Melanocortin 4 Receptor Gene Sequence Analyses in Diverse Populations

Abstract

Background: Melanocortin 4 receptor (MC4R) is a G-protein-coupled receptor involved in appetite regulation. Mutations in the MC4R gene are the most common cause of monogenic obesity. More than 200 sequence variants in the MC4R gene have been associated with obesity, although the vast majority of these data have been obtained from populations of European ancestry. The prevalence and mutation profile of MC4R is thus poorly characterized in other ancestries/ethnicities.

Materials and Methods: We surveyed the allele frequencies of the MC4R variants of multiple racial/ethnic populations represented in the Genome Aggregation Database (gnomAD) and sequenced the MC4R gene in a diverse population of 60 individuals with extreme obesity.

Results: Allele frequencies were similar for most classes of variants except for a higher rate of synonymous substitutions in the African gnomAD population. We also identified two apparently novel MC4R variants and two variants with much higher allele frequencies in African populations whose functional impacts are not yet known.

Conclusion: These results highlight the need for characterizing MC4R variants in diverse populations with extreme obesity.

Introduction

Obesity is considered a complex trait involving interactions among environmental, behavioral, and genetic factors. Rates of obesity vary across different populations. For example, the frequency of obesity is higher among African Americans (AA) (48.4%) and Hispanic Americans (HA) (42.6%) than among European Americans (EA) (36.4%) (Segal et al., 2016). The frequency of class III extreme obesity, that is, body mass index (BMI) >40 kg/m², is also highest among AA (12.4%) versus HA (7.1%) or EA (7.6%), and highest in AA women (16.8%). Why these differences exist is not known.

The degree to which genetic variation contributes to the observed phenotypic variation in BMI, that is, heritability, is substantial. Based mainly on family and twin data, heritability estimates range from 40% to 80% (Albuquerque et al., 2017). However, common genetic variants identified primarily through GWAS (genome-wide association studies), only explain a minor fraction of the estimates of heritability of obesity (Muller et al., 2018). In contrast, monogenic obesity was found to be caused by mutations in a much smaller set of genes that have strong phenotypic effects, although it appears to account for only a small percentage of cases. The most common cause of monogenic obesity is owing to mutations in the melanocortin 4 receptor (MC4R) gene (Loos, 2011), which encodes a G-protein-coupled receptor that plays a key role in the regulation of food intake and body weight.

Since the first mutations in MC4R in obese humans were reported over 20 years ago, >200 mutations have been identified, primarily heterozygous dominant acting missense variants (Kuhnen et al., 2019). Despite more than two decades of research on MC4R mutations in obesity, most of the data reported were derived from populations of European ancestry. Few data are available from other racial/ethnic populations and most studies have focused on pediatric age groups. We sought to address these knowledge gaps by sequencing the MC4R gene in a diverse population with extreme obesity and characterize the genotypic spectrum of variants in the MC4R gene that were present in a large populations of different racial/ethnic composition that were unselected for obesity.

Materials and Methods

Participants

Blood samples and clinical data were obtained from unselected consecutive patients seen in the Metabolic and Bariatric Surgery Program of the Temple University Health System. All the participants provided written informed consent and all study activities were conducted according to The Code of Ethics of the World Medical Association (Declaration of Helsinki). The Institutional Review Board (IRB) of Temple University approved the research.

Blood samples and DNA preparation

Genomic DNA for sequencing was isolated using the EZ1 DNA Blood 200 μL kit on an EZ1 Advanced XL instrument (Qiagen).

Polymerase chain reaction amplification and sequencing

DNA was amplified using a Veriti 96-well Thermal Cycler (Applied Biosystems). The forward and reverse primers to amplify the entire coding region and 5′ untranslated region (UTR) of the MC4R gene were as previously described (Saeed et al., 2012): forward primer, 5′-GTGAGCATGTGCGCACAGATTC-3′ and reverse primer, 5′-GATATTCTCAACCAGTACCCTACA-3′. The polymerase chain reaction (PCR) amplification conditions were initial denaturation for 3 min at 95°C, followed by 35 cycles of 30 s 95°C denaturation, 30 s of 60°C annealing, and 1 min of 72°C extension, followed by a final 15 min 72°C extension. The 1491 base pair PCR products were then electrophoresed and purified (QIAquick PCR Purification Kit). The purified PCR product was then subject to Sanger sequencing using the forward and reverse PCR primers on an ABI 3730xl DNA Analyzer for capillary electrophoresis and fluorescent dye terminator detection (GeneWiz, NJ). The onboard ABI 3730xl DNA Analyzer instrument software generates Fasta sequence files and fluorescent trace electrophoretograms.

Bioinformatics analysis

Because the automated base calling of sequence data using the ABI 3730xl DNA Analyzer Sequencing Analysis KB Basecaller software cannot distinguish coincident base peaks characteristic of heterozygous variants, each Sanger sequencing capillary electrophoretogram was visually inspected to identify and confirm variants (Supplementary Figs. S1, S2, S3, S4). The NCBI BLAST tool was used to align the DNA sequence to the MC4R sequence (NG_016441.1) and to translate the DNA sequence to its corresponding protein sequence and then align the protein sequences against the protein sequence of human MC4R (CCDS11976.1). Average allele frequencies of MC4R variants in the Genome Aggregation Database (gnomAD) (Lek et al., 2016) version 2.1.1 comprising 141,456 individuals were determined for 3′ UTR, 5′ UTR, frameshift, in-frame deletion, missense, start codon lost, stop codon gained, and synonymous variants by summing the number of variants for each population, African, Latino, Ashkenazi Jewish, East Asian, European (Finnish), European (non-Finnish), and South Asian, and dividing by the average number of chromosomes sequenced.

Results

DNA was isolated from the blood of 60 individuals enrolled in the Metabolic and Bariatric Surgery program at Temple University Health System. About three-fourth of the individuals were women (Table 1) with >85% not of European ancestry and a mean BMI of ∼43 kg/m² (range = 36.0–79.8 kg/m²). A total of four single nucleotide variants were identified in 10 individuals (Table 2). In each case, the variants were heterozygous found by visual inspection of the Sanger sequence electrophoretograms. One individual carried two of the four variants.

Table 1.

Patient Demographics (n = 60)

	Result
Females, %	74
European American, %	13
African American, %	39
Hispanic American, %	49
BMI (range), kg/m2	43.0 (36.0–79.8)
Age (range), years	43 (20–70)

BMI, body mass index.

Table 2.

Sequence Variants Identified in Individuals with Extreme Obesity

ID	Race/ethnic group	BMI, kg/m2	MAF gnomAD	MAF cohort	SNVa	Baseb	dbSNPrs number	Effect
1	EA	44.9	—	0.0714	18-58038543-G-A	6459	—	Intergenic
7	HA	49.4	—	0.0263	18-58038571-G-T	6431	—	3′ untranslated
15	HA	38.9	0.0019	0.0263	18-58038989-G-A	6013	rs61741819	Synonymous
38	AA	40.5	0.0382	0.0208	18-58038989-G-A	6013	rs61741819	Synonymous
24	AA	50.4	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
31	AA	52.6	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
34	AA	48.1	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
35	AA	36.0	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
38	AA	40.5	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
13	AA	44.9	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated
19	AA	45.1	0.1111	0.1458	18-58039760-T-G	5242	rs34114122	5′ untranslated

^aGRCh37/hg19.

^bBP on NG_016441.1.

AA, African Americans; EA, European Americans; gnomAD, Genome Aggregation Database; HA, Hispanic Americans; MAF, minor allele frequencies; SNV, single nucleotide variants.

Little information was available on the intergenic variant (18-58038543-G-A) that was found in an individual of EA ancestry. It is completely conserved across 12 primate species and in 56 of 58 mammalian species with no species carrying an A (Supplementary Fig. S5). It does not appear to be part of any known regulatory elements, although these exhibit tissue specificity and few data are available from specific brain regions.

The 18-58038571-G-A variant was located in the 3′ UTR of the MC4R gene 13 bases downstream from the stop codon (c.*13C>A). It was found in an individual of HA ancestry. The reference base G is conserved in 11 of 12 primate species but poorly conserved across other mammalian species (Supplementary Fig. S6).

The 18-58038989-G-A variant results in a synonymous amino acid, p.Ile198Ile. We identified the variant in one individual of AA ancestry and one of HA ancestry. The Ile is conserved in 12 of 12 primate species and highly conserved across a diversity of vertebrate species (Supplementary Fig. S7).

The 5′ UTR variant 18-58039760-T-G is located 178 bases upstream from the ATG start codon. The A is conserved in 12 of 12 primate species but poorly conserved across other mammalian species (Supplementary Fig. S8).

The intergenic variant (18-58038543-G-A) was not found in the gnomAD or dbSNP databases. A single individual of 56,845 European ancestry had a variant at the base position of the 3′ UTR 18-58038571-G-A variant in the gnomAD database but was a substitution with a T. The A substitution thus seems to be novel. The frequency of the 18-58038989-G-A synonymous variant was >10-fold higher in the gnomAD African ancestry population than in other populations (Supplementary Table S1). However, the frequency of 18-58038989-G-A was statistically higher (p = 0.00064 using a Z-test of proportions) in our HA cohort than in the gnomAD HA population, although not statistically different (p > 0.05) in our AA cohort than in the gnomAD AA population. The 5′ UTR variant 18-58039760-T-G is also more common in individuals of African ancestry (Supplementary Table S2), with an allele frequency over fourfold higher in the gnomAD African ancestry population than in other populations. The frequency of 18-58039760-T-G was not statistically different (p > 0.05) in our AA cohort than in the gnomAD AA population.

We then calculated the average allele frequencies of different classes of MC4R variants (Table 3) in the gnomAD database stratified by racial/ethnic group (Lek et al., 2016). Fewer than 10 variants were found for the in-frame deletion, start lost, and stop gain classes; thus, the estimates of their allele frequencies may be uncertain. The allele frequencies of the other classes of variants varied widely, although this reflects in part the different lengths of sequence representing the different classes of variants, for example, the 3′ and 5′ UTR lengths are different from the length of the coding region that reflects missense and synonymous variants, and different evolutionary pressures for sequence conservation. Among African, Latino, and non-Finnish European populations, the African group had the highest allele frequency for 3′ UTR variants but the lowest for 5′ UTR and frameshift variants. African and non-Finnish European groups had similar rates of missense variants, with ∼8% of the population carrying a variant that was about twice the frequency of the Latino group. The African group had a much higher rate of synonymous variants relative to Latino and non-Finnish European groups with ∼10% of individuals carrying a synonymous variant.

Table 3.

Frequency of MC4R Variants in Different Populations from gnomAD

Population	3′ UTR	5′ UTR	Frameshift	Inframe DEL	Missense	Start lost	Stop gain	Syn
African	2.4 × 10−4	6.5 × 10−5	5.8 × 10−5	0	4.4 × 10−2	2.3 × 10−4	1.1 × 10−4	6.1 × 10−2
Latino	9.7 × 10−5	1.1 × 10−4	1.1 × 10−4	0	2.1 × 10−2	0	2.9 × 10−5	2.7 × 10−3
Ashkenazi Jewish	0	1.3 × 10−4	7.9 × 10−4	4 × 10−4	1.9 × 10−2	0	0	1.0 × 10−4
East Asian	6.0 × 10−5	6.8 × 10−5	5.4 × 10−5	0	2.9 × 10−2	0	3.8 × 10−4	3.0 × 10−3
European (Finnish)	5.0 × 10−5	5.6 × 10−5	4.5 × 10−5	0	3.2 × 10−2	0	0	1.4 × 10−4
European (non-Finnish)	1.0 × 10−4	1.3 × 10−4	1.0 × 10−4	8.8 × 10−6	4.1 × 10−2	0	2.3 × 10−4	1.3 × 10−3
South Asian	0	2.0 × 10−4	9.8 × 10−5	0	3.0 × 10−2	0	9.8 × 10−5	4.9 × 10−3

MC4R, melanocortin 4 receptor; UTR, untranslated region.

Discussion

Variants/mutations in the MC4R gene are the most common cause of monogenic obesity (Meyre et al., 2009; Nordang et al., 2017). Variants with much higher frequency, that is, single nucleotide polymorphisms, located at the MC4R locus have been one of the most highly replicated genetic variants associated with complex (non-Mendelian) obesity (Beckers et al., 2011). The frequency of such variants and their potential role in extreme obesity in adults is not known in AA (Paolini et al., 2016) or HA. However, of the four variants found, two seem to be rare or novel, suggesting that there are likely a large number of MC4R variants yet to be discovered, consistent with recent data indicating that novel loci, secondary signals at known loci, and effect-size heterogeneity is observed when diverse ancestries are analyzed (Wojcik et al., 2019).

The locations of the MC4R variants suggest differential effects on protein function and different evolutionary pressures to maintain sequence conservation. The 3′ UTR and 5′ UTR regions are important in regulating gene expression through RNA stability and translation. Both variants we identified were relatively highly conserved among mammalian species but poorly conserved among other vertebrate species. The novel 3′ UTR variant is more likely to have a functional impact, although functional classifications of variant effects on MC4R protein function that have been developed are based on biochemical evidence of activation of G proteins, ligand binding, and cell-surface expression (Tao, 2005), as well as intracellular retention and constitutive versus ligand-stimulated activation (Govaerts et al., 2005). The development of synthetic MC4R agonists to treat obesity has led to further functional characterization (Kuhnen et al., 2019), although this requires in vitro systems that may not recapitulate human physiology. Variants affecting RNA stability and turnover are not yet part of these systems.

The higher rate of synonymous sequence variation in the MC4R gene in an African population is consistent with numerous other studies in which nucleotide diversity is lower in non-Africans than Africans because of a population bottleneck resulting from the ancestral migrations of humans out of Africa (Campbell and Tishkoff, 2008). The ancestral populations also experienced different selective pressures over evolutionary time scales because of differences in geography. The disparate environmental influences therefore likely selected for different variants in the MC4R gene that would have functional effects on appetite and body weight. This further highlights the need for characterizing diverse populations.

Data Sharing

The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available because of privacy or ethical restrictions.

Authors' Contributions

Study concept and design, G.S.G. and M.A.E.; acquisition of data, T.T., G.S., A.S., and S.K.; analysis and interpretation of data, G.S.G., M.A.E., T.T., G.S., A.S., and S.K.; drafting of the article, G.S.G. and M.A.E.; critical revision of the article for important intellectual content, G.S.G. and M.A.E.; obtained funding, G.S.G.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Glenn S. Gerhard was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (Grant No. DK107735). The funding sources did not play any role in the design of the study, collection, analysis, or interpretation of the data, or developing and composing the article.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Figure S8