This art icle was downloaded by: [ Huazhong Universit y of Science & Technology ] On: 22 March 2014, At : 05: 27 Publisher: Rout ledge I nform a Lt d Regist ered in England and Wales Regist ered Num ber: 1072954 Regist ered office: Mort im er House, 37- 41 Mort im er St reet , London W1T 3JH, UK Nutrition and Cancer Publicat ion det ails, including inst ruct ions f or aut hors and subscript ion inf ormat ion: ht t p: / / www. t andf online. com/ loi/ hnuc20 Sex-Specific Differences in Colon Cancer Associated With p53 Mutations Mart ha L. Slat t ery , Rachel Ballard-Barbash , John D. Pot t er , Khe-Ni Ma , Bet t e J. Caan , Krist in Anderson & Wade Samowit z Published online: 18 Nov 2009. To cite this article: Mart ha L. Slat t ery , Rachel Ballard-Barbash , John D. Pot t er , Khe-Ni Ma , Bet t e J. Caan , Krist in Anderson & Wade Samowit z (2004) Sex-Specif ic Dif f erences in Colon Cancer Associat ed Wit h p53 Mut at ions, Nut rit ion and Cancer, 49: 1, 41-48, DOI: 10. 1207/ s15327914nc4901_6 To link to this article: ht t p: / / dx. doi. org/ 10. 1207/ s15327914nc4901_6 PLEASE SCROLL DOWN FOR ARTI CLE Taylor & Francis m akes every effort t o ensure t he accuracy of all t he inform at ion ( t he “ Cont ent ” ) cont ained in t he publicat ions on our plat form . However, Taylor & Francis, our agent s, and our licensors m ake no represent at ions or warrant ies what soever as t o t he accuracy, com plet eness, or suit abilit y for any purpose of t he Cont ent . Any opinions and views expressed in t his publicat ion are t he opinions and views of t he aut hors, and are not t he views of or endorsed by Taylor & Francis. The accuracy of t he Cont ent should not be relied upon and should be independent ly verified wit h prim ary sources of inform at ion. Taylor and Francis shall not be liable for any losses, act ions, claim s, proceedings, dem ands, cost s, expenses, dam ages, and ot her liabilit ies what soever or howsoever caused arising direct ly or indirect ly in connect ion wit h, in relat ion t o or arising out of t he use of t he Cont ent . This art icle m ay be used for research, t eaching, and privat e st udy purposes. Any subst ant ial or syst em at ic reproduct ion, redist ribut ion, reselling, loan, sub- licensing, syst em at ic supply, or dist ribut ion in any form t o anyone is expressly forbidden. Term s & Condit ions of access and use can be found at ht t p: / / www.t andfonline.com / page/ t erm s- and- condit ions NUTRITION AND CANCER, 49(1), 41–48 Copyright © 2004, Lawrence Erlbaum Associates, Inc. Sex-Specific Differences in Colon Cancer Associated With p53 Mutations Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 Martha L. Slattery, Rachel Ballard-Barbash, John D. Potter, Khe-Ni Ma, Bette J. Caan, Kristin Anderson, and Wade Samowitz Abstract: Introduction: Sex-specific differences in observed incidence rates, tumor subsite, and diet and lifestyle associations with colon cancer have been observed. We evaluate sex-specific associations with p53 mutations in colon cancer to add to understanding of these differences. Data from a large population-based incident case-control study of colon cancer were used to evaluate age and gender associations with p53 mutations. To obtain a better understanding of gender-specific associations, we evaluated the role of estrogen as a mediator of risk. For these analyses, women were classified as estrogen positive or negative, based on menopausal status and use of hormone replacement therapy (HRT). Results: There was a significant interaction between age and sex and risk of an acquired p53 mutation compared with p53 Wt. Among men, there was an increase in p53 mutations with age, whereas among women the opposite was observed. Associations with parity, oral contraceptive use, and total ovulatory months were not associated with p53 mutations. However, recent use of HRT reduced risk of all tumors, as did being estrogen positive. Women who were estrogen positive (either premenopausal or recent users of HRT) were at a significantly increased risk of an acquired p53 mutation if they consumed a diet with a high sugar index (odds ratio = 2.94; 95% confidence interval = 1.47–5.89); similar increases in risk of p53 mutations were not observed for men or women who were estrogen negative. Conclusions: Although sex-specific associations were detected for acquired p53 mutations, they do not indicate a unique role of estrogens in the mutation of p53. These data are consistent with a role for estrogen in altering susceptibility to diet and lifestyle factors possibly via an insulin-related mechanism. Sex-specific associations with colon cancer have been repeatedly observed (1–6). Men have higher incidence rates of colon cancer; women have more proximal tumors than men; and women have an increased risk of polyp recurrence on a high-carbohydrate diet, whereas men do not (1–6). Observa- tions that hormone replacement therapy (HRT) reduces risk of colorectal cancer provide a biological basis for some of the observed sex-specific differences in colorectal cancer reported (7). Sex-specific associations have been attributed to many factors, including the effect of estrogen on tumor development (8) and differences in gut metabolism observed for men and women (9). Ability to examine tumors is helping to unravel our understanding of colon cancer by allowing us to observe unique diet and lifestyle associations with specific tumor characteristics. Our recent work and that by others suggest that estrogens may protect against tumors with microsatellite instability (MSI) (10,11). This observation helps to explain both patterns of MSI in tumors and possibly tumor subsite differences between men and women (11). Associations between p53 immunohistochemistry-positive tumors and oral contraceptive use also have been reported (12). Dietary glycemic index (GI) has been reported as being more strongly associated with p53 mutations in women than in men (13). Taken together, this may suggest that estrogen-related factors are associated with acquired p53 mutations. Explanations that account for both steroid (that is, estrogen) and growth hormones (that is, insulin) may better integrate the role of established risk factors. Body size is related to a variety of metabolic phenomena that may influence cellular proliferation. A large body mass index (BMI) is associated with both insulin resistance and alterations in estrogen metabolism and function (14,15). Likewise, physical activity appears to play an important role in the regulation of both estrogen and insulin (16,17). In this article, we evaluate sex-specific associations with acquired p53 mutations in colon tumors by examining both estrogen- and insulin-related variables. We hypothesize that estrogen is associated with p53 mutations that may involve regulation of insulin-related mechanisms. Physical activity, BMI, and dietary sugar may be further modified by estrogen as they relate to p53 tumor mutations. M. L. Slattery and K.-N. Ma are affiliated with the Health Research Center, Department of Family and Preventive Medicine, University of Utah, Salt Lake City, UT 84108. R. Ballard-Barbash is affiliated with the National Cancer Institute, Bethesda, MD 20892-7324. J. D. Potter is affiliated with the Fred Hutchinson Cancer Research Center, Seattle, WA 98109-1024. B. J. Caan, Kaiser Permanente Medical Care Program, Oakland, CA 94612. K. Anderson is affiliated with the University of Minnesota, Minneapolis, MN 55454-1015. W. Samowitz is affiliated with the Department of Surgical Pathology, University of Utah, Salt Lake City, UT 84108. Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 Methods Study participants were African American, white, or Hispanic and were from the Kaiser Permanente Medical Care Program (KPMCP) of northern California, an eight county area in Utah (Davis, Salt Lake, Utah, Weber, Wasatch, Tooele, Morgan, and Summit counties), or the Twin Cities metropolitan area in Minnesota. Eligibility criteria for cases included diagnosis with first-primary incident colon cancer (ICD-O, 2nd ed., codes 18.0 and 18.2–18.9) between October 1, 1991, and September 30, 1994; between 30 and 79 yr of age at time of diagnosis; and mentally competent to complete the interview. Cases with adenocarcinoma or carcinoma of the rectosigmoid junction or rectum (defined as the first 15 cm from the anal opening) with known familial adenomatous polyposis, ulcerative colitis, or Crohn’s disease were not eligible. Of all cases identified, 75.6% of those we were able to contact participated in the study. Controls, in addition to the eligibility criteria for cases, did not have a history of a colorectal tumor. Controls were selected from members’ lists for KPMCP, from driver’s license lists in Minnesota, and from driver’s license lists, random-digit dialing, or Health Care Finance Administration lists for Utah. These methods have been described in detail (18). The response rate for controls was 63.7%. The study population was approximately 90% non-Hispanic white, 5% African American, and 5% Hispanic. flect intake and metabolic response to carbohydrates and sugar. The SI was created by equally weighing individual GI index, sucrose-to-fiber ratio, and foods high in sugar; this summary variable was more robust than individual items and used in the analysis. Data on demographics, physical activity, and body size, including usual adult height and weight 2 and 5 yr prior to diagnosis, were obtained. A measure of long-term (past 20 yr) levels of vigorous leisure-time physical activity was used because this was shown to be a sensitive predictor of cancer risk in this population (22). BMI (weight in kg/height in m2) was used as an indicator of body size. A detailed reproductive history was obtained from women. The interview included questions on menopausal status and the use of exogenous hormones such as estrogen, progestin, or other steroid hormones for both contraceptive and noncontraceptive purposes. Dates of first and last use and duration of use of HRT were ascertained. Women were considered estrogen positive if they were premenopausal or were postmenopausal and were recent (within 2 yr prior to the referent period) users of HRT or oral contraceptives. Women who had a hysterectomy but did not have both ovaries removed were considered estrogen positive if less than age 55. All other women were considered estrogen negative. Ovulatory months were calculated based on age at first menses until age at menopause minus months pregnant and months on oral contraceptives. Tissue Ascertainment Questionnaire Data Data were collected by trained and certified interviewers using laptop computers (19). The referent period for the study was the calendar year, approximately 2 yr prior to date of diagnosis or date of interview for controls. Using this referent period, information was collected on dietary intake using a detailed diet history questionnaire (20). Information from the diet history was used to estimate the sugar content of the diet. Dietary GI was estimated so that carbohydrates could be evaluated taking into account their metabolic effect. We estimated GI from values determined in controlled studies for blood glucose level after consumption of 50-g carbohydrate portions (carrots, parsnips, and peanuts were tested on a 25-g portion) relative to the standard of 50 g carbohydrate from white bread or glucose. An individual’s dietary GI was estimated for a given food by determining the grams of carbohydrate consumed for that food, dividing that amount by 50 g (the carbohydrate food portion on which the GI was based except in those instances where a 25-g portion was used to determine the GI), and multiplying that value by the GI for the food as reported in the literature (21). Additionally, we evaluated sugar content of the diet by evaluating servings of food high in sugar as well as the sucrose-to-fiber ratio. Because all of the three indicators of dietary sugar and response to carbohydrate had comparable associations, rather than adjust for each variable in assessment of the other, we created a variable that took into account all of these factors, labeled a sugar index (SI), that we believed would more accurately re42 Methods for obtaining tumor tissue have been described (23). Colon cancer tissue was microdissected and DNA was extracted from formalin-fixed paraffin-embedded tissue blocks as described (23). In Utah and KPMCP, tissue and DNA were available from 97% of eligible cases in Utah and 85% of cases in KPMCP. In Minnesota, tumor DNA was available for those people who could be reconsented, representing approximately 35% of all cases identified. p53 Exons 5–8 of p53 and relevant intron/exon boundaries were polymerase chain reaction (PCR) amplified with the following primers (two sets of primers, a 5′ and 3′ set, were used to amplify the relatively large exon 5; “F” is the forward primer, “R” the reverse): 5–5′F:ttatctgttcacttgtgccc, 5–5′R:tcatgtgctgtgactgcttg, 5–3′F:ttccacacccccgcccggca, 5–3′R:accctgggcaaccagccctg, 6F:acgacagggctggttgccca, 6R:ctcccagagaccccagttgc, 7F:ggcctcatcttgggcctgtg, 7R:cagtgtgcagggtggcaagt, 8F:gtttctgcctcttgcttctctttt, and 8R:tctcctccaccgcttcttgt. Primers were dye labeled, and a single-strand conformational polymorphism analysis (SSCP) of the respective PCR products was performed by electrophoresis in a nondenaturing gel with evaluation on an ABI 373 (24). All tumor samples corresponding to any abnormally migrating SSCP bands were reamplified with the respective primers tailed with UP and RP (and without dye labeling) for sequencing. PCR products were sequenced using prism Big Dye terminators and cycle seNutrition and Cancer 2004 Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 quencing with Taq FS DNA polymerase. DNA sequence was collected and analyzed on an ABI prism 377 automated DNA sequencer. Any alterations were verified by sequencing in both directions. SSCP was also performed on the respective germline DNA sample of any tumor with a non–hot-spot missense mutation, insertion or deletion mutations that consisted of multiples of three bases, and splice-site mutations to verify that it was an acquired rather than inherited mutation. For hot-spot mutations in which the same mutation was found in 10 or more tumors, SSCP was performed on a random sample of 10 of the respective normal samples. If abnormally migrating SSCP bands were seen in any germline DNA sample, these samples were also sequenced to determine whether or not the germline DNA harbored the same genetic alteration as the respective tumor. Sequencing also was performed on the corresponding germline DNA for tumors in which multiple genetic alterations within a particular exon were identified by the initial sequencing. Statistical Methods Sex-specific associations with acquired p53 mutations were evaluated by examining age, tumor site, and estrogen-related factors. A total of 2,410 controls and 1,458 cases with p53 mutational data were available for analysis. Estrogen status, parity, ovulatory months, and use of exogenous hormones were used in the analyses as indicators of steroid-hormone status and its impact on p53 mutations in women. Recent HRT use was defined as having used HRT within the past 2 yr including current HRT users. Dietary sugar content or SI was used as an indicator of possible insulin status along with BMI [wt (kg)/ht (m)2] and long-term vigorous leisure physical activity level (PAL). Cut points for SI were based on distribution of the variable in the control population, giving the GI a rank from 1 to 5, the sucrose-to-fiber ratio a rank from 1 to 5, and servings of high-sugar foods a rank from 1 to 5. The summary SI ranged from 3 to 15, with 3 being individuals with very low dietary sugar as well as a low GI and 15 being individuals with a high level of all values. Low SI were people with scores of <6; high SI was ≥12. Estimates of association were derived from multiple logistic regression models where age was adjusted. Adjustment for other factors, such as energy intake, smoking cigarettes, alcohol intake, or other risk factors, did not appreciably alter results. In logistic regression models, cases with and without p53 mutations were compared with controls; cases with p53 mutations also were compared with cases without p53 mutations to provide insight into unique associations with acquired p53 mutations. When evaluating the combined associations between estrogen status and BMI, PAL, and dietary SI, the lowest risk group was used as the referent. Statistical testing of the overall improved fit of the model with an interaction term was calculated by taking –2 times the difference in the log-likelihood of those models with and without the cross-product term; additionally, multiplicative interaction was assessed using the cross-product of Vol. 49, No. 1 the two terms being analyzed in the logistic model. Site-specific analyses were done for proximal tumors defined as those in the cecum through transverse colon and distal tumors were those located in the splenic flexure, descending, and sigmoid colon. Results Acquired p53 mutations were detected in 48.4% of colon tumors in men and in 45.4% of colon tumors in women. A significant difference in age was observed for men and women when comparing those with a p53 mutation with those whose tumor did not have a p53 mutation (Table 1). Among men, the likelihood of a p53 mutation increased with age, whereas, among women, the likelihood of this type of mutation decreased with age. Assessment of the effects of gender and age on p53 mutations by tumor site showed that the interaction between gender and age was stronger for distal tumors than for proximal tumors (P interaction = 0.03). Further stratification by MSI status was not possible because few people who were MSI positive also had a p53 mutation; however, associations persisted after adjustment for MSI status. Parity and oral contraceptive use were not associated with p53 mutations (Table 2). Women who reported ever using HRT were at reduced risk of both p53-mutated and p53-Wt tumors. Associations were slightly stronger when use of HRT occurred within the past 2 yr (that is, recent use) and for women with a p53-mutated tumor. Additionally, there was a significant linear trend for p53-mutated tumors when comparing never, former, and recent HRT users. Relative to controls, women with few ovulatory months were at lower risk of having a cancer without a p53 mutation (odds ratio, OR = 0.56; 95% confidence interval, CI = 0.37–0.85). Simultaneous adjustment for parity and oral contraceptive use did not alter observed associations. Evaluation of overall estrogen status showed that being estrogen negative increased risk of both p53-mutated and p53-Wt tumors. Adjustment for MSI status did not alter the results. The associations among SI, BMI, and PAL were evaluated for men and women and for women based on estrogen status (Table 3). High SI was associated with increased risk of a p53 mutation in women (OR = 1.55; 95% CI = 1.05–2.29) but not in men (OR = 1.23; 95% CI = 0.88–1.73); neither the goodness of fit of the model (P = 0.30) nor the P for interaction was statistically significant (P = 0.63) nor were there tumor site-specific associations. Women who were estrogen positive were at statistically significantly increased risk of p53-mutated tumors associated when consuming a high-SI diet (OR = 2.94; 95% CI = 1.47–5.89), whereas women who were estrogen negative had a threefold increased risk from being estrogen negative (OR = 3.08; 95% CI = 1.54–6.14) and a slight increased risk from consuming a high-SI diet (P value for interaction between estrogen and SI = 0.04); on an additive scale, one would have expected an OR of 4.78 (2.94 + 2.84 – 1.0) rather than the observed 3.08 for being both estrogen negative and having a high-SI diet. Associations were 43 Table 1. Sex-Specific Description of Population With and Without p53 Tumor Mutations Men Everyone Proximalb Distal Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 Women Everyone Proximal Distal Age Controls (C) p53 M (N) p53 Wt (N) OR 95% CI p53 a M vs. p53 Wt OR 95% CI p53 M vs.C OR 95% CI p53 Wt vs.C ≤50 51–64 65–79 ≤50 51–64 65–79 ≤50 51–64 65–79 131 414 745 131 414 745 131 414 745 27 133 224 12 51 103 14 77 117 37 134 238 28 66 139 9 66 93 1.00 1.36 (0.78–2.36) 1.29 (0.76–2.19) 1.00 1.80 (0.84–3.89) 1.73 (0.84–3.56) 1.00 0.75 (0.31–1.84) 0.81 (0.34–1.95) 1.00 1.56 (0.99–2.46) 1.46 (0.94–2.27) 1.00 1.35 (0.70–2.60) 1.51 (0.81–2.82) 1.00 1.74 (0.95–3.18) 1.47 (0.82–2.64) 1.00 1.15 (0.76–1.73) 1.13 (0.76–1.68) 1.00 0.75 (0.46–1.21) 0.87 (0.56–1.36) 1.00 2.32 (1.13–4.78) 1.82 (0.89–3.69) ≤50 51–64 65–79 ≤50 51–64 65–79 ≤50 51–64 65–79 98 346 676 98 346 676 98 346 676 31 111 160 10 36 80 19 73 76 24 126 213 15 64 141 8 62 66 1.00 0.68 (0.37–1.23) 0.58 (0.33–1.03) 1.00 0.84 (0.34–3.07) 0.85 (0.37–1.98) 1.00 0.50 (0.20–1.21) 0.49 (0.20–1.18) 1.00 1.01 (0.64–1.61) 0.75 (0.48–1.16) 1.00 1.02 (0.49–2.13) 1.16 (0.58–2.13) 1.00 1.09 (0.63–1.89) 0.58 (0.34–1.00) 1.00 1.49 (0.91–2.43) 1.29 (0.80–2.06) 1.00 1.21 (0.66–2.21) 1.36 (0.77–2.42) 1.00 2.20 (1.02–4.74) 1.20 (0.56–2.57) a: Interaction between gender and age comparing p53 M (mutation) with Wt (wild type) is 0.05 for everyone, 0.22 for proximal tumors, and 0.37 for distal tumors with a p53 mutation the P for interaction between age and gender is 0.03. b: Numbers differ slightly for proximal and distal tumors because of unknown site for some individuals. stronger for distal rather than proximal tumors, with estrogen-negative women consuming a high-SI diet being at over a fivefold increased risk of distal tumors (OR = 5.39; 95% CI = 1.78–16.36) compared with women who were estrogen positive and consuming a low-SI diet. For p53-Wt tumors, a high-SI diet did not increase risk in either men or women; associations did not differ by tumor site. Among estrogen-negative women with p53-Wt tumors there was an approximate twofold increase in risk from being estrogen negative (OR = 1.98; 95% CI = 1.17–3.36), although the level of estimated risk decreased with increasing SI. Thus, the P for interaction was statistically significant at 0.01, but the direction of associated risk was different for the p53-mutated and -Wt tumors. BMI was associated with colon cancer overall in men, both for p53-mutated and p53-Wt tumors (Table 3). Associations were stronger for distal tumors. Men with a BMI of ≥30 were over twice as likely (OR = 2.11; 95% CI = 1.41–3.18) to have a p53 mutation than men with a BMI of <25 when compared with controls. The comparable risk for proximal tumors was 1.46 (95% CI = 0.94–2.25). Women who were estrogen positive had an increased risk of both p53-mutated and -Wt tumors associated with a larger BMI, with the P for interaction between estrogen status and BMI being 0.01 for both types of tumors. As with men, associations comparing p53 mutation with controls were stronger for distal tumors (OR = 3.16; 95% CI = 1.65–6.08) among estrogen-positive women with a BMI of ≥30 compared with women with a BMI of <25; the comparable association with proximal tumors was 1.42 (95% CI = 0.63–3.20). There was no difference in p53 tumor mutational status by level of physical ac44 tivity in either men or women; reporting no vigorous physical activity was associated with increased colon cancer risk in people with and without p53 tumors. Evaluation of PAL and SI suggests that among women a high dietary SI greatly increases the likelihood of having a p53 mutation rather than having a p53 Wt tumor (Table 4), whereas the same combination of factors among men decreased the likelihood of a p53-mutated tumor (P for goodness of fit for interaction term of PAL, SI, and gender was <0.01). For instance, women who had low PAL and high SI were 3.54 times more likely to have a p53-mutated tumor relative to those with high PAL and low SI (95% CI = 1.29–9.74); men with a low PAL and high SI were not more likely to have a p53-mutated tumor (OR = 1.16; 95% CI = 0.48–2.80). Assessment of these combined factors by estrogen status showed slightly stronger associations for estrogen-positive women, although sample sizes for some cells were too small to precisely determine meaningful differences in risk by estrogen status. Further assessment of these associations by tumor site gave only imprecise estimates of the three-way interaction. Discussion Molecular characteristics of tumors allow us to evaluate unique pathways of colon carcinogenesis, some of which appear to have sex-specific risks. In a previous study, we observed that diets high in sugar were associated with an increased risk of p53 mutations in women but not men (13). In Nutrition and Cancer 2004 Table 2. Age-Adjusted Associations Between Reproductive-Related Factors and p53 Tumor Alterations in Womena Parity 1–3 ≥4 P Linear Trend 139/178/514 1.00 (0.62–1.62) 0.87 (0.58–1.33) 138/148/513 1.00 (0.62–1.62) 0.73 (0.48–1.11) 0.997 0.07 0 N (M, Wt, C) OR 95% CI (M vs. C) OR 95% CI (Wt. vs. C) 25/37/93 1.00 1.00 Oral Contraceptive Use Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 N (M, Wt, C) OR 95% CI (M vs. C) OR 95% CI (Wt. vs. C) Never Ever P Linear Trend 220/272/817 1.00 1.00 75/80/280 0.82 (0.58–1.18) 0.91 (0.65–1.29) NA NA HRT Use (postmenopausal women) N (M, Wt, C) OR 95% CI (M vs. C) OR 95% CI (Wt. vs. C) b Never Ever Recent (<2 yr) Former 147/176/480 1.00 1.00 99/126/429 0.71 (0.53–0.95) 0.78 (0.60–1.02)c 58/76/285 0.58 (0.41–0.83) 0.69 (0.50–0.95) 41/50/144 0.97 (0.65–1.43) 0.96 (0.66–1.38) P Linear Trend 0.004 0.07 Total Ovulatory Months N (M, Wt, C) OR 95% CI (M vs. C) OR 95% CI (Wt. vs. C) ≥431 370–430 294–369 <294 P Linear Trend 77/107/302 1.00 1.00 85/104/309 1.05 (0.74–1.49) 0.90 (0.66–1.24) 67/90/257 0.98 (0.68–1.43) 0.97 (0.70–1.35) 63/39/196 1.14 (0.77–1.69) 0.56 (0.37–0.85) 0.55 0.04 Estrogen Status N (M, Wt, C) OR 95% CI (M vs. C) OR 95% CI (Wt. vs. C) Estrogen+ Estrogen– P Linear Trend 103/118/443 1.00 1.00 189/237/647 1.64 (1.18–2.27) 1.45 (1.06–1.96) NA NA a: Abbreviations are as follows: M, mutation; Wt, wild type; C, control. b: Linear trend compares never, former, and recent users of HRT; analyses are restricted to postmenopausal women; not applicable (NA) for dichotomous variables. c: P value is 0.07. the current article, we build on that observation and examine the interaction among age, gender, and risk of having a colon cancer that contains a p53 mutation. Older men were slightly more likely to have p53-mutated tumors than women. As in our previous study of MSI (11), there are suggestions in the data that estrogen-related factors might be relevant to p53 mutations. In contrast to the situation with MSI, however, we did not detect p53-specific associations with parity, HRT, oral contraceptive use, or total ovulatory months. Given the independent nature of colon tumors in regard to MSI and p53 status, it was impossible to stratify data by MSI status because almost all MSI-positive tumors did not have p53 mutations. We did, however, find suggestions of increased risk of p53 mutations among estrogen-positive women who consumed a diet with a high SI. These observations were not seen in men. Evaluation of reproductive and estrogen-related factors resulted in several important observations. First, some factors such as parity and ovulatory months were more associated with p53-Wt tumors than p53-mutated tumors. We believe that this most likely represents an association with MSI Vol. 49, No. 1 because MSI is a component of Wt tumors and we have demonstrated in the past that MSI may be associated with these reproductive factors and other reproductive indicators; recent use of HRT appears to be more important than use in the past (11). However, HRT use and estrogen status were associated with both p53-mutated and p53-Wt tumors. We believe that this is indicative of multiple roles of estrogen and sex steroids in the development of colon cancer. The sex-specific association of SI with p53 mutations and the interaction of estrogen status with variables such as SI and physical activity suggest that estrogens may have multiple roles in the etiology of colon cancer. We hypothesize that estrogen may be acting through insulin-related factors, given the association with dietary sugar, to increase the risk of acquired p53 mutations. Laboratory studies show that estrogen and insulin co-regulate the function of the other (25–30), lending support to our hypothesis. The co-regulator action has been seen in studies where estrogen-containing oral contraceptives resulted in increased plasma insulin levels and insulin resistance among users (27); both androgens and estrogens have been shown to induce insulin resistance (28); 45 Table 3. Age-Adjusted Associations Between SI, BMI, and PAL and p53-M and p53-Wt Colon Cancer in Men and Women and by Estrogen Status Among Womena SI (OR; 95% CI) Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 p53 M vs. C Men Women Estrogen+ Estrogen– p53 Wt vs. C Men Women Estrogen+ Estrogen– c b GOF P Value Case/Control (N) Low Intermediate High P Interaction 384/1290 302/1120 108/458 189/648 1.00 1.00 1.00 2.84 (1.42–5.69) 0.99 (0.73–1.33) 1.38 (0.98–1.93) 1.91 (1.01–3.63) 3.15 (1.66–5.96) 1.23 (0.88–1.73) 1.55 (1.05–2.29) 2.94 (1.47–5.89) 3.08 (1.54–6.14) 0.63 0.30 0.04 0.06 409/1290 363/1120 121/458 238/648 1.00 1.00 1.00 1.98 (1.17–3.36) 1.30 (0.96–1.75) 0.79 (0.60–1.06) 0.91 (0.55–1.52) 1.42 (0.87–2.33) 1.09 (0.77–1.56) 0.94 (0.68–1.33) 1.53 (0.88–2.70) 1.33 (0.76–2.32) 0.36 0.04 0.01 0.06 2 BMI (kg/m ) p53 M vs C Men Women Estrogen+ Estrogen– p53 Wt vs. C Men Women Estrogen+ Estrogen– c b <25 25–<30 ≥30 P Interaction 383/1290 300/1114 108/455 107/645 1.00 1.00 1.00 1.93 (1.23–3.02) 1.15 (0.88–1.52) 1.14 (0.82–1.60) 1.13 (0.67–1.89) 1.90 (1.19–3.04) 1.80 (1.31–2.47) 1.26 (0.93–1.70) 2.24 (1.32–3.98) 2.01 (1.24–3.25) 0.02 0.52 0.01 0.07 408/1280 359/1114 121/455 234/645 1.00 1.00 1.00 1.73 (1.14–2.62) 1.50 (1.15–1.96) 1.26 (0.95–1.65) 1.18 (0.73–1.91) 2.03 (1.33–3.10) 1.59 (1.15–2.21) 1.26 (0.92–1.71) 2.17 (1.30–3.62) 1.52 (0.95–2.41) 0.44 0.52 0.01 0.01 PAL (vigorous activity over past 20 years) p53 M vs C Men Women Estrogen+ Estrogen– p53 Wt vs. C Men Women Estrogen+ Estrogen– GOF P Value Case/Control (N) c b GOF P Value Case/Control (N) ~3.5 + hrs/wk Intermediate None P Interaction 384/1,290 302/1,120 108/458 189/648 1.00 1.00 1.00 2.07 (1.19–3.60) 1.22 (0.94–1.57) 1.09 (0.79–1.51) 1.49 (0.90–2.46) 1.80 (1.06–3.03) 1.44 (1.04–1.98) 1.57 (1.12–2.21) 1.55 (0.87–2.75) 3.10 (1.81–5.30) 0.98 0.74 0.81 0.17 409/1,290 363/1,120 121/458 238/648 1.00 1.00 1.00 1.80 (1.08–2.99) 1.54 (1.20–1.98) 1.03 (0.76–1.39) 1.38 (0.86–2.23) 1.51 (0.93–2.46) 1.57 (1.14–2.16) 1.45 (1.06–1.99) 1.41 (0.82–2.44) 2.51 (1.53–4.11) 0.24 0.09 0.83 0.17 a: Abbreviations are as follows: M, mutation; Wt, wild type; C, control. b: Multiplicative P value for interaction compares interaction between exposure and gender and between estrogen-positive and estrogen-negative women. c: GOF (goodness of fit) is the improved fit of the overall model by inclusion of the interactive term progestins have been shown to regulate insulin and IGF-receptor levels in human breast cancer cell lines (29); and studies have suggested that PR status is regulated by both estrogen and IGF-1 (30). Recent laboratory studies have indicated that obesity is directly related to insulin levels and that physical activity is inversely associated with insulin levels; diets that increase glycemic load have been directly associated with insulin levels (31–33). The complexity of the association among estrogen, obesity, and insulin has been previously discussed; in estrogen-negative women, obesity does not increase colon cancer risk, possibly because adipose tissue stores estrogen (34). Differences in p53 mutations observed for men and women for SI may be metabolic in nature as sex differences in carbo- 46 hydrate metabolism have been observed (35). It has been suggested that these differences are mediated through estrogen as well as being influenced by physical activity (36,37). Studies have shown that administration of 17-β-estradiol to men or amenorrheic women results in better maintenance of plasma glucose during endurance exercise (38), that glycogen storage is altered by menstrual cycle phase (39), and that glycogen metabolism is influenced by 17-β-estradiol (40). Women also have been shown to respond differently to physical activity, oxidizing less carbohydrate during endurance exercise than men (41). It also is possible that the sex-specific associations observed between dietary sugar and p53 mutations result from differences in dietary source of sugar. However, the only significant difference in source of sugar for men and women was Nutrition and Cancer 2004 Table 4. Age-Adjusted Associations Between SI, BMI, and PAL and Risk Associated With Having a p53-Mutated Tumor Versus Having a p53-Wt Tumor in Men and Womena SI (OR; 95% CI) Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 p53 M vs. Wt Women PAL High Intermediate Low BMI <25 25–29 ≥30 Men PAL High Intermediate Low BMI <25 25–29 ≥30 N (M/Wt) GOF P Value b Low Intermediate High E × SI E × SI × S 77/93 117/138 108/132 1.00 2.33 (0.86–6.30) 2.25 (0.83–6.08) 3.96 (1.56–10.08) 3.10 (1.24–7.73) 3.27 (1.31–8.16) 2.44 (0.86–6.92) 3.20 (1.23–8.36) 3.54 (1.29–9.74) 0.31 <0.01 131/150 89/125 80/84 1.00 1.23 (0.57–2.62) 0.60 (0.23–1.52) 1.59 (0.84–3.00) 1.47 (0.74–2.90) 2.06 (1.04–4.10) 1.83 (0.88–3.79) 1.00 (0.46–2.14) 2.10 (0.89–4.96) 0.14 <0.01 143/136 160/191 81/82 1.00 0.57 (0.27–1.18) 0.85 (0.34–2.16) 0.58 (0.30–1.13) 0.55 (0.29–1.05) 0.50 (0.25–1.02) 0.97 (0.45–2.08) 0.63 (0.30–1.30) 1.16 (0.48–2.80) 0.49 101/97 175/219 107/92 1.00 0.48 (0.21–1.12) 0.71 (0.28–1.80) 0.54 (0.25–1.21) 0.42 (0.20–0.89) 0.51 (0.23–1.14) 0.56 (0.22–1.40) 0.59 (0.26–1.35) 1.18 (0.48–2.93) 0.36 a: Abbreviations are as follows: M, mutation; Wt, wild type. b: GOF (goodness of fit) is the improved fit of the overall model by inclusion of the interactive term; E is either PAL or BMI * SI; second P value is for a three-way interaction that also includes sex. soft drink contributing to a higher proportion of sugar foods for men than for women. The association between SI and p53 mutations observed among women was consistent with that observed when cases with p53 mutations were compared with either controls or with cases who were p53 Wt. This suggests a p53 pathway-specific association with SI. However, it should be kept in mind, when conducting case-case comparison, that a subset of tumors that are p53 Wt includes mutations in other genes, such as Ki-ras and MSI. Ki-ras also is thought to be associated with the insulin-related disease pathway (42). It is possible that Ki-ras is associated with other factors along the insulin-related pathway, although not sugar. Likewise, MSI has been shown to be related to estrogen status (11), although, because it is also related to parity, it is possible that estrogen association with MSI is through a pathway directly involved in sex steroids. Disease pathways are complex and represent a variety of factors that are interrelated in their action. Using exposure data in conjunction with tumor data provides insights into these pathways. Because of the nature of the present work, it was impossible to directly measure insulin levels or metabolic responses to carbohydrates. However, we believe that these observations provide clues for further research and insight into factors that influence risk of colon cancer. These data raise the question of the involvement of estrogen in multiple pathways of colon carcinogenesis. We hypothesize that estrogen influences the association of insulin-related factors and their association with p53-positive colon cancer in women. Vol. 49, No. 1 Acknowledgments and Notes This study was funded by CA48998 and CA61757 to Dr. Slattery. This research was supported by the Utah Cancer Registry, which is funded by Contract N01-PC-67000 from the National Cancer Institute with additional support from the State of Utah Department of Health, the Northern California Cancer Registry, and the Sacramento Tumor Registry. We would like to acknowledge the contributions and support of Margaret Robertson at the University of Utah DNA Sequencing Core Facility, Linda Ballard at the University of Utah Genomics Core Facility, Melanie Nichols and Kristen Gruenthal for laboratory support, and Donna Schaffer, Judy Morse, Sandra Edwards, Leslie Palmer, and Karen Curtin for the data collection and management efforts of this study. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the National Cancer Institute. Address correspondence to M. Slattery, Department of Family and Preventive Medicine, University of Utah Health Research Center, 375 Chipeta Way, Suite A, Salt Lake City, UT 84108. Phone: 801–585–6955. E-mail: mslatter@hrc.utah.edu. Submitted 23 September 2003; accepted in final form 29 March 2004. References 1. McMichael AJ and Potter DJ: Reproduction, endogenous and exogenous sex hormones, and colon cancer: a review and hypothesis. JNCI 65, 1201–1207, 1980. 2. McMichael AJ and Potter JD: Host factors in carcinogenesis: certain bile-acid metabolic profiles that selectively increase the risk of proximal colon cancer. JNCI 75, 185–191, 1985. 3. McMichael AJ and Potter JD: Do intrinsic sex differences in lower alimentary tract physiology influence the sex-specific risk of bowel cancer and other biliary and intestinal diseases? Am J Epidemiol 118, 620–627, 1983. 47 Downloaded by [Huazhong University of Science & Technology ] at 05:27 22 March 2014 4. Potter JD and McMichael AJ: Large bowel cancer in women in relation to reproductive and hormonal factors: a case-control study. JNCI 71, 703–709, 1983. 5. Hebert-Croteau N: A meta-analysis of hormone replacement therapy and colon cancer in women. Cancer Epidemiol Biomarkers Prev 7, 653–660, 1998. 6. Schatzkin A, Lanza E, Corle D, Lance P, Iber F, et al.: The effect of a low-fat, high-fiber, fruit- and vegetable-enriched eating plan on colorectal adenoma recurrence: the Polyp Prevention Trial. N Engl J Med 342, 1149–1155, 2000. 7. Women’s Health Initiative Investigator’s Group: Risks and benefits of estrogen plus progestin in healthy post-menopausal women: principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288, 321–333, 2002. 8. Potter JD: Hormones and colon cancer. JNCI 87, 1039–1040, 1995. 9. Lampe JW, Fredstrom SB, Slavin JL, and Potter JD: Sex differences in colonic function: a randomized trial. Gut 34, 531–536, 1993. 10. Breivik J, Lothe RA, Meling GI, Rognum TO, Borresen-Dale AL, et al.: Different genetic pathways to proximal and distal colorectal cancer influenced by sex-related factors. Int J Cancer 74, 664–669, 1997. 11. Slattery ML, Potter JD, Curtin K, Edwards S, Ma K-Ni, et al.: Estrogens reduce and withdrawal of estrogens increases risk of microsatellite instability-positive colon cancer. Cancer Res 61, 126–130, 2001. 12. Freedman AN, Michalek AM, Troisi R, Mettlin CJ, Petrelli NJ, et al.: Oral contraceptives, reproductive factors and p53 gene expression in colorectal cancer. Carcinogenesis 18, 855–856, 1997. 13. Slattery ML, Curtin K, Ma K, Edwards S, Schaffer D, et al.: Diet, activity, and lifestyle associations with p53 mutations in colon tumors. Cancer Epidemiol Biomarkers Prev 11, 541–548, 2002. 14. Bjorntorp P: Metabolic implications of body fat distribution. Diabetes Care 14, 1132–1143, 1991. 15. Heber D: Interrelationships of high fat diets, obesity, hormones and cancer. Adv Exp Med Biol 399, 13–25, 1996. 16. Cooper GS, Sandler DP, Whelan EA, and Smith DR: Association of physical and behavioral characteristics with menstrual cycle patterns in women age 29–31 years. Epidemiology 7, 624–628, 1996. 17. Harding AH, Williams DE, Hennings SH, Mitchell J, and Wareham NJ: Is the association between dietary fat intake and insulin resistance modified by physical activity? Metabolism 50, 1186–1192, 2001. 18. Slattery ML, Potter JD, Caan BJ, Edwards SL, Coates A, et al.: Energy balance and colon cancer: beyond physical activity. Cancer Res 57, 75–80, 1997. 19. Edwards S, Slattery ML, Mori M, Berry TD, and Palmer P: Objective system for interviewer performance evaluation for use in epidemiologic studies. Am J Epidemiol 140, 1020–1028, 1994. 20. Slattery ML, Caan B, Duncan D, Berry TD, Caan B, et al.: A computerized diet history questionnaire for epidemiologic studies. J Am Diet Assoc 94, 761–766, 1994. 21. Slattery ML, Benson J, Berry TD, Duncan D, Edwards SL, et al.: Dietary sugar and colon cancer. Cancer Epidemiol Biomarkers Prev 6, 677–686, 1997. 22. Slattery ML, Edwards SL, Ma KN, Friedman FG, and Potter JD: Physical activity and colon cancer: a public health perspective. Ann Epidemiol 7, 137–145, 1997. 23. Slattery ML, Edwards SL, Palmer L, Curtin K, Morse J, et al.: Use of archival tissue in epidemiologic studies: collection procedures and assessment of potential sources of bias. Mutagen Res 432, 7–14, 2000. 24. Bonner MR and Ballard LW: Considerations in adding mutation detection services to a sequencing core facility. J Biomol Tech 10, 177–186, 1999. 48 25. Plymate SR, Hoop RC, Jones RE, and Matej LA: Regulation of sex hormone-binding globulin production by growth factors. Metabolism 39, 967–970, 1990. 26. Lipworth L, Adami H-O, Trichopoulos D, Carlstrom K, and Mantzoros C: Serum steroid hormone levels, sex hormone-binding globulin, and body mass index in the etiology of postmenopausal breast cancer. Epidemiology 7, 96–100, 1996. 27. Krauss RM and Burkman RT Jr: The metabolic impact of oral contraceptives. Am J Obstet Gynecol 167, 1177–1184, 1992. 28. Polderman KH, Gooren LJ, Asscheman H, Bakker A, and Heine RJ: Induction of insulin resistance by androgens and estrogens. J Clin Endocrinol Metab 79, 265–271, 1994. 29. Goldfine ID, Papa V, Vigneri R, Siiteri P, and Rosenthal S: Progestin regulation of insulin and insulin-like growth factor I receptors in cultured human breast cancer cells. Breast Cancer Res Treat 22, 69–79, 1992. 30. Katzenellenbogen BS and Norman MJ: Multihormonal regulation of the progesterone receptor in MCF-7 human breast cancer cells: interrelationships among insulin/insulin-like growth factor-I, serum, and estrogen. Endocrinology 126, 891–898, 1990. 31. McKeown-Eyssen G: Epidemiology of colorectal cancer revisited: are serum triglycerides and/or plasma glucose associated with risk? Cancer Epidemiol Biomarkers Prev 3, 687–696, 1994. 32. Giovannucci E: Insulin, insulin-like growth factors and colon cancer: a review of the evidence. J Nutr 131 Suppl, 3109S–3120S, 2001. 33. Liu S, Manson JE, Stampfer MJ, Holmes MD, Hu FB, et al.: Dietary glycemic load assessed by food-frequency in relation to plasma high-density-lipoprotein cholesterol and fasting plasma triacylglycerols in postmenopausal women. Am J Clin Nutr 73, 560–566, 2001. 34. Slattery ML, Ballard-Barbash R, Edwards S, Caan BJ, and Potter JD: Body mass index and colon cancer: an evaluation of the modifying effects of estrogen (United States). Cancer Causes Control 14, 75–84, 2003. 35. Friedlander AL, Casazza GA, Horning MA, Huie MJ, Piacentini MF, et al.: Training-induced alterations of carbohydrate metabolism in women: women respond differently from men. J Appl Physiol 85, 1175–1186, 1998. 36. Haglin L, Lindblad A, and Bygren LO: Hypophosphataemia in the metabolic syndrome. Gender differences in body weight and blood glucose. Eur J Clin Nutr 55, 493–498, 2001. 37. Perseghin G, Scifo P, Pagliato E, Battezzati A, Benedini S, et al.: Gender factors affect fatty acids-induced insulin resistance in nonobese humans: effects of oral steroidal contraception. J Clin Endocrinol Metab 86, 3188–3196, 2001. 38. Tarnopolsky MA, Ettinger S, MacDonald JR, Roy BD, and MacKenzie S: 17-Estradiol [E2] does not affect muscle metabolism in males. Med Sci Sports Exerc 29 Suppl, 93S, 1997. 39. Nicklas BJ, Hackney AC, and Sharp RL: The menstrual cycle and exercise: performance, muscle glycogen, and substrate responses. Int J Sports Med 10, 264–269, 1989. 40. Rooney TP, Kendrick ZV, Carlson J, Ellis GS, Matakevich B, et al.: Effect of estradiol on the temporal pattern of exercise-induced tissue glycogen depletion in male rats. J Appl Physiol 75, 1502–1506, 1993. 41. Tarnopolsky M: Gender differences in metabolism, nutrition and supplements. J Sci Med Sport 3, 287–298, 2000. 42. Dupont J and LeRoith D: Insulin and insulin-like growth factor 1 receptors: similarities and differences in signal transduction. Horm Res 55 Suppl, 22S–26S, 2001. Nutrition and Cancer 2004