Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Feb 10.
Published in final edited form as: Nutr Cancer. 2012 Sep 11;64(7):919–928. doi: 10.1080/01635581.2012.711418

Vitamin D intake is negatively associated with promoter methylation of the Wnt antagonist gene DKK1 in a large group of colorectal cancer patients

James B Rawson 1,2, Zhouyu Sun 3, Elizabeth Dicks 3, Darshana Daftary 4, Patrick S Parfrey 3, Roger C Green 3, Steven Gallinger 1,2,4,5, John R McLaughlin 2,4,6, Peizhong P Wang 3, Julia A Knight 2,6, Bharati Bapat 1,2,7
PMCID: PMC4323165  NIHMSID: NIHMS658916  PMID: 22966878

Abstract

Diet and lifestyle help mediate colorectal cancer (CRC) risk but the molecular events that mediate these effects are poorly characterized. Several dietary and lifestyle factors can modulate DNA methylation suggesting that they may influence CRC risk through epigenetic regulation of cancer-related genes. The Wnt regulatory genes DKK1 and Wnt5a are important contributors to colonic carcinogenesis and are often silenced by promoter hypermethylation in CRC; however, the dietary contributions to these events have not been explored. To investigate the link between dietary/lifestyle factors and epigenetic regulation of these Wnt signaling genes, we assessed promoter methylation of these genes in a large cohort of Canadian CRC patients from Ontario (n=549) and Newfoundland (n=443) and examined associations to dietary/lifestyle factors implicated in CRC risk and/or DNA methylation including intake of vitamins, fats, cholesterol, fibre, and alcohol as well as BMI and smoking status. Several factors were associated with methylation status including alcohol intake, BMI, and cigarette smoking. Most significantly however, dietary vitamin D intake was strongly negatively associated with DKK1 methylation in Newfoundland (p=0.001) and a similar trend was observed in Ontario. These results suggest that vitamin D and other dietary/lifestyle factors may alter CRC risk by mediating extracellular Wnt inhibition.

1. Introduction

Over the past several decades, epidemiological studies have underscored the importance of dietary and lifestyle factors on colorectal cancer (CRC) risk. Among the supported environmental risk factors for CRC are obesity, smoking, consumption of red meats and fats, and low intake of fruits and vegetables. Despite potential roles for these factors in CRC onset, little is known about the specific molecular events that may mediate these changes in tumor biology.

A broad mechanism with emerging importance by which these and other factors induce their risk-modifying effects is via DNA methylation. Methylation of CpG dinucleotides within gene promoters is a fundamental method of transcriptional silencing in eukaryotes and altered genomic methylation status is common to many cancers. In fact, aberrant promoter hypermethylation is a particularly well-established hallmark of CRC, defining the CpG island methylator phenotype (CIMP) and driving onset of most microsatellite unstable tumours through silencing of the critical DNA mistmatch repair gene MLH1. Several dietary and lifestyle risk factors for CRC are known to influence DNA methylation levels within the colon including folate, alcohol, and certain B vitamins [1]; however, the specific risk genes affected remain poorly characterized.

Among the most important genes in colorectal cancer are those encoding mediators of Wnt signaling. Canonical Wnt signaling mediates homeostatic proliferation in colonic crypts and hyperactivation of this pathway is the major driver of colorectal tumour growth. Diet-induced changes in Wnt activity within the colon have been observed during both development [2] and tumourigenesis [3] but the involvement of promoter methylation within Wnt genes in causing these important changes in Wnt activity has not been explored.

The DKK1 and Wnt5a genes encode pivotal Wnt mediators that are silenced in early CRC by promoter hypermethylation [4, 5]. DKK1 encodes a canonical Wnt suppressor that has strong anti-proliferative and pro-apoptotic effects in CRC cell lines [6, 7] and dramatically reduces tumour burden in mouse xenografts [7]. Wnt5a encodes a quintessential non-canonical Wnt ligand that has emerging roles in both colonic tumour-suppression and oncogenesis (reviewed in [8]). Expression levels of DKK1 and Wnt5a have been associated with disease outcome in several cancers including CRC [9-11] bringing into question whether epigenetic silencing of these genes may be responsible for changes in CRC risk. To investigate the potential role of DKK1 and Wnt5a promoter methylation in mediating diet-associated changes in CRC risk, we quantified promoter methylation levels of these genes in two large populations of CRC patients from Canada and examined their associations to dietary and lifestyle factors implicated in CRC risk and/or colonic methylation.

2. Materials and Methods

2.1 Study Participants

Participants included in this study were accrued through the population-based Ontario Familial Colorectal Cancer Registry (OFCCR) and Newfoundland Familial Colorectal Cancer Registry (NFCCR). Patient accrual, data collection, and biospecimen collection procedures at OFCCR have been previously detailed [12]. Briefly, Ontario residents diagnosed with pathology-confirmed CRC between the ages of 20 and 74 from 1997 to 2000 were eligible for recruitment. Apparent cases of familial adenomatous polyposis were excluded and recruitment was preferential for cases with positive familial risk based on patient response to a family history questionnaire. 1004 probands satisfying these conditions had blood and/or tissue biospecimens available at the time of study accrual. Due to the high prevalence of Caucasians (92.5%) and to limit the possible effects of population stratification, non-white patients and those with unknown or mixed ethnic background were excluded from this study. Of these 921 (91.7%) probands, 561 (55.8%) had tumour samples available for methylation analysis. Several probands were excluded during methylation analysis due to poor quality DNA (see Methods: MethyLight), leading to 545 (54.2%) final cases. An overview of patient clinicopathological features is provided in Table 1.

Table 1.

Overview of clinicopathological features among colorectal cancer cases from Ontario and Newfoundland.

No. of Cases (%)
Ontario Newfoundland
Cases of primary colorectal carcinoma 549 443
Mean Age (± Std.Dev.) 60.6 ± 8.8 61.3 ± 9.4
Sex
 Female 263 (47.9) 172 (38.8)
 Male 286 (52.1) 271 (61.2)
Tumor Locationa
 Proximal 209 (38.1) 175 (39.5)
 Distal 332 (60.5) 260 (58.7)
 Unavailable/Overlapping 8 (1.5) 8 (1.8)
TNM Stage
 1 96 (17.5) 88 (19.9)
 2 211 (38.4) 177 (40.0)
 3 168 (30.6) 141 (31.8)
 4 30 (5.5) 37 (8.4)
 Unavailable 44 (8.0) 0 (0.0)
Histological Grade
 Low 43 (7.8) 66 (15.1)
 Moderate 392 (71.4) 331 (74.7)
 High 64 (11.7) 38 (8.6)
 Unavailable 50 (9.1) 7 (1.6)
Histological Typeb
 Non-mucinous 446 (81.2) 395 (89.2)
 Mucinous 64 (11.7) 48 (10.8)
 Unavailable 39 (7.1) 0 (0.0)
MSI Status
 Stable 385 (70.1) 368 (83.1)
 Low 71 (12.9) 20 (4.5)
 High 88 (16.0) 52 (11.7)
 Unavailable 5 (0.9) 3 (0.7)
DKK1 Methylation
 Unmethylated 477 (87.0) 382 (86.2)
 Methylated 71 (13.0) 61 (13.8)
Wnt5a Methylation
 Unmethylated 438 (79.8) 361 (81.5)
 Methylated 107 (19.5) 75 (16.9)
 Unavailable 4 (0.7) 7 (1.6)
Open in a new tab
a

Locations proximal or distal to the splenic flexure. Tumors at the splenic flexure are included in the “Proximal” category.

b

Presence of any mucin within the tumor stroma is defined as “Mucinous”.

Patient accrual at NFCCR was similar to that of OFCCR and has also been described in detail [13]. Briefly, cases were accrued over a slightly later recruitment period (1999 to 2003) and included all incident CRC patients irrespective of family history. In addition, for deceased patients, NFCCR accepted proxy consent from living family members, which led to more frequent inclusion of late-stage patients. Of the 747 probands recruited by NFCCR at the time of study accrual, 721 (96.5%) probands were Caucasian (northern European origin) and had tumour samples available for methylation analysis. Methylation analysis was successfully performed on 687 (92.0%) cases after removal of poor quality samples. 253 (35.8%) probands were recruited through proxy consent of family members and removed from the analysis to increase the accuracy of dietary reporting.

2.2 Accrual, Selection, and Definition of Dietary and Lifestyle Variables

Dietary and lifestyle data for each population was abstracted through patient responses to two mailed questionnaires: personal history (PHQ) and food frequency (FFQ). The FFQ in Ontario (ON) was developed by the Epidemiology Program at the Cancer Research Center of Hawaii and has been previously described and validated [14]. The FFQ in Newfoundland (NFLD) was developed at Memorial University, St. John's, Newfoundland [15]. Several probands from each population did not respond to the FFQ (ON=35, NFLD=59) or PHQ (ON=39, NFLD=5). Dietary and lifestyle variables were selected for abstraction based on a literature search of risk factors for CRC and/or colonic methylation. Dietary variables examined were: vitamins B2, B6, B12, and D, folate, isoflavonoids, fibre, fat, saturated fat, and cholesterol. Lifestyle variables examined were: daily intake of alcohol (averaged over drinkable lifetime), history of cigarette smoking, and body mass index (BMI). Dietary fat and saturated intakes were reported in absolute (daily intake in weight units) and/or relative (daily intake as percent of total daily calories) format.

Dietary and lifestyle habits in Ontario and Newfoundland are considerably discordant. To remove possible bias due to outliers and to improve homogeneity between populations, all nutrient intake measurements were categorized into population-specific rounded quartiles except for vitamin D intake from supplements, which was categorized into users and non-users due to a limited number of users. Daily intake of alcohol (previously defined [16]) was grouped by median intake after segregation of never-drinkers, who were assigned an intake value of zero. Smoking status (previously defined [16]) was grouped into ever- and never-smokers. BMI was examined at age 20 (“BMI Age 20”) and at one year prior to cancer diagnosis (“BMI current”) and categorized according to World Health Organization guidelines with collapsed groups where appropriate to ensure statistical power. For sex-stratified and stage-stratified analyses of vitamin D intake, the upper three quartiles of intake were collapsed into a single comparison group to ensure statistical power.

2.3 Methylation Analysis of Candidate Genes

DKK1 and Wnt5a were selected as candidate genes based on their strong functional effects on Wnt signaling and previous associations with prognostically-relevant CRC subtypes. Methylation analysis of two other Wnt genes, SFRP1 and WIF-1, was also performed; however, these genes were not considered for analysis due to a limited number of unmethylated individuals in Ontario (SFRP1=28; WIF-1= 64) and Newfoundland (SFRP1=21; WIF-1= 39).

Methylation analysis was performed on DNA extracted from paraffin-embedded tumours using semi-quantitative MethyLight assay, as previously described [17]. Primers and probes were designed to amplify the promoter CpG island regions of DKK1 and Wnt5a [7, 18]. The extent of promoter methylation was calculated as percent methylated reference (PMR) as follows: PMR =[gene/reference]sample / [gene/reference]CpGenome × 100% [19]; where CpGenome DNA is a fully methylated positive control. Alu-C4 was used as the internal reference control amplicon, as recommended to maximize assay robustness [20]. For cases with multiple DNA samples due to synchronous baseline tumours (ON: n=34, NFLD: n=16), PMR values were averaged. Cases were categorized as “methylated” or “unmethylated” based on a tumour-specific methylation cutoff of PMR ≥ 10% as previously validated [5, 21]. Samples with PMR values ±2% of this cutoff value were reanalyzed and PMR values were averaged to more accurately define methylation status.

Validation was also performed on randomly selected samples to assess batch effects. PMR readings fluctuated <10% among different batches. During both screening and validation, samples in which the Alu-C4 internal control reaction reached threshold fluorescence above 22 cycles (C(t) > 22) were considered to be poor quality, retreated with sodium bisulphite, and subsequently reanalyzed to ensure robust amplification. Samples that remained poor quality in both independent analyses were removed from the analysis (ON: n=16; NFLD: n=34).

2.4 Statistical Analysis

Associations of dietary and lifestyle exposure variables to methylation status of DKK1 or Wnt5a (outcome variables categorized as “methylated” versus “unmethylated”, as detailed above) were assessed by unconditional logistic regression. Logistic regression models were adjusted for age, sex, and total energy intake. Adjustment for other clinical and pathological features was not performed in order to avoid overadjustment for factors not expected to impact the direct biological relationship between diet and methylation status. Two-sided p < 0.05 was considered statistically significant. Exposure variables were carefully selected from evidence of effects on colorectal cancer risk and/or gene methylation status; as a result, statistical adjustment for multiple hypothesis testing was not performed. All analyses were performed using SPSS v16.0 (SPSS Inc.).

3. Results

We examined the associations of dietary factors (Table 2) and lifestyle factors (Table 3) with methylation status of DKK1 and Wnt5a in 549 CRCs from Ontario and 443 CRCs from Newfoundland. There were no significant differences in the mean age at diagnosis, tumor stage, location, histological type, or methylation status of DKK1 or Wnt5a between these populations. However, in Newfoundland compared to Ontario there were a higher proportion of males (p=0.005), lower proportion of MSI tumor (p<10−6), and lower proportion of high grade tumors (p=0.002).

Table 2.

Associations between dietary factors and methylation status of DKK1 and Wnt5a in colorectal carcinomas from Ontario and Newfoundland. Models were adjusted for age, sex, and total energy intake.

Ontario Newfoundland
DKK1 Wnt5a DKK1 Wnt5a
Variable Total (%) OR (95% CI) p OR (95% CI) p Variable Total (%) OR (95% CI) p OR (95% CI) p
Vitamin D (ug) Vitamin D (ug)
 Overall 0.25 0.76  Overall 0.001 0.05
  0-4 132 (25.7) 1.00 (Referent) 1.00 (Referent)   0-4 101 (26.4) 1.00 (Referent) 1.00 (Referent)
  >4-7 131 (25.5) 0.47 (0.21, 1.01) 1.28 (0.69, 2.38)   >4-6 82 (21.5) 0.20 (0.07, 0.54) 0.25 (0.09, 0.67)
  >7-12 109 (21.2) 0.63 (0.29, 1.39) 0.93 (0.47, 1.86)   >6-11 103 (27.0) 0.27 (0.11, 0.64) 0.68 (0.31, 1.48)
  >12 142 (27.6) 0.61 (0.30, 1.23) 1.00 (0.53, 1.86)   >11 96 (25.1) 0.21 (0.08, 0.54) 0.56 (0.25, 1.27)
 Dietary 0.81 0.98  Dietary 0.001 0.11
  0-3 104 (20.2) 1.00 (Referent) 1.00 (Referent)   0-4 114 (29.7) 1.00 (Referent) 1.00 (Referent)
  >3-5 150 (29.2) 1.31 (0.59, 2.93) 0.92 (0.48, 1.75)   >4-6 100 (26.0) 0.27 (0.11, 0.65) 0.43 (0.19, 0.98)
  >5-7 127 (24.7) 1.24 (0.53, 2.90) 0.98 (0.50, 1.90)   >6-8 75 (19.5) 0.47 (0.20, 1.08) 0.96 (0.44, 2.09)
  >7 133 (25.9) 1.54 (0.65, 3.65) 0.88 (0.43, 1.79)   >8 95 (24.7) 0.13 (0.04, 0.39) 0.51 (0.22, 1.21)
 Supplements 0.79 0.90  Supplements 0.11 0.55
  No 346 (67.3) 1.00 (Referent) 1.00 (Referent)   No 316 (82.3) 1.00 (Referent) 1.00 (Referent)
  Yes 168 (32.7) 0.80 (0.45, 1.41) 0.97 (0.61, 1.55)   Yes 68 (17.7) 0.45 (0.17, 1.20) 0.79 (0.38, 1.68)
Vitamin B2 (mg) 0.76 0.76 Vitamin B2 (mg) 0.27 0.11
  0-2.2 148 (28.8) 1.00 (Referent) 1.00 (Referent)   0-1.9 88 (23.1) 1.00 (Referent) 1.00 (Referent)
  >2.2-3.0 115 (22.4) 0.92 (0.44, 1.93) 0.78 (0.42, 1.48)   >1.9-2.6 97 (25.5) 0.60 (0.23, 1.57) 0.35 (0.14, 0.89)
  >3.0-4.0 129 (25.1) 0.67 (0.31, 1.43) 0.83 (0.45, 1.54)   >2.6-3.7 99 (26.0) 1.28 (0.50, 3.23) 0.90 (0.39, 2.08)
  >4.0 122 (23.7) 0.81 (0.36, 1.84) 0.69 (0.34, 1.40)   >3.7 97 (25.5) 0.65 (0.21, 2.01) 0.64 (0.24, 1.71)
Vitamin B6 (mg) 0.71 0.79 Vitamin B6 (mg) 0.66 0.37
  0-2.0 133 (25.9) 1.00 (Referent) 1.00 (Referent)   0-1.7 97 (25.4) 1.00 (Referent) 1.00 (Referent)
  >2.0-2.7 116 (22.6) 1.32 (0.62, 2.81) 1.03 (0.55, 1.94)   >1.7-2.2 87 (22.8) 0.65 (0.25, 1.65) 0.48 (0.19, 1.24)
  >2.7-4.2 134 (26.1) 0.88 (0.40, 1.96) 0.81 (0.42, 1.56)   >2.2-3.2 104 (27.2) 0.96 (0.38, 2.43) 1.02 (0.44, 2.38)
  >4.2 131 (25.5) 0.90 (0.40, 2.01) 0.77 (0.39, 1.52)   >3.2 94 (24.6) 0.65 (0.22, 1.88) 0.95 (0.38, 2.40)
Vitamin B12 (ug) 0.69 0.41 Vitamin B12 (ug) 0.21 0.91
  0-5 124 (24.1) 1.00 (Referent) 1.00 (Referent)   0-5 92 (24.1) 1.00 (Referent) 1.00 (Referent)
  >5-8 133 (25.9) 1.41 (0.66, 3.00) 1.48 (0.79, 2.77)   >5-8 95 (24.9) 0.43 (0.18, 1.04) 0.84 (0.37, 1.89)
  >8-16 126 (24.5) 1.09 (0.48, 2.47) 0.94 (0.47, 1.90)   >8-13 99 (25.9) 0.51 (0.21, 1.24) 0.74 (0.32, 1.74)
  >16 131 (25.5) 0.91 (0.42, 2.00) 0.94 (0.49, 1.81)   >13 96 (25.1) 0.47 (0.19, 1.21) 0.90 (0.38, 2.11)
Folate (ug) 0.74 0.61 Folate (ug) 0.65 0.73
  0-550 130 (25.3) 1.00 (Referent) 1.00 (Referent)   0-300 84 (22.0) 1.00 (Referent) 1.00 (Referent)
  >550-840 129 (25.1) 0.94 (0.45, 1.97) 0.69 (0.36, 1.30)   >300-400 101 (26.5) 0.94 (0.36, 2.44) 1.26 (0.50, 3.20)
  >840-1240 127 (24.7) 0.68 (0.31, 1.49) 0.78 (0.41, 1.48)   >400-525 101 (26.5) 1.38 (0.52, 3.69) 1.68 (0.65, 4.30)
  >1240 128 (24.9) 0.73 (0.32, 1.65) 0.66 (0.33, 1.31)   >525 95 (24.9) 0.83 (0.23, 2.99) 1.36 (0.43, 4.29)
Isoflavonoids (mg) 0.64 0.15 Isoflavonoids (mg) 0.21 0.13
  0-0.3 113 (22.0) 1.00 (Referent) 1.00 (Referent)   0-0.1 106 (27.6) 1.00 (Referent) 1.00 (Referent)
  >0.3-0.5 123 (23.9) 1.03 (0.49, 2.16) 0.48 (0.25, 0.93)   >0.1-0.25 91 (23.7) 1.18 (0.44, 3.16) 2.00 (0.82, 4.90)
  >0.5-0.8 136 (26.5) 0.91 (0.43, 1.92) 0.61 (0.32, 1.15)   >0.25-0.32 88 (22.9) 1.99 (0.80, 4.92) 2.52 (1.05, 6.07)
  >0.8 142 (27.6) 0.62 (0.27, 1.47) 0.77 (0.40, 1.49)   >0.32 99 (25.8) 2.23 (0.93, 5.33) 2.65 (1.13, 6.20)
Fibre (g) 0.37 0.18 Fibre (g) 0.53 0.62
  0-18 124 (24.1) 1.00 (Referent) 1.00 (Referent)   0-15 101 (26.3) 1.00 (Referent) 1.00 (Referent)
  >18-24 131 (25.5) 1.50 (0.71, 3.13) 0.75 (0.41, 1.39)   >15-20 81 (21.1) 1.52 (0.62, 3.69) 1.22 (0.51, 2.92)
  >24-32 126 (24.5) 0.83 (0.36, 1.91) 0.46 (0.23, 0.92)   >20-27 99 (25.8) 0.80 (0.30, 2.12) 1.02 (0.42, 2.92)
  >32 133 (25.9) 0.84 (0.33, 2.12) 0.68 (0.32, 1.43)   >27 103 (26.8) 1.15 (0.39, 3.38) 1.52 (0.58, 4.00)
Fat Fat
 Daily Intake (g) 0.31 0.39  Daily Intake (g) 0.13 0.41
  0-60 146 (28.4) 1.00 (Referent) 1.00 (Referent)   0-60 103 (26.8) 1.00 (Referent) 1.00 (Referent)
  >60-80 129 (25.1) 1.65 (0.77, 3.55) 1.71 (0.92, 3.16)   >60-80 90 (23.4) 0.33 (0.12, 0.92) 0.51 (0.21, 1.26
  >80-100 102 (19.8) 1.87 (0.76, 4.58) 1.34 (0.64, 2.78)   >80-110 100 (26.0) 0.89 (0.37, 2.13) 0.93 (0.40, 2.15)
  >100 137 (26.7) 2.93 (0.95, 9.02) 1.36 (0.55, 3.36)   >110 91 (23.7) 0.48 (0.12, 1.92) 1.06 (0.32, 3.46)
 % Daily Calories 0.30 0.25  % Daily Calories 0.81 0.40
  0-28 142 (27.6) 1.00 (Referent) 1.00 (Referent)   0-27 93 (24.2) 1.00 (Referent) 1.00 (Referent)
  >28-32 108 (21.0) 1.78 (0.82, 3.89) 1.39 (0.75, 2.60)   >27-31 91 (23.7) 0.81 (0.35, 1.87) 0.88 (0.42, 1.86)
  >32-36 134 (26.1) 1.79 (0.85, 3.76) 0.88 (0.47, 1.65)   >31-35 108 (28.1) 0.78 (0.35, 1.75) 0.54 (0.25, 1.16)
  >36 130 (25.3) 1.94 (0.90, 4.18) 1.54 (0.84, 2.95)   >35 92 (24.0) 0.64 (0.26, 1.56) 0.67 (0.30, 1.49)
Saturated Fat Saturated Fat
 Daily Intake (g) 0.48 0.16  Daily Intake (g) 0.72 0.86
  0-20 146 (28.4) 1.00 (Referent) 1.00 (Referent)   0-20 98 (25.5) 1.00 (Referent) 1.00 (Referent)
  >20-25 104 (20.2) 1.65 (0.75, 3.62) 0.99 (0.50, 1.98)   >20-28 108 (28.1) 0.64 (0.28, 1.48) 0.87 (0.40, 1.91)
  >25-35 147 (28.6) 1.68 (0.77, 3.68) 1.84 (0.98, 3.46)   >28-38 81 (21.1) 0.64 (0.24, 1.72) 0.80 (0.32, 1.98)
  >35 117 (22.8) 2.30 (0.73, 7.28) 1.37 (0.53, 3.54)   >38 97 (25.3) 0.57 (0.16, 2.00) 0.60 (0.18, 1.93)
 % Daily Calories 0.32 0.03  % Daily Calories 0.63 0.12
  0-9.0 130 (25.3) 1.00 (Referent) 1.00 (Referent)   0-9 96 (25.0) 1.00 (Referent) 1.00 (Referent)
  >9.0-10.5 114 (22.2) 1.17 (0.52, 2.65) 0.96 (0.47, 1.98)   >9-11 113 (29.4) 0.70 (0.32, 1.54) 0.71 (0.36, 1.43)
  >10.5-12.0 135 (26.3) 1.43 (0.67, 3.03) 2.05 (1.10, 3.84)   >11-13 98 (25.5) 0.60 (0.26, 1.38) 0.40 (0.18, 0.89)
  >12.0 135 (26.3) 1.94 (0.93, 4.06) 1.96 (1.04, 3.70)   >13 77 (20.1) 0.66 (0.26, 1.65) 0.50 (0.21, 1.17)
Cholesterol (mg) 0.25 0.30 Cholesterol (mg) 0.50 0.61
  0-180 111 (21.6) 1.00 (Referent) 1.00 (Referent)   0-225 102 (26.6) 1.00 (Referent) 1.00 (Referent)
  >180-260 151 (29.4) 1.94 (0.92, 4.09) 1.31 (0.69, 2.48)   >225-300 93 (24.2) 0.72 (0.32, 1.63) 0.71 (0.33, 1.56)
  >260-340 130 (25.3) 1.34 (0.57, 3.13) 1.61 (0.81, 3.19)   >300-400 81 (21.1) 0.60 (0.24, 1.52) 0.60 (0.25, 1.42)
  >340 122 (23.7) 1.05 (0.35, 3.12) 0.91 (0.37, 2.23)   >400 108 (28.1) 0.43 (0.14, 1.30) 0.55 (0.21, 1.46)
Open in a new tab

Table 3.

Associations between lifestyle factors and methylation status of DKK1 and Wnt5a in colorectal carcinomas from Ontario and Newfoundland. Models were adjusted for age, sex, and total energy intake.

Ontario Newfoundland
DKK1 Wnt5a DKK1 Wnt5a
Variable Total (%) OR (95% CI) p OR (95% CI) p Variable Total (%) OR (95% CI) p OR (95% CI) p
Alcohol Alcohol
 Daily Intake (g) 0.16 0.02  Daily Intake (g) 0.98 0.89
  0 136 (26.4) 1.00 (Referent) 1.00 (Referent)   0 206 (51.6) 1.00 (Referent) 1.00 (Referent)
  >0-11 205 (39.7) 0.59 (0.31, 1.10) 0.45 (0.26, 0.77)   >0-9 74 (18.5) 0.94 (0.41, 2.17) 0.98 (0.46, 2.12)
  >11 175 (33.9) 0.56 (0.27, 1.14) 0.60 (0.33, 1.09)   >9 119 (29.8) 0.92 (0.42, 2.01) 0.85 (0.42, 1.70)
 % Daily Calories 0.34 0.85  % Daily Calories 0.85 0.84
  0-0.5 188 (36.6) 1.00 (Referent) 1.00 (Referent)   0 143 (37.2) 1.00 (Referent) 1.00 (Referent)
  >0.5-4.5 169 (32.9) 1.12 (0.61, 2.04) 1.00 (0.59, 1.72)   >0-3 124 (32.3) 0.91 (0.45, 1.82) 0.98 (0.52, 1.87)
  >4.5 157 (30.5) 0.65 (0.31, 1.33) 1.16 (0.66, 2.05)   >3 117 (30.5) 0.79 (0.35, 1.80) 0.81 (0.39, 1.69)
Smoking 0.83 0.04 Smoking 0.88 0.12
  No 190 (37.0) 1.00 (Referent) 1.00 (Referent)   No 115 (28.0) 1.00 (Referent) 1.00 (Referent)
  Yes 324 (63.0) 1.10 (0.63, 1.92) 0.62 (0.39, 0.98)   Yes 296 (72.0) 0.95 (0.49, 1.84) 0.63 (0.35, 1.13)
BMI BMI
 Current 0.05 0.62  Current 0.45 0.26
  <25 175 (36.2) 1.00 (Referent) 1.00 (Referent)   ≤5 118 (29.6) 1.00 (Referent) 1.00 (Referent)
  >25-30 192 (39.7) 0.44 (0.23, 0.87) 0.81 (0.47, 1.39)   >25-30 162 (40.6) 0.87 (0.43, 1.79) 0.57 (0.29, 1.12)
  >30 117 (24.2) 0.88 (0.45, 1.75) 1.07 (0.59, 1.94)   >30 119 (29.8) 0.59 (0.25, 1.36) 0.78 (0.38, 1.60)
 Age 20 0.76 0.02  Age 20 0.94 0.99
  <25 354 (74.4) 1.00 (Referent) 1.00 (Referent)   <25 307 (79.3) 1.00 (Referent) 1.00 (Referent)
  >25 122 (25.6) 1.11 (0.57, 2.17) 1.93 (1.12, 3.31)   >25 80 (20.7) 1.03 (0.44, 2.41) 1.01 (0.46, 2.19)
Open in a new tab

Several dietary and lifestyle factors were significantly associated with promoter methylation status of DKK1 or Wnt5a. Most significantly, among Newfoundland CRCs there was a strong negative association DKK1 methylation and overall vitamin D intake (p=0.001). Upon further stratification by intake source, a similar negative relationship was found between DKK1 methylation and vitamin D intake from dietary sources (p=0.001) and from vitamin supplements, although the latter did not reach statistical significance (p=0.11). Wnt5a methylation was also negatively associated with overall vitamin D intake in this population (p=0.05) and the directionality of this relationship was maintained when examining vitamin D intake from dietary sources (p=0.11). To further investigate this borderline relationship, individuals with low vitamin D intake (first quartile) were compared to individuals with high vitamin D intake (other quartiles). Individuals with high intake were significantly less likely to have Wnt5a methylation (p=0.04, OR=0.49 [0.25, 0.96]). In Ontario, vitamin D intake was not significantly associated with methylation of either gene, although directionality of this relationship was similar when examining overall/dietary vitamin D intake and DKK1 methylation as well as supplemental vitamin D intake and DKK1 methylation but not dietary vitamin D intake.

To investigate whether gender differences between Newfoundland and Ontario may explain the discordant relationships between vitamin D intake and DKK1 methylation in these populations, a sex-stratified analysis was conducted. In Newfoundland, the inverse association between overall vitamin D intake and DKK1 methylation were both maintained for males alone (p=0.03) and for females alone (p=0.001) although this relationship seemed stronger in females (Supplementary Table 1). Similar results were seen for dietary vitamin D intake and DKK1 methylation. In Ontario, vitamin D intake remained unassociated with DKK1 methylation in both sexes. Another possible reason behind the discordant vitamin D findings could be differential accumulation of stage-specific alterations that change the ability of vitamin D to influence DKK1 methylation. To examine this hypothesis, individuals in each population were stratified into early tumor stage (I/II) and late tumor stage (III/IV) and an analysis was conducted. In Newfoundland, overall vitamin D intake remained negatively associated with DKK1 methylation in early stage (p=0.003) and late stage tumors (p=0.006) (Supplementary Table 2). In Ontario, overall vitamin D intake became negatively associated with DKK1 methylation in early stage tumors (p=0.05) but not in late stage tumors (p=0.87).

Other factors that showed a modest association with methylation status of DKK1 or Wnt5a in Ontario include: negative associations between DKK1 methylation and current BMI (p=0.05), Wnt5a methylation and cigarette smoking (p=0.04), and Wnt5a methylation and daily alcohol intake (p=0.02). Positive associations were found between Wnt5a methylation and BMI at age 20 (p=0.02) as well as Wnt5a methylation and percent of daily calories from saturated fat (p=0.03). None of these comparisons reached statistical significance in Newfoundland; however, the negative directionality between Wnt5a methylation and smoking was consistent. No associations were found between methylation status and intake of B vitamins, cholesterol, fibre, folate, isoflavonoids, or other measurements of fat in either population.

4. Discussion

In order to fully understand the roles of diet and lifestyle in cancer risk and progression and to improve related guidelines for cancer prevention and treatment, there is a need to identify specific, functionally-relevant molecular events that may be responsible for these outcomes. Vitamin D is an emerging protective factor for CRC [22] that mediates expression of numerous cancer genes through conventional binding of the internalized vitamin D receptor to gene promoters or through epigenetic mediation of these gene promoters including histone modification and gene methylation [23]. We found that vitamin D intake, both overall and from dietary sources, was a strong negative indicator of DKK1 promoter methylation in Newfoundland CRCs. Due to the silencing effect of promoter methylation on DKK1 gene expression [6], this negative association suggests that DKK1 is preferentially expressed in CRC patients with high vitamin D intake. Vitamin D has previously been shown to induce DKK1 expression in CRC cells and in corresponding mouse xenografts through a mechanism independent of canonical promoter recruitment [24]. Consequently, our results suggest that the mechanism involved may be loss of DKK1 promoter methylation and that this mechanism may also be found in primary CRC. Due to the tumour suppressive functions of DKK1, this epigenetic event may represent an important molecular link between vitamin D and CRC protection. Interestingly however, DKK1 methylation was not significantly associated with vitamin D intake derived from vitamin supplements. The significance of this result is unknown due to the methods of data ascertainment used in this study but should be further explored to determine whether certain types of vitamin D may selectively influence methylation at the DKK1 promoter.

Vitamin D and DKK1 methylation were not significantly associated in Ontario CRCs despite this relationship maintaining similar directionality in both populations. Ontario and Newfoundland are clinically and genetically heterogeneous CRC populations that vary with respect to disease incidence and mortality [Health Canada, 25], family history [13], and tumour MSI status [26]. It is likely that components of the genetic and environmental variability that underlie these population-specific differences may also contribute to the discordant relationship between vitamin D and DKK1 methylation found in this study. Further population stratification suggested that this discordance is not due to gender differences but may be related to tumor stage in Ontario since early stage tumors but not late stage tumors exhibited the negative association between DKK1 methylation and overall vitamin D intake. Perhaps a specific alteration occurs during tumor development in Ontario that makes these individuals selectively fail to respond to vitamin D intake. Further genetic analysis of vitamin D metabolizing enzymes in this population may help to address this observation.

Differences in sunlight exposure may also influence the effect of measured vitamin D intake on methylation due to the vital role of UV-B radiation in catalyzing endogenous synthesis of the majority of this nutrient in individuals with adequate sunlight exposure. Since UV exposure in Ontario is abundant and much higher than in Newfoundland [27], vitamin D intake from dietary sources may be of limited biological significance among Ontario CRCs. This may help explain the limited significance of the relationship between measured vitamin D intake and DKK1 methylation in Ontario. Analysis of circulating 1,25-dihydroxyvitamin D levels may help to more directly scrutinize the relationship between vitamin D and DKK1 methylation in Ontario CRCs.

Other dietary and lifestyle factors for CRC that were weakly associated with methylation status of DKK1 or Wnt5a in Ontario included cigarette smoking and alcohol as negative indicators while BMI showed opposing associations with methylation status. The significance of these results varied by gene and method of data representation (e.g. alcohol intake measured in grams versus percent of daily calories). Nevertheless, the negative relationship between smoking and methylation is consistent with smoking as a negative indicator of Wnt antagonist promoter methylation in lung cancer [28] but inconsistent with the general methylation-promoting effects of smoking including its strong association with CIMP CRCs [29] – where the majority of DKK1 and Wnt5a methylation is found [4, 5]. Furthermore, it is unclear the extent to which this negative relationship between smoking and methylation may reflect increased expression of these genes since tobacco seems to have alternate mechanisms of silencing Wnt antagonist genes [30]. Additional functional analyses examining the effects of smoking on methylation and expression of Wnt antagonist genes will be required to assess the possible contribution of these epigenetic events to smoking-related changes in CRC risk.

Like smoking, BMI has been associated with CIMP CRC; however, BMI and several other measures of body mass have also been linked to risk of non-CIMP CRCs [31]. The relationship between body mass and methylation within CRC is therefore less certain than that of smoking. Conversely, alcohol has been linked to hypomethylated CRC via LINE-1 status [32] but is not correlated with CIMP status [33]. Consequently, it is uncertain whether the observed associations of BMI and alcohol with Wnt antagonist methylation may be related to their broader relationships to overall CRC methylation. Further scrutiny of these factors as specific mediators of Wnt antagonist gene methylation and expression will be necessary to determine whether these genes play a significant role in BMI-related or alcohol-related changes in CRC risk.

This is the first study to investigate the dietary and lifestyle contributions to promoter methylation of important Wnt regulatory genes in CRC, offering insight into the functionally-relevant epigenetic events through which these factors may influence CRC risk and outcome. We found a strong negative association between vitamin D intake and DKK1 methylation, providing a possible epigenetic link between vitamin D and CRC protection. Smoking, alcohol intake, and BMI were also modestly linked to methylation status; however, due to the lack of consistent findings between our populations, additional validation in other CRC cohorts is warranted. Nevertheless, this study paves the way for further examination of Wnt antagonist gene methylation in order to assess the contribution of these events to dietary and lifestyle-associated changes in CRC risk.

Supplementary Material

SupplementaryTable 1. Sex-stratified associations between vitamin D intake and DKK1 methylation in colorectal carcinomas from Ontario and Newfoundland.

Supplementary Table 2. Stage-stratified associations between vitamin D intake and DKK1 methylation in colorectal carcinomas from Ontario and Newfoundland.

Acknowledgments

This work was undertaken at the Department of Laboratory Medicine and Pathobiology at the University of Toronto, Toronto, ON, Canada and was conducted at the Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, ON, Canada. We sincerely thank the investigators, staff, and participants of the Colon Cancer Family Registry for their dedicated contributions leading to this work.

This work was supported by the National Cancer Institute, National Institutes of Health under RFA # CA-95-011 and through cooperative agreements with members of the Colon Cancer Family Registry and P.I.s. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the CFRs, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the CFR. This work was also supported by a Team Grant from the Canadian Institutes of Health Research (CTP-79845 awarded to JRM, BB, JAK, SSG, RCG, and PSP) and by funding within the Colon Cancer Familial Registry awarded to the Ontario Registry for Studies of Familial Colorectal Cancer (Grant no. U01 CA074783, Principle Investigator: SSG). JBR was supported by graduate studentships from the Team in Interdisciplinary Research on Colorectal Cancer funded by the Canadian Institutes of Health Research and by graduate studentships from the University of Toronto (Open Fellowships, administered by the Department of Laboratory Medicine and Pathobiology).

Abbreviations

CRC

colorectal cancer

CIMP

CpG island methylator phenotype

DKK1

dickkopft 1 homolog

MLH1

mutL homolog 1

MSI

microsatellite instability

NFCCR

Newfoundland Familial Colorectal Cancer

OFCCR

Ontario Familial Colorectal Cancer Registry

PMR

percent methylated reference

Wnt5a

wingless-related MMTV integration site 5A

References

  • 1.Arasaradnam RP, Commane DM, Bradburn D, Mathers JC. A review of dietary factors and its influence on DNA methylation in colorectal carcinogenesis. Epigenetics. 2008;3(4):193–8. doi: 10.4161/epi.3.4.6508. [DOI] [PubMed] [Google Scholar]
  • 2.Gray JD, Nakouzi G, Slowinska-Castaldo B, Dazard JE, Rao JS, et al. Functional interactions between the LRP6 WNT co-receptor and folate supplementation. Hum Mol Genet. 2010;19(23):4560–72. doi: 10.1093/hmg/ddq384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Liu Z, Choi SW, Crott JW, Keyes MK, Jang H, et al. Mild depletion of dietary folate combined with other B vitamins alters multiple components of the Wnt pathway in mouse colon. J Nutr. 2007;137(12):2701–8. doi: 10.1093/jn/137.12.2701. [DOI] [PubMed] [Google Scholar]
  • 4.Rawson JB, Mrkonjic M, Daftary D, Dicks E, Buchanan DD, et al. Promoter methylation of Wnt5a is associated with microsatellite instability and BRAF V600E mutation in two large populations of colorectal cancer patients. Br J Cancer. 2011 doi: 10.1038/bjc.2011.165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Rawson JB, Manno M, Mrkonjic M, Daftary D, Dicks E, et al. Promoter methylation of Wnt antagonists DKK1 and SFRP1 is associated with opposing tumor subtypes in two large populations of colorectal cancer patients. Carcinogenesis. 2011;32(5):741–7. doi: 10.1093/carcin/bgr020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Sato H, Suzuki H, Toyota M, Nojima M, Maruyama R, et al. Frequent epigenetic inactivation of DICKKOPF family genes in human gastrointestinal tumors. Carcinogenesis. 2007;28(12):2459–66. doi: 10.1093/carcin/bgm178. [DOI] [PubMed] [Google Scholar]
  • 7.Aguilera O, Fraga MF, Ballestar E, Paz MF, Herranz M, et al. Epigenetic inactivation of the Wnt antagonist DICKKOPF-1 (DKK-1) gene in human colorectal cancer. Oncogene. 2006;25(29):4116–21. doi: 10.1038/sj.onc.1209439. [DOI] [PubMed] [Google Scholar]
  • 8.Wang Y. Wnt/Planar cell polarity signaling: a new paradigm for cancer therapy. Mol Cancer Ther. 2009;8(8):2103–9. doi: 10.1158/1535-7163.MCT-09-0282. [DOI] [PubMed] [Google Scholar]
  • 9.Dejmek J, Dejmek A, Safholm A, Sjolander A, Andersson T. Wnt-5a protein expression in primary dukes B colon cancers identifies a subgroup of patients with good prognosis. Cancer Res. 2005;65(20):9142–6. doi: 10.1158/0008-5472.CAN-05-1710. [DOI] [PubMed] [Google Scholar]
  • 10.Weeraratna AT, Jiang Y, Hostetter G, Rosenblatt K, Duray P, et al. Wnt5a signaling directly affects cell motility and invasion of metastatic melanoma. Cancer Cell. 2002;1(3):279–88. doi: 10.1016/s1535-6108(02)00045-4. [DOI] [PubMed] [Google Scholar]
  • 11.Suzuki R, Onizuka M, Kojima M, Shimada M, Fukagawa S, et al. Preferential hypermethylation of the Dickkopf-1 promoter in core-binding factor leukaemia. Br J Haematol. 2007;138(5):624–31. doi: 10.1111/j.1365-2141.2007.06702.x. [DOI] [PubMed] [Google Scholar]
  • 12.Newcomb PA, Baron J, Cotterchio M, Gallinger S, Grove J, et al. Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol Biomarkers Prev. 2007;16(11):2331–43. doi: 10.1158/1055-9965.EPI-07-0648. [DOI] [PubMed] [Google Scholar]
  • 13.Green RC, Green JS, Buehler SK, Robb JD, Daftary D, et al. Very high incidence of familial colorectal cancer in Newfoundland: a comparison with Ontario and 13 other population‐based studies. Fam Cancer. 2007;6(1):53–62. doi: 10.1007/s10689-006-9104-x. [DOI] [PubMed] [Google Scholar]
  • 14.Cotterchio M, Boucher BA, Kreiger N, Mills CA, Thompson LU. Dietary phytoestrogen intake--lignans and isoflavones--and breast cancer risk (Canada) Cancer Causes Control. 2008;19(3):259–72. doi: 10.1007/s10552-007-9089-2. [DOI] [PubMed] [Google Scholar]
  • 15.Squires J, Roebothan B, Buehler S, Sun Z, Cotterchio M, et al. Pickled meat consumption and colorectal cancer (CRC): a case-control study in Newfoundland and Labrador, Canada. Cancer Causes Control. 2010;21(9):1513–21. doi: 10.1007/s10552-010-9580-z. [DOI] [PubMed] [Google Scholar]
  • 16.Poynter JN, Haile RW, Siegmund KD, Campbell PT, Figueiredo JC, et al. Associations between smoking, alcohol consumption, and colorectal cancer, overall and by tumor microsatellite instability status. Cancer Epidemiol Biomarkers Prev. 2009;18(10):2745–50. doi: 10.1158/1055-9965.EPI-09-0517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, et al. MethyLight: a high-throughput assay to measure DNA methylation. Nucleic Acids Res. 2000;28(8):E32. doi: 10.1093/nar/28.8.e32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ying J, Li H, Yu J, Ng KM, Poon FF, et al. WNT5A exhibits tumor-suppressive activity through antagonizing the Wnt/beta-catenin signaling, and is frequently methylated in colorectal cancer. Clin Cancer Res. 2008;14(1):55–61. doi: 10.1158/1078-0432.CCR-07-1644. [DOI] [PubMed] [Google Scholar]
  • 19.Weisenberger DJ, Siegmund KD, Campan M, Young J, Long TI, et al. CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat Genet. 2006;38(7):787–93. doi: 10.1038/ng1834. [DOI] [PubMed] [Google Scholar]
  • 20.Weisenberger DJ, Campan M, Long TI, Kim M, Woods C, et al. Analysis of repetitive element DNA methylation by MethyLight. Nucleic Acids Res. 2005;33(21):6823–36. doi: 10.1093/nar/gki987. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ogino S, Kawasaki T, Brahmandam M, Cantor M, Kirkner GJ, et al. Precision and performance characteristics of bisulfite conversion and real-time PCR (MethyLight) for quantitative DNA methylation analysis. J Mol Diagn. 2006;8(2):209–17. doi: 10.2353/jmoldx.2006.050135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Giovannucci E. Epidemiology of vitamin D and colorectal cancer: casual or causal link? J Steroid Biochem Mol Biol. 2010;121(1-2):349–54. doi: 10.1016/j.jsbmb.2010.03.085. [DOI] [PubMed] [Google Scholar]
  • 23.Stefanska B, Karlic H, Varga F, Fabianowska-Majewska K, Haslberger AG. Epigenetic mechanisms in anti-cancer actions of bioactive food components-the implications in cancer prevention. Br J Pharmacol. 2012 doi: 10.1111/j.1476-5381.2012.02002.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Aguilera O, Pena C, Garcia JM, Larriba MJ, Ordonez-Moran P, et al. The Wnt antagonist DICKKOPF-1 gene is induced by 1alpha,25-dihydroxyvitamin D3 associated to the differentiation of human colon cancer cells. Carcinogenesis. 2007;28(9):1877–84. doi: 10.1093/carcin/bgm094. [DOI] [PubMed] [Google Scholar]
  • 25.HealthCanada. Canadian Cancer Statistics. 2010 [Google Scholar]
  • 26.Raptis S, Mrkonjic M, Green RC, Pethe VV, Monga N, et al. MLH1 -93G>A promoter polymorphism and the risk of microsatellite-unstable colorectal cancer. J Natl Cancer Inst. 2007;99(6):463–74. doi: 10.1093/jnci/djk095. [DOI] [PubMed] [Google Scholar]
  • 27.Fioletov VE, McArthur LJ, Mathews TW, Marrett L. Estimated ultraviolet exposure levels for a sufficient vitamin D status in North America. J Photochem Photobiol B. 2010;100(2):57–66. doi: 10.1016/j.jphotobiol.2010.05.002. [DOI] [PubMed] [Google Scholar]
  • 28.Suzuki M, Shigematsu H, Nakajima T, Kubo R, Motohashi S, et al. Synchronous alterations of Wnt and epidermal growth factor receptor signaling pathways through aberrant methylation and mutation in non small cell lung cancer. Clin Cancer Res. 2007;13(20):6087–92. doi: 10.1158/1078-0432.CCR-07-0591. [DOI] [PubMed] [Google Scholar]
  • 29.Samowitz WS, Albertsen H, Sweeney C, Herrick J, Caan BJ, et al. Association of smoking, CpG island methylator phenotype, and V600E BRAF mutations in colon cancer. J Natl Cancer Inst. 2006;98(23):1731–8. doi: 10.1093/jnci/djj468. [DOI] [PubMed] [Google Scholar]
  • 30.Hussain M, Rao M, Humphries AE, Hong JA, Liu F, et al. Tobacco smoke induces polycomb-mediated repression of Dickkopf-1 in lung cancer cells. Cancer Res. 2009;69(8):3570–8. doi: 10.1158/0008-5472.CAN-08-2807. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Hughes LA, Simons CC, van den Brandt PA, Goldbohm RA, van Engeland M, et al. Body size and colorectal cancer risk after 16.3 years of follow-up: an analysis from the Netherlands Cohort Study. Am J Epidemiol. 2011;174(10):1127–39. doi: 10.1093/aje/kwr247. [DOI] [PubMed] [Google Scholar]
  • 32.Schernhammer ES, Giovannucci E, Kawasaki T, Rosner B, Fuchs CS, et al. Dietary folate, alcohol and B vitamins in relation to LINE-1 hypomethylation in colon cancer. Gut. 2010;59(6):794–9. doi: 10.1136/gut.2009.183707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Schernhammer ES, Giovannucci E, Baba Y, Fuchs CS, Ogino S. B vitamins, methionine and alcohol intake and risk of colon cancer in relation to BRAF mutation and CpG island methylator phenotype (CIMP) PLoS One. 2011;6(6):e21102. doi: 10.1371/journal.pone.0021102. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SupplementaryTable 1. Sex-stratified associations between vitamin D intake and DKK1 methylation in colorectal carcinomas from Ontario and Newfoundland.

Supplementary Table 2. Stage-stratified associations between vitamin D intake and DKK1 methylation in colorectal carcinomas from Ontario and Newfoundland.

NIHMS658916-supplement-supplement_1.pdf (179.3KB, pdf)

RESOURCES