Vitamin D receptor genetic variation and risk of non-Hodgkin lymphoma

2339

Vitamin D has been shown to have antiproliferative, antiangiogenic, and prodifferentiation properties, such that increased vitamin D may protect against various cancers. Ultraviolet (UV) exposure to the skin results in increased production of vitamin D, and its effects are largely mediated through the vitamin D receptor (VDR). Recent results suggest that UV exposure is protective against non-Hodgkin lymphoma (NHL). We hypothesize that the protective effect of UV exposure is due to increased vitamin D production, and this effect may be mediated through genetic variation of VDR. To examine the role of genetic variation in VDR and NHL, we genotyped variants in germline DNA provided by a sample of non-Hispanic white participants from a population-based case-control study conducted in the San Francisco Bay Area, 1988-1995. Controls (n = 684) were frequency matched to cases (n = 308) on sex, county of residence, and age within 5 years. Participants had no history of chemotherapy within three months prior to venipuncture. SNP-specific and haplotype block analyses were done using unconditional logistic regression, adjusting for age and sex. Seven single nucleotide polymorphisms (SNPs) were selected for genotyping: rs4516035 (T/C), rs3890734 (G/A), rs4334089 (G/A), FokI (G/A), BsmI (G/A), ApaI (C/A), and TaqI (T/C). The SNPs rs4516035, rs3890734, and rs4334089 were selected as haplotype-tagging SNPs for a haplotype block immediately upstream of the coding exons of VDR. Intronic SNPs BsmI and ApaI, along with synonymous coding SNP TaqI, are highly correlated, and located at the 3’ end of the coding region of VDR. FokI, a nonsynonymous SNP located in exon 2, is not in linkage disequilibrium with either flanking haplotype block. In SNP-specific analyses, the adjusted relative risk (RR) estimate for the follicular lymphoma (FL) subtype (n = 112) among rs4516035 variant C allele carriers was 1.6 (95% Confidence Interval (CI) 1.0-2.6) compared with TT homozygotes. This association was stronger among men (RR = 2.2, 95% CI 1.0-4.5) than women (RR = 1.2, 95% CI 0.66-2.3). The global test for association among common haplotypes (≥0.05) formed by BsmI-ApaI-TaqI genotypes was marginally significant (p = 0.07). Specifically, the G-A-C haplotype was inversely associated with FL (RR = 0.39, 95% CI 0.15-0.95). Further, the association of this haplotype and FL differed significantly (p = 0.01) between men (RR = 1.2, 95% CI 0.62-2.7) and women (RR = 0.31, 95% CI 0.13-0.72). There were no consistent SNP-specific or haplotype associations with all NHL or the diffuse large-cell lymphoma subtype. These results suggest an association between genetic variants spanning VDR and the FL subtype of NHL, and this association may be modified by sex. We currently are estimating UV exposure for individuals based on occupational and residential history to evaluate whether there is environmental modification of the observed genetic associations.