Abstract 05: A pilot gene-environment-wide interaction study (GEWIS) of neurocognitive impairment in childhood medulloblastoma survivors.

Abstract

Abstract Introduction: Neurocognitive impairment after radiation therapy is a significant complication in cancer patients, affecting every age group; however pediatric populations are the most vulnerable. Medulloblastoma is the most common central nervous system malignancy in children, and craniospinal radiation is an essential part of standard treatment. While 5-year survival is approximately 70% due to advances in treatment, at least half of the survivors develop progressive and permanent neurocognitive impairment. To date, few risk factors for these adverse outcomes have been identified. Neurocognitive impairment is believed to be a complex process affected by genes, environmental factors, and their interactions. The genome-wide association study (GWAS) is an important method for identifying main genetic effects, that is, associations between single nucleotide polymorphisms (SNPs) and prognosis. However, scanning for main effects might miss important genetic variants among those with unique environmental exposures. Therefore, we employed a 2-step approach utilizing GWAS data for detecting loci involved in gene-environment interactions to determine predictors of neurocognitive impairment. This approach (the gene-environment-wide interaction study or GEWIS) is a powerful alternative to single-step interaction tests. Methods: Childhood medulloblastoma survivors were recruited from the long-term survivor clinics at Texas Children's Cancer Center and MD Anderson Cancer Center between 2003 and 2010. Working memory was assessed using age-appropriate intelligence scales for children, which included scores for verbal IQ (VIQ), performance IQ (PIQ), and full scale IQ (FSIQ). We defined neurocognitive impairment as a score of <77.5 on any of the three IQ subscales. Demographic and clinical data were abstracted from medical records. Exposure to ambient levels of benzene (a suspected environmental neurotoxicant) was estimated using the U.S. Environmental Protection Agency's Assessment System for Population Exposure Nationwide (ASPEN) model. Genotyping was conducted using the Illumina Human 1M Quad SNP Chip. Exposure and genotype data were available on 47 subjects (20 with neurocognitive impairment and 27 without neurocognitive impairment). In step 1, we examined the association between each SNP (G) and benzene exposure (E) using a Wald test based on a logistic regression model. SNPs where p <0.001 were then included in step 2, a case-control analysis of GxE interactions using logistic regression models with the constrained maximum-likelihood method. Analyses were conducted using the ‘CGEN’ package in R. As this was a pilot study, the false discovery rate (FDR) was controlled at 0.25. Results: Benzene exposure was positively associated with neurocognitive impairment (OR = 2.0, 95% CI: 0.39-10.16); however, the association was not statistically significant. In step 1 of the preliminary analysis, seven SNPs were associated with neurocognitive impairment (p <0.001). Among this set, two GxE interactions were significantly associated with neurocognitive impairment (rs1223760-benzene, p = 0.014, q = 0.085 and rs11932152-benzene, p = 0.060, q = 0.182). Interestingly, rs1223760 is an intergenic SNP between PROX1 and SMYD2, two genes suspected of influencing DNA methylation. Conclusions: To our knowledge, this is the first GEWIS of neurocognitive impairment in survivors of childhood medulloblastoma. As the genetic architecture of neurocognitive impairment is complex, it is important to explore interactions using high-dimensional data. Our results are suggestive that GxE interactions may be important in prognosis; however, the results are preliminary and must be validated in a larger and independent population. Citation Format: Philip J. Lupo, M Fatih Okcu, Renke Zhou, Lynnette L. Harris, Bartlett D. Moore, III, Michael E. Scheurer. A pilot gene-environment-wide interaction study (GEWIS) of neurocognitive impairment in childhood medulloblastoma survivors. [abstract]. In: Proceedings of the AACR Special Conference on Post-GWAS Horizons in Molecular Epidemiology: Digging Deeper into the Environment; 2012 Nov 11-14; Hollywood, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2012;21(11 Suppl):Abstract nr 05.