Mismatch repair polymorphisms and risk of colon cancer, tumour microsatellite instability and interactions with lifestyle factors.

BACKGROUND: Germline mutations in DNA mismatch repair (MMR) genes cause Lynch syndrome colon cancers. Less understood is the risk of colon cancer associated with common polymorphisms in MMR genes and the potential interacting role of lifestyle factors known to damage DNA. METHODS: A study was conducted to examine whether MLH1 (-93G>A and Ile219Val) and MSH6 (Gly39Glu) polymorphisms were associated with risk of colon cancer in data from 1609 colon cancer cases and 1972 controls. Genotype data were further stratified by microsatellite instability status, smoking, alcohol, Western diet, alcohol and obesity, to investigate potential heterogeneity. RESULTS: The MSH6 39Glu allele was associated with increased risk of colon cancer among men (Gly/Glu or Glu/Glu vs Gly/Gly, OR 1.27; 95% CI 1.04 to 1.54). Neither MLH1 polymorphism was associated with colon cancer risk overall. When stratified by microsatellite stability status, however, the MLH1 -93A allele was associated with a more than doubling in microsatellite instability (MSI)-positive colon cancer risk (AA vs GG, OR 2.47; 95% CI 1.48 to 4.11); no associations were observed between the MMR polymorphisms examined and MSI-negative colon cancer. Statistically significant interactions were observed between: MLH1 -93G>A and smoking (MSI-negative colon cancer only, p value interaction: 0.005); and MLH1 Ile219Val and Western diet (p value interaction: 0.03). CONCLUSIONS: The MSH6 Gly39Glu and MLH1 -93G>A polymorphisms were associated with risk of overall colon and MSI-positive colon cancers, respectively. Risk for colon cancer, stratified by MMR genotype, was further modified by smoking and Western diet.

A prospective cohort of American Indian and Alaska Native people: study design, methods, and implementation.

In 2001, the National Cancer Institute funded three centers to test the feasibility of establishing a cohort of American Indian and Alaska Native people. Participating tribal organizations named the study EARTH (Education and Research Towards Health). This paper describes the study methods. A computerized data collection and tracking system was developed using audio computer-assisted survey methodology with touch screens. Data were collected on diet, physical activity, lifestyle and cultural practices, medical and reproductive history, and family history of heart disease, diabetes, and cancer. In addition, a small panel of medical measurements was obtained, including height, weight, waist and hip circumferences, blood pressure, and a lipid panel plus glucose. At the completion of the enrollment visit, data were used to provide immediate health feedback to study participants. During the initial funding period, the authors anticipate enrolling 16,000 American Indian and Alaska Native participants. The age distribution of the study population was similar to that reported in the 2000 US Census for the relevant populations. A component critical to the success of the EARTH Study has been the partnerships with tribal members. The study has focused on involvement of American Indian and Alaska Native communities in development and implementation and on provision of feedback to participants and communities.

Genetic variation in IGF1, IGFBP3, IRS1, IRS2 and risk of breast cancer in women living in Southwestern United States.

BACKGROUND: An insulin-related pathway to breast cancer has been hypothesized. METHODS: We examine the 19 CA repeat of the IGF1 gene, the -202 C > A IGFBP3, the G972R IRS, and the G1057D IRS2 polymorphisms among 1,175 non-Hispanic white (NHW) and 576 Hispanic newly diagnosed breast cancer cases and 1,330 NHW and 727 Hispanic controls living in Arizona, Colorado, New Mexico, and Utah. RESULTS: Among post-menopausal women not recently exposed to hormones, not having the 19 CA repeat of IGF1 gene was associated with breast cancer among NHW women [odds ratio (OR) 2.14, 95% confidence interval (CI) 1.21-3.79] and having an R allele of G972R IRS1 increased breast cancer risk among Hispanic women (OR 2.70, 95% CI 1.13-6.46). Among post-menopausal Hispanic women recently exposed to hormones the A allele of the -202 C > A IGFBP3 polymorphism increased risk of breast cancer (OR 1.57, 95% CI 1.06-2.33). The IGF1 19 CA repeat polymorphism interacted with hormone replacement therapy (HRT) among NHW post-menopausal women; women who had the 19/19 IGF1 genotype were at reduced risk of breast cancer (OR 0.64, 95% CI 0.47-0.88) if they did not use HRT. We also observed interaction between body mass index and IGF1 19 CA repeat (p=0.06) and between weight gain and the -202 C > A IGFBP3 polymorphism (p=0.05) in NHW post-menopausal women not recently exposed to hormones. CONCLUSIONS: Our data suggest that associations between insulin-related genes and breast cancer risk among women living in the Southwestern United States may be dependent on estrogen exposure and may differ by ethnicity.

Genetic, anthropometric, and lifestyle factors associated with IGF-1 and IGFBP-3 levels in Hispanic and non-Hispanic white women.

Circulating concentrations of IGF-I and IGFBP-3 are associated with risk of pre-menopausal breast cancer. Racial differences in levels of these factors have been reported, and determinants of IGF-I and IGFBP-3 levels within racial and ethnic groups are unclear. In this study we examine genetic, anthropometric, diet, and lifestyle factors that may predict serum levels of IGF-I and IGFBP-3 among Hispanic and non-Hispanic white women. A sample of healthy controls participating in the SHINE (Southwest Hormone, Insulin, Nutrition, and Exercise Study) case-control breast cancer in Arizona, Colorado, New Mexico, and Utah were included in these analyses. Subjects included 210 Hispanic and 284 non-Hispanic white women. Hispanic women had significantly lower levels of IGFBP-3 (mean=3764.3 mcg/ml) after adjusting for age, body size, physical activity, menopausal status, and dietary factors than non-Hispanic white women (mean = 4058.0 mcg/ml; p<0.01). The CC genotype of the -202 A>C polymorphism of the IGFBP3 gene was associated with lower IGFBP-3 levels in both ethnic groups. The frequency of the IGFBP3 C allele differed between Hispanic (0.65) and non-Hispanic white women (0.53), but serum levels of IGFBP-3 were lower for Hispanic women than non-Hispanic after accounting for IGFBP3 genotype. Body size indicators, vigorous physical activity, and dietary factors appeared to influence serum levels of IGF-1 and the ratio of IGF-1 to IGFBP-3 in pre-menopausal women more than in post-menopausal women. On the other hand, using aspirin/NSAIDs appeared to increase IGFBP-3 levels significantly among pre-menopausal Hispanic women. Results from this study suggest that differences in IGFBP-3 levels exist in Hispanic and non-Hispanic white women. These differences could be due to the combined effects of genetic and behavioral factors which could account for ethnic differences in the risk of breast cancer and other chronic diseases.

MTHFR variants reduce the risk of G:C->A:T transition mutations within the p53 tumor suppressor gene in colon tumors.

5,10-Methylene-tetrahydrofolate reductase (MTHFR) is a key enzyme in folate-mediated 1-carbon metabolism. Reduced MTHFR activity has been associated with genomic DNA hypomethylation. Methylated cytosines at CpG sites are easily mutated and have been implicated in G:C-->A:T transitions in the p53 tumor suppressor gene. We investigated 2 polymorphisms in the MTHFR gene (C677T and A1298C) and their associations with colon tumor characteristics, including acquired mutations in Ki-ras and p53 genes and microsatellite instability (MSI). The study population comprised 1248 colon cancer cases and 1972 controls, who participated in a population-based case-control study and had been analyzed previously for MSI, acquired mutations in Ki-ras, p53, and germline MTHFR polymorphisms. Multivariable-adjusted odds ratios are presented. Overall, MTHFR genotypes were not associated with MSI status or the presence of any p53 or Ki-ras mutation. Individuals with homozygous variant MTHFR genotypes had a significantly reduced risk of G:C-->A:T transition mutations within the p53 gene, yet, as hypothesized, only at CpG-associated sites [677TT vs. 677CC (referent group) OR = 0.4 (95% CI: 0.1-0.8) for CpG-associated sites; OR = 1.5 (0.7-3.6) for non-CpG associated sites]. Genotypes conferring reduced MTHFR activity were associated with a decreased risk of acquired G:C-->A:T mutations within the p53 gene occurring at CpG sites. Consistent with evidence on the phenotypic effect of the MTHFR C677T variant, we hypothesize that this relation may be explained by modestly reduced genomic DNA methylation, resulting in a lower probability of spontaneous deamination of methylated cytosine to thymidine. These results suggest a novel mechanism by which MTHFR polymorphisms can affect the risk of colon cancer.

Contacting controls: are we working harder for similar response rates, and does it make a difference?

Although obtaining high response rates is critical to epidemiologic studies, effort to achieve response rates is undocumented. The authors used three population-based case-control studies conducted in Utah between October 1991 and February 2003 to examine effort required for both initial contact and determination of final status. Differences in lifestyle characteristics between easy- or more-difficult-to-interview female controls were evaluated. Letter, phone, and in-person contacts were recorded to determine contact effort. Regarding effort required to achieve a final outcome, the number of contacts increased from eight to 14 over the 12-year study period. Compared with those in study A (conducted in 1991-1994), controls in studies B and C were twice as likely to require seven or more phone calls and controls in study B were twice as likely to require one or more in-person visit. Hispanic controls in study C were more likely than non-Hispanic White controls to receive an in-person visit and a noncontact letter. Compared with those more difficult to contact, those easy to contact were more likely to be overweight and less likely to have a family history of cancer. The amount of effort required to achieve similar or slightly lower response rates increased over time. This finding may in part depend on demographic characteristics of the population studied.

Factors associated with colorectal cancer screening in a population-based study: the impact of gender, health care source, and time.

INTRODUCTION: The effectiveness of colorectal cancer screening in reducing incident colorectal cancer and the risk of death has been shown. Despite campaigns to promote the benefits of and use of colorectal cancer screening, most people are not participating in screening. In this paper, we examine factors associated with screening behavior over time, by health care provider, and by gender and report associations between screening and development of colorectal cancer after adjusting for diet and lifestyle factors. METHODS: Data from two population-based case-control studies of colorectal cancer were used to examine risk associations with nonparticipation in colorectal cancer screening. Study participants were identified for the first study between 1991 and 1994 (N = 1,346 cases and 1,544 controls) and for the second between 1997 and 2001 (N = 952 cases and 1,205 controls) and were asked to complete a detailed in-person interviewer-administered diet and lifestyle questionnaire. The control population is used to examine changes in screening behavior and associations with screening over time. RESULTS: Significantly, fewer people reported fecal occult blood test (FOBT) in 1997-2001 than in 1991-1994 (62.5% in 1991-1994 vs. 47.2% in 1997-2001); a slight nonsignificant increase in sigmoidoscopy screening was reported for these periods among controls (33.9% vs. 36.6%). In the control population, during these periods, there was a statistically significant increase in the number of people who reported having had a sigmoidoscopy for screening rather than for problems (72.6% in 1997-2001 vs. 63.8% in 1991-1994). There were differences in factors associated with screening behavior by time, by sex, and by health care provider, although having a family history of colorectal cancer, having more education, and being male was associated with more screening in all settings. After adjusting for diet and lifestyle factors, we observed that non-sigmoidoscopy screening significantly influenced risk of incident cancer (rectal OR: 2.9; 95% CI, 2.3-3.7; distal tumor OR: 1.8; 95% CI, 1.4-2.3); proximal tumor: 1.4; 95% CI, 1.1-1.8). Nonuse of FOBT also was associated significantly with tumors in the rectal (OR: 1.6; 95% CI, 1.3-1.9) and distal (OR: 1.4; 95% CI, 1.1-1.8) sites. SUMMARY: These data reinforce the importance of screening to reduce risk of colorectal cancer development. However, flexible sigmoidoscopy screening is increasing only modestly over time, and primarily in settings where a significant investment in screening has been made. FOBT screening, which is effective for rectal cancer prevention, is actually decreasing.

Family history and colorectal cancer: predictors of risk.

INTRODUCTION: While the association between family history of colorectal cancer in first-degree relatives and risk of developing colon cancer has been well defined, the association with rectal cancer is much less clear. The purpose of this study is to define rectal cancer risk associated with family history of colorectal cancer in first-degree relatives. We also evaluate diet and lifestyle factors associated with developing colorectal cancer among participants with a positive family history. METHODS: Data were available from two population-based case--control studies of colon and rectal cancer. Participants were members of the Kaiser Permanente Medical Care Program (KPMCP) or residents of the state of Utah. Cases were first primary colon cancer diagnosed between 1991 and 1994 (n = 1308 cases and 1544 controls) or rectal cancer diagnosed between 1997 and 2001 (n = 952 cases and 1205 controls). RESULTS: A family history of colorectal cancer in any first-degree relatives slightly increased risk of rectal cancer (OR: 1.37 95% CI: 1.02-1.85). Family history of colorectal cancer was associated with the greatest risk among those diagnosed at age 50 or younger (OR: 2.09 95% CI: 0.94-4.65 for rectal tumors; OR: 3.00 95% CI: 0.98-9.20 for distal colon tumors; and OR: 7.88 95% CI: 2.62-23.7 for proximal colon tumors). Factors significantly associated with cancer risk among those with a family history of colorectal cancer, included not having a sigmoidoscopy (OR: 2.81 95% CI: 1.86-4.24): a diet not Prudent, i.e. high in fruits, vegetables, whole grains, fish and poultry, (OR: 2.79 95% CI: 1.40-5.56); smoking cigarettes (OR: 1.68 95% CI: 1.12-2.53), and eating a Western diet, i.e. a diet high in meat, refined grains, high-fat foods, and fast foods, (OR: 2.15 95% CI: 1.06-4.35). Physical inactivity was not associated with increased cancer risk among those with a positive family history of colorectal cancer. SUMMARY: These results confirm observations reported by others that a family history of colorectal cancer increases risk of cancer among those diagnosed at a younger age. Associations with family history are weakest for rectal cancer and strongest for proximal colonic tumors. Since several diet and lifestyle factors influence development of cancer among those with a family history of the disease, there appears to be practical approaches for individuals with a family history of colorectal cancer to reduce their cancer risk.

Physical activity and colorectal cancer.

Physical activity has been inconsistently associated with rectal cancer despite the consistent association between physical activity and colon cancer. In this study, the authors evaluated the association between physical activity and rectal cancer using the same questionnaire used to evaluate the previously reported association with colon cancer. A population-based study of 952 incident cases of cancer in the rectum and rectosigmoid junction and 1,205 age- and sex-matched controls was conducted in Utah and northern California at the Kaiser Permanente Medical Care Program between 1997 and 2002. Vigorous physical activity was associated with reduced risk of rectal cancer in both men and women (odds ratio (OR) = 0.60, 95% confidence interval (CI): 0.44, 0.81 for men; OR = 0.59, 95% CI: 0.40, 0.86 for women). Among men, moderate levels of physical activity also were associated with reduced risk of rectal cancer (OR = 0.70, 95% CI: 0.51, 0.97). Participation in vigorous activity over the past 20 years conferred the greatest protection for both men and women (OR = 0.55, 95% CI: 0.39, 0.78 for men; OR = 0.44, 95% CI: 0.30, 0.67 for women). In summary, physical activity was associated with reduced risk of rectal cancer in these data. The reduced risk was similar to that previously observed for colon cancer.

Body mass index and colon cancer: an evaluation of the modifying effects of estrogen (United States).

OBJECTIVE: The association between body mass index (BMI) and colon cancer has been reported to be different for men and women. No prior literature has examined if estrogen influences these differences. METHODS: Using data from an incident population-based case (n = 1,972) and control (n = 2386) study of colon cancer we evaluated if estrogen modifies the association between BMI and risk of colon cancer. RESULTS: Women who were estrogen-negative (postmenopausal women not taking hormone replacement therapy, HRT) were at increased risk of colon cancer regardless of indicator of estrogen status used (i.e. estrogen-negative compared to estrogen-positive women defined as either being premenopausal or postmenopausal women using HRT, OR 1.54, 95% CI 1.23-1.93; no recent exposure to estrogens compared to current or HRT use within the past 2 years, OR 1.58, 95% CI 1.24-2.00; postmenopausal women not currently using HRT compared to postmenopausal women taking HRT, OR 1.65, 95% CI 1.29-2.12). BMI (kg/m2) was not associated with an increased risk of colon cancer among women who were estrogen-negative. However, women who were estrogen-positive experienced a greater than two-fold increase in colon cancer risk if they had a BMI of > 30 relative to those who had a BMI of <23 (for estrogen-positive, OR, 2.50, 95% CI 1.51-4.13; premenopausal, OR 2.19, 95% CI 0.94-5.07; postmenopausal using HRT, OR 3.36, 95% CI 1.58-7.13). Among men the colon cancer risk associated with BMI decreased with advancing age. Physical activity modified the increased colon cancer risk associated with a large BMI. CONCLUSIONS: These data suggest the importance of estrogen in colon cancer etiology. Being estrogen-negative resulted in a significant increased risk of colon cancer. However, BMI significantly increased the risk of colon cancer among women who were estrogen-positive. We hypothesize that estrogen up-regulates IGF-I receptors and IRS-I levels in the colon, which in turn increases susceptibility to obesity-induced increased levels of insulin. We further hypothesize that androgens may have similar effects in men given the decline in colon cancer risk associated with BMI with advancing age.

GSTM-1 and NAT2 and genetic alterations in colon tumors.

OBJECTIVE: Phase II metabolizing enzymes such as glutathione S-transferases and N-acetyltransferase are involved in the detoxification of carcinogens. Genetic variants of genes coding for these enzymes have been evaluated as to their association with colon cancer, both as independent risk factors and as effect modifiers for associations with diet and cigarette smoking. In this study, we evaluate associations between the GSTM-1 genotype and the NAT2-imputed phenotype and acquired mutations in tumors. METHODS: Data is taken from a set of 1836 cases and 1958 controls with colon cancer who were part of a large case-control study of colon cancer and whose tumors were previously analyzed for Ki-ras, p53, and microsatellite instability (MSI). We also evaluate the modifying effects of these genetic variants with diet and cigarette smoking, factors previously identified as being associated with specific tumor alterations. RESULTS: Neither GSTM-1 nor the NAT2-imputed phenotype was independently associated with Ki-ras, p53, or MSI. Cigarette smoking significantly increased the risk of tumors involving the MSI pathway. Additionally, cigarette smoking doubled the risk of p53 transversion mutations among those who were GSTM-1 present. Cases were slightly more likely to have a p53 mutation if they frequently consumed red meat and had the imputed NAT2 intermediate/rapid phenotype relative to slow phenotype/infrequent consumers of red meat (OR 2.0, 95% CI 1.3-3.0 for intermediate/rapid). CONCLUSIONS: These data provide support that diet and cigarette smoking may be associated with specific disease pathways, although GSTM-1 and NAT2 do not independently appear to alter susceptibility to these diet and lifestyle factors.

Uncommon TGFBRI allele is not associated with increased susceptibility to colon cancer.

Previous studies have suggested that an allele of the transforming growth factor-beta type I receptor (TGFBRI) gene that codes for six instead of the usual nine alanines in a polyalanine repeat is associated with an increased susceptibility to colon cancer, and that the six-alanine homozygote is seen only in individuals with some form of cancer. We evaluated this TGFBRI polymorphism in a population-based sample of 252 individuals with colon cancer and 362 age- and gender-matched controls from the state of Utah. TGFBRI genotypes were determined by PCR amplification and length determination of the polyalanine repeat. In addition to the common nine-alanine (9A) allele, we identified six- (6A), eight- (8A), ten- (10A), eleven- (11A), and twelve-alanine (12A) TGFBRI alleles. 6A/9A heterozygotes were seen in similar percentages of colon cancer cases (18.3%) and controls (16.0%). 6A/6A homozygotes were slightly more common in controls than in colon cancer cases (1.4% vs. 0.8%), and none of the controls with the 6A/6A genotype had any of the non-colonic cancers reported in previous studies. We conclude that the 6A TGFBRI allele is not associated with an increased susceptibility to colon cancer at the population level, and that the 6A/6A homozygote is not restricted to individuals with some form of cancer.

Lifestyle factors and Ki-ras mutations in colon cancer tumors.

Heterogeneity in colon tumors implies that environmental, lifestyle, or genetic factors influence the type of mutations seen in tumors. In this study we evaluate the association between previously identified risk factors for colon cancer and Kirsten-ras (Ki-ras) mutations in tumors. The presence of Ki-ras mutations in codons 12 and 13 were determined in a population-based case-control study of colon cancer. Participants were between 30 and 79 years of age at time of diagnosis and include both men and women. Questionnaire data were used to obtain information on lifestyle factors. Valid study data and Ki-ras mutational status were available from 1428 cases of colon cancer, data from 2410 controls were available for comparative purposes. Participants with Ki-ras mutations were more likely to have proximal rather than distal tumors. Cigarette smoking, use of aspirin and/or NSAIDs, use of vitamin/mineral supplements, and consumption of caffeine were associated with both Ki-ras+ and Ki-ras- tumors; the associations were not confounded by dietary intake or other lifestyle factors. Among men, but not among women, those with low levels of physical activity were more likely to have a tumor with a Ki-ras mutation than one without a Ki-ras mutation. However, among women, those with a larger BMI were more likely to have a Ki-ras mutation in their tumor. Given the limited information available on what causes Ki-ras mutations, the information generated from this study indicates that these factors previously associated with colon cancer work through other disease pathways.

The colon cancer burden of genetically defined hereditary nonpolyposis colon cancer.

BACKGROUND & AIMS: Estimates of the frequency of hereditary nonpolyposis colon cancer (HNPCC) based on clinical criteria have varied widely. Recent studies of germline mismatch repair gene mutations have suggested that HNPCC accounts for close to 3% of all colon cancer, but this estimate may have been inflated by inclusion of founder effects peculiar to Finland. We therefore determined by genetic criteria the colon cancer burden associated with HNPCC in a population-based study of 1066 individuals from Utah and California. METHODS: The coding regions of mismatch repair genes hMSH2 and hMLH1 were sequenced from the germline of those individuals whose tumors exhibited microsatellite instability. RESULTS: Microsatellite instability was present in 16% (171/1066) of tumors. Pathogenic germline mismatch repair gene mutations were identified in 7 individuals, and missense amino acid changes of uncertain significance were identified in another 6 individuals. After adjusting for the availability of sufficient germline DNA for sequencing, the 7 clearly pathogenic mutations accounted for 0.86% of colon cancer at the population level. Individuals with these mutations were significantly younger, more likely to have a family history of colon and endometrial cancer, and more likely to have first-degree relatives with a young-age onset of colon cancer than individuals with unstable tumors but without germline mutations (P < 0.01). CONCLUSIONS: We conclude that genetically defined HNPCC accounts for a very small percentage of colon cancer at the population level, a percentage less than that estimated by most previous clinical studies.

Microsatellite instability in sporadic colon cancer is associated with an improved prognosis at the population level.

Some previous studies have reported an improved prognosis in sporadic colon cancers with microsatellite instability, whereas others have not. In addition, relatively few of those reporting an improved prognosis controlled for tumor stage or were population-based. Therefore, we evaluated the relationship between microsatellite instability and prognosis, tumor stage, and other clinical variables in a population-based study of 1026 individuals. Microsatellite instability was determined by the noncoding mononucleotide repeat BAT-26 and the coding mononucleotide repeat in transforming growth factor-beta receptor type II. Significant relationships were seen between microsatellite instability and proximal tumor location, female gender, young and old age at diagnosis, poor histological differentiation, and low tumor stage (P < 0.01). There was a significant relationship between microsatellite instability and improved prognosis, even after adjusting for stage, with a reduction in the risk of death attributable to colon cancer of approximately 60%. Most of this risk reduction occurred in individuals with American Joint Committee on Cancer stage III tumors, although transforming growth factor-beta receptor type II mutations were associated with a significant reduction in colon cancer death in tumors with distant metastases. We conclude that microsatellite instability in sporadic colon cancer is associated with an improved prognosis at the population level.

Dietary intake and microsatellite instability in colon tumors.

Microsatellite instability (MSI) occurs in approximately 15% of colon tumors. Other than relatively rare mutations in mismatch repair genes, the causes of MSI are not generally known. The purpose of this study was to determine if dietary intake of nutrients previously reported as being associated with colon cancer relate specifically to the MSI disease pathway. Data from a population-based case-control study of adenocarcinoma of the colon were used to evaluate associations between dietary intake and MSI. Participants were between 30 and 79 years of age at time of diagnosis and included both men and women. Dietary intake data were obtained from a computerized diet history questionnaire. MSI was evaluated in several ways: by a panel of 10 tetranucleotide repeats, and by 2 mononucleotide repeats, BAT-26 and TGFbetaRII. A total of 1,510 cases had valid study data and tumor DNA on which we were able to obtain MSI status. Cases with and without MSI were compared with dietary data reported by 2,410 population-based controls to determine dietary associations that may be different for these 2 subsets of cases. We compared dietary intake for cases with and without MSI to further determine associations that are specific to the MSI disease pathway. When comparing MSI+ to MSI- tumors we observed that long-term alcohol consumption, especially intake of liquor, increased the probability of having a tumor with MSI [odds ratio (OR) for MSI+ vs. MSI- tumors for alcohol 1.6, 95% confidence interval (CI) 1.0-2.5; OR for liquor 1.6, 95% CI 1.1-2.4]. The likelihood of having MSI in the tumor from the combined effects of high alcohol consumption and smoking cigarettes showed a 70% excess in risk from the additive model. There were some suggestions that high intakes of refined grain might also be associated with MSI+ tumors, although associations were less consistent. Risk estimates for most other dietary factors did not differ substantially by MSI status. Data from this large population-based case-control study of colon cancer indicate that alcohol consumption, especially consumption of liquor, may increase the odds of an MSI+ tumor. Most other dietary factors do not appear operate exclusively in the MSI+ disease pathway.

Inverse relationship between microsatellite instability and K-ras and p53 gene alterations in colon cancer.

Some studies have shown an inverse relationship between microsatellite instability in colon cancer and mutations in p53 and K-ras, whereas others have not. We therefore evaluated these features in a population-based sample of 496 individuals with colon cancer. Microsatellite instability was determined by a panel of 10 tetranucleotide repeats, the Bethesda consensus panel of mono- and dinucleotide repeats, and coding mononucleotide repeats in transforming growth factor-beta receptor type II, hMSH3, BAX, hMSH6, and insulin-like growth factor receptor type II. Mutations in codons 12 and 13 in K-ras were evaluated by sequencing. p53 overexpression (as detected by immunohistochemistry) was used as an indicator of p53 mutation; this was evaluated in 275 of the tumors. K-ras mutations were present in 33.2% of tumors, p53 overexpression in 51.5%, and microsatellite instability (as determined by the Bethesda consensus panel) in 12.5%. K-ras mutations were significantly less common in unstable tumors than stable tumors (11.8% versus 36.9%, P: < 0.001). p53 overexpression was significantly less common in unstable tumors than stable tumors (20.0% versus 55.7%, P: < 0.001). These inverse relationships between microsatellite instability and ras gene mutations and p53 overexpression were shown to be independent of tumor site in logistic regression analyses. All other measures of instability also showed statistically significant inverse relationships independent of tumor site with alterations in ras and p53, and instability results determined by the panel of 10 tetranucleotide repeats were highly significantly related to those determined by the Bethesda consensus panel. Coding mononucleotide repeat mutations were significantly more common in unstable tumors than stable tumors (85.7% versus 1.0%, P: < 0.001). We conclude that there is an inverse relationship between microsatellite instability and mutations in p53 and K-ras, and that the molecular profile of colon cancers with microsatellite instability is characterized by relatively infrequent mutations in K-ras and p53 and relatively frequent mutations in coding mononucleotide repeats.

Estrogens reduce and withdrawal of estrogens increase risk of microsatellite instability-positive colon cancer.

There are sex differences in the occurrence of microsatellite instability (MSI) in colon tumors. Taken together with the epidemiological evidence that hormone replacement therapy (HRT) and, less consistently, parity, are inversely associated with colon cancer, it has been hypothesized that estrogens are associated with MSI. The purpose of this study was to evaluate sex-specific differences in the prevalence of MSI in colon tumors and to determine whether reproductive history and hormonal exposures are associated with MSI. Using data from a population-based case-control study of 1836 cases with MSI data and 2410 population-based controls, we evaluated sex, reproductive factors, and hormone exposure in relation to the presence or absence of MSI in tumors. MSI was evaluated by a panel of 10 tetranucleotide repeats, the noncoding mononucleotide repeat BAT-26, and the coding mononucleotide repeat in transforming growth factor beta receptor type II (TGFbetaRII). Exposure data on reproduction, hormone use, obesity, and physical activity were obtained from an interviewer-administered questionnaire. Women were less likely then men to have MSI+ tumors at a young age and more likely to have unstable tumors at an older age; we observed a significant interaction (P < 0.01) between age, sex, and MSI. Evaluation of reproductive factors showed that women who had ever been pregnant had half the risk of MSI+ tumors compared with women who had never been pregnant. In complementary fashion, total ovulatory months were associated with an increased risk of MSI+ tumors [odds ratio (OR), 2.1; 95% confidence interval (CI), 1.1-4.0 comparing MSI+ versus MSI- tumors]. Age at first and last pregnancy did not influence the association. The observed associations were strongest among women <60 years of age at the time of diagnosis. Having used oral contraceptives was associated with a lower risk of MSI+ tumors (OR, 0.7; 95% CI, 0.4-1.2); recent users of HRT were at a reduced risk of MSI+ tumors (OR, 0.8; 95% CI, 0.5-1.4); and women who were former HRT users were at an increased risk of MSI+ tumors (OR, 1.8; 95% CI, 1.1-3.0). Obesity and lack of physical activity were associated with an elevated risk of both MSI+ (OR, 1.7; 95% CI, 0.7-3.3) and MSI- (OR, 2.2; 95% CI, 1.7-3.) tumors in men, but only with MSI- (OR, 1.5; 95% CI, 1.1-2.2) tumors in women. The excess of MSI+ tumors in women is explained by the excess of MSI+ tumors at older ages. Our data suggest that estrogen exposure in women protects against MSI, whereas the lack of estrogen in older women increases risk of instability. HRT in these older women may, again, reduce the risk of unstable tumors. A model for the way in which estrogens (endogenous, exogenous, and obesity-associated) modify the risk of MSI+ tumors is proposed.