Journal of Gastrointestinal Cancer

, Volume 43, Issue 3, pp 512–517

Vitamin D Receptor Gene Variants and Esophageal Adenocarcinoma Risk: A Population-Based Case–Control Study

Authors

  • C. K. Chang
    • University of Ulster
    • King’s College London
  • H. G. Mulholland
    • Cancer Epidemiology and Health Services Research Group, Centre for Public HealthQueen’s University Belfast
  • M. M. Cantwell
    • Cancer Epidemiology and Health Services Research Group, Centre for Public HealthQueen’s University Belfast
  • L. A. Anderson
    • Cancer Epidemiology and Health Services Research Group, Centre for Public HealthQueen’s University Belfast
  • B. T. Johnston
    • Belfast Health and Social Care Trust
  • A. J. McKnight
    • Cancer Epidemiology and Health Services Research Group, Centre for Public HealthQueen’s University Belfast
  • P. D. Thompson
    • University of Ulster
  • R. G. P. Watson
    • Belfast Health and Social Care Trust
    • Cancer Epidemiology and Health Services Research Group, Centre for Public HealthQueen’s University Belfast
  • on behalf of the FINBAR study group
Brief Communication

DOI: 10.1007/s12029-011-9322-9

Cite this article as:
Chang, C.K., Mulholland, H.G., Cantwell, M.M. et al. J Gastrointest Canc (2012) 43: 512. doi:10.1007/s12029-011-9322-9

Abstract

Purpose

Polymorphisms in the vitamin D receptor (VDR) gene may be of etiological importance in determining cancer risk. The aim of this study was to assess the association between common VDR gene polymorphisms and esophageal adenocarcinoma (EAC) risk in an all-Ireland population-based case–control study.

Methods

EAC cases and frequency-matched controls by age and gender recruited between March 2002 and December 2004 throughout Ireland were included. Participants were interviewed, and a blood sample collected for DNA extraction. Twenty-seven single nucleotide polymorphisms in the VDR gene were genotyped using Sequenom or TaqMan assays while the poly(A) microsatellite was genotyped by fluorescent fragment analysis. Unconditional logistic regression was applied to assess the association between VDR polymorphisms and EAC risk.

Results

A total of 224 cases of EAC and 256 controls were involved in analyses. After adjustment for potential confounders, TT homozygotes at rs2238139 and rs2107301 had significantly reduced risks of EAC compared with CC homozygotes. In contrast, SS alleles of the poly(A) microsatellite had significantly elevated risks of EAC compared with SL/LL alleles. However, following permutation analyses to adjust for multiple comparisons, no significant associations were observed between any VDR gene polymorphism and EAC risk.

Conclusions

VDR gene polymorphisms were not significantly associated with EAC development in this Irish population. Confirmation is required from larger studies.

Keywords

Esophageal adenocarcinomaVitamin D receptorPolymorphismSingle nucleotideCase–control studies

Introduction

Vitamin D has anti-carcinogenic effects in the body such as suppressing cell proliferation, promoting cell differentiation, regulating apoptosis and is likely to be involved in immunomodulation and regulation of inflammation [1, 2]. Vitamin D exerts many of its biological effects by binding to the vitamin D receptor (VDR). VDR gene polymorphisms may alter mRNA stability and transcriptional activity [3]. VDR gene polymorphisms have been related to tumour development [4, 5], although the majority of research to date has focused on a few specific polymorphisms and a large amount of common variation in the VDR gene has not been investigated. Esophageal adenocarcinoma (EAC) incidence has risen rapidly over recent decades [6], and while others have investigated VDR gene expression in EAC [7], the relationship between VDR gene polymorphisms and EAC risk has yet to be explored. The aim of this study was to investigate the association between variants in the VDR gene and EAC risk.

Materials and Methods

Study Design

The FINBAR (Factors Influencing the Barrett’s Adenocarcinoma Relationship) population-based case–control study set out to identify risk factors for EAC and its precursor conditions, as previously reported [8]. In brief, cases of histologically confirmed EAC were identified via electronic records from all pathology laboratories in Northern Ireland or via clinical records from main hospitals in the Republic of Ireland between March 2002 and December 2004. Incident EAC patients were excluded if they were aged >85 years, had in situ cancers, or were mentally impaired. Matched controls with no history of gastro-intestinal cancer were selected in Northern Ireland using the General Practice Master Index and in the Republic of Ireland via general practices by the frequency of age and gender in cases. Response rates were 74% and 42% for EAC patients and controls, respectively. The representativeness of controls has been described in previous manuscripts [9, 10]. This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients were approved by the Research Ethics Committees of Queen’s University, Belfast; Cork Teaching Hospital and St. James’ Hospital, Dublin. Written informed consent was obtained from all subjects/patients.

Interview and Measurement

Participants were interviewed by trained personnel who gathered background information on demographics, lifestyle and medical history. This included collection of habitual dietary habits 5 years prior to diagnosis via a 101-item food frequency questionnaire [11]. A venous blood sample was also obtained from participants and DNA was extracted from buffy coats stored at −80°C.

Single Nucleotide Polymorphism (SNP) Selection and Genotyping

Previous resequencing of the VDR gene identified 245 SNPs, of which 26 tag SNPs were identified as necessary to study all common variation within the VDR gene in a Caucasian population from the British Isles [12]. The four classically typed SNPs (BsmI, FokI, TaqI, ApaI) were genotyped [12]. Two SNPs were genotyped by Taqman assays (ABI, Foster City, CA, USA) because they were not amenable to genotyping with the other multiplexed SNPs, and the remaining 25 SNPs by Sequenom iPLEX assays (Sequenom, Hamburg, Germany), which facilitated more cost-effectively. Additionally, a poly(A) microsatellite was genotyped by fluorescent fragment analysis on an ABI 3730 Genetic Analyser (Applied Biosystems, Warrington, UK) [13] and was classified as “long” (L, A18-A23) or “short” (S, A13-A17). Hardy–Weinberg disequilibrium was identified for rs4303288 (p < 0.001) using the exact method integrated to Haploview [14] and it was excluded from further analysis.

Statistical Analyses

Independent t tests and chi-squared tests were applied to assess differences in descriptive characteristics between cases and controls. Univariate and multivariate logistic regression models were used to investigate the effects of VDR gene variants and EAC risk by generating odds ratio (OR) and 95% confidence intervals (CI), prior to and post adjustment for potential confounders. Interactions between two VDR polymorphisms, Fok I and poly(A), were explored by an interaction term in the logistic model based on an a priori hypothesis that they may interact to alter cancer risk [15]. Analyses were performed by Stata/SE version 8.0 (College Station, TX, USA). Permutation analysis was performed to address the issue of multiple comparisons by Haploview [16]. Haploview provides a framework for permuting association results in order to obtain a measure of significance corrected for multiple testing bias. We permuted the set of tests loaded from an external file by specifying 1,000 permutations.

Results

Of the 227 EAC cases and 260 controls recruited, 224 EAC cases and 256 controls were included in the current analysis as they provided blood samples for genotype analysis. Selected characteristics of cases and controls are presented in Table 1.
Table 1

Characteristics of esophageal adenocarcinoma cases and controls

Characteristics

Controls n = 256

Esophageal adenocarcinoma cases n = 224

p value

Age (years)

63.0 ± 12.9

64.3 ± 11.2

p = 0.25

Sex: males, n (%)

216 (84.4)

189 (84.4)

p = 1.00

Education (years)

12.0 ± 3.2

10.7 ± 2.6

p < 0.01

BMI 5 years prior (kg/m2)

27.1 ± 3.9

28.6 ± 4.9

p < 0.01

Energy intake (kcal)

2,576 ± 811

2,756 ± 813

p = 0.02

Vitamin D intake (μg/d)

2.8 ± 1.5

3.2 ± 1.5

p < 0.01

Calcium intake (mg/d)

1126 ± 368

1183 ± 390

p = 0.10

Alcohol intake (g/d)

26.3 ± 23.3

19.8 ± 22.0

p < 0.01

Occupation type, n (%)

 Manual

123 (49.4)

131 (60.1)

p = 0.02

 Non-manual

126 (50.6)

87 (39.9)

GOR symptomsa, n (%)

 Ever

49 (19.2)

108 (48.2)

p < 0.01

Smoking status, n (%)

 Current

44 (17.7)

76 (34.7)

p < 0.01

 Previous

105 (42.2)

98 (44.8)

 Never

100 (40.1)

45 (20.5)

Location, n (%)

 Northern Ireland

119 (46.5)

114 (50.9)

p = 0.34

 Republic of Ireland

137 (53.5)

110 (49.1)

GOR Gastro-oesophageal reflux

aHeartburn/acid reflux symptoms experienced at least once weekly or >50 times per year >5 years before interview date.

Table 2 illustrates the associations between the VDR polymorphisms and EAC risk. After adjusting for potential confounders, individuals with TT alleles at rs2238139 and rs2107301 had significantly lower EAC risk compared to those with CC alleles (OR 0.26, 95% CI 0.07–0.93; OR 0.19, 95% CI 0.06–0.67, respectively).
Table 2

VDR gene polymorphisms and esophageal adenocarcinoma risk

VDR polymorphism

Controls, n (%)

Esophageal adenocarcinoma cases, n (%)

Unadjusted OR (95% CI)

Multivariate adjusted OR (95% CI)

p for trend

BsmI (rs1544410)

 bb

80 (34.0)

66 (32.5)

Ref

Ref

0.77

 Bb

113 (48.1)

104 (51.2)

1.12 (0.73, 1.70)

0.96 (0.55, 1.66)

 BB

42 (17.9)

33 (16.3)

0.95 (0.54, 1.67)

1.59 (0.75, 3.40)

FokI (rs10735810)

 FF

84 (35.9)

78 (38.6)

Ref

Ref

0.96

 Ff

116 (49.6)

93 (46.0)

0.86 (0.57, 1.30)

0.99 (0.58, 1.70)

 Ff

34 (14.5)

31 (15.4)

0.98 (0.55, 1.75)

1.12 (0.50, 2.50)

ApaI (rs7975232)

 AA

63 (26.9)

54 (26.9)

Ref

Ref

0.42

 Aa

131 (56.0)

112 (55.7)

1.00 (0.64, 1.55)

0.79 (0.43, 1.43)

 Aa

40 (17.1)

35 (17.4)

1.02 (0.57, 1.83)

0.81 (0.37, 1.77)

TaqI (rs731236)

 TT

82 (35.0)

68 (33.7)

Ref

Ref

0.75

 Tt

111 (47.5)

101 (50.0)

1.10 (0.72, 1.67)

0.98 (0.56, 1.70)

 Tt

41 (17.5)

33 (16.3)

0.97 (0.55, 1.70)

1.70 (0.79, 3.65)

Poly(A) (long/short)

 SS

84 (35.7)

67 (32.7)

Ref

Ref

0.48

 SL

110 (46.8)

104 (50.7)

1.19 (0.78, 1.80)

1.04 (0.60, 1.81)

 LL

41 (17.5)

34 (16.6)

1.04 (0.60, 1.81)

1.79 (0.84, 3.83)

rs3923693

 CC

181 (76.4)

160 (78.1)

Ref

Ref

0.68

 CT

52 (21.9)

39 (19.0)

0.85 (0.52, 1.35)

0.95 (0.52, 1.76)

 TT

4 (1.7)

6 (2.9)

1.70 (0.47, 6,12)

2.25 (0.40, 12.58)

rs10875694

 AA

6 (2.6)

4 (2.0)

Ref

Ref

0.41

 TA

70 (30.2)

50 (24.6)

1.07 (0.29, 4.00)

1.38 (0.22, 8.53)

 TT

156 (67.2)

149 (73.4)

1.43 (0.40, 5.18)

1.67 (0.28, 9.93)

rs11168275

 AA

142 (60.9)

118 (58.1)

Ref

Ref

0.51

 GA

79 (33.9)

73 (36.0)

1.11 (0.75, 1.66)

1.12 (0.66, 1.90)

 GG

12 (5.2)

12 (5.9)

1.20 (0.52, 2.78)

1.46 (0.45, 4.71)

rs11168287

 AA

54 (23.1)

50 (24.6)

Ref

Ref

0.72

 AG

113 (48.3)

93 (45.8)

0.89 (0.55, 1.43)

0.95 (0.51, 1.77)

 GG

67 (28.6)

60 (29.6)

0.97 (0.58, 1.63)

1.12 (0.57, 2.21)

rs11168314

 AA

14 (6.0)

6 (3.0)

Ref

Ref

0.81

 GA

84 (35.7)

77 (38.1)

2.14 (0.78, 5.84)

2.09 (0.52, 8.36)

 GG

137 (58.3)

119 (58.9)

2.03 (0.76, 5.44)

1.90 (0.49, 7.44)

rs2071358

 AA

6 (2.6)

7 (3.5)

Ref

Ref

0.89

 CA

61 (26.1)

50 (24.7)

0.70 (0.22, 2.23)

1.37 (0.34, 5.56)

 CC

167 (71.3)

145 (71.8)

0.74 (0.25, 2.27)

1.21 (0.32, 4.67)

rs2189480

 AA

28 (12.0)

21 (10.3)

Ref

Ref

0.25

 CA

110 (47.0)

89 (43.9)

1.08 (0.57, 2.03)

1.46 (0.62, 3.46)

 CC

96 (41.0)

93 (45.8)

1.29 (0.69, 2.43)

1.71 (0.71, 4.10)

rs2238139

 CC

5 (2.1)

12 (5.9)_

Ref

Ref

0.09

 CT

81 (34.8)

68 (33.5)

0.35 (0.12, 1.04)

0.30 (0.08, 1.11)

 TT

147 (63.1)

123 (60.6)

0.35 (0.12, 1.02)

0.26 (0.07, 0.93)

rs2239181

 GG

3 (1.3)

5 (2.5)

Ref

Ref

0.96

 GT

39 (16.7)

35 (17.3)

0.54 (0.12, 2.42)

0.45 (0.06, 3.44)

 TT

192 (82.0)

162 (80.2)

0.51 (0.12, 2.15)

0.54 (0.08, 3.79)

rs2239182

 AA

60 (25.9)

45 (22.2)

Ref

Ref

0.30

 AG

108 (46.6)

104 (51.2)

1.28 (0.80, 2.06)

1.03 (0.55, 1.91)

 GG

64 (27.5)

54 (26.6)

1.13 (0.66, 1.91)

1.46 (0.71, 2.98)

rs3782905

 CC

91 (39.2)

76 (38.0)

Ref

Ref

0.75

 CG

109 (47.0)

98 (49.0)

1.08 (0.72, 1.62)

0.94 (0.55, 1.61)

 GG

32 (13.8)

26 (13.0)

0.97 (0.53, 1.77)

1.26 (0.55, 2.89)

rs3922882

 CC

95 (40.6)

88 (43.6)

Ref

Ref

0.42

 CG

108 (46.2)

87 (43.1)

0.87 (0.58, 1.30)

0.71 (0.41, 1.22)

 GG

31 (13.2)

27 (13.3)

0.94 (0.52, 1.70)

0.86 (0.39, 1.93)

rs4760674

 AA

37 (15.9)

35 (17.3)

Ref

Ref

0.59

 CA

108 (46.4)

86 (42.6)

0.84 (0.49, 1.45)

0.71 (0.35, 1.45)

 CC

88 (37.7)

81 (40.1)

0.97 (0.56, 1.69)

1.06 (0.52, 2.18)

rs6823

 CC

75 (32.2)

68 (33.5)

Ref

Ref

0.76

 GC

107 (45.9)

95 (46.8)

0.98 (0.64, 1.50)

0.96 (0.54, 1.71)

 GG

51 (21.9)

40 (19.7)

0.87 (0.51, 1.47)

0.89 (0.44, 1.82)

rs7299460

 CC

106 (45.3)

96 (47.5)

Ref

Ref

0.82

 CT

102 (43.6)

88 (43.6)

0.95 (0.64, 1.42)

1.09 (0.65, 1.83)

 TT

26 (11.1)

18 (8.9)

0.76 (0.40, 1.48)

0.74 (0.28, 1.93)

rs739837

 GG

52 (22.2)

38 (18.7)

Ref

Ref

0.21

 GT

119 (50.9)

99 (48.8)

1.14 (0.69, 1.87)

0.88 (0.46, 1.65)

 TT

63 (26.9)

66 (32.5)

1.43 (0.83, 2.47)

1.51 (0.74, 3.05)

rs7974708

 CC

34 (14.7)

36 (17.6)

Ref

Ref

0.62

 CT

114 (49.1)

93 (45.6)

0.77 (0.45, 1.33)

0.53 (0.25, 1.11)

 TT

84 (36.2)

75 (36.8)

0.84 (0.48, 1.48)

0.69 (0.32, 1.47)

rs2107301

 CC

128 (54.2)

113 (55.9)

Ref

Ref

0.06

 TC

87 (36.9)

79 (39.1)

1.03 (0.69, 1.53)

0.96 (0.56, 1.62)

 TT

21 (8.9)

10 (5.0)

0.54 (0.24, 1.19)

0.19 (0.06, 0.67)

rs2238135

 CC

12 (5.1)

8 (3.9)

Ref

Ref

0.37

 GC

92 (39.1)

82 (40.2)

1.34 (0.52, 3.43)

1.76 (0.53, 5.90)

 GG

131 (55.8)

114 (55.9)

1.31 (0.52, 3.31)

1.93 (0.59, 6.36)

rs2408876

 CC

48 (20.6)

30 (14.9)

Ref

Ref

0.34

 TC

103 (44.2)

89 (44.1)

1.38 (0.81, 2.37)

1.15 (0.57, 2.30)

 TT

82 (35.2)

83 (41.0)

1.62 (0.94, 2.80)

1.41 (0.69, 2.89)

rs2544038

 CC

68 (28.8)

50 (24.6)

Ref

Ref

0.18

 CT

118 (50.0)

89 (43.8)

1.03 (0.65, 1.62)

0.80 (0.43, 1.50)

 TT

50 (21.2)

64 (31.6)

1.74 (1.04, 2.93)

1.57 (0.78, 3.15)

rs4073729

 CC

162 (69.2)

145 (72.1)

Ref

Ref

0.59

 TC

66 (28.2)

53 (26.4)

0.90 (0.59, 1.37)

1.30 (0.73, 2.32)

 TT

6 (2.6)

3 (1.5)

0.56 (0.14, 2.27)

0.62 (0.09, 4.31)

Multivariate adjustments: age at interview (years), sex, alcohol consumption (grams per day), energy intake (kilocalories per day), BMI 5 years prior (kilograms per meter squared), regular gastro-esophageal reflux symptoms (ever/never), smoking status (current/former/never), education (years), location (Northern Ireland/Republic of Ireland), occupation (manual/non-manual).

We also explored associations after grouping recessive homozygotes and heterozygotes and comparing these with dominant homozygotes, and vice versa. This resulted in mostly non-significant associations (data not shown), although there was some evidence that SS alleles of the poly(A) microsatellite were associated with increased EAC risk compared with SL/LL genotypes (OR 2.30, 95% CI 1.01–5.26). However, once permutation analyses were performed to address multiple comparisons, no significant associations remained for any VDR gene polymorphism and EAC risk. No significant association was identified with EAC risk when investigating a potential interactive effect of FokI and poly(A) polymorphisms (p = 0.72).

Discussion

The results of this novel investigation suggest little evidence for an association between common VDR gene polymorphisms and EAC risk in an Irish population. Our non-significant findings are in line with those of a Chinese study that observed no role for VDR Taq1 genotypes in esophageal squamous cell carcinoma [17]. This study comprehensively examined the VDR gene and included, but was not limited to, well-characterized SNPs implicated in other diseases [12]. There were some indications for a protective role of TT homozygotes at rs2238139 and rs2107301 SNPs which are both intronic, while a positive association was observed for SS homozygotes of the poly(A) microsatellite, although these became non-significant after adjustment for multiple comparisons. Both rs2238139 and rs2107301 are VDR tag SNPs that, to our knowledge, do not have any functional effect on VDR activity. Previous in vitro work has demonstrated that the LL allele of the poly(A) microsatellite may produce more stable mRNA than the S allele [15]. It could therefore be speculated that SS homozygotes have less-stable VDR mRNA that may alter gene transcription to increase EAC risk. Our research team has demonstrated that vitamin D intake is associated with an increased EAC risk [18] and it could be speculated that vitamin D intake may interact with poly(A) microsatellite to influence tumour development. Studies investigating the poly(A) microsatellite and risk of other cancers have produced conflicting results [19] and it may well be that this observation was due to chance. Our study has several strengths, including its population-based design, strict inclusion criteria and ability to adjust for multiple confounders. However, a key limitation was the limited study size to enable nutrient–gene interactions to be analysed.

In conclusion, we found no significant associations between VDR gene polymorphisms and EAC risk in our population-based case–control study.

Acknowledgments

The FINBAR study group members include LJ Murray (Queen’s University Belfast), LA Anderson (Queen’s University Belfast), BT Johnston (Belfast Health & Social Care Trust), RGP Watson (Belfast Health & Social Care Trust), J McGuigan (Belfast Health & Social Care Trust), HR Ferguson (Belfast Health & Social Care Trust), SJ Murphy (St Vincent’s Hospital Dublin), JV Reynolds (St James’ Hospital, Dublin) and H Comber (National Cancer Registry of Ireland). We appreciate the contributions made by the study participants, their families and all who assisted with the study, particularly the Northern Ireland Cancer Registry and National Cancer Registry Cork. This research was supported by funding from the Ulster Cancer Foundation, the Northern Ireland R&D office and the Health Research Board.

Conflicts of Interest

The authors have no conflict of interest to declare.

Copyright information

© Springer Science+Business Media, LLC 2011