TLR6, TLR1, and TLR10) and Prostate Cancer

1982

Association between Toll-Like Receptor Gene Cluster (TLR6, TLR1, and TLR10) and Prostate Cancer

Yen-Ching Chen,1,4 Edward Giovannucci,1,2 Peter Kraft,3 Ross Lazarus,1 and David J. Hunter1,2,3 1Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School; 2Department of Nutrition and 3Program in Molecular and Genetic Epidemiology, Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts and 4Research Center for Genes, Environment, and Human Health, and Graduate Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan

Abstract

Background: Chronic inflammation may be a risk factor specific antigen test. We genotyped 19 common (>5%) for prostate cancer. Previously, we found significant haplotype-tagging SNPs chosen from the SNPs discov- associations between single nucleotide polymorphisms ered in a resequencing study spanning TLR6, TLR1, (SNPs) and haplotypes in Toll-like receptor (TLR) 4 and TLR10 to test for the association between sequence and the risk of prostate cancer. TLR6, TLR1, and TLR10 variants cluster and prostate cancer. are also involved in the pathogen-mediated inflamma- Results: Neither individual SNPs nor common haplo- tion pathway. A Swedish study observed associations types in the three gene regions were associated with between sequence variants in the TLR6-TLR1-TLR10 altered risk of prostate cancer or subgroups of aggres- gene cluster and the risk of prostate cancer. We sive prostate cancer. No effect modification was assessed if genetic polymorphisms of this gene cluster observed for age, body mass index, or family history were associated with the risk of prostate cancer in a of prostate cancer, except that TLR6_3649 showed U.S. population. nominally significant interaction with family history Methods: In a nested case-control design within the at the P < 0.05 level. Health Professionals Follow-Up Study, we identified Conclusion: Inherited sequence variants of the innate 700 participants with prostate cancer who were diag- immune gene cluster TLR6-TLR1-TLR10 were not nosed after they had provided a blood specimen in 1993 appreciably associated with the risk of prostate cancer and by January 31, 2000. Controls were 700 age-matched in this cohort. (Cancer Epidemiol Biomarkers Prev men without prostate cancer who had had a prostate- 2007;16(10):1982–9)

Introduction

Human Toll-like receptor (TLR) is a type I transmem- TLR1 (3). TLR6 is highly expressed in peripheral blood brane protein with an extracellular domain consisting of lymphocytes, the spleen, the thymus, the ovary, and the a leucine-rich repeat region and an intracellular domain lung (4, 5). TLR1 is expressed in peripheral blood lym- homologous to that of the Toll/interleukin-1 receptor phocytes, epithelial cells, and endothelial cells (5, 6). (1, 2). TLRs play a role in the innate immune response TLR10 is highly expressed in the spleen (7) and B cells (8) and are involved in a pathway similar to that used by the and expressed at lower levels in the lungs, the small Toll/interleukin-1 receptor. The human TLR family intestine, the stomach, the thymus, peripheral blood includes the TLR3, TLR4, TLR5, TLR2, and TLR9 lymphocytes, lymph nodes, and the tonsils (5, 7, 9). subfamilies. The TLR2subfamily is composed of TLR1, TLR1, TLR6, and TLR10 activate the nuclear factor-nB TLR2, TLR6, and TLR10 (3). Human TLR6, TLR1, and pathway (2). Additionally, TLR6 activates the c-Jun NH2- TLR10 are located on chromosome 4p spanning 54 kb terminal kinase pathway (4). TLR1 and TLR6 need to and have one, four, and three exons, respectively. TLR1 combine with TLR2to form dimers to recognize and TLR6 are close to each other because they evolved pathogens (10, 11) and produce subsequent cytokine from the same orthologous gene (3, 4). The amino acid induction (12). The TLR1-TLR2 dimers recognize pepti- sequence of human TLR6 shows 69% similarity with doglycans from Gram-positive bacteria and zymosan from yeast cell walls; the TLR2-TLR6 dimers recognize the lipoproteins from Mycoplasma (11, 12). TLR10 can

Received 4/9/07; revised 7/26/07; accepted 8/6/07. form either a homodimer or a heterodimer with TLR1 or Grant support: NIH grants UO1 CA98233 and CA55075. TLR2; its cytoplasmic regions are also different from The costs of publication of this article were defrayed in part by the payment of page those of TLR1 and TLR6 (7). Genetic variations in TLR1 charges. This article must therefore be hereby marked advertisement in accordance have been related to the risk of Lyme disease (10) and with 18 U.S.C. Section 1734 solely to indicate this fact. atherogenesis (13). TLR10 has been linked to asthma (14). Note: Supplementary data for this article are available at Cancer Epidemiology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/). Health effects related to sequence variants of TLR6 are Requests for reprints: Yen-Ching Chen, Channing Laboratory, Department of unclear. Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, 181 In TLR1-deficient mice, proinflammatory cytokine Longwood Avenue, Boston, MA 02115. Phone: 617-525-2279; Fax: 617-525-2008. E-mail: [email protected] was not produced normally by macrophages when Copyright D 2007 American Association for Cancer Research. stimulated by lipoprotein and a synthetic triacylated doi:10.1158/1055-9965.EPI-07-0325 lipopeptide (15). A murine model showed that a TLR2

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

and TLR6 heterodimer may be involved in immunomo- of prostate cancer in the Health Professionals Follow-Up dulation of macrophages mediated by an antigenic Study. SNP and haplotype analyses as well as case-only protein, rLcrV (16). The Toll/interleukin-1 receptor analyses were done to test this hypothesis. domain of TLR10 is important in activation of promoters of certain inflammation cytokines (7). Materials and Methods Chronic inflammation has been associated with some cancers (e.g., cervix, stomach, primary liver, and bladder Study Population. This is a case-control study nested cancers; refs. 17, 18). Substantial evidence, including within the ongoing Health Professionals Follow-Up studies on sexually transmitted infections, clinical Study, in which incident prostate cancer cases were prostatitis, and genetic and circulating markers of identified, with follow-up from 1986 to 2000. A total of inflammation and response to infection, supports a link 51,529 U.S. men aged 40 to 75 years were enrolled in between chronic intraprostatic inflammation and the 1986. Each participant completed a mailed questionnaire risk of prostate cancer (19). Although TLR6, TLR1, and on demographics, lifestyle, and medical history and a TLR10 are not expressed in prostate tissue, these genes semiquantitative food-frequency questionnaire at base- are involved in the signaling pathway of pathogen- line. Information on exposures and diseases was updated related innate immune responses, which may be every other year, and diet information was updated involved in prostate carcinogenesis. A study of a every 4 years. Deaths were identified through reports by Swedish population found that single nucleotide poly- family members, follow-up questionnaires, or a search of morphisms (SNP) and haplotypes in the TLR6-TLR1- the National Death Index (24). This study was approved TLR10 gene cluster were associated with the risk of by the institutional review board at the Harvard School prostate cancer (20). The same team later reported that of Public Health. Completion of the self-administered the combination of IRAK4-7987 CG/CC and the risk questionnaire was considered to imply informed consent. genotype at the TLR6-TLR1-TLR10 gene cluster was Between 1993 and 1995, blood samples from 18,018 of associated with a 9.7-fold risk of prostate cancer the participants were collected in tubes containing compared with men with wild-type genotypes in the sodium EDTA. Samples were shipped by overnight same Swedish population (21). Previous studies also courier and centrifuged; the aliquots, including plasma, showed significant associations between sequence variants erythrocytes, and buffy coat, were stored in liquid of TLR4 and prostate cancer risk (22, 23). We hypothe- nitrogen freezers. We used a Qiagen QIAamp blood sized that genetic polymorphisms of the TLR6-TLR1- extraction kit for DNA extraction. All DNA samples were TLR10 gene cluster are associated with the risk of whole genome amplified, and the quality control prostate cancer. Therefore, we explored the association samples had 100% genotype concordance rates. Among between the TLR6-TLR1-TLR10 gene cluster and the risk the men who gave a blood specimen, 95% responded to

Figure 1. TLR6-TLR1- TLR10 linkage disequilibrium plot. This linkage disequilibrium plot was generated by the Cauca- sian data from HapMap. The Gabriel et al. approach in Haploview program was used to define haplotype block and the 18 SNPs formed three blocks. The rs number on the top from right to left corre- sponds to the SNP name (Snp1, Snp2, etc.). The level of pairwise D¶,which indicates the degree of linkage disequilibrium between two SNPs, is shown in the linkage disequilibrium structure in red. Nine common haplotypes (frequency >0.05) were identified.

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

Table 1. Characteristics of TLR6, TLR1, and TLR10 SNPs among Caucasians

IIPGA SNP Nucleotide Location rs no. Minor allele HWE P value, Minor allele HWE P value, P* name name change (change of frequency, controls (%) frequency, cases (%) amino acid) controls (%) cases (%) TLR6_455 Snp1 C!G Promoter rs5743788 0.49 0.020.50 0.87 1.00 TLR6_1166 Snp2G !A Promoter rs5743795 0.21 0.43 0.19 0.27 0.91 TLR6_1894 Snp3 C!T Promoter rs5743806 0.30 0.21 0.31 0.69 1.00 TLR6_2065 Snp4 C!T Promoter rs1039599 0.46 0.01 0.46 0.10 1.00 TLR6_3311 Snp5 T!C Exon (P249S)rs5743810 0.41 0.020.420.16 1.00 TLR6_3649 Snp6 C!G Exon rs3821985 0.33 0.31 0.34 0.55 1.00 TLR6_3846 Snp7 T!C Exon (V427A) rs5743815 0.02 0.68 0.01 0.74 1.00 TLR1_1260 Snp8 A!G Promoter rs5743551 0.26 0.26 0.24 0.81 1.00 TLR1_2063 Snp9 T!C Promoter rs5743556 0.19 0.27 0.18 0.59 1.00 TLR1_7631 Snp10 T!C Intron 3 rs5743604 0.26 0.70 0.23 0.72 1.00 TLR1_8702 Snp11 C!G Exon 4 (N80T) rs5743611 0.09 0.0020.08 0.241.00 TLR1_11040 Snp12A !G3¶ rs4624663 0.04 0.03 0.04 0.33 1.00 TLR10_995 Snp13 A!G Promoter rs11466617 0.18 0.55 0.17 0.94 1.00 TLR10_2571 Snp14 C!T Promoter rs11466640 0.19 0.41 0.17 0.46 1.00 TLR10_4003 Snp15 A!G Exon 2 rs4274855 0.19 0.26 0.18 0.01 1.00 TLR10_4983 Snp16 A!C Exon 3 (N241H) rs11096957 0.36 0.20 0.33 0.26 1.00 TLR10_5367 Snp17 A!C Exon 3 (I369L) rs11096955 0.36 0.20 0.33 0.28 1.00 TLR10_5680 Snp18 C!T Exon 3 (I473T) rs11466657 0.04 0.25 0.04 0.81 1.00 TLR10_6585 Snp19 A!G Exon 3 (I775V) rs4129009 0.18 0.55 0.17 0.90 1.00

Abbreviation: IIPGA, Innate Immunity in Heart, Lung and Blood Disease-Programs for Genomic Applications. *m2 test P value for comparing minor allele frequency between cases and controls.

the year 2000 questionnaire; 18 died of prostate cancer and Snp19 are common haplotype-tagging SNPs in before the end of follow-up and were included in the TLR10, and Snp18 was forced in for comparison case series. purposes as well. Laboratory personnel were blinded to We identified 700 incident prostate cancer cases and case-control status. All case-control matched pairs were 700 controls, which were composed of 94% self-reported analyzed together using the Sequenom system. Multiplex Caucasians, 2.7% other races, and 3.5% without ethnicity PCRs were carried out to generate short PCR products data. To prevent the possible effect of population (>100 bp) containing one SNP. The details of PCR and stratification, all analyses were restricted to Caucasians matrix-assisted laser desorption ionization time-of-flight only (cases = 659, controls = 656). Each case was matched mass spectrometry are available on request. Six control with one control who was alive, had not been diagnosed DNA samples were used for optimization. One SNP in with cancer by the date of the case’s diagnosis, and had a TLR1 failed Sequenom assay design due to other SNPs prostate-specific antigen (PSA) test after the date of blood that might interfere with primer annealing or extension. draw. The latter criterion ensured that controls had the Another SNP in TLR1 was dropped after optimization opportunity to have an occult prostate cancer diagnosed. because of high discordance rates among its duplicate All controls had a PSA test within 2.5 years of the date samples. Finally, a total of 19 SNPs (Fig. 1; Table 1) were of diagnosis of their matched case. Cases and controls genotyped in seven plexes at the Harvard Partners were matched on year of birth (F1 year), PSA test before Center for Genetics and Genomics (Boston, MA). For blood draw (yes/no), and time (midnight to before 9 each SNP, genotyping data were missing in <5% of the a.m., 9 a.m. to before noon, noon to before 4 p.m., and study participants. Sixty-eight quality control samples 4 p.m. to before midnight), season (winter, spring, were obtained from 18 external participants and each of summer, and fall), and exact year of blood draw because them had two to six duplicates. These quality control plasma analyses were being done on the same case- samples were genotyped together with all other samples control set. in this study. All quality control samples passed the quality control test (discordance rate = 0). Laboratory Assays. Haplotype-tagging SNPs were chosen using resequencing data from the Innate Immu- Ascertainment of Prostate Cancer. Investigators nity in Heart, Lung and Blood Disease-Programs for reviewed the medical and pathology records for men Genomic Applications. The Innate Immunity in Heart, with prostate cancer reported from the follow-up Lung and Blood Disease-Programs for Genomic Appli- questionnaire or death certificate to confirm adenocarci- cations resequenced the TLR6, TLR1, and TLR10 genes of noma of the prostate and to document clinical presenta- 23 unrelated Europeans from Centre d’Etude du Poly- tion, stage, and Gleason sum of the tumor. Because the morphisme Humain families, including 2.5 kb 5¶ of the concordance between self-reports and cases confirmed genes, exons, and 1.5 kb 3¶ of the gene. Snp1 to Snp6 are by medical record was high (>90%), self-reported cases common haplotype-tagging SNPs in TLR6 with frequen- for which we were unable to acquire medical records cies >5%, chosen to tag haplotypes with frequencies were included. The clinical presentation was categorized >10% using the algorithm described in Sebastiani et al. into elevated serum PSA concentration only, abnormal (25). Snp7 was forced in because it is a nonsynonymous digital rectal examination with or without an elevated SNP. Snp9 and Snp10 are common haplotype-tagging serum PSA concentration, or other/unknown. The cases SNPs in TLR1; Snp8, Snp11, and Snp12were forced in for were categorized into regionally invasive or metastatic z comparison with the previous study (20). Snp13 to Snp17 (stage T3b,N1,orM1), organ-confined or minimal

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

extraprostatic extension (T1b to T3a and N0M0), higher Results grade (Gleason sum z7), and lower grade (Gleason TLR6-TLR1-TLR10 sum <7). Incidental microscopic focal tumors (stage T1a) Nineteen SNPs in were genotyped. were excluded because they are generally indolent and Among controls, Snp11 was out of HWE among controls P susceptible to detection bias due to differential rates of ( = 0.002) and was therefore dropped from all the surgery for benign prostatic hyperplasia. In addition, analyses (Table 1). Snp1, Snp4, Snp5, and Snp12were out P men with a previous cancer, except nonmelanoma skin of HWE ( = 0.02-0.03), but we retained them in the cancer, before the date of blood draw were excluded. analyses, as the less conservative Benjamini-Hochberg Confirmed non-T1a tumors between blood draw and step-up procedure to control the false discovery rate January 31, 2000 were included. In the blood subcohort, (<5%) indicated that the null hypothesis of HWE was 92% of cases were confirmed by medical records and 5% rejected only for Snp11. The internal blinded quality by other corroborating information; only 3% were based control specimens did not show evidence of genotyping on self-report (26). error. The study population included 659 incident prostate Statistical Analysis. The Hardy-Weinberg equilibrium cancer cases and 656 matched controls. Age and BMI (HWE) test was done for each SNP among controls. distributions were similar for cases and controls (Table 2). Haplotype block structure (Fig. 1) was determined by 5 6 Family history of prostate cancer was significantly using Haploview and LocusView. The expectation- different between cases and controls (P = 0.009). The maximization algorithm was applied to estimate haplo- mean age at starting smoking, lifetime average number type frequencies in each block using the tagSNP program of cigarettes/day, and alcohol consumption were similar (27). Conditional logistic regression models were used for cases and controls. Among cases, 78% were in tumor to estimate odds ratios (OR) for disease in participants carrying either one or two versus zero copies of the minor allele of each SNP and each multilocus haplotype; Table 2. Characteristics of study participants in the haplotype trend regression (28) was used to test global TLR6-TLR1-TLR10 Health Professionals Follow-Up Study among Cauca- association between haplotypes and sians prostate cancer. The type I error rate is controlled by the single multiple-degree-of-freedom test of association Variable Cases Controls between TLR6-TLR1-TLR10 haplotypes and prostate (n = 659), (n = 656), cancer. Given a significant global test, haplotype- and n (%) n (%) SNP-specific tests can provide some guidance as to Age, y (matching variable) which variant(s) contributes to the significant global V65 297 (45) 302 (46) test, although the nominal P values we present do not >65 362(55) 354 (54) control the family-wise error rate for these post hoc BMI V comparisons. 25 405 (61) 389 (59) 25-30 221 (34) 221 (34) Age and family history are known risk factors for >30 33 (5) 46 (7) prostate cancer (29, 30); body mass index (BMI) is related Family history of prostate cancer to the risk of prostate cancer, although not consistently No 528 (80) 557 (85) (31). We evaluated how these factors modified the Yes 131 (20) 99 (15) TLR6-TLR1-TLR10 Age started smoking 23.0 F 5.3 22.8 F 5.4 association between SNPs or haplo- F F types and the risk of prostate cancer by comparing a Lifetime average 10.5 6.3 11.0 6.7 cigarettes/day model with terms for main effects and interaction terms Alcohol (g/d) 11.4 F 14.9 10.5 F 14.7 to a model with terms for main effects only using the PSA at time of diagnosis 11.3 F 21.0 Not applicable likelihood ratio test. Because of the role of TLR6-TLR1- (ng/mL) (n = 479) TLR10 in the innate immune response, the aggressive- Stage ness of prostate cancer may relate to genetic variations T1b-T3a 513 (78) Not applicable in the TLR6-TLR1-TLR10 gene cluster. We tested the T3b or T4 or N1 or M1 or 55 (8) TLR6-TLR1-TLR10 death due to prostate cancer association between haplotypes and Missing 91 (14) aggressiveness among patients with prostate cancer by Gleason sum using two definitions for tumor aggressiveness (aggres- 2-4 44 (7) Not applicable siveness 1: stage T3b or T4 or N1 or M1 or death due to 5-7 484 (73) 8-10 55 (8) prostate cancer; aggressiveness 2: stage T3b or T4 or N1 or z Missing 76 (12) M1 or death due to prostate cancer or Gleason sum 7). Aggressiveness 1 Aggressiveness 1 is useful in evaluating participants No 513 (78) Not applicable lacking information for Gleason sum and indicates how Yes 55 (8) far a cancer has progressed independent of grade. Missing 91 (14) Aggressiveness 2indicates the potential of the tumor to Aggressiveness 2 progress by considering grade information. All analyses No 419 (64) Not applicable Yes 240 (36) were conducted with Statistical Analysis System release Missing 0 9.0 (SAS Institute), and all statistical tests were two sided. Death due to prostate cancer No 641 (97) Not applicable Yes 18 (3)

NOTE: Aggressiveness 1 defined as stage T3b or T4 or N1 or M1 or death 5 http://www.broad.mit.edu/mpg/haploview/index.php due to prostate cancer. Aggressiveness 2defined as stage T 3b or T4 or N1 6 z http://www.broad.mit.edu/mpg/locusview/ or M1 or death due to prostate cancer or Gleason sum 7.

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

Table 3. TLR6-TLR1-TLR10 SNP analysis by genotype among Caucasians

SNP 0 copies 1 copy 2copies P*

Case/control OR Case/control OR (95% CI) Case/control OR (95% CI) Snp1 157/179 1.00 317/287 1.26 (0.97-1.65) 156/167 1.06 (0.78-1.44) 0.19 Snp2 418/403 1.00 189/204 0.90 (0.70-1.14) 28/31 0.87 (0.51-1.47) 0.62 Snp3 307/326 1.00 277/255 1.15 (0.92-1.45) 58/63 0.98 (0.67-1.45) 0.43 Snp4 189/198 1.00 290/282 1.07 (0.83-1.39) 145/150 1.01 (0.74-1.36) 0.83 Snp5 219/236 1.00 293/281 1.13 (0.88-1.44) 123/122 1.08 (0.79-1.47) 0.64 Snp6 272/281 1.00 265/260 1.06 (0.84-1.35) 72/72 1.04 (0.72-1.50) 0.89 Snp7 629/624 1.00 14/20 0.69 (0.35-1.38) 0/0 — 0.29 Snp8 366/352 1.00 233/237 0.95 (0.75-1.19) 39/55 0.76 (0.48-1.18) 0.45 Snp9 421/421 1.00 185/187 0.99 (0.78-1.26) 23/26 0.89 (0.50-1.58) 0.92 Snp10 380/360 1.00 228/242 0.89 (0.71-1.13) 37/44 0.80 (0.51-1.27) 0.47 Snp12591/593 1.00 47/44 1.06 (0.69-1.63) 0/3 — 0.96 Snp13 435/424 1.00 180/195 0.90 (0.71-1.15) 19/19 0.98 (0.51-1.88) 0.70 Snp14 443/424 1.00 179/201 0.85 (0.67-1.09) 22/19 1.12 (0.60-2.09) 0.39 Snp15 425/413 1.00 168/184 0.89 (0.69-1.14) 31/27 1.12 (0.66-1.91) 0.54 Snp16 293/255 1.00 274/311 0.77 (0.61-0.97) 78/76 0.90 (0.63-1.28) 0.09 Snp17 291/256 1.00 274/312 0.78 (0.61-0.98) 78/76 0.91 (0.63-1.30) 0.10 Snp18 597/596 1.00 43/45 0.96 (0.62-1.48) 1/2 0.48 (0.04-5.31) 0.82 Snp19 437/425 1.00 179/195 0.89 (0.70-1.14) 19/19 0.98 (0.51-1.88) 0.66

Abbreviation: 95% CI, 95% confidence interval. P * value is for testing the null hypothesis: OR1 copy =OR2copies =1.

TLR6-TLR1-TLR10 stage T1b to T3a, 73% had Gleason grade 5 to 7, 8% had modifiers for the association between aggressive prostate cancer defined by aggressiveness 1, SNPs and prostate cancer. and 36% had aggressive prostate cancer defined by Case-only analysis (aggressive versus nonaggressive) aggressiveness 2. Eighteen cases died of prostate cancer was done among prostate cancer patients. No SNPs were before January 31, 2000. associated with tumor aggressiveness (data not shown). All the individual SNP tests for associations were null Results were not significant for haplotype analyses as (Table 3). Eighteen SNPs spanning TLR6-TLR1-TLR10 well (Table 5). formed three linkage disequilibrium blocks using a modified version of the Gabriel et al. algorithm (32), where blocks identified with the default settings in Discussion Haploview are merged if they have multiallelic D¶ >0.8, and the cumulative frequency of common (>5% frequen- Chronic intraprostatic inflammation has been reported to cy) haplotypes in the merged block is >80% (33). The first increase the risk of prostate cancer (19). TLRs play an block included seven SNPs from TLR6, which constitut- important role in pathogen-mediated innate immunity ed three common haplotypes (frequency >5%) with a and chronic inflammation. TLR6, TLR1, and TLR10 cumulative frequency of 89.9% in controls (Table 4). The belong to the TLR2 subfamily (3) and may influence P valuefortheglobaltestofthethreecommon carcinogenesis in the prostate by affecting the risk of haplotypes was 0.62. The second block included three chronic inflammation. In this study, we found no SNPs from TLR1, which constituted three common nominally significant (P < 0.05) associations with any haplotypes (frequency >5%) with a cumulative frequency haplotype-tagging SNPs or haplotypes in the TLR6- of 98.6% in controls (Table 4). The P value for the global TLR1-TLR10 gene cluster and risk of prostate cancer. test of the three common haplotypes was 0.50. Block 3 Prostate cancer family history modified the associations included eight SNPs, one SNP from TLR1 and seven between Snp6 in TLR6 and the risk of prostate cancer SNPs from TLR10, which constitute three common (P = 0.03), but considering the large number of tests we haplotypes with a cumulative frequency of 91.7% in conducted, the possibility that this result is a chance controls. The P value for the global test of the three finding cannot be excluded. common haplotypes was 0.59. Our results contrast with a Swedish study (20) that We also constructed the long-range 17-SNP haplotypes found that eight SNPs and six haplotypes in the same (without Snp4, Snp6, and Snp11) for comparison with gene cluster were significantly associated with prostate the risk haplotype in Sun et al. (20), GATCTGCCGAC- cancer risk. None of the individual SNPs that were TACCTG. Compared with the risk haplotype in Sun significantly associated with prostate cancer in the et al., ours lacked Snp11 because of its deviation from Swedish study were significantly associated in ours. HWE. We found four common haplotypes constituted Several factors could account for the disparity in results. with these SNPs had a cumulative frequency of 77.1% Although both populations are Caucasian, the Swedish (data not shown). No haplotypes were associated with one is relatively homogeneous compared with ours. The risk of prostate cancer. case mix may have been different, as the cases in our No SNPs showed nominally significant interaction population are heavily skewed toward early-stage PSA- with family history at the P < 0.01 level, and only one detected cases. The Swedish study had about twice as had a nominally significant interaction at the <0.05 level many cases (including prevalent cases), and we may (Supplementary Table S1). BMI and age were not effect have been underpowered to confirm the associations the

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

Swedish researchers observed. However, the direction of and was composed of 94% Caucasians. In addition, prostate cancer risk was opposite between the two results from inclusion of all participants and all Cauca- studies and thus statistical power may not be an issue. sians were similar. This suggests that population strati- In addition, there may have been some study design fication is not an issue in our population. For disease issues (e.g., SNP selection), although we believe that ascertainment, only 3% of cases were based on self- these were not major enough to account for the quite report. A high concordance rate (>90%) was found divergent results. Finally, if relevant, the TLR pathway between self-report and medical record–confirmed would influence prostate cancer risk through some cases. As a whole, these factors increase the validity of inflammatory pathway, and exposure to the presumed the results. An important limitation is that prostate etiologic infectious agent(s) could differ in the two cancer is a very heterogeneous cancer, and our case mix populations. Therefore, we believe that results from the was largely composed of early-stage PSA-detected Swedish study might not be generalizable to the U.S. cancers. Although we had reasonable power to examine Caucasian population. total prostate cancer as the end point, if an association Previous studies indicated that high BMI was related exists for an important subgroup of cases (e.g., metastatic with low blood testosterone level and thus lower risk of cancer), we were probably underpowered. early-onset prostate cancer (34). The other study, After recognition of microbial components, TLRs however, found that a low level of testosterone was activate not only innate immunity but also adaptive related to high BMI (z30), which reflects high insulin and immunity, largely by dendritic cells, which later express insulin-like growth factor-I levels and, thus, high risk of TLRs (e.g., TLR1, TLR2, TLR4, and TLR5; refs. 3, 39). high-grade prostate cancer (35). It is well known that the TLR1 and TLR6 need to combine with TLR2to form risk of prostate cancer significantly increases after age 50 dimers to recognize pathogens (10, 11) and elicit (36, 37). However, the associations between SNPs in the subsequent cytokine induction (12). Studies on mice TLR6-TLR1-TLR10 gene cluster and the risk of prostate have shown that CD4TLR10 could interact with MyD88 cancer were not significantly modified by BMI and age in and then activate nuclear factor-nB (2, 7). These mech- this study. anisms are very similar as TLR4 signaling pathway in Our study had several limitations. Although popula- response to pathogen recognition. However, data on tion stratification cannot be excluded in a case-control human TLR10 are limited. A previous study (22) shows study (38), our study population had a large sample size some evidence of association, supporting the contention

Table 4. ORs between TLR6-TLR1-TLR10 haplotypes and the risk of prostate cancer among Caucasians

Block 1

Haplotype Prevalence among Global test P = 0.62 P* controls, % (95% CI) 0 copies 1 copy 2copies

Case/control OR Case/control OR (95% CI) Case/control OR (95% CI) Hap1: GGTCTCT 39.9 (37.2-42.5) 242/258 1.00 294/273 1.15 (0.90-1.47) 123/125 1.04 (0.77-1.41) 0.50 Hap2: CGCTCGT 29.1 (26.6-31.5) 319/338 1.00 282/256 1.17 (0.93-1.47) 58/62 1.00 (0.68-1.47) 0.39 Hap3: CATTCCT 20.9 (18.7-23.1) 438/416 1.00 192/208 0.87 (0.68-1.10) 29/32 0.84 (0.50-1.42) 0.45

Block 2

Haplotype Prevalence among Global test P = 0.50 P* controls, % (95% CI) 0 copies 1 copy 2copies

Case/Control OR Case/control OR (95% CI) Case/control OR (95% CI)

Hap4: ATT 73.2 (70.8-75.6) 45/51 1.00 237/249 1.08 (0.70-1.67) 377/356 1.20 (0.78-1.84) 0.53 Hap5: GCC 19.0 (16.9-21.1) 445/433 1.00 190/196 0.95 (0.74-1.21) 24/27 0.87 (0.49-1.55) 0.83 Hap6: GTC 6.4 (5.1-7.7) 589/576 1.00 70/76 0.87 (0.61-1.25) 0/4 0.03 (<0.001-8.12) 0.08

Block 3

Haplotype Prevalence among Global test P = 0.59 P* controls, % (95% CI) 0 copies 1 copy 2copies

Case/Control OR Case/control OR (95% CI) Case/control OR (95% CI)

Hap7: AACGAATA 63.5 (60.9-66.1) 88/84 1.00 274/311 0.85 (0.60-1.20) 297/261 1.09 (0.77-1.54) 0.10 Hap8: AGTACCTG 14.7 (12.8-16.6) 510/491 1.00 138/153 0.87 (0.67-1.13) 11/12 0.89 (0.39-2.04) 0.57 Hap9: AACGCCTA 13.5 (11.7-15.4) 493/474 1.00 151/170 0.86 (0.67-1.11) 15/12 1.22 (0.55-2.67) 0.43

P * value is for testing the null hypothesis: OR1 copy =OR2copies =1.

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

Table 5. TLR6-TLR1-TLR10 haplotypes and risk of aggressive prostate cancer among Caucasian cases

Haplotype Noncarriers Carriers P*

Aggressive/nonaggressive OR Aggressive/nonaggressive OR (95% CI) Aggressiveness 1 Block 1 (Global test P = 0.71) Hap1: GGTCTCT 18/188 1.00 37/325 1.18 (0.65-2.13) 0.59 Hap2: CGCTCGT 36/249 1.00 29/264 1.05 (0.60-1.84) 0.86 Hap3: CATTCCT 40/344 1.00 15/169 0.76 (0.41-1.42) 0.38 Block 2(Global test P = 0.54) Hap4: ATT 6/35 1.00 49/478 0.60 (0.24-1.50) 0.30 Hap5: GCC 40/348 1.00 15/165 0.79 (0.42-1.47) 0.45 Hap6: GTC 47/459 1.00 8/54 1.47 (0.65-3.33) 0.38 Block 3 (Global test P = 0.42) Hap7: AACGAATA 5/71 1.00 50/4421.58 (0.61-4.10) 0.32 Hap8: AGTACCTG 43/395 1.00 12/118 0.96 (0.48-1.88) 0.89 Hap9: AACGCCTA 44/386 1.00 11/127 0.74 (0.37-1.49) 0.39 Aggressiveness 2 Block 1 (Global test P = 0.11) Hap1: GGTCTCT 81/161 1.00 159/258 1.24 (0.88-1.73) 0.22 Hap2: CGCTCGT 108/210 1.00 132/209 1.23 (0.90-1.70) 0.20 Hap3: CATTCCT 167/271 1.00 73/148 0.80 (0.57-1.12) 0.19 Block 2(Global test P =0.23) Hap4: ATT 19/26 1.00 221/393 0.77 (0.42-1.43) 0.41 Hap5: GCC 168/276 1.00 72/143 0.82 (0.58-1.16) 0.26 Hap6: GTC 210/381 1.00 30/38 1.49 (0.89-2.50) 0.13 Block 3 (Global test P =0.21) Hap7: AACGAATA 27/60 1.00 213/359 1.30 (0.80-2.11) 0.29 Hap8: AGTACCTG 193/317 1.00 47/1020.76 (0.51-1.13) 0.17 Hap9: AACGCCTA 187/306 1.00 53/113 0.77 (0.53-1.12) 0.16

P * value is for testing the null hypothesis: ORcarriers =1.

that variation in TLRs may reduce or even block the 6. Opal SM, Esmon CT. Bench-to-bedside review: functional relation- ships between coagulation and the innate immune response and signaling of the immune response (e.g., chronic inflam- their respective roles in the pathogenesis of sepsis. Crit Care 2003;7: mation) and thereby lower the risk of prostate cancer. But 23 – 38. this study did not find any association with TLR6-TLR1- 7. Hasan U, Chaffois C, Gaillard C, et al. Human TLR10 is a functional TLR10. Combined with the previous study, the results receptor, expressed by B cells and plasmacytoid dendritic cells, are mixed. The inconsistent results with the TLR6-TLR1- which activates gene transcription through MyD88. J Immunol 2005; TLR10 TLR4 174:2942 – 50. gene cluster as well as between the Swedish 8. Hornung V, Rothenfusser S, Britsch S, et al. Quantitative expression study and ours suggest the problem of low reproduc- of toll-like receptor 1-10 mRNA in cellular subsets of human ibility across genetic studies. Because of the limited peripheral blood mononuclear cells and sensitivity to CpG oligo- amount of research on inflammation and prostate cancer, deoxynucleotides. J Immunol 2002;168:4531 – 7. 9. Shinkai H, Muneta Y, Suzuki K, Eguchi-Ogawa T, Awata T, Uenishi more studies on these genes will be needed to confirm H. Porcine Toll-like receptor 1, 6, and 10 genes: complete sequencing these findings. Future studies related to biological path- of genomic region and expression analysis. Mol Immunol 2006;43: ways may help us understand the mechanisms of 1474 – 80. prostate cancer risk. 10. Abreu MT, Arditi M. Innate immunity and toll-like receptors: clinical implications of basic science research. J Pediatr 2004;144:421 – 9. 11. Janssens S, Beyaert R. Role of Toll-like receptors in pathogen recognition. Clin Microbiol Rev 2003;16:637 – 46. Acknowledgments 12. Ozinsky A, Underhill DM, Fontenot JD, et al. The repertoire for We thank Monica Coleman for her assistance; Pati Soule and pattern recognition of pathogens by the innate immune system is Ana-Tereza Andrade for DNA sample extraction; and Dr. David defined by cooperation between toll-like receptors. Proc Natl Acad Kwiatkowski, Alison Brown, and Maura Regan (Partners High- Sci U S A 2000;97:13766 – 71. Throughput Genotyping Center) for genotyping. 13. Edfeldt K, Swedenborg J, Hansson GK, Yan ZQ. Expression of toll- like receptors in human atherosclerotic lesions: a possible pathway for plaque activation. Circulation 2002;105:1158 – 61. 14. Lazarus R, Raby BA, Lange C, et al. TOLL-like receptor 10 genetic References variation is associated with asthma in two independent samples. Am 1. Medzhitov R, Preston-Hurlburt P, Janeway CA, Jr. A human J Respir Crit Care Med 2004;170:594 – 600. homologue of the Drosophila Toll protein signals activation of 15. Takeuchi O, Sato S, Horiuchi T, et al. Cutting edge: role of Toll-like adaptive immunity. Nature 1997;388:394 – 7. receptor 1 in mediating immune response to microbial lipoproteins. 2. Hasan UA, Dollet S, Vlach J. Differential induction of gene promoter J Immunol 2002;169:10 – 4. constructs by constitutively active human TLRs. Biochem Biophys 16. Sharma RK, Sodhi A, Batra HV. Involvement of TLR6/1 in rLcrV- Res Commun 2004;321:124 – 31. mediated immunomodulation of murine peritoneal macrophages 3. Takeda K, Kaisho T, Akira S. Toll-like receptors. Annu Rev Immunol in vitro. Mol Immunol 2005;42:695 – 701. 2003;21:335 – 76. 17. Kinlen L. Infections and immune factors in cancer: the role of 4. Takeuchi O, Kawai T, Sanjo H, et al. TLR6: a novel member of an epidemiology. Oncogene 2004;23:6341 – 8. expanding toll-like receptor family. Gene 1999;231:59 – 65. 18. Ernst P. Review article: the role of inflammation in the pathogenesis 5. Li L. Regulation of innate immunity signaling and its connection of gastric cancer. Aliment Pharmacol Ther 1999;13 Suppl 1:13 – 8. with human diseases. Curr Drug Targets Inflamm Allergy 2004;3: 19. Platz EA, De Marzo AM. Epidemiology of inflammation and prostate 81 – 6. cancer. J Urol 2004;171:S36 – 40.

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

20. Sun J, Wiklund F, Zheng SL, et al. Sequence variants in Toll-like discrete and continuous traits in samples of unrelated individuals. receptor gene cluster (TLR6-TLR1-TLR10) and prostate cancer risk. Hum Hered 2002;53:79 – 91. J Natl Cancer Inst 2005;97:525 – 32. 29. Kalish LA, McDougal WS, McKinlay JB. Family history and the risk 21. Sun J, Wiklund F, Hsu F-C, et al. Interactions of sequence variants in of prostate cancer. Urology 2000;56:803 – 6. interleukin-1 receptor-associated kinase4 and the Toll-like receptor 30. Cerhan JR, Parker AS, Putnam SD, et al. Family history and prostate 6-1-10 gene cluster increase prostate cancer risk. Cancer Epidemiol cancer risk in a population-based cohort of Iowa men. Cancer Biomarkers Prev 2006;15:480 – 5. Epidemiol Biomarkers Prev 1999;8:53 – 60. 22. Chen YC, Giovannucci E, Lazarus R, Kraft P, Ketkar S, Hunter DJ. 31. Giovannucci E, Rimm EB, Liu Y, et al. Body mass index and risk of prostate Sequence variants of Toll-like receptor 4 (TLR4) and susceptibility to cancer in U.S. health professionals. J Natl Cancer Inst 2003;95:1240 – 4. prostate cancer. Cancer Res 2005;65:11771 – 8. 32. Gabriel SB, Schaffner SF, Nguyen H, et al. The structure of haplotype 23. Zheng SL, Augustsson-Balter K, Chang B, et al. Sequence variants of blocks in the human genome. Science 2002;296:2225 – 9. toll-like receptor 4 are associated with prostate cancer risk: results 33. Florez JC, Burtt N, de Bakker PI, et al. Haplotype structure and from the CAncer Prostate in Sweden Study. Cancer Res 2004;64: genotype-phenotype correlations of the sulfonylurea receptor and 2918 – 22. the islet ATP-sensitive potassium channel gene region. Diabetes 2004; 24. Stampfer MJ, Willett WC, Speizer FE, et al. Test of the National Death 53:1360 – 8. Index. Am J Epidemiol 1984;119:837 – 9. 34. Giovannucci E. Nutrition, insulin, insulin-like growth factors and 25. Sebastiani P, Lazarus R, Weiss ST, Kunkel LM, Kohane IS, Ramoni cancer. Horm Metab Res 2003;35:694 – 704. MF. Minimal haplotype tagging. Proc Natl Acad Sci U S A 2003;100: 35. Neugut AI, Chen AC, Petrylak DP. The ‘‘skinny’’ on obesity and 9900 – 5. prostate cancer prognosis. J Clin Oncol 2004;22:395 – 8. 26. Platz EA, Leitzmann MF, Rifai N, et al. Sex steroid hormones and the 36. Franks LM. Proceedings: etiology, epidemiology, and pathology of androgen receptor gene CAG repeat and subsequent risk of prostate prostatic cancer. Cancer 1973;32:1092 – 5. cancer in the prostate-specific antigen era. Cancer Epidemiol 37. Holund B. Latent prostatic cancer in a consecutive autopsy series. Biomarkers Prev 2005;14:1262 – 9. Scand J Urol Nephrol 1980;14:29 – 35. 27. Stram DO, Leigh Pearce C, Bretsky P, et al. Modeling and E-M 38. Freedman ML, Reich D, Penney KL, et al. Assessing the impact of estimation of haplotype-specific relative risks from genotype data for population stratification on genetic association studies. Nat Genet a case-control study of unrelated individuals. Hum Hered 2003;55: 2004;36:388 – 93. 179 – 90. 39. Visintin A, Mazzoni A, Spitzer JH, Wyllie DH, Dower SK, Segal DM. 28. Zaykin DV, Westfall PH, Young SS, Karnoub MA, Wagner MJ, Ehm Regulation of Toll-like receptors in human monocytes and dendritic MG. Testing association of statistically inferred haplotypes with cells. J Immunol 2001;166:249 – 55.

Cancer Epidemiol Biomarkers Prev 2007;16(10). October 2007

Downloaded from cebp.aacrjournals.org on September 27, 2021. © 2007 American Association for Cancer Research. Association between Toll-Like Receptor Gene Cluster (TLR6, TLR1, and TLR10) and Prostate Cancer

Yen-Ching Chen, Edward Giovannucci, Peter Kraft, et al.

Cancer Epidemiol Biomarkers Prev 2007;16:1982-1989.

Updated version Access the most recent version of this article at: http://cebp.aacrjournals.org/content/16/10/1982

Cited articles This article cites 39 articles, 16 of which you can access for free at: http://cebp.aacrjournals.org/content/16/10/1982.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cebp.aacrjournals.org/content/16/10/1982.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/16/10/1982. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.