Contributions of Function-Altering Variants in Genes Implicated in Pubertal Timing and Body Mass for Self-Limited Delayed Puberty

Abstract

Context: Self-limited delayed puberty (DP) is often associated with a delay in physical maturation, but although highly heritable the causal genetic factors remain elusive. Genome-wide association studies of the timing of puberty have identified multiple loci for age at menarche in females and voice break in males, particularly in pathways controlling energy balance.

Objective/main outcome measures: We sought to assess the contribution of rare variants in such genes to the phenotype of familial DP.

Design/patients: We performed whole-exome sequencing in 67 pedigrees (125 individuals with DP and 35 unaffected controls) from our unique cohort of familial self-limited DP. Using a whole-exome sequencing filtering pipeline one candidate gene [fat mass and obesity-associated gene (FTO)] was identified. In silico, in vitro, and mouse model studies were performed to investigate the pathogenicity of FTO variants and timing of puberty in FTO+/- mice.

Results: We identified potentially pathogenic, rare variants in genes in linkage disequilibrium with genome-wide association studies of age at menarche loci in 283 genes. Of these, five genes were implicated in the control of body mass. After filtering for segregation with trait, one candidate, FTO, was retained. Two FTO variants, found in 14 affected individuals from three families, were also associated with leanness in these patients with DP. One variant (p.Leu44Val) demonstrated altered demethylation activity of the mutant protein in vitro. Fto+/- mice displayed a significantly delayed timing of pubertal onset (P < 0.05).

Conclusions: Mutations in genes implicated in body mass and timing of puberty in the general population may contribute to the pathogenesis of self-limited DP.

Figure 1.

Flowchart of WES filtering strategy to identify candidate genes. WES was initially performed on DNA extracted from peripheral blood leukocytes of 160 individuals from the 67 most extensive families from our cohort (125 with DP and 35 controls), with exome capture on a Nimblegen v2 or Agilent v5 platform and sequencing on the Illumina HiSeq 2000. The exome sequences were aligned to the University of California, Santa Cruz hg19 reference genome. Picard tools and the genome analysis toolkit were used to mark PCR duplicates, realign around indels, and recalibrate quality scores and call variants. Variants were then analyzed further and filtered for potential causal variants using filters for quality control, predicted functional annotation, minor allele frequency (MAF), segregation with trait, and GWAS relevance (see Materials and Methods for further information on filtering criteria). Targeted exome sequencing using a Fluidigm array of a candidate gene identified after filtering was then performed in a further 42 families from the same cohort (288 individuals, 178 with DP and 110 controls). Variants after targeted resequencing were filtered using the same criteria as the WES data. Functional annotation of the variants is as described elsewhere in Materials and Methods.

Figure 2.

Pedigrees and auxological data of the families with potentially pathogenic FTO variants. (a) Squares indicate male family members; circles indicate female family members. Black symbols represent clinically affected, gray represent unknown phenotype, and clear symbols represent unaffected individuals. The arrowhead with “P” indicates the proband in each family, and “us” indicates unsequenced due to lack of DNA from that individual. The mutation in each family is given next to the family number; a horizontal black line above an individual’s symbol indicates that they are heterozygous for the variant as confirmed by either WES or Fluidigm array, and verified by Sanger sequencing. A red dot indicates that the individual was underweight (thinness grade 2 or more significant) and ‘?’ indicates that BMI information for that individual is not available. (b–d) BMI and HSDS charts for the probands of each of the three pedigrees (family 1.III.2, family 2.III.5, and family 3.III.2). Underweight values are shown in red, green dots indicate a significant deflection from previous height measurements, and orange dots indicate significant deflection from target height. Normal values, based on data from >70,000 healthy Finnish children, have been previously published (22).

Figure 3.

Demethylation assay assessing kinetic activity of mutant vs WT FTO proteins. FTO activity is proportional to the concentration present in the reaction. Demethylase activity is likely to be related to the ability of FTO to function as a sensor for cellular metabolism (29). The R316Q mutant is enzymatically dead across all concentrations tested. The A163T and L44V mutants showed demethylase activity toward methylated uridine in a dose-dependent manner but with different affinities.

Figure 4.

Timing of VO in WT and FTO^+/−heterozygous mice. Cumulative percentages of mice displaying VO by postnatal day are shown for WT and FTO^+/− mice. WT mice, n = 24; FTO^+/− mice, n = 45. P < 0.05 by unpaired t test. het, heterozygous.

Figure 5.

Mean body weight (grams) for WT and Fto^+/− (heterozygous) mice in 7 days prior to VO. Mean body weight (grams) ± SEM: 11.64 ± 0.21 in WT mice (n = 24) vs 11.45 ± 0.14 in Fto^+/ mice (n = 45). P = 0.467 by unpaired t test. Error bars show SEM for each group each day. HET, heterozygous.