PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of amjepidLink to Publisher's site
 
Am J Epidemiol. 2010 February 1; 171(3): 267–276.
Published online 2010 January 4. doi:  10.1093/aje/kwp383
PMCID: PMC2842204

Tumor Necrosis Factor (TNF) and Lymphotoxin-α (LTA) Polymorphisms and Risk of Non-Hodgkin Lymphoma in the InterLymph Consortium

Abstract

In an International Lymphoma Epidemiology Consortium pooled analysis, polymorphisms in 2 immune-system-related genes, tumor necrosis factor (TNF) and interleukin-10 (IL10), were associated with non-Hodgkin lymphoma (NHL) risk. Here, 8,847 participants were added to previous data (patients diagnosed from 1989 to 2005 in 14 case-control studies; 7,999 cases, 8,452 controls) for testing of polymorphisms in the TNF –308G>A (rs1800629), lymphotoxin-α (LTA) 252A>G (rs909253), IL10 –3575T>A (rs1800890, rs1800896), and nucleotide-binding oligomerization domain containing 2 (NOD2) 3020insC (rs2066847) genes. Odds ratios were estimated for non-Hispanic whites and several ethnic subgroups using 2-sided tests. Consistent with previous findings, odds ratios were increased for “new” participant TNF –308A carriers (NHL: per-allele odds ratio (ORallelic) = 1.10, Ptrend = 0.001; diffuse large B-cell lymphoma (DLBCL): ORallelic = 1.23, Ptrend = 0.004). In the combined population, odds ratios were increased for TNF –308A carriers (NHL: ORallelic = 1.13, Ptrend = 0.0001; DLBCL: ORallelic = 1.25, Ptrend = 3.7 × 10−6; marginal zone lymphoma: ORallelic = 1.35, Ptrend = 0.004) and LTA 252G carriers (DLBCL: ORallelic = 1.12, Ptrend = 0.006; mycosis fungoides: ORallelic = 1.44, Ptrend = 0.015). The LTA 252A>G/TNF –308G>A haplotype containing the LTA/TNF variant alleles was strongly associated with DLBCL (P = 2.9 × 10−8). Results suggested associations between IL10 –3575T>A and DLBCL (Ptrend = 0.02) and IL10 –1082A>G and mantle cell lymphoma (Ptrend = 0.04). These findings strengthen previous results for DLBCL and the LTA 252A>G/TNF –308A locus and provide robust evidence that these TNF/LTA gene variants, or others in linkage disequilibrium, are involved in NHL etiology.

Keywords: lymphoma, lymphoma, non-Hodgkin, lymphotoxin-alpha, meta-analysis, polymorphism, genetic, polymorphism, single nucleotide, tumor necrosis factor-alpha

There is evidence that genetic susceptibility plays a role in lymphomagenesis. Population-based studies show 2-fold risk elevations for non-Hodgkin lymphoma (NHL) among persons with a family history of hematopoietic malignancies (1, 2). Furthermore, findings from case-control studies suggest associations between a small but growing number of common genetic variants and risk of NHL (3). Immune system deregulation is an established risk factor for NHL, and common variations among immune-response genes are of particular interest in understanding the pathogenesis of lymphoma. Specifically, inflammatory response genes may play a role in chronic antigenic stimulation, which is a key mechanism for lymphomagenesis. Whereas immune signatures have been reported for follicular and Burkitt's lymphoma (4, 5), specific expression signatures involving genes that mediate inflammation have been reported in diffuse large B-cell lymphoma (DLBCL) (6).

We previously reported statistically significant associations between variants in 2 proinflammatory cytokine genes, tumor necrosis factor-α (TNF –308G>A) and interleukin-10 (IL10 –3575T>A), and DLBCL among 3,586 Caucasian NHL cases and 4,018 Caucasian controls enrolled in 8 case-control studies that are part of the International Lymphoma Epidemiology (InterLymph) Consortium (7). An association between these single nucleotide polymorphisms (SNPs) and DLBCL, but not follicular lymphoma, indicated that genetic variation in inflammatory mediators such as TNF and IL10 may play an essential role in DLBCL etiology. In our initial pooled analysis, we also noted statistically significant associations between DLBCL and 2 other polymorphisms in genes that contribute to nuclear factor-κB activation, lymphotoxin-α (LTA) 252A>G and a C insertion that results in a frame-shift mutation in exon 11 of the nucleotide-binding oligomerization domain containing 2 (NOD2) gene (3020insC) (7). However, LTA 252A>G is in linkage disequilibrium with TNF –308G>A, and a haplotype analysis could not separate effects of risk for these 2 variants. The results for NOD2 were limited by the rarity of the SNP (<1% frequency of homozygote variants). Furthermore, the sample was not large enough for us to evaluate the above associations with less common NHL subtypes or in other ethnic groups.

To confirm our initial findings, we evaluated TNF –308G>A, IL10 –3575T>A, and LTA 252G>A in 6,601 human immunodeficiency virus (HIV)-negative non-Hispanic white participants (3,278 NHL cases and 3,323 controls) from 8 study centers not included in the previous pooled analyses (7). We then pooled all InterLymph study data (7,999 NHL cases and 8,452 controls) from 14 European, Canadian, Australian, and US case-control studies to increase precision and statistical power for analyzing the associations of the TNF –308G>A, IL10 –3575T>A, LTA 252G>A, and NOD2 3020insC SNPs with NHL. We further extended our analyses to include other, rarer subtypes of NHL and other ethnic groups—that is, black, Hispanic, and Asian populations.

MATERIALS AND METHODS

Participating InterLymph studies

Participating studies from the InterLymph Consortium included the York, United Kingdom, lymphoma study (8); the Connecticut Women's NHL Study from Yale University (9); the National Cancer Institute (NCI)/Surveillance, Epidemiology, and End Results (SEER) Multi-Center Case-Control Study (10); the Mayo Clinic NHL study (11); the New South Wales, Australia, lymphoma study (12); the British Columbia, Canada, NHL study (13); the University of California, San Francisco (UCSF)/University of California, Berkeley studies of NHL (2 studies, hereafter called UCSF1 and UCSF2) (1416); and studies from Spain, Germany, France, the Czech Republic, Ireland, and Italy (Sardinia) that were part of a multicenter European study (EpiLymph) (17) (Table 1). Demographic data, including age at diagnosis for cases and age at interview for controls, sex, self-reported race and Hispanic/Latino ethnicity (or geographic ancestry for the EpiLymph and British Columbia studies), HIV status, histologic subtype classification and/or International Classification of Diseases for Oncology code for lymphoma diagnoses, and genotype data for the TNF –308G>A (rs1800629), LTA 252A>G (rs909253), IL10 –3575T>A (rs1800890), IL10 –1082A>G (rs1800896), and NOD2 3020insC (rs2066847) polymorphisms were obtained for each study. Cases were diagnosed with incident lymphoma between 1989 and 2005. Analyses were restricted to participants aged ≥18 years and to patients diagnosed with NHL, as some studies included other lymphatic malignancies. (See Table 1 for individual study characteristics.) Data were further restricted to HIV-negative, non-Hispanic white participants for the primary analyses. Hispanic ethnicity was defined per the US definition as someone of Mexican, Puerto Rican, or Latin-American descent—that is, an admixed population of Native American, Caucasian, and black/African ancestry. Therefore, Caucasian participants from EpiLymph Spain were included in the analyses of non-Hispanic whites, with the exception of several participants who met the definition of Hispanic ethnicity based on their parental ancestry information.

Table 1.
Characteristics of Studies Included in Pooled Analyses of the Relation of TNF, LTA, IL10, and NOD2 Genetic Polymorphisms to Non-Hodgkin Lymphoma, International Lymphoma Epidemiology Consortium, 1989–2005

Histologic subtypes were grouped for analyses using the pathology coding scheme developed by InterLymph collaborators and study pathologists, which was based on the current World Health Organization classification (18). Outcomes for which results are presented include NHL and broad categories of NHL by cell lineage (B-cell, T-cell, and unknown), as well as specific subtypes, including DLBCL, follicular lymphoma, chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL), mantle cell lymphoma, marginal zone lymphoma (MZL) (further subdivided into extranodal MZL and other MZL), lymphoplasmacytic lymphoma, mycosis fungoides, and peripheral T-cell lymphoma.

Genotyping

SNPs were genotyped using either TaqMan (Applied Biosystems, Foster City, California) or Pyrosequencing (Qiagen NV, Hilden, Germany) (all EpiLymph studies). Assay conditions for TaqMan assays are available on the NCI's SNP500Cancer Web site (http://snp500cancer.nci.nih.gov). All laboratories genotyped a quality control set of DNA samples from 102 ethnically diverse individuals that had been previously sequenced and genotyped on 1 or more platforms to ensure consistent genotype calls. Genotyping was assessed and verified for accuracy across laboratories.

Statistical analyses

Statistical analyses were conducted using SAS, version 9 (SAS Institute Inc., Cary, North Carolina) and the haplo.stat 1.3.8 and metaplot packages in R, version 2.7.1 (R Foundation for Statistical Computing, Vienna, Austria). Risk estimates were obtained from unconditional logistic regression models, adjusted for age, sex, and study center. Heterogeneity of effects across studies was determined using a logit model in PROC GLIMMIX, where study center was included as a random variable. The contribution of individual studies to the overall effect estimates was assessed in sensitivity analyses. For each polymorphism, genotype was coded as an ordinal variable based on the number of variant alleles (0, 1, 2) and also as a dichotomous variable, assuming dominant inheritance (heterozygous/homozygous variant vs. homozygous wild-type genotypes). The most common homozygous genotype among control participants was used as the referent group. Tests for linear trend in risk estimates were based on the Wald chi-squared statistic from adjusted logistic regression models, where genotype was included as an ordinal variable. Adjusted polytomous generalized logit regression models for categorical outcomes were used to determine heterogeneity of SNP effects across the DLBCL and follicular lymphoma subtypes.

Haplotype analyses of TNF and IL10 genes were conducted using the R package haplo.stat (19). Gene-gene interaction between TNF and each of the IL10 SNPs was assessed by comparing the difference in the −2 log-likelihoods from the main-effects and interaction models to a chi-squared statistic with degrees of freedom equal to the difference in the number of terms between the nested models. We also analyzed categorical variables of the joint TNF and IL10 SNP genotypes grouped, where homozygous wild types for both SNPs were the referent group.

The large size of the pooled sample also allowed for exploratory analyses of genotype effects for NHL and common NHL subtypes (DLBCL and follicular lymphoma) among blacks/African Americans (NCI/SEER, Yale, and UCSF studies), Asians (British Columbia, NCI/SEER, New South Wales, and UCSF studies), and Hispanic whites (EpiLymph Spain, NCI/SEER, Yale, and UCSF studies). Of a total of 16,451 participants (7,999 cases and 8,452 controls), 14,498 HIV-negative non-Hispanic whites (7,034 cases and 7,464 controls), representing 88% of the total study population, were included in the overall analyses; of these, 6,601 (45.5%) had not been included in the earlier analyses (7). Investigators from several studies included in the initial pooled report (7) subsequently published additional study-specific findings (20, 21) that were included here. Data from the New South Wales and Mayo Clinic studies that were not included in the report were subsequently published separately (22, 23).

Participants at each study site provided written informed consent prior to interview and biospecimen collection. Procedures and protocols were approved by committees for the protection of human subjects at each institution.

Quality control assessment

To ensure that genotyping results were consistent across studies, each laboratory analyzed the same set of DNA samples from 102 ethnically diverse individuals as had previously been sequenced and genotyped for all SNPs on 1 or more platforms as part of the SNP500Cancer project (http://snp500cancer.nci.nih.gov) (24). The samples were obtained from the Coriell Institute for Medical Research (Camden, New Jersey; http://www.coriell.org/). We assessed concordance across laboratories and rechecked quality control data for assays not in Hardy-Weinberg equilibrium (defined as P < 0.05) to confirm accuracy.

RESULTS

Table 1 shows characteristics of the individual studies that provided data for this project. This includes studies that contributed data to the initial InterLymph pooled analysis (7) and studies contributing new data (8,847 participants) to the current pooled analysis that comprised the validation study population.

Validation study population

These analyses included 6,601 HIV-negative non-Hispanic white participants (3,278 cases and 3,323 controls) from 8 study centers (EpiLymph studies in France, Italy, Ireland, and the Czech Republic; New South Wales; British Columbia; Mayo Clinic; and UCSF2) who were not included in the initial pooled analyses (7). Results from analyses restricted to these newly included cases and controls showed increasing risks of NHL and DLBCL among carriers of the TNF –308AA genotype (odds ratio (OR) = 1.69 (Ptrend = 0.001) and OR = 1.58 (Ptrend = 0.004), respectively; Table 2). Although the TNF –308AA genotype also was associated with an increased risk of follicular lymphoma (OR = 1.69, 95% confidence interval (CI): 1.09, 2.60), no effect on risk was observed for the –308GA genotype, nor was there a significant risk increase by allele (Ptrend = 0.84). We also found that the LTA 252AG genotype was associated with an increased risk of DLBCL (OR = 1.22, 95% CI: 1.02, 1.46; Ptrend = 0.04). IL10 –1082A>G and –3575T>A SNP frequencies were similar between cases and controls. The positive associations between the TNF –308AA and LTA 252GG genotypes and risk of DLBCL are consistent with our previous report (7) (TNF –308AA: OR = 1.65, 95% CI: 1.16, 2.34; LTA 252GG: OR = 1.47, 95% CI: 1.18, 1.84) and justified pooling of the new and previous data to increase precision and statistical power, especially for analyses of rarer subtypes and smaller ethnic subgroups.

Table 2.
Adjusted Odds Ratios in Validation Studies for Risk of All Non-Hodgkin Lymphoma, Diffuse Large B-Cell Lymphoma, and Follicular Lymphoma Associated With Polymorphisms in the TNF and LTA Genes, International Lymphoma Epidemiology Consortium, 1989–2005 ...

Population demographic characteristics and NHL subtypes

Of the combined 14,498 HIV-negative non-Hispanic white participants, ages at diagnosis ranged from 18 years to 89 years, with a median age at diagnosis of 60 years (interquartile range, 19 years). There were slightly more men (52%) than women (48%) in the pooled case group. Per the InterLymph pathology group classification of NHL subtypes (18), 77% of cases were of B-cell lineage (of these, 34% were DLBCL, 26% were follicular lymphoma, and 16% were CLL/SLL), 5% of lymphomas were of T-cell lineage (of these, 35% were mycosis fungoides and 58% were peripheral T-cell lymphoma), and 18% were unclassified.

SNP and haplotype results among non-Hispanic whites

TNF –308G>A.

For each study center, all SNPs were in Hardy-Weinberg equilibrium in the non-Hispanic white controls (P ≥ 0.01). Increased NHL risk was observed for carriers of the variant allele for TNF –308G>A per allele, OR = 1.13; Ptrend = 0.0001 (see Web Table 1, which is posted on the Journal’s Web site (http://aje.oxfordjournals.org/))). Consistent with TNF –308G>A results for all NHL, increasing risks also were observed for B-cell NHL (per allele, OR = 1.12; Ptrend = 0.0007), DLBCL (per allele, OR = 1.25; Ptrend = 3.7 × 10−6), and other MZL (per allele, OR = 1.35; Ptrend = 0.004). When study center was modeled as a random variable, the EpiLymph France, EpiLymph Italy, NCI/SEER, and UCSF1 study effects were significantly different from the average effect across studies. However, exclusion of data from these studies resulted in small changes in the odds ratio estimates obtained from the adjusted models with study center included as a fixed-effect variable. Figure 1 shows the study-specific and pooled adjusted risk estimates from fixed-effect models for TNF –308G>A for all NHL and DLBCL. No associations with TNF –308G>A were observed with CLL/SLL or with lymphomas broadly grouped as being of T-cell lineage. However, results from analyses of specific T-cell subtypes showed some evidence of an increasing risk of mycosis fungoides with TNF –308A (per allele, OR = 1.46; P = 0.026). Furthermore, polytomous analyses showed that lymphoma risk for persons with the TNF –308 A allele differed between DLBCL patients and follicular lymphoma patients (P = 0.0002).

Figure 1.
Odds ratios for the risks of A) all non-Hodgkin lymphoma and B) diffuse large B-cell lymphoma associated with tumor necrosis factor (TNF) –308G>A genotypes in individual studies (squares) and a pooled analysis (diamonds), International ...

LTA 252A>G, IL10 –1082A>G, IL10 –3575T>A, and NOD2 3020insC.

Although there was no association between risk of all NHL and polymorphisms in LTA 252A>G, IL10 –1082A>G, IL10 –3575T>A, or NOD2 3020insC, there was evidence of associations with specific NHL subtypes (Web Table 1 (http://aje.oxfordjournals.org/)). Results showed increased DLBCL risk for carriers of variant alleles for LTA 252A>G (per allele, OR = 1.12; Ptrend = 0.006) and IL10 –3575T>A (per allele, OR = 1.10; Ptrend = 0.02). In addition, increasing risk estimates were observed for mantle cell lymphoma with IL10 –1082A>G (per allele, OR = 1.30; Ptrend = 0.04), where risk was increased 67% among persons with the GG genotype, and for extranodal MZL, where risk was increased 88% among persons with the IL10 –3575AA genotype (per allele, OR = 1.30). Furthermore, the LTA 252GG genotype was associated with mycosis fungoides (per allele, OR = 1.44; Ptrend = 0.015), where risk was increased more than 2-fold. A nearly 9-fold increased risk of MZL among persons who were homozygous-variant for NOD2 3020insC also was observed (OR = 8.82, 95% CI: 2.33, 33), but results were based on few homozygous insC cases and controls (per allele, OR = 1.49; Ptrend = 0.03). No associations were found between TNF, IL10, and NOD2 polymorphisms and CLL/SLL risk. In polytomous analyses, lymphoma risk differed between DLBCL and follicular lymphoma patients among carriers of the variant allele for LTA 252A>G (P = 0.002) but not for IL10 –3575T>A (P = 0.24).

Haplotypes.

LTA/TNF haplotype analyses showed that the most common haplotype (67%) in the study population included the wild-type allele (AG) for both SNPs (Table 3). The adjusted overall global score statistic indicated that haplotype frequencies differed between NHL cases and control participants (P = 4.1 × 10−4). Most pronounced were results for DLBCL, where we observed a 31% increased risk of DLBCL among carriers of both variant alleles (GA) as compared with AG carriers (Table 3; adjusted P = 2.9 × 10−8). Haplotype analyses of the IL10 –1082A>G and IL10 –3575T>A SNPs showed that the most common haplotype (52%) was that with the wild-type alleles (AT) of both SNPs (Table 3). Haplotype frequencies differed between NHL cases and controls (adjusted global score statistic: P = 0.04). Carriers of the IL10 GA haplotype versus the AT haplotype had an increased risk of DLBCL (OR = 1.14, 95% CI: 1.06, 1.23; P = 7.2 × 10−4).

Table 3.
Adjusted Odds Ratios for Risk of All Non-Hodgkin Lymphoma and Diffuse Large B-Cell Lymphoma Associated With LTA 252A>G/TNF –308G>A and IL10-1082A>G/IL10-3575T>A Haplotypes, International Lymphoma Epidemiology Consortium, ...

Analyses of ethnic subgroups

There were nearly 600 Asian participants (289 cases and 286 controls from the New South Wales, British Columbia, NCI/SEER, and UCSF studies), more than 350 black/African-American participants (140 cases and 227 controls from the NCI/SEER, Yale, and UCSF studies), and more than 350 white participants of Hispanic/Latino ethnicity (172 cases and 206 controls from the EpiLymph Spain, NCI/SEER, Yale, and UCSF studies) in the analysis, affording us an opportunity to explore associations between the variants under study and NHL (overall and the most common subtypes) within these ethnic groups. Exploratory analyses (see Web Tables 2–4 (http://aje.oxfordjournals.org/)) suggested that among black/African-American participants, the TNF-308A allele conferred a nearly 3-fold increased risk of DLBCL (Ptrend = 0.02) (Web Table 2). Among Hispanic white participants, exploratory analyses showed a nearly 2-fold increased risk of B-cell lymphoma (Web Table 3), largely for the DLBCL subtype, among persons who carried either IL10 variant allele. In contrast, reduced risk of B-cell lymphoma was observed among Asian participants with the TNF-308A allele (Web Table 4).

DISCUSSION

This study provides further evidence that variant alleles in LTA 252A>G, IL10 –3575T>A, and (particularly) TNF –308G>A are associated with an increased risk of DLBCL in non-Hispanic white populations, as indicated in a previous analysis by the InterLymph Consortium (7). By pooling all of the cases and controls in the largest genetic study of NHL to date, we were able to investigate the influence of these SNPs on rarer subtypes of NHL. Here, we found that the TNF –308A allele was positively associated with risks of MZL and mycosis fungoides in non-Hispanic whites. These findings suggest a potentially common proinflammatory etiology in the pathogenesis of DLBCL, MZL, and mycosis fungoides. Moreover, in analyses of other ethnic groups, the TNF –308A allele conferred an increased risk of DLBCL in blacks and reduced risks of all NHL and B-cell lymphoma in Asians. Among Hispanic white participants, a nearly 2-fold increased risk of B-cell lymphoma also was observed, particularly for DLBCL, in persons who carried the IL10 –1082A>G or IL10 –3575T>A variant alleles. This study highlights the effect of genetic variation in the TNF, LTA, IL10, and NOD2 genes on risk of DLBCL and rare subtypes of lymphoma, as well as risk of B-cell lymphoma in other ethnic groups.

Because of linkage disequilibrium, we were unable to unambiguously determine whether the LTA and TNF SNPs exerted independent effects on NHL risk (linkage disequilibrium: D = 1.0, r2 = 0.38). However, in haplotype analyses, we found that persons with the GA haplotype (variant alleles for LTA 252A>G and TNF-308G>A) had an increased risk of DLBCL (P = 2.9 × 10−8; Table 3), whereas the AA haplotype (wild type for LTA and variant for TNF) did not influence DLBCL risk. This suggests an independent effect of the TNF-308G>A SNP on DLBCL risk; functional studies will be needed to dissect the effects of these variants. Because the TNF/LTA locus lies in the major histocompatibility complex class III region and is in long-range linkage disequilibrium with a highly conserved ancestral haplotype, these alleles may be linked to causal major histocompatibility complex SNPs or allelotypes that are part of an extended haplotype. Studies show that Caucasian populations carry an 8.1 ancestral haplotype containing the TNF –308A allele (*2), HLA-A1-B8-TNF-308*2-DR3-DQ2 (25), that has been implicated in susceptibility to autoimmune diseases, including rheumatoid arthritis, Crohn's disease, and systemic lupus erythematosus (26, 27). In contrast, the TNF –308A allele occurs as part of a 58.1 ancestral haplotype, HLA-A33-B58-TNF-308*2-DR3, that is found in Asian populations (28). Further studies will be needed to determine whether these extended haplotypes might explain the differences observed between the TNF-308A allele and risk of lymphoma in Caucasians and Asians. Moreover, fine mapping and functional studies of SNPs across this region will be required to determine whether the TNF –308G>A and LTA 252A>G SNPs constitute distinct susceptibility alleles or whether they are linked to other causal major histocompatibility complex loci.

The TNF –308A allele enhances transcriptional activation (29, 30) and TNF-α production (30) and has been implicated as a pathogenic mediator of numerous cancers (31). As a major proinflammatory cytokine, TNF-α may function as a bridge between inflammation and cancer through its roles in promoting cell transformation, proliferation, survival, invasion, and angiogenesis. Our findings suggest that this proinflammatory mechanism may provide a T-cell-fueled proinflammatory milieu that may be involved in the pathogenesis of DLBCL, and possibly MZL and mycosis fungoides, although validation studies of this association will be required for these rare NHL subtypes. Interleukin-10 plays an antiinflammatory role by down-regulating proinflammatory cytokines, including TNF-α (32). Because the IL10 –3575A allele results in decreased interleukin-10 production (33), this allele also may provide a proinflammatory milieu that can favor neoplastic transformation to DLBCL and MZL, particularly extranodal MZL. Based on findings from a recent genome-wide association study of NHL, genetic variation in the class II major histocompatibility complex is associated with risk of follicular lymphoma but not DLBCL (34). Furthermore, sustained B-lymphocyte activation and proliferation may be involved in the pathogenesis of follicular lymphoma and not DLBCL. This is based on genetic association studies in which gene variants that augment humoral immune T-helper cell 2 responses and dampen inflammatory T-helper cell 1 responses are associated with increased risk of follicular lymphoma but not DLBCL (3).

Given the low frequency of the NOD2 mutation in our non-Hispanic white controls, a large consortial study such as this was required to detect the association between the homozygous variant genotype and NHL and MZL risk. NOD2 plays an important role in the inflammatory response (35). Gastric MZL is associated with Helicobacter pylori infection and chronic inflammation. Thus, the NOD2 mutation may provide a proinflammatory milieu that promotes MZL. Because this is a relatively rare mutation, the results of our analysis should be interpreted with caution.

Results from this large pooled study allowed us to test a priori hypotheses to confirm and validate our earlier findings. The larger sample size also afforded us a unique opportunity to explore associations within rare subtypes and by ethnicity. However, underlying population stratification, even in relatively homogeneous populations, may influence results. Restricting the data to non-Hispanic Caucasians and adjusting for study site in our analyses helped us diminish possible effects due to underlying population stratification. In addition, we tested multiple hypotheses in our analyses by ethnicity, with some analyses being constrained by small sample sizes. Thus, our exploratory results should be interpreted conservatively and warrant further investigation and validation.

This study provides novel insights into mechanisms of lymphomagenesis involving signaling pathways that affect inflammation and immunity by regulating nuclear factor-κB-mediated processes. These pathways contribute to the development, survival, and attenuation of specific subtypes of NHL. The findings also underline the importance of large, pooled consortial studies in generating robust genetic association results. Continued efforts to dissect linkage disequilibrium patterns, particularly between TNF and the major histocompatibility complex region, and potential risk differences by ethnicity will be critical for pinpointing the causal variant(s) in NHL etiology. Finally, the InterLymph consortium is now preparing to carry out studies of gene-environment interactions, with a particular focus on SNPs in the TNF/LTA and IL10 genes and environmental exposures (including obesity, sun exposure, and history of autoimmune disease) that have shown an association in previous InterLymph pooled analyses (3638). This ongoing, well-powered consortial effort is likely to further our understanding of the fundamental processes that contribute to lymphomagenesis.

Supplementary Material

[Web Tables 1-4]

Acknowledgments

Author affiliations: Division of Environmental Health Sciences, School of Public Health, University of California, Berkeley, Berkeley, California (Christine F. Skibola, Danica R. Skibola, Martyn T. Smith, Luoping Zhang, Renee Shiao); Department of Epidemiology and Biostatistics, School of Medicine, University of California, San Francisco, San Francisco, California (Paige M. Bracci, Elizabeth A. Holly); Division of Cancer Epidemiology, German Cancer Research Center, Heidelberg, Germany (Alexandra Nieters, Nikolaus Becker); British Columbia Cancer Research Center, Vancouver, British Columbia, Canada (Angela Brooks-Wilson, John J. Spinelli); Department of Biomedical Physiology and Kinesiology, Faculty of Science, Simon Fraser University, Vancouver, British Columbia, Canada (Angela Brooks-Wilson); Unit of Infections and Cancer, Cancer Epidemiology Research Programme, Institut Català d’ Oncologia, Institut d'Investigació Biomèdica de Bellvitge, Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública, Barcelona, Spain (Silvia de Sanjosé, Yolanda Benavente); Department of Health Sciences Research, Mayo Clinic College of Medicine, Rochester, Minnesota (James R. Cerhan, Susan L. Slager, Anne J. Novak); Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland (Qing Lan, Mark Purdue, Anneclaire J. De Roos, Patricia Hartge, Martha S. Linet, Stephen J. Chanock, Nathaniel Rothman); Department of Family Medicine and Public Health Sciences, School of Medicine, Wayne State University, Detroit, Michigan (Maryjean Schenk); Karmanos Cancer Institute, Wayne State University, Detroit, Michigan (Maryjean Schenk); Epidemiology and Genetics Unit, Department of Health Sciences, University of York, York, United Kingdom (Eleanor Kane, Eve Roman, Tracy Lightfoot, Alex Smith); Norris Comprehensive Cancer Center, Keck School of Medicine, University of Southern California, Los Angeles, California (Wendy Cozen); Department of Epidemiology and Public Health, School of Medicine, Yale University, New Haven, Connecticut (Qing Lan, Tongzhang Zheng, Yawei Zhang, Theodore R. Holford); Department of Cancer Epidemiology and Genetics, Masaryk Memorial Cancer Institute, Brno, Czech Republic (Lenka Foretova); International Agency for Research on Cancer, Lyon, France (Paolo Boffetta, Paul Brennan); School of Public Health, University of Sydney, Sydney, New South Wales, Australia (Anne Kricker, Bruce Armstrong, Ann Maree Hughes); Department of Public Health, Occupational Health Section, Institute of Occupational Medicine, University of Cagliari, Cagliari, Italy (Pierluigi Cocco); Registry of Hematological Malignancies, Burgundy University and University Hospital, Dijon, France (Marc Maynadie); School of Nursing, Dublin City University, Dublin, Ireland (Anthony Staines); and Division of Etiology, Department of Population Sciences, City of Hope, Duarte, California (Sophia S. Wang).

Drs. Christine F. Skibola and Paige M. Bracci are joint first authors of this article.

This work was supported by National Institutes of Health (NIH) grant CA104862 (C. F. Skibola, Principal Investigator (PI), University of California, San Francisco (UCSF)/University of California, Berkeley (UCB), studies); NIH grants CA45614, CA89745, and CA87014 (E. A. Holly, PI, UCSF/UCB studies); the National Cancer Institute of Canada, the Canadian Institutes of Health Research, and the Michael Smith Foundation for Health Research (British Columbia study); the Intramural Research Program of the NIH (P. Hartge, PI, National Cancer Institute (NCI)/Surveillance, Epidemiology, and End Results study); the European Commission (grant QLK4-CT-2000-00422, EpiLymph study); NIH contract NO1-CO-12400 (EpiLymph Italy study); Fondazione Cariverona (2004: A. Scarpa, PI, EpiLymph Italy study; 2005: P. S. Moore, PI, EpiLymph Italy study); Compagnia di San Paolo—Programma Oncologia (P. Cocco, PI, EpiLymph Italy study); José Carreras Leukemia Foundation grant DJCLS-R04/08 (A. Nieters, PI, EpiLymph Germany study); Federal Office for Radiation Protection grants StSch4261 and StSch4420 (N. Becker, PI, EpiLymph Germany study); Fondo Investigaciones Sanitarias grant PI040091 (S. de Sanjose, PI, EpiLymph Spain study); Network for Research in Epidemiology and Public Health grant 03/09 and Fondo Investigaciones Sanitarias grant PI041467 (R. Bosch, PI, EpiLymph Spain study); Leukaemia Research Fund (E. Roman, PI, United Kingdom study); the National Health and Medical Research Council of Australia, the Cancer Council New South Wales, and the University of Sydney Medical Foundation Program (B. Armstrong, PI, New South Wales study); the Intramural Research Program of the NCI (New South Wales study); NCI grant CA62 006 and the Intramural Research Program of the NCI (T. Zheng, PI, Yale study); NIH grant CA92153 (J. Cerhan, PI, Mayo Clinic study); the Health Research Board, Ireland (EpiLymph); and Cancer Research, Ireland (InterLymph) (A. Staines, PI, EpiLymph Ireland study).

Conflict of interest: none declared.

Glossary

Abbreviations

CI
confidence interval
CLL
chronic lymphocytic leukemia
DLBCL
diffuse large B-cell lymphoma
HIV
human immunodeficiency virus
IL10
interleukin-10
LTA
lymphotoxin-α
MZL
marginal zone lymphoma
NCI
National Cancer Institute
NOD2
nucleotide-binding oligomerization domain containing 2
OR
odds ratio
SEER
Surveillance, Epidemiology, and End Results
SLL
small lymphocytic lymphoma
SNP
single nucleotide polymorphism
TNF
tumor necrosis factor
UCSF
University of California, San Francisco

References

1. Altieri A, Bermejo JL, Hemminki K. Familial risk for non-Hodgkin lymphoma and other lymphoproliferative malignancies by histopathologic subtype: the Swedish Family-Cancer Database. Blood. 2005;106(2):668–672. [PubMed]
2. Wang SS, Slager SL, Brennan P, et al. Family history of hematopoietic malignancies and risk of non-Hodgkin lymphoma (NHL): a pooled analysis of 10 211 cases and 11 905 controls from the International Lymphoma Epidemiology Consortium (InterLymph) Blood. 2007;109(8):3479–3488. [PubMed]
3. Skibola CF, Curry JD, Nieters A. Genetic susceptibility to lymphoma. Haematologica. 2007;92(7):960–969. [PMC free article] [PubMed]
4. Dave SS, Wright G, Tan B, et al. Prediction of survival in follicular lymphoma based on molecular features of tumor-infiltrating immune cells. N Engl J Med. 2004;351(21):2159–2169. [PubMed]
5. Dave SS, Fu K, Wright GW, et al. Molecular diagnosis of Burkitt's lymphoma. N Engl J Med. 2006;354(23):2431–2442. [PubMed]
6. Monti S, Savage KJ, Kutok JL, et al. Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood. 2005;105(5):1851–1861. [PubMed]
7. Rothman N, Skibola CF, Wang SS, et al. Genetic variation in TNF and IL10 and risk of non-Hodgkin lymphoma: a report from the InterLymph Consortium. Lancet Oncol. 2006;7(1):27–38. [PubMed]
8. Willett EV, Skibola CF, Adamson P, et al. Non-Hodgkin's lymphoma, obesity and energy homeostasis polymorphisms. Br J Cancer. 2005;93(7):811–816. [PMC free article] [PubMed]
9. Zhang Y, Lan Q, Rothman N, et al. A putative exonic splicing polymorphism in the BCL6 gene and the risk of non-Hodgkin lymphoma. J Natl Cancer Inst. 2005;97(21):1616–1618. [PubMed]
10. Chatterjee N, Hartge P, Cerhan JR, et al. Risk of non-Hodgkin's lymphoma and family history of lymphatic, hematologic, and other cancers. Cancer Epidemiol Biomarkers Prev. 2004;13(9):1415–1421. [PubMed]
11. Cerhan JR, Ansell SM, Fredericksen ZS, et al. Genetic variation in 1253 immune and inflammation genes and risk of non-Hodgkin lymphoma. Blood. 2007;110(13):4455–4463. [PubMed]
12. Hughes AM, Armstrong BK, Vajdic CM, et al. Sun exposure may protect against non-Hodgkin lymphoma: a case-control study. Int J Cancer. 2004;112(5):865–871. [PubMed]
13. Spinelli JJ, Ng CH, Weber JP, et al. Organochlorines and risk of non-Hodgkin lymphoma. Int J Cancer. 2007;121(12):2767–2775. [PubMed]
14. Holly EA, Bracci PM. Population-based study of non-Hodgkin lymphoma, histology, and medical history among human immunodeficiency virus-negative participants in San Francisco. Am J Epidemiol. 2003;158(4):316–327. [PubMed]
15. Skibola CF, Bracci PM, Halperin E, et al. Polymorphisms in the estrogen receptor 1 and vitamin C and matrix metalloproteinase gene families are associated with susceptibility to lymphoma [electronic article] PLoS One. 2008;3(7):e2816. [PMC free article] [PubMed]
16. Skibola CF, Holly EA, Forrest MS, et al. Body mass index, leptin and leptin receptor polymorphisms, and non-Hodgkin lymphoma. Cancer Epidemiol Biomarkers Prev. 2004;13(5):779–786. [PubMed]
17. de Sanjosé S, Benavente Y, Nieters A, et al. Association between personal use of hair dyes and lymphoid neoplasms in Europe. Am J Epidemiol. 2006;164(1):47–55. [PubMed]
18. Morton LM, Turner JJ, Cerhan JR, et al. Proposed classification of lymphoid neoplasms for epidemiologic research from the Pathology Working Group of the International Lymphoma Epidemiology Consortium (InterLymph) Blood. 2007;110(2):695–708. [PubMed]
19. Schaid DJ. Relative efficiency of ambiguous vs. directly measured haplotype frequencies. Genet Epidemiol. 2002;23(4):426–443. [PubMed]
20. Wang SS, Cerhan JR, Hartge P, et al. Common genetic variants in proinflammatory and other immunoregulatory genes and risk for non-Hodgkin lymphoma. Cancer Res. 2006;66(19):9771–9780. [PubMed]
21. Forrest MS, Skibola CF, Lightfoot TJ, et al. Polymorphisms in innate immunity genes and risk of non-Hodgkin lymphoma. Br J Haematol. 2006;134(2):180–183. [PubMed]
22. Purdue MP, Lan Q, Kricker A, et al. Polymorphisms in immune function genes and risk of non-Hodgkin lymphoma: findings from the New South Wales Non-Hodgkin Lymphoma Study. Carcinogenesis. 2007;28(3):704–712. [PubMed]
23. Cerhan JR, Liu-Mares W, Fredericksen ZS, et al. Genetic variation in tumor necrosis factor and the nuclear factor-κB canonical pathway and risk of non-Hodgkin's lymphoma. Cancer Epidemiol Biomarkers Prev. 2008;17(11):3161–3169. [PMC free article] [PubMed]
24. Packer BR, Yeager M, Burdett L, et al. SNP500Cancer: a public resource for sequence validation, assay development, and frequency analysis for genetic variation in candidate genes. Nucleic Acids Res. 2006;34(database issue):D617–D621. [PMC free article] [PubMed]
25. Jacob CO, Fronek Z, Lewis GD, et al. Heritable major histocompatibility complex class II-associated differences in production of tumor necrosis factor α: relevance to genetic predisposition to systemic lupus erythematosus. Proc Natl Acad Sci U S A. 1990;87(3):1233–1237. [PubMed]
26. Schotte H, Willeke P, Tidow N, et al. Extended haplotype analysis reveals an association of TNF polymorphisms with susceptibility to systemic lupus erythematosus beyond HLA-DR3. Scand J Rheumatol. 2005;34(2):114–121. [PubMed]
27. Newton JL, Harney SM, Timms AE, et al. Dissection of class III major histocompatibility complex haplotypes associated with rheumatoid arthritis. Arthritis Rheum. 2004;50(7):2122–2129. [PubMed]
28. Price P, Bolitho P, Jaye A, et al. A Gambian TNF haplotype matches the European HLA-A1,B8,DR3 and Chinese HLA-A33,B58,DR3 haplotypes. Tissue Antigens. 2003;62(1):72–75. [PubMed]
29. Wilson AG, Symons JA, McDowell TL, et al. Effects of a polymorphism in the human tumor necrosis factor α promoter on transcriptional activation. Proc Natl Acad Sci U S A. 1997;94(7):3195–3199. [PubMed]
30. Abraham LJ, Kroeger KM. Impact of the –308 TNF promoter polymorphism on the transcriptional regulation of the TNF gene: relevance to disease. J Leukoc Biol. 1999;66(4):562–566. [PubMed]
31. Ho SY, Wang YJ, Chen HL, et al. Increased risk of developing hepatocellular carcinoma associated with carriage of the TNF2 allele of the –308 tumor necrosis factor-α promoter gene. Cancer Causes Control. 2004;15(7):657–663. [PubMed]
32. Katsikis PD, Chu CQ, Brennan FM, et al. Immunoregulatory role of interleukin 10 in rheumatoid arthritis. J Exp Med. 1994;179(5):1517–1527. [PMC free article] [PubMed]
33. Turner DM, Williams DM, Sankaran D, et al. An investigation of polymorphism in the interleukin-10 gene promoter. Eur J Immunogenet. 1997;24(1):1–8. [PubMed]
34. Skibola CF, Bracci PM, Halperin E, et al. Genetic variants at 6p21.33 are associated with susceptibility to follicular lymphoma. Nat Genet. 2009;41(8):873–875. [PMC free article] [PubMed]
35. Ogura Y, Bonen DK, Inohara N, et al. A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature. 2001;411(6837):603–606. [PubMed]
36. Willett EV, Morton LM, Hartge P, et al. Non-Hodgkin lymphoma and obesity: a pooled analysis from the InterLymph Consortium. Int J Cancer. 2008;122(9):2062–2070. [PubMed]
37. Kricker A, Armstrong BK, Hughes AM, et al. Personal sun exposure and risk of non Hodgkin lymphoma: a pooled analysis from the Interlymph Consortium. Int J Cancer. 2008;122(1):144–154. [PubMed]
38. Ekström Smedby K, Vajdic CM, Falster M, et al. Autoimmune disorders and risk of non-Hodgkin lymphoma subtypes: a pooled analysis within the InterLymph Consortium. Blood. 2008;111(8):4029–4038. [PubMed]

Articles from American Journal of Epidemiology are provided here courtesy of Oxford University Press