Associations between vitamin D, vitamin D receptor gene and the androgen receptor gene with colon and rectal cancer



The transcriptional activity of the vitamin D receptor (VDR) gene is regulated, at least in part, by the androgen receptor (AR) gene. We evaluate how the number of polyglutamine (CAG) repeats of the AR gene influence colorectal cancer in conjunction with vitamin D, sunshine exposure and VDR. Studies of colon (1,580 cases and 1,968 controls) and rectal (797 cases and 1,016 controls) cancer were used. Vitamin D intake and average hours of sunshine exposure interacted with AR genotype in men. Men with low vitamin D intake or low levels of sunshine exposure who had 23+ CAG repeats of the AR gene had the greatest risk of colon cancer. ORs for men with 23 or more CAG repeats of the AR gene and in the lowest tertile of vitamin D intake or sunshine exposure were 1.71 (95% CI 1.14, 2.56) and 1.51 (95% CI 1.09, 2.09). Men with high levels of sunshine exposure were at reduced risk of developing rectal cancer if they had 23 or more CAG repeats (OR 0.62 95% CI 0.39, 0.97) than if they had fewer than 23 CAG repeats. The FF genotype of the Fok1 VDR gene was associated with reduced risk of colon cancer among women with any allele of 23+ CAG repeats (OR 0.62 95% CI 0.44, 0.88), whereas men with the LL/bb VDR genotypes were at reduced risk of rectal cancer if they also had 23+ CAG repeats (OR 0.71 95% CI 0.48, 1.05) relative to men with fewer than 23 CAG repeats of the AR gene. These data provide support for the role of vitamin D and sunshine exposure in the etiology of colorectal cancer and suggest that AR gene may modulate the association. © 2006 Wiley-Liss, Inc.

Both 1,25-dihydroxyvitamin D3 [1,25-(OH2)D3], the active metabolite of vitamin D, and androgens are essential for the regulation of cell growth and differentiation in several tissues, including colorectal tissue.1, 2, 3, 4 The vitamin D receptor (VDR) mediates biological responses to [1,25-(OH2)D3].5 Both the VDR and the androgen receptor (AR) are nuclear receptors that are ligand-dependent transcriptional regulatory proteins. AR signaling and VDR action have been shown to be linked.6, 7 In prostate cancer cells, over expression of AR can lead to the suppression of VDR transactivation8; other studies suggest that androgen plays a role in the antiproliferative effects of [1,25-(OH2)D3] and that the AR and the VDR are both involved in ovarian cancer.9 Our understanding of how androgens and vitamin D influence the development of colorectal cancer is evolving, although the association with vitamin D and sunshine exposure with colorectal cancer has been of interest for many years.

Vitamin D has been associated inconsistently with colorectal cancer.10 Differences in relative risk could stem from many sources, including sunshine, a major source of vitamin D, not being considered as part of the risk factor equation. Several investigators have studied sunshine exposure independently to better understand the associations between vitamin D and colorectal cancer.11, 12, 13 Polymorphisms of the VDR gene have also been studied in conjunction with colorectal cancer and adenomas, since it may play a key role in vitamin D signaling, which could further influence colorectal cancer risk.14, 15, 16, 17, 18, 19 Polymorphisms of the VDR gene that have been studied include a poly A repeat at the 3′ untranslated region (3′ UTR) of the gene, 2 polymorphisms at intron 8 (Bsm I and Apa I) and 1 in exon 9 (Taq I) that are in linkage-disequilibrium with each other.17, 20 These polymorphisms, either alone or in combination, have been associated with increased mRNA expression of the VDR gene and increased serum levels of 1,25-dihydroxy vitamin D.21, 22 At the start site of the gene, a polymorphism detected with a Fok I digest has also been studied and has been shown to be not in linkage disequilibrium with the other variants.17 A study conducted in human fibroblast cell lines found higher vitamin D hormone activity for the F allele than for the f allele.23 One study showed that the ff genotype of the Fok I polymorphism increased risk of colorectal cancer,18 whereas another showed that it was protective.24 The Bsm1 B and short poly-A alleles have been reported to be protective against the development of colonic adenomas and prostate cancer.15, 16, 19, 25

The gene encoding the AR is located on the X chromosome. The AR gene contains 2 polymorphic tri-nucleotide repeat segments that encode polyglutamine (CAG) and polyglycine (GGC). The CAG repeat has been studied most extensively and has been shown to range from 6 to 39 repeats, with some studies showing that the number of repeats is inversely related to prostate cancer risk. Fewer CAG repeats result in higher transcription of AR mRNA.26 One study showed that in the presence of relatively long CAG repeats, decline in age-related serum androgen levels did not occur in men.27 Westberg et al.28 found that women with fewer CAG repeats in the AR gene had higher serum androgen levels and lower leutenizing hormone levels than women with more repeats. There is evidence that there is an acquired reduction in the number of CAG repeats in the AR gene in colon tumors.29

Although limited data exist, it is possible that the AR may influence vitamin D activity and interact with vitamin D level as well as the VDR to alter risk of colorectal cancer. The purpose of this study is to examine associations between the CAG repeat of the AR gene with vitamin D intake, sunshine exposure, the Bsm1, polyA, Fok1 VDR polymorphisms and incidence of colorectal cancer.


Study populations

All aspects of this study were approved by the University of Utah Institutional Review Board as well as Institutional Review Boards at the Kaiser Permanente Medical Care Program of Northern California (KPMCP) and the University of Minnesota. Participants were from the KPMCP, the state of Utah and the Twin City Metropolitan area of Minnesota (colon cancer study only). Study procedures for determining case and control eligibility were the same in all centers, since the study was funded to enable evaluation of interactions. All eligible cases within these defined areas were identified and recruited using the same study criteria. The colon study included cases of first primary colon cancer (ICD-O 2nd edition codes 18.0, 18.2–18.9) diagnosed between October 1, 1991 and September 30, 1994. Cases from the rectal cancer study were diagnosed with a first primary tumor in the rectosigmoid junction or rectum, and were identified between May 1997 and May 2001. The rectal cancer study included cases and controls from Utah and KPMCP. Case eligibility was determined by the Surveillance Epidemiology and End Results (SEER) Cancer Registries in Northern California and in Utah and the Minnesota Cancer Surveillance System (colon cancer cases only). Eligibility included being between 30 and 79 years of age at the time of diagnosis, English speaking, mentally competent to complete the interview, no previous history of colorectal cancer30 and no known (as indicated on the pathology report) familial adenomatous polyposis, ulcerative colitis or Crohn's disease. Of cases contacted, 83% participated at KPMCP, 76% in Utah and 67% in Minnesota. For the rectal cancer study, the cooperation rates were 75.4% of cases from KPCMP and 69.7% of cases from Utah.

Controls were frequency matched to cases by sex and by 5-year age groups and were selected using a sampling frame that would target the same population from where the cases came. At the KPMCP, controls were randomly selected from membership lists. In Utah, controls 65 years and older were randomly selected from lists provided by the Centers for Medicare and Medicaid Services (formerly HCFA), and controls younger than 65 were randomly selected from driver's license lists. In Minnesota, controls were randomly selected from driver's license lists. Of controls contacted for the colon cancer study, 73% participated at KPMCP, 53% participated from Minnesota and 69% participated from Utah. For the rectal cancer study, cooperation rates were 69.9% for KPMCP and 67.2% for Utah.

Data collection

Trained and certified interviewers collected diet and lifestyle data.31, 32 All interviewers were originally trained centrally in Utah. All quality control procedures were the same and quality control was done by the same person in Utah providing feedback to interviewers in other locations. Each interview was audio-taped, and in between a random sample of tapes for each interviewer was reviewed centrally to monitor data quality. The referent year for the study was the calendar year approximately 2 years prior to date of diagnosis (cases) or selection (controls). Information was collected on demographic factors such as age, sex and study center; diet, physical activity, exposure to sunshine, body size and other lifestyle factors, including medical, family and reproductive history. Vitamin D intake was obtained from a detailed diet history questionnaire that included over 300 possible food items. Participants also were asked to recall the average number of hours spent outdoors in the sunshine for each season of the referent year, to correspond to information collected on diet and physical activity. The average for the 4 seasons of number of hours spent outdoors during the year was used as an indication of individual sunshine exposure.


DNA was extracted from peripheral blood leukocytes. For quality control, controls representing all known polymorphic variants and blanks were included in each 96-well tray. All genotypes were scored by 2 individuals with any discrepancies being scored by a third reader.

The AR-CAG repeat genotyping was performed according the methods of Westberg et al.28 with minor modifications. Briefly, 20 ng of genomic DNA was PCR-amplified using oligonucleotide primers AR-F2-5′-GTTTCTGTGGGGCCTCTACGATGG-3′and AR-R2-5′-GTTTCTGCGCGAAGTGATCCAGAA-3′ (HEX labeled). The 14-μl PCR reaction contained 0.2 mM dNTPs, 1.5 mM Mg2+ and 0.4 units of Taq polymerase (Perkin–Elmer). The PCR reactions were initially denatured at 94°C for 1 min, then subjected to 35 cycles consisting of 15 sec at 94°C, 15 sec at 57°C, followed by 30 sec at 72°C. A final extension of 5 min at 72°C was performed. The Hex labeled products were then analyzed on an ABI 3700 automated sequencer. Allele sizes were converted into CAG repeat copy number for analysis, by sizing alleles that had been sequenced.

The intron 8 Bsm I VDR polymorphism (rs154410) was amplified from genomic DNA and digested as described previously.17 Presence of the restriction site was scored as allele “b,” and absence of the restriction site was scored as allele “B”. For the 3′UTR poly-A repeat, genomic DNA was amplified and allele length determined as described previously.17 Repeat length was classified as short (14–17 repeats) or S, or long (18–22 repeats) or L, as described by Ingles et al.25 The FokI initiation codon polymorphism (rs10735810) was determined through PCR amplification on 20 ng of genomic DNA, in the presence of 10% DMSO by an initial denaturation at 95°C for 2 min, followed by 35 cycles of 95°C for 10 sec, 60°C for 30 sec and 72°C for 30 sec. A final 5 min extension at 72°C was performed. Restriction digests of the PCR products using Fok1 were performed according to the manufacturer's specifications (New England Biolabs). Presence of the restriction site was scored as allele “f,” absence of the restriction site was scored as allele “F”.

Statistical analysis

SAS statistical package, version 8.2, was used to conduct the analyses. Analyses included evaluating the distribution of the alleles and genotypes in the population, and studying the joint effect of the AR gene with the VDR gene, calcium, vitamin D and sunshine exposure. The AR gene was categorized as being short if there were fewer than 23 CAG repeats and long if there were 23 or more CAG repeats. This cut-point was selected based on previous cut-points reported in the literature, and because it appeared to be the cut-point that discriminated relative risk for colon cancer. The B and b Bsm I alleles are highly associated with the short (S) and long (L) poly-A alleles of the VDR gene. Since we did not have data on both polymorphisms for all cases, we combined the polymorphism results and report the genotypes as BB or SS, bb or LL or other (most of which are Bb/SL). Separate analyses of Bsm I and poly-A polymorphisms showed no additional associations (data not shown). Among those that had data for both genotypes, concordance between the LL and bb alleles was 97%, between the SS and BB was 95% and between the LS and Bb was 96%. The Fok1 genotype was evaluated as any f allele and the FF genotype.

Multiple logistic regression models were used to determine associations, with results reported as odds ratio (OR) and 95% confidence intervals (CI). In these models, potentially confounding factors that were considered were age, race, BMI, physical activity, dietary composition, energy intake and cigarette smoking amount. All adjustment variables were treated as continuous variables in the model unless they were categorical variables such as race. Cut-points for vitamin D and average hours of sunshine exposure during the referent year were based on the sex-specific distribution of the control population. Tertiles of vitamin D and sunshine were used to obtain more stable risk estimates when assessing interactions. Several tests for interaction were performed. The relative excess risk from interaction (RERI) and corresponding 95% CI was calculated to provide insight into differences that might be expected on an additive scale of interaction, as described by Hosmer and Lemoshow33; multiplicative interaction was tested using the cross product of the 2 variables of interest in the logistic model. Analyses were conducted separately for colon and rectal cancer.


The age, sex and race/ethnic distribution of the study population is shown in Table I. Rectal cancer cases were slightly younger than colon cancer cases; however, the majority of both study groups were over 60 years of age. The majority of the population self-reported as being non-Hispanic white. There were more male cases than female cases. The most common number of CAG repeats of the AR was 21 repeats; few people had fewer than 19 CAG repeats or more than 25 CAG repeats of the AR gene. Overall, having 23 or more CAG repeats increased relative risk of colon cancer among men (OR 1.3 95% CI 1.1, 1.5) and decreased risk of colon cancer among women (OR 0.8, 95% CI 0.7–1.0), as determined from multiple logistic regression models that adjusted for age and race. The p value for interaction between CAG repeats of AR gene and sex was <0.01. Because of the significant interaction between sex and AR and risk of colon cancer, sex-specific analyses are presented.

Table I. Description of Study Population
  Colon cancer studyRectal cancer study
  • 1

    AIAN, American Indian or Alaska Native.

  • 2

    Indicates allele frequency.

  N (%)N (%)N (%)N (%)
Age<4023 (1.5)40 (2.0)20 (2.5)22 (2.2)
40–49102 (6.5)126 (6.4)100 (12.6)110 (10.8)
50–59297 (18.8)330 (16.8)209 (26.2)256 (25.2)
60–69549 (34.8)682 (34.7)270 (33.9)349 (34.4)
70–79609 (38.5)790 (40.1)198 (24.8)279 (27.5)
EthnicityWhite caucasian1446 (91.7)1835 (93.3)653 (83.0)861 (85.7)
Hispanic61 (3.9)78 (4.0)54 (6.9)71 (7.1)
African American70 (4.4)53 (2.7)30 (3.8)40 (4.0)
Asian0040 (5.1)31 (3.1)
AIAN10010 (1.3)2 (0.2)
SexMale887 (56.1)1051 (53.4)469 (58.9)572 (56.3)
Female693 (43.9)917 (46.6)328 (41.1)444 (43.7)
AR29–1661 (2.7)77 (2.7)30 (2.7)21 (1.5)
 (CAG repeats)1733 (1.5)41 (1.5)19 (1.7)26 (1.8)
18130 (5.8)184 (6.5)62 (5.6)76 (5.3)
19228 (10.2)292 (10.3)121 (11.0)152 (10.6)
20251 (11.3)372 (13.1)104 (9.5)147 (10.2)
21408 (18.3)467 (16.5)179 (16.3)259 (18.0)
22255 (11.4)310 (11.0)121 (11.0)159 (11.1)
23228 (10.2)316 (11.2)148 (13.5)148 (10.3)
24235 (10.5)280 (9.9)119 (10.8)182 (12.7)
25165 (7.4)211 (7.5)89 (8.1)118 (8.2)
26103 (4.6)137 (4.8)49 (4.5)64 (4.5)
2756 (2.5)49 (1.7)22 (2.0)37 (2.6)
2841 (1.8)54 (1.9)17 (1.6)19 (1.3)
29–4137 (1.7)42 (1.5)19 (1.7)23 (2.0)
  Mean (SE)Mean (SE)Mean (SE)Mean (SE)
Vitamin D intake (mcg)/day6.8 (0.12)6.8 (0.10)8.0 (0.19)7.9 (0.17)
Hours sunshine exposure/week17.2 (0.35)17.8 (0.31)17.4 (0.50)17.5 (0.45)

We observed a significant interaction between Vitamin D intake and sunshine exposure and the AR gene and risk of colon cancer among men (Table II). In both instances, those at greater risk were men with low vitamin D intake or low levels of sunshine exposure who had 23 or more CAG repeats of the AR gene. There were no significant associations among women.

Table II. Associations between AR, Vitamin D and Sunshine Exposure with Colon Cancer
AR CAG repeatsMenWomen
(<23)(23+)(<23)(23+)(<23/<23)(Any 23+)(<23)(Any 23+)
CasesControlsCasesControls(95% CI)(95% CI)CasesControlsCasesControls(95% CI)(95% CI)
  • 1

    Adjusted for age, sex, BMI, physical activity, energy intake, cigarette smoking amount, dietary calcium and fiber.

  • 2

    Cutpoints for vitamin D (mcg) intake are 5.1 and 8.8 for men and 3.8 and 7.0 for women; hours of sunshine exposure cut-points were 1.3. and 25.5 for men and 6.5 and 14.0 for women.

  • 3

    Interaction determined by Relative Excess Risk from Interaction (RERI), multiplicative interaction.

Vitamin D2
 High1532141201321.00 (Reference)1.29 (0.92, 0.79)871041111741.00 (Reference)0.74 (0.51, 1.08)
 Middle167200921341.12 (0.80, 1.57)0.94 (0.64, 1.38)911051281790.94 (0.60, 1.46)0.81 (0.54, 1.23)
 Low1762421441040.93 (0.64, 1.35)1.71 (1.14, 2.56)1051241592110.94 (0.59, 1.51)0.84 (0.54, 1.31)
 p interaction3RERI, 0.02; Multiplicative, 0.23RERI, 0.75; Multiplicative, 0.63
Sunshine exposures (Average hours in sun/day)
 High1772331001391.00 (Reference)0.94 (0.67, 1.30)881121231791.00 (Reference)0.85 (0.59, 12.30)
 Middle1442221161120.83 (0.62, 1.12)1.38 (0.99–1.93)991121281921.04 (0.70, 1.54)0.80 (0.56, 1.15)
 Low1702001391181.12 (0.83, 1.50)1.51 (1.09, 2.09)951051441921.01 (0.68, 1.510)0.86 (0.60, 1.23)
 p interactionRERI, 0.03; Multiplicative, 0.07RERI, 0.92; Multiplicative, 0.63

The interaction between hours of sunshine exposure and AR genotype was associated with rectal cancer among men (Table III). Having 23 or more CAG repeats of the AR gene was associated with reduced relative risk of rectal cancer in the presence of high sunshine exposure. There were no significant interactions between calcium (data not shown in table) and vitamin D intake in either men or women and rectal cancer.

Table III. Associations Between AR, Vitamin D and Sunshine Exposure and Rectal Cancer
AR CAG repeatsMenWomen
(<23)(23+)(<23)(23+)(<23)(Any 23+)(<23)(Any 23+)
CasesControlsCasesControls(95% CI)(95% CI)CasesControlsCasesControls(95% CI)(95% CI)
  • a

    Odds Ratios (OR) and 95% Confidence Intervals (CI) adjusted for age, BMI, physical activity, energy intake, cigarette smoking amount, dietary calcium and fiber.

  • 2

    Cutpoints for vitamin D (mcg) intake are 5.1 and 8.8 for men and 3.8 and 7.0 for women; Hours of sunshine exposure cut-points were 13.3 and 25.5 for men and 6.5 and 14.0 for women.

  • 3

    Interaction determined by Relative Excess Risk from Interaction (RERI), multiplicative interaction.

Vitamin D2
 High9111181831.00 (Reference)1.17 (0.77, 1.78)4155901271.00 (Reference)0.98 (0.59, 1.62)
 Middle9211463840.95 (0.61, 1.48)0.87 (0.54, 1.39)326175860.61 (0.31, 1.17)1.07 (0.59, 1.93)
 Low7910152700.90 (0.55, 1.47)0.86 (0.50, 1.46)313748671.03 (0.50, 2.14)0.92 (0.47, 1.79)
 p interaction3RERI, 0.73; Multiplicative, 0.60RERI, 0.15; Multiplicative, 0.98
 High889255901.00 (Reference)0.62 (0.39, 0.97)305677811.00 (Reference)1.74 (1.00, 3.02)
 Middle7710865770.75 (0.49, 1.15)0.90 (0.57, 1.40)365665971.03 (0.55, 1.91)1.17 (0.67, 2.03)
 Low9512575700.77 (0.51, 1.17)1.06 (0.68, 1.66)3840711011.41 (0.73, 2.69)1.14 (0.66, 1.98)
 p interactionRERI, 0.01; Multiplicative, 0.11RERI, 0.31; Multiplicative, 0.14

The Fok1 VDR genotype interacted with AR to alter relative risk of colon cancer in women. The long, 23 or more CAG repeats in the AR genotype was associated with reduced risk of colon cancer among women with the FF genotype (Table IV). Women with an f allele had a lower risk than women with the FF genotype, but in these women the CAG long allele did not further reduce risk of colon cancer. Among men, the Bsm1/polyA VDR genotypes significantly interacted with the AR gene to alter risk of rectal cancer. There was a reduced risk of the VDR LL/bb genotype conferred reduced risk only among men with 23 or more CAG repeats. There was no significant interaction between the Bsm1 and polyA VDR genotypes and number of CAG repeats in the AR gene among women.

Table IV. Association Between AR with VDR (Bsm1/poly A and Fok1) for Colon and Rectal Cancer
AR CAG repeatsMenWomen
(<23)(23+)(<23)(23+)(<23)(<23/23+; 23+/23+)(<23)(Any 23+)
CasesControlsCasesControls(95% CI)(95% CI)CasesControlsCasesControls(95% CI)(95% CI)
  • a

    Odds ratios (OR) and 95% Confidence Intervals were adjusted for age.

  • 2

    p interaction tested via relative excess risk from interaction (RERI) and multiplicative interaction.

Colon cancer
VDR Bsm1/polyA
LLbb1922511231201.00 (Reference)1.34 (0.98–1.83)1091171592161.00 (Reference)0.79 (0.57–1.10)
LSBb2232731621731.07 (0.82–1.38)1.22 (0.92–1.63)1231601762470.83 (0.58–1.17)0.76 (0.55–1.06)
SSBB8013173750.80 (0.57–1.12)1.27 (0.87–1.85)5356631031.02 (0.64–1.61)0.66 (0.44–0.99)
p interaction2RERI, 0.53; Multiplicative, 0.43RERI, 0.44; Multiplicative, 0.42
VDR Fok1
FF1982431481391.00 (Reference)1.31 (0.97–1.77)1251051381861.00 (Reference)0.62 (0.44–0.88)
Ffff2713811952040.88 (0.69–1.12)1.17 (0.90–1.54)1472142393500.58 (0.41–0.80)0.57 (0.42–0.78)
p interactionRERI, 0.96; Multiplicative, 0.90 RERI, 0.02; Multiplicative, 0.07 
Rectal cancer
VDR Bsm1/polyA
LLbb115106771011.00 (Reference)0.71 (0.48–1.05)4362791111.00 (Reference)1.03 (0.64–1.68)
LSBb10916691980.61 (0.43–0.87)0.87 (0.59–1.28)4160951241.00 (0.58–1.75)1.12 (0.70–1.80)
SSBB365028360.66 (0.40–1.10)0.73 (0.42–1.28)193039420.93 (0.46–1.87)1.40 (0.78–2.52)
p interactionRERI, 0.02; Multiplicative, 0.04 RERI, 0.63; Multiplicative, 0.69 
VDR Fok1
FF8212278991.00 (Reference)1.18 (0.79–1.78)4050831081.00 (Reference)0.97 (0.59–1.61)
Ffff1701871111291.35 (0.95–1.92)1.29 (0.88–1.89)57941211640.76 (0.45–1.30)0.93 (0.58–1.50)
p interactionRERI, 0.46; Multiplicative, 0.42 


Our results suggest that the AR operates in conjunction with vitamin D-related compounds to alter risk of colon and rectal cancer. Among men, having more CAG repeats in the AR gene interacted significantly with vitamin D intake from diet and sunshine exposure to alter risk of colon cancer; sunshine exposure and the Bsm1/polyA VDR polymorphism interacted with AR to alter rectal cancer risk. Among women, the Fok1 VDR polymorphism interacted with AR to alter colon cancer risk, although neither vitamin D intake nor sunshine exposure interacted with AR to alter risk of either colon or rectal cancer.

The reason for the sex-specific differences in association is not obvious. We have shown previously that the AR gene interacts significantly with gender.34 The reason for the significant interaction between the AR gene and vitamin D intake and sunshine exposure in altering colorectal cancer risk among men is not clear. This difference in association could be from differences in sunshine exposure in men and women, and that a certain dose of exposure is needed to have a meaningful effect. In our study, men reported more sunshine exposure than women. Our previous results for rectal cancer suggested a trend towards a protective effect of sunshine exposure in men only.11 However, the difference could be biological in nature in that Westburg et al. have suggested that serum androgen levels in women are regulated by both the AR and estrogen receptor (ER) genes,28 making a more complicated disease pathway in women than in men. It is also possible that the associations detected are spurious findings; thus other studies to replicate these results are needed.

Although, interest in the role of vitamin D and sunlight exposure on cancer risk is not new, few studies have actually attempted to examine these associations in detail. Garland and Garland hypothesized that vitamin D status was associated with colorectal cancer mortality in the 1980s, based on the geographic distribution of colorectal cancer deaths and levels of sunshine in those areas.35, 36 Since then, studies in animals have shown the beneficial effects of 1,25(OH2)D3 in reducing cancer progression.37, 38, 39, 40 However, the associations between Vitamin D and sunshine exposure and colorectal cancer are mixed, with inverse associations generally observed, although often not reaching statistical significance.10, 13 In our study, we observed stronger associations with vitamin D intake and sunshine exposure for colon cancer in men if they also had 23 or more CAG repeats of the AR gene.

Different polymorphisms of the VDR gene interacted with AR for colon cancer (Fok1 polymorphism in women) than for rectal cancer (Bsm1/polyA polymorphisms in men). Our previous work has suggested that the Fok1 polymorphism may be more associated with colon cancer, while the Bsm1/polyA polymorphism may have a greater influence on rectal cancer.24 It is possible that the transactivation efficiency associated with the various VDR genotypes is heterogeneous with respect to downstream genes, such that some genes are more efficiently transactivated by one genotype and other genes by another, and that the importance of the various transactivated genes varies by cancer site or metabolic process. For example, it has been reported that the “less active” VDR alleles f and t are capable of facilitating a stronger cell-mediated immune response as demonstrated by T lymphocyte response.23, 41, 42 The t allele is in linkage-disequilibrium with B and S alleles of the Bsm1 and polyA polymorphisms.17 Another possibility for the observed associations is that different VDR genotypes may differentially interact with vitamin D response elements (VDRE) that could lead to altered ability to transcriptionally regulate VDR target genes.43, 44, 45, 46

Less is known about the role of androgens in the etiology of colorectal cancer than vitamin D or VDR, although studies in rats suggest that androgens work as promoters in the development of colon cancer.47 Androgen receptors are also expressed in colon tissue, and changes acquired in the length of the CAG repeat of the AR gene in colon tumors have been detected.48 AR transcriptional activity is determined by the presence or absence of other co-factors that can alter AR activity.49 Clinical studies suggest the functional importance of the CAG repeat sequence of the AR gene, with fewer repeats being associated with higher transcriptional activity of the receptor; more repeats are associated with lower serum androgen levels.27

There are several limitations to the current study. Perhaps, the greatest limitation is the crude indicator of sunshine exposure used. We asked participants to recall average number of hours exposed to sunshine for each season of the year. Although we gave considerable work in developing the sunshine exposure questions and pre-tested them prior to conducting the study, the questions were not validated. Differences in dose could be influenced by many other factors including use of sunscreen and degree of clothing used to cover the skin and thereby reduce exposure. To obtain accurate recall of these factors by season would be difficult; it is recognized that potential misclassification of sunshine exposure exists from not having this information. Dietary data were obtained using a very detailed questionnaire, although inaccurate recall of diet could contribute to imprecise associations. Additionally, there is variation in vitamin D levels in foods throughout the year. Estimates of vitamin D intake also could be compromised by the variability in vitamin D levels in fortified foods. Genotype data were based on what we believed to be functional polymorphisms of the AR and VDR genes, however, other polymorphism or haplotypes of these genes could possibly yield different results or help refine those observed. We selected cut-points for the AR gene based on the data since there is limited information on the most appropriate cut-points to evaluate. It would have been preferable to not make post hoc cut-points, but there was limited information to use to make these a priori.

In summary, these results suggest that androgen and vitamin D work jointly in their association with colon and rectal cancer. Low levels of sunshine exposure and vitamin D intake may increase risk of colon cancer, especially among men who have 23 or more CAG repeats of the AR gene. Confirmation of these results by other researchers is needed to further our understanding of the role of vitamin D and sunshine exposure in the etiology of colorectal cancer.


The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the National Cancer Institute. We would like to acknowledge the contributions of Michael Hoffman and Thao Tran for genotyping and Sandra Edwards, Karen Curtin, Roger Edwards, Leslie Palmer, Donna Schaffer, Dr. Kristin Anderson and Judy Morse for data management and collection.