Genome-wide copy number variation association study in anorexia nervosa

Abstract

This study represents the first large-scale investigation of rare (<1% population frequency) copy number variants (CNVs) in anorexia nervosa (AN). Large, rare CNVs are reported to be causally associated with anthropometric traits, neurodevelopmental disorders, and schizophrenia, yet their role in the genetic basis of AN is unclear. Using genome-wide association study (GWAS) array data from the Anorexia Nervosa Genetics Initiative (ANGI), which included 7414 AN case and 5044 controls, we investigated the association of 67 well-established syndromic CNVs and 178 pleiotropic disease-risk dosage-sensitive CNVs with AN. To identify novel CNV regions (CNVRs) that increase the risk of AN, we conducted genome-wide association studies with a focus on rare CNV-breakpoints (CNV-GWAS). We found no net enrichment of rare CNVs, either deletions or duplications, in AN, and none of the well-established syndromic or pleiotropic CNVs had a significant association with AN status. However, the CNV-GWAS found 21 nominally associated CNVRs that contribute to AN risk, covering protein-coding genes implicated in synaptic function, metabolic/mitochondrial factors, and lipid characteristics, like the CD36 (7q21.11) gene, which transports long-chain fatty acids into cells. CNVRs intersecting genes previously related to neurodevelopmental traits include deletions of NRXN1 intron 5 (2p16.3), IMMP2L (7q31.1), and PTPRD (9p23). Overall, given that our study is well powered to detect the CNV burden level reported for schizophrenia, we can conclude that rare CNVs have a limited role in the etiology of AN, as reported for bipolar disorder. Our nominal associations for the 21 discovered CNVRs are consistent with AN being a metabo-psychiatric trait, as demonstrated by the common genetic architecture of AN, and we provide association results to allow for replication in future research.

Fig. 1. Total genome-wide rare CNV burden.

A We tested whether genome-wide rCNV burden was greater in AN cases than in controls. A Partitioning the genome-wide rCNV burden by frequency (based on 50% reciprocal overlap with the full rCNV call set), no enrichment of singletons or rCNVs with counts up to 125 (1% frequency) was observed. B Partitioning total rCNV burden across CNV types and CNV lengths, no significant rCNV burden enrichment was observed across any length range, except for a slight excess of large (>500 kb) duplications (OR: 1.40, CI: 0.99–1.32, one-sided P: 0.037).

Fig. 2. Rare CNV breakpoint association results for the ~40 kb NRXN1 intron 5 deletion.

The two vertical gray dashed lines within all panels illustrate the critical region of the AN-risk CNVR. The top panel illustrates SAIGE rCNVb associations from the deletion-only model within the defined ~40 kb critical CNVR boundary and a buffer region of ±40 kb. 4.6% of the base pairs in this CNVR are annotated as highly constrained, an annotation given to 3.54% of the genome. Gene tracks panel shows the intersecting protein-coding gene tracks extracted from the UCSC genome browser (v43). The bottom two panels illustrate the tracks for CNVs found within AN cases and controls intersecting the CNVR by at least 1 base-pair, with green CNVs indicating individuals from ANGI-ANZUS, and blue CNVs indicating individuals from ANGI-SWE. The number of CNVs per cohort contributes to the differences in the test statistic and -log10 p-value across the x-axis.