Variations in Chromosomes 9 and 6P21.3 with Risk of Non–Hodgkin Lymphoma

Published OnlineFirst December 10, 2010; DOI: 10.1158/1055-9965.EPI-10-0638 Cancer Epidemiology, Research Article Biomarkers & Prevention

Variations in Chromosomes 9 and 6p21.3 with Risk of Non–Hodgkin Lymphoma

Sophia S. Wang1, Idan Menashe2, James R. Cerhan3, Wendy Cozen4, Richard K. Severson5, Scott Davis6, Amy Hutchinson7, Nathaniel Rothman2, Stephen J. Chanock2,7, Leslie Bernstein1, Patricia Hartge2, and Lindsay M. Morton2

Abstract Background: There is growing evidence linking genetic variations to non–Hodgkin lymphoma (NHL) etiology. To complement ongoing agnostic approaches for identifying susceptibility genes, we evaluated 488 candidate gene regions and their relation to risk for NHL and NHL subtypes. Methods: We genotyped 6,679 tag single nucleotide polymorphisms (SNPs) in 947 cases and 826 population-based controls from a multicenter U.S. case–control study. Gene-level summary of associations were obtained by computing the minimum P value ("minP test") on the basis of 10,000 permutations. We used logistic regression to evaluate the association between genotypes and haplotypes with NHL. For NHL subtypes, we conducted polytomous multivariate unconditional logistic regression (adjusted for sex, race, age). We calculated P-trends under the codominant model for each SNP. Results: Fourteen gene regions were associated with NHL (P < 0.01). The most significant SNP associated with NHL maps to the SYK gene (rs2991216, P-trend ¼ 0.00005). The three most significant gene regions were on chromosome 6p21.3 (RING1/RXRB; AIF1; BAT4). Accordingly, SNPs in RING1/RXRB (rs2855429), AIF1 (rs2857597), and BAT4 (rs3115667) were associated with NHL (P-trends 0.0002) and both diffuse large B-cell and follicular lymphomas (P-trends < 0.05). Conclusions: Our results suggest potential importance for SYK on chromosome 9 with NHL etiology. Our results further implicate 6p21.3 gene variants, supporting the need for full characterization of this chromosomal region in relation to lymphomagenesis. Impact: Gene variants on chromosome 9 may represent a new region of interesting for NHL etiology. The independence of the reported variants in 6p21.3 from implicated variants (TNF/HLA) supports the need to confirm causal variants in this region Cancer Epidemiol Biomarkers Prev; 20(1); 42–9. 2011 AACR.

Introduction wide association studies (GWAS) have implicated a number of immune-related genes with NHL risk. Specifically, There is growing evidence that common genetic var- polymorphisms in interleukin 10 (IL10) and within sev- iants play an important role in non–Hodgkin lymphoma eral genes located in the 6p21.3 chromosomal region are (NHL) etiology. Large consortial efforts and genome- associated with NHL risk, including the tumor necrosis factor (TNF) and human leukocyte antigen (HLA) genes (1–3). In our multicenter study of NHL in the United HLA Authors' Affiliations: 1Division of Cancer Etiology, Department of Popu- States, we have further implicated specific Class I lation Sciences, Beckman Research Institute and City of Hope, Duarte, and II alleles with NHL risk, most notably the HLA California; 2Division of Cancer Epidemiology and Genetics, National Can- DRB1*0101 allele with follicular lymphoma, a finding cer Institute, NIH, DHHS, Rockville, Maryland; 3Division of Epidemiology, College of Medicine, Mayo Clinic, Rochester, Minnesota; 4Norris Com- that is consistent with recently published GWAS data prehensive Cancer Center, University of Southern California, Los Angeles, (3;4). We have also implicated genes in other pathways 5 California; Department of Family Medicine and Karmanos Cancer Insti- including the nuclear factor kappa beta (NF-kB) pathway tute, Wayne State University, Detroit, Michigan; 6Fred Hutchinson Cancer Research Center and University of Washington, Seattle, Washington; and (5), cell cycle (6), DNA repair (7), and the family of 7Core Genotyping Facility, SAIC-Frederick, Inc., National Cancer Institute, caspase genes (8). NIH, DHHS, Gaithersburg, Maryland To complement previous a priori and ongoing agnostic Note: Supplementary data for this article are available at Cancer Epide- approaches to identifying susceptibility genes in NHL miology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/) etiology, we conducted a detailed literature search to Corresponding Author: Sophia S. Wang, Division of Cancer Etiology, Department of Population Sciences, Beckman Research Institute and City identify additional potential genes of relevance for lym- of Hope, 500 Duarte Road, Duarte, CA 9101. Phone: 626-471-7316; phomagenesis. We included genes implicated from Fax: 626-471-7308. E-mail: [email protected] laboratory and clinical research, particularly those based doi: 10.1158/1055-9965.EPI-10-0638 on translational efforts such as gene expression and 2011 American Association for Cancer Research. clinical response data. We present data from a U.S.

42 Cancer Epidemiol Biomarkers Prev; 20(1) January 2011

SYK and 6p21.3 Genetic Variations in NHL

population-based case–control study of NHL where gen- Build 36.1 assembly, dbSNPb126) using the software otyping for nearly 488 genes (6,679 SNPs) were con- Tagzilla, which implements a tagging algorithm based ducted. Genes and SNPs evaluated are annotated in on the pairwise binning method of Carlson et al. (15). For Supplementary Materials (Supplementary Table S1). each original target gene, SNPs within the region span- ning 20-kb 50 of the start of transcription (exon 1) to 10-kb Materials and Methods 30 of the end of the last exon were grouped using a binning threshold of r2 > 0.8 to define a gene region. Methods When there were multiple transcripts available for genes, Study population. The National Cancer Institute-Sur- only the primary transcript was assessed. veillance, Epidemiology and End Results (NCI-SEER) In the current manuscript, we exclude data from 22 NHL case–control study has been previously described genes (5% of total SNPs) for presentation due to prior (9;10). Briefly, the study included 1,321 incident NHL commitments to ongoing pooling efforts within the Inter- cases without evidence of HIV infection identified in 4 Lymph Consortium. These exclusions do not alter the SEER registries (Iowa; Detroit, MI; Los Angeles, CA; main results, discussion or conclusion based on our Seattle, WA) aged 20 to 74 years. 1,057 population con- analysis. trols were identified by random digit dialing (<65 years) and from Medicare eligibility files (65 years). Overall Quality control participation rates were 76% in cases and 52% in controls; Tag SNPs that failed manufacturing (ordered but did overall response rates were 59% and 44%, respectively. not convert), failed validation (no amplification or clus- Written informed consent was obtained from each parti- tering) and assays that had less than 80% completion or cipant prior to interview. 80% concordance with the 90 Hapmap CEU samples used Histopathology. All cases were histologically con- for validation were excluded, resulting in a subset of firmed by the local diagnosing pathologist. NHL sub- 6,830 tag SNPs with data for further analysis. Of the 6,830 types were originally coded according to the tag SNPs, SNPs with low completion rate (<90% of International Classification of Diseases for Oncology, samples) were excluded, including 53 SNPs among sam- 2nd Edition (11) and then updated to the ICD-O-3/World ples with DNA extracted from blood and 82 SNPs among Health Organization classification. In addition to NHL samples with DNA extracted from buccal cells. Replicate overall, we evaluated 4 B-cell subtypes: diffuse large B- samples (n ¼ 62) from 2 blood donors each and duplicate cell lymphoma (DLBCL), follicular lymphoma, marginal samples from 93 participants processed in an identical zone lymphoma, and chronic lymphocytic leukemia and fashion were interspersed for all assays and blinded from small lymphocytic lymphoma (CLL/SLL). the laboratory. For each plate of 368 samples, genotype- Biological samples and DNA extraction. Study par- specific quality control (QC) samples were also included ticipants who did not provide a biological specimen, did and comprised 4 each: homozygote wild-type (WT), not have sufficient material for DNA extraction or suffi- heterozygote, homozygote variant, and DNA-negative cient DNA for genotyping, or whose genotyped sex was controls. SNPs with concordance less than 95% in QC discordant from the questionnaire data were excluded samples were excluded, including 1 SNP from samples from this analysis. Of the 1,231 cases (820 blood, 411 with DNA extracted from blood and 13 SNPs from buccal cell) and 992 controls (692 blood, 300 buccal cell) samples with DNA extracted from buccal cells. Concor- with biospecimens, 963 cases and 837 controls were dance for all remaining QC replicates and duplicates was genotyped. As previously described, DNA was extracted 99% or more for all assays. We further excluded samples from blood clots or buffy coats (BBI Biotech) using Pure- with a low completion rate (<90%; n ¼ 27). gene Autopure DNA extraction kits (Gentra Systems), Hardy–Weinberg equilibrium was evaluated among and from buccal cell samples by phenol–chloroform non-Hispanic Caucasian controls. SNPs showing evi- extraction methods (12). Genotype frequencies for indi- dence of deviation from Hardy–Weinberg proportions viduals who provided blood compared with buccal cells (P < 0.0001) are denoted in Supplementary Materials were equivalent (13). (Supplementary Table S1). Final analytic population. The final analytic study Genotyping population included 947 cases and 826 controls with data Genotyping of tag SNPs from 488 candidate gene for 6,679 SNPs. regions (Supplementary Table S1) hypothesized to be involved in lymphomagenesis was conducted at the Statistical methods NCI Core Genotyping Facility (Advanced Technology Gene region-based analyses. We obtained a gene Center; ref. 14) using a custom-designed Infinium assay region-level summary of association by computing the (Illumina). The Infinium included a total of 7,943 tag minP test, which assesses the statistical significance of the SNPs. Tag SNPs were chosen from the designable set smallest P-trend within each gene region (determined by of common SNPs [minor allele frequency (MAF) > 5%] dichotomous logistic regression, comparing NHL or genotyped in the Caucasian (CEU) population sample of NHL subtypes with controls) by permutation-based the HapMap Project (Data Release 20/Phase II, NCBI resampling methods (10,000 permutations) that automa-

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 43

Wang et al.

tically adjust for the number of tag SNPs tested within between SNPs. For blocks of linkage disequilibrium, we that gene and the underlying linkage disequilibrium obtained ORs and 95% CIs for the underlying haplotypes pattern (16;17). To account for multiple comparisons under the assumption of an additive model (haplo.glm, within the 488 candidate gene regions, we applied the minimum haplotype frequency 1%). All haplotype ana- false discovery rate (FDR) method of Benjamini and lyses were adjusted for age, sex, and study center. Hochberg (18) to the minP test separately for NHL and each subtype. Results SNP-based analyses. We calculated odds ratios (OR) and 95% CI estimating the relative risk of NHL and NHL Descriptive characteristics of the 947 cases and 826 subtypes in relation to each genotype using dichotomous controls are provided in Supplementary Table S2. The and polytomous unconditional logistic regression mod- distributions of cases and controls were similar with els, respectively. The homozygote of the most common respect to sex, age, and race/ethnicity, with our popula- allele in the pooled study population was used as the tion largely comprising non-Hispanic Caucasians. The 2 referent group. Tests for trend under the codominant predominant NHL subtypes were DLBCL (30%) and model used a 3-level ordinal variable for each SNP follicular lymphoma (25%). (0 ¼ homozygote common, 1 ¼ heterozygote, 2 ¼ homozygote variant). All models were adjusted for age, race/ SNP- and haplotype-based analyses ethnicity, sex, and study center (categories listed in Sup- There were 5 SYK SNPs statistically significant with a plementary Table S2). We also conducted all analyses P-trend less than 0.001 for all NHL, including the most restricted to non-Hispanic Caucasians. We assessed het- significant SNP observed in our overall analysis erogeneity among NHL subtypes in the polytomous (rs2991216, P-trend ¼ 0.00005; Table 1). The associations multivariate unconditional logistic regression models were consistently in the same direction across the 4 using the Wald chi-square statistic. Analyses were con- subtypes evaluated, though most significantly associated ducted using SAS version 9.1 (SAS Institute). with DLBCL. On the basis of the haplotype structure, Haplotype analyses. We conducted haplotype ana- statistically significant SNPs for SYK fall within 2 larger lyses among non-Hispanic Caucasians using 2 methods. blocks. The first block includes SYK SNPs rs2035072 and 0 First, we evaluated risk of NHL and NHL subtypes rs290213 (D ¼ 0.93) and the second block includes associated with haplotypes defined by SNPs within a rs2991216, rs965892, and rs290203 (D’ ¼ 0.88–0.99). Link- 0 sliding window of 3 loci across a gene (Haplo Stats, age between the SNPs in the 2 blocks was low (D ¼ 0.24– version 1.2.1). A global score statistic was used to sum- 0.36), suggesting that 2 independent regions within SYK marize the evidence of association of disease with the may be associated with NHL etiology. Haplotype ana- haplotypes for each window. Second, we visualized hap- lyses by hapwalk of 3 SNPs yielded the same 2 regions of lotype structures using Haploview, version 3.11 (19) high significance for NHL risk. Haplotype analyses of the based on measures of pairwise linkage disequilibrium blocks as defined by Gabriel et al. did not reveal any

Table 1. SNPs with a P-trend < 0.001 for NHL from the logistic regression analyses

Gene Chromosome SNP NHL DLBCL Follicular Marginal zone CLL/SLL

SYK 9 rs2991216 0.00005 0.00040 0.05855 0.11674 0.02407 RING1,RXRB 6 rs2855429 0.00010 0.01661 0.01414 0.18289 0.10849 AIF1 6 rs2857597 0.00024 0.00051 0.02623 0.03094 0.06917 BAT4 6 rs3115667 0.00025 0.00050 0.02544 0.03667 0.11315 SYK 9 rs290213 0.00026 0.00068 0.08478 0.01554 0.46632 SYK 9 rs965892 0.00026 0.00027 0.10240 0.19448 0.04582 BCL11A 2 rs2556377 0.00029 0.07663 0.00136 0.01596 0.06369 KRAS 2 rs6487465 0.00030 0.00645 0.05695 0.02318 0.12108 SYK 9 rs290203 0.00041 0.00173 0.03237 0.40725 0.04097 ITGB2 21 rs2838735 0.00045 0.11975 0.01900 0.12890 0.00036 SMARCA2 9 rs10964501 0.00052 0.00019 0.28971 0.25884 0.96541 UGT1A8 2 rs17863784 0.00057 0.00619 0.03811 0.22683 0.89960 RING1,RXRB 6 rs213213 0.00059 0.01218 0.04787 0.75352 0.02204 ITGB2 21 rs2838734 0.00065 0.09384 0.02203 0.42666 0.00400 BCL11A 2 rs2556376 0.00071 0.08290 0.00109 0.01870 0.12928 SYK 9 rs2035072 0.00080 0.01138 0.02624 0.04159 0.02572

NOTE: Corresponding P for NHL subtypes also listed. Data for all SNPs are shown in Supplementary Table S4.

44 Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 Cancer Epidemiology, Biomarkers & Prevention

SYK and 6p21.3 Genetic Variations in NHL

further regions of significance as all haplotypes asso- NHL (P ¼ 0.004, FDR P ¼ 0.27), and also with both ciated with NHL included the variant allele for DLBCL and follicular lymphoma (P ¼ 0.01). Results for rs2035072 (data not shown). all gene regions evaluated are shown in Supplementary Corresponding SNPs for the 3 gene regions in chromo- Table S4. some 6p21.3 were statistically significantly associated Other regions of potential importance include genes on with NHL at P < 0.001, including SNPs in RING1/RXRB chromosome 2 (BCL11A, UGT1A8), chromosome 12 (rs2855429, rs213213), AIF1 (rs2857597), and BAT4 (KRAS), and chromosome 21 (ITGB2) where SNPs and (rs3115667; Table 1). By subtype, all 4 SNP associations gene regions were both statistically significant at P-trend were significant in both DLBCL and follicular lymphoma < 0.001 and permutation P < 0.01, respectively, for NHL (Table 1). The AIF1 and BAT4 SNPs were also signifi- overall (Table 1 and 2). In subtype-specific results, the cantly associated with marginal zone lymphoma, but the association between PGR and DLBCL was also note- 2 SNPs in RING1/RXRB were not. All associations worthy (P-trend < 0.0001; permutation P < 0.001). remained statistically significant in models adjusted for the implicated TNF -308A and HLA DRB1*01010 alleles Discussion (RING1/RXRB (rs2855429 TNF/HLA-adjusted P-trend ¼ 0.0003, rs213213 TNF/HLA-adjusted P-trend ¼ 0.001), In our evaluation of a priori candidate genetic poly- AIF1 (rs2857597 TNF/HLA-adjusted P-trend ¼ 0.02), morphisms in NHL etiology, the SYK SNPs comprised and BAT4 (rs3115667 TNF/HLA-adjusted P-trend ¼ the most statistically significant SNP of all variants eval- 0.01)). We note, however, that power was limited in these uated and were associated with NHL with P-trends < adjusted models because they are restricted to the two- 0.0001. Our results require replication but if confirmed, third of samples with data for the HLA allele. In general, are of interest for several reasons. SYK is important in B- risk estimates for each SNP were modest (e.g., less than a cell development and recent evidence supports its poten- 2-fold increase or decrease in risk; Supplementary tial as a therapeutic target (20–22). SYK is a protein- Table S3). tyrosine kinase gene encoded on chromosome 9 that is We plotted the P-trends for the 288 SNPs from the 27 expressed in hematopoietic cells and predominantly in gene regions that were evaluated in the 6p21.3 region for the lymphoid tissues of the spleen and thymus. Syk is this analysis with NHL (Fig. 1) and NHL subtypes (Sup- activated by the B-cell receptor and mediates a variety of plementary Figure S1). For NHL and DLBCL, we show critical cellular responses. Syk expression is necessary for that the statistical significance of all 3 SNPs is greater than the survival of human NHL cell lines, and its disruption that for the confirmed TNF G-308A polymorphism leads to cellular apoptosis in NHL mouse models (21). (rs1800629) in our data. Hapwalk analysis of 3-SNPs Friedberg and colleagues demonstrated that an oral Syk yielded the same regions of significance in AIF1, BAT4, inhibitor provided to patients with B-cell NHL improved and RING1/RXRB for NHL (Fig. 1). Importantly, the progression-free survival, particularly for CLL/SLL (20). haplotype structure reveals that the implicated regions If our results are replicated, determining the functional are in distinct haplotypic blocks from the TNF/LTA consequences of SNPs in SYK is warranted. region. The haplotype structure further shows the Our results also provide further evidence that variation RING1/RXRB region to be distinct from the AIF1 and among additional genes in the 6p21.3 chromosomal BAT4 SNPs, which is consistent with our observation that region is important for NHL etiology across multiple SNPs in AIF1 and BAT4 are potentially distinct from NHL subtypes. Notably, the SNPs within the implicated RING1/RXRB based on their different subtype-specific gene regions — RING1/RXRB, AIF1, and BAT4 — are patterns. neither correlated nor in linkage disequilibrium with the Haplotype analyses of blocks as defined by Gabriel confirmed TNF G-308A polymorphism (1;2) and reside in et al. did not yield any further information; no haplotypes different haplotype blocks as shown in Fig. 1C. These were more significant than the individual SNP associa- genes are also not linked to the well-established ancestral tions (Fig. 1). haplotype 8.1 that comprises TNF-308A-HLA A*01-B- 07*DRB1*03 that is associated with NHL and remain Gene region-based analyses statistically significantly associated with NHL even in Our SNP-based results are supported by our evalua- logistic regression models that include both the impli- tion of gene region-based analyses. Of the 488 gene cated TNF-308A and HLA DRB1*0101 alleles. Still, our regions evaluated, 14 were significant with a P < 0.01 alleles may represent surrogates for other causal genes for all NHL. The 3 top gene regions of significance associations with NHL and NHL subtypes. These collec- (RING/RXRB; AIF1; and BAT4) are located on chromo- tive data along with emerging evidence for HLA (3;4) and some 6p21.3 (Table 2) and had FDR-adjusted P values non-HLA genes (3) located in the same region in NHL 0.25. By subtype, the RING1/RXRB region was signifi- etiology strongly suggest that a detailed evaluation of cantly associated with follicular lymphoma, the BAT4 genetic variation in the 6p21.3 chromosomal region is region with DLBCL, and the AIF1 region with both warranted. DLBCL and follicular lymphoma. The gene region Of the genes in 6p21.3 we found to be significantly encompassing SYK was also significantly associated with associated with NHL etiology, BAT4 has previously been

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 45

Wang et al.

A 6p21.3

PRL AIF1 MSH5 RDBP TNXB,CREBL1 TAP2 BRD2 DAXX,TAPBP GLO1 TNFRSF21 C6ORF32 BAT2 HSPA1A,HSPA1L CREBL1 TAP2,PSMB9 RING1,RXRB BAK1 CCND3 TNF/LTA BAT4 EHMT2 TNXB PBX2 TAP1,TAP2 RPS18 SRPK1 NFKBI3 0.0001 rs2855429 rs2857597rs2857597 rs3115667 AIF1,BAT2, 0.0002 BAT4, 0.0003 RING1,RXRB, 0.0001 AIF1, 0.0002 rs213213 RING1,RXRB, 0.0006 0.0010

0.0100 rs1800629 TNF,LTA

0.1000

1.0000 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 B

0.00001

0.0001

0.001

0.01

0.1

1.0

PRL AIF1 MSH5 RDBP TNXB,CREBL1 TAP2 BRD2 DAXX,TAPBP GLO1 TNFRSF21 C6ORF32 BAT2 HSPA1A,HSPA1L CREBL1 TAP2,PSMB9 RING1,RXRB BAK1 CCND3 TNF/LTA BAT4 EHMT2 TNXB PBX2 TAP1,TAP2 RPS18 SRPK1 NFKBI3

HLA-A HLA-C HLA-B MICB HLA-DR*A HLA-DR*B HLA-DQ*A HLA_DQ*B

Figure 1. Graphed P values for 288 SNPs (27 gene regions) in the 6p21.3 region for NHL via (A) p-trends for individual SNPs from logistic regression analysis, (B) haplowalk using 3-SNP models, and (C) haplotype structure as viewed on Haploview 4.1.

46 Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 Cancer Epidemiology, Biomarkers & Prevention

SYK and 6p21.3 Genetic Variations in NHL

Table 2. Significance levels (P values) for 14 gene regions with permutation P < 0.01 (based on 10,000 permutations and the minimum P-trend within each region), for the association with NHL

Gene Chromosome # SNPs/gene NHL DLBCL Follicular Marginal zone CLL/SLL

RING1,RXRB 6 9 0.0010 0.0708 0.0409 0.7518 0.0992 AIF1 6 5 0.0016 0.0023 0.0129 0.1237 0.2579 BAT4 6 9 0.0019 0.0035 0.2347 0.0659 0.5583 SYK 9 66 0.0040 0.0143 0.0122 0.3587 0.2286 KRAS 12 20 0.0046 0.1269 0.6163 0.3515 0.3285 CXCL10 4 11 0.0054 0.0880 0.1132 0.7726 0.7691 BCL11A 2 28 0.0065 0.5952 0.0076 0.1815 0.6041 UGT1A8 2 13 0.0074 0.0301 0.0685 0.5243 0.338 NOS1 12 30 0.0080 0.0043 0.7263 0.9082 0.5876 NFATC3 16 4 0.0085 0.2217 0.1603 0.3774 0.6858 CXCL9 4 2 0.0090 0.3131 0.0911 0.5258 0.4654 ITGB2 21 22 0.0090 0.5866 0.1592 0.1333 0.0052 MSH5 6 8 0.0090 0.0142 0.0174 0.0963 0.7022 CD8B 2 8 0.0099 0.1195 0.8109 0.1434 0.7614

NOTE: Corresponding P values also listed for NHL subtypes (DLBCL, follicular, marginal zone, and CLL/SLL). Data for all genes are shown in Supplementary Table S3 for all NHL and NHL subtypes).

Table 3. Description of top-ranked genes associated with NHL lymphoma risk

Gene Chromosome Gene descriptiona Additional commentsb

RING1,RXRB 6p21.3 Ring finger protein 1, retinoid X receptor, beta In MHC class II locus AIF1 6p21.3 Allograft inflammatory factor In TNF cluster of genes located in human MHC BAT4 6p21.3 HLA-B associated transcript 4 SYK 9q22 Spleen tyrosine kinase ZAP70-deficient patients express high SYK levels KRAS 12p12.1 v-Ki-ras2 Kirsten rat sarcoma viral oncogene RAS mutations involved in leukemia homolog CXCL10 4q21 Chemokine (C-X-C motif) ligand 10 Increased expression among autoimmune myasthenia gravis BCL11A 2p13 B-cell CLL/lymphoma 11A (zinc finger protein) Deregulated expression in B-cell malignancy, B-CLL and NHL UGT1A8 2q37 UDP glucuronosyltransferase 1 family, Metabolizes flavonoids polypeptide A8 NOS1 12q24.2-q24.31 Nitric oxide synthase 1 (neuronal) Mediates tumoricidal and bactericidal activity NFATC3 16 Nuclear factor of activated T cells, Deregulation contributes to clinical cytoplasmic, calcineurin-dependent 3 manifestation of H. pylori infection CXCL9 4q21 Chemokine (C-X-C motif) ligand 9 T-cell chemoattractant inducible by gamma interferon ITGB2 21q22.3 Integrin, beta 2 (complement component 3 Marginal zone B cells express elevated levels receptor 3 and 4 subunit) MSH5 6p21.3 mutS homolog 5 Genetic variations identified in patients with IgA deficiency or common variable immunodeficiency

aAs defined in NCBI Entrez Gene (http://www.ncbi.nlm.nih.gov/sites/entrez). bGene Ontology annotation (http://www.geneontology.org/). Online Mendelian Inheritance in Man.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 47

Wang et al.

linked to NHL susceptibility, specifically for CLL (23). In identifying specific regions of functional importance our data, we found a different BAT4 SNP (rs3115667) to to NHL etiology. Additional studies will be needed to be significantly associated with all NHL, DLBCL, folli- delineate gene–environment interactions, particularly as cular and marginal zone lymphomas but not CLL/SLL. a number of autoimmune conditions, some of which are The remaining gene regions implicated in our data have potential risk factors for NHL (26), are also associated not previously been implicated in NHL etiology, though with susceptibility alleles in HLA and chromosome they have been implicated in other aspects of lympho- 6p21.3 immune response genes. Our results for SYK magenesis. Lower expression of RXRB and other MHC require replication, but are of great interest due to its Class II genes were reported to be correlated with poor importance in lymphomagenesis and its growing poten- patient survival in mediastinal DLBCL (24) and the AIF1 tial as a therapeutic target. is one of a group of genes (including BCL6, LMO2, GCET1) found to define the germinal center B-cell sig- Disclosure of Potential Conflicts of Interest nature (25), also with implications for DLBCL survival. A brief description of all top-ranked genes associated with No potential conflicts of interest were disclosed. NHL is shown in Table 3. Our study strengths include the population-based Acknowledgments design of this study and the tagging algorithm intended We thank Peter Hui, Michael Stagner, and Mary McAdams of the to allow for gene-based and SNP-based analyses. We note Information Management Services, Inc. for their programming support. that our results are consistent when restricted to non- We also gratefully acknowledge the contributions of the staff and scien- tists at the SEER centers of Iowa, Los Angeles, Detroit, and Seattle for the Hispanic Caucasians and thus are unlikely to be biased conduct of the study’s field effort. by population stratification. Still, we cannot fully exclude the potential for false positive associations in our results Grant Support and they therefore require replication in other independent and larger populations such as in pooled efforts. The NCI-SEER study was supported by the Intramural Research Program of Other study limitations include loss of eligible subjects to the NIH (NCI), and by Public Health Service (PHS) contracts N01-PC-65064, N01- death, illness, and refusal to participate. We also had PC-67008, N01-PC-67009, N01-PC-67010, and N02-PC-71105. DNA extraction, genotyping and statistical analysis for this project were supported by the limited power to detect modest associations for rare Intramural Research Program of the NIH [National Cancer Institute (NCI)]. alleles and less common NHL subtypes. This project has been funded in part with federal funds from the NCI, NIH, under contract no. HHSN261200800001E. The content of this publication does not In summary, our data suggest genetic variations in the necessarily reflect the views or policies of the Department of Health and Human SYK gene on chromosome 9 and 3 additional genes in the Services, nor does mention of trade names, commercial products, or organiza- 6p21.3 chromosomal region to be associated with NHL tions imply endorsement by the U.S. Government. The costs of publication of this article were defrayed in part by the etiology. Our results further support the need payment of page charges. This article must therefore be hereby marked for in-depth analyses of genetic variants in the 6p21.3 advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate chromosomal region. In particular, combining HLA this fact. allele information with data emerging from genome-wide Received June 17, 2010; revised October 20, 2010; accepted association studies should contribute significantly to November 16, 2010; published OnlineFirst December 10, 2010.

References 1. Rothman N, Skibola CF, Wang SS, Morgan G, Lan Q, Smith MT, et al. 7. Shen M, Menashe I, Morton LM, Zhang Y, Armstrong B, Wang SS, Genetic variation in TNF and IL10 and risk of non-Hodgkin lymphoma: a et al. Polymorphisms in DNA repair genes and risk of non-Hodgkin report from the InterLymph Consortium. Lancet Oncol 2006;7:27–38. lymphoma in a pooled analysis of three studies. Br J Haematol 2. Skibola CF, Bracci PM, Nieters A, Brooks-Wilson A, de SS, Hughes 2010;151:239–44. AM, et al. Tumor Necrosis Factor (TNF) and Lymphotoxin-{alpha} 8. Lan Q, Morton LM, Armstrong B, Hartge P, Menashe I, Zheng T, et al. (LTA) Polymorphisms and Risk of Non-Hodgkin Lymphoma in the Genetic variation in caspase genes and risk of non-Hodgkin lym- InterLymph Consortium. Am J Epidemiol 2010. phoma: a pooled analysis of 3 population-based case-control studies. 3. Skibola CF, Bracci PM, Halperin E, Conde L, Craig DW, Agana L, et al. Blood 2009;114:264–7. Genetic variants at 6p21.33 are associated with susceptibility to 9. Chatterjee N, Hartge P, Cerhan JR, Cozen W, Davis S, Ishibe N, et al. follicular lymphoma. Nat Genet 2009;41:873–5. Risk of non-Hodgkin's lymphoma and family history of lymphatic, 4. Wang SS, Abdou AM, Morton LM, Thomas R, Cerhan JR, Gao X, et al. hematologic, and other cancers. Cancer Epidemiol Biomarkers Prev Human leukocyte antigen class I and II alleles in non-Hodgkin lym- 2004;13:1415–21. phoma etiology. Blood 2010. 10. Wang SS, Cerhan JR, Hartge P, Davis S, Cozen W, Severson RK, et al. 5. Wang SS, Purdue MP, Cerhan JR, Zheng T, Menashe I, Armstrong BK, Common genetic variants in proinflammatory and other immunore- et al. Common gene variants in the tumor necrosis factor (TNF) and gulatory genes and risk for non-Hodgkin lymphoma. Cancer Res TNF receptor superfamilies and NF-kB transcription factors and non- 2006;66:9771–80. Hodgkin lymphoma risk. PLoS One 2009;4:e5360. 11. Harris NL, Jaffe ES, Diebold J, Flandrin G, Muller-Hermelink HK, 6. Morton LM, Purdue MP, Zheng T, Wang SS, Armstrong B, Zhang Y, Vardiman J, et al. The World Health Organization classification of et al. Risk of non-Hodgkin lymphoma associated with germline varia- neoplastic diseases of the hematopoietic and lymphoid tissues. tion in genes that regulate the cell cycle, apoptosis, and lymphocyte Report of the Clinical Advisory Committee meeting, Airlie House, development. Cancer Epidemiol Biomarkers Prev 2009;18:1259–70. Virginia, November, 1997. Ann Oncol 1999;10:1419–32.

48 Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 Cancer Epidemiology, Biomarkers & Prevention

SYK and 6p21.3 Genetic Variations in NHL

12. Garcia-Closas M, Egan KM, Abruzzo J, Newcomb PA, Titus-Ernstoff 19. Barrett J, Fry B, Maller J, Daly M. Haploview: analysis and visualization L, Franklin T, et al. Collection of genomic DNA from adults in epide- of LD and haplotype maps. Bioinformatics 2005;21:263–5. miological studies by buccal cytobrush and mouthwash. Cancer 20. Friedberg JW, Sharman J, Sweetenham J, Johnston PB, Vose JM, Epidemiol Biomarkers Prev 2001;10:687–96. Lacasce A, et al. Inhibition of Syk with fostamatinib disodium has 13. Bhatti P, Sigurdson AJ, Wang SS, Chen J, Rothman N, Hartge P, et al. significant clinical activity in non Hodgkin's lymphoma and chronic Genetic variation and willingness to participate in epidemiologic lymphocytic leukemia. Blood 2009. research: data from three studies. Cancer Epidemiol Biomarkers Prev 21. Young RM, Hardy IR, Clarke RL, Lundy N, Pine P, Turner BC, et al. 2005;14:2449–53. Mouse models of non-Hodgkin lymphoma reveal Syk as an important 14. Packer BR, Yeager M, Burdett L, Welch R, Beerman M, Qi L, therapeutic target. Blood 2009;113:2508–16. et al. SNP500Cancer: a public resource for sequence validation, 22. Leonard JP, Martin P, Barrientos J, Elstrom R. Targeted treatment and assay development, and frequency analysis for genetic new agents in diffuse large B-cell lymphoma. Semin Hematol 2008;45 variation in candidate genes. Nucl Acids Res 2006;34Suppl 1: Suppl 2:S11–6. D617–21. 23. Rudd MF, Sellick GS, Webb EL, Catovsky D, Houlston RS. Variants in 15. Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA. the ATM-BRCA2-CHEK2 axis predispose to chronic lymphocytic Selecting a maximally informative set of single-nucleotide polymorph- leukemia. Blood 2006;108:638–44. isms for association analyses using linkage disequilibrium. Am J Hum 24. Roberts RA, Wright G, Rosenwald AR, Jaramillo MA, Grogan TM, Miller Genet 2004;74:106–20. TP, et al. Loss of major histocompatibility class II gene and protein 16. Westfall P, Zaykin D, Young S. Multiple tests for genetic effects in expression in primary mediastinal large B-cell lymphoma is highly coor- association studies. Methods Mol Biol 2002;184:143–68. dinated and related to poor patient survival. Blood 2006;108:311–8. 17. Chen B, Sakoda L, Hsing AW, Rosenberg PS. Resampling-based 25. Shaffer AL, Rosenwald A, Staudt LM. Lymphoid malignancies: the dark multiple hypothesis testing procedures for genetic case-control asso- side of B-cell differentiation. Nat Rev Immunol 2002;2:920–32. ciation studies. Genet Epidemiol 2006;30:495–507. 26. Wang SS, Hartge P. Etiology and epidemiology of non-Hodgkin 18. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a lymphoma. In:Mauch PM, Armitage JO, Harris NL, Coiffier B, Dalla- practical and powerful approach to multiple testing. J R Stat Soc B Favera R,editors. Non-Hodgkin Lymphoma. Baltimore, MD: Lippin- 1995;57:289–300. cott, Williams and Wilkens; 2010.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 20(1) January 2011 49

Variations in Chromosomes 9 and 6p21.3 with Risk of Non− Hodgkin Lymphoma

Sophia S. Wang, Idan Menashe, James R. Cerhan, et al.

Cancer Epidemiol Biomarkers Prev 2011;20:42-49. Published OnlineFirst December 10, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/1055-9965.EPI-10-0638

Supplementary Access the most recent supplemental material at: Material http://cebp.aacrjournals.org/content/suppl/2010/12/10/1055-9965.EPI-10-0638.DC1

Cited articles This article cites 22 articles, 9 of which you can access for free at: http://cebp.aacrjournals.org/content/20/1/42.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://cebp.aacrjournals.org/content/20/1/42.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/20/1/42. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.