4258

Background. Chromosomal translocations, deletions and amplification and point mutations affecting oncogenes are critical in development of non-Hodgkin lymphomas (NHL). During these events, genes break at exons. The biological mechanisms of chromosome deletions, insertions and translocations in the development of malignancy are largely unknown. Translocations may impart these genes with the capacity to support malignant growth and influence the pattern of gene expression. We hypothesize that germline variations in DNA repair genes responding to the double-and single-strand breaks induced by these processes may play a role in the development of NHL.

Methods. We used a clinic-based study of 475 NHL cases and 481 frequency-matched controls recruited from the Mayo Clinic from 2002-2005; for this report we restricted our study population to Caucasian participants only. We evaluated 16 single nucleotide polymorphisms (SNPs) in 3 genes related to three DNA repair pathways: base-excision repair (XRCC1), nucleotide-excision repair (ERCC2) and homologous recombination (RAD51). Tagged SNPs were selected from HapMap and were supplemented with candidate coding SNPs. Logistic regression was used to estimate odds ratios (ORs) for risk of NHL associated with each SNP, adjusting for age and gender. The most prevalent homozygous genotype was used as the reference group, and each polymorphism was modeled individually as having a log-additive effect in the regression model. We also jointly tested the main effects for all independent (r2<0.25) SNPs from a gene using a multivariate logistic regression (MLR) model and a likelihood ratio test.

Results. The mean age at diagnosis was 59.3 years for cases and 56% were male. The mean age at enrollment was 61.5 years for controls and 55% were male. Based on the global p-value from the MLR model, there was an association of ERCC2 (p=0.016), but not XRCC1 (p=0.36) or RAD51 (p=0.98), with risk of NHL. There were 7 SNPs in ERCC2, including one synonymous (rs1052555), one non-synonymous (rs13181), one flanking 5’-UTR (rs11878644) and 4 intronic SNPs. The flanking 5’-UTR SNP was common (MAF=0.42 in controls) and was inversely associated with NHL risk (OR=0.7; p=0.04). Furthermore, among the 7 SNPs in ERCC2, two intronic SNPs were also significantly associated with NHL (rs238415, OR=1.19, p=0.06; and rs50871, OR=0.64, p=0.003).

Conclusion. Genetic polymorphisms in the DNA repair gene ERCC2 was associated with risk of NHL. More comprehensive studies on the role of DNA repair genes, and the nucleotide-excision repair pathway in particular, may lead to better understanding of NHL etiology.

98th AACR Annual Meeting-- Apr 14-18, 2007; Los Angeles, CA