Novel genes involved in severe early-onset obesity revealed by rare copy number and sequence variants

Abstract

Obesity is a multifactorial disorder with high heritability (50-75%), which is probably higher in early-onset and severe cases. Although rare monogenic forms and several genes and regions of susceptibility, including copy number variants (CNVs), have been described, the genetic causes underlying the disease still remain largely unknown. We searched for rare CNVs (>100kb in size, altering genes and present in <1/2000 population controls) in 157 Spanish children with non-syndromic early-onset obesity (EOO: body mass index >3 standard deviations above the mean at <3 years of age) using SNP array molecular karyotypes. We then performed case control studies (480 EOO cases/480 non-obese controls) with the validated CNVs and rare sequence variants (RSVs) detected by targeted resequencing of selected CNV genes (n = 14), and also studied the inheritance patterns in available first-degree relatives. A higher burden of gain-type CNVs was detected in EOO cases versus controls (OR = 1.71, p-value = 0.0358). In addition to a gain of the NPY gene in a familial case with EOO and attention deficit hyperactivity disorder, likely pathogenic CNVs included gains of glutamate receptors (GRIK1, GRM7) and the X-linked gastrin-peptide receptor (GRPR), all inherited from obese parents. Putatively functional RSVs absent in controls were also identified in EOO cases at NPY, GRIK1 and GRPR. A patient with a heterozygous deletion disrupting two contiguous and related genes, SLCO4C1 and SLCO6A1, also had a missense RSV at SLCO4C1 on the other allele, suggestive of a recessive model. The genes identified showed a clear enrichment of shared co-expression partners with known genes strongly related to obesity, reinforcing their role in the pathophysiology of the disease. Our data reveal a higher burden of rare CNVs and RSVs in several related genes in patients with EOO compared to controls, and implicate NPY, GRPR, two glutamate receptors and SLCO4C1 in highly penetrant forms of familial obesity.

Fig 1. Workflow of the approach followed to study CNVs and RSVs in a subset of selected genes.

The different control cohorts used for each of the analysis are shown, in addition to the number of obese patients and controls studied in each step.

Fig 2. Detection, validation and inheritance of the duplication encompassing NPY in case Ob_12 and his family.

A: Ideogram showing the location of the CNV and the specific genomic interval included in the duplication. The plot shows the results of SNP array with the Log R Ratio represented by black dots and the B Allele Frequency (BAF) represented by red dots. Hg19 assembly. B: MLPA of the trio showing the maternal inheritance of the rearrangement represented by a single probe (indicated). C: Pedigree of the family showing several cases with severe obesity, as well as ADHD (attention-deficit/hyperactivity disorder). The two individuals carrying the duplication are labeled by an *; samples from additional relatives were not available.

Fig 3. Enrichment analyses of shared co-expressed partners between 14 known obesity-related genes and the candidate genes from this study.

A: Key players on the regulation of food intake and energy expenditure mostly act through the leptin-melanocortin pathway. B: Shared co-expressed partners between the four candidate genes (GRIK1, GRM7, GRPR, SLCO4C1) and 500 randomly generated gene sets. The number shared co-expressed partners of the candidate genes with the 14 obesity-related genes are indicated by a red arrow. C: Co-expression network including the selected obesity-related genes and GRIK1, GRM7, GRPR, SLCO4C1. This network was visualized using Cytoscape.