Diet, Lifestyle, and Genomic Instability in the North Carolina Colon Cancer Study

Introduction

Colon cancer is the second leading cause of cancer death in the United States and the third most common cancer among adults, accounting for ∼15% of all cancers diagnosed annually (1). Colon tumorigenesis results from a loss of genomic stability that leads to the transformation of normal colonic epithelial cells to colon adenocarcinoma cells (2). Microsatellite instability (MSI), defined as a change in DNA microsatellite length because of insertion or deletion of nucleotides in tumor DNA when compared with normal tissue (3), represents one form of genomic instability found in almost all hereditary nonpolyposis colorectal cancers and 10% to 20% of sporadic colon carcinomas (3-5).

In addition to family history, environmental and lifestyle factors have been consistently associated with colon cancer risk (6-9). In particular, dietary factors are among the most widely accepted etiologic risk factors, and it is estimated that 12% of colon cancer are attributable to consumption of a Western-style diet (8). Given the strong associations between diet and colon cancer risk, there is considerable scientific interest in determining whether dietary factors associated with colon cancer may influence genomic instability. In addition to providing further corroboration of a hypothesized relationship, knowledge of associations between diet and molecular alterations can yield valuable insight into the possible mechanisms by which a dietary factor may contribute to malignancy.

There have been very few epidemiologic investigations of diet and MSI, and to date, results have been somewhat inconsistent. For example, Wu et al. (10) found a positive association between red meat intake and MSI in a case-only study, whereas Slattery et al. (11) observed no such association in a case-control study. Diergaarde et al. (12) reported recently positive associations of red meat and alcohol intakes and inverse associations of fruit and vegetable consumption with MSI using data from a case-control study in the Netherlands. However, to our knowledge, these relationships have not been examined in racially heterogeneous study populations.

In this report, we examined associations between various dietary factors and genomic instability in a large population-based case-control study in North Carolina that was racially diverse with a sizeable proportion of African Americans. We focused on nutrients and food groups found previously to be linked with colon cancer risk in this population from North Carolina.

Materials and Methods

Study Design and Population

The North Carolina Colon Cancer Study (NCCCS) is a population-based case-control study of colon cancer in North Carolina. Details on the NCCCS design, population, and data collection methods have been published (13). In brief, study participants were from 33 counties in the central and eastern portions of North Carolina, an area that includes rural, suburban, and urban counties with a diverse socioeconomic mix of African Americans and Whites. The study was approved by the Institutional Review Board at the University of North Carolina School of Medicine and by equivalent committees at the collaborating hospitals.

Cases were persons with a first diagnosis of histologically confirmed invasive adenocarcinoma of the colon between October 1, 1996 and September 30, 2000 identified through the rapid ascertainment system of the North Carolina Central Cancer Registry (14). Other eligibility criteria included ages 40 to 80 years at the time of diagnosis, residence in the 33-county study area in North Carolina, ability to give informed consent and complete the interview, a North Carolina driver's license or identification card if <65 years old (because controls <65 years old were sampled from driver's license rosters), and permission to contact from the primary physician. Eligible and consenting patients were generally interviewed within 5 months of surgery. White cases were undersampled to increase the proportion of non-White cases in the study population.

The noninstitutionalized population-based controls were selected from two sources: North Carolina Division of Motor Vehicle records for cases <65 years old and Center for Medicare and Medicaid for cases ≥65 years old. These listings were used to randomly select potential controls within the same 5-year age group-, sex-, and race-defined strata. Study participants were selected using a randomized recruitment approach to achieve approximate frequency matching on age, sex, and race and to enhance the proportion of African Americans (15, 16). Those identified as eligible controls were contacted in a similar fashion to the cases to schedule in-person interviews.

Completed interviews were obtained from 1,691 participants. The overall study cooperation rate [interviewed / (interviewed + refused)] was 84% for cases and 63% for controls, whereas the response rate (interviewed/eligible) was 72% for cases and 61% for controls. For both cases and controls, the cooperation and response rates were slightly higher for Whites than for African Americans (13).

Data Collection

Data were collected in person by trained nurse interviewers at the participant's home or, occasionally, at another convenient location. The questionnaire collected detailed information on several factors that might relate to colon cancer, including diet, lifestyle factors, and medical history. The referent period for the interview was the year before diagnosis (cases) or interview date (controls).

Dietary Intake

Food (nutrient and food group) intake was assessed with a modified version of the previously validated 100-item semiquantitative Block food frequency questionnaire that was adapted to capture regional dietary practices by adding 29 foods commonly consumed in North Carolina, such as cooked greens, black-eyed peas, fried shellfish, hushpuppies, grits, and cobblers (17). In the present study, participants were asked to estimate their usual frequency of consumption of various foods and typical portion sizes for the year before diagnosis (for cases) or the year preceding the interview date (for controls). Each food item had nine options for frequency (ranging from “never or less than once per month” to “two or more times per day”) and three options for portion size. The food frequency questionnaire also included adjustment questions on types of foods used in cooking and preparation techniques and questions relating to restaurant eating, consumption of low-fat foods, fortified beverages, and fats used in cooking (18). Food groups and nutrient intakes were generated by an analysis program provided by the National Cancer Institute (19).

Dietary supplement use was assessed with closed-ended questions on the use of multiple vitamins (including antioxidant combinations) and single supplements. Participants were asked specifically about duration (years), frequency (days per week), and usual dose (e.g., 100, 250, 500, 750, and 1,000 mg for vitamin C) of supplements used over the past year. Because supplements contribute a large proportion of micronutrient intakes (18), we summed across intakes from food sources and supplements to obtain total micronutrient exposure.

For these analyses, we examined only nutrients and food groups found previously to be significantly associated with colon cancer risk in this study sample (Tables 2 and 3; refs. 13, 20, 21). Participants with reported energy intakes <800 and >5,000 kcal for men and <600 and >4,000 kcal for women were excluded because their food frequency questionnaires were considered unreliable (22).

Other Participant Characteristics

Data were collected on demographic characteristics, including age, sex, education, race, smoking history, physical activity, nonsteroidal anti-inflammatory drugs (NSAID) use over the last 5 years, and first-degree family history of colon cancer. Height and weight were measured by trained staff using a standardized protocol and used to compute body mass index (BMI) as weight (in kilograms) divided by height (in meters) squared. BMI (kg/m²) was further categorized as “normal” (18.5-24.9), “overweight” (25-29.9), and “obese” (≥30; ref. 23). Cases were also asked their weight 1 and 5 years before diagnosis. Physical activity was measured in metabolic equivalent task-minutes per day for combined occupational, nonoccupational, and nonwork/weekend activities (including duration, frequency, and intensity) using a modified version of a validated 7-day physical activity recall (24, 25).

Microsatellite Instability

DNA Extraction. DNA was extracted from formalin-fixed, paraffin-embedded tissues of each participant's colorectal tumor and surrounding noncancerous tissue. Each area to be microdissected was identified on a reference H&E-stained slide. Unstained tissue slices adjacent to the reference slide were microdissected according to areas identified on the reference slide using a surgical scalpel blade. The dissected specimen was deparaffinized in a microfuge tube with xylene, and the DNA was purified with ethanol and GeneReleaser (BioVentures, Murfeesboro, TN) according to the manufacturer's recommendations. The samples were then treated with 200 μg/mL proteinase K (Sigma, St. Louis, MO) and incubated at 55°C for 5 hours. The proteinase K was destroyed by heating the sample to 95°C for 15 minutes and the samples were immediately iced and stored for PCR analysis.

Microsatellite Analysis. For determination of MSI, we used the National Cancer Institute–recommended panel of five microsatellite markers (BAT25, BAT26, D5S346, D2S123, and D17S250) to classify tumors as MSI-high [MSI-H; associated with inactivation of DNA mismatch repair (MMR)] or MSI-low or microsatellite stable (MSI-L/MSS; not associated with DNA MMR inactivation; ref. 3). MSI-H was defined as ≥2 markers demonstrating novel alleles compared with nontumor tissue, MSI-L as 1 marker with a novel allele, and MSS as no marker with novel alleles. Because of the similar features of MSI-L and MSS tumors, these two groups were combined as non-MSI-H tumors for these analyses. One primer from each microsatellite marker pair was radio-end-labeled with ³²P in a reaction containing the primer, kinase buffer, T4 polynucleotide kinase, and [γ-³²P]ATP. PCR was done on the microdissected template DNA in a reaction containing 0.125 pmol each of the end-labeled and “cold” primers, 0.25 units Taq DNA polymerase, 40 mmol/L deoxynucleotide triphosphate stock solution, and final concentration of 1.5 to 2.0 mmol/L magnesium. PCR products were denatured in 95% formamide and electrophoresed on a 6% polyacrylamide gel containing 7.5 mol/L urea. The gels are then dried and exposed by autoradiography to X-ray film or in a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

Statistical Analyses

After all exclusions, the analytic data set included 875 Whites (267 cases and 608 controls) and 659 African Americans (219 cases and 440 controls). Analyses were initially stratified by race; however, there was no evidence of effect modification, so results are presented for all participants combined. Descriptive statistics (raw means, SDs, and percentages) stratified by case (MSI-H or MSI-L/MSS) and control status were used to describe the demographic/lifestyle characteristics and dietary intakes of study participants (Tables 1 and 2). Logistic regression models were used to generate Ps to examine whether there were statistically significant differences between cases/cases and cases/controls for the participant characteristics and dietary factors under examination after adjusting for potential covariates. For Table 1, potential confounders included all the characteristics (except current BMI), whereas covariates for Table 2 were variables in Tables 1 and 2 that changed the variable coefficient of the dietary factor under examination by at least 15% when added to the model.

We calculated odds ratios (OR) and 95% confidence intervals (95% CI) from unconditional logistic regression models to ascertain associations of the dietary variables (in tertiles) with MSI status. Offset terms were included in all models to correct for randomized recruitment sampling fractions (15, 16) and allow estimation of unbiased ORs. This was necessary because we conditioned recruitment on age, sex, and race in addition to disease status; thus, the ORs without the offset term will be biased compared with a traditional design in which recruitment is conditioned on disease status alone. Cut points for tertiles of nutrient intakes were determined based on the distributions among controls. All variables in Tables 1 and 2 (except current BMI) were evaluated as potential confounding factors; covariate inclusion was based on whether there was a ≥15% alteration in the variable coefficient of interest. Covariates that met this criterion were placed in the model simultaneously and a backward-stepwise regression was done to determine the final model. All ORs are reported for energy-adjusted and non-energy-adjusted nutrient intakes. The standard multivariate method was used to adjust for total energy intake; applying other energy adjustment approaches (e.g., the nutrient residual and energy partition models) did not alter the results (22). Statistical tests were two sided and Ps < 0.05 were considered statistically significant. All analyses were done using SAS 8.1 (SAS Institute, Inc., Cary, NC).

Results

Demographic and lifestyle characteristics of study participants, stratified by colon cancer and MSI status, are given in Table 1. Ten percent (49 of 486) of the cases were classified as MSI-H, of whom 29% were African American. There were 437 MSI-L/MSS cases (47% African American) and 1,048 controls (42% African American). The majority (78%) of MSI-H cases had proximal tumors, whereas MSI-L/MSS cases more often had distal tumors. MSI-H cases were older than those with MSI-L/MSS tumors but slightly younger than controls (P < 0.0001). Compared with MSI-L/MSS cases and controls, cases with MSI-H tumors were also more likely to be female, college educated, former or never smokers, have a family history of colon cancer, and never or occasional users of NSAIDs during the preceding 5 years (all Ps < 0.05). Compared with controls, MSI-L/MSS cases had a higher mean BMI in the year before diagnosis, were more likely to have a first-degree relative with colon cancer, and less likely to have used NSAIDs over the past 5 years or to have used vitamin and mineral supplements over the previous year (all Ps < 0.05).

Table 2 gives mean intakes of total energy, micronutrients, macronutrients, dietary fiber, and food groups by colon cancer and MSI status. Overall, colon cancer cases reported higher intakes of total energy and macronutrients and lower intakes of micronutrients than controls. Specifically, compared with controls, MSI-H cases reported significantly higher intakes of total energy, carbohydrate (g/d), effective carbohydrate (13), total fat (g/d), saturated fat (g/d), and red meat (servings/wk) but lower β-carotene and percentage energy from carbohydrate (all Ps < 0.05). Similarly, cases with MSI-L/MSS tumors had higher total energy, total fat (g/d), and alcohol intakes but reported significantly lower intakes of all micronutrients, dietary fiber, vegetables (servings/wk), and percentage energy from protein than controls (all Ps < 0.05). There were no statistically significant differences in reported intakes of these foods or nutrients between MSI-H and MSI-L/MSS cases.

Results of case-control and case-case comparisons examining associations (ORs and 95% CI) of dietary factors with MSI-H and MSI-L/MSS tumors are given in Table 3. Because previous analyses have shown appreciable differences in results based on whether ORs were adjusted for total energy (13, 20, 21), both sets of results are presented here. Compared with controls and regardless of energy adjustment, MSI-H tumors were not statistically significantly associated with any of the dietary factors examined, except for a strong inverse association with energy-adjusted β-carotene (OR, 0.4; 95% CI, 0.2-0.9) and a positive association with non-energy-adjusted refined carbohydrates (OR, 3.5; 95% CI, 1.5-8.2). MSI-H tumors were also associated with energy-adjusted dietary fiber intake (OR, 0.4; 95% CI, 0.2-1.0) and red meat without control for total energy (OR, 2.0; 95% CI, 0.9-4.2). When cases with MSI-L/MSS tumors were compared with controls, there were inverse associations of MSI-L/MSS with vitamin C and dark green vegetables regardless of energy adjustment and vitamin E, calcium, and dietary fiber after energy adjustment (all Ps for trend < 0.05). MSI-L/MSS tumors were also statistically significantly associated with total energy and non-energy-adjusted carbohydrate, effective carbohydrate, total fat, saturated fat, and red meat. In case-case comparisons, there was no evidence that any of the dietary factors were significantly differently related to MSI-H compared with MSI-L/MSS tumors.

Discussion

MSI is one of the molecular mechanisms leading to genomic instability in colon carcinogenesis. Few studies have examined the possible associations of dietary factors with MSI. Using data from participants enrolled in the NCCCS, we compared intakes of total energy, macronutrients, micronutrients, and food groups in sporadic colon cancer cases and population-based controls in relation to MSI status. According to our results, consumption of refined carbohydrates (cereals) and red meat may promote the development of MSI-H tumors whereas β-carotene was inversely associated with MSI-H. In case-case comparisons, no dietary factors were significantly differently related to MSI-H compared with MSI-L/MSS tumors. We also observed significant inverse associations between MSI-L/MSS and energy-adjusted vitamin C, vitamin E, calcium, dietary fiber, and dark green vegetables and positive associations with total energy and non-energy-adjusted intakes of red meat and most macronutrients.

In MSI-H versus control comparisons, β-carotene was significantly inversely associated with MSI-H. Epidemiologic studies suggest that high consumption of fruits and vegetables, particularly those high in carotenoids, may play a role in preventing colon cancer. In previous analysis in our study sample, β-carotene was significantly associated with a reduction in colon cancer risk (20). There are few studies that have examined β-carotene in relation to MSI and colon cancer. Similar to our findings, Slattery et al. (11) observed a trend toward a protective effect of β-carotene on MSI-H, but the results were not statistically significant. Although the mechanisms by which carotenoids may lower cancer risk are not entirely understood, possibilities include their antioxidant activities, inhibition of cell proliferation, induction of detoxifying enzymes (26), stimulation of gap junction communication, and enhancement of immune function (27). β-Carotene may also inhibit growth and promote apoptosis in colon cancer through the redox regulation of nuclear factor-κB activity (28).

Our finding of a positive association between refined carbohydrates and MSI-H is in agreement with a report by Slattery et al. (11), whose data were suggestive of an association between high refined grains and increased likelihood of having MSI-positive colonic tumors. High consumption of refined carbohydrates and factors related to insulin resistance have been associated with colon cancer. Epidemiologic studies that have shown associations of glycemic index and load, glucose intolerance, and insulin with colon cancer (29, 30) suggest a role for refined carbohydrates in the etiology of colon cancer. The mechanism by which refined carbohydrates influence MSI in colon cancer is unknown. One possibility is that glucose may induce changes in gene expression. For example, in a recent study, Fornoni et al. observed that exposure of mesangial cells to high glucose concentrations in vitro induced MSI. Thus, dietary sugars, such as refined glucose, may create conditions to produce significant structural and biological alterations that might include genomic instability within DNA (31).

High intake of red meat was positively associated with MSI-H tumors compared with controls, suggesting that red meat intake may contribute to microsatellite unstable tumors in the colon. Similar observations have been reported by some but not all studies. Wu et al. found a positive association between intake of well-done red meat and MSI-H in a case-only study but did not observe a significant association with red meat overall (10). Slattery et al. observed no associations between red meat intake and MSI status in a population-based case-control study of colon cancer (11). Other studies have reported positive associations between red meat consumption and MSI-L/MSS but not MSI-H colon tumors (12). Red meat intake has also been positively associated with MSI-H in gastric cancer (32). Reasons for the dissimilarities between our study and others may relate to differences in meat cooking practices and/or differences in composition of study populations: more specifically, our study population was racially diverse with a sizeable number of African Americans, whereas most studies in the literature consist predominantly of Caucasians.

The underlying mechanism between red meat intake and MSI-H is unclear; however, there are several possible explanations for the observed associations. Red meat prepared at high temperatures produces heterocyclic amines, which are bulk-adduct-forming agents that have been associated with increased cancer risk. Colon cells may also be more susceptible to DNA damage due to reduced MMR activity (33-35). In vitro cell studies suggest that dietary factors may modulate the relationship between MMR and cancer risk. In one study, murine cells that were deficient in MMR activity were more susceptible to DNA damage and likely to escape apoptosis following exposure to chronic oxidative stress (33, 36). Thus, in susceptible individuals with decreased MMR function, the genes involved in the MSI-H pathway may be more susceptible to mutagens in red meat that are produced during cooking. Another possible explanation for the association between red meat and MSI-H tumors relates to dietary heme in red meat. Heme has been proposed as the link between red meat intake and increased risk of colon cancer because of its cytotoxic and proliferative effects on colonic cells. The increased proliferation may lead to an elevated frequency of spontaneous mutations in target genes that may be repaired less efficiently in susceptible individuals (32).

Compared with controls, we observed significant associations between MSI-L/MSS and several nutrients as noted above. Reasons why different dietary factors may be associated with MSI-L/MSS versus MSI-H tumors are not clear. We also note that in case-case comparisons many of the dietary factors that were associated with MSI-H were also associated with MSI-L/MSS, but no dietary variables were significantly differently related to MSI-H compared with MSI-L/MSS tumors, which may be due to the small number of participants with MSI-H tumors (n = 49). Alternatively, this finding may suggest that factors other than those related to diet might be involved in the pathology of microsatellite unstable tumors.

Our findings did not suggest racial differences in the associations of dietary factors with MSI. This was somewhat surprising because the available evidence suggests that MSI may be more frequent among African Americans compared with Whites and the elevated risk of MSI tumors among African Americans has been postulated to be related to dietary differences (37). Possibly, the small number of MSI-H cases precluded our ability to detect statistically significant differences by race. However, there were no differences between African Americans and Whites for comparisons between controls and cases with MSI-L/MSS tumors, for which we had larger numbers. Nonetheless, given that there are racial differences in colon cancer risk and diet-colon cancer associations (1, 13, 20, 21, 38), the influence of diet on MSI in colon cancer merits further exploration in larger, racially diverse study samples.

The strengths of our study include a population-based study design, a racially diverse population with a large number of African Americans, rapid case ascertainment, and detailed exposure information obtained through in-person interviews. As with other case-control studies, selection and differential bias are some potential limitations. Specifically, selection bias is a concern due to the lower participation rates of controls, and differential recall between cases and controls is a possibility because exposure information was collected after diagnosis of the disease; in particular, cases may recall dietary exposures differently from controls because of the presence of their illness and/or symptoms. Nonetheless, we made every effort to reduce bias; for example, we limited most of our exposure information to 1 year before diagnosis for cases or the year before the interview for controls and exposure information was collected as soon as possible after diagnosis. Our response rates were also comparable with previous population-based case-control studies (11, 32, 39, 40).

In summary, our findings indicate that refined carbohydrates and red meat consumption may promote the development of MSI-H colon cancer, whereas β-carotene may be associated with a lower risk of MSI-H tumors. It is possible that differences between diet and colon cancer risk by phenotype may in part explain inconsistencies across studies of diet and cancer. Identification of the underlying mechanisms by which dietary factors affect tumor genomic stability might prove useful in the treatment of colon cancer as well as in the incorporation of dietary factors into colon cancer prevention strategies.