EPID-04. LEVERAGING GENOMIC DATA TO IDENTIFY RISK FACTORS FOR CHILDHOOD EPENDYMOMA. (5th November 2018)

• Record Type: Journal Article
• Authors: Zhang, Chenan; Hansen, Helen; Gonzalez-Maya, Julio; Smirnov, Ivan; Wiemels, Joseph; Walsh, Kyle
• Is Part Of: Neuro-oncology. Volume 20(2018)Supplement 6
• Page Start: vi80
• Page End: vi80
• Publication Date: 2018-11-05
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noy148.331
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: SNPs identified by previous GWAS can serve as instrumental variables for exposures of interest using a process known as Mendelian randomization. We sought to identify risk factors for childhood ependymoma using a Mendelian randomization approach. We identified 390 Californian children born after 1982 and diagnosed with intracranial ependymoma before age 19, and retrieved their archived newborn bloodspots from the state. DNA from 360 cases and 2176 controls was successfully extracted and genotyped on the Affymetrix Axiom array which, coupled with whole-genome imputation, yielded 4.5M SNP genotypes per subject. We constructed polygenic scores of anthropometric traits, telomere length, and immune cell profiles using SNPs discovered in prior GWAS. Case-control comparisons were adjusted for ancestry-informative principal components. A polygenic score for adult height (416 unlinked SNPs) indicated that a 2.5cm increase in predicted height was associated with a 1.23-fold increased risk of ependymoma (P=0.041). No associations were observed for birth length, pre-pubertal height, or head circumference. A polygenic score for inter-individual variation in telomere length (8 unlinked SNPs) indicated that longer predicted telomere length increases risk of ependymoma (P=0.037), particularly adolescent-onset ependymoma (P=7.1x10 -3 ). A polygenic score for lymphocyte count (156 unlinked SNPs) was not associated with ependymoma risk (P=0.14). However, a one standard deviation increase in the score for myeloid cell count (191 unlinked SNPs) was associated with a 0.79-fold decreased risk of ependymoma (P=9.0x10 -3 ). Within the granulocyte lineage, increased neutrophil count was associated with decreased risk of ependymoma (OR=0.82, P=0.027), while neither eosinophil count nor basophil count showed evidence of association (P=0.22 and 0.99, respectively). Because ependymoma therapy can affect height attainment, telomere length, and immune parameters, these exposures are ill-suited for traditional case-control study designs. Using a Mendelian randomization approach, we observe that genetic predisposition to taller height, longer telomere length, and decreased circulating myeloid cell counts may increase risk of ependymoma.