APC Asp1822Val and Gly2502Ser Polymorphisms and Risk of Colorectal Cancer and Adenoma

In the United States, colorectal cancer is expected to result in 146,000 newly diagnosed cases and over 56,000 deaths in 2004 (1). Colorectal cancers usually develop over many decades, with >90% of cases diagnosed after age 50 (1). Normal colonic epithelial cells develop into neoplastic cells via the adenoma-carcinoma sequence (2). Many genomic alterations take place during this sequence of events and usually involve the mutation or loss of the adenomatous polyposis coli (APC) tumor suppressor gene (2). APC is in part responsible for controlling the Wnt signaling pathway that regulates the proliferation, migration, and differentiation of cells in the intestinal epithelium. Wnt signaling leads to a nuclear accumulation of β-catenin, which interacts with T-cell factor/lymphoid enhancing factor to initiate transcription of genes controlling proliferation in the lower portion of the mucosal crypt (3). Cells undergoing differentiation in the upper crypt regulate β-catenin through a multiplex protein complex involving the APC tumor suppressor protein, axin, and glycogen synthase kinase that marks β-catenin for degradation (4). Disregulation of the Wnt/β-catenin signaling pathway leads to nuclear accumulation of β-catenin resulting in aberrant cellular proliferation in the upper crypt. Wnt/β-catenin overactivation, often resulting from mutated APC, leads to imbalanced cellular proliferation and differentiation and has been documented in tumors and is implicated in human cancer (5). APC has also been shown to play a role in apoptosis (6), cellular adhesion (7), and chromosome segregation (8).

Mutations in the APC gene initiate the early stages of the adenoma-carcinoma sequence in both familial and sporadic colon tumorigenesis. Truncating germ line mutations in the APC gene are responsible for most cases of familial adenomatous polyposis (9), which is characterized by early onset polyposis (the presence of hundreds to thousands of adenomatous polyps in the colon and rectum). Familial adenomatous polyposis is an autosomal dominant inherited disorder that accounts for about 1% of all colorectal cancers. The APC gene also acquires somatic mutations in 80% to 90% of sporadic colorectal adenomas (10, 11). In addition, several germ line APC mutations lead to a form of sporadic colorectal cancer that does not exhibit the polyposis phenotype, although most of these genetic variants are rare (<1%; refs. 11-16).

The most common APC variant is Asp1822Val with a variant allele frequency of ∼22% (17). The functional significance of the Asp1822Val substitution is unknown, although this amino acid change is located in the center of a β-catenin down-regulation domain (18). The clinical relevance of this polymorphism is also uncertain and previous work suggests that Asp1822Val may either be a low-penetrance allele that increases risk of developing colorectal cancer (16) or a common polymorphism without clinical implication (19). Slattery et al. (17) found that individuals consuming a low-fat diet were at reduced risk if they were homozygous for the 1822Val allele, suggesting that the influence of dietary and lifestyle factors on risk of colorectal cancer might vary by Asp1822Val genotype. Another APC variant, Gly2502Ser, has been reported in SNP500 (http://snp500cancer.nci.nih.gov) to have a variant allele frequency of 10% in Caucasians and has not been previously studied in relation to risk of colorectal neoplasia.

The aim of this study was to investigate the APC Asp1822Val and Gly2502Ser polymorphisms and their association with colorectal cancer and colorectal adenoma. We also examined whether the association of several dietary (including fat intake) and lifestyle factors with colorectal cancer and colorectal adenoma are modified by the Asp1822Val polymorphism. We analyzed samples from three case-control studies nested in the Nurses' Health Study (NHS), the Health Professionals Follow-up Study (HPFS), and the Physicians' Health Study (PHS) cohorts.

The NHS is an ongoing prospective study of 121,700 predominantly Caucasian American U.S. female registered nurses. Details of the design and follow-up of this cohort have been described previously (20). Briefly, upon enrollment in 1976, the participants, who were 30 to 55 years old, completed a mailed questionnaire providing information on risk factors for cancer and cardiovascular disease. Biennially, updated exposure and disease information is collected by mail, including reports of endoscopy and polyp diagnosis. Self-reported diagnoses of colorectal adenomas are confirmed through histopathologic reports reviewed by a study investigator. From 1989 to 1990, blood samples were collected from 32,826 of the NHS participants. To be eligible for selection as a case or control for these colorectal adenoma analyses, women must have supplied a blood sample between 1989 and 1990, must have undergone sigmoidoscopy or colonoscopy after the date of return of the blood sample, and not have had a cancer or adenoma diagnosis, excluding nonmelanoma skin cancer, before the date of endoscopy. One control was matched to each case according to year of birth, month of blood draw, fasting status, time period of endoscopy (within 2 years), and routine screening, gastrointestinal symptoms, or family history of colorectal cancer as indication(s) for endoscopy. A total of 557 cases and 557 matched controls were identified. Subsequently, one colon polyp case was identified as hyperplastic and removed from the final analysis, which included 556 cases and 557 controls. Among eligible women from whom blood samples were collected, 197 incident colorectal cancer cases were documented by the year 2000 and confirmed using medical records. Four hundred and ninety women who were free from cancer at the time of case assessment were selected as controls for the cancer cases and matched on year of birth and month of blood draw.

The HPFS is an ongoing prospective study of the causes of chronic diseases in men. The cohort began in 1986 when 51,529 U.S. predominantly Caucasian American male dentists, optometrists, osteopaths, podiatrists, pharmacists, and veterinarians, ages 40 to 75 years, responded to a mailed questionnaire (21). These men provided baseline information on age, marital status, height and weight, ancestry, medications, smoking history, medical history, physical activity, and diet. Exposure and medical history information are updated every 2 years. When a participant reported a diagnosis of colorectal adenomas on the follow-up questionnaires, we asked for permission to acquire the relevant medical records. All cases in this analysis were confirmed through histopathologic reports reviewed by a study investigator. Blood samples were collected between 1993 and 1994 from 18,025 participants among whom 367 had been diagnosed with adenomatous polyps between 1986 and 1994 and were confirmed by medical records. To be eligible for selection as a case or control, a man must have completed a valid dietary questionnaire in 1986, supplied a blood sample, and have undergone sigmoidoscopy or colonoscopy after the date of return of the blood sample, and not have had a cancer diagnosis, excluding nonmelanoma skin cancer, before the date of endoscopy. Seven hundred thirty-six men who were free from diagnosed polyps at the time of case assessment were selected as controls and matched to each case on year of birth, whether they had had a previous endoscopy (yes/no), and year of endoscopy. Subsequently, nine controls were identified as cases; thus, the final analysis included 376 cases and 725 controls.

The PHS is a randomized, double-blind trial of aspirin and β-carotene among 22,071 predominantly Caucasian American male physicians, 40 to 84 years of age in 1982. Blood samples were collected at baseline, in 1982, from 14,916 (68%) of the randomized physicians. The men were subsequently followed for incident cancer through annual mailed questionnaires. By the year 2000, 274 cases of colorectal cancer were identified and confirmed using medical records. Men who were free from diagnosed cancer at the time of cancer ascertainment (456) were selected as controls and were matched on age (±1 year up to ±5 years for older men) and on smoking history at baseline (current, former, and never smokers).

In the NHS and HPFS, venous blood samples were separated into plasma, buffy coat, and RBC and stored in liquid nitrogen. Genomic DNA was extracted from whole blood samples (PHS) or from 50 μL buffy coat diluted with 150 μL of PBS and using the QIAmp (Qiagen, Inc., Chatsworth, CA) 96-spin blood protocol according to the manufacturer's instructions. Genomic DNA concentrations were calculated in 96-well format using PicoGreen technology (Molecular Probes, Eugene, OR).

Genotyping of the APC Asp1822Val (rs459552) was carried out using the Taqman allelic discrimination system using forward (5′-CAGACAACAAAGATTCAAAGAAACAGA-3′) and reverse (5′-GCAAAACTTCCTCTGACTCTATCTTCA-3′) oligonucleotide primers and minor groove binder (MGB) probes (5′ FAM-CCAAGGTCTTCAATTG and 5′ VIC-CCAAGGACTTCAATTG). Taqman genotyping of APC Gly2502Ser (rs2229995) was carried out using forward (5′-GTCTCTATCCACACATTCGTCTGT-3′) and reverse (5′-GACTGAGATTAGGTGGGAGTTTTCG-3′) oligonucleotide primers and minor groove binder (MGB) probes (5′ FAM-CATCCACCAGCCTGA and 5′ VIC-CCATCCACTAGCCTGA). PE Applied Biosystems (Foster City, CA) synthesized Taqman primers and probes. Following PCR amplification, end-point fluorescence was read with the Applied Biosystems 7900HT instrument and genotypes were assigned using Allelic Discrimination Software (Applied Biosystems SDS Software v1.7a). Ten percent quality control samples were included and each analysis included no DNA template controls. Laboratory personnel were blinded to quality control and case-control status.

Colorectal cancer and adenoma risk was considered in relation to APC genotypes. APC Asp1822Val was categorized into the three genotypes (Asp/Asp, Asp/Val, and Val/Val) and as noncarriers versus carriers of the variant allele (Asp/Asp and Asp/Val + Val/Val). APC Gly2502Ser was categorized into wild-type and heterozygote genotypes (Gly/Gly and Gly/Ser). For analysis of the main effect of genotype, logistic regression was used to compute relative risks (RR) and 95% confidence intervals (95% CI), control for potentially confounding variables, and to test for gene-environment interactions. Although the data sets were initially based on matched cases and controls, the stratified analyses required unconditional analysis. For colorectal adenoma, separate analyses for men and women using unconditional logistic regression were controlled for age, indication for endoscopy, year of endoscopy, family history of colorectal cancer, smoking history, aspirin use, body mass index (BMI), postmenopausal hormone (PMH) use among postmenopausal women (NHS only), physical activity, and intake of red meat, folate, and alcohol. Men and women in the adenoma groups were combined for gene-environment interactions and were controlled for age, sex, and family history of colorectal cancer. Analyses involving NHS cancer cases and controls were controlled for age, family history of colorectal cancer, smoking history, aspirin use, BMI, postmenopausal status and PMH use, physical activity, and intake of red meat, folic acid, and alcohol. Analyses involving PHS cancer cases and controls were matched on age and smoking status and controlled for aspirin use, BMI, physical activity, and intake of red meat, vitamins, and alcohol.

The effects of lifestyle and dietary exposures on colon cancer and adenoma risk were tested in conjunction with APC genotype. In the NHS adenoma group, APC Asp1822Val genotypes were analyzed in combination with pack-years of smoking (0, <25, and ≥25 years), grams per day of alcohol (<7.5, 7.5-15.0, >15.0 g/d), micrograms per day of folate including supplements (≤310, 311-507, ≥508 g/d) based on median distributions in the control population, aspirin intake (<1/wk, 1-6/wk, ≥7/wk), family history of colorectal cancer (yes/no), PMH use among postmenopausal women (never/past and current), BMI (<22, 22 to <25, 25 to <30, and ≥30), energy expenditure in metabolic equivalent tasks measured in hours per week (<1.5, 1.5-5.9, 5.9-11.9, 11.9-19.9, >19.9 h/wk), intake of red meat categorized into two groups and intake of total fat, saturated fat, monounsaturated fat, polyunsaturated fat, transfatty acids, and total calories categorized into three groups based on median distributions in the control population. In the HPFS adenoma group, APC Asp1822Val genotypes were analyzed in combination with pack-years of smoking (0,<25, ≥25 years), grams per day of alcohol (<5, 5-30, >30 g/d), micrograms per day of folate including supplements (≤338, 339-496, ≥497 g/d) based on median distributions in the control population, regular aspirin intake (yes/no), family history of colorectal cancer (yes/no), BMI, and metabolic equivalent tasks categorized into four groups and intake of red meat, total fat, saturated fat, monounsaturated fat, polyunsaturated fat, transfatty acids, and total calories categorized into three groups based on distributions in the control population. In the NHS colon cancer group, APC Asp1822Val genotypes were analyzed in combination with pack-years of smoking before age 30 (0, ≤10, >10 years), grams per day of alcohol (<7.5 and ≥7.5 g/d), micrograms per day of folate including supplements (≤310, 311-507, ≥508 g/d) based on median distributions in the control population, aspirin intake (<1/wk, 1-6/wk, and ≥7/wk), family history of colorectal cancer (yes/no), PMH use among postmenopausal women (never/past, current), BMI (<22, 22 to <25, 25 to <30, and ≥30), metabolic equivalent tasks categorized into three groups and intake of red meat, total fat, saturated fat, monounsaturated fat, polyunsaturated fat, transfatty acids, and total calories categorized into two groups based on median distributions in the control population. In the PHS, APC Asp1822Val genotypes were analyzed in combination with smoking (never, past, and current), alcohol intake in drinks per day (1 and ≥1), vitamin intake (never, past, and current), aspirin assignment (yes/no), and BMI (<22, 22 to <25, 25 to <30, and ≥30). All dietary fat analyses were additionally controlled for intake of total calories. In the PHS, baseline dietary information did not permit analysis of dietary fat and other individual nutrients. All exposure information for the NHS adenoma was updated from 1990, 1992, 1994 to1996 and cancer was updated from 1990, 1992, 1994, 1996 to 1998 using the latest update of the information before diagnosis of disease (except for smoking history, which was cumulatively updated to the latest time period before diagnosis of disease). Exposure information for the HPFS and PHS was collected in at baseline in 1986 and 1982, respectively. The P for interaction was based on the Wald test for the cross-product term in a model containing the main effects of genotype and exposure variable. All Ps are based on two-sided tests. All statistical analyses were done using the SAS 6.0 statistical package (SAS Institute, Cary, NC).

We first examined whether the risk factors for colorectal cancer and adenoma were similar in the cases and controls that provided blood samples compared with previous observations for each cohort. For the HPFS nested case-control study, the risk patterns observed for adenoma cases in this sample were largely similar to those reported for the entire cohort, as previously reported (22). Family history of colorectal cancer, smoking, and alcohol intake of >30 g/d was associated with increased risk of colorectal adenoma whereas regular aspirin use was associated with a lower risk. For NHS nested case-control study, the risk patterns observed for colorectal cancer and adenoma cases in this sample were also largely similar to those reported for the entire cohort (21, 23, 24). Smoking ≥25 pack-years was associated with an increased risk of colorectal adenoma whereas regular aspirin use was associated with a decreased risk. In addition, the risk patterns observed for cancer cases were largely similar to those reported for the entire cohort (21, 23, 24). Smoking >10 pack-years before age 30, alcohol intake of ≥7.5 g/d, and past use of PMH was associated with an increased risk of colorectal cancer, whereas current PMH was inversely associated with risk of colorectal cancer. For the PHS nested case-control study, the risk factors analyzed were not associated with an increased risk of colorectal cancer.

The APC 1822Val allele frequency ranged from 22% to 25% in the NHS, HPFS, and PHS control populations and the genotype distributions were in Hardy-Weinberg equilibrium and genotypes were similar in frequency to those reported in Slattery et al. (ref. 17; Table 1). The APC 2502Ser allele frequency was 2% in the NHS, HPFS, and PHS control populations and the genotype distributions were in Hardy-Weinberg equilibrium (Table 2). The 2% 2502Ser allele frequency seen in this study was lower than the 10% Caucasian allele frequency reported by SNP500, which is based on a sample of 23 individuals. The APC Asp1822Val (Table 1) and Gly2502Ser (Table 2) polymorphisms were not associated with risk of either colorectal cancer or adenoma in both men and women.

We assessed the associations of dietary and lifestyle factors with risk of colorectal cancer or adenoma stratified by APC Asp1822Val genotype (Tables 3,Table 4,Table 5-6). Smoking, alcohol intake, and family history of colorectal cancer were associated with increased risk of adenoma in men and women for all genotypes, whereas regular aspirin use was inversely related to risk of colorectal cancer in women and colorectal adenoma in women and men regardless of genotype (Table 3). Similar relationships were not observed between this APC genotype and smoking, alcohol intake, family history, and aspirin use for risk of colorectal cancer for both PHS and NHS (data available from http://www.channing.harvard.edu/nhs/publications/tables/APC_2004_Tables.pdf). An assessment of dietary fat intake and APC Asp1822Val genotype did not reveal any patterns of association in relation to risk of colorectal cancer and colorectal adenoma (Tables 4-5). Current use of PMH among postmenopausal women was inversely related to risk of colorectal cancer and colorectal adenoma (Table 6) and a statistically significant (P = 0.03) gene-environment interaction was found for colorectal cancer. Current PMH use among postmenopausal women was associated with a statistically significant reduction in risk of colorectal cancer among carriers of one or two copies of the APC 1822Val allele (RR, 0.46; 95% CI, 0.24-0.88). Past or never PMH use among postmenopausal women was related to a statistically nonsignificant increase in risk of colorectal cancer among APC 1822Val allele carriers (RR, 1.38; 95% CI, 0.82-2.32). An inverse association between negative family history of colorectal cancer and risk of colorectal adenoma among men and women carrying two variant APC 1822Val alleles (RR, 0.65; 95% CI, 0.42-1.01), was not observed among individuals with positive family history (RR, 1.75; 95% CI, 0.46-6.68), although the interaction was not significant (Table 3). Other dietary and lifestyle factors assessed, including BMI, metabolic equivalent tasks, and intake of total calories, vitamins, folate, and red meat intake did not show any differences across APC Asp1822Val genotypes in relation to risk of colorectal cancer and colorectal adenoma (data available from http://www.channing.harvard.edu/nhs/publications/tables/APC_2004_Tables.pdf). The low APC 2502Ser allele prevalence did not permit the investigation of gene-environment interactions.

In the present study, we observed no association between the APC Asp1822Val and Gly2502Ser polymorphisms and risk of either colorectal cancer or adenoma in men and women. Evaluating the influence of the APC Asp1822Val polymorphism on the relationship of dietary and lifestyle factors with colorectal cancer risk revealed a statistically significant interaction with PMH use in postmenopausal women. Current PMH use by postmenopausal women was also inversely associated with adenoma risk for all APC Asp1822Val genotypes. Among men and women carrying two variant APC 1822Val alleles, having no family history of colorectal cancer was inversely associated with risk of colorectal adenoma. Men and women having a family history of colorectal cancer were not at reduced risk of colorectal adenoma for any APC Asp1822Val genotype. For other known risk factors, no statistically significant gene-environment interactions were observed for colorectal adenoma in this study, some consistent trends were observed in both women and men. Smoking and alcohol intake were associated with increased risk of adenoma in men and women, regardless of genotype. In addition, regular aspirin intake was inversely associated with colorectal adenoma in men and women.

Slattery et al. (17) observed that a low-fat diet was associated with a significantly decreased risk of colon cancer among homozygous carriers of the 1822Val allele and not among the wild-type or heterozygous carriers. This relationship was consistent for total, saturated, and unsaturated fat. In our study, a similar relationship between fat intake and Asp1822Val genotype was not observed for either colorectal cancer or adenoma. In their study of colorectal cancer, Slattery et al. (17) had greater power to detect gene-environment interactions due to the larger sample size of their study (1,590 cases and 1,945 controls), although our exposure information on diet was collected prospectively and thus less prone to recall bias. Slattery et al. (17) suggest that the consistent relationship seen in their study between total, saturated, and unsaturated fat intake and the APC codon 1822 substitution may indicate that the 1822Val homozygous variant has functional significance. The APC gene codes for a large protein (>8 kb) that is multifunctional and structurally complex. Within the APC gene are two amino acid repeat motifs responsible for binding and down-regulating β-catenin (25). The β-catenin binding site is located between codons 1020 and 1169 and consists of three imperfect 15 amino acid repeats. The region that coordinates β-catenin down-regulation is located between codons 1324 and 2075 and contains seven 20 amino acid repeats. Most acquired APC mutations occur between codons 1286 and 1513, called the mutation cluster region, and have some overlap with the β-catenin down-regulation repeat region. Mutations within the mutation cluster region usually result in truncated proteins, which lead to the severe familial adenomatous polyposis phenotype. The Asp1822Val polymorphism lies between the third and fourth amino acid repeats within the β-catenin down-regulation domain. Efficient β-catenin down-regulation requires a minimum of three of the seven binding repeats (coinciding with the 3′ limit of the mutation cluster region), suggesting that the Asp1822Val polymorphism may not have an appreciable affect on β-catenin degradation (25, 26). There is no known functional consequence of the Asp1822Val substitution and future analyses should examine the potential function of this polymorphism in apoptosis, cellular adhesion, and chromosome segregation.

Current use of PMH among postmenopausal women was inversely related to risk of colorectal cancer and adenoma regardless of APC Asp1822Val genotype; however, the strongest inverse association was observed for women carrying two copies of the 1822Val allele. These results are consistent with previous observational studies demonstrating that hormone replacement therapy reduces risk of colorectal adenoma and carcinoma (27). Several mechanisms have been proposed to explain the protective effect of PMH use on colorectal cancer. Exogenous estrogens decrease secondary bile acid production in the colonic epithelium, which reduces chronic irritation of the colonic mucosa (28, 29) and may protect against colorectal cancer (30). Methylation-associated inactivation of ER resulted in deregulated growth of the colonic epithelium, suggesting that the ER gene may act as a tumor suppressor (31). Methylation progresses with the natural decline of circulating estrogens in the aging colonic mucosa, resulting in an increased risk for neoplastic growth with age. In vitro, estrogens reduced ER gene methylation and inhibited cell proliferation (31), suggestive of another mechanism by which PMHs reduce the risk of colon cancer. Estrogens have also been shown to increase the expression of vitamin D receptors in a variety of tissues (28, 29). Vitamin D and several of its analogues are known to inhibit neoplastic growth and promote differentiation in colonic epithelial cells (28, 29) and increased vitamin D receptor activity may be an additional mechanism by which estrogens protect against colon carcinogenesis. Whereas estrogens act directly on colorectal cancer cells through ERβ (32, 33) and decreased ERβ expression in tumor mucosa compared with normal colon cells suggests its involvement in mediating the effects of estrogens on colon tissue (34), there is no clear association between aberrant β-catenin nuclear accumulation in the colonic epithelium and estrogen. Mutations in Wnt signaling genes have been reported in endometrial cancers (35) and abnormality in the Wnt signaling pathway, resulting in the nuclear accumulation of β-catenin, is a feature of the estrogen-related subset of endometrial carcinomas (36).

In summary, smoking, alcohol intake, and family history of colorectal cancer were associated with increased risk of adenoma in men and women. Regular aspirin intake was inversely associated with colorectal adenoma in men and women. Current PMH use by postmenopausal women was also inversely associated with colorectal cancer and adenoma risk. In addition, the effect of current PMH use on risk of colorectal cancer is modified by APC Asp1822Val genotype. Postmenopausal women carrying one or two copies of the variant APC 1822Val allele who were current PMH users were at a significantly reduced risk of colorectal cancer whereas women who were never or past users of PMH had a nonsignificant increase in risk of colorectal cancer. We observed no indication of interaction between the APC Asp1822Val polymorphism and previously hypothesized dietary factors, including dietary fat.

We thank the participants of the NHS, the PHS, and the HPFS for their cooperation and participation and Hardeep Ranu and Patrice Soule for technical assistance.