uk/spm). The first six volumes were discarded to allow for equilibration effects. The EPIs were corrected for slice time differences and realigned to the first scan by rigid body transformations to correct for head movements. EPI and structural scans were coregistered and normalized to the T1 standard template in Montreal Neurological Institute space (MNI-International Consortium for Brain Mapping) and smoothed with a Gaussian kernel of 8 mm FWHM. Data were analyzed using general linear models within SPM5. At the first level, condition effects were estimated using boxcar regressors convolved with the canonical hemodynamic response function, and movement parameters were included as regressors to account for residual movement-related variance. Data were high-pass filtered (128 s) to remove low-frequency signal drift. Statistical parametric maps were then generated for each individual subject by estimating activation contrasts between conditions (eg, appetizing vs bland). To determine group differences at the second level, we set up an ANOVA at the whole-brain level for each contrast and conducted a region-of-interest (ROI) analysis. Based on strong evidence for the involvement of dorsal and ventral striatum in food-reward processing, and more specifically in obesity, our primary ROIs were 8-mm radius spheres centered on the caudate/putamen (dorsal striatum) (x = −15, y = 18, z = 12) (7) and ventral striatum (x = −8, y = 16, z = −12) (6). Corrections for multiple comparisons were performed by small volume correction using a Bonferroni correction (family-wise error).

ally in obesity, our primary ROIs were 8-mm radius spheres centered on the caudate/putamen (dorsal striatum) (x = −15, y = 18, z = 12) (7) and ventral striatum (x = −8, y = 16, z = −12) (6). Corrections for multiple comparisons were performed by small volume correction using a Bonferroni correction (family-wise error). An additional exploratory analysis of other brain regions that have also been implicated in reward and in obesity, including amygdala, orbitofrontal cortex, and somatosensory cortex, as well as the insula/frontal operculum in disgust processing, was performed using the more lenient threshold of P < .001 uncorrected. For completeness, we also performed a whole-brain analysis for all contrasts at the uncorrected P < .001 level, with an extent threshold of 10 voxels (data shown in Supplemental Table 1). Significant or borderline group main effects in our ROIs were followed up with post hoc t tests between each of the groups.

rrected. For completeness, we also performed a whole-brain analysis for all contrasts at the uncorrected P < .001 level, with an extent threshold of 10 voxels (data shown in Supplemental Table 1). Significant or borderline group main effects in our ROIs were followed up with post hoc t tests between each of the groups. Results Images of appetizing foods shown to individuals undergoing fMRI scanning are associated with the activation of specific brain regions involved in a reward network that includes the striatum, amygdala, ventral tegmental area, and orbitofrontal and prefrontal cortex, which are the source and target of dopaminergic neurons that mediate food reward (8). In view of the limited number of MC4R-deficient individuals who conformed to our strict inclusion criteria and were thus available for this study, we performed a ROI analysis. Based on evidence for the involvement of dorsal and ventral striatum in food-reward processing, our primary regions of interest were the dorsal and ventral striatum. We first established group differences in the neural response to viewing appetizing foods in the dorsal and ventral striatum.

we performed a ROI analysis. Based on evidence for the involvement of dorsal and ventral striatum in food-reward processing, our primary regions of interest were the dorsal and ventral striatum. We first established group differences in the neural response to viewing appetizing foods in the dorsal and ventral striatum. A comparison of appetizing foods compared to non-foods produced a significant difference between the groups in the left dorsal striatum (caudate-putamen) [F(2,23) = 9.92; P = .04 small-volume correction (svc)] (Figure 1A and Table 3), reflecting reduced activation in obese controls relative to MC4R-deficient patients and lean controls (MC4R, T = 4.18, P = .01 svc; lean controls, T = 3.50, P = .02 svc). The dorsal striatum showed a similar group effect to appetizing foods vs bland foods [F(2,23) = 7.33; P = .07 svc], reflecting a significantly reduced response in the obese group compared with the MC4R-deficient patients and lean controls (MC4R—T = 3.81, P = .01 svc; lean controls—T = 3.58, P = .03 svc). A similar pattern was also apparent in the left ventral striatum [F(2,23) = 9.25; P = .05 svc] (Figure 1A), where obese controls showed reduced response to appetizing foods vs non-food objects relative to the MC4R-deficient group (T = 4.25; P = .008 svc) and a similar trend relative to the lean controls. For each of these contrasts, the MC4R group did not significantly differ from lean controls. There were no group differences in dorsal or ventral striatum for the contrast of disgusting foods with bland foods or non-foods, suggesting that the effects observed in response to appetizing foods were not related to arousal more generally.

se contrasts, the MC4R group did not significantly differ from lean controls. There were no group differences in dorsal or ventral striatum for the contrast of disgusting foods with bland foods or non-foods, suggesting that the effects observed in response to appetizing foods were not related to arousal more generally. Figure 1. Activation of dorsal and ventral striatum (A) and somatosensory cortex (B) in response to appetizing foods compared to non-food objects in MC4R deficiency and obese and lean controls. Coronal sections show the main effect of group at P < .005 uncorrected for display purposes. Extracted data plotted for each group are the average group parameter estimates at the peak voxel for ventral (x = −12, y = 18, z = −12) and dorsal striatum (x = −18, y = 14, z = 6) after multiple comparisons correction (small-volume correction), and at P < .001 uncorrected for the somatosensory cortex (x = 64, y = −20, z = 38). Error bars represent the standard error of the mean. Table 3. fMRI Striatal ROI Analysis in MC4R-Deficient Individuals, Lean and Obese Controls

Figure 1. Activation of dorsal and ventral striatum (A) and somatosensory cortex (B) in response to appetizing foods compared to non-food objects in MC4R deficiency and obese and lean controls. Coronal sections show the main effect of group at P < .005 uncorrected for display purposes. Extracted data plotted for each group are the average group parameter estimates at the peak voxel for ventral (x = −12, y = 18, z = −12) and dorsal striatum (x = −18, y = 14, z = 6) after multiple comparisons correction (small-volume correction), and at P < .001 uncorrected for the somatosensory cortex (x = 64, y = −20, z = 38). Error bars represent the standard error of the mean. Table 3. fMRI Striatal ROI Analysis in MC4R-Deficient Individuals, Lean and Obese Controls Brain Region P (FWE) F or T MNI Coordinates x y z Appetizing > non-food Dorsal striatum left (caudate/putamen) Main effect of group .04 F = 9.92 −18 14 6 MC4R > obese .01 T = 4.18 −18 14 6 LC > obese .02 T = 3.45 −20 14 8 Ventral striatum left Main effect of group .05 F = 9.25 −12 18 −12 MC4R > obese .01 T = 4.25 −12 20 −12 LC > obese .10 T = 2.87 −12 16 −12 Appetizing > bland Dorsal striatum left (caudate/putamen) Main effect of group .07 F = 7.33 −14 20 16 MC4R > obese .01 T = 3.81 −12 20 16 LC > obese .03 T = 3.26 −16 18 8 Abbreviations: LC, lean controls; FWE, family-wise error; MNI, Montreal Neurological Institute. Statistics and coordinates for analysis of group differences in striatal ROIs (small volume corrected for multiple comparisons). Main effects of group are listed, as well as post hoc t tests.

LC > obese .03 T = 3.26 −16 18 8 Abbreviations: LC, lean controls; FWE, family-wise error; MNI, Montreal Neurological Institute. Statistics and coordinates for analysis of group differences in striatal ROIs (small volume corrected for multiple comparisons). Main effects of group are listed, as well as post hoc t tests. Our additional exploratory analyses of the amygdala, orbitofrontal cortex, and somatosensory cortex, at a more lenient threshold of P < .001 uncorrected, revealed group differences in response to disgusting foods relative to non-foods in the right frontal operculum (x = 54, y = 30, z = 2) (Figure 2 and Supplemental Table 1), where MC4R-deficient individuals showed significantly greater response than obese controls (T = 4.88; P < .001), but no difference from lean controls. The frontal operculum is involved in processing taste information and relating it to motivation, emotion, and visceral activity (9). Studies in rodents indicate that forebrain MC4R stimulation influences taste responsiveness, potentially by modulation of the perceived intensity of taste stimuli (10, 11). In addition, we found significantly greater activity in the MC4R-deficient group [F(2,23) = 11.13; P < .001 uncorrected) for appetizing foods vs non-foods compared to obese (T = 3.70; P = .001) and lean controls (T = 4.43; P < .001) in the inferior parietal cortex (x = 62, y = −22, z = 40), which corresponds to the somatosensory cortex area that encodes sensation in the mouth, lips, and tongue (Figure 1A). We found no group differences in the amygdala or orbitofrontal cortex for any of the contrasts, even at our more lenient threshold of P < .001 uncorrected. Additional brain regions that showed a main effect of group at the whole brain uncorrected P < .001 level for each contrast are listed in Supplemental Table 1.

1A). We found no group differences in the amygdala or orbitofrontal cortex for any of the contrasts, even at our more lenient threshold of P < .001 uncorrected. Additional brain regions that showed a main effect of group at the whole brain uncorrected P < .001 level for each contrast are listed in Supplemental Table 1. Figure 2. Significant differences in activation in right frontal operculum (x = 54, y = 30, z = 2) in response to disgusting foods compared to non-foods in MC4R deficiency and obese and lean controls. Axial section shows the main effect of group at P < .005 uncorrected for display purposes. Extracted data plotted for each group are the average group parameter estimates at the peak voxel. Error bars represent the standard error of the mean. Differences in brain activation were not explained by differences in reward sensitivity between the groups as assessed by the Behavioral Activation Scale-drive scale (12) (F < 1; P = .8) which has previously been associated with variability in responses to food reward (6). There were also no group differences in how participants rated the images on pleasantness (F < 1; P = .57) and disgust (F < 1; P = .23) postscanning.

ups as assessed by the Behavioral Activation Scale-drive scale (12) (F < 1; P = .8) which has previously been associated with variability in responses to food reward (6). There were also no group differences in how participants rated the images on pleasantness (F < 1; P = .57) and disgust (F < 1; P = .23) postscanning. Discussion In this small study, we have been able to distinguish the brain response to food images between two highly comparable groups of obese people with and without genetic disruption of melanocortin signaling. In our study, all participants were studied in the satiated state to adjust for potential differences in motivational state between the groups (hunger ratings using visual analog scores were comparable before scanning).

hly comparable groups of obese people with and without genetic disruption of melanocortin signaling. In our study, all participants were studied in the satiated state to adjust for potential differences in motivational state between the groups (hunger ratings using visual analog scores were comparable before scanning). Several studies have reported changes in striatal activation in response to food-related stimuli in obese individuals. Some have proposed that a hyper-responsiveness of reward circuitry, when presented with food images, increases the risk for overeating (13, 14). Others hypothesize that obese individuals show hyporesponsiveness of reward circuitry after the consumption of food, which leads them to overeat to compensate for this deficiency (15). In a prospective study, weight gain was associated with a reduction in striatal activation in response to palatable food intake relative to baseline response (16). Our findings suggest that some obese individuals show hyporesponsivity of reward circuitry to visual food cues when studied in the satiated state (which may more closely parallel studies of food consumption). Our findings do not preclude that increased striatal activation may be seen in some obese individuals in response to food cues presented in the fasted/hungry state. Of note, in many studies, participants were studied after fasting for several hours, and motivational state may have modulated the striatal response to images of food and/or consumption of food. Our findings argue against a model where chronic overconsumption is sufficient to account for decreased striatal activation. Instead, our findings suggest that differences in brain responses to food cues in obese people are more likely to be due to differences in neural circuitry that occur with weight gain. We suggest that these neural changes require intact melanocortin signaling and, as such, are not seen in individuals with mutations that disrupt MC4R.

ndings suggest that differences in brain responses to food cues in obese people are more likely to be due to differences in neural circuitry that occur with weight gain. We suggest that these neural changes require intact melanocortin signaling and, as such, are not seen in individuals with mutations that disrupt MC4R. It has been suggested that differences in striatal activation may reflect differences in dopaminergic tone. Obese rats, relative to lean rats, show reduced D2 receptor density in the hypothalamus and in the striatum, and chronic excessive intake of high-calorie foods and attendant weight gain results in down-regulation of postsynaptic D2 receptors, increased D1 receptor binding, and decreased D2 sensitivity in animals (17, 18). Obese vs lean humans show reduced striatal D2 receptor density measured by positron emission tomography imaging in some studies (19). How might disruption of melanocortinergic signaling modulate dopaminergic circuits involved in food reward? Several studies in rodents have demonstrated that injection of melanocortins or of the endogenous MC4R antagonist, Agouti-related peptide, can modulate dopamine-containing neurons within brain regions known to mediate reward-based decision making and augment operant responding for palatable food (3). However, the direction of response seen varies depending on the anatomical site of injection, suggesting that multiple excitatory and inhibitory inputs on MC4R-expressing neurons are likely to contribute to the regulation of dopamine turnover. Interestingly, a recent case report suggested that pharmacological modification of dopaminergic tone by methylphenidate might stimulate weight loss in a patient with MC4R deficiency (20).

ultiple excitatory and inhibitory inputs on MC4R-expressing neurons are likely to contribute to the regulation of dopamine turnover. Interestingly, a recent case report suggested that pharmacological modification of dopaminergic tone by methylphenidate might stimulate weight loss in a patient with MC4R deficiency (20). The explanation for increased activation of oral somatosensory cortex in the MC4R-deficient group is unclear. Pharmacological studies in rodents suggest that melanocortins do not modulate taste (as measured by licking frequency for sucrose/quinine). It is plausible that increased oral somatosensory cortex activation in MC4R-deficient humans may reflect increased anticipatory orosensory activity or salivation in response to appetizing food images in this group, although this was not formally tested. Increased somatosensory cortex activation has been observed in response to food reward (but not to monetary reward) in some previous studies (21). Using positron emission tomography, obese people showed greater resting metabolic activity in the oral somatosensory cortex relative to lean people, and using fMRI, obese vs lean adolescents exhibited greater activation in the oral somatosensory cortex in response to receipt of a chocolate milkshake vs receipt of a tasteless solution (22). Interestingly, apart from the somatosensory cortex findings in our exploratory analysis, the brain response of the MC4R-deficient individuals was not different from lean volunteers. This might be due to compensatory mechanisms related to the signaling defect in MC4R deficiency or the possibility that any changes might be too subtle to detect with fMRI. Alternatively, we recognize that the sample size is small due to the nature of the condition studied, and these findings will have to be repeated in larger studies. Nonetheless, our findings from this preliminary study suggest that central melanocortin signaling is involved in modulating the neural response to food cues in human obesity. Further studies will be needed to investigate the precise neural circuits that explain these observations. Understanding the complex interplay of biological and behavioral factors involved in the response to rewarding food cues may suggest interventions aimed at limiting the overconsumption of highly palatable foods associated with weight gain.

eded to investigate the precise neural circuits that explain these observations. Understanding the complex interplay of biological and behavioral factors involved in the response to rewarding food cues may suggest interventions aimed at limiting the overconsumption of highly palatable foods associated with weight gain. Abbreviations: EPIecho planar image fMRIfunctional magnetic resonance imaging MC4Rmelanocortin-4 receptor ROIregion of interest svcsmall-volume correction. Acknowledgments We thank Paul Fletcher, Department of Psychiatry, University of Cambridge, Cambridge, for helpful discussions. This work was supported by the Wellcome Trust (to A.A.v.d.K., I.S.F.), the National Institute for Health Research Cambridge Biomedical Research Centre (to S.O., I.S.F.), the Bernard Wolfe Health Neuroscience Fund (to A.A.v.d.K., I.S.F.), and the Medical Research Council UK, Grant MC_US_A060_5PQ50 (to A.J.C.). Disclosure Summary: The authors have nothing to declare.

Primary hyperparathyroidism (PHPT) is a common endocrine disease; the prevalence is 1–4 per 1000, and it increases to 21 per 1000 in age groups 55–75 years (1). In 5–10% of cases, PHPT is part of a genetic syndrome, such as multiple endocrine neoplasia type 1 or 2, hyperparathyroidism-jaw tumor syndrome (HPT-JT), familial isolated hyperparathyroidism (FIHP), or familial hypocalciuric hypercalcemia (1–5). The HPT-JT syndrome is an autosomal dominant disease characterized by parathyroid tumors and ossifying tumors of the jaw (4). Some patients also develop renal and uterine tumors (4). The HPT-JT syndrome is due to mutations of the cell division cycle protein 73 homolog (CDC73) gene, located at 1q31.2 (3, 6–8). CDC73 has 17 exons, acts as a tumor suppressor gene, and encodes parafibromin, a 531-amino acid protein predominantly expressed in the nucleus (4). Parafibromin serves as a parathyroid carcinoma marker because it is expressed in normal parathyroid glands, parathyroid hyperplasia, and adenomas but is usually absent in parathyroid carcinomas (9–11) and occasionally in atypical adenomas. To date, more than 60 CDC73 germline mutations have been reported, the majority being frameshift, nonsense, and missense mutations (12). Approximately 55% of CDC73 mutations are associated with HPT-JT and over 20% with FIHP (12). HPT-JT and FIHP patients who do not have CDC73 mutations may have intragenic or whole deletions of CDC73 (13–15), and >5% of PHPT patients without CDC73 mutations have been reported to have large CDC73 deletions (15). Thus, it is recommended that deletion analysis of the CDC73 gene should be performed in HPT-JT, FIHP, parathyroid carcinoma, or severe early-onset PHPT patients who do not have CDC73 point mutations (15–17). We report a three-generation family with FIHP, in whom CDC73 mutations were not identified in the original report (8), but in whom we have now identified an intragenic deletion involving exons 1 to 10.

PT-JT, FIHP, parathyroid carcinoma, or severe early-onset PHPT patients who do not have CDC73 point mutations (15–17). We report a three-generation family with FIHP, in whom CDC73 mutations were not identified in the original report (8), but in whom we have now identified an intragenic deletion involving exons 1 to 10. Subjects and Methods Eleven members of a three-generation family with PHPT were ascertained (Figure 1 and Table 1). Informed written consent was obtained from all the patients. Clinical data from the two older generations as well as linkage to the HPT-JT locus on 1q21-q32 were reported previously (3). The affected relatives in the second generation shared haplotypes with the father (Figure 1, I-1) but not the mother (Figure 1, I-2) and thus had inherited the disease from the father (3). Both the father and the mother (Figure 1, I-1 and I-2) suffered from PHPT, but the father, who died in 2001 with a ruptured thoracic aneurysm at the age of 69 years, did not have parathyroid surgery. He was not willing to undergo regular screening for PHPT during his lifetime. His serum calcium was documented to be increased a year before he died. The mother (member I-2) was operated on for PHPT in 2011. Histopathology was compatible with parathyroid hyperplasia. Her serum ionized calcium has increased to preoperative concentrations and is currently 1.40 mmol/L (reference, 1.16–1.30 mmol/L). Plasma biochemistry using morning fasting samples was performed at the Helsinki University Central Hospital Laboratory using standard methods.

y was compatible with parathyroid hyperplasia. Her serum ionized calcium has increased to preoperative concentrations and is currently 1.40 mmol/L (reference, 1.16–1.30 mmol/L). Plasma biochemistry using morning fasting samples was performed at the Helsinki University Central Hospital Laboratory using standard methods. Figure 1. Family tree. White denotes normal biochemistry; light gray, PHPT and underlying hyperplasia or adenoma; dark gray, PHPT and atypical adenoma; and black, PHPT and parathyroid carcinoma. +/+, Wild type (no deletion of CDC73 gene exons 1–10); +/−, heterozygous deletion of CDC73 gene exons 1–10; NA, not available; squares, male family members; circles, female family members; slash, deceased family member. Table 1. Characteristics of Family Members ID No. Sex (F/M) Year of Birth PHPT (Age at Surgery, y) Calcium Ion in Nonoperated Subjects, mmol/L Histology Renal Cysts Aortic Aneurysm CDC73 Deletion I-1 M 1932 Yes NA No Yes NA I-2 F 1934 Yes (77) HP No No No II-1 F 1953 No 1.24 No II-2 M 1955 Yes (36) A No No Yes II-3 F 1956 Yes (37) AA Yes No Yes II-4 M 1959 Yes (38) A No No Yes II-5 M 1961 Yes (32) CA Yes Yes Yes III-1 F 1974 No 1.21 Yes III-2 F 1978 Yes (31) A No NA III-3 M 1993 Yes (18) CA, AA No No Yes III-4 F 1993 No 1.26 No III-5 F 1987 No 1.30 Yes III-6 F 1995 No 1.26 No Abbreviations: F, female; M, male; A, parathyroid adenoma; AA, atypical adenoma; CA, parathyroid carcinoma; NA, not available; HP, hyperplasia. The reference range for serum calcium ion is 1.16–1.30 mmol/L.

d endometriosis. RQ, relative quantification. Statistical analysis was performed using a one-way ANOVA combined with a Newman Keuls post hoc test. *, P < .05; **, P < .01; ***, P < .001. Asterisks denote statistical significance compared with control, and lines and asterisks denote significance between specific groups. mRNA concentrations of nociceptive genes are regulated by estrogens in ES cell-derived sensory neurons Sensory neurons immunopositive for neurofilament were generated in vitro (15) (Figure 2, A–C). Differentiation was accompanied by a reduction in octamer transcription factor-4 mRNA concentrations and an increase in TAC1, SCN9A, and SCN11A (Figure 2D), consistent with differentiation of sensory neurons with a nociceptor-like phenotype. When the differentiated cells were incubated with capsaicin (activates TRPV1), an intracellular calcium flux was detected (Figure 2E). The expression of ER subtypes changed during differentiation; ERα mRNA concentrations decreased and ERβ mRNAs increased (Figure 2F). Stimulation of sensory neurons with the estrogen ligands E2, DPN, and PPT revealed ER-dependent regulation of nociceptive ion channels. TAC1 and P2RX3 mRNAs were increased after the incubation with DPN compared with cells incubated with DPN in the presence of ICI (Figure 2G; P < .001) or vehicle control (DMSO; Figure 2H; P < .05), respectively. TAC1 was also elevated by PPT compared with PPT + ICI (Figure 2G; P < .01). Although SCN9A increased with DPN (P < .05), this was not blocked by the addition of ICI (Figure 2I), and the expression of SCN11A (Figure 2J) and TRPA1 (Figure 2K) did not appear to be ER dependent. In contrast, TRPV1 mRNA concentrations were elevated by both E2 (P < .05) and DPN (Figure 2I; P < .01), and this effect was abrogated by ICI.

A III-3 M 1993 Yes (18) CA, AA No No Yes III-4 F 1993 No 1.26 No III-5 F 1987 No 1.30 Yes III-6 F 1995 No 1.26 No Abbreviations: F, female; M, male; A, parathyroid adenoma; AA, atypical adenoma; CA, parathyroid carcinoma; NA, not available; HP, hyperplasia. The reference range for serum calcium ion is 1.16–1.30 mmol/L. Multiplex ligation-dependent probe amplification (MPLA) analysis Genomic DNA extracted from whole blood was used in an MLPA assay, which is a dosage-based technique for detection of deletions and duplications of one or more exons. An MLPA kit P200-A1 (MRC Holland) was used as a reference kit, with the addition of an in-house-designed synthetic probe mix to detect deletions or duplications of CDC73. This assay did not include probes for exons 5, 12, 14, or 16, but a probe for all other exons (1–4, 6–11, 13, 15, and 17) was included. Immunohistochemistry Immunostaining was performed using deparaffinized tissue section utilizing mouse monoclonal Ki-67 antibody (clone MIB-1; Dako), 1:100 for proliferation index, as well as mouse monoclonal parafibromin antibody (clone sc-33638; Santa Cruz Biotechnology), 1:1000 with the polymer detection kit EnVision (Dako) in a LabVision Autostainer (Thermo Scientific); sections were counterstained with Mayer's Hematoxylin (Lillie's Modification) (Dako) and mounted with Mountex (Histolab). Ki-67 proliferation indices were assessed in 2000 neoplastic cells.

; Santa Cruz Biotechnology), 1:1000 with the polymer detection kit EnVision (Dako) in a LabVision Autostainer (Thermo Scientific); sections were counterstained with Mayer's Hematoxylin (Lillie's Modification) (Dako) and mounted with Mountex (Histolab). Ki-67 proliferation indices were assessed in 2000 neoplastic cells. Results CDC73 gene analyses MPLA analysis of the CDC73 gene revealed a heterozygous deletion of exons 1–10 in seven of the available 11 family members (Figure 1 and Table 1). Five of these individuals with the CDC73 deletion had been operated on for PHPT at the ages of 18–38 years (Figure 1 and Table 1), whereas the other two (Table 1, III-1 and III-5), aged 39 and 26 years at the last biochemical screening, did not have PHPT.

seven of the available 11 family members (Figure 1 and Table 1). Five of these individuals with the CDC73 deletion had been operated on for PHPT at the ages of 18–38 years (Figure 1 and Table 1), whereas the other two (Table 1, III-1 and III-5), aged 39 and 26 years at the last biochemical screening, did not have PHPT. Phenotypes of family members affected with PHPT Family members II-5 and III-3 (Figure 1 and Table 1) presented with severe PHPT, headaches, and hypertension at the ages of 32 and 18 years, respectively. Blood pressure normalized in both subjects after primary parathyroidectomy. Family member II-5 was most severely affected. He was diagnosed with severe PHPT at age 32 years, the initial symptoms being headaches and hypertension (blood pressure values were 160–180/110–130 mm Hg). Surgery revealed a 3-cm encapsulated parathyroid carcinoma of the left lower parathyroid gland. The right parathyroid glands were macroscopically normal and were left intact. He had his first recurrence <2 years later when a 2-cm lymph node metastasis was resected from the left side of the neck. Recurrence was preceded by rising serum calcium and PTH concentrations and recurrent hypertension. A second neck exploration was performed at age 36 years, when an adenoma of the right upper parathyroid gland, a normal left lower parathyroid, and a lymph node metastasis were resected. He has since had more than nine operations because of recurrent hypercalcemia. Distal metastases located close to the spine were resected at ages 38, 39, and 49 years. Family member II-5 was also diagnosed with a thoracic aortic aneurysm, which was operated at the age of 51 years, and family member I-1 died at the age of 69 because of a ruptured thoracic aortic aneurysm (Figure 1 and Table 1).

a. Distal metastases located close to the spine were resected at ages 38, 39, and 49 years. Family member II-5 was also diagnosed with a thoracic aortic aneurysm, which was operated at the age of 51 years, and family member I-1 died at the age of 69 because of a ruptured thoracic aortic aneurysm (Figure 1 and Table 1). Family member III-3 was diagnosed at the age of 18 years with severe hypertension (blood pressure, 210/110–115 mm Hg) and headaches. He had severe PHPT, with serum ionized calcium of 1.88 mmol/L (normal, 1.16–1.30 mmol/L) and serum PTH of 693 ng/L (normal, 8–73 ng/L). Investigations revealed that he did not have pheochromocytoma or primary hyperaldosteronism. Bilateral neck exploration was performed; a 1.5-cm enlarged upper right parathyroid was resected, and the right lower parathyroid gland was left intact. The left upper parathyroid was also enlarged at 1.4 cm and was resected, as well as the left lower parathyroid gland, which was macroscopically normal. Serum calcium ion normalized immediately postoperatively (from 1.30 to 1.19 mmol/L), and serum PTH was 10 ng/L on the first postoperative day. The headaches and hypertension resolved. He had carcinoma of the right upper parathyroid gland and atypical adenoma of the left upper gland. The left lower parathyroid was normal. Among the other affected family members, family member II-3 had an atypical adenoma (Table 1), whereas members II-2, II-4, and III-2 had parathyroid adenomas. Two family members (II-3 and II-5; Table 1) have bilateral renal cysts. X-ray did not reveal any jaw tumors in this family.

he left lower parathyroid was normal. Among the other affected family members, family member II-3 had an atypical adenoma (Table 1), whereas members II-2, II-4, and III-2 had parathyroid adenomas. Two family members (II-3 and II-5; Table 1) have bilateral renal cysts. X-ray did not reveal any jaw tumors in this family. Ki67 proliferation index and parafibromin stain in parathyroid carcinoma Histopathological examination of the right upper gland of family member III-3 (18-year-old male; Figure 1 and Table 1) demonstrated parathyroid carcinoma with vascular invasion (Figure 2, A and B). The Ki-67 proliferation index was 5% (Figure 2C), and nuclear parafibromin immunostaining was negative (Figure 2D). The left upper parathyroid was an atypical adenoma that did not immunostain for parafibromin. In family member II-5 (Figure 1 and Table 1), the Ki-67 of the primary parathyroid carcinoma resected at age 32 years was 14.5%, and that of a neck lymph node metastasis resected at age 49 years was 20%. These parathyroid carcinoma and lymph node metastasis did not immunostain for parafibromin. Figure 2. Parathyroid carcinoma. A, Monotonous growth of mainly chief cell-like tumor cells, with slide nuclear atypia. Arrow indicates mitotic figure (hematoxylin and eosin; magnification, ×400). B, Vascular invasion, arrow indicating tumor thrombus. C, Proliferation index (Ki67) of 5%, evaluated immunohistochemically by Mib-1 antibody. D, Negative parafibromin staining of neoplastic cells. Arrow indicates parafibromin-positive vascular endothelial cells that serve as an internal positive control.

ation, ×400). B, Vascular invasion, arrow indicating tumor thrombus. C, Proliferation index (Ki67) of 5%, evaluated immunohistochemically by Mib-1 antibody. D, Negative parafibromin staining of neoplastic cells. Arrow indicates parafibromin-positive vascular endothelial cells that serve as an internal positive control. Discussion MPLA analysis revealed that a previously unreported heterozygous deletion of CDC73 exons 1–10 is the cause of PHPT in this three-generation family, thereby highlighting the need for performing CDC73 deletion analysis in parathyroid carcinoma patients, and FIHP and HPT-JT families who do not have point mutations of CDC73. This represents the fourth report of a CDC73 deletion. Whole CDC73 deletions have been previously reported in a 25-year-old Portuguese man with severe, sporadic PHPT due to a single parathyroid adenoma (17) and a Spanish HPT-JT family, in which the index patient was an 18-year-old female who had jaw tumors and PHPT due to an adenoma that stained negative for parafibromin (13). Analysis of 250 PHPT patients reported 7% (n = 20 patients) to have a germline CDC73 abnormality; CDC73 deletions were found in seven patients, with three having whole gene deletions and four having intragenic deletions involving exons 3, 2 and 3, 4–6, and 7–13, respectively (15).

tained negative for parafibromin (13). Analysis of 250 PHPT patients reported 7% (n = 20 patients) to have a germline CDC73 abnormality; CDC73 deletions were found in seven patients, with three having whole gene deletions and four having intragenic deletions involving exons 3, 2 and 3, 4–6, and 7–13, respectively (15). The most severe PHPT and phenotypes were encountered in two males with the exon 1–10 deletion of CDC73, and both of these had parathyroid carcinoma, at ages 32 and 18 years; they also had headaches and severe hypertension as the initial manifestations of PHPT, which resolved after primary surgery. Hypertension is not a usual feature of PHPT; today, PHPT is often a biochemically mild, asymptomatic disorder, or it involves diffuse symptoms such as obstipation and fatigue (1). Genotype-phenotype correlations were not apparent within this family, consistent with previous reports (4, 5). Thus, deletion of CDC73 exons 1–10 resulted in parathyroid carcinoma diagnosed at the ages of 18 and 32 years in two male family members (Table 1; members III-3 and II-5, respectively), whereas other members have parathyroid adenomas or atypical adenomas. In addition, two affected males had thoracic aortic aneurysms, which have not been previously reported to be associated with FIHP, HPT-JT, or CDC73 mutations, and it is possible that aortic aneurysms may be a novel feature of these disorders. Alternatively, thoracic aneurysms represent another trait in this family unrelated to the CDC73 deletion. Two mutation carriers (members III-1 and III-5; ages 39 and 26 y) have not yet developed PHPT. The older one has a serum ionized calcium of 1.21 mmol/L and may represent a case of incomplete penetrance. The younger one might still develop PHPT because her current serum calcium is 1.30 mmol/L, at the upper reference limit. There is not much data on unaffected mutation carriers in the literature. Guarnieri et al (18) reported a family with FIHP caused by a CDC73 frameshift mutation, where three of 10 mutation carriers had either an atypical parathyroid adenoma (n = 2) or parathyroid carcinoma (n = 1), but only one of the mutation carriers was hypercalcemic. The authors concluded that longitudinal surveillance including neck ultrasound is important (18).

FIHP caused by a CDC73 frameshift mutation, where three of 10 mutation carriers had either an atypical parathyroid adenoma (n = 2) or parathyroid carcinoma (n = 1), but only one of the mutation carriers was hypercalcemic. The authors concluded that longitudinal surveillance including neck ultrasound is important (18). The parathyroid carcinomas and the atypical adenoma did not immunostain for parafibromin, and the Ki-67 proliferation indices of the parathyroid carcinomas were 5 and 14.5%. The aggressive and recurrent behavior of the parathyroid carcinoma is consistent with observations that indicate that the presence of a CDC73 mutation and loss of parafibromin immunostaining predicts significantly higher recurrence rates and lower overall 10-year survival (17). Loss of parafibromin immunostaining has been reported to be a better predictor of clinical outcome and mortality rates than CDC73 mutation (20), and our findings are consistent with these observations. Further studies are needed to evaluate the predictive value of these tumor markers (17, 19–21) in patients with CDC73 germline mutations in sporadic cases of parathyroid carcinomas and atypical adenomas.

nical outcome and mortality rates than CDC73 mutation (20), and our findings are consistent with these observations. Further studies are needed to evaluate the predictive value of these tumor markers (17, 19–21) in patients with CDC73 germline mutations in sporadic cases of parathyroid carcinomas and atypical adenomas. In summary, we report a previously unreported intragenic deletion involving exons 1–10 of the CDC73 gene in familial PHPT associated with parathyroid carcinoma. In addition to parathyroid tumors, this deletion was associated in two males with thoracic aortic aneurysms, which possibly represents a new manifestation of CDC73 mutations. The findings of the present study underline the need for performing CDC73 deletion analysis in HPT-JT, FIHP, parathyroid carcinoma, and/or severe early-onset PHPT patients who do not have point mutations. Abbreviations: CDC73cell division cycle protein 73 homolog FIHPfamilial isolated hyperparathyroidism HPT-JThyperparathyroidism-jaw tumor syndrome MPLAmultiplex ligation-dependent probe amplification PHPTprimary hyperparathyroidism. Acknowledgments We thank Ms Helena Ranta-aho and Dr Otto Knutar for kind and skillful technical help. This work was supported by research funds from the University Central Hospital in Helsinki (to C.S.-J., E.R., K.A.), Finska Läkaresällskapet (to C.S.-J.), and from the United Kingdom Medical Research Council (Grants G9825289 and G1000467; to R.V.T.), and the National Institute for Health Research Oxford Biomedical Research Centre Programme (to R.V.T.). Disclosure Summary: The authors have nothing to disclose.

Endometriosis is associated with debilitating chronic pelvic pain (CPP). It is an estrogen-dependent inflammatory disorder characterized by growth of endometrium (lesions) in ectopic locations, typically on the peritoneum and ovaries. Lesions become innervated by small diameter nerve fibers typical of afferent sensory innervation (1), and stimulation of these nerves by the inflammatory milieu within a lesion may be a direct cause of endometriosis-associated pain. Women with endometriosis often suffer from cyclical pain, and pain may occur in association with other conditions, such as irritable bowel syndrome and migraine (2, 3). Visceral, mechanical, and generalized hypersensitivity is extremely common in women with endometriosis (4), although there is no correlation between the extent of disease and reported pain scores (5). Stratton and Berkley (6) recently provided an overview of evidence for recruitment of nerves to endometriosis lesions, and they suggested that their stimulation and subsequent dialogue with the central nervous system may explain why women with endometriosis suffer from chronic pain.

isease and reported pain scores (5). Stratton and Berkley (6) recently provided an overview of evidence for recruitment of nerves to endometriosis lesions, and they suggested that their stimulation and subsequent dialogue with the central nervous system may explain why women with endometriosis suffer from chronic pain. Studies in patients with altered pain responses due to rare conditions as well as genetically modified mice (7) have identified key genes expressed by nerves that detect noxious stimuli (nociceptors). These include the TRP channel family (TRPA1, TRPV1), sodium voltage-gated ion channels (SCN9A, SCN11A) and purinergic X (P2X) receptors. An increase in the expression of neuropeptides, neurotrophins, and ion channels leading to an increased responsiveness of nociceptors has been suggested as an important contributor to development of hypersensitivity (8). For example, the TRPV1 channel (capsaicin receptor) has been implicated in development of hypersensitivity to heat associated with inflammation in chronic pain conditions (9). Recently studies have reported that TRPV1 immunoreactivity is increased in the peritoneum of women with CPP (10) and in lesions of women with endometriosis (11). The neuropeptide substance P (encoded by the TAC1 gene) has been immunolocalized to endometriosis lesions (12), and TRPA1 has been implicated in thermal and mechanical hypersensitivity to pain (13). The purinergic X family of receptors are voltage-gated ion channels that open in response to ATP released from cells during inflammation and have also been associated with mechanical hyperalgesia (14). Sodium channels are also critical in producing an acute response to noxious stimuli; Nav1.7 (SCN9A) and Nav1.9 (SCN11A) are both reported to be increased in a range of inflammatory pain conditions (7). Notably, induction of a hypoestrogenic state often ameliorates CPP, and although the mechanisms responsible for this remain poorly understood, estrogens are reported to influence the sensitivity of peripheral sensory neurons and central neuronal activity in pain states (2).

a range of inflammatory pain conditions (7). Notably, induction of a hypoestrogenic state often ameliorates CPP, and although the mechanisms responsible for this remain poorly understood, estrogens are reported to influence the sensitivity of peripheral sensory neurons and central neuronal activity in pain states (2). In this study we measured the concentrations of mRNAs encoded by genes known to regulate nociception in tissue samples from women with CPP, some of whom were diagnosed with endometriosis. We also explored the impact of estrogens on their expression using a newly established in vitro model of human embryonic stem (ES) cell-derived sensory neurons.

e concentrations of mRNAs encoded by genes known to regulate nociception in tissue samples from women with CPP, some of whom were diagnosed with endometriosis. We also explored the impact of estrogens on their expression using a newly established in vitro model of human embryonic stem (ES) cell-derived sensory neurons. Materials and Methods Patients and recovery of tissue samples The study was approved by the Lothian Research Ethics Committee (LREC 11/AL/0376). Peritoneal biopsies [∼1.0 × 0.5 cm size: stored in RNAlater (Applied Biosystems) at −20°C] were obtained from 3 groups of women (aged 18–45 y) at the time of surgery with informed consent: 1) women undergoing laparoscopic investigation for CPP with histological evidence of pelvic endometriosis (n = 12); 2) women undergoing laparoscopic investigation for CPP without evidence of endometriosis (n = 10); and 3) women undergoing laparoscopic sterilization without CPP and without evidence of endometriosis (n = 5). Endometriosis lesions were also collected from women with endometriosis (n = 18). Peritoneal biopsies from women without endometriosis were recovered from a site prone to endometriosis (pouch of Douglas) and in women with endometriosis the peritoneal biopsy was recovered from a site adjacent to the endometriosis lesion. None of the women had taken exogenous hormones for at least 3 months at the time of sampling.

opsies from women without endometriosis were recovered from a site prone to endometriosis (pouch of Douglas) and in women with endometriosis the peritoneal biopsy was recovered from a site adjacent to the endometriosis lesion. None of the women had taken exogenous hormones for at least 3 months at the time of sampling. Differentiation of sensory neurons from human embryonic stem cells Human ES cells, strain H9 (WiCell) were cultured on inactivated mouse embryonic fibroblasts maintained in Knockout DMEM/F12 supplemented with 20% knockout serum replacement (Gibco), 1 mM L-glutamine, 100 μM MEM nonessential amino acids, and 0.1 μM β-mercaptoethanol, with 6 ng/mL FGF-2 (R&D Systems). Differentiation of ES cells was induced using a cocktail of small molecule inhibitors as detailed by Chambers et al (15). RNA was extracted from cells at selected time points for the characterization and verification of phenotype. On day 22 the sensory neurons were trypsinized and reseeded into 12-well plates at equal densities (5 × 10−5 cells/well). On day 25, the cells were treated for 24 hours with 10−8 M 17β-estradiol (E2; Sigma), the estrogen receptor (ER)-α selective agonist 4,4′,4′-[4-propyl-(1H)-pyrazole-1,3,5-tryl] trisphenol (PPT), or the ERβ selective agonist 2,3-bis(4-hydroxy-phenyl)-propionitrile (DPN; Tocris) alone or in combination with the antiestrogen fulvestrant (ICI; 10−7 M; Tocris) dissolved in dimethylsulfoxide (DMSO).

; Sigma), the estrogen receptor (ER)-α selective agonist 4,4′,4′-[4-propyl-(1H)-pyrazole-1,3,5-tryl] trisphenol (PPT), or the ERβ selective agonist 2,3-bis(4-hydroxy-phenyl)-propionitrile (DPN; Tocris) alone or in combination with the antiestrogen fulvestrant (ICI; 10−7 M; Tocris) dissolved in dimethylsulfoxide (DMSO). Immunocytochemistry Expression of neurofilament protein was detected using neurofilament H chicken antineurofilament H (1:1000; Covance); nuclei were stained with 4′,6′-diamino-2-phenylindole and images captured on an Axiovert microscope (Carl Zeiss Inc). Calcium assay Neurons were stimulated with 4 nM capsaicin (Sigma), and intracellular calcium was measured using a calcium indicator kit (BD Biosciences) with calcium flux captured using a NOVOstar microplate fluorometer (BMG Labtech).

Immunocytochemistry Expression of neurofilament protein was detected using neurofilament H chicken antineurofilament H (1:1000; Covance); nuclei were stained with 4′,6′-diamino-2-phenylindole and images captured on an Axiovert microscope (Carl Zeiss Inc). Calcium assay Neurons were stimulated with 4 nM capsaicin (Sigma), and intracellular calcium was measured using a calcium indicator kit (BD Biosciences) with calcium flux captured using a NOVOstar microplate fluorometer (BMG Labtech). Quantitative real-time PCR RNA was extracted from human tissues by homogenization in TRI reagent, chloroform phase separation and the lysates processed using an RNAeasy kit (QIAGEN). RNA was extracted from cells using RLT (lysis) buffer and an RNAeasy kit. Samples were deoxyribonuclease treated (QIAGEN) and concentration and purity assessed using a NanoDrop ND 1000. cDNA was synthesized using SuperScript VILO enzyme (Invitrogen) with 100 ng starting template in a 20-μL reaction. PCRs (15 μL) were performed using the Roche Universal Probe Library (Roche Applied Science) and Express quantitative PCR supermix (Invitrogen) with primers added at a concentration of 20 μM and thermal cycling performed on a 7900 Fast real-time PCR machine (Applied Biosystems) with 18S selected as the reference gene because we have confirmed this gene is not altered by estrogenic treatments. cDNA was added at 1.5 μL per reaction and duplicate technical replicates performed.

ded at a concentration of 20 μM and thermal cycling performed on a 7900 Fast real-time PCR machine (Applied Biosystems) with 18S selected as the reference gene because we have confirmed this gene is not altered by estrogenic treatments. cDNA was added at 1.5 μL per reaction and duplicate technical replicates performed. Data were analyzed with RQ manager software (Applied Biosystems) using the δδCt method, and samples were normalized to either one laparoscopic sterilization sample (patients) or one vehicle control sample (cells). Primer sequences included the following: TRPA1, forward, 5′-tggacaccttcttcttgcatt-3′, reverse, 5′-tcatccatttcatgcagcac-3′; TRPV1, forward, 5′-agagtcacgctggcaacc-3′, reverse, 5′-ggcagagactctccatcacac-3′; SCN9A, forward, 5′-caacttttaagggatggacga-3′, reverse, 5′-tcatatttgggctgcttgtct-3′; SCN11A, forward, 5′-acctgagcctgaacaacagg-3′, reverse, 5′-tttgaactctctggctcgtg-3′; P2RX3, forward, 5′-ggcctttacttctgtgggagt-3′, reverse, 5′-aaacttcttggctttgtactggtc-3′; TAC1, forward, 5′-gcctcagcagttctttggat-3′, reverse, 5′-agcctttaacagggccactt-3′; ERα, forward, 5′-ttactgaccaacctggcaga-3′, reverse, 5′-atcatggagggtcaaatcca-3′; and ERβ, forward, 5′-atcatggagggtcaaatcca-3′, reverse, 5′-tgggcattcagcatctcc-3′. Statistical analysis Quantitative PCR (QPCR) data were analyzed using a one-way ANOVA and a Newman Keuls post hoc multiple comparison test. Statistics were generated using GraphPad Prism 6 software.

Data were analyzed with RQ manager software (Applied Biosystems) using the δδCt method, and samples were normalized to either one laparoscopic sterilization sample (patients) or one vehicle control sample (cells). Primer sequences included the following: TRPA1, forward, 5′-tggacaccttcttcttgcatt-3′, reverse, 5′-tcatccatttcatgcagcac-3′; TRPV1, forward, 5′-agagtcacgctggcaacc-3′, reverse, 5′-ggcagagactctccatcacac-3′; SCN9A, forward, 5′-caacttttaagggatggacga-3′, reverse, 5′-tcatatttgggctgcttgtct-3′; SCN11A, forward, 5′-acctgagcctgaacaacagg-3′, reverse, 5′-tttgaactctctggctcgtg-3′; P2RX3, forward, 5′-ggcctttacttctgtgggagt-3′, reverse, 5′-aaacttcttggctttgtactggtc-3′; TAC1, forward, 5′-gcctcagcagttctttggat-3′, reverse, 5′-agcctttaacagggccactt-3′; ERα, forward, 5′-ttactgaccaacctggcaga-3′, reverse, 5′-atcatggagggtcaaatcca-3′; and ERβ, forward, 5′-atcatggagggtcaaatcca-3′, reverse, 5′-tgggcattcagcatctcc-3′. Statistical analysis Quantitative PCR (QPCR) data were analyzed using a one-way ANOVA and a Newman Keuls post hoc multiple comparison test. Statistics were generated using GraphPad Prism 6 software. Results mRNAs encoding nociceptive proteins exhibit altered expression patterns in the peritoneum of women with CPP Measurement of mRNAs encoded by TAC1 and nociceptive ion channels revealed differences between tissue samples; TAC1 was elevated only in endometriosis lesions (Figure 1A; P < .05), and P2RX3 was increased in the peritoneum of women with CPP, regardless of whether they were diagnosed with endometriosis compared with the peritoneum of healthy women (Figure 1B; P < .05). SCN9A (Nav1.7) was elevated in lesions from women with endometriosis (Figure 1C; P < .05), whereas SCN11A was significantly higher in the peritoneum of women with CPP and endometriosis compared with the peritoneum of women with CPP alone (Figure 1D; P < .001). mRNA encoded by TRPA1 was significantly increased in the peritoneum of women with endometriosis compared with the peritoneum of healthy women and those with CPP alone (Figure 1E; P < .001). TRPV1 was elevated in peritoneum (P < .01) and in lesions (P < .05) of women with endometriosis compared with the peritoneum of healthy women (Figure 1F).

s significantly increased in the peritoneum of women with endometriosis compared with the peritoneum of healthy women and those with CPP alone (Figure 1E; P < .001). TRPV1 was elevated in peritoneum (P < .01) and in lesions (P < .05) of women with endometriosis compared with the peritoneum of healthy women (Figure 1F). Figure 1. The neuropeptide TAC1 and nociceptive ion channels are differentially expressed in CPP and in endometriosis. A–F, mRNA concentrations of the neuropeptide TAC1 (A) and the nociceptive ion channels P2RX3 (B), SCN9A (C), SCN11A (D), TRPA1 (E), and TRPV1 (F) were analyzed using QPCR. Concentrations of mRNAs were measured in the peritoneum of women with no pain (control; n = 5), compared with the peritoneum of women with CPP but no obvious underlying pathology (PP; n = 10), and the peritoneum (EP; n = 12) and peritoneal lesions (EL; n = 18) of women with confirmed endometriosis. RQ, relative quantification. Statistical analysis was performed using a one-way ANOVA combined with a Newman Keuls post hoc test. *, P < .05; **, P < .01; ***, P < .001. Asterisks denote statistical significance compared with control, and lines and asterisks denote significance between specific groups.

sed with DPN (P < .05), this was not blocked by the addition of ICI (Figure 2I), and the expression of SCN11A (Figure 2J) and TRPA1 (Figure 2K) did not appear to be ER dependent. In contrast, TRPV1 mRNA concentrations were elevated by both E2 (P < .05) and DPN (Figure 2I; P < .01), and this effect was abrogated by ICI. Figure 2. mRNA concentrations of nociceptive markers are regulated by estrogens in an in vitro model of human sensory neurons with a nociceptor-like phenotype derived from ES cells. A, Human ES cells were differentiated into sensory neurons with a nociceptor-like phenotype using combined small molecule inhibitors. Sensory neurons developed projections (B) and stained positively for neurofilament using immunocytochemistry (C). Scale bar, 200 μM. D, During the differentiation method, RNA was extracted from cells at different time points: pluripotent stem cells (d −7, n = 3), day 1 (n = 4), day 3 (n = 4), day 5 (n = 6), day 7 (n = 4), day 9 (n = 3), day 11 (n = 3), day 17 (n = 3), and day 21 (n = 3). QPCR analysis revealed that mRNA concentrations of the pluripotency marker octamer transcription factor-4 (Oct4) declined as the differentiation procedure progressed. The nociceptive markers TAC1, SCN9A, and SCN11A increased. E, Sensory neuron functionality was determined by stimulating cells with 4 nM capsaicin and recording calcium flux. F, During differentiation from ES cells to mature sensory neurons, ERα mRNAs declined, whereas ERβ inclined. G–L, Mature sensory neurons were incubated with DMSO (vehicle), E2, DPN, or PPT, with or without ICI, and RNA was extracted after 24h. G–I, Using QPCR TAC1 mRNAs were elevated by DPN and PPT compared to DPN + ICI and PPT + ICI. P2RX3 and SCN9A mRNAs were elevated by DPN compared with vehicle control (DMSO). J and K, No significant differences were detected in SCN11A and TRPA1. L, TRPV1 mRNAs were elevated by E2 and DPN. RQ, relative quantification. Statistical analysis was performed using a one-way ANOVA and Newman Keuls post hoc test. *, P < .05; **, P < .01; ***, P < .001. Asterisks denote statistical significance compared with vehicle control (DMSO), and lines and asterisks denote significance between specific treatment groups.

PN. RQ, relative quantification. Statistical analysis was performed using a one-way ANOVA and Newman Keuls post hoc test. *, P < .05; **, P < .01; ***, P < .001. Asterisks denote statistical significance compared with vehicle control (DMSO), and lines and asterisks denote significance between specific treatment groups. Discussion We believe this is the first study to detect differences in the expression of mRNAs encoded by the ion channels TRPA1, P2RX3, SCN9A, and SCN11A in the peritoneum and lesions of women suffering from CPP, some of whom had endometriosis. We believe these insights may be indicative of sensitization of nerves present within the peritoneal lining of women with pain and therefore shed new light on the mechanisms responsible for development of chronic hypersensitivity. In women with active endometriosis, we discovered that the concentrations of mRNAs encoded by SCN11A, TRPA1, and TRPV1 were all significantly increased in the samples of peritoneum. We believe these novel findings are consistent with nociceptive changes taking place within the peritoneum of women with endometriosis, suggesting an increased sensitization of the peritoneum adjacent to the lesion. We detected significant increases in TRPV1 mRNAs in both the peritoneum and in lesions, confirming and extending a previous study (11). Notably, P2RX3 mRNAs were significantly elevated in the peritoneum of women with CPP, regardless of whether they had active endometriosis and would be consistent with a neuropathic component to idiopathic CPP.

ncreases in TRPV1 mRNAs in both the peritoneum and in lesions, confirming and extending a previous study (11). Notably, P2RX3 mRNAs were significantly elevated in the peritoneum of women with CPP, regardless of whether they had active endometriosis and would be consistent with a neuropathic component to idiopathic CPP. To date, researchers have used rat or chick dorsal root ganglia (DRG) neurons as an in vitro model to investigate the function of sensory neurons. However, because DRG neurons represent a mixed population of somatosensory neurons, the development of a differentiation protocol allowing the generation of a population of human sensory neurons with a nociceptor-like phenotype (15) is particularly welcome. In the current study, we detected expression of estrogen receptor mRNAs in our ES cell-derived cultures. Treatment of cells with E2 or the ERβ agonist DPN induced a significant increase in TRPV1 mRNAs that was blocked by the addition of the ER antagonist ICI. In addition, P2RX3, SCN9A, and TAC1 mRNAs were all increased when cells were incubated with the ERβ agonist DPN, and TAC1 was also increased by PPT. We have previously demonstrated DPN-dependent changes in gene expression in endometrial endothelial cells and have shown that these were mediated via specificity protein-1 tethered transcription (16). Bioinformatics revealed the presence of specificity protein-1 transcription factor binding sites in the promoter regions of TRPV1, P2RX3, SCN9A, and TAC1 (Match; Biobase international.com). TRPV1 mRNA was the only transcript that was increased in response to E2 treatment and abrogated by the antiestrogen ICI. Notably, bioinformatics detected several estrogen response elements in the promoter region of TRPV1 and throughout the length of the transcript (Dragon estrogen response element locator version 6.0; National Institute of Health) that could facilitate direct binding of E2-activated ER receptor dimers. Further studies are required to validate these findings.

al estrogen response elements in the promoter region of TRPV1 and throughout the length of the transcript (Dragon estrogen response element locator version 6.0; National Institute of Health) that could facilitate direct binding of E2-activated ER receptor dimers. Further studies are required to validate these findings. Estrogens are reported to have both pro- and antinociceptive properties. Mice exhibit hyperalgesia when estrogens are depleted after ovariectomy (17), and nociceptive responses are lower in ERβ knockout mice during the early stages of inflammation (18). The results of our study suggest estrogens may have a direct impact on nociception by up-regulating the expression of ion channels and would be consistent with reports that the expression of Trpv1 and P2x3 proteins are decreased in DRGs from mice with targeted deletions of Esr2 and Esr2 (19). In summary, the results of this study indicate that therapies targeting P2RX3 may be useful in treating CPP in women with a diverse range of painful etiologies, whereas women with active endometriosis could benefit from treatments targeting TRPV1, TRPA1, SCN9A, or SCN11A. Studies using sensory neurons suggest regulation of TRPV1 by estrogens may provide an explanation for the reduction in pain experienced when ovarian steroids are suppressed and an opportunity for development of novel therapies. Abbreviations: CPPchronic pelvic pain DMSOdimethylsulfoxide DPN2,3-bis(4-hydroxy-phenyl)-propionitrile DRGdorsal root ganglia E217β-estradiol ERestrogen receptor ESembryonic stem ICIfulvestrant PPT4,4′,4′-[4-propyl-(1H)-pyrazole-1,3,5-tryl] trisphenol

In summary, the results of this study indicate that therapies targeting P2RX3 may be useful in treating CPP in women with a diverse range of painful etiologies, whereas women with active endometriosis could benefit from treatments targeting TRPV1, TRPA1, SCN9A, or SCN11A. Studies using sensory neurons suggest regulation of TRPV1 by estrogens may provide an explanation for the reduction in pain experienced when ovarian steroids are suppressed and an opportunity for development of novel therapies. Abbreviations: CPPchronic pelvic pain DMSOdimethylsulfoxide DPN2,3-bis(4-hydroxy-phenyl)-propionitrile DRGdorsal root ganglia E217β-estradiol ERestrogen receptor ESembryonic stem ICIfulvestrant PPT4,4′,4′-[4-propyl-(1H)-pyrazole-1,3,5-tryl] trisphenol QPCRquantitative PCR. Acknowledgments We are grateful to Ronnie Grant for the preparation of figures and our research nurses Ann Doust and Helen Dewart as well as the women who gave informed consent for the use of their material in this study. This work was supported by a Medical Research Council Programme Grant G1100356/1 (to P.T.K.S.) and a Wellbeing of Women Grant R42533 (to A.W.H.). Disclosure Summary: The authors have nothing to disclose.

e; PPi, pyrophosphate. B and C, Homology model of human PAPSS2 based on the 1XNJ structure of PAPSS1 (12), with one high-affinity dimer represented in gray, the other in cyan, visualizing protein truncation by the frameshift mutation p.W462Cfs*3 (B) and the glycine residue affected by the missense mutation p.G270D (C). Interestingly, homozygous PAPSS2 mutations had already been described in 1998 in a consanguineous Pakistani family presenting with spondyloepimetaphyseal dysplasia (SEMD) (5, 6), a subgroup of the large and heterogeneous group of bone dysplasias (7), whereas no clinically overt bone phenotype was found in our female patient (4), with only mild radiological evidence of platyspondyly within the thoracic spine. The individuals in the Pakistani family, 11 men and five women, did not undergo endocrine investigations, and no access was granted to the women for clinical assessment. Three recent papers have described 24 additional individuals with PAPSS2 deficiency (8–10), all of them presenting with clinically overt bone dysplasia. However, serum androgens were measured in only five of them, unanimously revealing low DHEAS but normal circulating active androgens.

Vitamin D is essential for human health, and certain groups may require supplementation to prevent vitamin D deficiency (1). Vitamin D3 (cholecalciferol) is formed endogenously in the skin on exposure to UVB light and is also available from some foods, either naturally or through fortification. Vitamin D2 (ergocalciferol) is present in some fortified foods, supplements, and a small number of natural foods. Either form of the vitamin is used for prophylaxis and/or treatment. However, there is uncertainty over the relative effectiveness of the two forms of vitamin D (2, 3).

r naturally or through fortification. Vitamin D2 (ergocalciferol) is present in some fortified foods, supplements, and a small number of natural foods. Either form of the vitamin is used for prophylaxis and/or treatment. However, there is uncertainty over the relative effectiveness of the two forms of vitamin D (2, 3). Both vitamin D2 and vitamin D3 are effective in the prevention and treatment of vitamin D deficiency rickets and osteomalacia (4–6). Some studies have noted differences in the PTH response after single oral doses of vitamin D2 or vitamin D3 (7, 8), whereas others, with regular oral doses, have reported no difference in changes in PTH levels (4, 9, 10), bone turnover markers (4), or calcium absorption (11). Less clear is the ability of vitamin D2 compared with vitamin D3 to maintain plasma 25-hydroxyvitamin D [25(OH)D], particularly after a single bolus dose. The initial rise in plasma 25(OH)D concentration in response to a single dose of vitamin D2 or vitamin D3 is similar (7), but the subsequent decline may be more rapid after a vitamin D2 dose (2, 3, 7). Differences in the 25(OH)D plasma response to vitamin D2 and vitamin D3 may be due to differences between vitamin D2 and vitamin D3 or their metabolites in affinity for vitamin D binding protein (DBP), hydroxylases, or the vitamin D receptor (VDR). Vitamin D metabolites are primarily transported in plasma by DBP, and their binding affinities for DBP, which may also be altered by genetic variation in DBP, are important determinants of plasma half-life (12). Accordingly, half-lives of vitamin D and 1,25-dihydroxyvitamin D [1,25(OH)2D] are shorter than that of 25(OH)D. Similarly, a shorter half-life for vitamin D2 metabolites may be expected due to their lower DBP binding affinities (13).

ered by genetic variation in DBP, are important determinants of plasma half-life (12). Accordingly, half-lives of vitamin D and 1,25-dihydroxyvitamin D [1,25(OH)2D] are shorter than that of 25(OH)D. Similarly, a shorter half-life for vitamin D2 metabolites may be expected due to their lower DBP binding affinities (13). We have developed a method to measure 25(OH)D plasma half-life using stable isotope-labeled compounds. The purpose of this experimental study was to measure the plasma half-lives of 25(OH)D2 and 25(OH)D3 simultaneously. The study was performed in a rural West African setting and in Cambridge, United Kingdom, countries that differ markedly in vitamin D status, calcium intake, and markers of vitamin D metabolism [eg, 1,25(OH)2D and PTH] (14). In addition, these groups differ in their predominant DBP genotypes (15). Therefore, performing the study in these two populations provided contrasting environments to investigate environmental and genetic influences on vitamin D metabolism.

ium intake, and markers of vitamin D metabolism [eg, 1,25(OH)2D and PTH] (14). In addition, these groups differ in their predominant DBP genotypes (15). Therefore, performing the study in these two populations provided contrasting environments to investigate environmental and genetic influences on vitamin D metabolism. Materials and Methods Study setting The study took place in May and June 2010 at the rural field station of the UK Medical Research Council (MRC) in Keneba, The Gambia (latitude 13°N), where there is little seasonal variability in vitamin D synthesis or status (16), and at MRC Human Nutrition Research, Cambridge, United Kingdom (latitude 52°N) between January and April 2011, when there is little cutaneous vitamin D synthesis. The study was conducted according to the Declaration of Helsinki. In each country, trained staff explained the study to participants and informed, written consent was obtained. All procedures were approved by the joint Gambian Government-MRC Ethics Committee or the UK National Research Ethics Service, Cambridge Committee.

study was conducted according to the Declaration of Helsinki. In each country, trained staff explained the study to participants and informed, written consent was obtained. All procedures were approved by the joint Gambian Government-MRC Ethics Committee or the UK National Research Ethics Service, Cambridge Committee. Participants Thirty-six healthy, nonsmoking males, aged 24–39 years, with a body mass index (BMI) less than 27 kg/m2 participated. One additional Gambian participant was removed from the analysis because his estimated half-lives were more than 3 interquartile ranges beyond the 75th centile and had a major impact on the relationships between the half-life and other variables. Their 25(OH)D, 1,25(OH)2D, and 24,25-dihydroxyvitamin D [24,25(OH)2D] plasma concentrations were typical of the other participants and suggested nothing abnormal about their vitamin D metabolism. Plasma tracer concentrations were, however, variable across the time course of the experiment, suggesting the long half-life was an artifact. Exclusion criteria were recent illness (within 2 weeks); broken bone within the last 3 years; known bone, kidney, or liver disease; taking any prescription medicines; or a hemoglobin level less than 10 g/dL. UK participants were all self-classified as white European and black Gambians were all of Mandinka ethnicity.

Participants Thirty-six healthy, nonsmoking males, aged 24–39 years, with a body mass index (BMI) less than 27 kg/m2 participated. One additional Gambian participant was removed from the analysis because his estimated half-lives were more than 3 interquartile ranges beyond the 75th centile and had a major impact on the relationships between the half-life and other variables. Their 25(OH)D, 1,25(OH)2D, and 24,25-dihydroxyvitamin D [24,25(OH)2D] plasma concentrations were typical of the other participants and suggested nothing abnormal about their vitamin D metabolism. Plasma tracer concentrations were, however, variable across the time course of the experiment, suggesting the long half-life was an artifact. Exclusion criteria were recent illness (within 2 weeks); broken bone within the last 3 years; known bone, kidney, or liver disease; taking any prescription medicines; or a hemoglobin level less than 10 g/dL. UK participants were all self-classified as white European and black Gambians were all of Mandinka ethnicity. Dose preparation Deuterated (three deuterium atoms at positions 6, 19, and 19) 25(OH)D2 [d3-25(OH)D2] and 25(OH)D3 [d3-25(OH)D3] (product numbers 705497 and 705888) (both 97 atom percentage; purity 98%; Sigma-Aldrich) were dissolved in vegetable oil. The solution was protected from light and incubated in a water bath for 1 hour at 35°C and then mixed thoroughly. Aliquots were frozen at −20°C until use. Each 1000 μL dose contained 40 nmol of both deuterated 25(OH)D2 and deuterated 25(OH)D3.

atom percentage; purity 98%; Sigma-Aldrich) were dissolved in vegetable oil. The solution was protected from light and incubated in a water bath for 1 hour at 35°C and then mixed thoroughly. Aliquots were frozen at −20°C until use. Each 1000 μL dose contained 40 nmol of both deuterated 25(OH)D2 and deuterated 25(OH)D3. Study protocols The same study protocols were followed in The Gambia and the United Kingdom. On day 1 and day 21 of the study and after an overnight fast and voiding of the first morning urine, a 2-hour fasting urine was collected from approximately 7:00 am as described previously (17). EDTA and lithium-heparin (LH) (Sarstedt Ltd) blood samples were collected after 1 hour and placed on ice. Height (Leicester Stadiometer; Chasmoors Ltd) and weight (Tanita HD305 scale; Chasmoors Ltd) were measured. After completion of the urine collection, the dose was pipetted onto a small piece of bread and eaten by the participant under supervision, followed by a standardized breakfast (17). Breakfasts were different between countries but were designed to have equal energy content and percentage energy from fat, protein, and carbohydrate. Water was permitted ad libitum. After the dose, participants were asked to refrain from lying down, exercising, or eating for the following 5 hours. Fasted blood samples were collected on day 6 and (±2 d) days 9, 21, 24, 27, 30, and 33 to measure plasma half-lives (17).

ntage energy from fat, protein, and carbohydrate. Water was permitted ad libitum. After the dose, participants were asked to refrain from lying down, exercising, or eating for the following 5 hours. Fasted blood samples were collected on day 6 and (±2 d) days 9, 21, 24, 27, 30, and 33 to measure plasma half-lives (17). Sample processing and laboratory analyses Plasma and urine were treated and stored as described previously (17). Samples from The Gambia were shipped on dry ice to MRC Human Nutrition Research for analysis. EDTA plasma was used for analysis of PTH in singleton by immunoassay (Immulite; Siemens Healthcare Diagnostics Ltd). Between-assay coefficient of variation (CV) was 4.7%. All other assays were performed in duplicate with LH plasma. Albumin was measured on the Konelab 20i (Kone). Within- and between-assay CVs were less than 2% and less than 4%, respectively. Within- and between-assay CVs for total plasma 1,25(OH)2D were 7.5% and 9.0% (IDS Ltd). Performance was monitored using kit and in-house controls and under strict standardization according to ISO 9001:2000. Quality assurance of 25(OH)D, 1,25(OH)2D, and PTH assays were performed as part of the Vitamin D External Quality Assessment Scheme (www.deqas.org) and the National External Quality Assessment Scheme (www.ukneqas.org.uk) and were within acceptable limits. DBP was measured by a radial immunodiffusion assay with a polyclonal antibody (18), and 24,25(OH)2D was analyzed by ultraperformance liquid chromatography and tandem mass spectrometry (UPLC-MS/MS) (19) at Katholieke Universiteit, Leuven, Belgium.

Assessment Scheme (www.ukneqas.org.uk) and were within acceptable limits. DBP was measured by a radial immunodiffusion assay with a polyclonal antibody (18), and 24,25(OH)2D was analyzed by ultraperformance liquid chromatography and tandem mass spectrometry (UPLC-MS/MS) (19) at Katholieke Universiteit, Leuven, Belgium. Derivatized 25(OH)D2/D3 and tracers were measured by UPLC-MS/MS as described previously (17, 20). Tracer and endogenous 25(OH)D2/D3 extraction and analysis were performed separately with different levels of internal standards, d6-25(OH)D2 and d6-25(OH)D3 (Chemaphor Inc). 25(OH)D2, 25(OH)D3, d3-25(OH)D2, and d3-25(OH)D3 were purchased from Sigma-Aldrich. Plasma [200 μL for 25(OH)D3/D2 and 250 μL for tracers] was mixed with 500 or 600 μL acetonitrile, respectively, to release protein-bound vitamin D metabolites. Sample pretreatment and UPLC-MS/MS operating parameters were as described previously (20) with slight modifications applied for the Acquity ultraperformance liquid chromatography module (Waters), interfaced to a 3200 tandem mass spectrometer (AB Sciex). Mass transitions were (mass to charge ratio) 607→298 for 25(OH)D3, 619→298 for 25(OH)D2, 610→301 for d3-25(OH)D3, 622→301 for d3-25(OH)D2, 613→298 for d6-25(OH)D3, and 625→298 for d6-25(OH)D2. Calibrations used a seven-point standard curve containing 25(OH)D3 (0–130 nM), 25(OH)D2 (0–12 nM), or tracers (0–4 nM). Intra- and interassay CVs were less than 10%.

ratio) 607→298 for 25(OH)D3, 619→298 for 25(OH)D2, 610→301 for d3-25(OH)D3, 622→301 for d3-25(OH)D2, 613→298 for d6-25(OH)D3, and 625→298 for d6-25(OH)D2. Calibrations used a seven-point standard curve containing 25(OH)D3 (0–130 nM), 25(OH)D2 (0–12 nM), or tracers (0–4 nM). Intra- and interassay CVs were less than 10%. Genetic analysis DNA was extracted from LH blood pellets with QIAGEN QIAamp DNA Blood maxi kit. The genotyping of DBP and vitamin D-hydroxylase gene variants known to be associated with vitamin D status and of well-described VDR variants was performed at the Vesalius Research Center (Katholieke Universiteit, Leuven, Belgium) by iPLEX technology on a MassARRAY compact analyzer (Sequenom Inc) (21). Derived variables and data analysis The slope of plasma tracer disappearance (kB) was calculated from the line of best fit of the natural log of d3-25(OH)D2 and d3-25(OH)D3 concentrations against time from day 6 to day 33. Half-lives were then calculated in Microsoft Excel 2010 (Microsoft Corp) using the following equation: T12=In(2)kB Statistics were performed in Datadesk 6.3 (Data Description, Inc). Normally distributed data are presented as mean and SD. Skewed data were log transformed and are presented as geometric means and 95% confidence interval. Total 25(OH)D concentration [t25(OH)D] was the sum of 25(OH)D2 and 25(OH)D3.

Derived variables and data analysis The slope of plasma tracer disappearance (kB) was calculated from the line of best fit of the natural log of d3-25(OH)D2 and d3-25(OH)D3 concentrations against time from day 6 to day 33. Half-lives were then calculated in Microsoft Excel 2010 (Microsoft Corp) using the following equation: T12=In(2)kB Statistics were performed in Datadesk 6.3 (Data Description, Inc). Normally distributed data are presented as mean and SD. Skewed data were log transformed and are presented as geometric means and 95% confidence interval. Total 25(OH)D concentration [t25(OH)D] was the sum of 25(OH)D2 and 25(OH)D3. Country differences in half-lives were examined using a linear model with ln d3-25(OH)D concentration as the dependent variable and participant, country and sample time point as independent variables, with an interaction term between country and participant. In paired Student's t tests, there were no differences between day 1 and day 21 values, so biochemical data were used in statistical analysis and are presented as the day 1–21 mean. DBP and 24,25(OH)2D were measured on day 21 and day 1 only, respectively. Country differences were determined using unpaired t tests. Associations between half-lives and t25(OH)D, DBP, bioavailable 25(OH)D [b25(OH)D], or free-25(OH)D [f25(OH)D] were explored using linear regression and with interaction terms between the independent variable and country [eg, t25(OH)D × country]. Results were considered significant when P < .05.

using unpaired t tests. Associations between half-lives and t25(OH)D, DBP, bioavailable 25(OH)D [b25(OH)D], or free-25(OH)D [f25(OH)D] were explored using linear regression and with interaction terms between the independent variable and country [eg, t25(OH)D × country]. Results were considered significant when P < .05. f25(OH)D (nonprotein bound) and b25(OH)D (free and albumin-bound portion) were calculated according to Powe et al (22) and Chun et al (23). The model by Chun et al also accounts for affinity differences due to DBP genotype [Gc-f25(OH)D and Gc-b25(OH)D], but both models use the same association constants for DBP and albumin. DBP isoforms (haplotypes) were defined on the basis of allele combinations of rs7041 and rs4588 single nucleotide polymorphisms (SNPs): Gc1f, T-C; Gc1s, G-C, and Gc2, T-A. To investigate the relationships between half-lives, DBP concentration, and t25OHD with relevant SNPs (21), we used ANOVA with Scheffé post hoc tests and analysis of covariance with the inclusion of country as a confounder (see Table 3). Due to the small group sizes, the relationships with half-life were also examined by combined genotype (diplotype), ie, Gc1f/1f (T-C/T-C) diplotype vs the other diplotypes combined [(Gc1f/1s (T-C, G-C) (n = 10), Gc1f/2 (T-C, T-A) (n = 2), Gc1s/1s (G-C, G-C) (n = 6), and Gc1s/2 (G-C, T-A) (n = 1)].

ounder (see Table 3). Due to the small group sizes, the relationships with half-life were also examined by combined genotype (diplotype), ie, Gc1f/1f (T-C/T-C) diplotype vs the other diplotypes combined [(Gc1f/1s (T-C, G-C) (n = 10), Gc1f/2 (T-C, T-A) (n = 2), Gc1s/1s (G-C, G-C) (n = 6), and Gc1s/2 (G-C, T-A) (n = 1)]. Results Differences between countries were observed in most biochemical parameters including t25(OH)D, 1,25(OH)2D, and PTH (Table 1). There was no difference in the DBP concentration between countries, but f25(OH)D and b25(OH)D were significantly higher in Gambian participants and albumin significantly lower (Table 1). Country differences in tracer concentration at the same time points revealed significant differences at day 30 for 25(OH)D2 (P = .01) and day 33 for 25(OH)D3 (P = .02). The 25(OH)D2 half-life was significantly shorter than the 25(OH)D3 half-life for the countries combined, 13.9 (2.6) days and 15.1 (3.1) days, respectively (P = .001) (Figure 1A). When countries were analyzed separately, there was a significant difference between 25(OH)D2 and 25(OH)D3 half-lives in Gambian (P = .0007) but not UK participants (P = .3) (Figure 1B). Table 1. Anthropometric and Biochemical Characteristics for 18 Gambian and 18 UK Participants and Country Differencesa

Results Differences between countries were observed in most biochemical parameters including t25(OH)D, 1,25(OH)2D, and PTH (Table 1). There was no difference in the DBP concentration between countries, but f25(OH)D and b25(OH)D were significantly higher in Gambian participants and albumin significantly lower (Table 1). Country differences in tracer concentration at the same time points revealed significant differences at day 30 for 25(OH)D2 (P = .01) and day 33 for 25(OH)D3 (P = .02). The 25(OH)D2 half-life was significantly shorter than the 25(OH)D3 half-life for the countries combined, 13.9 (2.6) days and 15.1 (3.1) days, respectively (P = .001) (Figure 1A). When countries were analyzed separately, there was a significant difference between 25(OH)D2 and 25(OH)D3 half-lives in Gambian (P = .0007) but not UK participants (P = .3) (Figure 1B). Table 1. Anthropometric and Biochemical Characteristics for 18 Gambian and 18 UK Participants and Country Differencesa The Gambia UK P Valueb Age, y 29.2 (3.2) 29.3 (4.4) 1.00 Weight, kg 64.5 (8.3) 73.3 (10.9) .01 Height, m 1.74 (0.06) 1.80 (0.07) <.01 BMI, kg/m2 21.3 (2.0) 22.6 (2.3) .08 25(OH)D2, nmol/L 0.6 (0.2) 2.1 (1.0) <.0001 25(OH)D3, nmol/L 68.4 (13.1) 26.5 (10.8) <.0001 t25(OH)D, nmol/L 69.0 (13.2) 28.6 (11.0) <.0001 1,25(OH)2D, pmol/Lc 181 (165, 197) 120 (109, 132) <.0001 24,25(OH)2D, nmol/Ld 6.5 (1.3) 1.9 (1.4) <.0001 PTH, μg/Lc 50.1 (41.9, 60.0) 32.8 (27.4, 39.3) <.01 DBP, mg/Ld 259 (33) 268 (23) .4 Albumin, g/L 36.4 (2.2) 41.0 (2.4) <.0001 b25(OH)D, nmol/L 6.7 (1.2)e 3.0 (1.3)f <.0001 f25(OH)D, pmol/L 20.4 (4.2)e 8.1 (3.3)f <.0001 Gc-b25(OH)D, nmol/Lg 6.6 (1.6)e 4.4 (2.1)f .005 Gc-f25(OH)D, pmol/Lg 20.1 (5.5)e 12.0 (5.6)f .0008 Abbreviations: Gc-b25(OH)D, Gc-genotype-corrected bioavailable 25(OH)D; Gc-f25(OH)D, Gc-genotype-corrected free 25(OH)D.

<.0001 b25(OH)D, nmol/L 6.7 (1.2)e 3.0 (1.3)f <.0001 f25(OH)D, pmol/L 20.4 (4.2)e 8.1 (3.3)f <.0001 Gc-b25(OH)D, nmol/Lg 6.6 (1.6)e 4.4 (2.1)f .005 Gc-f25(OH)D, pmol/Lg 20.1 (5.5)e 12.0 (5.6)f .0008 Abbreviations: Gc-b25(OH)D, Gc-genotype-corrected bioavailable 25(OH)D; Gc-f25(OH)D, Gc-genotype-corrected free 25(OH)D. a Other than age, weight, height, and BMI, data are presented as means (SD) of data collected on days 1 and 21 unless marked. b Country differences as tested by unpaired, two-tailed Student's t test. c Geometric means and 95% confidence interval are given. d 24,25(OH)2D and DBP were measured at one time point only. e There were no differences between f25(OH)D and Gc-f25(OH)D or b25(OH)D and Gc-b25(OH)D in Gambians (both P = .9). f In UK participants, Gc-f25(OH)D and Gc-b25(OH)D were significantly higher than f25(OH)D and b25(OH)D (P = .003 and P = .002, respectively). g Calculated on a subset of the population (Gambia, n = 16; UK, n = 12). Figure 1. Mean, SD, and individual values for the half-lives of 25(OH)D2 (▴) and 25(OH)D3 (●) for countries combined [18 Gambian (solid symbols) and 18 UK (open symbols) participants] (A) and the countries separated (B). Differences between the half-lives of 25(OH)D2 and 25(OH)D3 and country differences are displayed as P values.

ean, SD, and individual values for the half-lives of 25(OH)D2 (▴) and 25(OH)D3 (●) for countries combined [18 Gambian (solid symbols) and 18 UK (open symbols) participants] (A) and the countries separated (B). Differences between the half-lives of 25(OH)D2 and 25(OH)D3 and country differences are displayed as P values. Comparison of half-lives between countries revealed that the 25(OH)D2 half-life was significantly shorter [mean difference (SE)] [2.3 (0.8) day ; P = .005] in Gambian participants than in UK participants, but there was no difference in the 25(OH)D3 half-life [0.9 (1.0) d; P = .4) (Figure 1B]. In the regression analysis, there were no significant relationships between half-lives and t25(OH)D, f25(OH)D, or b25(OH)D (Table 2). Relationships between half-lives and the f25(OH)D index [the molar ratio of t25(OH)D and DBP concentrations (24)] (data not shown) were very similar to those reported for f25(OH)D and b25(OH)D. Plasma DBP concentration was significantly and positively associated with both 25(OH)D2 and 25(OH)D3 half-lives in Gambians and for the countries combined but not in the UK participants (Figure 2 and Table 2). However, DBP concentration × country interactions were not significant. DBP isoform (haplotype) frequencies differed between countries (χ2 test, P < .0001) [Gc1f (T-C): 0.78 vs 0.21; Gc1s (G-C): 0.19 vs 0.71; Gc2 (T-A): 0.03 vs 0.08 for The Gambia and the United Kingdom, respectively]. Table 2. Linear Regression Data for Associations Between Half-Lives and t25(OH)D, DBP, and b25(OH)D and f25(OH)D in 18 Gambian and 18 UK Participants

Comparison of half-lives between countries revealed that the 25(OH)D2 half-life was significantly shorter [mean difference (SE)] [2.3 (0.8) day ; P = .005] in Gambian participants than in UK participants, but there was no difference in the 25(OH)D3 half-life [0.9 (1.0) d; P = .4) (Figure 1B]. In the regression analysis, there were no significant relationships between half-lives and t25(OH)D, f25(OH)D, or b25(OH)D (Table 2). Relationships between half-lives and the f25(OH)D index [the molar ratio of t25(OH)D and DBP concentrations (24)] (data not shown) were very similar to those reported for f25(OH)D and b25(OH)D. Plasma DBP concentration was significantly and positively associated with both 25(OH)D2 and 25(OH)D3 half-lives in Gambians and for the countries combined but not in the UK participants (Figure 2 and Table 2). However, DBP concentration × country interactions were not significant. DBP isoform (haplotype) frequencies differed between countries (χ2 test, P < .0001) [Gc1f (T-C): 0.78 vs 0.21; Gc1s (G-C): 0.19 vs 0.71; Gc2 (T-A): 0.03 vs 0.08 for The Gambia and the United Kingdom, respectively]. Table 2. Linear Regression Data for Associations Between Half-Lives and t25(OH)D, DBP, and b25(OH)D and f25(OH)D in 18 Gambian and 18 UK Participants 25(OH)D2 Half-Life, d 25(OH)D3 Half-Life, d β-Coefficient (SE) P Value β-Coefficient (SE) P Value t25(OH)D, nmol/L·U Gambia 0.058 (0.039) .2 0.090 (0.063) .2 United Kingdom −0.024 (0.055) .7 −0.024 (0.057) .7 Countries combined 0.025 (0.033) .5 −0.043 (0.043) .3 Country .04 .2 25(OH)D*country interaction .2 .2 DBP, mg/L Gambia 0.037 (0.014) .02 0.061 (0.022) .01 United Kingdom 0.017 (0.026) .5 0.004 (0.027) .9 Countries combined 0.031 (0.013) .03 0.042 (0.017) .02 Country .008 .6 DBP*country interaction .1 .1 b25(OH)D, nmol/L Gambia −0.116 (0.469) .8 −0.166 (0.740) .8 United Kingdom −0.268 (0.481) .6 −0.198 (0.500) .7 Countries combined −0.192 (0.337) .6 −0.182 (0.436) .7 Country .3 .9 b25(OH)D*country interaction .8 1.00 f25(OH)D, pmol/L Gambia −452e-6 (0.134) 1.00 −0.010 (0.211) 1.00 United Kingdom 0.100 (0.181) .6 −0.075 (0.187) .7 Countries combined −0.039 (0.107) .7 −0.035 (0.141) .8 Country .2 .8 f25(OH)D*country interaction .7 .8 Gc-b25(OH)D, nmol/L Gambia 0.621 (0.292) .052 0.615 (0.547) .3 United Kingdom −0.389 (0.173) .2 −0.247 (0.215) .3 Countries combined 0.056 (0.216) .8 0.133 (0.303) .7 Country .01 .8 Gc-b25(OH)D*country interaction .02 .2 Gc-f25(OH)D, pmol/L Gambia 0.175 (0.086) .06 0.173 (0.160) .3 United Kingdom −0.143 (0.094) .2 −0.092 (0.077) .3 Countries combined 0.030 (0.070) .7 0.052 (0.098) .6 Country .01 .7 Gc-f25(OH)D*country interaction .02 .2 Abbreviations: Gc-b25(OH)D, Gc-genotype-corrected bioavailable 25(OH)D; Gc-f25(OH)D, Gc-genotype -corrected free 25(OH)D.

ambia 0.175 (0.086) .06 0.173 (0.160) .3 United Kingdom −0.143 (0.094) .2 −0.092 (0.077) .3 Countries combined 0.030 (0.070) .7 0.052 (0.098) .6 Country .01 .7 Gc-f25(OH)D*country interaction .02 .2 Abbreviations: Gc-b25(OH)D, Gc-genotype-corrected bioavailable 25(OH)D; Gc-f25(OH)D, Gc-genotype -corrected free 25(OH)D. Figure 2. Country-specific relationships between the half-lives and total 25(OH)D, DBP, and b25(OH)D concentrations for 25(OH)D2 (▴, Gambia; ▵, UK) and 25(OH)D3 (●, Gambia; ○, UK) in 18 Gambian and 18 UK participants, and Gc-b25(OH)D in 16 Gambian and 12 UK participants. The charts show individual values and country-specific regression lines. Regression line statistics are shown in Table 2. 25(OH)D2 half-lives were significantly different between the rs7041 genotypes and the combined genotypes (Table 3), and there was a trend for the same pattern for 25(OH)D3 half-life; inclusion of country decreased the significance of the relationships with 25(OH)D2. t25(OH)D was significantly different between genotypes, and this was strengthened by correction for country (Table 3). Gc-f25(OH)D and Gc-b25(OH)D were similar to f25(OH)D and b25(OH)D in the Gambians but were higher in the UK participants (Table 1). Relationships between half-lives and either f25(OH)D and b25(OH)D or Gc-f25(OH)D and Gc-b25(OH)D were similar (Figure 2 and Table 2), but relationships were different between countries for the 25(OH)D2 half-life and Gc-f25(OH)D and Gc-b25(OH)D as confirmed by the significant country interaction (Table 2).

ts (Table 1). Relationships between half-lives and either f25(OH)D and b25(OH)D or Gc-f25(OH)D and Gc-b25(OH)D were similar (Figure 2 and Table 2), but relationships were different between countries for the 25(OH)D2 half-life and Gc-f25(OH)D and Gc-b25(OH)D as confirmed by the significant country interaction (Table 2). Table 3. Genotype Frequencies and 25(OH)D2 and 25(OH)D3 Half-Lives, t25(OH)D and DBP Concentration Group Values and Differences by Genotypea

ts (Table 1). Relationships between half-lives and either f25(OH)D and b25(OH)D or Gc-f25(OH)D and Gc-b25(OH)D were similar (Figure 2 and Table 2), but relationships were different between countries for the 25(OH)D2 half-life and Gc-f25(OH)D and Gc-b25(OH)D as confirmed by the significant country interaction (Table 2). Table 3. Genotype Frequencies and 25(OH)D2 and 25(OH)D3 Half-Lives, t25(OH)D and DBP Concentration Group Values and Differences by Genotypea Gene, Variant Genotype nb 25(OH)D2 Half-Life, d P Valuea 25(OH)D3 Half-Life, d P Valuea t25(OH)D, nmol/L P Valuea DBP mg/L P Valuea DBP, rs7041 TT 11 12.5 (1.8) .0009 (.052)c 14.1 (2.4) .07 (.07) 58.3 (14.2) .03 (.02)d 257 (41) .1 (.2) GT 13 14.9 (2.0) 16.6 (2.9) 52.7 (28.0) 264 (20) GG 6 16.1 (1.4) 15.2 (1.4) 29.3 (12.0) 287 (14) Country .3 .6 <.0001 .6 DBP, rs4588 CC 27 13.7 (2.6) .8 (.9) 15.0 (3.1) .6 (.6) 52.4 (24.6) .5 (.9) 264 (32) .8 (.9) CA 3 14.1 (2.4) 14.0 (0.6) 42.2 (17.2) 269 (13) AA 0 Country .03 .9 <.0001 .3 DBP, Gc1f/1f (TT, CC) 1f/1fe 9 12.0 (1.2) .0001 (.007) 14.1 (2.7) .1 (.08) 62.4 (9.0) .1 (.03) 253 (44) .1 (.2) Otherse 19 15.1 (1.9) 15.8 (2.7) 47.0 (25.8) 272 (20) Country .3 .3 <.0001 .9 CYP24A1, rs6013897 TT 17 14.8 (2.3) .2 (.08) 15.9 (3.1) .1 (.1) 55.3 (26.2) .4 (.2) AT 11 13.5 (2.9) 14.1 (2.4) 47.7 (19.9) AA 0 Country .0003 .3 <.0001 CYP2R1, rs10741657 GG 10 13.7 (2.9) .3 (.4) 15.9 (3.8) .6 (.5) 64.4 (18.5) .02 (.95)f GA 16 13.7 (2.1) 14.8 (2.4) 43.3 (22.6) AA 2 16.1 (0.6) 13.9 (0.3) 27.2 (5.5) Country .08 .7 <.0001 VDR, rs731236 (Taq1) TT 11 12.8 (2.1) .1 (.5) 14.3 (2.9) .4 (.5) 59.6 (19.4) .2 (.6) TC 18 14.6 (2.8) 15.5 (3.1) 47.6 (25.7) CC 1 16.5 (0) 17.6 (0) 17.3 (0) Country .03 .9 <.0001 VDR, rs739837 (Apa1) TT 8 14.0 (2.3) .6 (.8) 14.6 (2.3) .4 (.4) 52.7 (21.7) 1.00 (.4) GT 23 14.0 (2.7) 15.4 (3.1) 48.9 (25.8) GG 1 11.4 (0) 11.3 (0) 53.3 (0) Country .02 .9 <.0001 VDR, rs10735810 (Fok1) CC 16 13.4 (2.5) .3 (.5) 14.1 (2.4) .06 (.09) 59.6 (21.1) .08 (.3) TC 13 15.0 (2.5) 16.6 (2.9) 40.7 (23.7) TT 1 14.5 (0) 15.5 (0) 62.4 (0) Country .002 .8 a ANOVA for genotype differences by genotype and Scheffé post hoc test for paired differences. Bracketed P values are for analysis of covariance models that include country.

.4 (2.5) .3 (.5) 14.1 (2.4) .06 (.09) 59.6 (21.1) .08 (.3) TC 13 15.0 (2.5) 16.6 (2.9) 40.7 (23.7) TT 1 14.5 (0) 15.5 (0) 62.4 (0) Country .002 .8 a ANOVA for genotype differences by genotype and Scheffé post hoc test for paired differences. Bracketed P values are for analysis of covariance models that include country. b DNA was available for 35 participants, but low DNA yield prevented genotyping in three individuals. The remaining variation in sample sizes is due to not all SNPs being called. c TT vs GT, P = .01; TT vs GG, P = .002. d TT vs GG, P = .04. e Gc1f/1f combined genotype is assigned by TT and CC in SNPs rs7041 and rs4588, respectively. “Others” consists of Gc1f/1s (TG, CC) (n = 10), Gc1f/2 (TT, CA) (n = 2), Gc1s/1s (GG, CC) (n = 6), and Gc1s/2 (TG, CA) (n = 1). f No significant individual comparisons (GG vs GA, P = 0.06; GG vs AA, P = 0.09).

d TT vs GG, P = .04. e Gc1f/1f combined genotype is assigned by TT and CC in SNPs rs7041 and rs4588, respectively. “Others” consists of Gc1f/1s (TG, CC) (n = 10), Gc1f/2 (TT, CA) (n = 2), Gc1s/1s (GG, CC) (n = 6), and Gc1s/2 (TG, CA) (n = 1). f No significant individual comparisons (GG vs GA, P = 0.06; GG vs AA, P = 0.09). Discussion 25(OH)D half-life is a measure of 25(OH)D expenditure and is determined by CYP27B1 and CYP24A1 enzyme activity and factors that affect 25(OH)D transport and delivery to cells. In this study 25(OH)D2 half-life was shorter than 25(OH)D3 half-life for the countries combined and in Gambian participants when the countries were examined separately. 25(OH)D2 half-life, but not 25(OH)D3 half-life, was shorter in the Gambian compared with the UK participants. The DBP concentration significantly predicted 25(OH)D2 and 25(OH)D3 half-lives in the combined and Gambian models but not in the UK participants, although the DBP concentration country interaction was not significant. These different country relationships may be related to differences in DBP genotype.

UK participants. The DBP concentration significantly predicted 25(OH)D2 and 25(OH)D3 half-lives in the combined and Gambian models but not in the UK participants, although the DBP concentration country interaction was not significant. These different country relationships may be related to differences in DBP genotype. These data may partly explain the previous findings that equal doses of vitamin D2 and vitamin D3 do not equally maintain plasma 25(OH)D. The affinity of vitamin D2 metabolites for DBP is lower than that of vitamin D3 metabolites (13) and may result in proportionally higher f25(OH)D2 available for metabolism. This, together with differences in hydroxylase affinity for vitamin D2 and vitamin D3 metabolites, may explain the shorter half-life of 25(OH)D2 compared with 25(OH)D3. However, our data also suggest that these differences may not be consistent between populations, whether due to environmental or genetic factors. Our data support the assertion that the initial 25(OH)D plasma response after doses of vitamin D2 and vitamin D3 is comparable (7), reflecting similar intestinal absorption, but that differences in attained 25(OH)D concentration are observed after a period of days or weeks (3, 4, 7, 8, 25), suggesting differences in clearance. However, these observations may also be related to dosing frequency because studies that have given daily doses of vitamin D2 or vitamin D3 have observed a small (4) or no (2, 10, 26) difference in the 25(OH)D response.

are observed after a period of days or weeks (3, 4, 7, 8, 25), suggesting differences in clearance. However, these observations may also be related to dosing frequency because studies that have given daily doses of vitamin D2 or vitamin D3 have observed a small (4) or no (2, 10, 26) difference in the 25(OH)D response. Other factors may also influence the conclusions of these earlier studies and include inadequate analytical specificity, not controlling for UVB exposure and the use of large doses (26), as well as different dose vehicles (4, 7, 25) or foods (10, 26). Labeled tracer compounds can be differentiated from endogenous vitamin D and permit the use of small doses that do not perturb normal calcium, phosphate, or vitamin D metabolism or status (17). The nature of intervention studies and the difficulty in performing crossover studies in which there is a strong seasonal influence has meant that earlier studies have not directly compared vitamin D2 and vitamin D3 or 25(OH)D2 and 25(OH)D3. In this study, tracers were administered together and thereby interindividual and seasonal effects were eliminated.

d the difficulty in performing crossover studies in which there is a strong seasonal influence has meant that earlier studies have not directly compared vitamin D2 and vitamin D3 or 25(OH)D2 and 25(OH)D3. In this study, tracers were administered together and thereby interindividual and seasonal effects were eliminated. DBP concentration and genotype may modify vitamin D metabolism and function, eg, affecting bone mineral density (22) and immune cell activity (23). Two SNPs in the DBP gene, rs4588 and rs7041, give rise to three polymorphic isoforms, designated Gc1s, Gc1f, and Gc2, that may predict plasma DBP and 25(OH)D concentrations (27). The biological relevance of differences in binding affinity between these isoforms is not clear (23, 28). We found differences in half-lives, t25(OH)D and DBP concentration when analyzed by DBP genotype. Gc1f/1f homozygotes, hypothetically considered to have the highest binding affinity for 25(OH)D, had a shorter half-life. Associations with genetic variability in DBP may be confounded by race (29). In our study, although all Gc1f/1f homozygotes were Gambian, the difference between genotypes remained significant for 25(OH)D2 half-life after correction for t25OHD, DBP concentration, and country. DBP genotypes are hypothesized to vary between populations as an adaptation to lower UVB exposures at higher latitudes (15). In our study, DBP isoform frequencies were similar to those published previously from the same countries (15) with a higher frequency of Gc1f in Gambians and a higher frequency of Gc1s in the UK participants. The Gambian participants had a higher 25(OH)D status, and the UK group, with the supposedly lower-affinity genotype and expected lower UVB exposure, a lower 25(OH)D status. The different relationships between half-life and DBP concentration between these ethnic groups may be related to differences in the frequencies of the DBP genotypes and in the vitamin D supply and/or modified by the relationship between them (29, 30).

y genotype and expected lower UVB exposure, a lower 25(OH)D status. The different relationships between half-life and DBP concentration between these ethnic groups may be related to differences in the frequencies of the DBP genotypes and in the vitamin D supply and/or modified by the relationship between them (29, 30). Free-25(OH)D or b25(OH)D have been suggested as better markers of vitamin D function than t25(OH)D (31) and are calculated using DBP and albumin concentrations and affinity constants for 25(OH)D. Correction factors may also be included for DBP genotype (23). Gambian participants had higher calculated f25(OH)D and b25(OH)D because of higher plasma t25(OH)D and lower plasma albumin, but as a proportion of t25(OH)D, there was no difference between the countries, and this may explain the absence of larger differences in the 25(OH)D3 half-life between countries. In contrast to other reports (32), DBP concentration did not influence country differences in f25(OH)D and b25(OH)D. Inclusion of DBP genotype increased estimates of f25(OH)D and b25(OH)D in the UK group, which had a lower frequency of the Gc1f allele but had little effect on the levels in The Gambia. However, differences between countries remained. The associations between half-lives and DBP concentration and DBP genotype, rather than t25(OH)D, f25(OH)D, or b25(OH)D, suggest that DBP concentration and genotype are more important determinants of the 25(OH)D half-life than t25(OH)D. When corrected for genotype, there was a trend for a positive association between the Gc-f25(OH)D, Gc-b25(OH)D, and 25(OH)D2 half-life in the Gambian participants. The lack of a relationship with f25(OH)D, and a relationship with DBP concentration, may be explained by an overall higher 25(OH)D expenditure through pathways that are dependent on internalization of the 25(OH)D-DBP complex, such as in the kidney and muscle, rather than pathways dependent on f25(OH)D (eg, immune cells).

e lack of a relationship with f25(OH)D, and a relationship with DBP concentration, may be explained by an overall higher 25(OH)D expenditure through pathways that are dependent on internalization of the 25(OH)D-DBP complex, such as in the kidney and muscle, rather than pathways dependent on f25(OH)D (eg, immune cells). We hypothesized that the lower calcium intake, higher plasma PTH, 1,25(OH)2D, and 25(OH)D in the Gambian population (14) would result in higher production rates of both 1,25(OH)2D and 24,25(OH)2D and consequently a shorter 25(OH)D half-life (33–35). Although 1,25(OH)2D and 24,25(OH)2D were higher in Gambians, we observed a shorter half-life for 25(OH)D2 only. This may be because differences were detected more readily with 25(OH)D2 due to its lower binding affinity for DBP, which may be more affected by genetic variation in DBP or to differences in the rates of hydroxylation. It might be hypothesized that in The Gambia, the more common, higher-affinity Gc1f-variant may influence the availability of vitamin D and 25(OH)D for hepatic hydroxylation to 25(OH)D and renal production of 1,25(OH)2D, respectively, to ensure 25(OH)D supply in an environment with low calcium intake.

of hydroxylation. It might be hypothesized that in The Gambia, the more common, higher-affinity Gc1f-variant may influence the availability of vitamin D and 25(OH)D for hepatic hydroxylation to 25(OH)D and renal production of 1,25(OH)2D, respectively, to ensure 25(OH)D supply in an environment with low calcium intake. Half-lives measured in this, and our previous study (17) (10–24 days), tended toward the lower end of those previously reported in healthy individuals using radiolabeled tracers (summarized in reference 17). These differences may be due to differences in metabolism, dose used, timing of sample collection, or analytical methods. They are also shorter than estimates of tissue 25(OH)D half-life of approximately 3 months from depletion-type experiments in male submariners (36). However, a longer half-life might be expected in depletion studies because of the reduction in 24,25(OH)2D production when 25(OH)D concentration decreases (14, 37) and because it is likely that 25(OH)D produced or released from body stores (38) may contribute to the plasma pool, thereby lengthening the estimated half-life. Differences between tracer and depletion experiments may also be related to the extent of tracer equilibration between body pools (39). Our working model for half-life estimations using tracers consists of two exchanging pools that could be identified as the plasma pool and extravascular pool, between which there is assumed to be free exchange of protein-bound 25(OH)D. In contrast, studies performed over a longer period may include the mobilization of 25(OH)D from a third, deeper body pool and/or synthesis of 25(OH)D from stored vitamin D (38, 40). However, such differences in methodology should not affect the utility of shorter-term tracer methods to investigate changes in 25(OH)D half-life due to, for example, acute or chronic changes in 1,25(OH)2D production (34).

5(OH)D from a third, deeper body pool and/or synthesis of 25(OH)D from stored vitamin D (38, 40). However, such differences in methodology should not affect the utility of shorter-term tracer methods to investigate changes in 25(OH)D half-life due to, for example, acute or chronic changes in 1,25(OH)2D production (34). This study has limitations. The sample size was small and may have limited the ability to find significant relationships, particularly for different genotypes. Based on our previous study with 25(OH)D2 (17), this study was powered to detect a difference of 2.5 or more days between countries (5% significance and 80% power). To reduce variance due to factors other than those directly related to vitamin D metabolism, we restricted our study to healthy, young men of similar BMI. Analyses may be confounded by the associations between country, genotype, and 25(OH)D status, among others. We have not compared calculated f25(OH)D with direct measurements, but confirmation of differences in f25(OH)D2 and f25(OH)D3 would be difficult. Further studies are required to determine whether vitamin D metabolism is affected by sex, age, diet, or body composition, and larger studies are necessary to confirm the influence of genetic polymorphisms suggested in this study.

surements, but confirmation of differences in f25(OH)D2 and f25(OH)D3 would be difficult. Further studies are required to determine whether vitamin D metabolism is affected by sex, age, diet, or body composition, and larger studies are necessary to confirm the influence of genetic polymorphisms suggested in this study. In conclusion, our results suggest there are differences in 25(OH)D2 and 25(OH)D3 plasma half-life, but these differences may differ between populations. DBP concentration and genotype may influence the 25(OH)D half-life. It is likely that both DBP-mediated (including renal/endocrine effects) and f25(OH)D cellular uptake (extrarenal) are reflected in measures of vitamin D expenditure. The dual-isotope approach allowed the direct comparison of 25(OH)D2 and 25(OH)D3 half-lives in the same individual. Oral doses, relatively infrequent sampling, and small sample volumes make this a field-friendly method that could be applied to larger cohorts. These factors, combined with sensitive and specific liquid chromatography and tandem mass spectrometry, provide a robust method with which to explore vitamin D metabolism. Abbreviations: BMIbody mass index b25(OH)Dbioavailable 25(OH)D CVcoefficient of variation DBPvitamin D binding protein f25(OH)Dfree-25(OH)D LHlithium-heparin MRCMedical Research Council 1,25(OH)2D1,25-dihydroxyvitamin D 24,25(OH)2D24,25-dihydroxyvitamin D 25(OH)D25-hydroxyvitamin D SNPsingle-nucleotide polymorphism t25(OH)Dtotal 25(OH)D concentration UPLC-MS/MSultraperformance liquid chromatography and tandem mass spectrometry VDRvitamin D receptor.

DBPvitamin D binding protein f25(OH)Dfree-25(OH)D LHlithium-heparin MRCMedical Research Council 1,25(OH)2D1,25-dihydroxyvitamin D 24,25(OH)2D24,25-dihydroxyvitamin D 25(OH)D25-hydroxyvitamin D SNPsingle-nucleotide polymorphism t25(OH)Dtotal 25(OH)D concentration UPLC-MS/MSultraperformance liquid chromatography and tandem mass spectrometry VDRvitamin D receptor. Acknowledgments We thank the staff of Medical Research Council Keneba, The Gambia, especially Buba Sise, Ebrima Sise, and Ebrima Danso; the Medical Research Council-Human Nutrition Research: Janet Bennett, Shailja Nigdikar (biochemistry analysis), and Jenny Thompson (fieldwork coordination). We also thank Erik van Herck and Ivo Jans (Katholieke Universiteit, Leuven, Belgium) for DBP and 24,25(OH)2D analyses and Rene Chun (University of California, Los Angeles, Los Angeles, California) for sharing the free vitamin D calculator. This work was jointly supported by the Medical Research Council (Program Grants U105960371, U123261351, and MC-A760-5QX00) and the Department for International Development under the Medical Research Council/Department for International Development Concordat agreement. Disclosure Summary: The authors have nothing to declare.

Dehydroepiandrosterone (DHEA) can be converted to its inactive sulfate ester, DHEA sulfate (DHEAS), or toward active androgens via androstenedione and T to the most potent androgen, 5α-dihydrotestosterone (DHT). It was previously assumed that DHEA and DHEAS are continuously interconverted, with DHEAS serving as a circulating pool for reactivation to DHEA, and ultimately sex steroids. However, this concept was called into question by studies suggesting that DHEA sulfation by the enzyme DHEA sulfotransferase, SULT2A1, is the predominant reaction, and the conversion back to DHEA through the enzyme steroid sulfatase is only a rare occurrence (1, 2), except for distinct tissues with ample steroid sulfatase activity, such as placenta and cancers of prostate, breast, endometrium, and colon (3).

the enzyme DHEA sulfotransferase, SULT2A1, is the predominant reaction, and the conversion back to DHEA through the enzyme steroid sulfatase is only a rare occurrence (1, 2), except for distinct tissues with ample steroid sulfatase activity, such as placenta and cancers of prostate, breast, endometrium, and colon (3). We previously described a female patient with clinical and biochemical evidence of androgen excess and concurrently very low serum DHEAS (4). She presented with premature pubarche at 6 years of age and then progressed to a clinically overt polycystic ovary syndrome (PCOS) phenotype, with acne, hirsutism, and eventually secondary amenorrhea at the age of 13 years. We hypothesized that impaired DHEA sulfation would explain the concurrent findings of low DHEAS and increased active androgens. Genetic analysis revealed compound heterozygous mutations in the PAPSS2 gene encoding human PAPS synthase 2, which provides the sulfate donor PAPS (3′-phospho-adenosine-5′-phosphosulfate) to all human sulfotransferases including SULT2A1 (Figure 1A). Functional in vitro analysis of the mutant PAPSS2 proteins demonstrated significantly impaired DHEA sulfation (4).

ous mutations in the PAPSS2 gene encoding human PAPS synthase 2, which provides the sulfate donor PAPS (3′-phospho-adenosine-5′-phosphosulfate) to all human sulfotransferases including SULT2A1 (Figure 1A). Functional in vitro analysis of the mutant PAPSS2 proteins demonstrated significantly impaired DHEA sulfation (4). Figure 1. In silico analysis of the mutant PAPSS2 proteins. A, Either DHEA is converted via to T and DHT, activating the androgen receptor, or DHEA is sulfated by DHEA sulfotransferase (SULT2A1), which requires provision of the universal sulfate donor PAPS, generated by successive ATP sulfurylase and APS kinase activities of PAPSS2. A, androstenedione; APS, adenosine 5′-phosphosulfate; PAPS, 3′-phosphoadenosine 5′-phosphosulfate; PAP, 3′-phosphoadenosine 5′-phosphate; PPi, pyrophosphate. B and C, Homology model of human PAPSS2 based on the 1XNJ structure of PAPSS1 (12), with one high-affinity dimer represented in gray, the other in cyan, visualizing protein truncation by the frameshift mutation p.W462Cfs*3 (B) and the glycine residue affected by the missense mutation p.G270D (C).

he women for clinical assessment. Three recent papers have described 24 additional individuals with PAPSS2 deficiency (8–10), all of them presenting with clinically overt bone dysplasia. However, serum androgens were measured in only five of them, unanimously revealing low DHEAS but normal circulating active androgens. Here we have studied the biochemical and clinical consequences of PAPSS2 deficiency in a family with two brothers compound heterozygous for two novel PAPSS2 mutations, who presented with clinically overt SEMD, low serum DHEAS, but normal serum androgens. We carried out an integrated in silico and in vitro mutation analysis as well as detailed in vivo studies of DHEA sulfation capacity and androgen synthesis in the patients and their heterozygous parents. Patients and Methods Patients The first index case, P1, was born after an uneventful pregnancy at 37 weeks gestation (birth weight, 2580 g; −1.0 SD score [SDS]); birth length not recorded). Short stature with disproportionately short arms and legs was noticed during the first few months after birth. Psychomotor development was normal. At 4.6 years of age, he underwent detailed review by a clinical geneticist, documenting disproportionate growth and short stature with scoliosis, consistent with the phenotypic features of SEMD. His growth curve is shown in Supplemental Figure 1A. A skeletal survey showed scoliosis, platyspondyly with coronal clefts, horizontal acetabular roofs with small iliac wings, small femoral epiphyses, and slight metaphyseal changes, but normal hands (Supplemental Figure 2, A–D).

nt with the phenotypic features of SEMD. His growth curve is shown in Supplemental Figure 1A. A skeletal survey showed scoliosis, platyspondyly with coronal clefts, horizontal acetabular roofs with small iliac wings, small femoral epiphyses, and slight metaphyseal changes, but normal hands (Supplemental Figure 2, A–D). The patient reported normal age of pubarche (Tanner stage P2 at 12.5 y). At the age of 17.6 years, his height was 145.8 cm (−5.0 SDS), weight was 48.2 kg (body mass index [BMI], 0.7 SDS), head circumference was 56.8 cm (−0.1 SDS), arm span was 150.8 cm, sitting height was 78.5 cm, providing a sitting height/height ratio of 0.54 (+1.9 SDS), and his bone age was 15.5 years. Pubertal development was normal (Tanner stage A2P5G5; testicular volume, 25 mL bilaterally). Serum DHEAS was detectable but low (0.5 μmol/L; male reference range, 2.0–15.0 μmol/L); serum and T were within the reference range (Supplemental Table 1).

eight/height ratio of 0.54 (+1.9 SDS), and his bone age was 15.5 years. Pubertal development was normal (Tanner stage A2P5G5; testicular volume, 25 mL bilaterally). Serum DHEAS was detectable but low (0.5 μmol/L; male reference range, 2.0–15.0 μmol/L); serum and T were within the reference range (Supplemental Table 1). His brother, P2, who was 2.5 years younger, was found to have short extremities upon prenatal ultrasound. He was born at 38 weeks gestation (birth weight, 2860 g [−0.8 SDS]; length not recorded). Psychomotor development was normal. His growth curve is shown in Supplemental Figure 1B. Age of pubarche was normal (12–13 y). A skeletal survey showed skeletal abnormalities similar to his brother, P1 (Supplemental Figure 2, E–H). At the age of 15.1 years, his height was 143.3 cm (−4.1 SDS), weight was 45.3 kg (BMI, 1.3 SDS), span was 150.0 cm, sitting height was 73.7 cm (sitting height/height ratio, +0.3 SDS), and his bone age was 13.5 years. Pubertal development was normal (Tanner stage A1P4G4; testicular volume, 14 mL bilaterally). Serum DHEAS was low (0.44 μmol/L); circulating androgens were normal (Supplemental Table 1).

BMI, 1.3 SDS), span was 150.0 cm, sitting height was 73.7 cm (sitting height/height ratio, +0.3 SDS), and his bone age was 13.5 years. Pubertal development was normal (Tanner stage A1P4G4; testicular volume, 14 mL bilaterally). Serum DHEAS was low (0.44 μmol/L); circulating androgens were normal (Supplemental Table 1). The nonconsanguineous parents had normal body proportions and height (father, 183 cm, −0.1 SDS; mother, 168 cm, −0.4 SDS). The 43-year-old mother reported a normal age of menarche (13 y), but irregular periods from the start. Due to anovulatory oligomenorrhea, all her pregnancies were only achieved after ovulation induction with clomiphene. She did not suffer from acne or hirsutism, and her ovaries did not appear polycystic upon imaging. At the age of 40 years, she developed psoriatic arthritis, and subsequently intolerance to several drugs was noted (for details, see Supplemental Data). The older sister of the two index patients was of normal height (166 cm, −0.7 SDS) and weight (BMI, 23.2 kg/m2 at 19 years); however, she had failed to develop a spontaneous menarche and therefore had been initiated on oral contraceptives for cycle regulation at 13 years of age.

The nonconsanguineous parents had normal body proportions and height (father, 183 cm, −0.1 SDS; mother, 168 cm, −0.4 SDS). The 43-year-old mother reported a normal age of menarche (13 y), but irregular periods from the start. Due to anovulatory oligomenorrhea, all her pregnancies were only achieved after ovulation induction with clomiphene. She did not suffer from acne or hirsutism, and her ovaries did not appear polycystic upon imaging. At the age of 40 years, she developed psoriatic arthritis, and subsequently intolerance to several drugs was noted (for details, see Supplemental Data). The older sister of the two index patients was of normal height (166 cm, −0.7 SDS) and weight (BMI, 23.2 kg/m2 at 19 years); however, she had failed to develop a spontaneous menarche and therefore had been initiated on oral contraceptives for cycle regulation at 13 years of age. Genetic analysis Exome sequencing was undertaken after obtaining appropriate informed consent. Genomic DNA was isolated from peripheral blood using the AUTOPURE LS Instrument (Gentra Systems). Cytogenetic microarray analysis was performed using the CytoScan HD Array (Affymetrix) according to the manufacturer's protocol. Copy number was assessed using Chromosome Analysis Suite software (Affymetrix). Whole exome sequencing was performed on DNA fragmented into 200- to 400-bp fragments using Adaptive Focused Acoustics (Covaris Inc) shearing according to the manufacturer's instructions. The exome was captured by Nimblegen SeqCap EZ V2 kit (Roche Nimblegen, Inc) in combination with Illumina paired end library preparation and 2 × 100 bp sequencing with at least 70x mean coverage. Downstream analyses included demultiplexing (CASAVA software; Illumina Inc), sequence quality control, capture quality control, single nucleotide polymorphism calling, and indel (insertions and deletions) calling using different software applications as described by Santen (11). Identified mutations were confirmed by Sanger sequencing (primer sequences available on request).

are; Illumina Inc), sequence quality control, capture quality control, single nucleotide polymorphism calling, and indel (insertions and deletions) calling using different software applications as described by Santen (11). Identified mutations were confirmed by Sanger sequencing (primer sequences available on request). In silico modeling of the mutant PAPSS2 protein A homology model of human PAPSS2 was built on the known crystal structure 1XNJ of PAPSS1 (12), as described previously (13), and the affected Gly270 residue was visualized using YASARA (14). A sequence alignment of 101 vertebrate and invertebrate sequences (15) was extended to plant, fungal, and yeast sequences of ATP sulfurylase. The alignment was created using Clustal W (16). In vitro functional analysis of the mutant PAPSS2 proteins Expression vectors containing the human PAPSS2 coding sequence harboring the newly identified mutations were created by site-directed mutagenesis with DpnI selection in the pcDNA6 and pEGFP-N1 vector backgrounds, as previously described (17).

In silico modeling of the mutant PAPSS2 protein A homology model of human PAPSS2 was built on the known crystal structure 1XNJ of PAPSS1 (12), as described previously (13), and the affected Gly270 residue was visualized using YASARA (14). A sequence alignment of 101 vertebrate and invertebrate sequences (15) was extended to plant, fungal, and yeast sequences of ATP sulfurylase. The alignment was created using Clustal W (16). In vitro functional analysis of the mutant PAPSS2 proteins Expression vectors containing the human PAPSS2 coding sequence harboring the newly identified mutations were created by site-directed mutagenesis with DpnI selection in the pcDNA6 and pEGFP-N1 vector backgrounds, as previously described (17). To assess the impact of the mutant proteins on DHEA sulfation, an in-house clone of HEK293 cells with very low endogenous PAPSS2 and SULT2A1 expression was cotransfected with PAPSS2 and SULT2A1; in addition to wild-type PAPSS2 and the two novel PAPSS2 mutants, we expressed the previously described p.T48R mutant (4) for comparison. Cells were incubated with 250 nm DHEA and 0.2 μCi 3H-DHEA for 2 hours at 37°C; all assays were performed in triplicate. Steroids were extracted as previously described (4) and analyzed on a LablogicAR2000 bioscanner. The conversion rate observed after expression of SULT2A1 alone was subtracted from those observed after coexpression with PAPSS2. For SULT2A1 detection by Western blotting, we used a polyclonal rabbit antibody ab38416 (Abcam), employing the horseradish peroxidase-linked β-actin monoclonal antibody ab20272 (Abcam) for confirmation of equal loading. PAPSS2 protein expression was tested using the anti-PAPSS2 monoclonal antibody ab56393 (Abcam) raised against amino acids 513–613 of PAPSS2, thus not suitable for detection of p.W462Cfs*3. Expression of p.W462Cfs*3 was confirmed at the mRNA level by real-time PCR using the Hs00989921_m1TaqMan probe for PAPSS2 (Life Technologies).

ein expression was tested using the anti-PAPSS2 monoclonal antibody ab56393 (Abcam) raised against amino acids 513–613 of PAPSS2, thus not suitable for detection of p.W462Cfs*3. Expression of p.W462Cfs*3 was confirmed at the mRNA level by real-time PCR using the Hs00989921_m1TaqMan probe for PAPSS2 (Life Technologies). To test the impact of ubiquitination on the expression levels of the p.T48R and p.G270D mutants, these were separately expressed in HEK293 cells, followed by 6-hour incubation with 10 μm of the proteasome inhibitor MG-132 (Sigma) dissolved in dimethylsulfoxide or dimethylsulfoxide alone. DHEA challenge test We employed our previously established in vivo physiology assessment tool, the DHEA challenge test (18, 19), to study in vivo DHEA sulfation and androgen synthesis in the two affected brothers and their heterozygous parents; the sister did not participate because she continued on an oral contraceptive, which precluded detailed analysis of androgen status. In brief, after obtaining written informed consent and with approval from the local ethics committee, all four individuals underwent blood sampling at baseline (9 am) and 30, 60, 90, 120, 180, and 240 minutes after the oral administration of 100 mg DHEA. All four participants collected two consecutive 24-hour urine samples, the first during the 24 hours preceding the DHEA challenge test, and the second starting at the time of DHEA administration.

nt blood sampling at baseline (9 am) and 30, 60, 90, 120, 180, and 240 minutes after the oral administration of 100 mg DHEA. All four participants collected two consecutive 24-hour urine samples, the first during the 24 hours preceding the DHEA challenge test, and the second starting at the time of DHEA administration. Serum and urine steroid measurements Serum steroids were measured by liquid chromatography/tandem mass spectrometry employing a Waters Xevo mass spectrometer with Acquity uPLC system as described previously (20). In brief, serum steroid oxime analysis, which facilitates enhanced detection of DHEA by formation of oxime derivatives of the steroid oxo-groups (21), was employed for the measurement of DHEA, androstenedione, T, and DHT and carried out in positive mode, whereas measurement of serum DHEAS was performed in negative mode. DHEA, androstenedione, T, and DHT were extracted from serum via liquid-liquid extraction followed by derivatization into steroid oximes. Steroids were identified by matching retention times and two mass transitions in comparison to deuterated reference compounds.

asurement of serum DHEAS was performed in negative mode. DHEA, androstenedione, T, and DHT were extracted from serum via liquid-liquid extraction followed by derivatization into steroid oximes. Steroids were identified by matching retention times and two mass transitions in comparison to deuterated reference compounds. Urinary steroid metabolite excretion analysis was carried out by quantitative gas chromatography/mass spectrometry in selected-ion-monitoring analysis mode, as described previously (22). In brief, free and conjugated urinary steroids were extracted by solid-phase extraction, and the conjugates were enzymatically hydrolyzed, followed by recovery of the hydrolyzed steroids by Sep-Pak (Waters) extraction. Specifically, we quantified 5α-reduced androsterone and 5β-reduced etiocholanolone, the two major metabolites of active androgens, which were used to assess net systemic 5α-reductase activity by calculating the ratio of androsterone/etiocholanolone (An/Et). We also measured urinary DHEA, which represents the sum of urinary DHEA and DHEAS excretion, which cannot be distinguished with this method because the hydrolysis step removes the sulfate group. However, in preceding experiments (data not shown) we found that >99% of urinary DHEA originates from DHEAS, ie, proportionate to their respective circulating serum concentrations in the nanomolar and micromolar range, respectively. Thus, for clarity purposes, we have labeled the sum of urinary DHEA and DHEAS in results and figures as urinary DHEAS.

ts (data not shown) we found that >99% of urinary DHEA originates from DHEAS, ie, proportionate to their respective circulating serum concentrations in the nanomolar and micromolar range, respectively. Thus, for clarity purposes, we have labeled the sum of urinary DHEA and DHEAS in results and figures as urinary DHEAS. Statistical analysis SPSS version 21 software (SPSS Inc) was used for data analysis. All data were expressed as mean ± SD unless otherwise stated. Independent samples t tests or Mann-Whitney tests were used as appropriate for comparison between two groups. Differences were considered statistically significant at P < .05. Results Identification of novel PAPSS2 mutations FGFR3 and COL2A1 mutations as distinct causes of the disproportionate short stature had been excluded by direct sequencing during childhood. When the affected brothers were 17 and 15 years old, whole exome sequencing was undertaken in all five family members. This identified compound heterozygosity for two novel PAPSS2 mutations in both boys: a missense mutation (c.809G>A; p.Gly270Asp; p.G270D) and a frameshift mutation (c.1369del; p.W462Cfs*3). The mother was identified as a heterozygous carrier of p.W462Cfs*3, whereas the father and the sister were heterozygous for p.G270D.

is identified compound heterozygosity for two novel PAPSS2 mutations in both boys: a missense mutation (c.809G>A; p.Gly270Asp; p.G270D) and a frameshift mutation (c.1369del; p.W462Cfs*3). The mother was identified as a heterozygous carrier of p.W462Cfs*3, whereas the father and the sister were heterozygous for p.G270D. In silico analysis of the predicted mutant PAPSS2 proteins Both novel mutations affect the ATP sulfurylase domain of the PAPSS2 protein. The p.W462Cfs*3 frameshift mutation will result in a severely truncated protein (Figure 1B). The p.G270D missense mutation affects a buried glycine residue within the ATP sulfurylase domain (Figure 1C), which is positioned in a short and bent β-strand of the PAPSS2 protein, 25Å apart from the bound adenosine 5′-phosphosulfate nucleotide in the 1XNJ structure of PAPSS1 (12) (http://www.rcsb.org/pdb/explore/explore.do?structureId=1xnj). This glycine is invariant in an alignment of 101 vertebrate and invertebrate PAPS synthases (15). Exchanging this highly conserved, small amino acid for a larger and negatively charged aspartic acid, as in p.G270D, is highly likely to impact on protein stability and function.

g/pdb/explore/explore.do?structureId=1xnj). This glycine is invariant in an alignment of 101 vertebrate and invertebrate PAPS synthases (15). Exchanging this highly conserved, small amino acid for a larger and negatively charged aspartic acid, as in p.G270D, is highly likely to impact on protein stability and function. In vitro functional analysis of the mutant PAPSS2 proteins In vitro analysis of DHEA sulfation capacity after coexpression of wild-type and mutant PAPSS2 with SULT2A1 showed a significant reduction for p.G270D, similar to the previously described p.T48R mutant (4), retaining some residual catalytic activity (Figure 2A). By contrast, coexpression of SULT2A1 with the p.W462Cfs*3 mutant almost completely abolished DHEA sulfation (Figure 2A). Western blot analysis confirmed equal SULT2A1 expression; however, the missense mutants p.T48R and p.G270D consistently showed reduced protein expression (Figure 2B), whereas mRNA expression levels determined by real-time PCR were similar (data not shown). Therefore, we reasoned that both mutations might result in destabilization of the resulting mutant PAPSS2 proteins, leading to accelerated degradation by the ubiquitin-proteasome system. To test this, we investigated whether mutant protein expression could be enhanced by the proteasome inhibitor MG-132. This was indeed the case (Figure 2C), confirming enhanced ubiquitination as an explanation for the observed reduced expression of the mutant proteins.

ted degradation by the ubiquitin-proteasome system. To test this, we investigated whether mutant protein expression could be enhanced by the proteasome inhibitor MG-132. This was indeed the case (Figure 2C), confirming enhanced ubiquitination as an explanation for the observed reduced expression of the mutant proteins. Figure 2. Functional in vitro assessment of the mutant PAPSS2 proteins. A, Residual enzyme activity expressed as percentage of wild-type (WT) activity, defined as 100%, based on measurements of the conversion of DHEA to DHEAS in HEK293 cells cotransfected with SULT2A1 and WT or mutant PAPSS2. Error bars represent the mean ± SEM of three independent triplicate experiments. B, Representative Western blot demonstrating equal loading and equal SULT2A1 protein expression, but significantly lower expression of mutant PAPSS2 proteins in HEK293 cells cotransfected with SULT2A1 and WT or mutant PAPSS2. C, Treatment with the proteasome inhibitor MG-132 enhances protein expression of the PAPSS2 mutants p.T48R and p.G270D, confirming increased ubiquitination of the mutant proteins. In vivo analysis of DHEA sulfation by the DHEA challenge test The two brothers carrying compound heterozygous PAPSS2 mutations and their heterozygous parents underwent in vivo assessment to scope their capacity for DHEA sulfation and androgen synthesis, utilizing an oral challenge with 100 mg DHEA with frequent serum sampling and 24-hour urine collection before and after DHEA administration (Figure 3).

ng compound heterozygous PAPSS2 mutations and their heterozygous parents underwent in vivo assessment to scope their capacity for DHEA sulfation and androgen synthesis, utilizing an oral challenge with 100 mg DHEA with frequent serum sampling and 24-hour urine collection before and after DHEA administration (Figure 3). Figure 3. DHEA sulfation and androgen synthesis after an oral challenge with 100 mg DHEA. A–C, Serum concentrations of DHEAS (A), DHEA (B), and the ratio of serum DHEA/DHEAS (C) in the two brothers with compound PAPSS2 mutations (closed symbols) and their heterozygous parents (mother, empty circle; father, empty square) in comparison to healthy female controls (n = 20). D, Serum androstenedione after DHEA in patients, parents and healthy controls. E, Percentage of 24-hour urinary DHEAS excretion in relation to active androgen metabolite excretion after oral DHEA administration and the percentage of 5α-reduced androsterone to 5β-reduced etiocholanolone excretion, demonstrating reduced DHEAS generation and enhanced production of 5α-reduced androgens in the patients and their mother. The excretion pattern in the father resembled that observed in healthy controls (n = 8).

ral DHEA administration and the percentage of 5α-reduced androsterone to 5β-reduced etiocholanolone excretion, demonstrating reduced DHEAS generation and enhanced production of 5α-reduced androgens in the patients and their mother. The excretion pattern in the father resembled that observed in healthy controls (n = 8). At baseline, serum DHEAS was decreased in the two patients and their mother, who carries the major loss-of-function mutation p.W462Cfs*3, whereas the father (heterozygous for p.G270D) had normal DHEAS levels. Serum DHEA was in the higher normal range, and T were within the normal range for age and sex. Urinary steroid metabolite excretion analysis revealed a significantly enhanced baseline 5α-reductase activity in both patients, as represented by the ratio of 5α-reduced over 5β-reduced androgen metabolites, androsterone/etiocholanolone (An/Et) (Table 1). Similarly, their baseline excretion of 5α-reduced androgens was increased, whereas their urinary DHEAS excretion was very low. Androgen excretion in the parents was normal, except for a very low DHEAS excretion and a high 5α-reductase activity in the mother (Table 1). Table 1. Urinary Steroid Metabolite Excretion at Baseline and After Oral Administration of 100 mg DHEA in the Two Brothers Affected by PAPSS2 Deficiency, in Their Heterozygous Parents, and in Healthy Controls

At baseline, serum DHEAS was decreased in the two patients and their mother, who carries the major loss-of-function mutation p.W462Cfs*3, whereas the father (heterozygous for p.G270D) had normal DHEAS levels. Serum DHEA was in the higher normal range, and T were within the normal range for age and sex. Urinary steroid metabolite excretion analysis revealed a significantly enhanced baseline 5α-reductase activity in both patients, as represented by the ratio of 5α-reduced over 5β-reduced androgen metabolites, androsterone/etiocholanolone (An/Et) (Table 1). Similarly, their baseline excretion of 5α-reduced androgens was increased, whereas their urinary DHEAS excretion was very low. Androgen excretion in the parents was normal, except for a very low DHEAS excretion and a high 5α-reductase activity in the mother (Table 1). Table 1. Urinary Steroid Metabolite Excretion at Baseline and After Oral Administration of 100 mg DHEA in the Two Brothers Affected by PAPSS2 Deficiency, in Their Heterozygous Parents, and in Healthy Controls Urinary Steroid Metabolite P1, G270D/W462Cfs*3 P2, G270D/W462Cfs*3 Father, G270D/WT Mother, WT/W462Cfs*3 Postpubertal Boys, n = 10 Adult Men, n = 21 Adult Women, n = 30 Adult Women, n = 8 Age, y 17.6 15.1 48 43 14–19 25–48 25–49 23–39 24-h Urine excretion (μg/24 h) at baseline Androsterone 5751 4325 4450 1249 2491 (986–4928) 2286 (1332–7672) 1163 (324–2819) 893 (150–1746) Etiocholanolone 1363 746 4819 597 1663 (313–2393) 1958 (613–5981) 1199 (507–3205) 893 (94–1990) DHEAS 45 21 602 63 94 (31–1170) 1550 (60–6598) 280 (57–4697) 679 (49–4077) An/Et ratio 4.22 5.80 0.92 2.09 1.71 (0.95–3.15) 1.30 (0.67–3.04) 1.03 (0.35–2.25) 1.11 (0.73–1.60) 24-h Urine excretion (μg/24 h) after oral ingestion of 100 mg DHEA Androsterone 35 755 35 603 13 787 12 888 4556 (778–9756) Etiocholanolone 12 472 5982 14 650 6579 5728 (2516–12 647) DHEAS 415 456 3444 301 6064 (463–27 835) An/Et ratio 2.87 5.95 0.94 1.96 0.77 (0.27–1.11) Abbreviation: WT, wild-type. Data for healthy controls are expressed as median (range).

l ingestion of 100 mg DHEA Androsterone 35 755 35 603 13 787 12 888 4556 (778–9756) Etiocholanolone 12 472 5982 14 650 6579 5728 (2516–12 647) DHEAS 415 456 3444 301 6064 (463–27 835) An/Et ratio 2.87 5.95 0.94 1.96 0.77 (0.27–1.11) Abbreviation: WT, wild-type. Data for healthy controls are expressed as median (range). After administration of DHEA, both patients and their mother had a subnormal rate of DHEAS generation, whereas serum DHEAS in the father increased to the lower reference range (Figure 3A). Serum DHEA peaked in both patients well above the reference range, whereas the parents showed increases comparable to controls (Figure 3B). Serum DHEA/DHEAS ratios were highly increased throughout the challenge test in both affected boys, whereas their father showed values similar to healthy controls, and the mother's results ranged in between (Figure 3C). Serum after DHEA did not differ from levels in healthy controls (Figure 3D). Serum DHT increased after oral DHEA in both boys, which was not observed in their parents. Area under the curve analysis showed reduced values for DHEAS in both patients and their mother (Table 2). Table 2. Area Under the Curve of Four Sampling Hours (AUC0–4 h) for Serum DHEAS, DHEA, Androstenedione, T, and DHT After Oral Administration of 100 mg DHEA at Time 0 Min in the Two Brothers With PAPSS2 Deficiency, Their Heterozygous Parents, and Healthy Female Controls

After administration of DHEA, both patients and their mother had a subnormal rate of DHEAS generation, whereas serum DHEAS in the father increased to the lower reference range (Figure 3A). Serum DHEA peaked in both patients well above the reference range, whereas the parents showed increases comparable to controls (Figure 3B). Serum DHEA/DHEAS ratios were highly increased throughout the challenge test in both affected boys, whereas their father showed values similar to healthy controls, and the mother's results ranged in between (Figure 3C). Serum after DHEA did not differ from levels in healthy controls (Figure 3D). Serum DHT increased after oral DHEA in both boys, which was not observed in their parents. Area under the curve analysis showed reduced values for DHEAS in both patients and their mother (Table 2). Table 2. Area Under the Curve of Four Sampling Hours (AUC0–4 h) for Serum DHEAS, DHEA, Androstenedione, T, and DHT After Oral Administration of 100 mg DHEA at Time 0 Min in the Two Brothers With PAPSS2 Deficiency, Their Heterozygous Parents, and Healthy Female Controls Steroid AUC0–4 h P1, G270D/W462Cfs*3 P2, G270D/W462Cfs*3 Father, G270D/WT Mother, WT/W462Cfs*3 Healthy Female Controls, n = 20 Age, y 17.6 15.1 48 43 21–45 DHEAS, μmol/L*h 517 442 2268 1116 4356 (2426–9145) DHEA, nmol/L*h 10 646 8885 5516 4688 6667 (3755–14 604) Androstenedione, nmol/L*h 2240 1867 1825 1789 1794 (1031–3120) T, nmol/L*h 6320 3639 7343 464 341 (72–850) DHT, nmol/L*h 1019 983 688 a a Abbreviation: WT, wild-type.

Controls, n = 20 Age, y 17.6 15.1 48 43 21–45 DHEAS, μmol/L*h 517 442 2268 1116 4356 (2426–9145) DHEA, nmol/L*h 10 646 8885 5516 4688 6667 (3755–14 604) Androstenedione, nmol/L*h 2240 1867 1825 1789 1794 (1031–3120) T, nmol/L*h 6320 3639 7343 464 341 (72–850) DHT, nmol/L*h 1019 983 688 a a Abbreviation: WT, wild-type. a Serum DHT concentrations measured with liquid chromatography/tandem mass spectrometry were below the limit of detection in female controls. DHT levels in the heterozygous mother were detected, but below the limit of quantitation; therefore, no numbers are provided here. Urine steroid analysis after administration of DHEA revealed significantly reduced generation of DHEAS and enhanced output of androsterone, the major metabolite of DHT, in both patients and their mother. By contrast, the father showed excretion patterns comparable to controls (Figure 3E). Androsterone excretion in the two boys was 2.6-fold higher than in their father, whereas etiocholanolone excretion was normal (Table 1); urinary excretion of 5α-17-hydroxypregnanolone, indicative of the alternative androgen pathway (22), did not increase after DHEA administration (data not shown). After DHEA ingestion, the mother showed reduced DHEAS generation and increased androsterone generation compared to healthy controls (Table 1). The An/Et ratio was significantly raised in both patients and their mother, indicative of increased 5α-reductase activity.

t increase after DHEA administration (data not shown). After DHEA ingestion, the mother showed reduced DHEAS generation and increased androsterone generation compared to healthy controls (Table 1). The An/Et ratio was significantly raised in both patients and their mother, indicative of increased 5α-reductase activity. Discussion Here we carried out an integrated in vitro and in vivo analysis in a family with two brothers affected by PAPSS2 deficiency due to compound heterozygosity for two novel PAPSS2 mutations, p.W462Cfs*3s and p.G270D; their parents and sister were heterozygous carriers. In silico analysis predicted an obvious deleterious impact of the frameshift p.W462Cfs*3 mutation due to truncation of the PAPSS2 ATP sulfurylase domain. This was confirmed in our in vitro cell-based assay, whereas the p.G270D missense mutant exhibited some residual activity. However, Western blotting demonstrated reduced protein expression for p.G270D and the previously characterized missense mutation p.T48R (4). This suggests that the mutant proteins are able to maintain some function but are subject to accelerated degradation due to reduced protein stability. We confirmed this assumption by demonstrating that inhibition of ubiquitination stabilized the mutant proteins in vitro; thus, enhanced ubiquitination represents a mechanism likely to contribute to the disruption of enzymatic activity.

unction but are subject to accelerated degradation due to reduced protein stability. We confirmed this assumption by demonstrating that inhibition of ubiquitination stabilized the mutant proteins in vitro; thus, enhanced ubiquitination represents a mechanism likely to contribute to the disruption of enzymatic activity. Interestingly, the two affected boys had primarily presented with a severe bone phenotype, SEMD, and no apparent evidence of androgen excess. This phenotypic presentation is very similar to the first reported family with PAPSS2 deficiency, the consanguineous Pakistani family with 16 individuals affected by overt SEMD (5, 6). The bone phenotype is thought to be due to impaired proteoglycan sulfation disrupting extracellular matrix formation. The overt SEMD phenotype in the two current patients contrasts with the phenotype in the young girl we previously described (4), who presented with clinical and biochemical signs of androgen excess, manifesting as premature adrenarche and PCOS, but very mild bone abnormalities only visible on x-ray. However, all three patients had low circulating DHEAS, suggesting that in vivo DHEA sulfotransferase activity was disrupted, which should result in enhanced androgen generation from DHEA.

ical signs of androgen excess, manifesting as premature adrenarche and PCOS, but very mild bone abnormalities only visible on x-ray. However, all three patients had low circulating DHEAS, suggesting that in vivo DHEA sulfotransferase activity was disrupted, which should result in enhanced androgen generation from DHEA. To examine the genotype-phenotype variation in more detail, we summarized all reported cases with PAPSS2 deficiency (Table 3). Three recent publications (8–10) have added 24 additional individuals, yielding a total of 43 patients with clinical and radiological phenotype description, whereas functional in vitro analysis of mutant protein activity was only provided by this study and our previous study (4). Table 3 illustrates that in PAPSS2 deficiency, four different bone phenotype variants are observed: 1) overt SEMD with both vertebrae and long bones affected (n = 23); 2) overt brachyolmia with dysplastic changes confined to the spine (n = 15); 3) overt brachyolmia with dysplastic changes confined to the spine with additional minimal epimetaphyseal changes only visible on x-ray (n = 4); and 4) subclinical brachyolmia with radiological changes only, as observed in the first patient we described (4). Table 3. Genotype-Phenotype Spectrum in the 43 Patients With PAPSS2 Deficiency Reported to Date

To examine the genotype-phenotype variation in more detail, we summarized all reported cases with PAPSS2 deficiency (Table 3). Three recent publications (8–10) have added 24 additional individuals, yielding a total of 43 patients with clinical and radiological phenotype description, whereas functional in vitro analysis of mutant protein activity was only provided by this study and our previous study (4). Table 3 illustrates that in PAPSS2 deficiency, four different bone phenotype variants are observed: 1) overt SEMD with both vertebrae and long bones affected (n = 23); 2) overt brachyolmia with dysplastic changes confined to the spine (n = 15); 3) overt brachyolmia with dysplastic changes confined to the spine with additional minimal epimetaphyseal changes only visible on x-ray (n = 4); and 4) subclinical brachyolmia with radiological changes only, as observed in the first patient we described (4). Table 3. Genotype-Phenotype Spectrum in the 43 Patients With PAPSS2 Deficiency Reported to Date First Author, Year (Ref.) Mutant Allele 1 Mutant Allele 2 Sex Country of Origin Clinical Signs of Androgen Excess Serum Androgens (Routine Biochemistry) Bone Phenotype Bone Age Noordam, 2009 (4) p.T48R p.R329* F Turkey PA, AC, HI, SAa Low DHEAS; high A, T, DHT Subclinical BO Advanced Iida, 2013 (10) p.L76Q p.R129Lfs*25 F Turkey AC, HI, CM “Normal” (no details) Overt BO Normal Iida, 2013 (10) p.V540D p.V540D F Turkey AC, HI “Normal” (no details) Overt BO Advanced Iida, 2013 (10) p.V364Rfs*18 p.V364Rfs*18 F Turkey No NR Overt BO Normal Iida, 2013 (10) p.R129Lfs*25 p.R129Lfs*25 2 F Turkey No NR Overt BO Advanced Iida, 2013 (10) p.R129Lfs*25 p.R129Lfs*25 F Turkey No NR Overt BO Normal Iida, 2013 (10) p.Q211Cfs*11 p.Q211Cfs*11 F Turkey No NR Overt BO Normal Iida, 2013 (10) c.27 + 3A>C c.27 + 3A>C F Turkey No NR Overt BO Normal Iida, 2013 (10) p.F125Sfs*24 p.F125Sfs*24 F Syria PP High DHEA Overt BO Advanced Miyake, 2012 (8) p.A113Gfs*18 p.A113Gfs*18 F Turkey No NR Overt BO NR Miyake, 2012 (8) p.A113Gfs*18 p.A113Gfs*18 F Turkey No Low DHEAS Overt BO, MEMC Advanced Miyake, 2012 (8) p.V206Sfs*9 p.R437Gfs*19 F Japan No NR Overt BO, MEMC Advanced Miyake, 2012 (8) c.381 + 2delT c.381 + 2delT F Japan No NR Overt BO, MEMC Advanced Tüysüz, 2013 (9) p.R329* p.R329* F Turkey HI, OM Low DHEAS, normal DHEA, A, T Overt SEMD NR Tüysüz, 2013 (9) p.R329* p.R329* F Turkey No Low DHEAS, normal DHEA, A, T Overt SEMD NR Ahmad, 1998 (6) p.S480* p.S480* 11 M, 5 F Pakistan NR NR Overt SEMD NR Iida, 2013 (10) p.L76Q p.R129Lfs*25 M Turkey No NR Overt BO Normal Iida, 2013 (10) p.C43Y p.C43Y M Turkey PA NR Overt BO Advanced Iida, 2013 (10) p.V364Rfs*18 p.V364Rfs*18 M Turkey No NR Overt BO Advanced Iida, 2013 (10) p.A113Gfs*18 p.A113Gfs*18 2 M Turkey No NR Overt BO NR Miyake, 2012 (8) p.A113Gfs*18 p.A113Gfs*18 M Turkey No NR Overt BO Delayed Miyake, 2012 (8) p.K161Rfs*6 p.I221Sfs*40 M Korea No NR Overt BO, MEMC Advanced Tüysüz, 2013 (9) p.R329* p.R329* 3 M Turkey No Low DHEAS, normal DHEA, A, T Overt SEMD NR Present study p.G270D p.W462Cfs*3 2 M Netherlands Nob Low DHEAS, normal DHEA, A, T Overt SEMD Delayed Abbreviations: F, female; M, male; NR, not reported; PA, premature adrenarche; AC, acne;

M Korea No NR Overt BO, MEMC Advanced Tüysüz, 2013 (9) p.R329* p.R329* 3 M Turkey No Low DHEAS, normal DHEA, A, T Overt SEMD NR Present study p.G270D p.W462Cfs*3 2 M Netherlands Nob Low DHEAS, normal DHEA, A, T Overt SEMD Delayed Abbreviations: F, female; M, male; NR, not reported; PA, premature adrenarche; AC, acne; HI, hirsutism; SA, secondary amenorrhea; CM, clitoromegaly; PP, precocious puberty; OM, oligomenorrhea; A, androstenedione; BO, brachyolmia (short trunk-short stature due to platyspondyly); MEMC, minimal epimetaphyseal changes; SEMD, spondyloepimetaphyseal dysplasia (brachyolmia plus additional clinically overt epimetaphyeseal changes). Overt indicates clinical and radiological changes, and subclinical indicates radiological changes only. Mutation nomenclature is according to Human Genome Variation Society convention and employs PAPSS2 reference sequence NM_001015880.1. a Mother reported PCOS with oligomenorrhea and anovulation. b Mother reported chronic anovulation requiring ovulation induction with clomiphene for conception; older sister had primary amenorrhea.

Overt indicates clinical and radiological changes, and subclinical indicates radiological changes only. Mutation nomenclature is according to Human Genome Variation Society convention and employs PAPSS2 reference sequence NM_001015880.1. a Mother reported PCOS with oligomenorrhea and anovulation. b Mother reported chronic anovulation requiring ovulation induction with clomiphene for conception; older sister had primary amenorrhea. Only 16 of the 43 patients reported to date were assessed for androgen excess; clinical signs including premature adrenarche and PCOS were reported for six of them (Table 3). Serum DHEAS was measured in eight patients and invariably found to be low. Circulating androgens were measured in 10 individuals and increased in two female patients, whereas eight patients had normal androgens (four males, four females) (Table 3). Although circulating androgens were normal in the two affected male patients in this study, urinary steroid metabolite analysis revealed increased 5α-reduced androgens in both of them, indicative of enhanced generation of active androgens in peripheral tissues from DHEA. We confirmed this employing oral administration of the androgen precursor DHEA to assess DHEA sulfation and androgen synthesis capacity, which demonstrated that in the two patients, DHEA was converted to active androgens at a much higher rate than in their father or in healthy controls. The biological consequences of this androgen excess will need to be monitored prospectively, and it is interesting to note that neither boy had signs of premature adrenarche. Our control group consisted of adult women; however, we demonstrated previously that the conversion pattern toward DHEA, DHEAS, and after oral DHEA is identical in men and women (18, 23). In healthy men, DHEA administration does not elicit increases in T or DHT (23); therefore, it is interesting to note that both patients showed a clear increase in circulating DHT after oral DHEA. Using an oral DHEA challenge as a dynamic test, we provided the first direct functional in vivo evidence that impaired DHEA sulfation results in androgen excess.

oes not elicit increases in T or DHT (23); therefore, it is interesting to note that both patients showed a clear increase in circulating DHT after oral DHEA. Using an oral DHEA challenge as a dynamic test, we provided the first direct functional in vivo evidence that impaired DHEA sulfation results in androgen excess. The mother of the two patients studied here reported clinical features consistent with a PCOS phenotype, specifically chronic anovulation requiring ovulation induction. Androgen analysis showed evidence of enhanced 5α-reductase activity at baseline and increased 5α-reduced androgen production after the oral DHEA challenge, with low DHEAS at baseline and subnormal DHEAS generation after DHEA. The mother of our first patient (4) had actually reported a history of PCOS with oligomenorrhea and anovulation, but at the time had not provided serum or urine for more detailed androgen analysis. Coincidentally, both mothers were heterozygous carriers of a major loss-of-function PAPSS2 mutation, and thus it appears reasonable to assume that heterozygosity for a mutant allele does not only result in a subclinical phenotype with low DHEAS but can also impact clinically with features resembling PCOS. Two recent studies looked at the association of common genetic variants (minor allele frequency [MAF] > 5%) in SULT2A1 and PAPSS2 with androgen status (24, 25); future studies with a more detailed phenotyping and genotyping approach will need to investigate impaired DHEA sulfation as a predisposing factor to PCOS by looking for low frequency (MAF < 5%) and rare variants (MAF < 0.5%).

iants (minor allele frequency [MAF] > 5%) in SULT2A1 and PAPSS2 with androgen status (24, 25); future studies with a more detailed phenotyping and genotyping approach will need to investigate impaired DHEA sulfation as a predisposing factor to PCOS by looking for low frequency (MAF < 5%) and rare variants (MAF < 0.5%). In summary, we have presented conclusive in vivo evidence for androgen excess as a consequence of impaired DHEA sulfation in PAPSS2 deficiency. Further studies are needed to determine the effect on drug metabolism in affected patients because sulfation is a key process in the inactivation of drugs and xenobiotics, which may explain the multiple drug intolerances reported by the mother. Studies in PCOS cohorts will need to look more closely at the pattern of androgen excess as a differentiating factor, with mounting evidence that androgen status is closely linked to metabolic risk in PCOS patients (20). Abbreviations: An/Etandrosterone/etiocholanolone BMIbody mass index DHEAdehydroepiandrosterone DHEASDHEA sulfate DHT5α-dihydrotestosterone MAFminor allele frequency PAPS3′-phospho-adenosine-5′-phosphate PAPSS2PAPS synthase 2 PCOSpolycystic ovary syndrome SDSstandard deviation score SEMDspondyloepimetaphyseal dysplasia. Acknowledgments We thank the patients and their parents for their willingness to participate in this study.

DHEAdehydroepiandrosterone DHEASDHEA sulfate DHT5α-dihydrotestosterone MAFminor allele frequency PAPS3′-phospho-adenosine-5′-phosphate PAPSS2PAPS synthase 2 PCOSpolycystic ovary syndrome SDSstandard deviation score SEMDspondyloepimetaphyseal dysplasia. Acknowledgments We thank the patients and their parents for their willingness to participate in this study. This work was supported by the Wellcome Trust (Project Grant 092283, to W.A.; Research Training Fellowship 099909, to M.W.O.), the Medical Research Council UK (Research Training Fellowship G1001964, to J.I.), the European Commission (Marie Curie Fellowship 625451, to J.W.M.), and the Royal College of Physicians (Wolfson Bursary, to P.J.H.). W.O. received unrestricted grant support from Novo Nordisk for whole exome sequencing. Disclosure Summary: The authors have nothing to disclose.

Following identification of the therapeutic effect of glucagon-like peptide-1 (GLP-1), the dipeptidyl-peptidase-4 (DPP-4) inhibitors were developed specifically to delay its rapid degradation in plasma, and hence to enhance the incretin effect in type 2 diabetes (1–3). Vildagliptin achieves prolong and almost complete DPP-4 inhibition, resulting in the extension of meal induced increases in GLP-1 and gastric inhibitory peptide over 24 hours. GLP-1 and gastric inhibitory peptide increase the sensitivity of the α- and β-cells to glucose, which is accepted as their major mechanism of action (4, 5). However, vildagliptin brings about changes that would not be predicted from its actions in the pancreas. It decreases postprandial triglyceride levels and decreases lipolysis as assessed in vivo by palmitate dilution more than can be accounted for by change in plasma insulin concentration (6, 7). This could result in a decrease in liver triglyceride concentration. Vildagliptin has also been shown to increase glucose utilization, as assessed during a two-step hyperinsulinemic euglycemic clamp at the high insulin dose (80 mU), and this could potentially be secondary to a reduction in liver fat (1, 2, 8). Whether hepatic lipid metabolism is specifically affected has not been examined, and there is no information on any modulation of liver triglyceride concentration.

two-step hyperinsulinemic euglycemic clamp at the high insulin dose (80 mU), and this could potentially be secondary to a reduction in liver fat (1, 2, 8). Whether hepatic lipid metabolism is specifically affected has not been examined, and there is no information on any modulation of liver triglyceride concentration. The present randomized, placebo-controlled study was designed to examine the possible effects of vildagliptin on hepatic steatosis and insulin sensitivity. To minimize any indirect metabolic effects due to a large change in ambient plasma glucose levels, people with type 2 diabetes well controlled on metformin alone were studied. Patients and Methods Study protocol A single-center, randomized, double-blind, placebo-controlled, parallel-group study was conducted. Forty-four patients with type 2 diabetes and glycated hemoglobin (HbA1c) ≤ 7.6%, who were treated with metformin, were randomized equally to the DPP-4 inhibitor vildagliptin (50 mg twice a day) and placebo. Two patients from the vildagliptin-treated group (one with multiple myeloma, and the other with atrial fibrillation related to chest infection) and three patients from the placebo group (one with marked deterioration in glycemic control, another with metastatic prostate cancer, and a third who withdrew consent) were withdrawn. None of these events were thought to be drug related.

tiple myeloma, and the other with atrial fibrillation related to chest infection) and three patients from the placebo group (one with marked deterioration in glycemic control, another with metastatic prostate cancer, and a third who withdrew consent) were withdrawn. None of these events were thought to be drug related. Each participant attended one screening visit (wk −4; ie, 4 weeks before baseline assessments) for assessment of inclusion/exclusion criteria. Measurement of liver triglyceride and peripheral and hepatic insulin sensitivity and anthropometric tests were carried out at the Magnetic Resonance Centre on three occasions, each separated by at least 3 days. Randomization to vildagliptin or placebo was then carried out. Each person attended for eight additional visits over the 6-month period of treatment with vildagliptin 50 mg twice a day or placebo. Thereafter, measurements of liver triglyceride and peripheral insulin sensitivity and anthropometric tests were repeated. To permit judgment on the metabolic significance of any changes in liver triglyceride levels, a group of individuals with normal glucose tolerance defined by World Health Organization criteria were matched for age, weight, and sex with the randomized patients with type 2 diabetes and studied (n = 14). The study was approved by the Newcastle and North Tyneside 2 Research Ethics Committee and was registered on the database of clinical trials (ClinicalTrials.gov, ID no. NCT01356381).

Each participant attended one screening visit (wk −4; ie, 4 weeks before baseline assessments) for assessment of inclusion/exclusion criteria. Measurement of liver triglyceride and peripheral and hepatic insulin sensitivity and anthropometric tests were carried out at the Magnetic Resonance Centre on three occasions, each separated by at least 3 days. Randomization to vildagliptin or placebo was then carried out. Each person attended for eight additional visits over the 6-month period of treatment with vildagliptin 50 mg twice a day or placebo. Thereafter, measurements of liver triglyceride and peripheral insulin sensitivity and anthropometric tests were repeated. To permit judgment on the metabolic significance of any changes in liver triglyceride levels, a group of individuals with normal glucose tolerance defined by World Health Organization criteria were matched for age, weight, and sex with the randomized patients with type 2 diabetes and studied (n = 14). The study was approved by the Newcastle and North Tyneside 2 Research Ethics Committee and was registered on the database of clinical trials (ClinicalTrials.gov, ID no. NCT01356381). Subjects Subjects with type 2 diabetes (28 males and 16 females; HbA1c, 6.4 ± 0.1%); with a mean duration of diabetes of 5.7 ± 0.7 years were studied. All were taking metformin and no other oral hypoglycemic agent. None were taking insulin. A group of normal glucose-tolerant subjects with no first-degree relative with diabetes, matched for age, weight, and body mass index, were recruited as controls to provide comparative data. Normal glucose tolerance was demonstrated in all control subjects by a 75-g oral glucose tolerance test (mean fasting plasma glucose, 5.3 mmol/L; 2-h plasma glucose, 5.5 mmol/L). The characteristics of both groups are shown in Table 1. Informed written consent was obtained from all volunteers. All subjects underwent a 10-hour overnight fast and were advised to abstain from vigorous exercise and smoking for 3 days before test days.

ing plasma glucose, 5.3 mmol/L; 2-h plasma glucose, 5.5 mmol/L). The characteristics of both groups are shown in Table 1. Informed written consent was obtained from all volunteers. All subjects underwent a 10-hour overnight fast and were advised to abstain from vigorous exercise and smoking for 3 days before test days. Table 1. Characteristics of Study Groups Vildagliptin Group Placebo Group Type 2 Diabetes Group Controls n 22 22 44 14 Age, y 65.2 ± 0.7 58.9 ± 1.6 61.4 ± 1.0 59.0 ± 2.2 Weight, kg 83.0 ± 3.2 91.8 ± 2.4 87.2 ± 2.0 86.7 ± 3.7 BMI, kg/m2 29.4 ± 0.8 31.1 ± 0.6 30.3 ± 0.5 29.6 ± 1.0 Body fat, % 31.5 ± 1.6 30.7 ± 1.8 31.1 ± 1.2 27.9 ± 2.6 WHR 0.95 ± 0.01 0.95 ± 0.01 0.95 ± 0.01 0.90 ± 0.02a FPG, mmol/L 7.7 ± 0.2 7.8 ± 0.3 7.9 ± 0.2 5.1 ± 0.1b FPI, mU/L 13.0 ± 1.8 14.8 ± 1.6 13.5 ± 1.2 7.9 ± 0.9b HOMA-IR, μU/mol/L3 4.4 ± 0.6 5.3 ± 0.7 4.7 ± 0.4 1.9 ± 0.2b HbA1c, % 6.6 ± 0.1 6.5 ± 0.1 6.4 ± 0.1 - Abbreviations: BMI, body mass index; WHR, waist/hip ratio; FPG, fasting plasma glucose; FPI, fasting plasma insulin; HOMA-IR, homeostasis model of assessment for insulin resistance. Data are shown as mean ± SEM. a P ≤ .05; and b P ≤ .01.

Vildagliptin Group Placebo Group Type 2 Diabetes Group Controls n 22 22 44 14 Age, y 65.2 ± 0.7 58.9 ± 1.6 61.4 ± 1.0 59.0 ± 2.2 Weight, kg 83.0 ± 3.2 91.8 ± 2.4 87.2 ± 2.0 86.7 ± 3.7 BMI, kg/m2 29.4 ± 0.8 31.1 ± 0.6 30.3 ± 0.5 29.6 ± 1.0 Body fat, % 31.5 ± 1.6 30.7 ± 1.8 31.1 ± 1.2 27.9 ± 2.6 WHR 0.95 ± 0.01 0.95 ± 0.01 0.95 ± 0.01 0.90 ± 0.02a FPG, mmol/L 7.7 ± 0.2 7.8 ± 0.3 7.9 ± 0.2 5.1 ± 0.1b FPI, mU/L 13.0 ± 1.8 14.8 ± 1.6 13.5 ± 1.2 7.9 ± 0.9b HOMA-IR, μU/mol/L3 4.4 ± 0.6 5.3 ± 0.7 4.7 ± 0.4 1.9 ± 0.2b HbA1c, % 6.6 ± 0.1 6.5 ± 0.1 6.4 ± 0.1 - Abbreviations: BMI, body mass index; WHR, waist/hip ratio; FPG, fasting plasma glucose; FPI, fasting plasma insulin; HOMA-IR, homeostasis model of assessment for insulin resistance. Data are shown as mean ± SEM. a P ≤ .05; and b P ≤ .01. Liver triglyceride quantitation A Philips 3.0 T Achieva scanner with a six-channel cardiac coil (Philips Healthcare) was used to measure liver triglyceride content. Data were acquired using a three-point Dixon method (9) with three gradient-echo scans acquired with adjacent out-of-phase and in-phase echoes during a 17-second breath hold (repetition time/echo time/averages/flip angle = 50 ms/3.45, 4.60, 5.75 ms/1/5°). A matrix size of 160 × 109 and with a field view of 400–480 mm according to volunteer size was used. The fat and water contributions of the magnetic resonance imaging signal were separated using an in-house program written in MATLAB, with the triglyceride content in the images expressed as a percentage of the total signal from fat and water in each pixel. The interscan Bland-Altman repeatability coefficient has previously been reported as 0.5% (10). The polygon tool in the imaging software Image J (11) was used to define regions of interest within the homogenous liver parenchyma on five separate slices of each scan, clearly avoiding contamination of data from blood vessels, the gall bladder, or any peripheral tissue so that regions of interest solely represented intrahepatic triglyceride. The five slice measurements were then averaged.

regions of interest within the homogenous liver parenchyma on five separate slices of each scan, clearly avoiding contamination of data from blood vessels, the gall bladder, or any peripheral tissue so that regions of interest solely represented intrahepatic triglyceride. The five slice measurements were then averaged. Euglycemic hyperinsulinemic clamp and endogenous glucose production On the patient's arrival at the study center, an 18-gauge cannula was inserted into a wrist or hand vein. This hand was placed in a hand-warming device and heated to 55°C to achieve arterialization of the venous blood samples. A second cannula was inserted into the antecubital vein of the contralateral arm for infusion of [6′6′-2H] glucose (98% enriched; Cambridge Isotope Laboratories). Basal hepatic glucose production (HGP) was calculated during the last 30 minutes of the basal period (−180 to 0 min), and clamp HGP was calculated during the final 30 minutes of the clamp period (12). Plasma glucose was clamped at 5.5 mmol with an insulin infusion rate of 40 mU m−2min−1, and whole-body insulin sensitivity was determined during the last 30 minutes of the 180-minute clamp (13). To prevent marked fluctuations in plasma [6′6′-2H] glucose atom percentage excess, glucose clamping was carried out using 10% dextrose enriched with the isotope as previously reported (10). Whole-body glucose and lipid oxidation was measured before and between 120 and 150 minutes of the clamp using a Quark RMR indirect calorimeter (Cosmed). The equations upon which the calculations are based are not made available by the manufacturer and, although in widespread use, may be subject to error. Protein oxidation was calculated using previously validated assumptions and was not measured (14). It was assumed not to change during the 6-month study.

ter (Cosmed). The equations upon which the calculations are based are not made available by the manufacturer and, although in widespread use, may be subject to error. Protein oxidation was calculated using previously validated assumptions and was not measured (14). It was assumed not to change during the 6-month study. Body composition and anthropometry A bioelectrical impedance device (Bodystat 1500; Bodystat Ltd) was used to determine percentage body fat. The repeatability of the Bodystat 1500 machine was assessed for a single volunteer on 10 different occasions. The intraindividual coefficient of variation was 6.0%. Waist and hip circumferences were measured with a nondistensible tape measure with subjects in a relaxed standing position. The waist was defined as the midpoint between the lower edge of the rib cage superiorly and the anterior superior iliac spine inferiorly. The hip circumference was taken to be at the level of the greater trochanter. The waist and hip measurements were expressed as a ratio. Metabolite and hormone assay Plasma glucose concentration was measured with a Yellow Springs glucose analyzer (YSI Inc). HbA1c was measured by HPLC (Bio-Rad). Plasma insulin was measured with Dako Insulin ELISA (DAKO) using a spectrophotometric analyzer. Glucagon concentration was measured with a Millipore Glucagon RIA Kit (Millipore Corporation).

Plasma glucose concentration was measured with a Yellow Springs glucose analyzer (YSI Inc). HbA1c was measured by HPLC (Bio-Rad). Plasma insulin was measured with Dako Insulin ELISA (DAKO) using a spectrophotometric analyzer. Glucagon concentration was measured with a Millipore Glucagon RIA Kit (Millipore Corporation). The 2H atom percentage excess in plasma glucose was determined using a Thermo “Voyager” single quadruple mass spectrometer with Thermo “Trace” gas chromatograph (Thermo Scientific). Plasma triglyceride was measured with the Triglyceride GPO-PAP spectrophotometric assay (Roche Diagnostics), using Roche/Hitachi Modular Analyzer. Nonesterified fatty acid (NEFA) was measured with an enzymatic colorimetric method assay using the Wako NEFA-HR (2) reagent (Wako Chemical). High-density lipoprotein (HDL) cholesterol was measured by Roche WAKO Direct Homogenous assay. Very low-density lipoprotein cholesterol was calculated from the total HDL cholesterol measurements. Data analysis Data are expressed as mean ± SEM. The adjusted mean changes from baseline to endpoint (with the last observation carried forward for data obtained at or after the mo 3 visit) were compared between treatments using analysis of covariance. Student's two-tailed paired t test or the Mann-Whitney test was used to compare within-group and between-group changes at different time points. Pearson correlation was used to assess the relationship between variables. The safety data were summarized descriptively by treatment.

Data analysis Data are expressed as mean ± SEM. The adjusted mean changes from baseline to endpoint (with the last observation carried forward for data obtained at or after the mo 3 visit) were compared between treatments using analysis of covariance. Student's two-tailed paired t test or the Mann-Whitney test was used to compare within-group and between-group changes at different time points. Pearson correlation was used to assess the relationship between variables. The safety data were summarized descriptively by treatment. Results Intrahepatic triglyceride Mean fasting liver triglyceride decreased during vildagliptin treatment from 7.3 ± 1.0% at baseline to 5.3 ± 0.9% at endpoint (P = .001). There was no change in the placebo group (5.4 ± 0.7% to 5.4 ± 1.0%; P = .48). The between-group difference in change from baseline was significant (P = .013), representing a clinically relevant improvement in liver triglyceride. The time course of change is shown in Figure 1A. Within the vildagliptin group, intrahepatic triglyceride fell from baseline by 12% at 1 month (P = .04), 29% at 3 months (P = .001), and 27% at 6 months (P = .0006). Figure 1. The effect of 6 months of vildagliptin on change in: A, hepatic triglyceride (TG) content (expressed as absolute change in percentage hepatic fat, not relative change); and B, fasting plasma glucose in the vildagliptin-treated and placebo groups, respectively. *, P < .05; **, P < .005; ***, P < .0005.

Results Intrahepatic triglyceride Mean fasting liver triglyceride decreased during vildagliptin treatment from 7.3 ± 1.0% at baseline to 5.3 ± 0.9% at endpoint (P = .001). There was no change in the placebo group (5.4 ± 0.7% to 5.4 ± 1.0%; P = .48). The between-group difference in change from baseline was significant (P = .013), representing a clinically relevant improvement in liver triglyceride. The time course of change is shown in Figure 1A. Within the vildagliptin group, intrahepatic triglyceride fell from baseline by 12% at 1 month (P = .04), 29% at 3 months (P = .001), and 27% at 6 months (P = .0006). Figure 1. The effect of 6 months of vildagliptin on change in: A, hepatic triglyceride (TG) content (expressed as absolute change in percentage hepatic fat, not relative change); and B, fasting plasma glucose in the vildagliptin-treated and placebo groups, respectively. *, P < .05; **, P < .005; ***, P < .0005. At baseline, intrahepatic triglyceride was higher in the whole type 2 diabetes group compared with the normal control group (6.7 ± 0.7 vs 3.4 ± 0.8%; P = .0008). Plasma alanine aminotransferase (ALT) Mean plasma ALT fell from 27.2 ± 2.8 to 20.3 ± 1.4 IU/L in the vildagliptin group (P = .0007) and did not change in the placebo group (29.6 ± 3.0 to 29.6 ± 3.7 IU/L; P = .44). There was a positive correlation between the fall in ALT and the fall in liver fat in the vildagliptin group (r = 0.83; P < .0001) (Figure 2). The correlation between these remained significant if the outlier was excluded (r = 0.64; P = .003).

id not change in the placebo group (29.6 ± 3.0 to 29.6 ± 3.7 IU/L; P = .44). There was a positive correlation between the fall in ALT and the fall in liver fat in the vildagliptin group (r = 0.83; P < .0001) (Figure 2). The correlation between these remained significant if the outlier was excluded (r = 0.64; P = .003). Figure 2. Relationship between change in liver triglyceride concentration and change in plasma ALT. The correlation remained similarly significant if the outlying point was excluded (ie, r = 0.64; P = .003).

id not change in the placebo group (29.6 ± 3.0 to 29.6 ± 3.7 IU/L; P = .44). There was a positive correlation between the fall in ALT and the fall in liver fat in the vildagliptin group (r = 0.83; P < .0001) (Figure 2). The correlation between these remained significant if the outlier was excluded (r = 0.64; P = .003). Figure 2. Relationship between change in liver triglyceride concentration and change in plasma ALT. The correlation remained similarly significant if the outlying point was excluded (ie, r = 0.64; P = .003). Whole-body insulin sensitivity assessed by euglycemic (5.5 mmol/L) hyperinsulinemic (40 mU insulin min−2min−1) clamp Baseline mean glucose disposal rates were 3.24 ± 0.30 mg−1kg−1min−1 in the vildagliptin group and 3.19 ± 0.38 mg−1kg−1min−1 in the placebo group. Glucose disposal rates did not change meaningfully in either group from baseline to endpoint (3.24 ± 0.30 to 3.50 ± 0.31 and 3.19 ± 0.38 to 3.52 ± 0.40 mg−1kg−1min−1, respectively). There was no between-group difference in the change from baseline (−0.06 mg−1kg−1min−1; P = .86). When glucose disposal was expressed as a ratio to the clamp plasma insulin at steady state (M/I, calculated as glucose disposal (M) divided by steady state plasma insulin concentration), insulin sensitivity did not change at 6 months in the vildagliptin (0.0065 ± 0.0006 to 0.0071 ± 0.0006 mL/kg/min/pm) and placebo (0.0074 ± 0.0013 to 0.0078 ± 0.0013 mL/kg/min/pm) groups. There was no between-group difference in the change from baseline (0.0001 mL/kg/min/pm; P = .89). Steady-state concentrations of plasma glucose, insulin, and 6′6′ diduterated glucose were achieved at baseline and 6 months for the vildagliptin and placebo groups (Supplemental Figures 1 and 2).

0013 mL/kg/min/pm) groups. There was no between-group difference in the change from baseline (0.0001 mL/kg/min/pm; P = .89). Steady-state concentrations of plasma glucose, insulin, and 6′6′ diduterated glucose were achieved at baseline and 6 months for the vildagliptin and placebo groups (Supplemental Figures 1 and 2). Hepatic glucose production Fasting HGP at baseline was the same in the vildagliptin and placebo groups (1.95 ± 0.10 vs 1.95 ± 0.07 mg/kg/min). Percentage suppression of hepatic glucose production during the hyperinsulinemic euglycemic clamp at baseline and endpoint was similar in the vildagliptin (42.2 ± 0.07 vs 42.8 ± 0.05%; P = .89) and placebo (54.7 ± 0.06 vs 48.4 ± 0.07%; P = .42) groups. Plasma HbA1c and plasma glucose Mean HbA1c changed by −0.5 ± 0.1% (P < .0001) from a baseline of 6.5 ± 0.1% in the vildagliptin group, whereas a small numerical increase (0.2 ± 0.1% from a baseline of 6.4 ± 0.1%; P = .14) was seen in the placebo group, resulting in a significant between-group difference of −0.7 ± 0.1% (P < .001). In the vildagliptin group, most of the HbA1c decrease was seen in the first 3 months (6.1 ± 0.1% at 3 mo and 6.0 ± 0.1% at 6 mo, respectively), whereas in the placebo group there was little change over the entire study period (6.4 ± 0.1% at 3 mo and 6.5 ± 0.2% at 6 mo, respectively).

fference of −0.7 ± 0.1% (P < .001). In the vildagliptin group, most of the HbA1c decrease was seen in the first 3 months (6.1 ± 0.1% at 3 mo and 6.0 ± 0.1% at 6 mo, respectively), whereas in the placebo group there was little change over the entire study period (6.4 ± 0.1% at 3 mo and 6.5 ± 0.2% at 6 mo, respectively). Mean fasting plasma glucose changed by −0.9 ± 0.3 mmol/L (P = .001) in the vildagliptin group (baseline, 7.9 mmol/L) and by 0.2 ± 0.3 mmol/L (P = .24) in the placebo group (baseline, 7.5 ± 0.2 mmol/L) over the study period. The between-group difference in change from baseline was significant (P = .018). The time course of change is shown in Figure 1B. In the vildagliptin-treated group, the decrease in fasting plasma glucose positively correlated with the decrease in fasting liver triglyceride content at 3 months (r = 0.47; P = .02) and 6 months (r = 0.44; P = .03). Plasma insulin Mean fasting plasma insulin decreased over the study period by 0.2 ± 0.2 mU/L (P = .22) in the vildagliptin group (baseline, 12.0 ± 1.6 mU/L) and by 1.4 ± 0.2 mU/L (P = .21) in the placebo group (baseline, 14.6 ± 2.0 mU/L). The between-group difference in change from baseline to endpoint was insignificant (P = .95). There was a decrease in fasting plasma insulin within the vildagliptin group by 4 months (12.0 ± 1.6 to 8.6 ± 1.3 mU/L; P = .04).

.0 ± 1.6 mU/L) and by 1.4 ± 0.2 mU/L (P = .21) in the placebo group (baseline, 14.6 ± 2.0 mU/L). The between-group difference in change from baseline to endpoint was insignificant (P = .95). There was a decrease in fasting plasma insulin within the vildagliptin group by 4 months (12.0 ± 1.6 to 8.6 ± 1.3 mU/L; P = .04). Whole-body glucose and lipid oxidation Fasting whole-body glucose oxidation rate did not differ between the vildagliptin and placebo groups either at baseline (0.99 ± 0.16 vs 0.87 ± 0.10 mg/kg/min) or at study endpoint (0.77 ± 0.19 vs 0.95 ± 0.14 mg/kg/min, respectively; P = .21 for the between-group difference in change from baseline). Similarly, the insulin-stimulated whole-body glucose oxidation rate did not differ between the vildagliptin and placebo groups at baseline (1.52 ± 0.21 vs 1.50 ± 0.08 mg/kg/min) or at the end of the study (1.55 ± 0.16 vs 1.65 ± 0.14 mg/kg/min, respectively; P = .65 for the between-group change from baseline). The fasting whole-body lipid oxidation rate did not differ between the vildagliptin and placebo groups either at baseline (1.08 ± 0.07 vs 1.13 ± 0.05 mg/kg/min) or at study endpoint (1.13 ± 0.06 vs 1.05 ± 0.08 mg/kg/min, respectively, for the between-group difference in change from baseline). There was no significant difference in clamp lipid oxidation between both groups at baseline (0.89 ± 0.09 vs 0.86 ± 0.08 mg/kg/min) and the end of the study (0.91 ± 0.07 vs 0.94 ± 0.07 mg/kg/min).