https://orcid.org/0000-0003-4850-7444
Abstract
Rare partial/complete loss-of-function mutations in the melanocortin-4 receptor (MC4R) gene are the most common cause of Mendelian obesity in European populations, but their contribution to obesity in the Mexican population is unclear.
We investigated whether deleterious mutations in MC4R contribute to obesity in Mexican children and adults.
We provide evidence that the MC4R p.Ile269Asn (rs79783591) mutation may have arisen in modern human populations from a founder event in native Mexicans. The MC4R Isoleucine 269 is perfectly conserved across 184 species, which suggests a critical role for the amino acid in MC4R activity. Four in silico tools (SIFT, PolyPhen-2, CADD, MutPred2) predicted a deleterious impact of the p.Ile269Asn substitution on MC4R function. The MC4R p.Ile269Asn mutation was associated with childhood (Ncontrols = 952, Ncases = 661, odds ratio (OR) = 3.06, 95% confidence interval (95%CI) [1.94–4.85]) and adult obesity (Ncontrols = 1445, Ncases = 2,487, OR = 2.58, 95%CI [1.52–4.39]). The frequency of the MC4R p.Ile269Asn mutation ranged from 0.52 to 0.59% and 1.53 to 1.59% in children and adults with normal weight and obesity, respectively. The MC4R p.Ile269Asn mutation co-segregated perfectly with obesity in 5 multigenerational Mexican pedigrees. While adults with obesity carrying the p.Ile269Asn mutation had higher BMI values than noncarriers, this trend was not observed in children. The MC4R p.Ile269Asn mutation accounted for a population attributable risk of 1.28% and 0.68% for childhood and adult obesity, respectively, in the Mexican population.
The MC4R p.Ile269Asn mutation may have emerged as a founder mutation in native Mexicans and is associated with childhood and adult obesity in the modern Mexican population.
Worldwide, the prevalence of obesity has nearly tripled since 1975, but certain countries are more heavily impacted than others (1). For instance, a 2019 report by the Organization for Economic Co-operation and Development found that Mexico has among the highest rates of obesity in the world, second to the United States. According to the 2016 national survey of health and nutrition in Mexico, the prevalence of obesity was 15.3% in children between the ages of 5 and 11 years, 13.9% in teenagers between the ages of 12 and 19 years, and 33.3% in adults (2). Obesity, especially in its early-onset forms, is an important risk factor for multiple comorbidities (eg, type 2 diabetes [T2D], cardiovascular disease, cancer) and premature all-cause mortality (3–5). Efforts have been made to tackle childhood and adult obesity in Mexico, but they have failed at curbing the epidemic thus far (1, 2, 6, 7). In this context, more research on the causes of obesity is needed in this high-risk population to improve obesity prediction, prevention, and care (8).
The recent surge in obesity in Mexico can be largely explained by the rapid nutritional transition to Western diets and a decrease in physical activity promoted by fast urbanization (9, 10). However, not everyone exposed to an “obesogenic” environment becomes obese in Mexico (2), and the large inter-individual differences in body mass index (BMI) observed in this high-risk population are also attributable to biological factors: in utero exposure, sex, age, ancestral background, pre-existing medical conditions, gut microbiome, epigenetic and genetics (8, 11).
Twin and family studies in diverse ethnic groups, including Latino Americans, suggest that 40% to 75% of BMI variation can be explained by genetic factors (12). Rare and common genetic variations responsible for monogenic, oligogenic and polygenic forms of obesity have been identified, predominantly in European populations (12). Several studies have demonstrated the partial transferability of European-derived polygenic obesity loci to the Mexican population (11, 13, 14). By contrast, identification of the genetic contributors to oligogenic and monogenic obesity in the Mexican population remains elusive. As partial and complete MC4R deficiency represents the most common cause of monogenic/oligogenic obesity in European populations (15), we investigated whether deleterious founder mutations in the melanocortin-4 receptor (MC4R) gene contribute to obesity in Mexican children and adults.
Materials and Methods
Participants
The Genome Aggregation Database (GnomAD) includes 141 456 unrelated individuals sequenced (125 748 exomes and 15 708 genomes) as part of disease-specific and population-based genetic studies (16, 17). We included only individuals with exome data in our analysis to maintain a consistent sample size for estimating the prevalence of rare mutations and common polymorphisms. Exome-sequenced individuals were grouped into seven ethnic groups using principal component analysis: Europeans (N = 67 709), Latino (N = 17 296), South Asians (N = 15 308), East Asians (N = 9197), Africans/African Americans (N = 8128), Ashkenazi Jewish (N = 5040), and other (N = 3070). We excluded individuals in the “other” category, as they belonged to several ethnicities and details on their geographic origin were not reported. We accessed data from gnomAD on November 30, 2018.
The 1000 Genomes Project (1000G) provides a comprehensive description of common human genetic variation identified by sequencing a diverse set of individuals from multiple populations (18). The most recent version of 1000G includes 2504 individuals from 5 ethnic groups in 26 populations using a combination of low-coverage whole-genome sequencing, deep exome sequencing, and dense microarray genotyping (18).
A total of 1613 unrelated Mexican children (952 with normal weight and 661 with obesity) from 2 case-control studies were included in this study. The first study, conducted from June 2016 to October 2018, enrolled 1029 children (685 with normal weight and 344 with obesity) between the ages of 6 and 12 years from 3 states of Mexico (Campeche, Oaxaca, and Mexico City). As part of the National Obesity Network Mexico initiative, we collected genetic and anthropometric data in a second case-control study of 584 children (267 with normal weight and 317 with obesity) between the ages of 6 and 12 years, enrolled from 12 Mexican states (Baja California Sur, Estado de Mexico, Guanajuato, Hidalgo, Michoacán, Nayarit, Nuevo León, Puebla, Queretaro, Sinaloa, Sonora and Tamaulipas; replication sample). Recruitment occurred from June 2016 to October 2018.
A total of 3932 unrelated Mexican adults were enrolled in this study. Of these participants, 1445 were normal weight and 2487 were obese. Cases and controls were recruited from primary health care facilities in Mexico City and from the central blood bank of the National Medical Center “Siglo XXI” in Mexico City. Participants with 2 identical last names were considered related and only 1 randomly selected family member was included in the children and adult case control studies.
Research was approved by local institutional review boards. Written informed consent was obtained from each subject prior to participation in accordance with the Declaration of Helsinki (19).
Conservation of the MC4R Isoleucine 269 amino acid across species
MC4R protein sequences were obtained from the UniProtKb database (https://www.uniprot.org/help/uniprotkb). The Basic Local Alignment Search Tool (BLAST) was used to search for protein sequences with the human MC4R amino acid sequence (332 amino acids) as the input (20). Of the 250 hits, 184 species remained after removal of redundant entries. The sequences were imported in JalView for visualization (21). MUSCLE was used for multiple sequence alignment with the human MC4R protein sequence set as the reference (22). Conservation scores were calculated using the method described by Livingstone et al (22). Scores ranged from 1 to 11, with a score of 11 indicating no amino acid substitutions and perfect conservation among the 184 species (23).
MC4R p.Ile269Asn (rs79783591) ancestral and derived alleles in primates, archaic and modern humans
Allelic status of the MC4R rs79783591 polymorphism in 10 primate (chimpanzee, gorilla, orangutan, macaque, marmoset, gibbon, bonobo, vervet, crab-eating macaque, and olive baboon) genomes was obtained from the Ensembl Genome Browser (http://useast.ensembl.org/index.html) (24), and in archaic human genomes (Neandertals and Denisovans) from the UCSC Genome Browser (https://genome.ucsc.edu/) (25–27). The ancestral/derived status of the alleles was determined from the erythropoietin alignment (28).
DNA extraction, sequencing, and genotyping procedures
Genomic DNA of all participants was isolated from peripheral blood using the AutoGen Flex STAR system (Auto-Gen, Holliston, MA, US), and purity and integrity were verified by 260/280 nm measurements (BioTek Instruments, Winooski, VT, US) and by electrophoresis in 0.8% agarose gels stained with ethidium bromide. Genotyping of the MC4R p.Ile269Asn/rs79783591 mutation was performed by real-time polymerase chain reaction using TaqMan® allelic discrimination assay C_103977380_10 (Thermo Fisher Scientific) on a 7900HT Fast Real-Time PCR system (Applied Biosystems, CA, US), following standard protocols. Genotype discrimination was evaluated using the SDS software (Applied Biosystems, CA, US). Genotyping call rates of 98.6% and 98.0% were observed for the MC4R p.Ile269Asn mutation in the child and adult samples, respectively. No deviation from Hardy–Weinberg equilibrium was observed for the mutation in the control groups of the child and adult samples (P between .82 and .87) (29). Genotypes were duplicated in 10% of the discovery and replication samples and genotyping concordance was 100%. We compared allele frequencies of the children and adult samples with that of the adult Mexican American reference population in 1000G (29). Allele frequencies were not significantly different from the reported frequencies in 1000G.
Functional characterization of coding mutations
Nonsense, splice acceptor, splice donor, and frameshift mutations were considered as loss of function. To test whether missense mutations significantly impacted protein function, we used the PolyPhen 2 and SIFT software in gnomAD (30, 31). We further investigated the functional consequences of the MC4R p.Ile269Asn mutation using 6 in silico analysis tools: SIFT (30), PolyPhen-2 (31), CADD (32), REVEL (33), MutPred2 (34), and FATHMM (35).
Anthropometric and biochemical measurements
Weight was measured using a digital scale (Seca, Hamburg, Germany). Height was measured using a portable stadiometer (Seca 225, Hamburg, Germany). BMI was calculated as weight (kg)/height (m)2. In adults, normal weight, overweight, and obesity were defined as a BMI < 25 kg/m2, 25 kg/m2 ≤ BMI < 30 kg/m2, and BMI ≥ 30 kg/m2, respectively. In children, BMI was converted to age- and gender-adjusted standard deviation scores (BMI SDS) using guidelines from the Centers for Disease Control (36). Age- and gender- specific BMI percentiles were calculated using the Centers for Disease Control 2000 reference to classify children as normal weight (BMI between the 5th to 85th percentile), overweight (BMI between 85th to 95th percentile) and obese (BMI at or above the 95th percentile) (37). Fasting blood and 2-h postoral glucose tolerance test plasma glucose levels were measured using the ILab 350 Clinical Chemistry System (Instrumentation Laboratory IL. Barcelona, Spain). Diabetes was diagnosed using the American Diabetes Association criteria (ie, fasting plasma glucose ≥7.0 mmol/L or OGTT ≥ 11.1 mmol/L for the 2 h sample).
Data analysis
Differences between cases and controls for continuous and categorical traits were tested using Student’s t and chi-squared tests, respectively. Linear and logistic regression models adjusted for age, sex, and recruitment center (adults) were used to assess genetic associations. Generalized estimating equations were used to further adjust genetic associations for Mexican state in children. A logistic regression model adjusted for age, sex, and the first 10 principal components computed from SMARTPCA (38) was performed in a subset of 139 and 125 children with normal weight and obesity. Genetic analyses of the MC4R p.Ile269Asn mutation were performed under an additive model using the minor allele as the effect allele. Penetrance values were estimated from case-control studies using the Bayes’ theorem, as previously described (15). The population-attributable risk (PAR) of obesity due to the MC4R p.Ile269Asn mutation was calculated using a tool from the Centre for Clinical Research and Biostatistics (https://www2.ccrb.cuhk.edu.hk). Statistical analyses were conducted using SPSS software (version 22.0, IBM, Armonk, NY, US). A 2-sided P-value < .05 was considered significant.
Results
Identification of the mutation MC4R p.Ile269Asn in Latino populations
We investigated the presence of founder mutations in MC4R in 17 296 exome-sequenced Latino individuals in gnomAD, restricting analyses to nonsense, splice acceptor, splice donor, frameshift, and missense mutations. Mutations were considered founder mutations if they (i) had a minor allele frequency (MAF) higher or equal to 0.5% in the Latino population (39) and (ii) were absent from other ethnic groups. Only the coding missense mutation p.Ile269Asn (rs79783591) in MC4R met these criteria. The mutation was relatively common in the Latino population (MAF: 0.75%; 5 homozygous [0.029%] and 248 heterozygous [1.43%] carriers out of 17 295 individuals), but was absent in the other ethnic groups (European, N = 67 650; African, N = 8128; South Asian, N = 15 307; East Asian, N = 9190; Ashkenazi Jewish, N = 5037). Therefore, we focused subsequent analyses on the MC4R p.Ile269Asn mutation.
The MC4R p.Ile269Asn mutation may have resulted from a founder event in native Mexicans
In line with the observation in gnomAD, analysis of the MC4R p.Ile269Asn mutation in 1000G revealed an absence of the mutation in African (N = 661), East Asian (504), South Asian (N = 489), and European (N = 503) individuals. We found 1 heterozygous carrier in 64 people of Mexican ancestry living in Los Angeles, California, US (MAF: 0.78%). No other carriers were identified in three other Latino populations, comprising 94 Columbians from Medellin, 85 Peruvians from Lima, and 104 Puerto Ricans from Puerto Rico. Collectively, our data suggest that the mutation did not come from a population migration gene flow event but rather emerged as a founder mutation in the native Mexican population.
The MC4R p.Ile269Asn mutation rose in modern human populations
We found that the ancestral A allele of MC4R rs79783591 is shared by modern and archaic (Neandertal and Denisovan) humans and 10 primates (29). By contrast, the derived T allele was only found in modern humans, which suggests that it arose de novo in modern human populations.
Perfect phylogenetic conservation of the MC4R Isoleucine 269 across 184 species
The isoleucine amino acid at position 269 in MC4R was found in 100% of the aligned sequences across 184 species. The amino acid conservation score was 11 (23), suggesting a perfect phylogenic conservation and a critical role for the amino acid in the normal function of MC4R.
In silico predictions suggest a deleterious effect of the MC4R p.Ile269Asn mutation on MC4R function
The MC4R p.Ile269Asn mutation was predicted to have deleterious functional consequences by a majority (67%) of the in silico prediction tools (Table 1): deleterious (SIFT), probably damaging (PolyPhen-2), likely deleterious (CADD), likely pathogenic (MutPred2), likely benign (REVEL), and tolerated (FATHMM). Overall, these data suggest that the MC4R p.Ile269Asn mutation is likely to negatively impact MC4R function.
In silico Analyses of Ile269Asn in the MC4R Gene
| Software | SIFT | PolyPhen2 | FATHMM | MutPred2 | CADD | REVEL |
|---|---|---|---|---|---|---|
| Output and score values for the MC4R p.Ile269Asn mutation | Deleterious (0) | Probably damaging (0.992) | Tolerated (1.16) | Likely pathogenic (0.880) | Deleterious (27.1) | Likely benign (0.487) |
| Machine learning method | Neural network | Naïve Bayes classifier | Hidden Markov models with support vector machine classifier | Neural network | Support vector machine | Random forests |
| Software | SIFT | PolyPhen2 | FATHMM | MutPred2 | CADD | REVEL |
|---|---|---|---|---|---|---|
| Output and score values for the MC4R p.Ile269Asn mutation | Deleterious (0) | Probably damaging (0.992) | Tolerated (1.16) | Likely pathogenic (0.880) | Deleterious (27.1) | Likely benign (0.487) |
| Machine learning method | Neural network | Naïve Bayes classifier | Hidden Markov models with support vector machine classifier | Neural network | Support vector machine | Random forests |
In silico Analyses of Ile269Asn in the MC4R Gene
| Software | SIFT | PolyPhen2 | FATHMM | MutPred2 | CADD | REVEL |
|---|---|---|---|---|---|---|
| Output and score values for the MC4R p.Ile269Asn mutation | Deleterious (0) | Probably damaging (0.992) | Tolerated (1.16) | Likely pathogenic (0.880) | Deleterious (27.1) | Likely benign (0.487) |
| Machine learning method | Neural network | Naïve Bayes classifier | Hidden Markov models with support vector machine classifier | Neural network | Support vector machine | Random forests |
| Software | SIFT | PolyPhen2 | FATHMM | MutPred2 | CADD | REVEL |
|---|---|---|---|---|---|---|
| Output and score values for the MC4R p.Ile269Asn mutation | Deleterious (0) | Probably damaging (0.992) | Tolerated (1.16) | Likely pathogenic (0.880) | Deleterious (27.1) | Likely benign (0.487) |
| Machine learning method | Neural network | Naïve Bayes classifier | Hidden Markov models with support vector machine classifier | Neural network | Support vector machine | Random forests |
The MC4R p.Ile269Asn mutation is associated with childhood obesity in Mexicans
We assessed the association of the MC4R p.Ile269Asn mutation with obesity in 952 and 661 Mexican children with normal weight and obesity, respectively. Anthropometric and genotyping data are presented in Table 2. Age and sex ratios were not significantly different between the normal-weight and obesity groups. On average, children with obesity had a 7.8 kg/m2 higher BMI and a 1.8 higher BMI SDS than children with normal weight. The MC4R p.Ile269Asn mutation was 3 times more prevalent in children with obesity than children with normal weight (MAF: 1.59% vs 0.52%, P = .0023). The frequency of heterozygous carriers of the MC4R 269Asn allele was also higher in children with obesity than in children with normal weight (3.2% vs 1.1%; P = .002). No homozygous carrier of the 269Asn allele was identified. The MC4R p.Ile269Asn variant was associated with childhood obesity under an additive model (odds ratio [OR] = 3.06, 95% confidence interval [95%CI] [1.43–6.56], P = .004, model adjusted for sex and age). Further adjustment for Mexican state using a generalized estimating equation model did not sensibly change the association (OR = 3.06, 95%CI [1.94–4.85], P = 1.7 × 10–6). To exclude the possibility of a spurious association caused by population stratification, we compared the association between the MC4R p.Ile269Asn mutation and childhood obesity before and after adjustment for population structure in a subset of 139 and 125 children with normal weight and obesity for whom genome-wide SNP genotyping data were available. Age, sex ratio and BMI SDS did not differ between the genome-wide genotyped and other children (.52 ≤ P ≤ .97), to the exception of BMI SDS in children with normal-weight (genome-wide genotyped: BMI SDS = 0.223 ± 0.684; others: BMI SDS = 0.440 ± 0.820, P = .003). The association between the MC4R p.Ile269Asn mutation and childhood obesity, while less precisely estimated in this modestly powered subset of children, was significant and did not sensibly change before (OR = 9.10, 95%CI [1.12–73.98], P = .039, model adjusted for sex and age) and after adjustment for population stratification (OR = 9.57, 95%CI [1.15–79.66], P = .037, model adjusted for sex, age, and 10 principal components from SMARTPCA). No significant difference in BMI SDS was observed between noncarriers and heterozygous carriers of the MC4R p.Ile269Asn mutation in the obesity group (Ile269Ile: SDS BMI = 2.01 ± 0.50 vs Ile269Asn: BMI SDS = 2.05 ± 0.44, β = .048 ± .068, P = .481, model adjusted for sex, age and Mexican state). Using Bayes’ theorem, the estimated penetrance of obesity in heterozygous children was 27.5%. The PAR of childhood obesity due to the MC4R p.Ile269Asn mutation was 1.28%.
General Characteristics of Mexican Children and Adults with Normal-Weight and Obesity
| Trait | Normal-weight | Obesity | p-value |
|---|---|---|---|
| Children sample | N = 952 | N = 661 | |
| Female, N (%) | 490 (51.5) | 327 (49.5) | .429 |
| Age (years) | 8.9 ± 2.0 | 9.0 ± 1.7 | .266 |
| BMI (kg/m2) | 17.0 ± 2.5 | 24.8 ± 3.2 | 2.4 × 10 –38 |
| BMI SDS | 0.218 ± 1.03 | 2.013 ± 0.5 | 3.9 × 10 –32 |
| Rs79783591 A/A, N (%) | 942 (98.9) | 640 (96.8) | .002 |
| Rs79783591 A/T, N (%) | 10 (1.1) | 21 (3.2) | |
| Adult sample | N = 1445 | N = 2487 | |
| Female, N (%) | 511 (35.3) | 1,015 (40.8) | .001 |
| Age (years) | 47.1 ± 11.6 | 49.6 ± 10.0 | 5.2 × 10 –9 |
| BMI (kg/m2) | 23.2 ± 1.5 | 33.6 ± 3.6 | 6.4 × 10 –45 |
| T2D, N (%) | 430 (29.8) | 1,093 (43.9) | 6.5 × 10 –27 |
| Rs79783591 A/A, N (%) | 1,428 (98.8) | 2,413 (97.0) | .001 |
| Rs79783591 A/T, N (%) | 17 (1.2) | 72 (2.9) | |
| Rs79783591 T/T, N (%) | 0 (0) | 2 (0.1) |
| Trait | Normal-weight | Obesity | p-value |
|---|---|---|---|
| Children sample | N = 952 | N = 661 | |
| Female, N (%) | 490 (51.5) | 327 (49.5) | .429 |
| Age (years) | 8.9 ± 2.0 | 9.0 ± 1.7 | .266 |
| BMI (kg/m2) | 17.0 ± 2.5 | 24.8 ± 3.2 | 2.4 × 10 –38 |
| BMI SDS | 0.218 ± 1.03 | 2.013 ± 0.5 | 3.9 × 10 –32 |
| Rs79783591 A/A, N (%) | 942 (98.9) | 640 (96.8) | .002 |
| Rs79783591 A/T, N (%) | 10 (1.1) | 21 (3.2) | |
| Adult sample | N = 1445 | N = 2487 | |
| Female, N (%) | 511 (35.3) | 1,015 (40.8) | .001 |
| Age (years) | 47.1 ± 11.6 | 49.6 ± 10.0 | 5.2 × 10 –9 |
| BMI (kg/m2) | 23.2 ± 1.5 | 33.6 ± 3.6 | 6.4 × 10 –45 |
| T2D, N (%) | 430 (29.8) | 1,093 (43.9) | 6.5 × 10 –27 |
| Rs79783591 A/A, N (%) | 1,428 (98.8) | 2,413 (97.0) | .001 |
| Rs79783591 A/T, N (%) | 17 (1.2) | 72 (2.9) | |
| Rs79783591 T/T, N (%) | 0 (0) | 2 (0.1) |
Data are expressed as mean ± standard deviation and N (%). SDS: metabolic traits age- and sex-adjusted standard deviation scores. Difference in sex ratios was analyzed using the χ 2 test. Differences in means were analyzed using Student’s t-tests. Significant p values (P < .05) are reported in bold.
Aabbreviations: BMI, body mass index; T2D, type 2 diabetes.
General Characteristics of Mexican Children and Adults with Normal-Weight and Obesity
| Trait | Normal-weight | Obesity | p-value |
|---|---|---|---|
| Children sample | N = 952 | N = 661 | |
| Female, N (%) | 490 (51.5) | 327 (49.5) | .429 |
| Age (years) | 8.9 ± 2.0 | 9.0 ± 1.7 | .266 |
| BMI (kg/m2) | 17.0 ± 2.5 | 24.8 ± 3.2 | 2.4 × 10 –38 |
| BMI SDS | 0.218 ± 1.03 | 2.013 ± 0.5 | 3.9 × 10 –32 |
| Rs79783591 A/A, N (%) | 942 (98.9) | 640 (96.8) | .002 |
| Rs79783591 A/T, N (%) | 10 (1.1) | 21 (3.2) | |
| Adult sample | N = 1445 | N = 2487 | |
| Female, N (%) | 511 (35.3) | 1,015 (40.8) | .001 |
| Age (years) | 47.1 ± 11.6 | 49.6 ± 10.0 | 5.2 × 10 –9 |
| BMI (kg/m2) | 23.2 ± 1.5 | 33.6 ± 3.6 | 6.4 × 10 –45 |
| T2D, N (%) | 430 (29.8) | 1,093 (43.9) | 6.5 × 10 –27 |
| Rs79783591 A/A, N (%) | 1,428 (98.8) | 2,413 (97.0) | .001 |
| Rs79783591 A/T, N (%) | 17 (1.2) | 72 (2.9) | |
| Rs79783591 T/T, N (%) | 0 (0) | 2 (0.1) |
| Trait | Normal-weight | Obesity | p-value |
|---|---|---|---|
| Children sample | N = 952 | N = 661 | |
| Female, N (%) | 490 (51.5) | 327 (49.5) | .429 |
| Age (years) | 8.9 ± 2.0 | 9.0 ± 1.7 | .266 |
| BMI (kg/m2) | 17.0 ± 2.5 | 24.8 ± 3.2 | 2.4 × 10 –38 |
| BMI SDS | 0.218 ± 1.03 | 2.013 ± 0.5 | 3.9 × 10 –32 |
| Rs79783591 A/A, N (%) | 942 (98.9) | 640 (96.8) | .002 |
| Rs79783591 A/T, N (%) | 10 (1.1) | 21 (3.2) | |
| Adult sample | N = 1445 | N = 2487 | |
| Female, N (%) | 511 (35.3) | 1,015 (40.8) | .001 |
| Age (years) | 47.1 ± 11.6 | 49.6 ± 10.0 | 5.2 × 10 –9 |
| BMI (kg/m2) | 23.2 ± 1.5 | 33.6 ± 3.6 | 6.4 × 10 –45 |
| T2D, N (%) | 430 (29.8) | 1,093 (43.9) | 6.5 × 10 –27 |
| Rs79783591 A/A, N (%) | 1,428 (98.8) | 2,413 (97.0) | .001 |
| Rs79783591 A/T, N (%) | 17 (1.2) | 72 (2.9) | |
| Rs79783591 T/T, N (%) | 0 (0) | 2 (0.1) |
Data are expressed as mean ± standard deviation and N (%). SDS: metabolic traits age- and sex-adjusted standard deviation scores. Difference in sex ratios was analyzed using the χ 2 test. Differences in means were analyzed using Student’s t-tests. Significant p values (P < .05) are reported in bold.
Aabbreviations: BMI, body mass index; T2D, type 2 diabetes.
The MC4R p.Ile269Asn mutation is associated with adult obesity in Mexicans
We assessed the association of the MC4R p.Ile269Asn mutation with obesity in 1,445 and 2,487 Mexican adults with normal weight and obesity, respectively. Adults with obesity were more often women and had a higher age, BMI, and frequency of T2D than adults with normal weight (Table 2). The MC4R p.Ile269Asn mutation was about 3 times more prevalent in adults with obesity than in adults with normal weight (MAF: 1.53% vs 0.59%, P = .0002). A higher frequency of heterozygous carriers of MC4R 269Asn allele was also observed in adults with obesity (2.9% vs 1.2%; P = .001) (Table 2). We identified 2 homozygous carriers of the 269Asn allele in the obesity group. The MC4R p.Ile269Asn variant allele was associated with adult obesity under an additive model (OR = 2.58, 95%CI [1.52–4.39], P =.00043, model adjusted for sex, age, and recruitment center). Additional adjustment for T2D status did not sensibly change the association (OR = 2.55, 95%CI [1.50–4.33], P = .00053). We observed an increase in the BMI of noncarriers compared to heterozygous carriers and those homozygous for the mutation in the obesity group (Ile269Ile: BMI = 33.6 ± 3.5 vs Ile269Asn = 34.5 ± 5.0, vs Asn269Asn = 36.9 ± 5.6, β = 1.004 ±.411, P = .015, model adjusted for sex, age, and recruitment center). Using Bayes’ theorem, the estimated penetrance of obesity in heterozygous and homozygous adults was 67.9% and 100%, respectively. The PAR of adult obesity due to the MC4R p.Ile269Asn mutation was 0.68%.
Meta-analysis of association of the MC4R p.Ile269Asn mutation with obesity in Mexicans
We then meta-analyzed the data from Mexican children and adults and confirmed a strong association between the MC4R p.Ile269Asn variant allele and childhood/adult obesity (OR = 2.85, 95%CI [2.01–4.03], P = 8.4 × 10–9). No significant genetic heterogeneity was found between children and adults (I2 = 0%, P = .63).
Co-segregation of the MC4R p.Ile269Asn mutation with obesity in Mexican multigenerational pedigrees
We studied the co-segregation of the MC4R p.Ile269Asn mutation with obesity in 5 nonconsanguineous multigenerational Mexican pedigrees totaling 25 individuals (14 noncarriers, 10 heterozygous carriers, 1 homozygous carrier; Fig. 1). While 7.1% (N = 1), 42.9% (N = 6), and 50% (N = 7) of noncarriers displayed normal weight, overweight, and obesity, respectively, 100% of heterozygous and homozygous carriers of the MC4R p.Ile269Asn mutation had obesity. The mutation co-segregated perfectly with obesity in the pedigrees, and heterozygous and homozygous carriers of the mutation exhibited complete familial penetrance of obesity.
Description of 5 Mexican multigenerational nonconsanguineous pedigrees. Individuals with obesity, overweight, and normal weight are colored in black, gray, and white, respectively. Age in years old (y.o.), body mass index (BMI, adults) and BMI SDS (children) are reported for each individual. Homozygous carriers of A allele of MC4R rs79783591 (AA) are reported as “NN.” Heterozygous carriers TA and homozygous carriers TT are reported as “NM” and “MM,” respectively. Individuals with unavailable clinical and genetic data are reported as “?”.
Discussion
In this study, we provide evidence that the MC4R p.Ile269Asn mutation may have arisen de novo in modern human populations from a founder event in the native Mexican population. MC4R Isoleucine 269 is perfectly conserved across 184 species, consistent with a critical role for the amino acid in MC4R activity. Four in silico analysis tools predicted a deleterious impact of the p.Ile269Asn substitution on MC4R function. The MC4R p.Ile269Asn mutation is relatively widespread in the Mexican population (frequency of 0.52–0.59% in children and 1.53–1.59% and adults with normal weight and obesity, respectively) and is conclusively associated with childhood and adult obesity with a consistent effect size (OR = 2.85, 95%CI [2.01–4.03], P = 8.43 × 10–9). We also excluded the possibility that this association may result from a population stratification bias. Heterozygous and homozygous carriers of the MC4R p.Ile269Asn mutation exhibited complete familial penetrance of obesity in five multigenerational Mexican pedigrees. By contrast, a penetrance of 27.5 to 67.9% and 100% was estimated from the case-control data in heterozygous and homozygous mutation carriers, respectively. Whereas children with obesity had comparable BMI SDS values regardless of MC4R p.Ile269Asn mutation status, a progressive increase in the BMI of noncarriers, heterozygous carriers, and homozygous carriers of the mutation was observed in the adult obesity group, in line with our previous findings in the European population (15). The MC4R p.Ile269Asn founder mutation was estimated to account for a PAR for childhood and adult obesity of 1.28% and 0.68% in the Mexican population.
The MC4R p.Ile269Asn mutation was first described in 2009 in two heterozygous Hispanic children with severe obesity living in the United Kingdom (40). In vitro functional experiments demonstrated that the p.Ile269Asn mutation reduced cell surface expression of the MC4R by 30% and impaired its capacity to bind alpha-MSH by 70%, resulting in the complete inability of the mutant MC4R to induce cAMP production in response to its agonist (40). Structural modeling of the MC4R suggests that the p.Ile269Asn mutation may change the packing of the extracellular ends of TM6 and TM5 that is essential for activation and, thus, may affect the receptor stability, leading to misfolding and loss of function (40). The same year, the p.Ile269Asn mutation was identified in 2 heterozygous patients with morbid obesity living in California, US (41). While their ancestries were not reported, the obesity cohort was only 85% white, and the geographical proximity between California and Mexico is consistent with the possibility of Mexican ancestry for these two individuals (41). The authors demonstrated that the p.Ile269Asn mutation decreased MC4R function in vitro (41). In 2012, the p.Ile269Asn mutation was also reported in five Pima Indian heterozygous carriers who had an average BMI of 31.5 ± 4.0 kg/m2, above the 30 kg/m2 obesity cutoff (42). The frequency of the MC4R p.Ile269Asn mutation was 1 order of magnitude lower in Pima Indians (MAF: 0.062%) than in the Mexican population (0.52–0.59%), consistent with a migration gene flow between the 2 geographically close populations (Pima Indians reside in Mexico and the United States) (42, 43). In vitro functional experiments in this third study confirmed the deleterious consequences of the p.Ile269Asn mutation on MC4R function (42). The MC4R p.Ile269Asn mutation was recently described in a 6-year-old Hispanic boy with severe obesity (44). The frequency of the MC4R p.Ile269Asn mutation in the studied group of 109 Hispanic children with severe obesity (MAF: 0.46%) is substantially lower than that of our sample of Mexican children with obesity (MAF: 1.59%), likely due to the fact that only a subset of the Hispanic children living in the United States had Mexican ancestry (44). Lastly, a study identified an exome-wide significant association between the MC4R p.Ile269Asn mutation and T2D in Hispanic/Latino individuals (MAF: 0.89%, OR = 2.17 [95%CI 1.63–2.89]) (45). Conditioning on BMI attenuated the association, suggesting that the effect of the mutation on T2D may be primarily driven by an increase in body weight (45). Collectively, these previous reports from the literature strengthen our conclusions that the MC4R p.Ile269Asn mutation may have resulted from a founder event in the native Mexican population, impair MC4R function, and confer high risk of childhood and adult obesity.
To the best of our knowledge, the MC4R p.Ile269Asn mutation is the most important genetic contributor to obesity in the Mexican population to date. At least 1 in every 100 Mexican individuals carries 1 or 2 copies of the p.Ile269Asn mutation and has a high chance of developing obesity during their lifetime. The mutation may account for 0.68 and 1.28% of obesity cases in Mexican adults and children, respectively. In comparison, only 1 in 1,100 individuals carries a deleterious mutation in MC4R in European populations (46). While the clinical utility of obesity-associated polygenic variants is still debated (47), early screening for Mendelian forms of obesity holds considerable promise for personalized care of affected high-risk patients (48). As an illustration, children with obesity who carry MC4R mutations lose weight when subjected to lifestyle interventions but have much greater difficulty maintaining lost weight than noncarriers, highlighting the need for constant monitoring in the context of MC4R genetic deficiencies (49). Adults with obesity who are heterozygous for deleterious mutations in MC4R lose more weight (3.48 kg) than obese noncarriers (3.07 kg) after 1 month of treatment with the MC4R agonist setmelanotide (50). By contrast, adults with severe obesity who carry MC4R mutations have a higher risk of major complications and reoperation rates than noncarriers following bariatric surgery (51).
Our results are consistent with an autosomal additive mode of inheritance of obesity for the MC4R p.Ile269Asn mutation in the Mexican population and are in line with observations made in European populations (15, 50). While heterozygous carriers of the p.Ile269Asn mutation exhibit a partially penetrant/oligogenic form of obesity, homozygous carriers develop a fully penetrant/monogenic form of obesity. The difference in penetrance of obesity observed between children and adults carrying 1 copy of the p.Ile269Asn mutation (27.5% vs 67.9%) can be explained largely by the increase in the prevalence of obesity over the life course. Consistent with this view, no significant difference in the genetic effect of the mutation on obesity was observed between children and adults.
Strengths of this study include the originality of the research question, the hypothesis-driven nature of the work, the novelty of the results, use of sophisticated experimental (ie, case-control, case-only, pedigree) designs and methods (ie, estimation of odd ratios, penetrance, PAR), adequate statistical power, assessment of children and adults in an underinvestigated population at high risk of obesity, and the concordance between our data and previous literature. Our study also presents several limitations. We focused our work on the p.Ile269Asn founder mutation and did not study rarer, potentially deleterious mutations in the MC4R gene. We also did not investigate the contribution of other well-established monogenic obesity genes in the Mexican population (48). Furthermore, we did not study other relevant phenotypes. Investigating the association of the MC4R p.Ile269Asn mutation with additional clinical features related to MC4R deficiency in humans (eg, food intake, height, blood pressure, bone mineral density and content) in the general Mexican population would nicely complement our study (52, 53). Resequencing the MC4R coding region in indigenous Mexican populations is also critical to accurately characterize the origin of the MC4R p.Ile269Asn mutation. The p.Ile269Asn mutation is not in substantial (r2 > 0.1) linkage disequilibrium with another coding variant in the Mexican population in 1000G (N = 64; data not shown). However, resequencing the MC4R region in a larger number of Mexican individuals is essential to exclude the possibility that the p.Ile269Asn mutation tags another causal variant. We acknowledge that additional in vitro (eg, investigation of the basal activity; β-arrestin recruitment and response to β-melanocyte-stimulating hormone, γ 2-melanocyte-stimulating hormone, and adrenocorticotropin agonists; and hAGRP antagonist of the mutant MC4R) and in vivo (creation of a “humanized” p.Ile269Asn mouse model) experiments are needed to further delineate the functional consequences of the p.Ile269Asn mutation on MC4R activity and its role in obesity (54–56). While we cannot totally exclude the possibility of having individuals with some degree of relatedness in our case control studies, we have previously used genome-wide SNP genotyping data to demonstrate that cryptic relatedness is exceptionally rare in the Mexican population (57).
In conclusion, our data suggest that the MC4R p.Ile269Asn mutation may have resulted from a founder event in native Mexicans and may confer high risk of childhood and adult obesity in the Mexican population. Our study adds to the growing body of evidence that founder mutations in Mendelian genes account for a nonnegligible fraction of cases of obesity in specific populations (58) and paves the way for large-scale exome resequencing projects in diverse ethnic groups to close the missing heritability gap for obesity.
Acknowledgments
We are indebted to all participants of this study. The authors would like to thank the Exome Aggregation Consortium (ExAC), the Genome Aggregation Database (gnomAD), and the groups that provided exome and genome variant data to these resources (16, 17).
Financial Support: This work was supported by grants from the Instituto Mexicano del Seguro Social (IMSS) under the program of Priority Health Topics 2017 (Grant No. FIS/IMSS/PROT/PRIO/17/062). MVM (Ciencias Médicas Odontológicas y de la Salud PhD program from Universidad Nacional Autónoma de México) and DLM (Ciencias Biomédicas PhD program from Universidad Autónoma de Guerrero) were supported by PhD fellowships from the Consejo Nacional de Ciencia y Tecnología (CONACYT) and IMSS (Mexico). DM is supported by a Canada Research Chair in Genetics of Obesity.
Author Contributions: MVM, MC, and DM designed the experiment. MVM, JPR, DLM, MRF, GSB, SMM, AVS, NWR, MC, and members of the National Obesity Network Mexico contributed to the recruitment of participants and the clinical aspects of the study. MVM and DLM performed laboratory experiments. MVM, HZ, HA, AM, VT, AMB, and DM prepared the dataset and conducted analyses. MVM, AMB, and DM wrote the manuscript and prepared all tables and figures. JPR, DLM, HZ, HA, AM, VT, MRF, GSB, SMM, AVS, NWR, and MC critically reviewed the manuscript. MC and DM had primary responsibility for final content.
Additional Information
Disclosure Summary: No potential conflicts of interest relevant to this article were reported.
Data Availability: The data set generated and/or analyzed in the current study is available from the corresponding authors on reasonable request.
Author’s Appendix
National Obesity Network Mexico
Baja California Sur State: Andrea S. Álvarez-Villaseñor; Kelly G. Acosta; Raquel Flores-Torrecillas; Uriel Flores-Osuna; Mariell G. García-Avilés.
Campeche State: Roxana del S. González-Dzib.
Chihuahua State: René A. Gameros-Gardea.
Mexico State: María L. Pizano-Zárate; Jorge A. Núñez-Hernández; Verónica de León-Camacho;
Mexico City: Roberto Karam-Araujo; Perla Corona-Salazar; Fernando Suarez-Sánchez; Jaime Gómez-Zamudio; Eugenia Flores-Alfaro.
Guanajuato State: Arturo Reyes-Hernández; Catalina Peralta-Cortázar; Emmanuel G. Martínez-Moralesvalla; Luz V. Díaz de León Morales; Irma L. del C. González-González; Arturo M. Reyes-Sosa; Sonia Lazcano-Bautista.
Hidalgo State: María G. Arteaga-Alcaraz; Nandy García-Silva; Moisés Herrera-Lemus; Gress M. Gómez-Arteaga.
Michoacán State: Anel Gómez-García; Martha V. Urbina-Treviño; Diana C. Villalpando-Sánchez; Cleto Álvarez-Aguilar.
Nayarit: Ramón E. Jiménez-Arredondo.
Monterrey State: Martha I. Dávila-Rodríguez; Francisco González-Salazar; Laura H. de la Garza-Salinas.
Oaxaca State: Aleyda Pérez-Herrera.
Puebla State: Jorge Martínez-Torres; Elizabeth Méndez-Fernández; Víctor A. Segura-Bonilla; Mariana Gutiérrez-Hernández.
Querétaro State: Lilia S. Gallardo-Vidal; Leticia Blanco-Castillo; José J. García-González.
Sinaloa State: Julio M. Medina-Serrano; Adrián Canizalez-Román.
Sonora State: Cruz M. López-Morales; Jaime G. Valle-Leal.
Tamaulipas State: Martin Segura-Chico; Rafael Violenté-Ortiz; Verónica Fernández-Jiménez; Norma A. Sánchez-Hernández.
References
