Skip Navigation


American Journal of Epidemiology Advance Access originally published online on August 24, 2006
American Journal of Epidemiology 2006 164(10):984-989; doi:10.1093/aje/kwj294
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
164/10/984    most recent
kwj294v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (6)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Liu, X.
Right arrow Articles by Witte, J. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, X.
Right arrow Articles by Witte, J. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Copyright © 2006 by the Johns Hopkins Bloomberg School of Public Health All rights reserved; printed in U.S.A.

Original Contribution

Nonsteroidal Antiinflammatory Drugs and Decreased Risk of Advanced Prostate Cancer: Modification by Lymphotoxin Alpha

Xin Liu1, Sarah J. Plummer2, Nora L. Nock3, Graham Casey2 and John S. Witte1

1 Department of Epidemiology and Biostatistics and Center for Human Genetics, University of California, San Francisco, San Francisco, CA
2 Department of Cancer Biology, Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, OH
3 Department of Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, OH

Correspondence to Dr. John S. Witte, Department of Epidemiology and Biostatistics and Center for Human Genetics, University of California, San Francisco, San Francisco, CA 94143-0794 (wittej{at}humgen.ucsf.edu).

Received for publication November 23, 2005. Accepted for publication April 13, 2006.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
The potentially protective effect of nonsteroidal antiinflammatory drugs (NSAIDs) on prostate cancer may only exist among certain subgroups of men, such as those with particular variants in inflammatory response genes. To investigate this, the authors undertook a case-control study (n = 1,012) of the association between NSAIDs and more advanced prostate cancer in Ohio men recruited between 2001 and 2004 and evaluated whether this association was modified by a functional polymorphism in the lymphotoxin alpha (LTA) gene (LTA C+80A, where the CC genotype results in higher LTA production). The authors observed an inverse association between aspirin or ibuprofen use and disease (odds ratio = 0.67, 95% confidence interval: 0.52, 0.87). This was modified by the LTA C+80A variant (p for interaction = 0.03): Among men with the CC genotype, the inverse association between NSAIDs and prostate cancer was substantially stronger (odds ratio = 0.43, 95% confidence interval: 0.28, 0.67). For men without the CC genotype, NSAID use was not associated with disease (p = 0.30). The authors observed similar associations when examining dose/duration of NSAID use. This suggests that any potential chemoprevention of prostate cancer by NSAIDs may be most appropriate for men with the LTA +80CC genotype.

anti-inflammatory drugs, non-steroidal; case-control studies; lymphotoxin; prostatic neoplasms

Abbreviations: CI, confidence interval; LTA, lymphotoxin alpha; NSAID(s), nonsteroidal antiinflammatory drug(s)


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Nonsteroidal antiinflammatory drugs (NSAIDs) have been shown experimentally to induce apoptosis, inhibit angiogenesis and cell proliferation, and suppress prostate cancer metastasis (1Go–3Go). Epidemiologic findings also suggest that NSAIDs may protect against prostate cancer (4Go–6Go). However, the results of some studies have been equivocal (4Go), indicating that any protective effect of NSAIDs may only occur among certain subgroups of men (e.g., those with variants in genes involved in the inflammatory process and/or advanced disease).

Numerous proteins affect inflammation, including lymphotoxin alpha (LTA), which modulates the immune and inflammatory response to pathogens (7Go). We focus here on the variant LTA C+80A (rs2239704) because it is the main predictor of LTA production in human B cells (8Go). Specifically, among 12 common polymorphisms in LTA and the tumor necrosis factor alpha (TNFA) gene, the LTA C+80A polymorphism (and other variants in high linkage disequilibrium) was the only statistically significant predictor of LTA production. In addition, in vitro and in vivo studies showed that activated B-cell factor 1, a transcription suppressor, was preferentially recruited to the LTA +80A allele (8Go). Therefore, LTA C+80A is more likely than other variants to functionally influence LTA levels and subsequent inflammatory response to pathogens. We hypothesized that LTA C+80A might modify the protective effect of NSAID use on prostate cancer. To examine this potential relation, we evaluated the interactive effects of NSAIDs and this functional variant using a case-control population.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Study subjects
The study population comprised 1,012 men: 506 cases diagnosed with advanced prostate cancer and an equal number of age- and ethnicity-matched controls. All subjects were recruited from and frequency-matched on the major medical institutions in Cleveland, Ohio (i.e., the Cleveland Clinic Foundation, University Hospitals of Cleveland, and their affiliates) between 2001 and 2004. Physicians at these institutions see a large majority of men diagnosed with prostate cancer in the Greater Cleveland area. Hence, while the sample was not formally population-based, the cases were fairly representative of men diagnosed with prostate cancer in the Cleveland region.

The cases were newly diagnosed with histologically confirmed disease, with any one of the following: Gleason score ≥7; tumor stage ≥T2c; or a prostate-specific antigen level greater than 10 ng/ml at diagnosis. Cases were contacted as quickly as possible following diagnosis with prostate cancer (median time between diagnosis and recruitment, 4.7 months). Studying advanced cases allowed us to focus on men with the most clinically relevant disease.

To help ensure that the controls were representative of the cases' source population, we chose men who underwent standard annual medical examinations at the collaborating medical institutions. To avoid the potential for bias arising from using "hospital-based" controls, we excluded from the study men who had been diagnosed with any nonskin cancer. Since the different medical institutions may have had different populations and varying diagnostic criteria, controls were matched to cases on medical institution. To confirm that the controls were correctly classified as not having prostate cancer, we tested their prostate-specific antigen levels. Controls with prostate-specific antigen levels above 4 ng/ml were informed of this, were counseled to discuss this result with their personal physician, and if deemed appropriate, were urged to see a urologist for further evaluation. We followed up on 50 controls exhibiting high prostate-specific antigen levels. Two were ultimately diagnosed with prostate cancer; both of these men had more advanced disease and were hence reclassified as cases in our study.

Institutional review board approval was obtained from the participating medical institutions, and all study participants gave informed consent.

NSAIDs
Men were asked about the amount and duration of previous aspirin- and ibuprofen-containing drug use (prior to diagnosis for cases and prior to enrollment for controls) during an in-person computer-aided personal interview. We defined any NSAID consumption as use of either aspirin or ibuprofen at least twice a week for more than a month, and we defined never use as no reported use for both medications. For the dose/duration of use, we determined "pill-years," which is the product of number of pills taken per day and years of drug use, for aspirin and ibuprofen. The summation of pill-years of aspirin use and ibuprofen use is the dose/duration of any NSAID consumption.

Genotypes
The LTA C+80A variant was genotyped using a TaqMan single nucleotide polymorphism assay (C_2451912_1; Applied Biosystems, Foster City, California). Assays were performed using TaqMan Universal PCR Master Mix, and results were read on a 7900HT Sequence Detection System according to the manufacturer's instructions (Applied Biosystems). Instrument software (SDS 2.2) was used to analyze the data and assign genotypes (Applied Biosystems). Genotypes unassigned by TaqMan (3 percent) were determined by DNA sequencing and found to contain an adjacent rare variant (rs3093546). The genotyping percentage was greater than 99.5 percent (three samples failed to amplify when using the TaqMan assay).

Statistical analysis
We evaluated the association between NSAIDs, LTA C+80A genotypes, and more advanced prostate cancer using unconditional logistic regression that adjusted for the matching factors. We first evaluated the main effects of aspirin, ibuprofen, any NSAID consumption, and the dose/duration of use. For the latter, we compared the ensuing tertiles (based on the distribution among controls) with never use. We then examined LTA C+80A, comparing men with the CC genotype to those with CA or AA genotypes, since this reflects the expected functional activity of LTA C+80A (8Go). To investigate potential effect modification, we stratified the NSAID analyses by LTA C+80A genotype, and we tested for multiplicative interaction by including main effects and a cross-product term in a logistic regression model.

Statistical significance was assessed via both Wald test and likelihood ratio test comparing full and reduced models (i.e., with and without the cross-product term). We also conducted permutation testing to determine empirical p values for our results. In particular, case-control status was randomly permuted 10,000 times within strata defined by age group (≤60, 61–70, or ≥71 years), ethnicity, and medical institution. The proportion of test statistics larger than that observed provides an empirical p value. All p values were from two-sided tests, and all analyses were undertaken with SAS software (version 9.1; SAS Institute, Inc., Cary, North Carolina).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
A total of 834 subjects were Caucasian (82 percent), and 178 were African-American (18 percent). Per the recruitment criteria, there were no case-control differences by age, ethnicity, or institution. At diagnosis, 84 percent of the cases had a Gleason score ≥7 and 10 percent had a tumor stage ≥T2C. As expected, controls had much lower prostate-specific antigen levels than cases (mean prostate-specific antigen level = 1.7 ng/ml vs.14.2 ng/ml). Controls were more likely to have a higher educational level than cases. The LTA C+80A genotypes were in Hardy-Weinberg equilibrium in both Caucasian and African-American controls (p > 0.01).

We observed an inverse association between NSAID use and advanced prostate cancer (odds ratio = 0.67, 95 percent confidence interval (CI): 0.52, 0.87) (table 1). Increasing numbers of pill-years of NSAID use were also inversely associated with disease (table 1). Specifically, in comparison with no NSAID consumption, the first, second, and third tertiles of pill-years gave odds ratios equal to 0.70 (95 percent CI: 0.50, 0.99), 0.78 (95 percent CI: 0.56, 1.10), and 0.50 (95 percent CI: 0.35, 0.71) (p-trend = 0.0004). When we examined particular NSAIDs, similar protective effects were observed for the use of aspirin, though for ibuprofen there was only an inverse association for the highest tertile of pill-years (table 1).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Association between use of nonsteroidal antiinflammatory drugs (NSAIDs) and advanced prostate cancer in a case-control study, Cleveland, Ohio, 2001–2004

 
When we evaluated the main effect of LTA C+80A alone, there was no apparent association with advanced prostate cancer. The allele frequencies were relatively similar among cases and controls: The A allele was observed in 41 percent of Caucasian cases versus 42 percent of Caucasian controls (p = 0.89) and in 32 percent of African-American cases versus 27 percent of African-American controls (p = 0.27). In the logistic regression analyses, comparing men with the AA/CA genotype with those with the CC genotype gave an odds ratio (adjusted for age, ethnicity, and medical institution) of 1.19 (95 percent CI: 0.92, 1.54).

The LTA C+80A variant, however, substantially modified the protective effect of NSAIDs on disease (table 2). In particular, the inverse association between any NSAID use and disease was stronger for men carrying the CC genotype (odds ratio = 0.43, 95 percent CI: 0.28, 0.67) and weaker for those with the AA/AC genotypes (odds ratio = 0.84, 95 percent CI: 0.61, 1.17) (p for interaction = 0.03).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Association between use of nonsteroidal antiinflammatory drugs (NSAIDs) and advanced prostate cancer in a case-control study, by lymphotoxin alpha (LTA) C+80A genotype, Cleveland, Ohio, 2001–2004

 
The LTA C+80A genotypes also modified the inverse association between NSAID dose/duration and disease: Among men with CC genotypes, the odds ratios comparing tertiles 1–3 of NSAID pill-years with no use of NSAIDs were 0.52 (95 percent CI: 0.30, 0.92), 0.47 (95 percent CI: 0.27, 0.84), and 0.29 (95 percent CI: 0.15, 0.53), respectively (p-trend < 0.0001). In contrast, among men with AA/CA genotypes, the influence of NSAID pill-years was largely attenuated (p-trend = 0.17). The protective effect of aspirin alone was also modified by LTA C+80A (table 2). Furthermore, ibuprofen now exhibited a strong inverse association among men with the CC genotypes (p-trend = 0.004) (table 2). Restricting our analyses to Caucasians slightly strengthened our results (see table 2 footnotes). The results were not materially changed when we analyzed years of use and pills/day separately (not shown). Finally, the empirical p values for our findings were extremely similar to the asymptotic p values presented in tables 1 and 2. For example, empirical p values for the effects of interactions between LTA C+80A and any NSAIDs, aspirin, and ibuprofen on advanced prostate cancer were 0.029, 0.089, and 0.008, respectively; the corresponding p values for dose/duration were 0.023, 0.036, and 0.018, respectively.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
This relatively large study of cases with incident advanced prostate cancer and matched controls found an inverse association between NSAIDs and disease. There was no association between the functional variant LTA C+80A and advanced prostate cancer. However, the inverse association between NSAID use and advanced prostate cancer was substantially modified by LTA C+80A genotype.

The biologic mechanism underlying the potential interaction between NSAIDs and LTA C+80A remains unclear. LTA can activate nuclear factor {kappa}B through receptors shared with tumor necrosis factor alpha (9Go) and subsequently induce the expression of antiapoptotic and inflammatory genes, including cyclooxygenase 2 (9Go, 10Go). NSAIDs have been shown to regulate the inflammatory response by inhibiting cyclooxygenase 2 activity and nuclear factor {kappa}B (11Go–13Go). The higher LTA producer genotype (CC) may define a subgroup of persons for whom the antiinflammatory effect of NSAIDs is most beneficial. Specifically, LTA can promote apoptosis through induction of a caspase cascade (7Go, 9Go), and it is possible that NSAID use might enhance LTA-induced apoptosis in tumor cells through inhibition of nuclear factor {kappa}B.

If our findings are confirmed by others, this finding may have substantial public health implications because of the widespread regular use of NSAIDs (i.e., 30–40 percent among men) (14Go) and the high frequency of the LTA +80CC genotype (40 percent among our controls). On the basis of these percentages and the odds ratios observed here, population attributable risk calculations suggest that—if the effect is truly protective—up to 20 percent of advanced prostate cancers might be prevented by NSAID use among carriers of the LTA +80CC genotype. Moreover, the efficacious dose of NSAIDs may be lower for men with the LTA +80CC genotype. This is especially important in light of the numerous potential side effects of long-term prophylactic NSAID use, including gastrointestinal mucosal damage and nephrotoxicity (15Go, 16Go). Hence, it might be worthwhile for investigators conducting risk-benefit analyses of prostate cancer prevention by NSAID use to incorporate information about modification by genes.

Our study had approximately 60 percent power to detect the interaction between NSAID use and LTA genotype observed here (QUANTO, version 1.0 (17Go)). This is not overly surprising in light of the large sample sizes generally required to detect gene-environment interactions. While some of the observed interactions exhibited modest p values, they were very similar to empirical p values obtained from permutation testing. Another limitation of the current study is potential recall bias regarding NSAIDs use. The lack of general knowledge about the potentially protective effects of NSAIDs on prostate cancer should have made any such bias nondifferential among cases and controls. Moreover, we did not ask participants to specify the exact pill dose, which may also have resulted in nondifferential misclassification bias: This bias is generally expected to lead to an underestimate of the effect estimates. A related concern is that controls might not have been fully representative of the cases' source population because they were selected from men who underwent standard annual medical examinations. In comparison with the cases' source population, such controls might have cared more about their health, been more highly educated, had a higher income, or been more likely to have medical insurance. As a proxy for these potential case-control differences, we adjusted our results for education and income; doing so did not materially alter our findings (not shown). Of course, since these factors are only proxies, adjusting for them may not have fully accounted for the potential case-control differences.

Finally, in any study of genetic factors, investigators must be concerned about potential bias due to population stratification. We addressed this issue first by adjusting for ethnicity. In addition, when restricting our analysis to Caucasians—a population within which this potential bias is likely to be small (18Go)—we observed results similar to those presented here. This suggests that our findings were not due to population stratification. Furthermore, since any such bias is expected to be the same across environmental exposures, the potential for population stratification is further reduced when evaluating gene-environment interactions (18Go, 19Go).

In summary, we found that NSAID use is inversely associated with advanced prostate cancer and that this effect is substantially modified by LTA C+80A. If replicated, our findings could suggest that any potential chemoprevention of prostate cancer by NSAIDs may be most effective in men with the LTA +80CC genotype.


   ACKNOWLEDGMENTS
 
This work was supported by National Institutes of Health grants CA88164, CA94211, and CA98683.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Jones MK, Wang H, Peskar BM, et al. Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implications for cancer growth and ulcer healing. Nat Med 1999;5:1418–23.[CrossRef][Web of Science][Medline]
  2. Andrews J, Djakiew D, Krygier S, et al. Superior effectiveness of ibuprofen compared with other NSAIDs for reducing the survival of human prostate cancer cells. Cancer Chemother Pharmacol 2002;50:277–84.[CrossRef][Web of Science][Medline]
  3. Lloyd FP Jr, Slivova V, Valachovicova T, et al. Aspirin inhibits highly invasive prostate cancer cells. Int J Oncol 2003;23:1277–83.[Web of Science][Medline]
  4. Mahmud S, Franco E, Aprikian A. Prostate cancer and use of nonsteroidal anti-inflammatory drugs: systematic review and meta-analysis. Br J Cancer 2004;90:93–9.[CrossRef][Web of Science][Medline]
  5. Garcia Rodriguez LA, Gonzalez-Perez A. Inverse association between nonsteroidal anti-inflammatory drugs and prostate cancer. Cancer Epidemiol Biomarkers Prev 2004;13:649–53.[Abstract/Free Full Text]
  6. Jacobs EJ, Rodriguez C, Mondul AM, et al. A large cohort study of aspirin and other nonsteroidal anti-inflammatory drugs and prostate cancer incidence. J Natl Cancer Inst 2005;97:975–80.[Abstract/Free Full Text]
  7. Schneider K, Potter KG, Ware CF. Lymphotoxin and LIGHT signaling pathways and target genes. Immunol Rev 2004;202:49–66.[CrossRef][Web of Science][Medline]
  8. Knight JC, Keating BJ, Kwiatkowski DP. Allele-specific repression of lymphotoxin-alpha by activated B cell factor-1. Nat Genet 2004;36:394–9.[CrossRef][Web of Science][Medline]
  9. Aggarwal BB. Signaling pathways of the TNF superfamily: a double-edged sword. Nat Rev Immunol 2003;3:745–56.[CrossRef][Web of Science][Medline]
  10. Nakanishi C, Toi M. Nuclear factor-{kappa}B inhibitors as sensitizers to anticancer drugs. Nat Rev Cancer 2005;5:297–309.[CrossRef][Web of Science][Medline]
  11. Takada Y, Bhardwaj A, Potdar P, et al. Nonsteroidal anti-inflammatory agents differ in their ability to suppress NF-{kappa}B activation, inhibition of expression of cyclooxygenase-2 and cyclin D1, and abrogation of tumor cell proliferation. Oncogene 2004;23:9247–58.[Web of Science][Medline]
  12. Yin MJ, Yamamoto Y, Gaynor RB. The anti-inflammatory agents aspirin and salicylate inhibit the activity of I{kappa}B kinase-beta. Nature 1998;396:77–80.[CrossRef][Medline]
  13. Vane JR, Botting RM. The mechanism of action of aspirin. Thromb Res 2003;110:255–8.[CrossRef][Web of Science][Medline]
  14. Kim C, Beckles GL. Cardiovascular disease risk reduction in the Behavioral Risk Factor Surveillance System. Am J Prev Med 2004;27:1–7.[Medline]
  15. Slattery J, Warlow CP, Shorrock CJ, et al. Risks of gastrointestinal bleeding during secondary prevention of vascular events with aspirin—analysis of gastrointestinal bleeding during the UK-TIA trial. Gut 1995;37:509–11.[Abstract/Free Full Text]
  16. Gambaro G, Perazella MA. Adverse renal effects of anti-inflammatory agents: evaluation of selective and nonselective cyclooxygenase inhibitors. J Intern Med 2003;253:643–52.[CrossRef][Web of Science][Medline]
  17. Gauderman WJ. Sample size requirements for association studies of gene-gene interaction. Am J Epidemiol 2002;155:478–84.[Abstract/Free Full Text]
  18. Wacholder S, Rothman N, Caporaso N. Population stratification in epidemiologic studies of common genetic variants and cancer: quantification of bias. J Natl Cancer Inst 2000;92:1151–8.[Abstract/Free Full Text]
  19. Thomas DC, Witte JS. Point: population stratification: a problem for case-control studies of candidate-gene associations? Cancer Epidemiol Biomarkers Prev 2002;11:505–12.[Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
E. D. Flick, L. A. Habel, K. A. Chan, S. K. Van Den Eeden, V. P. Quinn, R. Haque, E. J. Orav, J. D. Seeger, M. C. Sadler, C. P. Quesenberry Jr., et al.
Statin Use and Risk of Prostate Cancer in the California Men's Health Study Cohort
Cancer Epidemiol. Biomarkers Prev., November 1, 2007; 16(11): 2218 - 2225.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
I. Cheng, L. M. Krumroy, S. J. Plummer, G. Casey, and J. S. Witte
MIC1 and IL1RN Genetic Variation and Advanced Prostate Cancer Risk
Cancer Epidemiol. Biomarkers Prev., June 1, 2007; 16(6): 1309 - 1311.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
164/10/984    most recent
kwj294v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (6)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Liu, X.
Right arrow Articles by Witte, J. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liu, X.
Right arrow Articles by Witte, J. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?