Session: 634. Myeloproliferative Syndromes: Clinical: Poster III
Hematology Disease Topics & Pathways:
Diseases, Pediatric, Study Population, Myeloid Malignancies
Methods: LCH cases (n=117) for this discovery GWAS were recruited from Texas Children’s Hospital, of which 52 patients experienced a relapse event and 65 patients did not. Genotyping was performed on the Illumina Omni5 Quad BeadChip. We tested the association between common variants (minor allele frequency >5%) and LCH relapse risk in PLINK. A genome-wide threshold of significance was applied at P-value<5.0x10-8, and threshold of suggestive significance at P-value<1.0x10-5. Principal component analysis was performed to account for genetic ancestry, with the top two principal components (PCs) accounting for 85% of variability. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) controlling for age at diagnosis, sex, and the top two PCs were estimated using logistic regression. An independent replication set of LCH cases (n=124) were recruited from Texas Children’s Hospital to validate discovery GWAS findings.
Results: High-quality genotype data for 583,173 germline variants were tested for an association with LCH relapse risk in the discovery GWAS stage. We identified a variant in high linkage disequilibrium with a cluster of loci on Chromosome 9 that surpassed our threshold of suggestive significance (non-coding RNA LOC100506532 rs2182640; P-value=6.98x10-6). This intronic variant was associated with a decreased risk of LCH relapse that remained statistically significant after adjusting for age at diagnosis, sex, and the top two PCs (aOR: 0.16; 95% CI: 0.07-0.35). While this non-coding RNA gene is largely uncharacterized, non-coding RNAs function to regulate gene expression at the transcriptional and post-transcriptional level. Results from validation of this risk locus in an independent replication set are forthcoming.
Conclusions: In this genome-wide assessment of inherited genetic variation and LCH relapse, we identified a genomic region that may be associated with LCH relapse. It is unclear which variant in the LOC100506532 cluster is a potentially causal allele. One of these variants may be a proxy for the causal locus, or may act through an effect on other genes in the same region or at distal sites. Notably RXRA, a retinoic acid receptor gene, is located in close proximity (~400kb) to LOC100506532 and regulates gene expression in various biologic processes. Risk variants within this gene have been identified to modulate disease-specific survival in melanoma, another BRAF-driven malignancy. While we are in the process of validating the association between LOC100506532 rs2182640 and LCH relapse risk in an independent replication set, our discovery GWAS results provide initial evidence to suggest that inherited risk factors influence LCH disease severity.
Disclosures: Allen: Novimmune: Membership on an entity's Board of Directors or advisory committees.
See more of: Oral and Poster Abstracts