{"paper_id":"c1f60098-9c35-4736-a825-5b9e8bef9e37","body_text":"We studied patients with severe obesity (BMI, weight in kilograms divided by height in meters squared) SD score > 3) of early-onset (age < 10 years) recruited to the Genetics of Obesity Study cohort ( www.goos.org.uk ). These studies were approved by the Cambridge Local Research Ethics Committee (03/103, 03/104, 18/EE/0032) and conducted in accordance with the principles of the Declaration of Helsinki. Each participant or their legal guardian (for children age < 16 years) provided written, informed consent, and minors provided verbal or written consent. Exome sequencing and targeted resequencing were performed as previously described ( 5 ); all variants were verified by Sanger sequencing.\nWe performed a series of studies to describe the clinical features of  SRC-1  variant carriers. We compared data from adult variant carriers with those obtained on obese volunteers recruited by advertisement who were carefully matched for age, sex, and BMI and in whom variants in  SRC-1  were excluded by Sanger sequencing. Those with comorbidities (type 2 diabetes, autoimmune hypothyroidism and hyperthyroidism and/or taking levothyroxine treatment) and concomitant medications that affect body weight and/or metabolism were excluded from clinical studies. Some measurements were not performed on all eligible cases and controls for a variety of reasons (weight limit for dual x-ray absorptiometry [DEXA] and/or magnetic resonance imaging [MRI], challenging vascular access, medications); numbers studied are indicated for each measurement. Eighteen children with variants in  SRC-1  were compared to 11 age-matched controls. Whereas adult controls were matched for BMI, pediatric controls were not, as equally severely obese children often have other genetic conditions.\nAdipose tissue samples (Paris obesity cohort) were obtained from people with severe obesity prospectively recruited to the Prospective Bariatric Surgery Cohort of the Nutrition Department at Pitié-Salpêtrière Hospital, Paris, France, between 2014 and 2018. They are part of several studies registered at ClinicalTrials.gov (P050318,  NCT01655017 , and  NCT01454232 ). Informed consent was obtained to perform paired omental and subcutaneous surgical adipose tissue biopsies obtained during bariatric surgery. In the surgical department in Paris, patients are asked to have a stabilized weight before bariatric surgery and to not routinely consume a restrictive diet preoperatively in contrast to clinical practice at other centers. The details of adipose tissue handling have been reported previously ( 15 ).\nWeight and height were measured barefoot in light clothing. DEXA (DPX software; Lunar Corp) was used to determine body composition, bone mineral content and density (BMD) (whole body). Subcutaneous and visceral fat mass were quantified in a subset of cases and controls by MRI at the level of L1 and analyzed using OsiriX Lite software (Pixmeo). In a subset of 13  SRC-1  variant carriers and 15 controls, MRI scans were carried out on a wide-bore 1.5-T whole-body system (GE Medical Systems) and included localizer images, T2-weighted fat-saturated images, proton density fat fraction (PDFF), and magnetic resonance elastography (MRE). Participants fasted for 4 hours before the procedure and were positioned in the supine position with an acoustic driver positioned superficial to the liver. Liver volume: T2-weighted images were used to draw regions of interest around the liver perimeter on consecutive slices and volume computed using OsiriX Lite (Pixmeo). Images for PDFF were acquired using a gradient echo sequence with a low flip angle that were then analyzed using the validated IDEAL-IQ software (GE Healthcare) to generate an unbiased quantitative measure of percentage fat fraction. MRE was used to quantify hepatic fibrosis (> 2.9 kPa correlates with statistically significant hepatic fibrosis on histology and a METAVIR score of ≥ F2) ( 16 ). MRE images were acquired using a phase-contrast gradient echo sequence. Where participants were unable to have MRI (weight ≥ 230 kg, abdominal girth ≥ 70 cm), acoustic radiation force impulse (ARFI) elastography or mechanical impulse elastography (FibroScan, Echosens) was performed. ARFI was performed using Canon shear wave elastography by a trained radiologist on an Aplio 500 (version 2 software) or Aplio i800. Mechanical impulse elastography was performed with FibroScan using the XL probe. Results were accepted if they met recognized reliability criteria (interquartile range/median < 0.3) ( 17 ). Operators were blinded to genotype. Two participants were unable to have an MRI scan (weight ≥ 230 kg, abdominal girth ≥ 70 cm); as such ARFI or transient elastography was performed.\nAd libitum energy intake was assessed using an 18-MJ meal of known macronutrient content (50% carbohydrate, 30% fat, 20% protein) administered after an overnight fast; energy intake was expressed per kilogram of lean body mass as measured by DEXA to allow comparison between individuals of different body weights and compositions. Basal metabolic rate and respiratory quotient were determined by indirect calorimetry after a 10-hour overnight fast using an open circuit, ventilated, canopy measurement system (Europa Gas Exchange Monitor; NutrEn Technology). After adjustment for body composition, basal metabolic rate was compared to predicted metabolic rate based on standard age- and sex-specific equations. Blood pressure was measured in the rested, fasted state using automated wrist (OMRON Healthcare) monitors. We used a DINAMAP automated brachial monitor with the Dura large adult cuff (GE Healthcare) in 3 people whose wrist circumference precluded use of this device. Heart rate was measured using a digital portable heart rate monitor (Actiheart, Cambridge Technologies). Heart rate data collected using the portable heart rate monitor were exported to MS Excel 2010 via Actiheart software (version 4.0.116, CamNtech Ltd). Overnight readings were taken between 0000 h and 0500 h when each participant was asleep; awake readings were taken during a 30-minute window directly after the participant had woken up. Heart rate variability was analyzed with Kubios HRV Premium (version 3.4.1, Kubios) by 2 researchers who were blinded to genotype.\nBlood samples were obtained in the fasting state and analyzed for lipids, thyrotropin (TSH), free thyroxine (FT4), free 3,5,3′-triiodothyronine, luteinizing hormone, follicle-stimulating hormone, estradiol, testosterone, and cortisol with the use of standard assays. In addition, an oral glucose tolerance test examining the glucose and insulin response to 75-g oral glucose load over 180 minutes was performed in 15 adult variant carriers and 14 controls who did not have a diagnosis of type 2 diabetes.\nIn 13 case participants and 14 age-, sex-, and BMI-matched controls who consented to the procedure, subcutaneous adipose tissue (SAT) biopsies were obtained from the left iliac fossa, using a nondiathermy surgical biopsy method, under local anesthesia (1% lidocaine with 1:200 000 adrenaline). A 2- to 3-cm skin incision was made, skin flaps were raised on either side of the incision, and the superficial fascia incised. Approximately 2 g of adipose tissue was excised and washed in phosphate-buffered saline. Separate sections (4 μm) of SAT were stained with hematoxylin and eosin (H&E) and Picosirius red and scanned using the Axio Scan.Z1 slide scanner (Zeiss) into whole digital images. To evaluate adipocyte morphology and measure adipocyte surface area, H&E-stained slides were processed using HALO image analysis software (Indica Labs). All adipocytes in the section were included in the analysis; adipocyte surface area was measured using the inbuilt Vacuole v2.2 algorithm. Scoring of anonymized samples from cases with  SRC-1  variants and controls (UK and Paris groups) was performed by the same observer (P.B.L.). Values were used to generate a frequency distribution with a bin width of 500 μm. To quantify adipose tissue fibrosis, Picosirius red–stained slides were assessed using the fibrosis score of adipose tissue (FAT score) ( 18 ) by an experienced observer (P.B.L.) blinded to the identity of the samples. Severe adipose tissue fibrosis was defined as a FAT score greater than or equal to 2 as previously described ( 18 ).\nWe measured DNL in 14 case participants and 16 age-, sex-, and BMI-matched controls. Fasting blood samples were collected at 2000 h, and a priming dose of deuterium-labeled water (3 ml × 0.5 × body weight [kg] for females; 3 mL × 0.6 × body weight [kg] males) was made up to 200 mL with tap water and provided in 2 equal portions at 2000 h and 2200 h. Ad libitum deuterated water (4.5 ml/L) was provided overnight; no food was permitted. A blood sample was collected at 0800 h the following morning. Participants were then provided with breakfast comprising 50% of energy requirements (60% carbohydrate, 30% fat, 10% protein), a further portion of deuterated water at 1000 h and a final blood sample collected at 1200 h.\nTotal plasma lipids were extracted using chloroform-methanol and triglycerides separated by solid-phase extraction ( 19 ). Fatty acid methyl esters were prepared using methanolic sulfuric acid and fatty acid relative abundance (mol %) was determined by gas chromatography ( 19 ). Fasting and postprandial hepatic DNL was assessed based on the incorporation of deuterium from  2 H 2 O in plasma water (Finnigan GasBench-II, Thermo Fisher Scientific) into plasma TAG palmitate using gas chromatography–mass spectrometry, monitoring ions with mass-to-charge ratios of 270 (M + 0) and 271 (M + 1) ( 20 ). Background isotopic enrichment in plasma water was measured in a fasting blood sample taken before participants consumed deuterated water. Plasma metabolites (3-hydroxybutyrate, triglyceride, nonesterified fatty acids [NEFAs]) were analyzed enzymatically (ILab 650 clinical chemistry; Werfen).\nStatistical analysis was performed using R and GraphPad prism v9 statistical packages. All data sets were checked for normality of distribution using the Shapiro-Wilk, Kolmogorov-Smirnov tests, and equality of variance using the  F  test. Data sets with normal distributions and equal variance were compared using an unpaired  t  test, and nonparametric data were compared using the Mann-Whitney  U  test. Where multiple comparisons were performed, the Holm-Sidak correction for multiple comparisons was used to compute the adjusted  P  value. Statistical tests are 2-tailed unless otherwise stated and statistical significance of an individual test was declared at  P  less than .05.\n\nIn this study, we performed exome sequencing and targeted resequencing of 2462 people with severe obesity and identified 23 rare heterozygous variants in  SRC-1  ( Fig. 1B ); a number of variants were found in multiple unrelated individuals and some were present in publicly available exomes ( Table 1 ). In a previous study, we identified 15 variants in 16 unrelated individuals among 2548 people with severe obesity studied using the same methods ( 6 ) and showed they caused an LOF in cells using a POMC reporter assay (see  Table 1 ). While we did not perform functional studies of the 23 variants identified in this new study, the variants identified here affect the domain known to interact with STAT3 and the domain that interacts with NHRs (nuclear receptor interacting domain, NRID) (see  Fig. 1B ). We have previously shown that variants affecting these domains can cause an LOF ( 6 ).\nSteroid receptor coactivator-1 ( SRC-1 ) variants identified in a cohort with severe obesity (Genetics of Obesity Study)\nVariant numbering based on Ensembl transcript ID ENST00000406961. Allele frequencies obtained from gnoMAD v2.1 ( https://gnomad.broadinstitute.org/ ).\nWe performed family cosegregation studies and found that rare variants in  SRC-1  did not consistently cosegregate with severe obesity in a mendelian manner ( Fig. 2 ). This was true even for variants shown to cause a clear LOF in cells previously ( 6 ). We conclude that rare  SRC-1  variants are not fully penetrant but rather may interact with other genetic and/or environmental factors to modulate the phenotype. These findings align with those reported for other obesity-associated genes ( MRAP2 ,  KSR2 , and  PHIP ) ( 10 ,  21 ,  22 ), where rare heterozygous variants are associated with obesity but are not always causative, in contrast to the classical monogenic obesity syndromes that follow an autosomal recessive ( LEP ,  LEPR ,  POMC , and  PCSK1 ) or dominant ( MC4R ,  SIM1 , and  GNAS ) mode of inheritance ( 1 ).\nPedigrees of  SRC-1  cases. Cosegregation of  SRC-1  variants is shown where male (squares) and female (circles) family members consented to genotyping. Heterozygous (filled) and wild-type (empty)  SRC-1  variant carriers indicated; some individuals were not available for genotyping (gray). Body mass index (BMI) (> 25 kg/m2 = overweight; > 30 kg/m2 = obesity) for adults and BMI SDS for children are shown where available. Arrow indicates the proband. Rhombus = unknown. * = Homozygous. SRC-1, steroid receptor coactivator-1.\nWe invited all probands with rare variants in  SRC-1  and their affected family members to take part in clinical studies. Twenty-nine adults and 18 children (age < 18 years) with rare variants in  SRC-1  consented to take part ( Fig. 3A ). Twenty-nine adult  SRC-1  variant carriers were studied using a core phenotyping protocol ( Table 2 ); mean (± SD) age 31.4 ± 2.1 years (range, 18.2-49.8 years); their data were compared to 30 controls (mean (± SD) age 30.9 ± 1.7 years (range, 19.0-48.4 years,  P  = .94). The mean BMI of the adult  SRC-1  variant carriers was 41.8 ± 13.3 (23.3-74.9) and controls 43.2 ± 12.7 (26.8-85.2,  P  = .62). A subset of these individuals, 15 SRC-1 case participants and 16 controls, consented to take part in a metabolic substudy that involved liver imaging, adipose tissue biopsies, and stable-isotope studies of DNL ( Fig. 3A  and  Table 2 ).\nSteroid receptor coactivator-1 ( SRC-1 ) variant carriers included in clinical studies\nForty-seven  SRC-1  variant carriers (22 probands, 25 family members) consented to participate in physiological studies. All 29 adults participated in a core phenotyping protocol; a subset of 15 adults participated in metabolic studies. Eighteen children were studied on a limited pediatric protocol.\nAbbreviations: –, information not available; BMI, body mass index; F, female, M, male.\na \n Proband.\nClinical features of  SRC-1  variant carriers. A, Forty-seven people with  SRC-1  variants consented to phenotypic studies. Twenty-nine adult  SRC-1  cases and 30 age-, sex-, and body mass index–matched controls underwent core phenotyping studies; a subset of 15  SRC-1  cases and 16 controls consented to additional metabolic phenotyping studies. Eighteen children with variants in  SRC-1  were compared to 11 age-matched controls. B, Additional clinical features seen in  SRC-1  variant carriers; details of age, sex and  SRC-1  variant are shown. C and D, Ad libitum energy intake assessed using a 18-MJ meal of known macronutrient content after an overnight fast in C, children, and D, adults; food intake expressed per kilogram lean body mass measured by dual-energy x-ray absorptiometry. E, Basal metabolic rate (BMR; megajoules, MJ/d) determined by indirect calorimetry after an overnight fast. Measured BMR was compared to BMR predicted by age- and sex-specific equations. SRC-1, steroid receptor coactivator-1.\nA high proportion (40%; n = 19) of  SRC-1  variant carriers had a history of fractures in response to minor injuries in contrast to patients with other genetic obesity syndromes ( 1 ) ( Fig. 3B ). Whole-body BMD was comparable in cases and controls (BMD  Z  score 0.6 ± 0.2 vs 0.7 ± 0.2, respectively). While none of the  SRC-1  variant carriers had thyroid function tests suggestive of RTH (high TSH/high FT4,  Table 3 ), interestingly, T4 treatment for autoimmune hypothyroidism of the women carrying the L1376P and A715T variants was complicated by failure to adequately suppress TSH, despite high-dose (> 200 μg) levothyroxine, unless FT4 levels were above 25 pmol/L (normal range, 10-19.8 pmol/L). We carefully reviewed compliance with T4 therapy. Notably, there were no features of malabsorption or bowel disease at colonoscopy in the patient carrying L1376P  SRC-1 .\nMetabolic and endocrine measurements in adult steroid receptor coactivator-1 ( SRC-1 ) variant carriers compared to age-, sex-, and body mass index–matched controls\nValues are presented as mean ± SEM.\nAbbreviations: ~ a subset of people underwent detailed metabolic phenotyping (see  Table 2 ); ALP, alkaline phosphatase; ALT, alanine aminotransferase; AUC, area under the curve; BP, blood pressure; DBP, diastolic blood pressure; DNL, de novo lipogenesis; HDL, high-density lipoprotein; HF, high frequency; HR, heart rate; LDL, low-density lipoprotein; LF, low frequency; NEFA, nonesterified fatty acid; OGTT, oral glucose tolerance test; RMSSD, root mean square of successive differences; SBP, systolic blood pressure; TSH, thyrotropin.\na \n \n P  less than .05; otherwise there was no statistically significant difference between cases and controls.\nMenstrual irregularities are common in women with obesity. We found that 13% (2/15) of women (age 16-50 years) with  SRC-1  variants had a history of polycystic ovary syndrome (with hyperandrogenism and hirsutism); this was comparable to severely obese women with MC4R deficiency (9%; 2/22) of comparable BMI. In addition, 8 out of 18 female  SRC-1  variant carriers older than 16 years (44%) reported menorrhagia with persistent bleeding (see  Fig. 3B ). One female  SRC-1  variant carrier was diagnosed with complex atypical endometrial hyperplasia in her early 20s that responded well to progestogen therapy. One female variant carrier of the L1376P variant had primary amenorrhea. Women with  SRC-1  variants were more likely to be prescribed tranexamic acid for persistent bleeding (13.5%; 4/15) than women with MC4R deficiency (0 of 22) we have studied.\nFour out of 12 adult male  SRC-1  variant carriers had low testosterone and gonadotropin levels consistent with hypogonadotropic hypogonadism (see  Fig. 3B ). Three out of these 4 variant carriers had signs of primary hypogonadism (undescended testis, reduced penile length). In comparison, 3 out of 8 male controls had low testosterone and gonadotropin levels, consistent with secondary hypogonadotropic hypogonadism. None of the variant carriers had overt clinical signs or symptoms of mineralocorticoid or corticosteroid hormone resistance.\nA number of  SRC-1  variant carriers had been diagnosed with hepatic fibrosis or cirrhosis at a young age (see  Fig. 3B ). Liver function tests were within the normal range in case and control participants (see  Table 3 ); liver volume determined by T2-weighted MRI (2126 ± 545 cm 3  in case participants v 2388 ± 661 cm 3  in controls,  P  = .29) and liver fat assessed using MRI-PDFF was comparable in case and control individuals (7.1% vs 10.9%, respectively,  P  = .38). We found that 5 out of 13 (38%) adult  SRC-1  variant carriers had significant hepatic fibrosis compared to 2 out of 13 (15%) adult controls; odds ratio equal to 3.4 ( P  = .38). Fourteen  SRC-1  variant carriers (30%) reported frequent and/or persistent diarrhea, described as pale, loose stools, typically after fatty food. There was no biochemical evidence of macronutrient or micronutrient malabsorption (see  Table 3 ). These features could be consistent with bile acid malabsorption; further studies will be required to test this.\nTwelve  SRC-1  variant carriers (25%) had moderate to severe asthma requiring home nebulizers, frequent courses of oral corticosteroids, and/or hospital admissions (see  Fig. 3B ). STAT3 is a major regulator of the differentiation and function of Th17 cells, a subset of CD4 +  T cells implicated in inflammation in patients with severe asthma. Three variant carriers had pulmonary valve stenosis, a rare congenital abnormality (population prevalence 3/10 000 births). Further studies will be needed to explore the mechanisms underlying these observations.\nChildren demonstrated increased food intake at an 18-MJ ad libitum test meal compared to age-matched controls ( Fig. 3C ); food intake was not increased in adults with  SRC-1  variants compared to controls ( Fig. 3D ). In contrast to findings in SRC-1–deficient mice, there was no deficit in measured basal metabolic rate, which correlated tightly with that predicted by age, sex and body composition ( R  = 0.6989,  P  = .0013,  Fig. 3E ). Heart rate and heart rate variability (markers of autonomic nervous system tone) were comparable between the groups (see  Table 3 ). We quantified the ratio of visceral adipose tissue (VAT) to SAT using MRI. VAT:SAT was comparable in  SRC-1  variant carriers and controls (1.49 ± 0.91 in cases vs 1.27 ± 0.48 in controls,  P  = .9).\nTwo out of 29  SRC-1  variant carriers had a diagnosis of type 2 diabetes. Although SRC-1 has been implicated in hepatic glucose production ( 23 ), we did not find any differences in fasting glucose or insulin levels or in the area under the curve during a 75-g oral glucose tolerance test between  SRC-1  variant carriers and controls ( Fig. 4A  and  4B ; see  Table 3 ). Fasting lipid profiles were within the normal range in case and control participants (see  Table 3 ). We measured fasting and postprandial DNL using deuterated water. There was no difference between cases and controls in fasting hepatic DNL, expressed as a percentage or a concentration (see  Table 3 ). With meal consumption, DNL increased in both groups to a similar extent (see  Table 3 ). We found no difference in fasting plasma concentrations of 3-hydroxybutyrate, a marker of hepatic fatty acid oxidation, between case and control participants (see  Table 3 ). As expected, there was a statistically significant decrease in plasma 3-hydroxybutyrate concentrations postprandially in both groups (see  Table 3 ). Fasting NEFA concentrations were not statistically significantly different between case and control individuals (see Table 3). While postprandial NEFA concentrations decreased in both groups, this decline was attenuated in case vs control participants (191.7 ± 15.6 µmol/l vs 137.1 ± 18.7 µmol/L, respectively,  P  = .02), which may suggest impaired insulin sensitivity of the adipose tissue.\nGlucose homeostasis and adipose tissue morphology in  SRC-1  variant carriers. A and B, Insulin and glucose levels during a 75-g oral glucose tolerance test over 180 minutes in SRC-1 cases and controls. C and D, Representative histology of subcutaneous adipose tissue biopsies taken from  C, SRC-1  variant, carriers, and D, controls, stained for collagen, consistent with fibrosis score of adipose tissue (FAT) 3 and 0, respectively. Picosirius red staining, scale bar = 500 μm. E,  SRC-1  variant carriers (cases) have a statistically significantly higher proportion of severe adipose tissue fibrosis (FAT 2-3), compared to age-, sex-, and body mass index–matched controls. F, Distribution of FAT score by age in  SRC-1  variant carriers (black circles), matched controls (open circles), and a large cohort of individuals with severe obesity undergoing bariatric surgery (gray circles, “Paris obesity cohort”). SRC-1, steroid receptor coactivator-1.\nOmental and SAT can both undergo fibrosis in people with obesity ( 24 ,  25 ). Increased adipose tissue fibrosis is associated with inflammation ( 26 ) and may lead to decreased tissue plasticity and thus a failure to expand fat mass with sustained weight gain ( 27 ,  28 ). In addition, adipose tissue fibrosis is associated with reduced weight loss following bariatric surgery in people with severe obesity ( 18 ,  29 ).  SRC-1  variant carriers had a higher proportion of smaller adipocytes (500 µm 2 ) compared to controls (21.7 ± 4.3% in case individuals vs 19.6 ± 6.0% in control individuals, adjusted  P  = .02). Further analysis of the tissue demonstrated marked fibrosis in biopsies taken from several case participants ( Fig. 4C  and  4D ). The degree of fibrosis was assessed using a previously reported FAT score ( 18 ). A greater proportion of  SRC-1  variant carriers had severe fibrosis (FAT score 2-3) compared to age-, sex-, and BMI-matched controls (46.2% vs 7.1%,  P  = .03) ( Fig. 4E ). There was no correlation between FAT score and BMI in either group. In addition, we compared our findings to those from a cohort of 265 similarly obese patients (mean BMI ± SD 46.3 ± 14.1 vs 46.6 ± 6.4,  P  = .93) attending for bariatric surgery in Paris, France (data on 183 of these patients were previously published) ( 18 ). The prevalence of severe fibrosis was comparable in both groups (46.2% v 39.2%,  P  = .7), although the Paris control group was statistically significantly older than the  SRC-1  variant carriers in our study (31.4 ± 12.1 vs 42.9 ± 12.1 years, respectively,  P  < .01) ( Fig. 4F ).\n\nIn this study, we describe the clinical features seen in people with severe obesity carrying rare variants in  SRC-1 . In keeping with findings in patients with other genetic disorders affecting the leptin-melanocortin pathway,  SRC-1  variant carriers experience hyperphagia in childhood, but have a normal basal metabolic rate and mild insulin resistance. These findings are consistent with previous findings showing that SRC-1 interacts with phosphorylated STAT3 to modulate leptin-mediated POMC transcription and with findings in a mouse model of a human LOF  SRC-1  variant ( 6 ).\nWe observed a very high prevalence of multiple fractures with minimal trauma in males and females starting from childhood, in contrast to findings in patients with other genetic obesity syndromes such as MC4R deficiency ( 30 ). SRC-1 plays an important role in the development and maintenance of bone density, primarily through its interaction with the estrogen receptor ( 31 ,  32 ). The skeletal effect of  SRC-1  variants may also be mediated by resistance to the vitamin D receptor and the TH receptor, which function as heterodimers with retinoid X receptor. Although we did not observe a reduction in whole-body BMD in SRC-1 case vs control participants, it is important to note that deletion of  SRC-1  results in trabecular osteopenia in male and female mice ( 13 ). Trabecular bone accounts for only 20% of total bone mass in humans. Targeted spine, femur, femoral neck, and distal radius DEXA scans and/or high-resolution peripheral quantitative computed tomography or bone histology may be needed to investigate the effect of SRC-1 deficiency on bone architecture.\nSRC-1 enhances TH receptor–mediated signaling ( 33 ).  SRC-1  knockout mice display RTH with a 2.5-fold elevation of TSH levels, despite a 50% increase in free TH levels ( 11 ). Whilst none of the  SRC-1  variant carriers in our study had overt classical RTH, treatment of concomitant autoimmune hypothyroidism was complicated by failure to suppress TSH in 2 variant carriers, despite high-dose T4. There was no suggestion of noncompliance or intermittent compliance, and there were no features of malabsorptive disorders such as celiac disease or bowel disease. Interestingly, the decline of TSH observed in mice treated with L-triiodothyronine is blunted in  SRC-1  knockout mice, indicating that SRC-1 enhances the sensitivity by which TH downregulates TSH ( 11 ,  34 ). Studies in larger cohorts of patients will be needed to establish whether partial RTH, as seen in  SRC-1  knockout mice, is a feature in some people with  SRC-1  variants.\nMultiple female  SRC-1  variant carriers reported severe menorrhagia that significantly affected their quality of life. SRC-1 is expressed in the endometrium; its expression increases during menstruation both in the glandular epithelium and stroma, where it may play a role in endometrial remodeling ( 35 ,  36 ) that depends on the balance of estrogen and progesterone signaling ( 12 ,  37 ). Interestingly, a truncated SRC-1 isoform has been implicated in the pathophysiology of endometriosis ( 38 ). The complex atypical endometrial hyperplasia of one of the  SRC-1  variant carriers did respond to progestogens, suggesting that any subtle defects in progesterone signaling due to  SRC-1  variants may be overcome by pharmacotherapy with progesterone.\nGenerally, increased adipose tissue mass is characterized by adipocyte hypertrophy followed by hyperplasia. In chronic obesity (as seen in the older patients studied in the Paris cohort), adipose tissue undergoes major remodeling, resulting in inflammation and fibrosis, which represents a physical constraint to adipose tissue expansion ( 39 ). Studies in rodents ( 40 ) (and insights from patients with lipodystrophy) ( 41 ) indicate that this reduction in adipose expandability promotes the deposition of excess lipid in liver and skeletal muscle leading to metabolic complications ( 42 ). Here, glucose tolerance was comparable in case and control individuals tightly matched for age and BMI. Whether increased adipose tissue fibrosis in young patients with  SRC-1  variants predisposes them to the early development of metabolic complications such as type 2 diabetes and fatty liver disease is an important question that needs to be addressed in larger longitudinal studies. The observation that several patients have hepatic fibrosis (but not steatosis) is not readily explained and will require further investigation and long-term monitoring.\nLarger studies will be needed to test for genotype-phenotype correlations. While a common variant in the gene encoding SRC-1 ( NCOA1 ) has been associated with reduced hip and lumbar BMD in women receiving tamoxifen treatment ( 32 ), to date variants at this locus have not been associated with BMI in genome-wide association studies.\nIn conclusion, the clinical spectrum associated with  SRC-1  variants is characterized by hyperphagia in childhood and severe obesity and encompasses a range of other metabolic and endocrine features. The interpretation of these findings is challenging because several factors including the functional consequences of different  SRC-1  variants, the presence of residual SRC-1 activity in carriers of heterozygous variants, and compensation by the closely related molecule SRC-2 ( 7 ) will contribute to phenotypic heterogeneity. Further functional characterization of the  SRC-1  variants reported here will be needed to establish their pathogenicity. All are very rare, affect highly conserved residues, and are located in domains in which we have previously identified LOF variants. Nonetheless, as seen with the first clinical descriptions of functional variants affecting TH receptor-α and -β and estrogen receptor-α, these findings illustrate important aspects of physiology and pathophysiology. Human variants affecting SRC-1 will be useful molecular tools with which to dissect NHR signaling and its regulation by coactivators and corepressors, studies that will provide insights into underlying disease mechanisms and ultimately inform the clinical investigation and management of patients.","source_license":"CC-BY-4.0","license_restricted":false}