WHOLE GENOME SEQUENCING OF CYCLIC VOMITING PATIENTS REVEALS NOVEL RISK LOCI

Introduction: Cyclic vomiting syndrome (CVS) is a chronic condition characterized by intermittent episodes of recurrent vomiting lasting less than one week that affects approximately 2% of American adults and children. The pathogenic mechanisms by which CVS develops are not completely understood. However, current evidence suggests that the development of the syndrome is likely due to environmental and individual characteristics that lead to autonomic dysfunction and neuronal hyperexcitability. Several studies have indicated that deleterious nuclear and mitochondrial genetic variation plays an important role in the development of CVS. However, no studies have been performed to date investigating the association of rare protein-coding variation with CVS on a genome-wide scale. To address this, we performed genome-wide genetic burden analysis, which identifies genes that are enriched for nonsynonymous genetic variants in cases compared to controls, to identify novel CVS-associated genes.

Methods: We performed whole-genome sequencing for 50 CVS patients cared for at Massachusetts General Hospital. As most patients were of European ancestry, we limited the analysis to rare (Genome Aggregation Database (gnomAD) non-Finnish European (NFE) minor allele frequency < 0.1%) nonsynonymous single nucleotide polymorphisms. We performed a genetic burden analysis using the Test Rare vAriants with Public Data (TRAPD)method in 50 CVS cases and 34,029 NFE gnomAD control individuals.

Results: After Bonferroni multiple test correction, we found six significant genes under a dominant inheritance model (AHNAK2, BLTP1, DYNC2I1, FHIP1A, ABTB3, ARGLU1) and two significant genes under a recessive inheritance model (AHNAK2 and DUXB). None of these genes have been previously reported to be associated with CVS. We examined 35 genes that have been hypothesized to be either directly or indirectly involved in pathogenesis and found no evidence of significant (p < 0.05) genetic burden in any of these genes.

Conclusions: Genetic variants of FHIP1A (dyslexia and intelligence), DYNC2I1 (short-rib polydactyly syndrome), ARGLU1 (amyotrophic lateral sclerosis), BLTP1 (Alkuraya-Kucinskas syndrome), and AHNAK2 (Charcot-Marie Tooth) have been associated with neurological phenotypes. Interestingly, ARGLU1 is a glucocorticoid receptor co-activator, a hormonal axis dysregulated in CVS, in neuronal cells that leads to abnormal gene splicing when genetically disrupted. Together, our results suggest that there are nuclear germline variants that may contribute toward CVS pathogenesis. Further studies will be required to confirm these findings in a larger CVS cohort and characterize the functional role of the associated genes in CVS.