Dominant and recessive inheritance of morbid obesity associated with melanocortin 4 receptor deficiency (original) (raw)
Identification of MC4R mutations. In the cohort of 243 unrelated subjects with severe early-onset obesity, no subject had the CTCT deletion at codon 211 that had been described previously by us in a British family. The sequence variants detected in this study and their location in the MC4R protein structure are described in Table 1 and Figure 1, respectively. A novel frameshift mutation consisting of a di-nucleotide (GT) insertion at codon 279 was found in two unrelated probands. Seven different point mutations that result in a change of amino acid sequence were found. T112M, R165Q, V253I, and C271Y were each found as heterozygous changes in single probands. N62S was found as a homozygous mutation in one proband. Other than T112M, which was reported previously in both a lean (11) and an obese subject (12), all the above mutations are novel. None of the above variants were found in 108 alleles from nonobese U.K. Caucasian controls. Two previously described sequence variants were found: V103I in three and I251L in seven subjects, respectively. These two variants were found at allelic frequencies of 0.9% and 2.7%, respectively, in our Caucasian control population. We also identified an obese proband who was heterozygous for a deletion of a C nucleotide in the 3′ untranslated region, 28-base pairs downstream of the stop codon.
Sequence variants detected in MC4R. The positions of the sequence variants described in Table 1 are indicated on a model of the MC4R protein structure. Amino acids are indicated as circles in single-letter code. Amino acids affected by mutations are shaded, and specific mutations are then indicated in a shaded oval. Shaded ovals outlined in black indicate mutations where phenotypic studies on patients have been performed. Codon number 1 refers to the initiator methionine.
Sequence variants of MC4R detected in 243 subjects with severe early-onset obesity
Family members were available for study in the case of three of the above probands (Figure 2). Also included in Figure 2 is the family tree of the subject with the CTCT deletion reported by us previously (9) because phenotypic data on this family will be described further. In the case of the GT insertion, this was transmitted to the proband by her affected father, who also had severe early-onset obesity, whereas the lean mother was wild-type at this locus. The C271Y mutation also appears to segregate with obesity in a dominant fashion. N62S was found in homozygous form in five children with severe obesity from a highly consanguineous extended family of Pakistani origin. In this pedigree, none of the four heterozygote parents had a body mass index of 30 kg/m2 or more. Notably, this is the first description of homozygous MC4R mutations in human subjects and of a recessive pattern of inheritance for mutations in this gene. Finally, the family of one subject with the common V103I variant was studied, and there was evidence against cosegregation of this variant with obesity, which is consistent with the reported normal signaling properties of this mutant in vitro (11, 19).
Inheritance of mutations in MC4R. Cosegregation of mutational status with BMI is shown for the five families in the mutational screening study for whom family data were available. Also shown is the family tree for the family with the CTCT deletion reported previously by us (9). Filled symbols represent subjects with severe obesity. The first line of the symbols corresponds to the identification number. The second line shows the genotype: N, normal allele; M, mutant allele. The third and fourth lines show current age of subject (years) and BMI standard deviation score, respectively.
Phenotypic studies. Affected and unaffected members from three families in whom mutations in MC4R cosegregate with obesity (CTCTΔ, GTins, and N62S) were subjected to a more-detailed evaluation to better define the clinical, metabolic, and endocrine phenotype of human MC4R deficiency. By definition, all probands were obese from an early age. For the purposes of further discussion, the term “affected” refers to heterozygotes in the case of the frameshift mutations and homozygotes in the case of the N62S mutation. In total, four affected adults (two heterozygous frameshift and two homozygous N62S) and five affected children (two heterozygous frameshift and three homozygous N62S) were studied. The nonobese subjects from these families that were studied comprised two wild-type and four heterozygous N62S adults and two wild-type children.
Table 2 describes the clinical features and results of anthropometry, body composition, basal energy expenditure, and energy intake measurements of all subjects in these three families. In all cases the parents reported affected children to be excessively hungry from 6–8 months of age, with persistent food-seeking behavior and distress if food was not provided. Apart from obesity and hyperphagia, the affected children had no other developmental, intellectual, or behavioral problems. On clinical examination, apart from obesity, there were no dysmorphic features. No subject had acanthosis nigricans. Blood pressure was elevated in two of four affected adults and systolic blood pressure was greater than the 95th percentile for age, gender, and height in two of five affected children. Of note, however, two of four nonobese heterozygous adults were markedly hypertensive. Body temperature was normal in all subjects. Given the increased linear growth of _Mc4r_-KO animals, it is of note that the affected children were on average on the 91st percentile for height. In the N62S family, in particular, the two homozygous wild-type children (subjects 11 and 15) are on the ninth and 25th percentiles for height, respectively, whereas the homozygous mutant children are on the 91st–99th percentiles for height. Growth charts available on several affected subjects confirmed an increased growth velocity in childhood (data not shown). Bone age was assessed by radiography of the wrist in all children in these pedigrees. Bone age exceeded chronological age in four of five affected children by a range of 1.0–4.9 years.
Clinical features, anthropometry, body composition, and energy balance in affected and unaffected members of families with the CTCTΔ, GT ins, and N62S MC4R mutations
Body composition was assessed by DXA scanning. Of particular note is the finding that total bone mineral density (BMD) in affected subjects was consistently greater than expected from the age-adjusted population range (Z score, mean ± SD: 2.32 ± 0.69). In contrast, the total body BMD measurements in nonobese subjects from the three families were similar to age- and gender-related reference values (Z score, mean ± SD: 0.72 ± 1.16). Total body bone mineral content (BMC) was also increased in all affected subjects compared with age- and gender-specific data (20).
BMR was measured in all subjects by indirect calorimetry. In the case of the Caucasian subjects with heterozygous frameshift mutations, there was no significant deviation of these measurements from those predicted by the Hayter and Henry equations relating BMR to body weight (13). In contrast, all adult subjects in the N62S family (unaffected and affected) had BMRs considerably lower than that predicted by the same equation (data not shown), which may reflect ethnic differences reported previously (21). Predicted BMRs for adults of Pakistani origin (subjects 7–10, 13, and 14) were obtained using equations derived from two different data sets based on males and females living in India (14, 15). Compared with these predictions, all four nonobese adults and two affected adults had measured BMRs only slightly below predicted levels. Measured BMRs for affected children were similar to values predicted on the basis of age, gender, height, and weight.
Given the limitations of methods used to assess energy intake in free-living humans, we undertook semiquantitative studies of ingestive behavior in all children in these families. A large test meal of known nutrient and energy content was provided after an overnight fast, and the quantity and macronutrient composition of the food ingested was recorded. All affected subjects had greatly increased energy intake when compared with wild-type siblings (mean for affected subjects 4.1 MJ and mean for unaffected subjects 1.9 MJ). When energy intake was expressed as kilojoules per kilogram of lean body mass (as an approximate adjustment for the body size of the subjects studied), the mean energy intake of affected subjects was observed to be more than twice that of unaffected subjects (124 kJ/kg and 50 kJ/kg, respectively; unpaired t test: P < 0.05). There was no evidence for macronutrient preference between affected subjects (65% carbohydrate, 15% protein, 20% fat) and wild-type siblings (72% carbohydrate,10% protein,18% fat).
Secondary sexual characteristics and age of onset of puberty were normal in all adult subjects studied. Consistent with these clinical findings, circulating levels of sex steroids and gonadotropins were appropriate for Tanner stage in all affected subjects (Table 3). Of note, the absence of clinical and biochemical signs of puberty in some subjects with markedly advanced bone age was surprising and suggests that the advance in bone age in these cases was being driven by something other than sex steroids.
Metabolic and endocrine measurements in affected and unaffected members of families with the CTCTΔ, GTins, and N62S MC4R mutations
Hyperinsulinemia, but not diabetes, has been described in the _Mc4r_-KO mouse. Congruent with this, none of the affected subjects in this study were hyperglycemic. Fasting insulin levels were elevated in all subjects when compared with similarly obese subjects matched for age (Table 3; I.S. Farooqi, unpublished observations), with four of the homozygous N62S children having markedly elevated fasting plasma insulin values (185-437 pmol/L). In none of these cases were proinsulin or split proinsulin levels excessively high, which contrasts with murine and human obesity syndromes associated with defects in carboxypeptidase or prohormone convertase 1 (4, 5). Plasma leptin levels appeared to reflect fat mass (Table 3).
The fact that all subjects with heterozygous frameshift mutations in MC4R had severe early-onset obesity whereas heterozygotes for the N62S mutation were not obese was striking. To explore the molecular basis for this phenotypic heterogeneity, vectors expressing wild-type, CTCTΔ, GTins**,** N62S, and C271Y were transiently transfected into HEK293 cells. The cells were then exposed to varying concentrations of αMSH, and the generation of cAMP was assayed indirectly through activation of a cotransfected cAMP-responsive reporter construct expressing luciferase (Figure 3). Cells expressing wild-type receptor showed a sigmoidal dose response to αMSH. In contrast, cells expressing the CTCTΔ**,** GTins, and C271Y receptors showed no response. Cells expressing the N62S receptor showed a response intermediate between the wild-type and nonsense mutant receptors.
Signaling properties of mutant MC4Rs, N62S, GTins, CTCTΔ, and C271Y. HEK293 cells were transiently transfected with either empty pcDNA3 vector or the same vector expressing wild-type, CTCTΔ, GTins, N62S, and C271Y-mutant MC4Rs. The response to increasing concentrations of ligand (αMSH) was assessed by cotransfection with a cAMP-responsive reporter plasmid as described previously (18) and by measuring luciferase activity in a luminometer (see Methods). Data obtained were normalized for transfection efficiency by cotransfection of an internal control plasmid, pRL-CMV, which constitutively expresses Renilla luciferase. Data is expressed as a fold induction of luciferase activity. Each point represents the mean (± SE) of at least three independent experiments performed in quadruplicate.