Characterization of Rare Variants in MC4R in African American and Latino Children With Severe Early-Onset Obesity

Abstract

Context

Mutations in melanocortin receptor (MC4R) are the most common cause of monogenic obesity in children of European ancestry, but little is known about their prevalence in children from the minority populations in the United States.

Objective

This study aims to identify the prevalence of MC4R mutations in children with severe early-onset obesity of African American or Latino ancestry.

Design and Setting

Participants were recruited from the weight management clinics at two hospitals and from the institutional biobank at a third hospital. Sequencing of the MC4R gene was performed by whole exome or Sanger sequencing. Functional testing was performed to establish the surface expression of the receptor and cAMP response to its cognate ligand α-melanocyte–stimulating hormone.

Participants

Three hundred twelve children (1 to 18 years old, 50% girls) with body mass index (BMI) >120% of 95th percentile of Centers for Disease Control and Prevention 2000 growth charts at an age <6 years, with no known pathological cause of obesity, were enrolled.

Results

Eight rare MC4R mutations (2.6%) were identified in this study [R7S, F202L (n = 2), M215I, G252D, V253I, I269N, and F284I], three of which were not previously reported (G252D, F284I, and R7S). The pathogenicity of selected variants was confirmed by prior literature reports or functional testing. There was no significant difference in the BMI or height trajectories of children with or without MC4R mutations in this cohort.

Conclusions

Although the prevalence of MC4R mutations in this cohort was similar to that reported for obese children of European ancestry, some of the variants were novel.

This study identifies pathogenic variants in the MC4R gene in children with severe early-onset obesity from African American race and Latino ancestry.

The leptin melanocortin axis plays a crucial role in weight regulation, with mutations in gene members being identified as causal for human monogenic obesity (1). The melanocortin 4 receptor (MC4R), which is most abundantly expressed in the paraventricular nucleus of the hypothalamus, is the sensor for orexigenic, as well as anorexigenic, signals in this pathway (2). Binding of α-melanocyte–stimulating hormone (α-MSH) to the G protein–coupled receptor MC4R induces satiety. Mutations in the MC4R gene and genes encoding its trafficking pathway have been identified as the most common cause of monogenic obesity in children of European ancestry, with prevalence estimates ranging from 1.5% to 5.8% (3–5). Individuals with MC4R mutations have been reported to have hyperphagia and early onset of obesity, even in the heterozygous state (3–5).

Although the obesity epidemic has disproportionately affected non-Hispanic African American and Hispanic youth in the United States (6), little is known about the prevalence of MC4R mutations in children of these ethnicities. In a recently published study of 25 unrelated Afro-Caribbean children with obesity from the Guadeloupe Island, one 11.8-year-old boy had a p.Ile301Thr mutation in MC4R (7), and in a study of South African adults of mixed ancestry that included 63% Black Africans (n = 187), 2.6% harbored a rare MC4R variant (8). These studies provide early clues on the expected prevalence of MC4R variants in children of non-European ancestry. In this study, we catalog the frequency of MC4R mutations in children with severe early-onset obesity of African American or Latino ancestry in the United States.

Materials and Methods

Clinical study recruitment

The study cohort was recruited from Boston Children’s Hospital (BCH), Children’s Hospital of Philadelphia (CHOP), and Columbia University Medical Center (CUMC). The research ethics board of each participating institution approved the study, and informed consent or assent was obtained before enrollment. The inclusion criteria were the presence of severe early-onset obesity, defined as body mass index (BMI) >120% of the 95th percentile (BMI_95pct) on the Centers for Disease Control and Prevention (CDC) 2000 BMI charts for age and sex (9) documented at <6 years of age with no previously known underlying genetic or other pathogenic cause; for the sequencing study reported herein, we focused on children with self-reported African American or Latino ancestry with age at recruitment between 2 and 18 years. The subjects at BCH were prospectively identified with a validated electronic algorithm designed to identify children with severe obesity (10), and families were approached for recruitment in the weight management or primary care clinic under the Genetics of Early Childhood Obesity study (n = 225; NCT01998750). In this cohort, 48.4% (n = 109) children were of self-reported African American or Latino ancestry. Latino to the electronic data warehouse at CHOP to identify children with severe obesity retrospectively. After biologically implausible data points were filtered and individuals with BMI z score <−5, obesity duration <6 months, or lack of data in early childhood were removed, 1648 participants with severe early-onset obesity remained in the study. Then we analyzed longitudinal BMI trajectory with Super Imposition by Translation and Rotation (SITAR, R version 3.1.3) to generate the size, tempo, and velocity estimates for each subject (n = 1120) of self-reported African American ancestry (11). Children who had the highest velocity of the BMI trajectory and a DNA sample in the institutional biobank (Center for Applied Genomics; n = 157) were prioritized for genetic sequencing. At CUMC, children were recruited from the 110 families attending the Families Improving Health Together Program, a New York state–supported clinical and research program designed for children with early-onset obesity (n = 47).

Phenotype data for longitudinal BMI, blood pressure, and laboratory studies were extracted from the electronic health records. Class of obesity was assigned according to the recent classification of the CDC 2000 BMI growth charts, with class 2 defined as 120% to 140% of BMI_95pct and class 3 as >140% of BMI_95pct (9). No additional clinical testing was performed for the study.

Molecular genetic analyses

DNA was extracted from the peripheral blood lymphocytes by standard methods or by saliva collection with DNA Genotek® kits as per manufacturer’s instructions. The coding exon (332 amino acids) and adjacent intronic regions of MC4R were amplified by PCR with previously reported primer pairs (12). The PCR products were purified and directly sequenced with the Big Dye Sequencing kit (Applied Biosystems, Foster City, CA) on an ABI 3100 automated DNA sequencer (Applied Biosystems). The chromatograms from the sequencing studies were analyzed with Chromaseq software (version 1.7).

A subset of the samples was sequenced by whole exome sequencing (n = 157). Library preparation was performed with SureSelect XT Human All Exon V5 kit (Agilent Technologies, Santa Clara, CA), and sequencing was performed on a HiSeq platform (Illumina, Inc, San Diego, CA) as paired-end 2 × 125-bp runs with minimum 20× coverage. The reads were mapped to the human genome assembly (hg19; UCSC browser) with Burrows-Wheeler Alignment (0.6.2, http://bio-bwa.sourceforge.net/). Optical and PCR duplicates were marked and removed with Picard. Local realignment of reads in the indel sites and quality recalibration were performed with the Genome Analysis Tool Kit (version 2.1, http://www.broadinstitute.org/gatk/). Single nucelotide polymorphisms and small indels were called with a GATK UnifiedGenotyper. Initial variants were filtered according to GATK best practice by excluding variants that had SNPs “QD < 2.0,” “MQ < 40.0,” “FS > 60.0,” “MQRankSum < -12.5,” or “ReadPosRankSum < -8.0,” or indels “QD < 2.0,” “ReadPosRankSum < -20.0,” or “FS > 200.0.”

ANNOVAR (version 2-1-2016, http://www.openbioinformatics.org/annovar/) was used to annotate the variants. Variants were additionally filtered to include nonsynonymous, splice site, and indel variants with an allele frequency <0.001 in Exome Aggregation Consortium (12). The variants were confirmed by Sanger sequencing. Testing was also performed in parents of four children with the MC4R variant where DNA samples were available.

Functional studies

Selected MC4R variants were functionally characterized in vitro with cell-based assays to determine the cAMP levels upon stimulation with α-MSH as a measure of MC4R signaling activity and the MC4R surface expression. Both assays were performed in HEK293 cells transiently transfected with wild type (WT) or mutant MC4R. HEK293 cells were maintained in DMEM (cat. no. 11995065; ThermoFisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (cat. no. 10082147; ThermoFisher Scientific). Transfections were performed with Lipofectamine 2000 (cat. no. 11668019; ThermoFisher Scientific) and Opti-MEM I (cat. no. 31985062; ThermoFisher Scientific) according to the manufacturer’s instructions. WT MC4R cDNA construct (N-terminal FLAG tag, fused to EGFP at the C-terminus) was a gift from Dr. Christian Vaisse. Mutations were introduced via site-directed mutagenesis with the QuickChange Lightning Site-Directed Mutagenesis Kit (cat. no. 210518, Agilent Technologies) according to the manufacturer’s recommendations.

• cAMP ELISA: Cells were seeded in 24-well plates (200,000 cells in 500 μL medium per well) 20 hours before transfection with 0.5 μg DNA in 25 μL Opti MEM and 3 μL lipofectamine 2000 in 25 μL Opti-MEM I per well. Then, 24 hours after transfection the medium was aspirated, cells were rinsed once in Krebs-Ringer bicarbonate buffer containing glucose (KRBG; cat. no. K4002-10X1l; Sigma-Aldrich, St. Louis, MO). α-MSH (cat. no. M4135; Sigma-Aldrich), dissolved in KRBG, was added at various concentrations (0, 0.001, 0.01, 0.1, 1, 10, and 100 μM). Cells were incubated for 5 minutes at 37°C. For sample collection, the medium was aspirated, cells were rinsed once with Dulbecco’s phosphate-buffered saline and 200 μL 0.1 M HCl was added, followed by an incubation at room temperature for 10 minutes. Then cells were scraped off with a P1000 pipette and transferred to an Eppendorf tube and placed on ice, followed by vortexing and spinning for 15 minutes at 4° at high speed. The supernatant was transferred into one precooled Eppendorf tubes (200 μL) and stored at −80°C. The ELISA was performed with the cAMP complete ELISA kit (cat. no. ADI-900-163; Enzo Life Sciences, Farmingdale, NY) according to the manufacturer’s guidelines, and 100 μL of each sample was used per reaction. Protein concentration was determined with Pierce™ BCA Protein Assay Kit (cat. no. 23225; ThermoFisher Scientific). Four-parameter log-logistic dose-response curves were analyzed via the drc package in R statistical software. The EC50 was compared with ANOVA (13).

• MC4R surface expression via surface biotinylation: Cells were seeded in 150-mm tissue culture dishes (cat. no. 08-772-24; ThermoFisher Scientific) at a concentration of 16.9 × 10⁶ cells in 42 mL per dish. After 20 hours, transfection was performed with 42 μg DNA in 2111 μL Opti-MEM I and 253 μL Lipofectamine 2000 in 2111 μL Opti-MEM I per dish (one dish per experimental condition). Cells were incubated for 24 hours at 37°C. MC4R cell surface localization was performed with the Pierce Cell Surface Protein Isolation Kit (cat. no. 89881; ThermoFisher Scientific) according to the manufacturer’s instructions. Briefly, cells were rinsed once in KRBG, followed by surface biotinylation by a water-soluble and membrane-impermeable sulfo-NHS-biotin reagent and isolation of biotinylated proteins. Total proteins from each extract were loaded on a 4% to 12% gradient Bis-Tris gel (cat. no. NP0335BOX; ThermoFisher Scientific) and transferred onto nitrocellulose membrane with an iBlot 2 Dry Blotting System (ThermoFisher Scientific). Membrane was blocked for 1 hour at room temperature with 5% nonfat dry milk in Tris-buffered saline with 0.1% Tween-20 (cat. no. 1706531; Bio-Rad, Hercules, CA) and then incubated overnight at 4°C with primary antibody against eGFP Tag (F56-6A1.2.3) from ThermoFisher Scientific (cat. no. MA1-952; 1:1000) to assess the expression of MC4R in the biotinylated cell surface fraction, washed three times in Tris-buffered saline with 0.1% Tween-20 and incubated with secondary antibody anti-mouse horseradish peroxidase (HRP) (1:10,000; cat. no. 7076S; Cell Signaling, Danvers, MA) for 1 hour at room temperature. Specific bands were then detected by electrochemiluminescence analysis with SuperSigna West Pico PLUS Chemiluminescent Substrate (cat. no. 34577; ThermoFisher Scientific). Antibody against ADAM17 (cat. no. ab2051; Abcam, Cambridge, UK; 1:1000) with secondary antibody anti-rabbit HRP (1:10,000; cat. no. 7074S; Cell Signaling) was used as the loading control for the surface fraction. Antibody against β-actin (cat. no. ab8227; Abcam; 1:1000) with secondary antibody anti-rabbit HRP (1:10,000; cat. no. 7074S; Cell Signaling) was used as the loading control for total protein.

Results

This study included a total of 312 unrelated subjects (BCH = 109, CHOP = 157, CUMC = 46). The demographic distribution of the subjects is provided in Table 1. Briefly, 86% of the children were of self-reported African American and 20% of Latino ancestry, with 50% being female. All participating children had severe early-onset obesity (see Methods).

Table 1.

Characteristics of the Study Population

	BCH	CUMC	CHOP
Subjects	109	46	157
Female, n (%)	58 (53.2)	25 (54.3)	74 (47)
Race or ethnicity
Black	56	5	157
Hispanic	67	42	—
Black and Hispanic	14	1	NA
Age at initial visit, y, median (range)	2.25 (0.3–4.8)	0.1 (0–2.1)	0.1 (0–1.1)
Age at most recent contact, y, median (range)	11.07 (8.3–13.1)	9 (6.2–12.2)	7.7 (4.8–11.4)
Duration of follow-up, y, median (range)	5.5 (3.7–10)	4.5 (2.3–10.5)	6.7 (4.3–10.1)
Number of measurements, median (IQR)	32 (35.3)	15 (7)	9 (7)
BMI within individual over all measurements, % of BMI95pct, median (IQR)	140 (124–155)	135 (124–145)	152 (135–173)
Maximal BMI within individual over all measurements, % of BMI95pct, median (IQR)	158 (146–138)	153 (130–165)	152 (135–173)
Median height z score within individual over all measurements, median (IQR)	1.3 (0.5–1.6)	1.3 (0.6–1.5)	1.2 (0.5–1.8)

Abbreviations: IQR, interquartile range; NA, not applicable.

Of the 312 children who underwent sequencing of MC4R, eight were heterozygous for rare MC4R nonsynonymous protein coding variants with minor allele frequency (MAF) <1%. The variants identified were R7S, F202L (n = 2), M215I, G252D, V253I, I269N, and F284I, three of which have not been previously described in the literature (M215I, G252D, and F284I). Another 15 children harbored common (MAF >1%) synonymous or nonsynonymous coding variants (V103I, I198= , Q156= , I251L).

The BMI trajectories of the children carrying these variants did not differ from those of the rest of the cohort [Fig. 1(a], and there was no increase in the height SD when compared with those of comparable obesity without the mutation, unlike in a prior report (14) [Fig. 1(b)]. The clinical features of the subjects are summarized in Table 2. There was no history of consanguinity in any of the families [Fig. 1(c)–1(h)]. In the four families where parental DNA was available, autosomal dominant inheritance was present in three [Fig. 1(c), 1(d), and 1(g)], and one child had a de novo mutation [Fig. 1(e)].

Figure 1.

(a) BMI trajectories fitted with SITAR model of the children enrolled in the study. The colored lines represent children with mutations in MC4R, and gray lines are those without. The dashed lines represent the BMI curves based on CDC 2000 growth charts. (b) Height trajectories fitted with SITAR model of the children enrolled in the study. The colored lines represent children with mutations in MC4R, and gray lines are those without. The dashed lines represent the height curves based on CDC 2000 growth charts. (c–h) Pedigrees of 6 subjects with mutations in MC4R where family data were available. Gray-filled figures are individuals with obesity. Green color represents documented mutation in MC4R gene.

Table 2.

Phenotype and Variant Details of Individuals With MC4R Mutations

Phenotype details
Amino acid change	p. G252D	p. F284I	p.R7S	p.M215I	p. V253I	p. F202L	p. I269N	p.F202L
c.DNA change	c.755 G>A	c.850 T>A	c.19C>T	c.645G>A	c.757 G>A	c.606 C>A	c.806 T>A	c.606 C>A
Center	BCH	BCH	BCH	BCH	CHOP	CHOP	CUMC	CUMC
Race or ethnicity	AA	Latino	AA	Latino	AA	AA	Latino	Latino
Sex	F	M	F	F	F	F	M	M
Birth weight, kg							3.93	3.375
Obesity class	3	2	2	3	2	3	3	3
First available age, mo	48	24	28	37	36	81	0	0
First available BMI, %BMI95pct	153.1	119.5	119.7	133.3	165	124	90	50
Last available age, mo	203	87.7	87.5	160	99	123	80	33.5
Last available height, cm	169.5	125.2	133.5	167.3	154	145.5	119	92.7
Last available height, percentile	84.6	60.65	95.7	86.8	>99	80	56	40
Last available BMI, %BMI95pct	152.3	126.6	140.3	200.6	132	174	130	148
Last available BMI, kg/m2	45.01	24.23	27.94	56.9	34.87	33.4	22.3	26.3
Age, mo	215	99	87	160	98	123	71	33.5
Height, cm	169	130	134	167.3	152	145.5	116	92.7
BP
SBP, mm Hg (percentile)	130 (94)	99 (48)	98 (40)	127 (92)	104 (53)	129 (99)	97 (59)	95 (72)
DBP, mm Hg percentile)	67 (50)	50 (20)	72 (86)	65 (44)	56 (24)	60 (45)	56 (65)	49 (67)
Variant details
gnomAD total allele frequency	4.95E-05	Absent	4E-06	8E-06	6E-05	0.0009	0.00092	0.000898
gnomAD allele frequency Africans	0	0	0	0	0	0.00942	0	0.009415
gnomAD allele frequency Latino	0.00011	0	0	0	0.00011	0.00045	0.00731	0.000452
Polyphen	Probably damaging	Probably damaging	Benign	Probably damaging	Probably damaging	Benign	Probably damaging	Benign
SIFT	Deleterious	Deleterious	Tolerated	Deleterious	Deleterious	Tolerated	Deleterious	Tolerated
CADD score	28.6	26	12.25	28.7	25	12.17	26.8	12.17
Functional	Novel	Novel	Novel	Known	Known	Known	Known	Known
References				Unpublished	Yeo, 2003; Farooqi, 2003; Nijenhuis, 2003; Lubrano-Berthelier, 2003	Tao, 2005; Stutzmann, 2008; Hughes, 2009; Calton, 2009; Xiang, 2010; Hohenadel, 2014	Tan, 2009; Calton, 2009; Thearle, 2012; Hohenadel, 2014;	Tao, 2005; Stutzmann, 2008; Hughes, 2009; Calton, 2009; Xiang, 2010; Hohenadel, 2014

Class 1 obesity, BMI 95%–120% of BMI_95pct; class 2 obesity, BMI 120%–140% of BMI_95pct; class 3 obesity, BMI >140% of BMI_95pct. In silico prediction tools: CADD, Combined Annotation Dependent Depletion (https://cadd.gs.washington.edu/). A scaled CADD score of 20 means that a variant is among the top 1% of deleterious variants in the human genome. A scaled CADD score of 30 means that the variant is in the top 0.1%. Polyphen, Polymorphism Phenotyping (http://genetics.bwh.harvard.edu/pph2/); SIFT, Sorting Intolerant From Tolerant (https://sift.bii.a-star.edu.sg/) (5, 16–24).

Abbreviations: AA, African American; DBP, diastolic blood pressure; F, female; gnomAD, the Genome Aggregation Database (15); M, male; NA, not available; SBP, systolic blood pressure; TG, triglycerides.

A subset of the study subjects had data for cardiometabolic risk factors: blood pressure (n = 140), hemoglobin A1c (n = 121), lipid panel (n = 111), and liver enzymes (n = 122). The proportion of subjects with diabetes was higher in those with MC4R mutations (12.5% vs 0.8%, P = 0.01), and prediabetes was not different (25% vs 15%, P = 0.45); whereas the abnormalities in lipid panel and liver enzymes were not different. Blood pressure in the MC4R-positive subjects was not lower, as previously reported (15); on the contrary, the systolic BP >95th percentile for age, sex, and height was noted in 2 subjects (Table 2). The proportion of children with age of onset of severe obesity <2 years was higher in those with MC4R mutation (75% vs 43%, P = 0.07).

Clinical features of the subjects

p.G252D, 18:58038828 C>T, c.755 G>A

This 17-year-old African American girl with history of severe obesity since 9 months of age was first seen in the medical system to rule out precocious puberty at 4 to 4-1/2 years of age with class 3 obesity (BMI z = 3.45, 153% of BMI_95pct) that continued at 17 years of age. Her linear growth tracked along the 90th percentile until 12 years of age, at which time it reached a plateau with the onset of puberty, with adult height in the 80th percentile. Her mother has a history of obesity, she carries the same MC4R variant, and the maternal family has a history of cardiometabolic diseases. The patient is being treated for polycystic ovary syndrome, severe obesity, and prediabetes. The variant, G252D, is absent in ClinVar and reported primarily in Latino (AF_Lat 0.0001) individuals in gnomAD (25).

p.F284I, 18:58038733 A>T, c.850 T>A

This 8-year-old boy of Puerto Rican descent had a birth weight of 4 kg. He was first seen at the hospital at 2 years of age with BMI of 22.2 kg/m² (BMI z = 3.37, 120% of BMI_95pct). He received a diagnosis of speech delay that resolved with therapy and café-au-lait spots that have grown with age. He does not have any other signs of neurofibromatosis-1 and is negative for mutation in NF-1 gene. His BMI continues to be in the class 2 obesity category with nutritional and psychological intervention, the most recent BMI being 26.60 kg/m² (BMI z = 2.39, 127% of BMI_95pct). His height is currently within the 50th to 60th percentile. Both parents are obese, with high cardiometabolic morbidity in the maternal family. The variant is absent in ClinVar, gnomAD, and HGMD.

p.R7S, 18:58039564 G>T, c.19C>T

This is a 7-year-old African American girl who started gaining weight at 22 months of age. Her BMI was 22.50 kg/m² (BMI z = 3.27, 120% of BMI_95pct) at 2 years of age, and she has consistently remained in obesity class 2 or 3, with the current BMI at 27.90 kg/m² at 7.25 years of age (BMI z = 2.66, 140% of BMI_95pct). Her height was in the 72nd percentile in early childhood and the 95th percentile at 7 years. There is an extensive family history of type 2 diabetes, but the child does not have any cardiometabolic abnormalities. The MC4R variant is inherited from her mother. The variant is absent in ClinVar and reported in 1/248,964 alleles in gnomAD in an individual of South Asian inheritance (AF_SEA 0.00003).

p.M215I. 18:58038938 C>T, c.645G>A

This 14-year-old girl of Latino ancestry had a birth weight of 3.5 kg. The family reports a history of obesity since 8 months of age, and her first recorded BMI at 3 years is 24.30 kg/m² (BMI z = 3.60, 132% of BMI_95pct). Her height has consistently been in the 97th percentile, and BMI at 14 years is 57.50 kg/m² (BMI z = 2.91, >200% of BMI_95pct). At 14 years of age, she has hypertension treated with lisinopril, dyslipidemia, prediabetes with an elevated HbA1c and impaired fasting glucose, hyperinsulinemia, and extensive acanthosis. The variant is absent in ClinVar, and reported in 2/251,254 heterozygous alleles in individuals of non-Finnish European descent in gnomAD (AF_NFE 0.000008) only.

p.V253I, 18: 58038826 C>T, c.757 G>A

This is an 8.5-year-old African American girl with severe obesity since 2.3 years of age (BMI z = 4.7, 174% of BMI_95pct). The available growth data demonstrate tall stature (height z score 1.80 to 2.37 between ages 2.3 and 3.1 years). At study intake, the only noted medical problem was obesity, and at age 7 she was being treated for type 2 diabetes mellitus requiring insulin. This variant has been reported as pathogenic in heterozygous state in ClinVar and 17/282,740 alleles in gnomAD, primarily in Latinos (AF_Lat 0.0001).

p.F202L, 18: 58038977 G>T, c.606 C>A

The subject is a 10-year-old African American girl who had an increase in BMI after 2 years of age to a maximal observed at 7 years (BMI z = 2.63, 132% of BMI_95pct). She was noted to be shorter than typical at younger ages (height z = −1.77 at age 3.2 years), but she changed to average (height z = −0.41) at age 7 years. Aside from obesity, the only documented comorbidity was asthma. This variant has been reported with conflicting interpretations of pathogenicity and seems to be limited to individuals of African American and Latino ancestry in gnomAD Allele Count 254/282,718 alleles, of which 235 were reported in Africans (AF_Afr 0.0094, 1 homozygote) and Latinos (AF_Lat 0.00045) and none in Europeans.

p.I269N, 18:58038777 A>T, c.806 T>A

This child is a 6-year-old boy of Latino ancestry, with BMI of 18.1 kg/m² at 16 months of age (class 1 obesity). At 5 years of age, his BMI is 24.4 kg/m² (BMI z = 3.4, 135% of BMI_95pct). His height has been between the 34th and 60th percentile, and aside from acanthosis, there are no other cardiometabolic risks secondary to obesity. He received a diagnosis of speech delay that has resolved, but he continues to receive therapy for cognitive delay and management of obesity. This variant has been reported with conflicting evidence of pathogenicity in ClinVar and 260/282,764 alleles (AF_total 0.0009195) in gnomAD, with 259 alleles in individuals of Latino descent (AF_Lat 0.0073, 5 homozygotes) and none in Europeans or African.

p.F202L, 18:58038977 G>T, c.606 C>A

This 3-year-old boy of Latino ancestry was born to a mother with severe obesity (reported BMI 40 kg/m²). Pregnancy was complicated by maternal preeclampsia, polyhydramnios, and gestational diabetes, and the baby was delivered at 34 weeks by cesarean section with a birth weight of 3.4 kg and macrosomia. The child received a diagnosis of gastrointestinal reflux, moderate persistent asthma, and toe walking. Weight gain was noted at 12 months, with BMI of 19.45 kg/m², and has continued at 3 years of age with BMI 26.3 kg/m² (BMI z = 4.69, 148% of BMI_95pct). Height has been along the 50th percentile for age. Karyotype and chromosomal microarray are normal, with normal DNA methylation studies for Prader-Willi and Angelman syndrome. Both parents are obese, and mother has undergone vertical sleeve gastrectomy after the birth of the child.

Functional follow-up

To understand the clinical relevance of the identified rare variants, we performed functional assays on four variants, one novel (F284I), one previously characterized but as yet unpublished (M215I) (www.mc4r.uk.org), and two published V253I and G252S (16, 17). Because the primary signaling pathway for MC4R involves the activation of Gα_s subunit and, consequently, adenylyl cyclase (26–28), we measured the half-maximal effective concentration (EC₅₀) of ligand-induced cAMP in transiently transfected HEK293 cells expressing WT or mutant MC4R after exposure to a range of concentrations of endogenous ligand α-MSH [Fig. 2(a)]. In this experiment, the initiation in activating Gα_s signaling required 10-fold higher α-MSH concentration in two variants—M215I (EC₅₀ ratio 3.9) and G252S (EC₅₀ ratio 5.9)—and 100-fold higher concentration in the F284I variant (EC₅₀ ratio 20.9) compared with the WT MC4R. Of note, none of the tested variants produced a peak cAMP response as high as that achieved with WT MC4R protein, suggesting that even the variants with normal EC₅₀ ratios have a functional defect. Cell surface localization of the receptor in MC4R-expressing HEK293 cells was assessed by biotinylation of MC4R [Fig. 2(b); see Methods]. Western blot analysis with antibodies against EGFP to detect the MC4R-EGFP fusion protein showed diminished levels for the mutants as compared with WT, suggesting that decreased expression or protein half-life could contribute to the observed in vitro functional defects for all of these variants. For the variant V253I, although the dose-response curve was close to normal, the surface expression was diminished, indicating intracellular retention as the likely mechanism of pathogenicity (29).

Figure 2.

(a) Dose-response curves for the tested MC4R mutants along with WT MC4R as the positive control and mock as the negative control in HEK-293 cellular overexpression system. The curves are shifted to the right for all the mutants and statistically different in all tested variants except V253I by 4-parameter log logistic dose-response curves. n = 2 for each condition, except for 10 and 100 μM for M215I (n = 1). (b) EC₅₀ (mean ± SE) for optimal cAMP response measured by ELISA. All variants except V253I have P values <0.05 when compared with WT via ANOVA followed by Tukey post hoc test. N = 2 for each condition. (c) Western blot analysis of the surface expression of MC4R protein in HEK-293 cells as determined by surface biotinylation. All mutants have lower expression of the protein compared with WT. ADAM17 and β-actin are loading controls for surface fraction and total protein, respectively. N = 1 for each condition. (d) Quantitative measurement of the ratio of EGFP/ADAM17/β-actin protein expression via ImageJ software for surface expression of the protein. N = 1 for each condition. &, inestimable for mock sample.

Discussion

Mutations in the leptin-melanocortin pathway are the best-defined causes of monogenic obesity, of which autosomal dominant mutations in MC4R are the most common. Although the prevalence of severe obesity is highest in children from non-European ancestry in the United States, there is little information on the monogenic causes in these populations.

To our knowledge, this cohort of 312 children with severe early-onset obesity from self-reported African American or Latino ancestry is the largest reported cohort of children with severe obesity from these populations tested for MC4R variants. Rare pathogenic MC4R variants, shown either by functional testing or previous literature reports, were found in eight (2.6%) children. This prevalence is comparable to that reported in recent studies in children of European ancestry (4, 30) but lower than that reported by Farooqi et al. (5). There are two previous reports of rare variants in MC4R in individuals of African American descent: in a cohort of 25 children from Guadeloupe Island in the Caribbean, one 11.8-year-old boy with severe obesity (BMI 30.5 kg/m², BMI z = 2.39), insulin resistance, and blood lipid abnormalities was found to have a variant in MC4R, I301T that is absent in gnomAD and ClinVar (7); in a report of 167 adults of African descent in South Africa, Logan et al. (8) identified six rare variants (R7H, R165Q, I170V, I198T, F202L, I251V). In gnomAD, low-frequency variants in MC4R such as I251L and V103I are equally distributed across populations of different ancestries, whereas some rare variants are more common in individuals of African descent, such as N240S (AF_Afr 0.002), F202L (AF_Afr 0.009), R7H (AF_Afr 0.0004); Latino descent, such as L304F (AF_Lat 0.0003), T276C (AF_Lat 0.0004), Ile269Asn (AF_Lat 0.0073), A259V (AF_Lat 0.0002), T150I (AF_Lat 0.0005); or East Asian descent, such as M218T (AF_EastAs 0.0005) and Y35C (AF_EastAs 0.001). These data and our results suggest that the repertoire and distribution of rare functional variants in MC4R in obese individuals of African American or Latino ancestry differ from those seen in the studies of individuals of European ancestry.

The study sample is enriched for severe early-onset obesity (BMI >120% of BMI_95pct with an onset of obesity before 6 years of age), and the onset of obesity in most children with the MC4R mutations was before 2 years of age, as in other recent reports (31, 32). We did not observe any differences in the BMI trajectories of the children with the mutation in our cohort. This observation is in keeping with the finding by Kohlsdorf et al. (31), where they observed a much higher BMI at an early age in children with LEP and LEPR mutations, as compared with those with MC4R. Additionally, we did not observe differences in stature of the subjects with or without the mutations. It is possible that this phenotype is more likely to be observed later in life or is different across ancestries or that we are insufficiently powered to observe small differences in height.

Overall, the MC4R gene has been extensively studied, and the recently curated atlas of MC4R variants (www.mc4r.uk.org) in humans is a valuable resource to review the functional relevance of rare variants. In this study, three of the eight (38%) variants observed (M215I, G252D, and F284I) have not been previously described in the literature or in publicly available clinical resources for genetic variants (https://www.ncbi.nlm.nih.gov/clinvar/), probably reflecting both the genetic diversity in children across different racial and ethnic groups and the substantial allelic heterogeneity for pathogenic variants in this gene.

In this cohort of children with severe early-onset obesity, none of the patients had undergone clinical genetic testing for MC4R despite their extreme phenotype seen at an early age. With the advent of safer melanocortin agonist therapies that have been used for patients with POMC, LEPR, and MC4R mutations (28, 33) and the increasing focus on personalized medicine, clinical genetic testing of children with severe early-onset obesity is becoming increasingly relevant. Studies on the relationship between genetic and phenotypic variation have been historically carried out on people of European ancestry, with much lower representation of other racial and ethnic groups. Appropriately, more attention is being paid to the valuable insights that can be gleaned from studies in genetically diverse populations. The clinical role for genetic testing in severe early-onset obesity has been discussed in the guidelines for the management of obesity in children by the Endocrine Society (34), and we agree that, where indicated, opportunities for testing should be made available across the wide range of affected children. Specifically, to interpret genetic testing in the populations most affected, it will be critical to have allele frequency and functional data on variants from children of diverse backgrounds.

Glossary

Abbreviations:

BCH

Boston Children’s Hospital

BMI

body mass index

BMI_95pct

body mass index 95th percentile

CDC

Centers for Disease Control and Prevention

CHOP

Children’s Hospital of Philadelphia

CUMC

Columbia University Medical Center

EC₅₀

half-maximal effective concentration

HRP

horseradish peroxidase

KRBG

Krebs-Ringer bicarbonate buffer containing glucose

MAF

minor allele frequency

MC4R

melanocortin receptor

α-MSH

α-melanocyte–stimulating hormone

SITAR

Super Imposition by Translation and Rotation

WT

wild type