Endocrine-Related Cancer (2011) 18 265–276

Association of genetic polymorphisms in ESR2, HSD17B1, ABCB1, and SHBG with colorectal cancer risk

J Sainz1*, A Rudolph 2*, R Hein 2, M Hoffmeister 3, S Buch4,5, W von Scho¨nfels6, J Hampe4, C Schafmayer 5,6,HVo¨lzke7, B Frank 3, H Brenner 3,AFo¨rsti1,8, K Hemminki1,8 and J Chang-Claude 2

1Division of Molecular Genetic Epidemiology, German Cancer Research Center, Im Neuenheimer Feld 580, 69120 Heidelberg, Germany 2Division of Cancer Epidemiology, German Cancer Research Center, Im Neuenheimer Feld 581, 69120 Heidelberg, Germany 3Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Bergheimer Straße 20, 69115 Heidelberg, Germany 4Department of General Internal Medicine, University Hospital Schleswig-Holstein, Campus Kiel, House 6, Arnold-Heller-Straße 3, 24105 Kiel, Germany 5POPGEN biobank project, University Hospital Schleswig-Holstein, Campus Kiel, House 3, Arnold-Heller-Straße 3, 24105 Kiel, Germany 6Department of General and Thoracic Surgery, University Hospital Schleswig-Holstein, Campus Kiel, House 18, Arnold-Heller-Straße 3, 24105 Kiel, Germany 7Institute for Community Medicine, University Hospital of the Ernst Moritz Arndt University, Walter Rathenau Strasse 48, 17475 Greifswald, Germany 8Center for Primary Health Care Research, Clinical Research Center, SUS Malmo¨, House 28, 205 02 Malmo¨, Sweden (Correspondence should be addressed to J Chang-Claude; Email: [email protected]) *(J Sainz and A Rudolph contributed equally to this work)

Abstract The incidence rates and relative risks for colorectal cancer (CRC) are higher in men than in women. Sex steroids may play a role in this gender-associated difference in CRC risk. This study was conducted to explore the relationship of single nucleotide polymorphisms (SNPs) in steroid hormone signaling (ESR1, ESR2, PGR, NR1I2,andSHBG), phase I- and II-metabolizing enzyme (COMT, HSD17B1, CYP1A1, CYP17A1, CYP1A2, CYP1B1, CYP2C9, CYP3A4, CYP2C19,andGSTP1), and hormone transporter (ABCB1) genes with the risk of CRC in German women and men, separately. From the population-based DACHS study (South Germany), 47 putatively functional SNPs were genotyped in 1798 CRC cases (746 women and 1052 men) and 1810 controls (732 women and 1078 men). Significant allele dose–response associations were observed with ESR2_rs1255998, ESR2_rs928554, HSD17B1_rs605059, and ABCB1_rs2229109 in women (P trendZ0.004, 0.05, 0.03, and 0.05 respectively) and with ABCB1_rs1045642, ABCB1_rs9282564, and SHBG_rs6259 in men (P trendZ0.01, 0.03, and 0.02 respectively). The ESR2_rs1255998_G allele showed the most significant association with risk for CRC in women, with a per-allele odds ratio (OR) of 0.68 (95% confidence interval (CI) 0.52–0.88). This finding was replicated in an independent study from North Germany including 1076 female CRC cases and 1151 controls (ORZ0.84, 95% CI 0.71–1.04), yielding a per-allele OR of 0.80 (95% CI 0.69–0.93, P trendZ0.003) in the pooled sample. These findings implicate a role of ESR2 in the risk for developing CRC in women and suggest that HSD17B1, ABCB1,andSHBG genes may contribute to sex steroid-mediated effects on CRC development. Endocrine-Related Cancer (2011) 18 265–276

Introduction of all newly diagnosed carcinomas in the European Colorectal cancer (CRC) is the third most common population in 2008. Men are more likely to develop fatal solid cancer worldwide, and it comprised 13.4% CRC than women with a male–female ratio of w1.4

Endocrine-Related Cancer (2011) 18 265–276 Downloaded from Bioscientifica.comDOI: 10.1530/ERC-10-0264 at 09/23/2021 01:31:09PM 1351–0088/11/018–265 q 2011 Society for Endocrinology Printed in Great Britain Online version via http://www.endocrinology-journals.orgvia free access J Sainz et al.: metabolism-related genes and CRC

(Ferlay et al.2010). It has also been demonstrated that In an attempt to characterize sex-specific genetic women have a better survival than men (Hendifar et al. factors that may contribute to the known gender 2009). A possible explanation for this sex-associated differences in CRC risk, we conducted a comprehensive difference is that either endogenous or exogenous case–control study (Darmkrebs: Chancen der Verhu¨- estrogen influences normal colonic functionality and tung durch Screening (DACHS) study in Germany) and CRC progression (Koo & Leong 2010). The involve- examined 47 genetic variants within genes that are ment of estrogen in breast, ovarian, and endometrial linked to sex steroid metabolism. Polymorphisms in the cancers is well established (Amant et al. 2005, Reeves ATP-binding cassette, sub-family B (MDR/TAP), et al. 2006, Beral et al. 2007); however, there are many member 1 (ABCB1), COMT, hydroxysteroid 17b controversial hypotheses about how estrogen exerts its dehydrogenase 1 (HSD17B1), CYP1A1, CYP17A1, biological action on colonic epithelial cells. Initially, CYP1A2, CYP1B1, CYP2C9, CYP3A4, GSTP1, cyto- estrogen was suggested to promote proliferation of chrome P450, family 2, subfamily C, polypeptide 19 colon epithelial cells through binding to estrogen (CYP2C19), ESR1, ESR2, PGR, NR1I2,andsex (ER)-1, also called ER-a (Xu & Thomas hormone-binding globulin (SHBG) genes were selected 1994). The identification of ER2 (ER-b)asthe and association with CRC was evaluated separately predominant subtype present in both colon epithelial in women and men. To confirm the main findings, cell lines and human colon epithelium samples (Arai we carried out a replication study in an independent et al.2000, Xie et al. 2004) as well as its down- case–control study from Northern Germany. regulated expression on tumor cells compared with normal have pointed out that ER2 might have a selective and inhibitory effect on the proliferation of Materials and methods colon epithelial cells (Foley et al. 2000). Later on, Study participants and data collection epidemiological and clinical studies have reported that The DACHS study is a population-based case–control postmenopausal women undergoing hormone replace- study conducted in Southwest of Germany, which has ment therapy (HRT) have a decreased risk of previously been described in detail (Brenner et al. developing CRC (La Vecchia et al. 2005, Hoffmeister 2006, Hoffmeister et al. 2009, Frank et al. 2010). et al. 2007). In support of this anti-proliferative action Briefly, CRC cases were recruited from patients who of in colonic cells, Foley et al. (2000) also received in-patient treatment in a hospital of the reported that ESR2 expression level correlate with Rhein–Neckar–Odenwald region due to a first CRC patient survival. diagnosis of CRC. Controls were frequency-matched Intracellular actions of sex steroids are modulated by according to gender, 5-year age groups, county of receptors and local receptor concentrations at the tissue residence, and were then contacted by mail and and, at a cellular level, strictly controlled by sex steroid follow-up calls. Demographic information as well as transporters as well as phase I- and II-metabolizing information on colonoscopies, diet, anthropometry, enzymes. Increasing evidence suggests that the physical activity, medication (including statins, non- presence of single nucleotide polymorphisms (SNPs) steroidal anti-inflammatory drugs (NSAIDs), meno- within genes coding sex steroid hormone receptors and pausal HRT), reproductive factors, lifestyle factors, transporters and phase I- and II-metabolizing enzymes and family history was collected during a face-to-face might play a relevant role determining CRC risk interview by trained interviewers using a standardized (Bethke et al. 2007), although no genetic variants have questionnaire. To be eligible, participants had to be at clearly been demonstrated to explain the inter- least 30 years old and capable to complete the individual differences in mRNA, protein expression, interview. From January 01, 2003 until December 31, or enzyme activity. So far, epidemiologic studies 2007, 1798 cases and 1810 controls were recruited. investigating polymorphisms in hormone-related genes Written informed consent was obtained from all study and CRC have yielded mostly inconclusive results participants. Approximately, 41% of the participants (Slattery et al. 2005, Bethke et al. 2007, Kiss et al. were women (746 cases and 732 controls). Genomic 2007, Lin et al.2010). A significant gender difference DNA was extracted from peripheral blood mono- was reported in one study where an odds ratio (OR) of nuclear cells (PBMCs) or mouthwash using Flexigene 2.1 (95% confidence interval (CI) 1.2–3.6) for colon Kit 250 (Qiagen) and Qiagen Mini Kit (Qiagen) cancer was found in women carrying two alleles with respectively. For genotyping, CRC cases and controls at least 25 CA repeats in the ESR2 , but not in men were randomly distributed on 384-well plates (Slattery et al. 2005). (Applied Biosystems, Foster City, CA, USA).

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access 266 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 265–276

Participants with DNA samples that failed in more done using SNPlex chemistry (Applied Biosystems) than 25% of genotyping assays were excluded (28 on an automated platform in the Institute for cases and 20 controls), assuming poor DNA quality Clinical Molecular Biology at the University Hospital and therefore unreliable measurements. This study Schleswig-Holstein in Kiel. was approved by the Ethics Committees of the University of Heidelberg (Germany) and the State Medical Boards of Baden-Wurttemberg and Rhine- Statistical analysis land-Palatinate (Germany). Statistical analyses were conducted separately in women and men using SAS, version 9.2 (SAS Institute, SNP selection and genotyping Inc., Cary, NC, USA). Genotype frequencies were An extensive literature search concerning the mecha- assessed in cases and controls separately and tested for nism of action of estrogen and , deviation from Hardy–Weinberg equilibrium (HWE) hormone transporter, and hormone-metabolizing in controls at the aZ1% level. To evaluate the enzyme genes was performed to select candidate association between CRC and genotypes, multivariable genes that might affect the risk of CRC. SNPs were unconditional logistic regression models were used. assessed on the basis of NCBI data and were selected Basic regression models were adjusted for matching according to their known or putative functional variables (age and county of residence). Fully adjusted consequences, i.e. their modifying influence on the models additionally comprised the following estab- structure of proteins, transcription level, or alternative lished CRC risk factors: former colonoscopy, diagnosis splicing mechanisms. SNPs within the same gene of diabetes, body mass index (BMI), regular use of were also selected on the basis of linkage disequi- NSAIDs, and additionally use of HRT when assessing librium (LD) data. In total, 47 SNPs in 16 genes associations in women. Adjustment for additional risk were selected for this study (Table 1). Genotyping factors such as a family history of CRC, smoking, or of the SNPs was performed using KASPar assays red/processed meat consumption did not substantially (KBiosciences, Hoddesdon, Hertfordshire, UK) in change estimates. Estimates out of basic and fully a 384-well plate format. PCR products were analyzed adjusted models were similar, so that we report results with the ABI Prism 7900HT detection system that were adjusted for additional risk factors. Associ- using the SDS 2.4 software (Applied Biosystems). ations of genotypes and CRC risk are reported as ORs For internal quality control, R5% of samples with 95% CIs. The trend in risk for the additive model were randomly selected and included as duplicate. of inheritance was assessed by the means of Wald Concordance between the original and the duplicate statistics derived from the corresponding estimates of R samples for the 47 SNPs analyzed was 99.2%. the logistic regression model. To assess effect R Call rates for all SNPs were 94.5% in both case and modification by gender, multiplicative interaction control populations. terms were utilized in multivariate regression models. A log-likelihood ratio test was performed to compare Replication study population models with and without the interaction term. For the To determine whether the main findings in the DACHS explorative assessment of the association of genotypes population can be replicated, we used an independent with cancer sites, we used a multinomial logistic population-based case–control study from North regression model. The outcome was differentiated into Germany. This Northern German population com- colon (ICD 10: C18–C19) and rectum (ICD 10: C20) prised 2431 cases (1076 women and 1355 men) and cases. All tests were two sided and were considered 2200 controls (1151 women and 1049 men; Lascorz statistically significant with a P value!0.05. P values et al. 2010). Cases were identified through the regional were not corrected for multiple hypotheses testing. cancer registry of Schleswig-Holstein or through First of all, our tests were hypothesis driven as we records of surgical departments in Northern Germany investigated SNPs that were selected on the basis of and recruited by the POPGEN biobank. Controls prior knowledge regarding their putative biologic (nZ514) were identified through random selection functionality. Furthermore, due to the dependency out of population registries and also recruited by the between the investigated SNPs, Bonferroni correction POPGEN biobank. Additional controls (nZ1686) would be overly conservative. Therefore, we decided were recruited for Study of Health in Pomerania to replicate our findings in an independent study rather (SHIP) in Northeastern Germany and were cancer-free than directly correcting P values for multiple at recruitment (Volzke et al. 2010). Genotyping was hypotheses testing.

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access www.endocrinology-journals.org 267 J Sainz et al.: Estrogen metabolism-related genes and CRC

Table 1 Selected SNPs of hormone receptors and transporters as well as phase I- and II-metabolizing enzymes

Nucleotide Amino acid Gene_SNP dbSNP rs# substitution MAF (%) substitution

ABCB1_c.3435 rs1045642 T/C 49 I1145I ABCB1_c.1199 rs2229109 G/A 5 S400N ABCB1_IVS7C139 rs1202168 C/T 41 Intronic ABCB1_c.61 rs9282564 A/G 10 N21D ABCB1_c.-1 rs2214102 G/A 7 50-UTR SHBG_c.1066 rs6259 G/A 11 D356N ESR2_c.55833 rs1255998 C/G 12 30-UTR ESR2_c.55509 rs928554 A/G 41 30-UTR ESR2_c.49888 rs4986938 G/A 38 30-UTR ESR2_c.-12214 rs1271572 G/T 42 50-UTR ESR1_c.-105857 rs851984 C/T 41 50-UTR ESR1_c.-9929 rs2881766 T/G 18 50-UTR ESR1_c.-2224 rs2071454 T/G 10 50-UTR ESR1_c.30 rs2077647 A/G 48 S10S ESR1_IVS1-6610 rs827421 T/C 49 Intronic ESR1_IVS1-397 rs2234693 T/C 46 Intronic ESR1_IVS1-351 rs9340799 A/G 36 Intronic ESR1_c.292083 rs3798577 T/C 46 30-UTR ESR1_c.303855 rs910416 T/C 48 30 near gene PGR_c.1978 rs1042838 C/A 14 V660L PGR_c.1179 rs1379130 G/A 35 G393G PGR_c.-413 rs10895068 G/A 5 50-UTR PGR_c.-700 rs518162 G/A 8 50-UTR NR1I2_c.-24381 rs1523127 A/C 40 50-UTR NR1I2_IVS1C160 rs2276706 G/A 40 Intronic NR1I2_IVS2C55 rs1464603 T/C 33 Intronic NR1I2_IVS5-93 rs6785049 A/G 39 Intronic NR1I2_IVS6-17 rs2276707 C/T 18 Intronic NR1I2_c.10799 rs1054191 G/A 15 30-UTR NR1I2_c.11156 rs3814057 A/C 18 30-UTR COMT_c.472 rs4680 A/G 49 V158M HSD17B1_c.937 rs605059 T/C 46 G313S CYP1B1_c.1358 rs1800440 A/G 19 N453S CYP1B1_c.1294 rs1056836 C/G 41 L432V CYP1B1_c.355 rs1056827 G/T 30 A119S CYP1B1_c.142 rs10012 C/G 30 R48G CYP1A1_c.3801 rs4646903 T/C 10 30-near gene CYP1A2_IVS1-154 rs762551 A/C 30 Intronic CYP2C19_c.-806 rs12248560 C/T 23 50-UTR CYP2C19_c.681 rs4244285 G/A 15 P227P CYP2C9_c.430 rs1799853 C/T 12 R144C CYP2C9_c.1075 rs1057910 A/C 6 I359L CYP3A4_c.-392 rs2740574 A/G 4 50-UTR CYP3A4_c.-747 rs11773597 G/C 7 50-UTR CYP17A1_c.-34 rs743572 A/G 41 50-UTR GSTP1_c.313 rs1695 A/G 33 I105V GSTP1_c.341 rs1138272 C/T 9 A114V

SNP, single nucleotide polymorphism; MAF, minor allele frequency; UTR, untranslated region.

Results patients had a higher BMI, although they less often had This study is based on 746 female and 1052 male a former colonoscopy or took NSAIDs regularly. patients with CRC and 732 female and 1078 male Compared with female controls, male cases were unaffected controls respectively. Relevant epidemio- more likely to smoke than controls, but not female logical characteristics for women and men are shown cases. The average lifetime ethanol intake was lower in in Table 2. Overall, compared with controls, CRC female cases compared with controls but higher in

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access 268 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 265–276 male cases compared with controls. Female cases were non-significantly reduced risk for CRC was observed also more likely to be diagnosed with diabetes in women (ORZ0.79, 95% CI 0.61–1.03, per-allele), compared with controls, and this difference was not yielding a significant effect modification by gender as distinct in males. Although the cancer stages were (P interactionZ0.04). similarly distributed in women and men, the proportion To confirm our strongest findings for an effect of of rectal cancer was smaller in women. genetic polymorphisms in ESR2 on risk of CRC in The association of risk of CRC with the individual 47 women, we performed a replication study for SNPs in the sex steroid hormone signaling, transport, or ESR2_rs1255998 and ESR2_rs928554 in an indepen- metabolism genes is shown separately for women and dent study from North Germany including 2431 CRC menintheSupplementary Table 1, see section on cases (1076 female) and 2200 controls (1151 female). supplementary data given at the end of this article. In this study population, neither the ESR2_rs1255998 The minor allele frequencies of the 47 SNPs ranged from nor the ESR2_rs928554 was significantly associated 4 to 49%. All genotype frequencies in the control with CRC risk among women (Table 4). However, the population were in HWE, except for four SNPs, i.e. ORs for ESR2_rs1255998 in women were similar to CYP2C19_rs12248560 and ESR1_rs3798577 in men, those for women in the DACHS study. Women CYP1A1_rs4646903 in women, and CYP3A4_rs2740574 carrying the rs1255998_G allele had a non-significant in both men and women (P value!0.01). These SNPs decreased risk of CRC (ORZ0.86, 95% CI 0.71–1.04, were excluded from further analyses. per-allele). A pooled analysis confirmed the associ- Several genetic polymorphisms were found to be ation of the ESR2_rs1255998_G allele with a significantly associated with CRC risk in women decreased risk of CRC (ORZ0.80, 95% CI and/or in men (Table 3). The ESR2_rs1255998_G 0.69–0.93, per-allele; Table 4). As in the primary allele showed the strongest association with risk of analysis, the ESR2 variants were not associated with CRC in women, with a per-allele OR of 0.68 (95% CI CRC risk among men, neither in the North German 0.52–0.88). Another ESR2 variant, the rs928554_G study nor in the two study populations combined (data allele, was also associated with an increased risk not shown). The heterogeneity in effect by gender was (per-allele ORZ1.18, 95% CI 1.00–1.39). The respec- neither statistically significant in the North German tive associations were not significant in men (Table 3), study nor in the pooled study sample. and there was no significant effect modification by We also investigated whether there were tumor site- gender for those two ESR2 polymorphisms (P for specific associations between these polymorphisms interactionZ0.16 and 0.25 respectively). Similarly, and risk for CRC. Our results showed that there was no the HSD17B1_rs605059 C allele was associated heterogeneity in risk associations by tumor site (data with a significantly increased risk of CRC in women not shown). (ORZ1.20, 95% CI 1.02–1.41) but not in men, without showing significant effect heterogeneity (P for inter- actionZ0.06, Table 3). Discussion For all the three variants in ABCB1, the association with CRC risk was significant only in women or in This study is one of the largest studies assessing the men, although the risk association was statistically gender-specific association of genetic variants within the different by only gender for ABCB1_rs2229109 sex steroid hormone receptors, transporters, and hor- (P interactionZ0.04). Women carrying the rs2229109_A mone metabolism-related genes and CRC risk. In allele had a decreased risk of CRC (per-allele ORZ0.69, women, two variants within the ESR2 gene (rs1255998 95% CI 0.48–0.99). This variant was not associated and rs928554) as well as variants within ABCB1 with CRC in men (ORZ1.02, 95% CI 0.75–1.38, (rs2229109) and HSD17B1 (rs605059) genes showed per-allele). Men bearing an ABCB1_rs1045642_C evidence of an association with CRC risk in the DACHS allele had a decreased risk for developing CRC study population. In men, two variants within ABCB1 (ORZ0.85, 95% CI 0.74–0.97, per-allele) whereas (rs1045642 and rs9282564) and a variant within SHBG those harboring the_rs9282564_G allele had a higher (rs6259) were significantly associated with CRC risk. risk of CRC (per-allele ORZ1.26, 95% CI 1.02–1.57). The effect on CRC risk was statistically different between These two variants were not found to be associated men and women only for the ABCB1_rs2229109 and with risk in women (Table 3). SHBG_rs6259 variants. For most of the hormone-related For SHBG_rs6259_A allele, a statistically signi- variants, we did not observe significant effect hetero- ficant increase in CRC risk was observed in men geneity by gender. Despite a rather large sample size, we (ORZ1.29, 95% CI 1.04–1.60, per-allele), whereas a still had limited power to detect heterogeneity of genetic

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access www.endocrinology-journals.org 269 J Sainz et al.: Estrogen metabolism-related genes and CRC

Table 2 Distribution of selected epidemiological variables in women and men in the DACHS study population

Women Men

Value Cases N (%) Controls N (%) P value Value Cases N (%) Controls N (%) P value

Age (years) 30 to !40 7 (0.9) 2 (0.3) 0.13 30 to !40 10 (1.0) 3 (0.3) !0.01 40 to !50 30 (4.0) 20 (2.7) 40 to !50 28 (2.7) 40 (3.7) 50 to !60 103 (13.8) 93 (12.7) 50 to !60 167 (15.9) 167 (15.5) 60 to !70 229 (30.7) 260 (35.5) 60 to !70 396 (37.6) 315 (29.2) 70 to !80 232 (31.1) 231 (31.6) 70 to !80 337 (32.0) 384 (35.6) 80C 145 (19.4) 126 (17.2) 80C 114 (10.8) 169 (15.7) BMI (kg/m2)R5 years before diagnosis/date of interview !18.5 16 (2.1) 13 (1.8) !0.01 !18.5 1 (0.1) 2 (0.2) 0.02 18.5 to !25 305 (40.9) 350 (47.8) 18.5 to !25 263 (25.0) 332 (30.8) 25 to !30 248 (33.2) 263 (35.9) 25 to !30 582 (55.3) 574 (53.2) 30 to !35 121 (16.2) 79 (10.8) 30 to !35 163 (15.5) 144 (13.4) 35 to !40 30 (4.0) 16 (2.2) 35 to !40 30 (2.9) 17 (1.6) O40 6 (0.8) 7 (1.0) O40 5 (0.5) 3 (0.3) Missing 20 (2.7) 4 (0.5) Missing 8 (0.8) 6 (0.6) Average lifetime ethanol intake per day (g/day) None 256 (34.3) 209 (28.6) 0.02 None 52 (4.9) 61 (5.7) !0.01 0! to !1.9 145 (19.4) 122 (16.7) 0! to !5.4 187 (17.8) 248 (23.0) 1.9 to !3.9 132 (17.7) 142 (19.4) 5.4 to !11.3 247 (23.5) 258 (23.9) 3.9 to !7.3 96 (12.9) 128 (17.5) 11.3 to !20.2 246 (23.4) 262 (24.3) R7.3 117 (15.7) 131 (17.9) R20.2 320 (30.4) 249 (23.1) Average physical activity in METs in last 12 months (h/week) !84.9 215 (28.8) 179 (24.5) 0.12 !56.6 225 (21.4) 259 (24.0) 0.10 84.9 to !122.5 164 (22.0) 177 (24.2) 56.6 to !101.4 229 (21.8) 263 (24.4) 122.5 to !183.0 162 (21.7) 181 (24.7) 101.4 to !168.9 255 (24.2) 263 (24.4) R183.0 164 (22.0) 185 (25.3) R168.9 301 (28.6) 266 (24.7) Missing 41 (5.5) 10 (1.4) Missing 42 (4.0) 27 (2.5) Average lifetime pack years of regular smoking Non-smoker 517 (69.3) 526 (71.9) 0.41 Non-smoker 344 (32.7) 417 (38.7) !0.01 O0to!10 93 (12.5) 90 (12.3) O0to!10 193 (18.3) 221 (20.5) 10 to !20 55 (7.4) 46 (6.3) 10 to !20 163 (15.5) 156 (14.5) 20 to !30 37 (5.0) 39 (5.3) 20 to !30 150 (14.3) 121 (11.2) R30 43 (5.8) 28 (3.8) R30 189 (18.0) 154 (14.3) Missing 1 (0.1) 3 (0.4) Missing 13 (1.2) 9 (0.8) First degree family history of CRC No 608 (81.5) 612 (83.6) 0.19 No 880 (83.7) 942 (87.4) 0.01 Yes 116 (15.5) 96 (13.1) Yes 143 (13.6) 106 (9.8) Missing 21 (2.9) 24 (3.3) Missing 29 (2.8) 30 (2.8) Diagnosis of diabetes ever (through a physician) No 598 (80.2) 646 (88.3) !0.01 No 858 (81.6) 912 (84.6) 0.07 Yes 138 (18.5) 85 (11.6) Yes 188 (17.9) 162 (15.0) Missing 10 (1.4) 1 (0.1) Missing 6 (0.6) 4 (0.4) Ever colorectal endoscopy No/unknown 599 (80.3) 351 (48.0) !0.01 No/unknown 825 (78.4) 498 (46.2) !0.01 Yes 146 (19.6) 381 (52.0) Yes 227 (21.6) 580 (53.8) Missing 1 (0.1) 0 (0.0) Missing 0 (0.0) 0 (0.0) Ever use of hormone replacement therapy No 496 (66.5) 357 (48.8) !0.01 No NA NA NA Yes 245 (32.8) 373 (51.0) Yes NA NA Missing 5 (0.7) 2 (0.3) Missing NA NA Ever regular use of NSAIDs 2C times/weekR1 year No 573 (76.8) 518 (70.8) 0.01 No 812 (77.2) 733 (68.0) !0.01 Yes 171 (22.9) 210 (28.7) Yes 234 (22.2) 342 (31.7) Missing 2 (0.3) 4 (0.5) Missing 6 (0.6) 3 (0.3)

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access 270 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 265–276

Table 2 continued

Women Men

Value Cases N (%) Controls N (%) P value Value Cases N (%) Controls N (%) P value

UICC stage Stage I 164 (22.0) NA NA Stage I 271 (25.8) NA NA Stage II 235 (31.5) NA Stage II 306 (29.1) NA Stage III 247 (33.1) NA Stage III 320 (30.4) NA Stage IV 99 (13.3) NA Stage IV 152 (14.4) NA Missing 1 (0.1) NA Missing 3 (0.3) NA CRC localization Colon 500 (67.0) NA NA Colon 597 (56.7) NA NA Rectum 246 (33.0) NA Rectum 455 (43.3) NA

CRC, colorectal cancer; BMI, body mass index; MET, metabolic equivalent task; NSAID, non-steroidal anti-inflammatory drug; UICC, The International Union Against Cancer. effects by gender for polymorphisms with low minor early stages of CRC, whereas high levels of ESR2 allele frequency. have been associated with CRC patient survival An interesting and robust finding in this study was the (Caiazza et al. 2007). Our findings in women showed association of the ESR2_rs1255998 variant with a that the ESR2_rs1255998_G allele was significantly decreased CRC risk in women, suggesting that estrogen associated with a decreased CRC risk, whereas the through ER2-related mechanisms may play an import- ESR2_rs928554_G allele increased risk of CRC. ant role in the pathogenesis of CRC. In our investigation In addition, Slattery et al.(2005)reported an association of an independent replication study from Northern of a CA repeat within the ESR2 gene with an increased Germany, the association of ESR2_rs1255998 with risk of CRC. Because the CA repeat is located in a CRC did not appear to be significant. However, the different gene region, it may independently from observed lower P value for the inverse association ESR2_rs1255998 and ESR2_rs928554 lead to alterations with ESR2_rs1255998 in the pooled sample compared in the ESR2 expression and terminal biological effects with that in the initial analysis gives some assurance of estrogen. Thus, on the basis of the available evidence that the observed association is not solely driven by the and our own findings, it is conceivable to suggest a initial study population. direct relationship between ESR2 variants and CRC. Most of the observed findings are novel and ABCB1 encodes a multidrug resistance support data regarding the role of these genes as CRC P-glycoprotein, an efflux transporter that actively risk factors (Fig. 1). ER1 and ER2 present a high extrudes a wide variety of substances from cells homology in the DNA-binding domain and similar (Thiebaut et al. 1987), including estrogen (Siest et al. patterns of gene activation through estrogen-response 2005). Pharmacogenetics studies have revealed that element (ERE) binding. However, ER2 is highly ABCB1 variants influence the level of expression of the expressed in both colon epithelial cell lines and protein, its function, or specificity (Aird et al. 2008) human colon epithelium samples (Arai et al. 2000, and that they correlate with steroid hormone levels Xie et al. 2004, Newcomb et al. 2007), whereas ER1 (Nakamura et al. 2009). In this study, we found that is lacking or showing a lower level of expression. women carrying the allele G of ABCB1_rs2229109 had In addition, ER2 contrary to ER1 seems to inhibit a decreased risk of CRC. Likewise, men harboring the the non-classical AP-1-mediated transcription after minor allele for the rs1045642 SNP had a significantly estrogen binding (Webb et al. 1999, Liu et al. 2002). decreased risk of developing CRC, whereas men It has been suggested that ER2 is responsible of carrying the minor allele of the rs9282564 had an triggering pro-apoptotic and anti-proliferative increased risk of CRC. Based on these results, the mechanisms through a wide variety of genomic and ABCB1-mediated estrogen efflux from the colonic non-genomic pathways (Li et al. 2001, Safe 2001, epithelial cells might be influenced by these variants Bjornstrom & Sjoberg 2005, Kennelly et al. 2008). and, thus, modify the risk for CRC. However, as the A strong down-regulation of ER2 expression in human ABCB1 transports several substances, the associations colon adenocarcinoma has been reported, suggesting we observed with gene variants are not necessarily that ER2 might be a tumor suppressor. Interestingly, a attributable solely to an altered estrogen transport, significant reduction of ESR2 has been reported during especially in men.

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access www.endocrinology-journals.org 271 272 CRC and genes metabolism-related Estrogen al.: et Sainz J Table 3 Associations of polymorphisms involved in hormone transport, binding, and metabolism among women and men in the DACHS study

Women Men

Gene_rs number Genotype Controls (%) Cases (%) OR (95% CI)a P trend Controls (%) Cases (%) OR (95% CI)a P trend P interaction*

ABCB1_rs2229109 Missing 0 (0.0) 4 (0.5) 4 (0.4) 1 (0.1) G/G 629 (87.1) 658 (89.8) 1.00 967 (90.5) 929 (89.6) 1.00 G/A 92 (12.7) 70 (9.5) 0.68 (0.47–0.98) 96 (9.0) 104 (10.0) 0.97 (0.71–1.33) A/A 1 (0.1) 1 (0.1) 1.25 (0.05–33.2) 1 (0.1) 3 (0.3) 3.80 (0.36–40.1) Per A allele 0.69 (0.48–0.99) 0.05 1.02 (0.75–1.38) 0.89 0.04 ABCB1_rs1045642 Missing 3 (0.4) 1 (0.1) 3 (0.3) 4 (0.4) T/T 196 (27.1) 202 (27.6) 1.00 285 (26.7) 289 (27.9) 1.00 T/C 346 (47.9) 362 (49.4) 1.03 (0.78–1.36) 513 (48.0) 546 (52.7) 1.06 (0.85–1.32) C/C 177 (24.5) 168 (22.9) 0.95 (0.69–1.31) 267 (25.0) 198 (19.1) 0.70 (0.53–0.91) Per C allele 0.98 (0.83–1.15) 0.78 0.85 (0.74–0.97) 0.01 0.27 ABCB1_rs9282564 Missing 6 (0.8) 7 (1.0) 7 (0.7) 2 (0.2) A/A 582 (80.6) 586 (79.9) 1.00 875 (81.9) 816 (78.7) 1.00 A/G 122 (16.9) 131 (17.9) 1.08 (0.80–1.45) 177 (16.6) 204 (19.7) 1.26 (0.99–1.60) G/G 12 (1.7) 9 (1.2) 0.68 (0.25–1.86) 9 (0.8) 15 (1.4) 1.65 (0.68–4.00) Per G allele 1.00 (0.77–1.31) 0.97 1.26 (1.02–1.57) 0.03 0.17 ESR2_rs1255998 Missing 4 (0.6) 9 (1.2) 25 (2.3) 18 (1.7) C/C 555 (76.9) 598 (81.6) 1.00 807 (75.6) 810 (78.1) 1.00 C/G 147 (20.4) 120 (16.4) 0.72 (0.53–0.97) 223 (20.9) 196 (18.9) 0.93 (0.74–1.17) G/G 16 (2.2) 6 (0.8) 0.34 (0.12–0.93) 13 (1.2) 13 (1.3) 1.14 (0.49–2.67) Per G allele 0.68 (0.52–0.88) 0.004 0.96 (0.78–1.18) 0.68 0.16 ESR2_rs928554 Missing 4 (0.6) 7 (1.0) 21 (2.0) 12 (1.2) A/A 263 (36.4) 240 (32.7) 1.00 363 (34.0) 336 (32.4) 1.00 A/G 337 (46.7) 343 (46.8) 1.11 (0.86–1.43) 506 (47.4) 524 (50.5) 1.11 (0.90–1.37) G/G 118 (16.3) 143 (19.5) 1.43 (1.02–2.00) 178 (16.7) 165 (15.9) 0.93 (0.70–1.23) Per G allele 1.18 (1.00–1.39) 0.05 0.99 (0.86–1.13) 0.87 0.25 HSD17B1_rs605059 Missing 6 (0.8) 4 (0.5) 5 (0.5) 10 (1.0)

Downloaded fromBioscientifica.com at09/23/202101:31:09PM T/T 217 (30.1) 182 (24.8) 1.00 298 (27.9) 297 (28.6) 1.00 T/C 347 (48.1) 368 (50.2) 1.31 (1.00–1.73) 537 (50.3) 504 (48.6) 0.98 (0.78–1.21) C/C 152 (21.1) 179 (24.4) 1.43 (1.03–1.98) 228 (21.3) 226 (21.8) 1.01 (0.77–1.31) Per C allele 1.20 (1.02–1.41) 0.03 1.00 (0.88–1.14) 0.97 0.06 www.endocrinology-journals.org SHBG_rs6259 Missing 5 (0.7) 6 (0.8) 13 (1.2) 6 (0.6) G/G 559 (77.4) 593 (80.9) 1.00 847 (79.3) 801 (77.2) 1.00 G/A 150 (20.8) 124 (16.9) 0.75 (0.56–1.01) 201 (18.8) 220 (21.2) 1.30 (1.03–1.64) A/A 8 (1.1) 10 (1.4) 0.87 (0.31–2.44) 7 (0.7) 10 (1.0) 1.59 (0.58–4.37) Per A allele 0.79 (0.61–1.03) 0.08 1.29 (1.04–1.60) 0.02 0.04

*P value for testing of effect modification by gender utilizing an interaction term of gender and genetic polymorphism assuming a log-additive mode of inheritance (1df). OR, odds ratio; CI, confidence interval. aModels adjusted for age, county of residence, former colonoscopy (yes/no), ever diagnosis of diabetes (yes/no), body mass index (in categories of !23 kg/m2,23to!25 kg/m2,25to !27 kg/m2,27to!30 kg/m2,andO30 kg/m2), regular use of NASIDs (yes/no), and ever use of hormone replacement therapy (yes/no, for women only). via freeaccess Endocrine-Related Cancer (2011) 18 265–276

HSD17B1, an enzyme catalyzing the conversion

trend from into (E2), is highly expressed in P the colonic epithelium suggesting a central role of E2 in the epithelial cell biology, differentiation, and prolifer-

c ation (Mustonen et al. 1998, Saloniemi et al. 2010). Accordingly, an increased expression of HSD17B1 might enhance estrogen action at the colonic tissue by increasing the E2/estrone ratio. Likewise, HSD17B1 participates in the conversion of DHEA to (Tsuchiya et al. 2005), which may also influence on carcinogenesis (Aka et al. 2010). We observed that trend OR (95% CI)

P women carrying the HSD17B1_rs605059_C allele had an increased risk of developing CRC, although no

b association was seen in men. Although there is no evidence indicating that this non-synonymous poly- morphism affects the biological properties of the enzyme, we cannot rule out this possibility. Alter- natively, this polymorphism may be in LD with unidentified functional variants. The SHBG gene encodes the SHBG, which binds

trend OR (95% CI) and transports and estrogens in the blood P and regulates their biological availability (Kahn et al.

a 2002). We found a significantly differential effect of the SHBG_rs6259 SNP for women and men. The rs6259 is a non-synonymous SNP in exon 8 that results in an amino acid substitution of aspartic acid for asparagine (D356N). The A (N) allele of SHBG rs6259 has been associated with elevated circulating levels of SHBG in post-menopausal women (Cousin et al. 2004, Dunning et al.2004) and a reduction of SHBG clearance in an animal model (Cousin et al. 1999). However, a study among postmenopausal women did not find an association of SHBG serum levels with CRC risk (Clendenen et al.2009). In men, the

gene in women association of the rs6259 variant with serum SHBG levels has been less conclusive. Overall, there is much

ESR2 heterogeneity in studies investigating the association of SHBG variants and SHBG levels with common diseases (Xita & Tsatsoulis 2010). Multiple factors seem to influence SHBG variation and biological effects of SHBG that may also lead to the gender- related differences that we observed, but we cannot exclude the possibility that our findings with SHBG_rs6259 might be due to chance. Our study has both strengths and weaknesses. Strengths include its candidate gene study design, using a well-defined and homogeneous study popu-

DACHS population KIEL population DACHS population KIEL populationlation Total population with a sufficient size and another population-

Controls (%) Cases (%) Controls (%) Cases (%) OR (95% CI) based study in the same ethnic group for replication. We had a power of 80% to detect an OR of 1.25 in Replication of rs1255998 and rs928554 in the women assuming a log-additive mode of inheritance and a minor allele frequency of 30%. With the same HWE 0.56 0.31 0.27 0.32 Per G allele 1.14 (0.98–1.32) 0.09 0.92 (0.81–1.03) 0.15 1.01 (0.92–1.10) 0.88 HWE 0.09 0.99 0.29 0.91 Per G allele 0.72 (0.57–0.91) 0.007 0.86 (0.71–1.04) 0.13 0.80 (0.69–0.93) 0.003 Models adjusted for ageModels and adjusted county for of age. residence. Models adjusted for age and population. MissingA/AA/GG/G 4 (0.6)p 263 (36.4) 337 (46.7) 7 240 118 (1.0) (32.7) (16.3) 343 (46.8) 332 143 (28.8) (19.5) 13 582 (1.1) (50.6) 348 224 (32.4) (19.5) 500 (46.5) A/A 24 (2.2) 203 (18.9) A/G G/G 1.00 1.11 (0.88–1.40) 1.31 (0.97–1.78) 0.82 (0.68–0.99) 0.86 (0.68–1.10) 1.00 0.93 (0.80–1.08) 1.04 (0.86–1.25) 1.00 MissingC/CC/GG/G 4 (0.6)p 555 (76.9) 147 (20.4) 9 598 (1.2) (81.6) 16 (2.2) 120 (16.4) 876 (76.1) 21 242 (1.8) (21.0) 852 6 (79.3) (0.8) 197 (18.3) C/C 15 (1.4) C/G 12 (1.0) 1.00 11 (1.0) 0.76 (0.58–1.00) G/G 0.37 (0.14–0.96) 0.84 (0.68–1.03) 1.00 0.95 (0.42–2.16) 0.81 (0.68–0.95) 0.60 (0.32–1.09) 1.00 Table 4 a b c OR, odds ratio; CI, confidence interval. Only basic models used as additional variables were not available for the Kiel population. ESR2_rs928554 ESR2_rs1255998 assumptions, we were able to detect an OR of 1.21 in

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access www.endocrinology-journals.org 273 J Sainz et al.: Estrogen metabolism-related genes and CRC

Androstendione

Testosterone DHEA

HSD3B HSD17B1 16-OH-E2 Estrogen CYP19A1 17-OH-E2 degradation E2 E2 Estrone-3-sulphate E2 CYP2C19 E2 SHBG HSD17B1 E2-ER2 binding

ABCB1 ABCB1 ABCB1 ER2 Proliferative membrane-mediated Apoptosis effects of estradiol E E 2 Estrone-3-sulphate 2 Induction of P53 signaling Caspase Cascade

ER1 ER2 BAX ER1 ER2 Apaf-1

E2-ER2 binding

E2-ER1 binding

Non-Classical pathway Classical pathway

Interaction with transcription factors ER1 ER2 AP-1 Cyclin D1, etc... Transcription Transcription Transcription ERE ERE

Induction of proliferative effects Inhibition of the non-classical AP-1-transcription

Figure 1 Hypothetical scheme for the role of estrogen-related genes in CRC development based on current knowledge and discussed in detail in the discussion. men. On the other hand, the major limitation of the Declaration of interest study may be that only a few SNPs were studied in The authors declare that there is no conflict of interest that every gene. Therefore, other SNPs in or nearby the could be perceived as prejudicing the impartiality of the selected genes might be associated with risk for CRC. research reported. In conclusion, our findings provide evidence that genetic variants in the ESR2 play a role in the risk of Funding developing CRC in women and suggest that poly- morphisms in the HSD17B1, ABCB1, and SHBG genes The DACHS study was supported by grants from the German Research Council (Deutsche Forschungsgemeinschaft, grant may modulate sex steroid-mediated effects on CRC numbers BR 1704/6-1, BR 1704/6-3, BR 1704/6-4, and CH development. However, the generally weak ORs are 117/1-1) and the German Federal Ministry of Education and associated with a relatively low predictive value of Research (grant numbers 01KH0404 and 01ER0814). SHIP these polymorphisms on CRC risk and, therefore, these is conducted by the Research Network of Community results should be prudently interpreted. Likewise, our Medicine at the University of Greifswald, which is funded findings need to be confirmed by additional indepen- by the German Federal Ministry of Education and Research dent studies. and the Federal State of Mecklenburg-Vorpommern. This study was also supported by NGFNC (Nationales Genomforschungsnetz, grant number 01GS08181).

Supplementary data Acknowledgements This is linked to the online version of the paper at http://dx. We thank all participants in the DACHS case–control study doi.org/10.1530/ERC-10-0264. as well as those participated in the replication study located in

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access 274 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 265–276

Schleswig-Holstein and Mecklenburg–West Pomerania. We Cousin P, Dechaud H, Grenot C, Lejeune H, Hammond GL & acknowledge support by interviewers, the recruiting hospi- Pugeat M 1999 Influence of glycosylation on the tals, and physicians of the study regions as well as their clearance of recombinant human sex hormone-binding collaborating pathologists. We wish to thank Ute Handte- globulin from rabbit blood. Journal of Steroid Daub and Belinda-Su Kaspereit for their excellent technical Biochemistry and Molecular Biology 70 115–121. assistance. (doi:10.1016/S0960-0760(99)00101-6) Cousin P, Calemard-Michel L, Lejeune H, Raverot G, Yessaad N, Emptoz-Bonneton A, Morel Y & Pugeat M 2004 Influence of SHBG gene pentanucleotide TAAAA References repeat and D327N polymorphism on serum sex hormone- Aird RE, Thomson M, Macpherson JS, Thurston DE, Jodrell DI binding globulin concentration in hirsute women. & Guichard SM 2008 ABCB1 genetic polymorphism Journal of Clinical Endocrinology and Metabolism 89 influences the pharmacology of the new pyrrolobenzodia- 917–924. (doi:10.1210/jc.2002-021553) zepine derivative SJG-136. Pharmacogenomics Journal 8 Dunning AM, Dowsett M, Healey CS, Tee L, Luben RN, Folkerd 289–296. (doi:10.1038/sj.tpj.6500465) E, Novik KL, Kelemen L, Ogata S, Pharoah PD et al. 2004 Aka JA, Mazumdar M, Chen CQ, Poirier D & Lin SX 2010 Polymorphisms associated with circulating sex hormone 17b-Hydroxysteroid dehydrogenase type 1 stimulates levels in postmenopausal women. Journal of the National breast cancer by dihydrotestosterone inactivation Cancer Institute 96 936–945. (doi:10.1093/jnci/djh167) in addition to estradiol production. Molecular Ferlay J, Shin HR, Bray F, Forman D, Mathers C & Parkin Endocrinology 24 832–845. (doi:10.1210/me.2009-0468) DM 2010 Estimates of worldwide burden of cancer in Amant F, Moerman P, Neven P, Timmerman D, 2008: GLOBOCAN 2008. International Journal of Van Limbergen E & Vergote I 2005 Endometrial cancer. Cancer 127 2893–2917. (doi:10.1002/ijc.25516) Lancet 366 491–505. (doi:10.1016/S0140-6736(05) Foley EF, Jazaeri AA, Shupnik MA, Jazaeri O & Rice LW 67063-8) 2000 Selective loss of beta in malignant Arai N, Strom A, Rafter JJ & Gustafsson JA 2000 Estrogen human colon. Cancer Research 60 245–248. receptor beta mRNA in colon cancer cells: growth effects Frank B, Hoffmeister M, Klopp N, Illig T, Chang-Claude J & of estrogen and genistein. Biochemical and Biophysical Brenner H 2010 Colorectal cancer and polymorphisms in Research Communications 270 425–431. (doi:10.1006/ DNA repair genes WRN, RMI1 and BLM. Carcinogenesis bbrc.2000.2444) 31 442–445. (doi:10.1093/carcin/bgp293) Bethke L, Webb E, Sellick G, Rudd M, Penegar S, Withey L, Hendifar A, Yang D, Lenz F, Lurje G, Pohl A, Lenz C, Ning Y, Qureshi M & Houlston R 2007 Polymorphisms in the Zhang W & Lenz HJ 2009 Gender disparities in metastatic cytochrome P450 genes CYP1A2, CYP1B1, CYP3A4, colorectal cancer survival. Clinical Cancer Research 15 CYP3A5, CYP11A1, CYP17A1, CYP19A1 and color- 6391–6397. (doi:10.1158/1078-0432.CCR-09-0877) ectal cancer risk. BMC Cancer 7 123. (doi:10.1186/1471- Hoffmeister M, Raum E, Winter J, Chang-Claude J & 2407-7-123) Brenner H 2007 Hormone replacement therapy, body Beral V, Bull D, Green J & Reeves G 2007 Ovarian cancer mass, and the risk of colorectal cancer among post- and hormone replacement therapy in the Million Women menopausal women from Germany. British Journal of Study. Lancet 369 1703–1710. (doi:10.1016/S0140- Cancer 97 1486–1492. (doi:10.1038/sj.bjc.6604066) 6736(07)60534-0) Hoffmeister M, Raum E, Krtschil A, Chang-Claude J & Bjornstrom L & Sjoberg M 2005 Mechanisms of estrogen Brenner H 2009 No evidence for variation in colorectal receptor signaling: convergence of genomic and cancer risk associated with different types of postmeno- nongenomic actions on target genes. Molecular pausal hormone therapy. Clinical Pharmacology and Endocrinology 19 833–842. (doi:10.1210/me.2004-0486) Therapeutics 86 416–424. (doi:10.1038/clpt.2009.134) Brenner H, Chang-Claude J, Seiler CM, Sturmer T & Kahn SM, Hryb DJ, Nakhla AM, Romas NA & Rosner W Hoffmeister M 2006 Does a negative screening colono- 2002 Sex hormone-binding globulin is synthesized in scopy ever need to be repeated? Gut 55 1145–1150. target cells. Journal of Endocrinology 175 113–120. (doi:10.1136/gut.2005.087130) (doi:10.1677/joe.0.1750113) Caiazza F, Galluzzo P, Lorenzetti S & Marino M 2007 Kennelly R, Kavanagh DO, Hogan AM & Winter DC 2008 17b-Estradiol induces ERbeta up-regulation via Oestrogen and the colon: potential mechanisms for cancer p38/MAPK activation in colon cancer cells. Biochemical prevention. Lancet Oncology 9 385–391. (doi:10.1016/ and Biophysical Research Communications 359 102–107. S1470-2045(08)70100-1) (doi:10.1016/j.bbrc.2007.05.059) Kiss I, Orsos Z, Gombos K, Bogner B, Csejtei A, Tibold A, Clendenen TV, Koenig KL, Shore RE, Levitz M, Arslan AA Varga Z, Pazsit E, Magda I, Zolyomi A et al. 2007 & Zeleniuch-Jacquotte A 2009 Postmenopausal levels of Association between allelic polymorphisms of meta- endogenous sex hormones and risk of colorectal cancer. bolizing enzymes (CYP 1A1, CYP 1A2, CYP 2E1, mEH) Cancer Epidemiology, Biomarkers & Prevention 18 and occurrence of colorectal cancer in Hungary. 275–281. (doi:10.1158/1055-9965.EPI-08-0777) Anticancer Research 27 2931–2937.

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access www.endocrinology-journals.org 275 J Sainz et al.: Estrogen metabolism-related genes and CRC

Koo JH & Leong RW 2010 Sex differences in epidemiolo- Saloniemi T, Jarvensivu P, Koskimies P, Jokela H, gical, clinical and pathological characteristics of color- Lamminen T, Ghaem-Maghami S, Dina R, Damdimo- ectal cancer. Journal of Gastroenterology and Hepatology poulou P, Makela S, Perheentupa A et al. 2010 Novel 25 33–42. (doi:10.1111/j.1440-1746.2009.05992.x) hydroxysteroid (17beta) dehydrogenase 1 inhibitors Lascorz J, Forsti A, Chen B, Buch S, Steinke V, Rahner N, reverse estrogen-induced endometrial hyperplasia in Holinski-Feder E, Morak M, Schackert HK, Gorgens H transgenic mice. American Journal of Pathology 176 et al. 2010 Genome-wide association study for colorectal 1443–1451. (doi:10.2353/ajpath.2010.090325) cancer identifies risk polymorphisms in German familial Siest G, Marteau JB, Maumus S, Berrahmoune H, Jeannesson cases and implicates MAPK signalling pathways in E, Samara A, Batt AM & Visvikis-Siest S 2005 disease susceptibility. Carcinogenesis 31 1612–1619. Pharmacogenomics and cardiovascular drugs: need for (doi:10.1093/carcin/bgq146) integrated biological system with phenotypes and La Vecchia C, Gallus S & Fernandez E 2005 Hormone proteomic markers. European Journal of Pharmacology replacement therapy and colorectal cancer: an update. 527 1–22. (doi:10.1016/j.ejphar.2005.10.011) Journal of the British Menopause Society 11 166–172. Slattery ML, Sweeney C, Murtaugh M, Ma KN, Wolff RK, (doi:10.1258/136218005775544264) Potter JD, Caan BJ & Samowitz W 2005 Associations Li C, Briggs MR, Ahlborn TE, Kraemer FB & Liu J 2001 between ERalpha, ERbeta, and AR genotypes and colon and Requirement of Sp1 and interaction rectal cancer. Cancer Epidemiology, Biomarkers & Preven- in 17beta-estradiol-mediated transcriptional activation of the tion 14 2936–2942. (doi:10.1158/1055-9965.EPI-05-0514) low density lipoprotein receptor . Endo- Thiebaut F, Tsuruo T, Hamada H, Gottesman MM, Pastan I crinology 142 1546–1553. (doi:10.1210/en.142.4.1546) & Willingham MC 1987 Cellular localization of the Lin J, Zee RY, Liu KY, Zhang SM, Lee IM, Manson JE, multidrug-resistance gene product P-glycoprotein in Giovannucci E, Buring JE & Cook NR 2010 Genetic normal human tissues. PNAS 84 7735–7738. (doi:10. variation in sex-steroid receptors and synthesizing enzymes 1073/pnas.84.21.7735) and colorectal cancer risk in women. Cancer Causes & Tsuchiya Y, Nakajima M & Yokoi T 2005 Cytochrome Control 21 897–908. (doi:10.1007/s10552-010-9518-5) P450-mediated metabolism of estrogens and its regulation Liu MM, Albanese C, Anderson CM, Hilty K, Webb P, in human. Cancer Letters 227 115–124. (doi:10.1016/ Uht RM, Price RH Jr, Pestell RG & Kushner PJ 2002 j.canlet.2004.10.007) Opposing action of estrogen receptors alpha and beta on Volzke H, Alte D, Schmidt CO, Radke D, Lorbeer R, cyclin D1 gene expression. Journal of Biological Friedrich N, Aumann N, Lau K, Piontek M, Born G et al. Chemistry 277 24353–24360. (doi:10.1074/jbc. 2010 Cohort Profile: The Study of Health in Pomerania. M201829200) International Journal of Epidemiology [in press]. Mustonen MV, Poutanen MH, Kellokumpu S, de Launoit Y, (doi:10.1093/ije/dyp394) Isomaa VV, Vihko RK & Vihko PT 1998 Mouse 17 beta- Webb P, Nguyen P, Valentine C, Lopez GN, Kwok GR, hydroxysteroid dehydrogenase type 2 mRNA is predomi- McInerney E, Katzenellenbogen BS, Enmark E, nantly expressed in hepatocytes and in surface epithelial cells Gustafsson JA, Nilsson S et al. 1999 The estrogen of the gastrointestinal and urinary tracts. Journal of Molecular receptor enhances AP-1 activity by two distinct Endocrinology 20 67–74. (doi:10.1677/jme.0.0200067) mechanisms with different requirements for receptor Nakamura T, Okamura N, Yagi M, Omatsu H, Yamamori M, Molecular Endocrinology 13 Kuwahara A, Nishiguchi K, Horinouchi M, Okumura K & transactivation functions. Sakaeda T 2009 Effects of ABCB1 3435COT genotype 1672–1685. (doi:10.1210/me.13.10.1672) on serum levels of cortisol and aldosterone in women with Xie LQ, Yu JP & Luo HS 2004 Expression of estrogen normal menstrual cycles. Genetics and Molecular receptor beta in human colorectal cancer. World Research 8 397–403. (doi:10.4238/vol8-2gmr574) Journal of Gastroenterology 10 214–217. Newcomb PA, Zheng Y, Chia VM, Morimoto LM, Doria- Xita N & Tsatsoulis A 2010 Genetic variants of sex hormone- Rose VP, Templeton A, Thibodeau SN & Potter JD 2007 binding globulin and their biological consequences. Estrogen plus progestin use, microsatellite instability, and Molecular and Cellular Endocrinology 316 60–65. the risk of colorectal cancer in women. Cancer Research (doi:10.1016/j.mce.2009.08.025) 67 7534–7539. (doi:10.1158/0008-5472.CAN-06-4275) Xu X & Thomas ML 1994 Estrogen receptor-mediated Reeves GK, Beral V, Green J, Gathani T & Bull D 2006 direct stimulation of colon cancer cell growth in vitro. Hormonal therapy for menopause and breast-cancer risk Molecular and Cellular Endocrinology 105 197–201. by histological type: a cohort study and meta-analysis. (doi:10.1016/0303-7207(94)90170-8) Lancet Oncology 7 910–918. (doi:10.1016/S1470- 2045(06)70911-1) Safe S 2001 Transcriptional activation of genes by 17 beta- Received in final form 4 January 2011 estradiol through estrogen receptor-Sp1 interactions. Accepted 10 February 2011 Vitamins & Hormones 62 231–252. (doi:10.1016/S0083- Made available online as an Accepted Preprint 6729(01)62006-5) 11 February 2011

Downloaded from Bioscientifica.com at 09/23/2021 01:31:09PM via free access 276 www.endocrinology-journals.org