This study confirms that mutations in MC4R
are a significant cause of severe human obesity and extends this finding to severely obese North American adults. Indeed, we find the same combined prevalence of such mutations [2.25% (CI95%
: 1.44–3.47%)] in our severely obese adult patient populations as that previously described in French adults [2.35% (CI95%
: 0.90–3.80%)] with severe obesity (BMI > 35 kg/m2
) and in young adult Danish males (BMI ≥ 31 kg/m2
). This prevalence also matches the frequency of MC4R
mutations detected in cohorts of patients with childhood obesity of different origins (1–6%) (5
). Rare MC4R
mutations are also found in non-obese controls albeit with a significantly lower frequency. The prevalence of such mutations in our control population (0.64%) is strikingly similar to that described for a very large population-based sample of adults from Germany [0.66% (CI95%
: 0.44–0.96%)] (35
) and in most non-obese control populations. Although the significant difference in the prevalence of rare MC4R
mutations between cases and controls supports a causative role for such mutations in the severe obesity of these patients, this observation underlines the importance of systematically evaluating the functional consequences of such mutations. Indeed, restricting the association analysis to only mutations with a significant effect on the function of MC4R significantly increases the odds ratio of association of rare MC4R mutations with severe obesity.
When compared with MC4R, the importance and role of MC3R mutations in the pathogenesis of obesity in rodents and humans has been less clear. When compared with mc4r
−/− mice, mc3r
−/− mice have a milder phenotype, limited to an increase in body fat, and unlike in heterozygous mc4r +/−
mice, no energy homeostasis phenotype has been observed in heterozygous mc3r +/−
To date only two common variants (28
) and four rare mutations in heterozygous carriers (21
) have been described in the coding region of the MC3R
. The two common MC3R
variants Thr6Lys and Val81Ile, which were also detected in the current study, were not associated with any obesity-related phenotypes, therefore likely representing benign polymorphisms (28
With respect to rare mutations in MC3R
, one mutation, Ile183Asn, has been detected in one severely obese girl and her obese father, which was absent in control subjects and was characterized to be functionally inactive (21
). More recently, three MC3R mutations were identified in a study of 290 severely obese Italian adults, mean BMI of 44.2 ± 5.9 kg/m2
. These three mutations were absent in 215 non-obese controls (37
). In that study, however, controls were only screened for mutations found in cases, but the MC3R
gene was not systematically sequenced in these controls. Two of these mutations were studied in the two families and segregated with obesity in the family members (total of four relatives). Only one of the mutations, Ile335Ser, demonstrated in vitro
functional abnormalities (37
Our study, the largest yet to evaluate the prevalence of MC3R mutations in severe human obesity, does not support a significant role for mutations in this gene in this condition. The prevalence of such mutations in cases is 0.67% (CI95%: 0.27–1.50) and is the same in controls. Systematic in vitro study of these mutations also demonstrates that functionally significant mutations are as frequent in both cohorts. Interestingly, the prevalence of rare MC3R variants found in both groups is also the same as the prevalence of rare MC4R variants (whether considering all or only restricted to functionally significant) found in controls or in the general population, further suggesting that it corresponds to the ‘background’ frequency of rare mutations in these genes.
Our data do not eliminate a possible role for some of the individual MC3R mutations in the obesity of the carriers, in particular if some of these mutations have a pleiotropic or dominant negative effect. However, formal demonstration of the phenotypical effect of such specific mutations would require the study of very large multigenerational pedigrees to reach statistical significance. Such pedigrees are rarely available.
The recent completion of large genome scans has demonstrated that a certain number of common variants are associated with severe obesity but has also strongly suggested that, in aggregate, such common variants will only account for a small portion of the overall genetic predisposition to this condition (2
). This observation has led to the re-visiting of the hypothesis that rare variants could account for the majority of one individual's predisposition to severe obesity and has led to the suggestion that large-scale systematic sequencing of patients will be required to detect genes in which rare mutations predispose to the disease. Our results underline some of the possible limitations in the outcomes and interpretations of this approach. First, it is clear that the mere presence of rare heterozygous variants in cases is not sufficient to implicate a gene in the condition as such variant can be present at the same prevalence in controls, reflecting the background level of benign mutations in the gene. Second, differentiating functionally relevant mutations from functionally neutral mutations through specific in vitro
assays might be required to strengthen the association of rare mutations in a particular gene with a common phenotype. Finally, at an individual level, demonstration of the role of a specific mutation in a common condition might be limited.