Document mpJbJMbvevX8rxq3XbYKJVnYO
Polymorphisms and Haplotypes in the Cytochrome P450 17A1, Prolactin, and Catechol-O-Methyltransferase Genes and Non-Hodgkin Lymphoma Risk
2391
Christine F. Skibola,1 Paige M. Bracci,2 Randi A. Paynter,1 Matthew S. Forrest,1 Luz Agana,1 Trevor Woodage,3 Karl Guegler,3 Martyn T. Smith,1 and Elizabeth A. Holly2
1Division of Environmental Health Sciences, School of Public Health, University of California, Berkeley, California; 2Department of Epidemiology and Biostatistics, University of California, San Francisco, California; and 3Advanced Research and Technology, Applied Biosystems, Foster City, California
Abstract
Expression of prolactin and of prolactin and estrogen receptors in lymphocytes, bone marrow, and lymphoma cell lines suggests that hormonal modulation may influence lymphoma risk. Prolactin and estrogen promote the proliferation and survival of B cells, factors that may increase non-Hodgkin lymphoma risk, and effects of estrogen may be modified by catechol-O-methyltransferase (COMT), an enzyme that alters estrogenic activity. Cytochrome P450 17A1 (CYP17A1), a key enzyme in estrogen biosynthesis, has been associated with increased cancer risk and may affect lymphoma susceptibility. We studied the polymorphisms prolactin (PRL) 1149G>T, CYP17A1 34T>C, and COMT 108/158Val>Met, and predicted haplotypes among a subset of participants (n = 308 cases, n = 684 controls) in a San Francisco Bay Area population-based non-Hodgkin lymphoma study (n = 1,593 cases, n = 2,515 controls) conducted from 1988 to 1995. Oral contraceptive and other hormone use also was analyzed. Odds ratios (OR) for nonHodgkin lymphoma and follicular lymphoma were reduced
for carriers of the PRL 1149TT genotype [OR, 0.64; 95% confidence interval (95% CI), 0.41-1.0; OR, 0.53; 95% CI, 0.26-1.0, respectively]. Diffuse large-cell lymphoma risk was increased for those with CYP17A1 polymorphisms including CYP17A1 34CC (OR, 2.0; 95% CI, 1.1-3.5). ORs for all non-Hodgkin lymphoma and follicular lymphoma among women were decreased for COMT IVS1 701A>G [rs737865; variant allele: OR, 0.53; 95% CI, 0.34-0.82; OR, 0.42; 95% CI, 0.23-0.78, respectively]. Compared with never users of oral contraceptives, a 35% reduced risk was observed among oral contraceptive users in the total population. Reduced ORs for all non-Hodgkin lymphoma were observed with use of exogenous estrogens among genotyped women although 95% CIs included unity. These results suggest that PRL, CYP17A1, and COMT may be relevant genetic loci for non-Hodgkin lymphoma and indicate a possible role for prolactin and estrogen in lymphoma pathogenesis. (Cancer Epidemiol Biomarkers Prev 2005;14(10):2391 401)
Introduction
Extensive cross-talk exists between the endocrine and immune systems where hormones and their respective receptors influence immune function and, in turn, immune responses influence neuroendocrine changes (reviewed in ref. 1). Expression of prolactin and of prolactin and estrogen receptors found in normal B and T lymphocytes, bone marrow, and in leukemia and lymphoma cell lines (2, 3) suggests their importance in the lymphopoeitic system and that hormonal modulation may influence risk of lymphopoeitic diseases such as lymphoma. Prolactin and estrogens play important roles in women's reproductive physiology and they also function in both sexes as immune modulators that affect apoptosis, activation, and proliferation of immune cells and modulate B-cell development. Elevated prolactin levels have been implicated in the progression of hematologic diseases such as multiple myeloma, acute myeloid leukemia, and nonHodgkin lymphoma (4). However, there is less agreement in
Received 5/13/05; revised 8/10/05; accepted 8/15/05.
Grant support: NIH grants RO1-CA104862 (M.T. Smith, P.I.) and CA45614, CA89745, and CA87014 (E.A. Holly, P.I.) from the National Cancer Institute, and by the National Foundation for Cancer Research.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Epidemiology Biomakers and Prevention Online (http://cebp.aacrjournals.org/).
Requests for reprints: Christine F. Skibola, Division of Environmental Health Sciences, School of Public Health, 140 Earl Warren Hall, University of California, Berkeley, CA 94720-7360. Phone: 510-643-5041; Fax: 510-642-0427. E-mail: chrisfs@uclink.berkeley.edu
Copyright D 2005 American Association for Cancer Research.
doi:10.1158/1055-9965.EPI-05-0343
the literature about the association between exogenous estrogens (using postmenopausal hormones or oral contraceptives as a proxy of estimated exposure) and lymphoma/leukemia risk with reports of positive (5, 6), null (7), and inverse associations (8). Because of the many postmenopausal hormone formulations available, misclassification and differential recall may affect the results. Further, direct measurement of circulating hormones provides data only for a single time point and levels may be affected by genetic variability in pathways involved in their production and metabolism. Thus, investigation of genetic polymorphisms that may influence prolactin and estrogen production will contribute to our knowledge and add to the interpretation of the exposure data.
The prolactin (PRL) gene maps to regions linked to rheumatoid arthritis and systemic lupus erythematosus (9), in close proximity to the MHC on chromosome 6p. Multiple promoters and start sites are present in the PRL gene. PRL gene expression in lymphocytes and other extrapituitary tissues is directed by a promoter region that lies f6 kb upstream of the pituitary-specific start site of transcription (10). A single-nucleotide polymorphism (SNP) in this region (rs1341239: PRL 1149G>T) that regulates lymphocyte prolactin production recently has been identified (11). Stevens et al. reported that the PRL 1149G allele was overrepresented in a cohort of systemic lupus erythematosus patients and was associated with enhanced promoter activity and elevated prolactin mRNA levels in T lymphocytes. Cytochrome P450 17A1 (CYP17A1), which catalyzes the conversion of pregnenolone and progesterone to 17a-hydroxypregnenolone and 17ahydroxyprogesterone, respectively, is one of the key enzymes
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
2392 CYP17A1, PRL, and COMT Polymorphisms and Non-Hodgkin Lymphoma Risk
involved in estrogen and testosterone biosynthesis. A SNP in the 5V-untranslated region of the CYP17A1 gene, 34 bp upstream of the initiation site of translation (rs743572, 34T>C; ref. 12), has been speculated to enhance CYP17A1 transcriptional efficiency and enzyme activity. This SNP has been associated with earlier age at menarche, increased risk for breast and prostate cancers (13-16), and elevated serum estrogen levels (reviewed in refs. 17, 18). Furthermore, allelic variation in the catechol-O-methyltransferase (COMT) gene that expresses an intracellular enzyme involved in estrogen metabolism can alter circulating estrogen concentrations. The COMT gene encodes both a soluble protein (S-COMT) expressed in blood, liver, and kidneys and a membrane-bound protein (MB-COMT) expressed in brain neurons (19). A G>A SNP in exon 4 (rs4680) causes a valine to methionine substitution in S-COMT (108Val>Met) and MB-COMT (158Val>Met) that results in enzyme thermolability and 2- to 4-fold lower catalytic activity (20, 21). Consequently, this polymorphism could alter estrogenic activity in various target tissues.
We hypothesized that SNPs or haplotypes in the PRL, CYP17A1, and COMT genes associated with elevated prolactin and estrogen levels (i.e., PRL 1149G, CYP17A1 34C, and COMT 108/158Met alleles) promote B- and T-cell activation, survival, and proliferation, factors that may contribute to the pathogenesis of non-Hodgkin lymphoma. To test this, we evaluated these and other PRL, CYP17A1, and COMT SNPs and haplotypes in a population-based casecontrol study conducted in the San Francisco Bay Area between 1988 and 1995.
Materials and Methods
Study Population. Briefly, non-Hodgkin lymphoma patients were identified by the Northern California Cancer Center rapid case ascertainment. Eligible patients were between 21 and 74 years of age, were residents of one of the six Bay Area counties at the time of diagnosis, and could complete an interview in English. A total of 1,593 eligible patients (284 HIV positive) completed in-person interviews (72% response rate). Control participants were identified by random-digit dial and by random sampling of the Health Care Financing Administration lists to supplement recruitment of participants aged z65 years. Controls were frequency matched to patients by age within 5 years, sex, and county of residence. No proxy interviews were conducted. There were 2,515 (78% response rate) eligible control participants (111 HIV positive) who completed inperson interviews. The study population reported their race/ ethnicity as white Hispanic (6%), white non-Hispanic (84%), Black (4%), Asian (5%), and other (1%). Race/ethnicity distribution was similar for case and control participants. Detailed methods have been published previously (22-24).
Patients and control participants who had no history of chemotherapy within the past 3 months and no contraindications to venipuncture were asked to provide a blood specimen for the laboratory portion of the study. Almost all study patients (97%) had their pathology reports and diagnostic materials rereviewed by an expert pathologist and these were classified using the Working Formulation (NonHodgkin's Lymphoma Classification Project). To better reflect the Revised European American Lymphoma Classification and WHO Classification systems, Working Formulation diffuse large-cell and immunoblastic lymphoma were combined for the diffuse large-cell lymphoma subtype and Working Formulation follicular small, mixed, and large-cell lymphomas were combined for the follicular lymphoma subtype (25, 26) in these analyses. Study protocols were approved by the University of California San Francisco Committee on Human Research and participants provided written informed consent before interview and collection of blood specimens.
Isolation of DNA. DNA was isolated from peripheral blood mononuclear cells using a modified QIAamp DNA Blood Maxi Kit protocol (Qiagen, Inc., Santa Clarita, CA), and DNA was quantified using PicoGreen dsDNA Quantitation kits (Molecular Probes, Eugene, OR) according to the specifications of the manufacturers.
SNP Selection. PRL, CYP17A1, and COMT SNPs are listed in Table 1 and were identified using SNP (http://www.ncbi. nlm.nih.gov/SNP/) and SNPper (http://snpper.chip.org/). In addition, all available Applied Biosystems TaqMan SNP genotyping assays (Applied Biosystems, Foster City, CA) were identified (http://www.appliedbiosystems.com). SNPs were chosen for investigation based on a minor allele frequency of z5% and location, with a preference given to coding and untranslated region SNPs. Where no suitable exonic SNPs were found, intronic SNPs were chosen to ensure adequate gene coverage.
Genotyping. DNA was available and genotyping was done on 376 case and 801 control participants using the 5Vnuclease allelic discrimination assay on the ABI Prism 7700 Sequence Detection System or GeneAmp PCR System 9700. TaqMan SNP genotyping products and custom SNP genotyping assays were used. Reactions were done with the following protocol: 95jC for 10 minutes, then 40 cycles of 95jC for 15 seconds, and 60jC for 1 minute. A post-PCR plate read on the 7700 Sequence Detection System was used to determine genotype. Probes and primer sets used for the PRL, CYP17A1, and COMT SNPs are listed in Supplementary Table 1. Replicate, blinded quality control samples were included to assess reproducibility of the genotyping procedure.
Statistical Analysis. Because of differences in SNP and haplotype frequencies across self-identified race and Hispanic ethnicity categories, we restricted all analyses to only those HIV-negative individuals who reported their race/ethnicity as white non-Hispanic (cases, n = 308; controls, n = 684). All regression analyses were conducted using SAS statistical software (SAS version 8, SAS Institute, Cary, NC). Unconditional logistic regression models were used to compute odds ratios (OR), expressed in the text as ``risk'' for non-Hodgkin lymphoma, and corresponding 95% confidence intervals (95% CI) adjusted for age in 5-year groups and sex. All SNP-specific analyses used the homozygous wild-type category as the reference group.
Linkage disequilibrium was computed for each pair of polymorphisms and linkage disequilibrium plots were generated using Haploview (27). Haplotype frequencies were estimated from phase-unknown genotypes using the tagSNPs implementation of the estimation-maximization algorithm (28). ORs and 95% CIs were estimated for haplotype associations with non-Hodgkin lymphoma by unconditional logistic regression using the single imputation approach of Zaykin et al. (29). Haplotypes with estimated frequencies <5% were considered to be low frequency and were pooled into a single category labeled ``Other''. The global test for association between common haplotypes and non-Hodgkin lymphoma was evaluated using a likelihood ratio test.
Associations between non-Hodgkin lymphoma and hormone-related factors including oral contraceptive use, menopausal status, and non-oral-contraceptive hormone use were evaluated among all HIV-negative, white non-Hispanic women (n = 451 patients, n = 678 controls) and for the subset of women for whom DNA had been genotyped for these analyses (n = 134 cases, n = 220 controls). Oral contraceptive use was analyzed by ever/never use and by duration of use (V5 and >5 years). Women were classified as postmenopausal if they met any of the following conditions: age 55 years or older, had prior hysterectomy or oophorectomy, or reported
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
Cancer Epidemiology, Biomarkers & Prevention 2393
Table 1. ORs and 95% CIs for all non-Hodgkin lymphoma, diffuse large-cell lymphoma, and follicular lymphoma associated with SNPs in PRL, CYP17A1, and COMT genes among HIV-negative white non-Hispanics, San Francisco Bay Area, 1988-1995
SNP database Genotype identifier
Controls (N = 684)
n (%)*
All non-Hodgkin lymphomas (N = 308)
n (%)* OR (95% CI)c
Diffuse large-cell lymphoma (N = 98)
n (%)* OR (95% CI)c
Follicular lymphoma (N = 112)
n (%)* OR (95% CI)c
PRL SNP1 rs1341239 SNP2 rs849877 SNP3 rs7739889 SNP4 rs6239
CYP17A1 SNP1 rs743572 SNP2 rs6162 SNP3 rs6163 SNP4 rs3781287 SNP5 rs743575 SNP6 rs1004467 SNP7 rs3740397 SNP8 rs10883783 SNP9 rs4919685
COMT SNP1 rs737865
SNP2 rs4633
SNP3 rs4680
SNP4 rs165599
1149GG 1149GT 1149TT 1488AA 1488AG 1488GG
214CC
214CT
214TT
570GG
570GA/AA
253 (37) 326 (48) 102 (15) 267 (39) 319 (47) 98 (14) 401 (59) 242 (36) 36 (5) 652 (95) 32 (5)
34TT 34TC 34CC
137GG 137GA 137AA 195CC 195CA 195AA 270AA 270AC 270CC 105AA 105AC 105CC 35TT 35TC 35CC 75CC 75CG 75GG 114AA 114AT 114TT 2930GG 2930GT 2930TT
249 (36) 341 (50) 94 (14) 237 (35) 339 (50) 106 (16) 244 (36) 343 (50) 95 (14) 226 (33) 353 (52) 105 (15) 338 (50) 287 (42) 52 (8) 556 (81) 119 (17)
8 (1) 253 (37) 343 (50) 88 (13) 310 (45) 284 (42) 89 (13) 340 (50) 285 (42) 58 (8)
701AA 701AG 701GG 701AG/GG 186CC 186CT 186TT 186CT/TT 108/158 VV 108/158 VM 108/158 MM 108/158
VM/MM 6731AA 6731AG 6731GG 6731AG/GG
309 (45) 300 (44) 73 (11) 373 (55) 194 (29) 314 (46) 169 (25) 483 (71) 193 (28) 323 (48) 163 (24) 486 (72)
318 (47) 285 (42) 78 (11) 363 (53)
138 (45) 130 (43) 37 (12) 133 (43) 140 (45) 35 (11) 187 (61) 109 (36) 11 (4) 300 (97)
8 (3)
113 (37) 137 (44) 58 (19) 99 (32) 149 (48) 60 (19) 114 (37) 132 (43) 62 (20) 96 (31) 149 (48) 63 (20) 149 (49) 124 (40) 34 (11) 247 (80) 56 (18)
4 (1) 107 (35) 142 (46) 59 (19) 147 (48) 124 (40) 37 (12) 148 (48) 124 (40) 35 (11)
157 (51) 126 (41) 22 (7) 148 (49) 80 (26) 144 (47) 80 (26) 224 (74) 75 (25) 153 (50) 77 (25) 230 (75)
152 (50) 124 (40) 31 (10) 155 (50)
1.0 (reference) 0.73 (0.54-0.98) 0.64 (0.41-1.0)
1.0 (reference) 0.88 (0.65-1.2) 0.70 (0.45-1.1)
1.0 (reference) 0.91 (0.68-1.2) 0.63 (0.31-1.3)
1.0 (reference) 0.51 (0.23-1.2)
1.0 (reference) 0.88 (0.65-1.2)
1.4 (0.95-2.1) 1.0 (reference) 1.1 (0.78-1.4) 1.4 (0.95-2.1) 1.0 (reference) 0.83 (0.61-1.1) 1.5 (0.99-2.2) 1.0 (reference) 0.99 (0.73-1.4) 1.5 (0.99-2.2) 1.0 (reference) 1.0 (0.76-1.4) 1.5 (0.93-2.5) 1.0 (reference) 1.1 (0.76-1.6) 0.86 (0.25-3.0) 1.0 (reference) 0.97 (0.71-1.3) 1.6 (1.1-2.5) 1.0 (reference) 0.96 (0.71-1.3) 0.97 (0.62-1.5) 1.0 (reference) 1.1 (0.77-1.4) 1.5 (0.90-2.4)
1.0 (reference) 0.85 (0.63-1.1) 0.60 (0.36-1.0) 0.80 (0.61-1.1)
1.0 (reference) 1.1 (0.78-1.5) 1.1 (0.78-1.7) 1.1 (0.81-1.5) 1.0 (reference) 1.2 (0.86-1.7) 1.2 (0.80-1.8) 1.2 (0.87-1.7)
1.0 (reference) 0.91 (0.68-1.2) 0.87 (0.55-1.4) 0.90 (0.68-1.2)
38 (39) 46 (47) 14 (14) 42 (43) 44 (45) 12 (12) 63 (64) 31 (32) 4 (4) 97 (99) 1 (1)
35 (36) 39 (40) 24 (24) 32 (33) 42 (43) 24 (24) 35 (36) 37 (38) 26 (27) 32 (33) 43 (44) 23 (23) 46 (47) 38 (39) 14 (14) 78 (80) 19 (19) 1 (1) 34 (35) 38 (39) 26 (27) 46 (47) 38 (39) 14 (14) 45 (46) 38 (39) 15 (15)
50 (51) 42 (43) 6 (6) 48 (49) 24 (25) 48 (50) 25 (26) 75 (75) 23 (24) 48 (50) 25 (26) 73 (76)
52 (54) 38 (39) 7 (7) 45 (46)
1.0 (reference) 0.95 (0.59-1.5) 0.91 (0.47-1.8)
1.0 (reference) 0.88 (0.56-1.4) 0.77 (0.39-1.5)
1.0 (reference) 0.78 (0.49-1.2) 0.66 (0.23-1.9)
1.0 (reference) 0.20 (0.03-1.5)
1.0 (reference) 0.83 (0.51-1.4)
2.0 (1.1-3.5) 1.0 (reference) 0.94 (0.57-1.5) 1.8 (1.0-3.3) 1.0 (reference) 0.77 (0.47-1.3) 2.1 (1.2-3.6) 1.0 (reference) 0.87 (0.53-1.4) 1.7 (0.92-3.0) 1.0 (reference) 1.0 (0.65-1.6) 2.1 (1.1-4.2) 1.0 (reference) 1.2 (0.67-2.0) 0.72 (0.09-6.0) 1.0 (reference) 0.83 (0.51-1.4) 2.3 (1.3-4.1) 1.0 (reference) 0.95 (0.60-1.5) 1.3 (0.65-2.4) 1.0 (reference) 1.0 (0.66-1.7) 2.2 (1.1-4.3)
1.0 (reference) 0.87 (0.56-1.4) 0.51 (0.21-1.2) 0.80 (0.52-1.2)
1.0 (reference) 1.2 (0.72-2.1) 1.2 (0.66-2.2) 1.2 (0.74-2.0) 1.0 (reference) 1.2 (0.72-2.1) 1.3 (0.70-2.4) 1.3 (0.76-2.1)
1.0 (reference) 0.82 (0.52-1.3) 0.57 (0.25-1.3) 0.77 (0.50-1.2)
51 (46) 48 (43) 12 (11) 51 (46) 50 (45) 11 (10) 65 (58) 43 (38)
4 (4) 110 (98)
2 (2)
45 (40) 51 (46) 16 (14) 39 (35) 55 (49) 18 (16) 45 (40) 49 (44) 18 (16) 39 (35) 53 (47) 20 (18) 57 (51) 45 (40) 10 (9) 94 (84) 17 (15)
1 (1) 46 (41) 50 (45) 16 (14) 57 (51) 44 (39) 11 (10) 57 (51) 45 (40) 10 (9)
60 (54) 47 (42)
5 (4) 52 (46) 30 (27) 52 (47) 29 (26) 81 (73) 26 (23) 59 (53) 27 (24) 86 (77)
59 (53) 44 (39)
9 (8) 53 (47)
1.0 (reference) 0.69 (0.45-1.1) 0.53 (0.26-1.0) 1.0 (reference) 0.78 (0.51-1.2) 0.55 (0.27-1.1) 1.0 (reference) 1.0 (0.65-1.5) 0.65 (0.22-1.9) 1.0 (reference) 0.34 (0.08-1.5)
1.0 (reference) 0.83 (0.53-1.3) 1.0 (0.55-2.0) 1.0 (reference) 0.98 (0.62-1.5) 1.1 (0.61-2.1) 1.0 (reference) 0.79 (0.50-1.2) 1.1 (0.62-2.1) 1.0 (reference) 0.85 (0.54-1.3) 1.2 (0.68-2.3) 1.0 (reference) 1.0 (0.65-1.5) 1.2 (0.56-2.5) 1.0 (reference) 0.88 (0.50-1.6) 0.55 (0.07-4.6) 1.0 (reference) 0.80 (0.51-1.2) 1.1 (0.58-2.1) 1.0 (reference) 0.91 (0.59-1.4) 0.75 (0.37-1.5) 1.0 (reference) 1.0 (0.66-1.6) 1.1 (0.53-2.4)
1.0 (reference) 0.85 (0.56-1.3) 0.37 (0.14-0.97) 0.78 (0.50-1.1) 1.0 (reference) 1.0 (0.64-1.7) 1.1 (0.63-1.9) 1.1 (0.67-1.7) 1.0 (reference) 1.3 (0.79-2.2) 1.2 (0.66-2.1) 1.3 (0.79-2.1)
1.0 (reference) 0.82 (0.54-1.3) 0.64 (0.30-1.4) 0.78 (0.52-1.2)
*Numbers may not add to 308 cases and 684 controls due to missing genotypes. cORs and 95% CIs computed using unconditional logistic regression, adjusted for age and sex.
non-oral-contraceptive hormone use before age 55. Non-oralcontraceptive hormone use among postmenopausal women also was analyzed by ever/never use and by duration of use (V5 and >5 years). Never users composed the reference category for all analyses of oral contraceptives and non-oralcontraceptive hormones. m2 tests for linear trend in duration of use were conducted using the b coefficients computed from adjusted logistic regression models that included duration coded as an ordinal categorical variable.
Interactions between haplotypes and sex and body mass index (ordinal categories; <25, 25 to <30, z30) were evaluated for men and women combined. Gene-environment interaction terms were created by multiplying each environmental factor by each predicted haplotype as a continuous variable. Body mass index-haplotype interaction terms were generated by multiplying the ordinal body mass index category by each predicted haplotype. The Wald test was used to evaluate each haplotype interaction term. All models for women and men
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
2394 CYP17A1, PRL, and COMT Polymorphisms and Non-Hodgkin Lymphoma Risk
combined were adjusted for age and sex, whereas all analyses among women were adjusted for age alone. Results were considered statistically significant for two-sided P V 0.05 and borderline significant for 0.05 < P V 0.10.
Results
PRL, CYP17A1, and COMT Genotypes and NonHodgkin Lymphoma Risk. Figure 1 illustrates the scaled locations of the PRL, CYP17A1, and COMT SNPs genotyped. Control genotype distributions of all SNPs were in HardyWeinberg equilibrium. For PRL, inverse associations with nonHodgkin lymphoma were observed for SNP1 (heterozygotes: OR, 0.73; 95% CI, 0.54-0.98; homozygous variants: OR, 0.64; 95% CI, 0.41-1.0) and SNP2 (heterozygotes: OR, 0.88; 95% CI, 0.65-1.2; homozygous variants: OR, 0.70; 95% CI, 0.45-1.1; Table 1). ORs showed inverse associations with follicular lymphoma for SNP1 (heterozygotes: OR, 0.69; 95% CI, 0.45-1.1; homozygous variants: OR, 0.53; 95% CI, 0.26-1.0) and SNP2 (heterozygotes: OR, 0.78; 95% CI, 0.51-1.2; homozygous variants: OR, 0.55; 95% CI, 0.27-1.1). There was no evidence of interaction between PRL SNPs and sex (Tables 2 and 3).
For CYP17A1, an increased non-Hodgkin lymphoma risk was observed among homozygous variant carriers of SNP3 (OR, 1.5; 95% CI, 0.99-2.2), SNP4 (OR, 1.5; 95% CI, 0.99-2.2), and SNP7 (OR, 1.6; 95% CI, 1.1-2.5; Table 1). Increased ORs for diffuse large-cell lymphoma were observed among homozygous variant allele carriers of SNP1 (OR, 2.0; 95% CI, 1.1-3.5), SNP2 (OR, 1.8; 95% CI, 1.0-3.3), SNP3 (OR, 2.1; 95% CI, 1.2-3.6), SNP5 (OR, 2.1; 95% CI, 1.1-4.2), SNP7 (OR, 2.3; 95% CI, 1.3-4.1), and SNP9 (OR, 2.2; 95% CI, 1.1-4.3). Diffuse large-cell lymphoma was increased among women who were carriers of the homozygous variant alleles for SNP1 (OR, 2.4; 95% CI, 1.0-5.7), SNP3 (OR, 2.4; 95% CI, 1.0-5.8), and SNP7 (OR, 2.5; 95% CI, 1.0-5.9; Table 3). Further, diffuse largecell lymphoma risk was elevated among men who were
homozygous variant allele carriers for SNP7 (OR, 2.3; 95% CI, 1.0-4.9). Although 95% CIs overlapped unity, in men SNP3 (OR, 1.9; 95% CI, 0.88-4.1), SNP5 (OR, 2.3; 95% CI, 0.94-5.8), and SNP9 (OR, 2.3; 95% CI, 0.94-5.7) followed the same trend of increased ORs observed for all non-Hodgkin lymphoma.
For COMT, SNP1 was inversely associated with nonHodgkin lymphoma (heterozygotes: OR, 0.85; 95% CI, 0.631.1; homozygous variants: OR, 0.60; 95% CI, 0.36-1.0) and with follicular lymphoma (heterozygotes: OR, 0.85; 95% CI, 0.56-1.3; homozygous variants: OR, 0.37; 95% CI, 0.14-0.97; Table 1). In women, SNP1 was inversely associated with non-Hodgkin lymphoma (heterozygotes: OR, 0.57; 95% CI, 0.36-0.91; homozygous variants: OR, 0.36; 95% CI, 0.15-0.89; Table 2) and follicular lymphoma (heterozygotes: OR, 0.50; 95% CI, 0.260.94; Table 3), but not in men. Furthermore, increased risk estimates for non-Hodgkin lymphoma and follicular lymphoma approached statistical significance among women who were homozygous variant carriers for SNP3 (OR, 1.6; 95% CI, 0.86-3.1; OR, 2.0; 95% CI, 0.84-4.9, respectively).
PRL, CYP17A1, and COMT Haplotypes and NonHodgkin Lymphoma Risk. Common haplotypes for PRL, CYP17A1, and COMT are listed in Table 4. Linkage disequilibrium measures between SNPs for each gene studied are presented in Fig. 2. PRL haplotypes were estimated excluding SNP4 due to its low allele frequency (2.4%) and because all major haplotypes contained only the wild-type allele, rendering SNP4 uninformative to the haplotype analysis. Using PRL SNP1 to SNP3, four common haplotypes were predicted. Using the highest-frequency haplotype HapA (all wild-type alleles) as the reference group, HapB-D were inversely associated with non-Hodgkin lymphoma, although the global test for association was not statistically significant (P = 0.12). Notably, 59% of non-Hodgkin lymphoma cases were predicted to carry HapA compared with 55% of controls. HapA was associated with non-Hodgkin lymphoma (OR, 1.2; 95% CI, 1.0-1.5) when compared with all other haplotypes.
Figure 1. A diagrammatic representation of the SNPs investigated in the PRL, CYP17A1, and COMT genes.
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
Cancer Epidemiology, Biomarkers & Prevention 2395
Table 2. ORs and 95% CIs for all non-Hodgkin lymphoma associated with SNPs in PRL, CYP17A1, and COMT genes among San Francisco Bay Area HIV-negative white non-Hispanics, stratified by sex
SNP Genotype
Controls
All non-Hodgkin lymphoma
Men (N = 463) n (%)*
Women (N = 221) n (%)*
Men (N = 174)
n (%)*
OR (95% CI)c
Women (N = 134)
n (%)*
OR (95% CI)c
PRL SNP1 SNP2 SNP3 SNP4
CYP17 SNP1 SNP2 SNP3 SNP4 SNP5 SNP6 SNP7 SNP8 SNP9
COMT SNP1
SNP2
SNP3
SNP4
1149GG 1149GT 1149TT 1488AA 1488AG 1488GG
214CC 214CT 214TT 570GG 570GA/AA
34TT 34TC 34CC
137GG 137GA 137AA 195CC 195CA 195AA 270AA 270AC 270CC 105AA 105AC 105CC 35TT 35TC 35CC 75CC 75CG 75GG 114AA 114AT 114TT 2930GG 2930GT 2930TT
701AA 701AG 701GG 701AG/GG 186CC 186CT 186TT 186CT/TT 108/158 VV 108/158 VM 108/158 MM 108/158 VM/MM 6731AA 6731AG 6731GG 6731AG/GG
176 (38) 216 (47) 69 (15) 179 (39) 217 (47) 67 (14) 279 (61) 155 (34) 25 (5) 443 (96) 20 (4)
167 (36) 228 (49) 68 (15) 157 (34) 229 (50) 76 (16) 161 (35) 231 (50) 69 (15) 150 (32) 234 (51) 79 (17) 225 (49) 198 (43) 36 (8) 373 (81) 85 (18)
5 (1) 171 (37) 229 (49) 63 (14) 212 (46) 189 (41) 61 (13) 228 (49) 193 (42) 41 (9)
208 (45) 204 (43) 49 (11) 253 (55) 140 (30) 205 (45) 115 (25) 320 (70) 139 (30) 208 (45) 112 (24) 320 (70) 214 (47) 194 (42) 52 (11) 246 (53)
77 (35) 110 (50) 33 (15) 88 (40) 102 (46) 31 (14) 122 (55) 87 (40) 11 (5) 209 (95) 12 (5)
82 (37) 113 (51) 26 (12) 80 (36) 110 (50) 30 (14) 83 (38) 112 (51) 26 (12) 76 (34) 119 (54) 26 (12) 113 (52) 89 (41) 16 (7) 183 (83) 34 (15)
3 (1) 82 (37) 114 (52) 25 (11) 98 (44) 95 (43) 28 (13) 112 (51) 92 (42) 17 (8)
101 (46) 96 (43) 24 (11) 120 (54) 54 (25) 109 (50) 54 (25) 163 (75) 54 (25) 115 (52) 51 (23) 166 (75) 104 (47) 91 (41) 26 (12) 117 (53)
74 (43) 82 (47) 18 (10) 76 (44) 82 (47) 16 (9) 109 (63) 59 (34) 5 (3) 172 (99) 2 (1)
64 (37) 78 (45) 32 (18) 58 (33) 84 (48) 32 (18) 65 (37) 75 (43) 34 (20) 52 (30) 88 (51) 34 (20) 79 (46) 74 (43) 20 (12) 145 (84) 27 (16) 1 (0.6) 59 (34) 82 (47) 33 (19) 78 (45) 75 (43) 21 (12) 80 (46) 74 (43) 19 (11)
75 (44) 82 (48) 15 (9) 97 (56) 53 (30) 78 (45) 43 (25) 121 (70) 49 (28) 84 (49) 40 (23) 124 (72) 85 (49) 69 (40) 19 (11) 88 (51)
1.0 0.93 (0.63-1.4) 0.66 (0.36-1.2) 1.0 0.95 (0.65-1.4) 0.62 (0.33-1.2) 1.0 1.0 (0.70-1.5) 0.51 (0.19-1.4) 1.0 0.27 (0.06-1.2)
1.0 0.90 (0.60-1.3) 1.3 (0.76-2.2) 1.0 0.99 (0.66-1.5) 1.2 (0.71-2.1) 1.0 0.81 (0.54-1.2) 1.3 (0.76-2.2) 1.0 1.1 (0.71-1.6) 1.3 (0.75-2.2) 1.0 1.1 (0.75-1.6) 1.6 (0.83-2.9) 1.0 0.85 (0.52-1.4) 0.50 (0.06-4.5) 1.0 1.0 (0.67-1.5) 1.5 (0.90-2.7) 1.0 1.1 (0.75-1.6) 1.1 (0.60-1.9) 1.0 1.1 (0.76-1.6) 1.4 (0.77-2.7)
1.0 1.1 (0.74-1.6) 0.78 (0.40-1.5) 1.0 (0.71-1.5) 1.0 1.1 (0.69-1.6) 1.0 (0.62-1.7) 1.0 (0.70-1.5) 1.0 1.2 (0.79-1.9) 1.0 (0.61-1.7) 1.1 (0.76-1.7) 1.0 0.90 (0.61-1.3) 0.98 (0.53-1.8) 0.91 (0.63-1.3)
64 (49) 48 (37) 19 (15) 57 (43) 58 (43) 19 (14) 78 (58) 50 (37)
6 (4) 128 (96)
6 (4)
49 (37) 59 (44) 26 (19) 41 (31) 65 (49) 28 (21) 49 (37) 57 (43) 28 (21) 44 (33) 61 (46) 29 (22) 70 (52) 50 (37) 14 (10) 102 (76) 29 (22)
3 (2) 48 (36) 60 (45) 26 (19) 69 (51) 49 (37) 16 (12) 68 (51) 50 (37) 16 (12)
82 (62) 44 (33)
7 (5) 51 (38) 27 (21) 66 (51) 37 (28) 103 (79) 26 (20) 69 (52) 37 (28) 106 (80) 67 (50) 55 (41) 12 (9) 67 (50)
1.0 0.51 (0.32-0.83) 0.63 (0.32-1.2) 1.0 0.84 (0.52-1.3) 0.86 (0.44-1.7) 1.0 0.81 (0.51-1.3) 0.80 (0.28-2.3) 1.0 0.80 (0.29-2.2)
1.0 0.86 (0.54-1.4) 1.7 (0.87-3.2) 1.0 1.2 (0.71-1.9) 1.8 (0.94-3.4) 1.0 0.85 (0.53-1.4) 1.8 (0.96-3.5) 1.0 0.87 (0.53-1.4) 1.9 (0.99-3.6) 1.0 0.93 (0.58-1.5) 1.4 (0.64-3.1) 1.0 1.5 (0.87-2.7) 1.5 (0.29-7.6) 1.0 0.89 (0.55-1.4) 1.8 (0.92-3.4) 1.0 0.76 (0.48-1.2) 0.85 (0.42-1.7) 1.0 0.91 (0.57-1.4) 1.5 (0.71-3.2)
1.0 0.57 (0.36-0.91) 0.36 (0.15-0.89) 0.53 (0.34-0.82) 1.0 1.3 (0.74-2.3) 1.5 (0.80-2.9) 1.4 (0.80-2.3) 1.0 1.3 (0.75-2.3) 1.6 (0.86-3.1) 1.4 (0.83-2.4) 1.0 0.91 (0.57-1.4) 0.71 (0.33-1.5) 0.87 (0.56-1.3)
*Numbers may not add to 308 cases and 684 controls due to missing genotypes. cORs and 95% CIs calculated using unconditional logistic regression, adjusted for age.
According to the additive model for the single imputation approach to modeling HapA, those predicted to carry one copy had an OR for non-Hodgkin lymphoma of 1.2, whereas those predicted to carry two copies had an OR of 1.5. Similar results were observed for follicular lymphoma with ORs of 1.5 and 2.4 for those predicted to carry one or two copies, respectively. There was no evidence of sex-specific associations with any of the PRL haplotypes (Supplementary Table 2).
Strong pairwise linkage disequilibrium was observed among all CYP17A1 SNPs (Fig. 2) that resulted in three
common haplotypes (Table 4). The estimated haplotype structure and the low haplotype diversity across CYP17A1 were comparable to what has been reported for Caucasian populations (28, 30). A global test for association between these common CYP17A1 haplotypes and non-Hodgkin lymphoma was not statistically significant (P = 0.34). HapB (composed of variant alleles for all SNPs except SNP6) was found in 32% of diffuse large-cell lymphoma cases and 24% of controls. The global test confirmed an association between CYP17A1 haplotypes and diffuse large-cell lymphoma (P = 0.002). Using
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
2396 CYP17A1, PRL, and COMT Polymorphisms and Non-Hodgkin Lymphoma Risk
HapA (all wild-type alleles) as the reference group and assuming an additive model, one copy of HapB conferred a 1.5-fold increased risk for diffuse large-cell lymphoma, whereas two copies conferred a 2.1-fold increased risk. Further, among women, 11% of non-Hodgkin lymphoma patients were predicted to carry HapC compared with 7% of controls, whereas among men, 7% of patients and 8% of controls carried HapC (Supplementary Table 2). However, the
global test for association was not significant among women (P = 0.12) or among men (P = 0.66).
For COMT, six haplotypes were predicted with z5% frequency. We found an inverse association between HapC (composed of variant alleles at SNP1 and SNP4) and all non-Hodgkin lymphoma and diffuse large-cell lymphoma present in 14% of controls, 11% of all non-Hodgkin lymphoma, and 9% of diffuse large-cell lymphoma cases (Table 4). These
Table 3. ORs and 95% CIs for diffuse large-cell lymphoma and follicular lymphoma associated with SNPs in PRL, CYP17A1, and COMT genes among San Francisco Bay Area HIV-negative white non-Hispanics, stratified by sex
Genotype
Controls
Diffuse large-cell lymphoma
Follicular lymphoma
Men
Women Men
(N = 463) (N = 221) (N = 57)
Women (N = 41)
Men (N = 55)
Women (N = 57)
n (%)* n (%)* n (%)* OR (95% CI)c n (%)* OR (95% CI)c n (%)* OR (95% CI)c n (%)* OR (95% CI)c
PRL
SNP1 1149GG
176 (38)
1149GT
216 (47)
1149TT
69 (15)
SNP2 1488AA
179 (39)
1488AG
217 (47)
1488GG
67 (14)
SNP3 214CC
279 (61)
214CT
155 (34)
214TT
25 (5)
SNP4 570GG
443 (96)
570GA/AA 20 (4)
CYP17
SNP1 34TT
167 (36)
34TC
228 (49)
34CC
68 (15)
SNP2 137GG
157 (34)
137GA
229 (50)
137AA
76 (16)
SNP3 195CC
161 (35)
195CA
231 (50)
195AA
69 (15)
SNP4 270AA
150 (32)
270AC
234 (51)
270CC
79 (17)
SNP5 105AA
225 (49)
105AC
198 (43)
105CC
36 (8)
SNP6 35TT
373 (81)
35TC
85 (18)
35CC
5 (1)
SNP7 75CC
171 (37)
75CG
229 (49)
75GG
63 (14)
SNP8 114AA
212 (46)
114AT
189 (41)
114TT
61 (13)
SNP9 2930GG
228 (49)
2930GT
193 (42)
2930TT
41 (9)
COMT
SNP1 701AA
208 (45)
701AG
204 (43)
701GG
49 (11)
701AG/GG 253 (55)
SNP2 186CC
140 (30)
186CT
205 (45)
186TT
115 (25)
186CT/TT 320 (70)
SNP3 108/158 VV 139 (30)
108/158 VM 208 (45)
108/158 MM 112 (24)
108/158
320 (70)
VM/MM
SNP4 6731AA
214 (47)
6731AG
194 (42)
6731GG
52 (11)
6731AG/GG 246 (53)
77 (35) 110 (50) 33 (15) 88 (40) 102 (46) 31 (14) 122 (55) 87 (40) 11 (5) 209 (95) 12 (5)
82 (37) 113 (51) 26 (12) 80 (36) 110 (50) 30 (14) 83 (38) 112 (51) 26 (12) 76 (34) 119 (54) 26 (12) 113 (52) 89 (41) 16 (7) 183 (83) 34 (15)
3 (1) 82 (37) 114 (52) 25 (11) 98 (44) 95 (43) 28 (13) 112 (51) 92 (42) 17 (8)
101 (46) 96 (43) 24 (11) 120 (54) 54 (25) 109 (50) 54 (25) 163 (75) 54 (25) 115 (52) 51 (23) 166 (75)
104 (47) 91 (41) 26 (12) 117 (53)
20 (35) 1.0 29 (51) 1.3 (0.69-2.4) 8 (14) 1.1 (0.47-2.7) 24 (42) 1.0 27 (47) 1.0 (0.56-1.9) 6 (11) 0.76 (0.29-2.0) 38 (67) 1.0 17 (30) 0.89 (0.48-1.6) 2 (4) 0.56 (0.13-2.5) 57 (100) 1.0 0 (0) --
18 (44) 1.0 17 (41) 0.65 (0.31-1.3) 6 (15) 0.73 (0.26-2.0) 18 (44) 1.0 17 (41) 0.77 (0.37-1.6) 6 (15) 0.87 (0.31-2.4) 25 (61) 1.0 14 (34) 0.72 (0.35-1.5) 2 (5) 0.83 (0.17-4.0) 40 (98) 1.0 1 (2) 0.42 (0.05-3.3)
29 (53) 1.0 22 (40) 0.62 (0.34-1.1)
4 (7) 0.39 (0.13-1.2) 30 (55) 1.0 22 (40) 0.63 (0.35-1.1)
3 (5) 0.30 (0.09-1.0) 35 (64) 1.0 18 (33) 0.99 (0.54-1.8)
2 (4) 0.65 (0.15-2.9) 55 (100) 1.0
0 (0) --
22 (39) 1.0 26 (46) 0.82 (0.43-1.6) 8 (14) 0.73 (0.29-1.8) 21 (37) 1.0 28 (49) 1.1 (0.57-2.1) 8 (14) 0.98 (0.39-2.5) 30 (53) 1.0 25 (44) 1.1 (0.58-2.0) 2 (4) 0.67 (0.14-3.2) 55 (96) 1.0 2 (4) 0.63 (0.14-2.9)
19 (33) 26 (46) 12 (21) 16 (28) 29 (51) 12 (21) 19 (33) 24 (42) 14 (25) 16 (28) 30 (53) 11 (19) 24 (42) 25 (44) 8 (14) 48 (84) 9 (16) 0 (0) 18 (32) 25 (44) 14 (25) 24 (42) 25 (44) 8 (14) 24 (42) 25 (44) 8 (14)
1.0 1.0 (0.56-2.0) 1.7 (0.78-3.8)
1.0 1.3 (0.66-2.4) 1.7 (0.77-3.9)
1.0 0.92 (0.49-1.8)
1.9 (0.88-4.1) 1.0
1.2 (0.63-2.3) 1.4 (0.61-3.2) 1.0 1.3 (0.69-2.3) 2.3 (0.94-5.8) 1.0 0.84 (0.39-1.8) -- 1.0 1.0 (0.55-2.0) 2.3 (1.0-4.9) 1.0 1.2 (0.67-2.3) 1.5 (0.62-3.6) 1.0 1.3 (0.72-2.4) 2.3 (0.94-5.7)
16 (39) 1.0 13 (32) 0.59 (0.27-1.3) 12 (29) 2.4 (1.0-5.7) 16 (39) 1.0 13 (32) 0.60 (0.27-1.3) 12 (29) 2.0 (0.85-4.8) 16 (39) 1.0 13 (32) 0.60 (0.27-1.3) 12 (29) 2.4 (1.0-5.8) 16 (39) 1.0 13 (32) 0.52 (0.24-1.2) 12 (29) 2.2 (0.91-5.2) 22 (54) 1.0 13 (32) 0.76 (0.36-1.6) 6 (14) 1.9 (0.68-5.5) 30 (73) 1.0 10 (24) 1.8 (0.79-4.0) 1 (2) 1.8 (0.18-18) 16 (39) 1.0 13 (32) 0.59 (0.27-1.3) 12 (29) 2.5 (1.0-5.9) 22 (54) 1.0 13 (32) 0.63 (0.30-1.3) 6 (15) 1.0 (0.37-2.8) 21 (51) 1.0 13 (32) 0.76 (0.36-1.6) 7 (17) 2.2 (0.81-6.0)
24 (44) 24 (45)
7 (13) 21 (38) 25 (46)
9 (16) 24 (44) 24 (44)
7 (13) 21 (38) 24 (44) 10 (18) 26 (47) 25 (45)
4 (7) 50 (91)
5 (9) 0 (0) 24 (44) 24 (44) 7 (13) 26 (47) 25 (45) 4 (7) 26 (47) 25 (45) 4 (7)
1.0 0.76 (0.41-1.4) 0.82 (0.33-2.0) 1.0 0.82 (0.44-1.5) 0.99 (0.43-2.3) 1.0 0.73 (0.40-1.3) 0.78 (0.32-1.9) 1.0 0.73 (0.39-1.4) 0.99 (0.44-2.2) 1.0 1.2 (0.64-2.1) 1.1 (0.35-3.3) 1.0 0.47 (0.18-1.2) -- 1.0 0.77 (0.42-1.4) 0.88 (0.36-2.2) 1.0 1.15 (0.64-2.1) 0.65 (0.22-2.0) 1.0 1.2 (0.68-2.2) 1.0 (0.34-3.2)
21 (37) 1.0 27 (47) 0.93 (0.49-1.8) 9 (16) 1.4 (0.56-3.4) 18 (32) 1.0 30 (53) 1.2 (0.62-2.3) 9 (16) 1.3 (0.54-3.3) 21 (37) 1.0 25 (44) 0.87 (0.45-1.7) 11 (19) 1.7 (0.71-4.0) 18 (32) 1.0 29 (51) 1.0 (0.52-1.9) 10 (18) 1.6 (0.66-4.0) 31 (54) 1.0 20 (35) 0.85 (0.45-1.6) 6 (11) 1.3 (0.47-3.7) 44 (77) 1.0 12 (21) 1.5 (0.69-3.1) 1 (2) 1.2 (0.12-12) 22 (39) 1.0 26 (46) 0.84 (0.44-1.6) 9 (16) 1.4 (0.56-3.4) 31 (54) 1.0 19 (33) 0.67 (0.35-1.3) 7 (12) 0.84 (0.33-2.1) 31 (54) 1.0 20 (35) 0.82 (0.44-1.5) 6 (10) 1.3 (0.45-3.5)
25 (44) 28 (49) 4 (7) 32 (56) 18 (32) 24 (42) 15 (26) 39 (68) 17 (30) 25 (44) 15 (26) 40 (70)
1.0 1.1 (0.61-2.0)
0.63 (0.21-1.9) 1.0 (0.57-1.8)
1.0 0.94 (0.49-1.8)
1.1 (0.50-2.2) 0.98 (0.54-1.8) 1.0
1.0 (0.52-2.0) 1.1 (0.52-2.4) 1.1 (0.57-1.9)
25 (61) 1.0 14 (34) 0.60 (0.29-1.2) 2 (5) 0.35 (0.08-1.6) 16 (39) 0.55 (0.28-1.1) 6 (15) 1.0 24 (60) 2.1 (0.80-5.5) 10 (25) 1.8 (0.60-5.4) 34 (85) 2.0 (0.79-5.1) 6 (15) 1.0 23 (59) 1.9 (0.71-4.9) 10 (26) 1.8 (0.62-5.5) 33 (85) 1.9 (0.73-4.8)
22 (40) 29 (53)
4 (7) 33 (60) 18 (33) 25 (45) 12 (22) 37 (67) 16 (29) 29 (53) 10 (18) 39 (71)
1.0 1.3 (0.74-2.4) 0.78 (0.25-2.4) 1.23 (0.69-2.2) 1.0 0.97 (0.50-1.9) 0.79 (0.36-1.7) 0.90 (0.49-1.7) 1.0 1.2 (0.64-2.4) 0.73 (0.32-1.7) 1.1 (0.56-2.0)
38 (67) 1.0 18 (32) 0.50 (0.26-0.94) 1 (2) 0.11 (0.01-0.82) 19 (33) 0.42 (0.23-0.78) 12 (21) 1.0 27 (48) 1.3 (0.58-2.7) 17 (30) 1.6 (0.70-3.8) 44 (79) 1.4 (0.67-2.8) 10 (18) 1.0 30 (53) 1.6 (0.70-3.5) 17 (30) 2.0 (0.84-4.9) 47 (82) 1.7 (0.80-3.7)
28 (50) 23 (41) 5 (9) 28 (50)
1.0 0.91 (0.50-1.6) 0.78 (0.28-2.1) 0.88 (0.50-2.1)
24 (59) 1.0 15 (37) 0.71 (0.35-1.4) 2 (5) 0.33 (0.07-1.5) 17 (41) 0.63 (0.32-1.2)
32 (58) 17 (31)
6 (11) 23 (42)
1.0 27 (47) 1.0 0.60 (0.32-1.1) 27 (47) 1.1 (0.60-2.0) 0.81 (0.32-2.1) 3 (5) 0.43 (0.12-1.5) 0.64 (0.36-1.1) 30 (53) 0.96 (0.53-1.7)
*Numbers may not add to 308 cases and 684 controls due to missing genotypes. cORs and 95% CIs calculated using unconditional logistic regression, adjusted for age.
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
Cancer Epidemiology, Biomarkers & Prevention 2397
Table 4. ORs and 95% CIs for non-Hodgkin lymphoma associated with haplotypes in PRL, CYP17A1, and COMT genes among HIV-negative, white non-Hispanic men and women, San Francisco Bay Area, 1988-1995
Haplotype*,c
Controls N = 684
All non-Hodgkin lymphoma
N = 308
OR (95% CI)b
Diffuse large-cell lymphoma
N = 98
OR (95% CI)b
Follicular lymphoma N = 112 OR (95% CI)b
PRL A B C D Otherx
CYP17A1 A B C Otherx
COMT A B C D E F Otherx
0-0-0 1-1-1 1-1-0 0-1-0 Pooled
(low frequency)
0-0-0-0-0-0-0-0-0 1-1-1-1-1-0-1-1-1 1-1-1-1-0-1-1-0-0 Pooled
(low frequency)
0-1-1-0 1-0-0-0 1-0-0-1 0-0-0-1 0-0-0-0 0-1-1-1 Pooled
(low frequency)
0.55 0.20 0.13 0.06 0.07
0.51 0.24 0.08 0.16
0.37 0.14 0.14 0.12 0.12 0.06 0.06
0.59 0.17 0.11 0.05 0.08
0.50 0.27 0.09 0.14
0.40 0.13 0.11 0.13 0.12 0.05 0.05
1.0 (reference) 0.78 (0.60-1.0) 0.85 (0.61-1.2) 0.81 (0.52-1.3) 1.2 (0.84-1.8)
1.0 (reference) 1.2 (0.95-1.5) 1.1 (0.79-1.6) 0.91 (0.71-1.2)
1.0 (reference) 0.89 (0.65-1.2) 0.74 (0.53-1.0) 1.2 (0.87-1.7) 0.93 (0.66-1.3) 0.84 (0.53-1.3) 1.0 (0.65-1.6)
0.60 0.18 0.14 0.05 0.03
0.50 0.32 0.10 0.07
0.41 0.12 0.09 0.14 0.13 0.04 0.07
1.0 (reference) 0.88 (0.59-1.3) 1.2 (0.75-1.9) 0.80 (0.39-1.6) 0.40 (0.16-1.0)
1.0 (reference) 1.6 (1.1-2.2) 1.3 (0.80-2.2) 0.46 (0.27-0.79)
1.0 (reference) 0.86 (0.52-1.4) 0.54 (0.30-0.97) 1.3 (0.77-2.1) 1.1 (0.64-1.7) 0.58 (0.24-1.4) 1.2 (0.64-2.3)
0.63 0.20 0.09 0.03 0.05
0.57 0.28 0.08 0.07
0.42 0.12 0.11 0.11 0.14 0.04 0.05
1.0 (reference) 0.94 (0.65-1.4) 0.66 (0.39-1.1) 0.41 (0.18-0.96) 0.67 (0.34-1.3)
1.0 (reference) 1.2 (0.90-1.7) 1.0 (0.59-1.7) 0.48 (0.29-0.78)
1.0 (reference) 0.87 (0.54-1.4) 0.72 (0.44-1.2) 0.93 (0.56-1.5) 1.2 (0.76-1.9) 0.63 (0.29-1.4) 0.97 (0.48-2.0)
*Jointly adjusted haplotypes estimated using tagSNPs implementation of the EM algorithm. CYP17A1 haplotypes are described left to right: SNP1-SNP9, wild-type allele designated ``0'' and variant allele designated ``1''. cPRL haplotypes are described left to right: SNP1-SNP3, wild-type allele designated ``0'' and variant allele designated ``1''. SNP4 is not included because only the
wild-type allele cosegregates with haplotypes z0.05. bHaplotype ORs and 95% CIs estimated using a single-imputation approach, modeled using unconditional logistic regression adjusted for age and sex. All haplotype
categories for a gene are included in the same model using the highest-prevalence haplotype as the reference category. xHaplotypes with estimated frequencies <0.05 are pooled into a single category.
associations seemed to be due to the difference in haplotype frequencies in women, but the global test of association was not statistically significant for women (P = 0.20) or men (P = 0.41; Supplementary Table 2).
Oral Contraceptive and Non-Oral Contraceptive Hormone Use and Non-Hodgkin Lymphoma Risk among Women. Among all white non-Hispanic women in our study population, non-Hodgkin lymphoma risk was reduced by 35% among those who ever had used oral contraceptives compared with never users (Table 5). There also was a decreasing trend in ORs with increasing years of oral contraceptive use (P for trend = 0.001). Postmenopausal status and ever use of non-oral contraceptive hormones were not associated with non-Hodgkin lymphoma. Because long-term use of non-oralcontraceptive hormones may be related to hysterectomy, analyses were stratified by history of hysterectomy or oophorectomy. Among women with no history of hysterectomy/oophorectomy, ORs decreased with increasing years of use, whereas among women who had a history of
hysterectomy/oophorectomy, the OR was increased for shorter duration of use. In general, risk estimates from analyses restricted to genotyped women were only somewhat consistent with results from analyses among all women. In this restricted group of women, ORs for non-Hodgkin lymphoma associated with use of exogenous estrogens were imprecise and were consistently less than unity, but not different from a chance occurrence. The small number of exposed patients restricted more detailed analyses of duration of hormone use in this group. Due to sparse data, we also did not evaluate duration of use by non-Hodgkin lymphoma subtype or gene-environment interactions.
Discussion
Here we report an association between common genetic variants in the CYP17A1, PRL, and COMT genes and risk of non-Hodgkin lymphoma. Among both men and women, we observed increased risk for all non-Hodgkin lymphoma and
Figure 2. Pairwise measures of linkage disequilibrium (DV 100) among PRL (A), CYP17A1 (B), and COMT (C) loci genotyped in the control population. These linkage disequilibrium plots were generated using Haploview (27).
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
2398 CYP17A1, PRL, and COMT Polymorphisms and Non-Hodgkin Lymphoma Risk
Table 5. ORs and 95% CIs for non-Hodgkin lymphoma associated with exogenous estrogens among all HIV-negative, white non-Hispanic women and restricted to those for whom DNA was genotyped for hormone-related SNPs, San Francisco Bay Area
Characteristic
All women
Genotyped women
Cases (N = 451)
Controls (N = 678)
OR (95% CI)*
Cases (N = 134)
Controls (N = 220)
OR (95% CI)*
n (%)
n (%)
n (%)
n (%)
Ever oral contraceptive use No 270 (60) Yes 180 (40)
Duration of oral contraceptive use in years V5 125 (28) >5 55 (12)
354 (52) 324 (48)
206 (30) 118 (17)
Postmenopausal status Premenopausal Postmenopausal
100 (22) 351 (78)
155 (23) 523 (77)
1.0 (reference) 0.65 (0.49-0.86)
0.71 (0.52-0.97) P0.t5re4nd(0=.370-.00.0810c)
1.0 (reference) 1.1 (0.69-1.6)
74 (56) 59 (44)
39 (29) 20 (15)
24 (18) 110 (82)
119 (54) 101 (46)
58 (26) 43 (20)
50 (23) 170 (77)
1.0 (reference) 0.93 (0.56-1.6)
1.1 (0.61-1.4) 0.74 (0.38-1.4) Ptrend = 0.45c
1.0 (reference) 1.8 (0.81-4.2)
Postmenopausal women
Ever non-oral contraceptive hormone use
No
124 (35)
176 (34)
1.0 (reference)
43 (39)
Yes
227 (65)
346 (66)
0.93 (0.70-1.2)
67 (61)
Duration of non-oral contraceptive hormone use in years
V5
116 (33)
139 (27)
1.2 (0.85-1.7)
28 (26)
>5
110 (31)
204 (39)
0.77 (0.55-1.1)
38 (35)
P trend = 0.11c
Duration of non-oral contraceptive hormone use among women without a hysterectomy/oophorectomy
No use
91 (48)
120 (41)
1.0 (reference)
30 (43)
V5 y
60 (32)
89 (31)
0.83 (0.53-1.3)
20 (29)
>5 y
38 (20)
81 (28)
0.60 (0.37-0.96)
19 (28)
P trend = 0.04c
Duration of non-oral contraceptive hormone use among women with hysterectomy/oophorectomy
No use
33 (20)
56 (24)
1.0 (reference)
13 (32)
V5 y
56 (35)
50 (22)
1.9 (1.1-3.4)
8 (20)
>5 y
72 (45)
123 (54)
0.95 (0.56-1.6) P trend = 0.44c
19 (48)
54 (32) 116 (68)
48 (29) 66 (39)
36 (38) 32 (34) 27 (28)
18 (25) 16 (22) 39 (53)
1.0 (reference) 0.70 (0.42-1.2)
0.66 (0.35-1.2) 0.72 (0.41-1.3) Ptrend = 0.26c
1.0 (reference) 0.63 (0.29-1.4) 0.81 (0.38-1.7) Ptrend = 0.53
1.0 (reference) 0.66 (0.22-2.0) 0.69 (0.28-1.7) Ptrend = 0.45
*ORs and 95% CIs computed using unconditional logistic regression adjusted for age. Reference group is never users. cP trend based on m2 statistic for ordinal duration of use from age-adjusted unconditional logistic regression.
for diffuse large-cell lymphoma, particularly with the CYP17A1 34CC genotype. In CYP17A1 haplotype analyses, a high-risk haplotype for diffuse large-cell lymphoma (HapB) was more frequent among cases than controls. These data are consistent with our recent findings in another large nonHodgkin lymphoma case-control study conducted in the United Kingdom where the CYP17A1 34CC genotype was associated with a similar elevated risk (31). The replication of this finding in both men and women in a second study suggests that an association exists between the CYP17A1 34CC genotype and non-Hodgkin lymphoma risk. Further, the similar magnitudes of effect in both sexes suggest that testosterone and other cholesterol metabolites downstream of CYP17A1, or other factors common to both sexes, may be involved in the pathogenesis of non-Hodgkin lymphoma. The higher incidence of diffuse large-cell lymphoma among men compared with women (32) is consistent with the notion that steroids in this pathway other than estrogens influence diffuse large-cell lymphoma risk.
CYP17A1 exhibits both 17a-hydroxylase and 17,20-lyase enzymatic activities in ovarian theca cells, testicular Leydig cells, and in the adrenal cortex, which are essential for sex steroid and glucocorticoid production (33). Through the y5 pathway, CYP17A1 converts pregnenolone to dehydroepiandrosterone, the precursor for estrogen and testosterone (Fig. 3; ref. 34). Whereas the CYP17A1 34CC genotype has been associated with elevated estrogen levels in women, an association with increased estrogen or testosterone levels in men is uncertain (reviewed in refs. 18, 35). Thus, further studies may be warranted to test whether testosterone or its major metabolite, 5a-dihydrotestosterone, potentiates lymphoma
risk. Through the y4 pathway, CYP17A1 also converts progesterone to 17a-hydroxyprogesterone, a substrate in the production of cortisol (Fig. 3; ref. 34). Cortisol can either suppress or stimulate immune function in a dose-dependent manner, so modulation of its production could potentially influence nonHodgkin lymphoma risk. Currently, no functional studies have reported whether the CYP17A1 34T>C polymorphism alters glucocorticoid production. Additional studies of SNPs in genes involved in glucocorticoid and sex hormone production such as CYP21A2, CYP11B1, 3h-hydroxysteroid dehydrogenase (3b-HSD), 17b-HSD, CYP19, and 5a-reductase type 2 (SRD5A2) may clarify this pathway in lymphomagenesis.
We also observed that genetic variants in COMT were associated with increased risk of non-Hodgkin lymphoma in women. Specifically, the COMT SNP3 variant (108/158Met), which was related to reduced COMT enzyme activity, elevated circulating estradiol (36) and 2-hydroxyestrone levels (37), and increased breast cancer risk (38, 39), was associated with borderline elevated risk of non-Hodgkin lymphoma, particularly follicular lymphoma, in women. Reduced COMT activity decreases the detoxification of catechol estrogens to the less toxic methoxy derivatives (40), notably 2-methoxyestradiol, an anticarcinogen that induces apoptosis and inhibits angiogenesis and tumor cell growth (41). In contrast, the intronic COMT SNP1 variant allele was associated with a reduced risk for non-Hodgkin lymphoma in women, which was likely driven by the reduced risk observed for follicular lymphoma. Whether the effect of this SNP is due to a function of enhanced COMT expression or is linked to an unknown causal variant remains to be determined. Nonetheless, these findings suggest a possible role of
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
Cancer Epidemiology, Biomarkers & Prevention
catechol estrogens in the pathogenesis of follicular lymphoma in women through genotoxic mechanisms that involve oxidative DNA damage, DNA double-strand breaks, and/or tumor initiation. Alternatively, the increased risk of diffuse large-cell lymphoma associated with the CYP17 34CC genotype in women and men suggests enhanced B-cell activation, proliferation, and survival as a possible mechanism through estrogen receptor- or testosterone receptor-mediated effects.
In the haplotype analyses, COMT HapC was identified as a low-risk haplotype for non-Hodgkin lymphoma in both men and women. This same haplotype recently was described as a high-risk haplotype for schizophrenia (42), where the population frequency was similar to that for controls in our population. This haplotype was associated with reduced MBCOMT expression and elevated dopamine levels in the brain. Dopamine exerts profound effects on immune function, is produced by lymphocytes (43), and its receptors are found on lymphocytes, macrophages, and neutrophils (44). Thus, it is possible that interactions between the nervous and immune systems that involve dopamine and/or other neurotransmitters alter the risk for non-Hodgkin lymphoma.
Prolactin also regulates lymphocyte function and is synthesized by these cells (45). In the present study, the PRL 1149T variant (SNP1) was inversely associated with all non-Hodgkin lymphoma and with follicular lymphoma both in men and in women. Multiple promoters and start sites present in the PRL gene modulate pituitary and extrapituitary expression (10). The PRL 1149T allele, located in the extrapituitary promoter, is associated with reduced promoter activity and prolactin mRNA levels in lymphocytes (11), whereas the 1149G allele may abrogate the effect of prolactin on lymphoproliferation (46). Prolactin promotes both cell-mediated and humoral immune responses through signaling pathways, including Jak/Stat and mitogen-activated protein kinase, resulting in target gene expression (47), stimulation of B- and T-cell proliferation, proinflammatory cytokine production, and B-cell growth arrest (reviewed in ref. 1). Alternatively, estradiol exerts predominantly a humoral immune response via T-cell suppression and B-cell proliferation, enhanced antibody production, and B-cell survival (1). In animal studies, treatment with either estradiol (48) or prolactin (49) leads to the rescue of autoreactive B-cells from apoptosis by up-regulating BCL-2 expression (50), indicating a role of these hormones in autoimmune disease. Furthermore, testosterone and its major endogenous metabolite 5a-dihydrotestosterone may also exert pleiotropic effects on the immune system. 5a-dihydrotestosterone promotes proliferation of prostate epithelial cells through up-regulation of the BCL-2 and nuclear factor nB pathways (51), but little is known about its proliferative and antiapoptotic effects on B-cells.
Reduced ORs for non-Hodgkin lymphoma in long-term oral contraceptive users in our analyses are somewhat consistent with the results of two other studies (7, 8) but different from that of one study (52). Although the epidemiologic data have been inconsistent, it is biologically plausible that long-term oral contraceptive use and/or women's exposure to exogenous estrogens during the reproductive years may alter non-Hodgkin lymphoma risk. Oral contraceptive use inhibits ovulation and the cyclic fluxes in estrogen and progesterone production during the menstrual cycle. Furthermore, oral contraceptive use is associated with significantly reduced levels of serum testosterone and dehydroepiandrosterone sulfate and elevated levels of serum hormone binding globulin (53), a protein that binds to and restricts the biological action of estradiol and testosterone. It is plausible that long-term oral contraceptive use reduces the overall lifetime exposure to estrogens, thus reducing proliferation and enhanced survival of B-cells and risk for nonHodgkin lymphoma.
Although imprecise, the magnitude of the ORs associated with history of non-oral contraceptive hormone use among the genotyped and nongenotyped postmenopausal women tended to be consistent with the borderline reduced estimates published in most studies (7, 8, 52, 54). Exceptions to these results that show a somewhat inverse relationship are the increased risks for follicular lymphoma associated with hormone therapy among postmenopausal women in the Iowa Women's Health study (6) and for all non-Hodgkin lymphomas among women in Los Angeles County (5). The estimates from these two studies were somewhat similar to our results among women who had had a hysterectomy or oophorectomy and used non-oral-contraceptive hormones for 5 or fewer years. In general, the estimates from most previous studies and our study were imprecise and based on a small number of exposed patients. Studies that include a large number of exposed women and detailed information about hormone use are required to determine whether these observed associations are true. However, given that estrogens influence immune function, these epidemiologic results are biologically plausible and are consistent with our genetic data.
As with all exposure data collected in case-control studies, these data are subject to recall bias and exposure misclassification. To address these known problems, hormone-related information was collected from both case and control participants in a consistent manner, with photographs of the hormone types, brands, and manufacturers' packaging shown to all participants to assist recall. Unless patients perceived that oral contraceptive or non-oral contraceptive hormone use was associated with their disease, we would expect the misclassification to be nondifferential and the recall bias to be minimal. Thus, the estimated ORs are likely to be biased
Figure 3. Schematic of the synthesis and metabolism of estradiol and testosterone. CYP19, cytochrome P450 19; SRD5A2, 5a-reductase type 2; 3b-HSD, 3h-hydroxysteroid dehydrogenase; 17b-HSD, 17hhydroxysteroid dehydrogenase; DHT, 5a-dihydrotestosterone.
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
2400 CYP17A1, PRL, and COMT Polymorphisms and Non-Hodgkin Lymphoma Risk
toward the null especially for details about oral contraceptive and non-oral-contraceptive hormone use. Furthermore, the potential heterogeneity of non-oral-contraceptive hormone use related to other characteristics, including reason for use and type of hormone used, may have affected the estimates for these factors. Power to test associations for more detailed analyses in the restricted population of genotyped women was low. Analyses of gene-environment interactions were not pursued because estimates obtained from the analyses of exogenous hormone use in the restricted population of women were not entirely consistent with those obtained for the complete group of women and may have resulted in spurious gene-environment effects. Although these results are consistent with those from some previous epidemiologic investigations of hormone use and non-Hodgkin lymphoma, confirmation in larger studies is required.
Compared with all HIV-negative patients (regardless of eligibility) who did not provide a blood specimen, patients who gave blood were less likely to have had high-grade lymphomas. If treatment or prognosis for patients with highgrade lymphomas was related to blood collection, then our results may be comparable only to patients with better prognosis or less urgent treatment regimens. In addition, compared with noninterviewed patients, patients who were interviewed had a higher proportion of low-grade lymphomas (55). If all HIV-negative patients had been interviewed, the overall proportion of low-grade lymphomas would have been somewhat lower, whereas there would have been little change in the proportion of high-grade lymphomas. Given that lowgrade lymphomas are somewhat overrepresented among HIVnegative patients in our overall study population and among those who gave blood, our estimates for all non-Hodgkin lymphoma may be biased slightly away from the null for factors related to low-grade disease.
Additional limitations of this study are similar to other casecontrol studies of genetic associations and complex diseases. Like many polygenic diseases, the risk alleles studied are not likely to be sufficient to induce non-Hodgkin lymphoma and require replication and confirmation in additional larger studies. However, we have attempted to address some of the shortcomings of genetic association studies by investigating haplotypes in addition to SNPs, assessing the extent of linkage disequilibrium, considering haplotypes and SNPs at loci that function in the same or related biological pathways, restricting analyses to white non-Hispanics, and including epidemiologic measures of estrogen exposure to provide a more comprehensive evaluation of the potential role of estrogen in the development of non-Hodgkin lymphoma.
Overall, our observations suggest PRL, CYP17A1, and COMT as non-Hodgkin lymphoma susceptibility genes and provide support for the role of prolactin, estrogens, and possibly testosterone, cortisol, and/or dopamine in the pathogenesis of lymphoma. Our findings suggest that in both men and women, lymphocyte prolactin and circulating estrogen levels may be inversely associated with follicular lymphoma and diffuse large-cell lymphoma risk, respectively. These effects may be promoted through similar pathways involving enhanced B-cell activation, proliferation, and survival, although prolactin also can elicit a strong proinflammatory cytokine response. Our results among women suggest a role for catechol estrogens, possibly through genotoxic mechanisms, in the initiation of follicular lymphoma. The positive association between diffuse large-cell lymphoma and the CYP17 34CC genotype among men and women raises the question of whether this SNP has an effect on testosterone or cortisol production (not measured in this study) and whether these hormones influence lymphoma risk. Functional studies will be needed to address these questions. Finally, the inverse association between diffuse large cell lymphoma and COMT HapC, related to elevated dopamine levels, suggests that
although lymphoma is not considered a classic endocrinological tumor, interactions involving aberrant cross-talk between the neuroendocrine-immune networks may play a role in nonHodgkin lymphoma pathogenesis. Further investigation of these ideas is warranted in independent studies, ideally as part of a large consortium such as InterLymph.
Acknowledgments
We thank Katherine Lazaruk and Tony Dodge (Applied Biosystems, Inc., Foster City, CA) for assistance with COMT assays.
References
1. McMurray RW. Estrogen, prolactin, and autoimmunity: actions and interactions. Int Immunopharmacol 2001;1:995 1008.
2. Ben-Jonathan N, Liby K, McFarland M, Zinger M. Prolactin as an autocrine/ paracrine growth factor in human cancer. Trends Endocrinol Metab 2002;13: 245 50.
3. Jakob F, Tony HP, Schneider D, Thole HH. Immunological detection of the oestradiol receptor protein in cell lines derived from the lymphatic system and the haematopoietic system: variability of specific hormone binding in vitro . J Endocrinol 1992;134:397 404.
4. Gado K, Pallinger E, Kovacs P, et al. Prolactin influences proliferation and apoptosis of a human IgE secreting myeloma cell line, U266. Immunol Lett 2002;82:191 6.
5. Bernstein L, Ross RK. Prior medication use and health history as risk factors for non-Hodgkin's lymphoma: preliminary results from a case-control study in Los Angeles County. Cancer Res 1992;52:5510 5.
6. Cerhan JR, Vachon CM, Habermann TM, et al. Hormone replacement therapy and risk of non-Hodgkin lymphoma and chronic lymphocytic leukemia. Cancer Epidemiol Biomarkers Prev 2002;11:1466 71.
7. Beiderbeck AB, Holly EA, Sturkenboom MC, Coebergh JW, Stricker BH, Leufkens HG. No increased risk of non-Hodgkin's lymphoma with steroids, estrogens and psychotropics (Netherlands). Cancer Causes Control 2003;14: 639 44.
8. Nelson RA, Levine AM, Bernstein L. Reproductive factors and risk of intermediate- or high-grade B-cell non-Hodgkin's lymphoma in women. J Clin Oncol 2001;19:1381 7.
9. Brennan P, Hajeer A, Ong KR, et al. Allelic markers close to prolactin are associated with HLA-DRB1 susceptibility alleles among women with rheumatoid arthritis and systemic lupus erythematosus. Arthritis Rheum 1997;40:1383 6.
10. Ben-Jonathan N, Mershon JL, Allen DL, Steinmetz RW. Extrapituitary prolactin: distribution, regulation, functions, and clinical aspects. Endocr Rev 1996;17:639 69.
11. Stevens A, Ray D, Alansari A, et al. Characterization of a prolactin gene polymorphism and its associations with systemic lupus erythematosus. Arthritis Rheum 2001;44:2358 66.
12. Carey AH, Waterworth D, Patel K, et al. Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet 1994;3:1873 6.
13. Bergman-Jungestrom M, Gentile M, Lundin AC, Wingren S. Association between CYP17 gene polymorphism and risk of breast cancer in young women. Int J Cancer 1999;84:350 3.
14. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Henderson BE. Cytochrome P 450c17a gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 1998;58:585 7.
15. Gsur A, Bernhofer G, Hinteregger S, et al. A polymorphism in the CYP17 gene is associated with prostate cancer risk. Int J Cancer 2000;87:434 7.
16. Kittles RA, Panguluri RK, Chen W, et al. Cyp17 promoter variant associated with prostate cancer aggressiveness in African Americans. Cancer Epidemiol Biomarkers Prev 2001;10:943 7.
17. Hong CC, Thompson HJ, Jiang C, et al. Association between the T27C polymorphism in the cytochrome P 450 c17a (CYP17) gene and risk factors for breast cancer. Breast Cancer Res Treat 2004;88:217 30.
18. Sharp L, Cardy AH, Cotton SC, Little J. CYP17 gene polymorphisms: prevalence and associations with hormone levels and related factors. a HuGE review. Am J Epidemiol 2004;160:729 40.
19. Huh MM, Friedhoff AJ. Multiple molecular forms of catechol-O -methyltransferase. Evidence for two distinct forms, and their purification and physical characterization. J Biol Chem 1979;254:299 308.
20. Dawling S, Roodi N, Mernaugh RL, Wang X, Parl FF. Catechol-O methyltransferase (COMT)-mediated metabolism of catechol estrogens: comparison of wild-type and variant COMT isoforms. Cancer Res 2001;61: 6716 22.
21. Syvanen AC, Tilgmann C, Rinne J, Ulmanen I. Genetic polymorphism of catechol-O-methyltransferase (COMT): correlation of genotype with individual variation of S-COMT activity and comparison of the allele frequencies in the normal population and parkinsonian patients in Finland. Pharmacogenetics 1997;7:65 71.
22. Holly EA, Bracci PM. Population-based study of non-Hodgkin lymphoma, histology, and medical history among human immunodeficiency virusnegative participants in San Francisco. Am J Epidemiol 2003;158:316 27.
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005
Cancer Epidemiology, Biomarkers & Prevention 2401
23. Holly EA, Lele C. Non-Hodgkin's lymphoma in HIV-positive and HIVnegative homosexual men in the San Francisco Bay Area: allergies, prior medication use, and sexual practices. J Acquir Immune Defic Syndr Hum Retrovirol 1997;15:211 22.
24. Holly EA, Lele C, Bracci PM, McGrath MS. Case-control study of nonHodgkin's lymphoma among women and heterosexual men in the San Francisco Bay Area, California. Am J Epidemiol 1999;150:375 89.
25. Harris NL, Jaffe ES, Stein H, et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 1994;84:1361 92.
26. Jaffe ES, Harris NL, Stein H, Vardiman JV, editors. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. IARC Press: Lyon: 2001.
27. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of linkage disequilibrium and haplotype maps. Bioinformatics 2005;21:263 5. Epub 2004 Aug 5.
28. Stram DO, Leigh Pearce C, Bretsky P, et al. Modeling and E-M estimation of haplotype-specific relative risks from genotype data for a case-control study of unrelated individuals. Hum Hered 2003;55:179 90.
29. Zaykin DV, Westfall PH, Young SS, Karnoub MA, Wagner MJ, Ehm MG. Testing association of statistically inferred haplotypes with discrete and continuous traits in samples of unrelated individuals. Hum Hered 2002; 53:79 91.
30. Loukola A, Chadha M, Penn SG, et al. Comprehensive evaluation of the association between prostate cancer and genotypes/haplotypes in CYP17A1, CYP3A4, and SRD5A2. Eur J Hum Genet 2004;12:321 32.
31. Skibola CF, Lightfoot T, Agana L, et al. Polymorphisms in cytochrome P 450 17A1 and risk of non-Hodgkin lymphoma. Br J Haematol 2005;129:618 21.
32. Groves FD, Linet MS, Travis LB, Devesa SS. Cancer surveillance series: nonHodgkin's lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 2000;92:1240 51.
33. Nakajin S, Shively JE, Yuan PM, Hall PF. Microsomal cytochrome P -450 from neonatal pig testis: two enzymatic activities (17a-hydroxylase and c17,20-lyase) associated with one protein. Biochemistry 1981;20:4037 42.
34. Conley AJ, Bird IM. The role of cytochrome P 450 17a-hydroxylase and 3hhydroxysteroid dehydrogenase in the integration of gonadal and adrenal steroidogenesis via the y5 and y4 pathways of steroidogenesis in mammals. Biol Reprod 1997;56:789 99.
35. Ntais C, Polycarpou A, Ioannidis JP. Association of the CYP17 gene polymorphism with the risk of prostate cancer: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2003;12:120 6.
36. Worda C, Sator MO, Schneeberger C, Jantschev T, Ferlitsch K, Huber JC. Influence of the catechol-O-methyltransferase (COMT) codon 158 polymorphism on estrogen levels in women. Hum Reprod 2003;18:262 6.
37. Tworoger SS, Chubak J, Aiello EJ, et al. Association of CYP17, CYP19, CYP1B1, and COMT polymorphisms with serum and urinary sex hormone concentrations in postmenopausal women. Cancer Epidemiol Biomarkers Prev 2004;13:94 101.
38. Huang CS, Chern HD, Chang KJ, Cheng CW, Hsu SM, Shen CY. Breast
cancer risk associated with genotype polymorphism of the estrogenmetabolizing genes CYP17, CYP1A1, and COMT: a multigenic study on cancer susceptibility. Cancer Res 1999;59:4870 5. 39. Lavigne JA, Helzlsouer KJ, Huang HY, et al. An association between the allele coding for a low activity variant of catechol-O-methyltransferase and the risk for breast cancer. Cancer Res 1997;57:5493 7. 40. Weinshilboum RM, Otterness DM, Szumlanski CL. Methylation pharmacogenetics: catechol O-methyltransferase, thiopurine methyltransferase, and histamine N -methyltransferase. Annu Rev Pharmacol Toxicol 1999;39:19 52. 41. Klauber N, Parangi S, Flynn E, Hamel E, D'Amato RJ. Inhibition of angiogenesis and breast cancer in mice by the microtubule inhibitors 2methoxyestradiol and taxol. Cancer Res 1997;57:81 6. 42. Bray NJ, Buckland PR, Williams NM, et al. A haplotype implicated in schizophrenia susceptibility is associated with reduced COMT expression in human brain. Am J Hum Genet 2003;73:152 61. 43. Bergquist J, Josefsson E, Tarkowski A, Ekman R, Ewing A. Measurements of catecholamine-mediated apoptosis of immunocompetent cells by capillary electrophoresis. Electrophoresis 1997;18:1760 6. 44. Basu S, Dasgupta PS. Dopamine, a neurotransmitter, influences the immune system. J Neuroimmunol 2000;102:113 24. 45. Sabharwal P, Glaser R, Lafuse W, et al. Prolactin synthesized and secreted by human peripheral blood mononuclear cells: an autocrine growth factor for lymphoproliferation. Proc Natl Acad Sci U S A 1992;89:7713 6. 46. Stevens A, Ray DW, Worthington J, Davis JR. Polymorphisms of the human prolactin gene--implications for production of lymphocyte prolactin and systemic lupus erythematosus. Lupus 2001;10:676 83. 47. Yu-Lee L, Luo G, Moutoussamy S, Finidori J. Prolactin and growth hormone signal transduction in lymphohaemopoietic cells. Cell Mol Life Sci 1998;54:1067 75. 48. Grimaldi CM, Cleary J, Dagtas AS, Moussai D, Diamond B. Estrogen alters thresholds for B cell apoptosis and activation. J Clin Invest 2002;109: 1625 33. 49. Peeva E, Michael D, Cleary J, Rice J, Chen X, Diamond B. Prolactin modulates the naive B cell repertoire. J Clin Invest 2003;111:275 83. 50. Buckley AR. Prolactin, a lymphocyte growth and survival factor. Lupus 2001;10:684 90. 51. Coffey RN, Watson RW, O'Neill AJ, Mc Eleny K, Fitzpatrick JM. Androgenmediated resistance to apoptosis. Prostate 2002;53:300 9. 52. Zhang Y, Holford TR, Leaderer B, et al. Prior medical conditions and medication use and risk of non-Hodgkin lymphoma in Connecticut United States women. Cancer Causes Control 2004;15:419 28. 53. Wiegratz I, Kutschera E, Lee JH, et al. Effect of four different oral contraceptives on various sex hormones and serum-binding globulins. Contraception 2003;67:25 32. 54. Altieri A, Gallus S, Franceschi S, et al. Hormone replacement therapy and risk of lymphomas and myelomas. Eur J Cancer Prev 2004;13:349 51. 55. Holly EA, Gautam M, Bracci PM. Comparison of interviewed and noninterviewed non-Hodgkin's lymphoma (NHL) patients in the San Francisco Bay Area. Ann Epidemiol 2002;12:419 25.
Cancer Epidemiol Biomarkers Prev 2005;14(10). October 2005