Skip to main content
NCBI home page
Journal of Cancer Research and Clinical Oncology logoLink to Journal of Cancer Research and Clinical Oncology
. 2015 Jan 22;141(9):1533–1544. doi: 10.1007/s00432-015-1915-4

Association between co-inhibitory molecule gene tagging single nucleotide polymorphisms and the risk of colorectal cancer in Chinese

Jie Ge 1, Lin Zhu 1, Junde Zhou 2, Guangxiao Li 1, Ye Li 1, Shuying Li 1, Zhiwei Wu 1, Jiesheng Rong 3, Huiping Yuan 4, Yanhong Liu 5, Qiang Chi 2, Daxun Piao 6, Yashuang Zhao 1,✉, Binbin Cui 7,✉
PMCID: PMC11823739  PMID: 25604582

Abstract

Purpose

T lymphocyte immune responses are controlled by both co-stimulatory and co-inhibitory signaling through T cell co-receptors. Cytotoxic T lymphocyte antigen-4 (CTLA-4), programmed death 1 (PD-1) and B and T lymphocyte attenuator (BTLA) are all co-inhibitory molecules that negatively regulate the activation of T cells. In this study, we investigated the relationship between ten tagging SNPs in three co-inhibitory molecule genes and colorectal cancer (CRC).

Methods

We conducted a hospital-based case–control study consisting of 601 cases with CRC and 627 CRC-free individuals from the Heilongjiang Province of China.

Results

The rs7421861 CT genotype was significantly associated with the risk of colorectal cancer compared to the wild-type TT genotype (adjusted OR 1.314, 95 % CI 1.012–1.706, P = 0.041). The rs2705535 TT genotype was associated with the risk of rectal cancer [OR 1.819 (1.093–3.027), P = 0.021]. There was statistical interaction between the PD-1/rs2227982 (CT + TT) genotypes and high seafood intake (>once/week), as well as the CTLA-4/rs231777 variant and high pungent food intake (>3 times/week). The AG + AA genotypes of CTLA-4/rs3087243 statistically and antagonistically interacted with soybeans, pork and alcohol intake and were associated with CRC risk. Analogously, BTLA/rs1844089 interacted with pork intake, PD-1/rs7421861 with beef and lamb consumption and PD-1/rs6710479 with barbecue consumption. Haplotype G-C-G-A-T-T-A was significantly associated with CRC risk (OR 1.221 P = 0.034).

Conclusions

These data indicate potential associations between BTLA and PD-1 polymorphisms and CRC susceptibility. Additionally, the three co-inhibitory molecule gene SNPs have environmental interactions associated with CRC risk.

Electronic supplementary material

The online version of this article (doi:10.1007/s00432-015-1915-4) contains supplementary material, which is available to authorized users.

Keywords: Cytotoxic T lymphocyte antigen-4 (CTLA-4), Programmed death 1 (PD-1), B and T lymphocyte attenuator (BTLA), Single nucleotide polymorphisms (SNPs), Colorectal cancer (CRC)

Introduction

Colorectal cancer (CRC) is a malignant neoplasm that arises from the lining of the large intestine (colon and rectum). Its incidence and mortality rates have increased rapidly in developed and developing countries. According to statistical data from WHO, colorectal cancer is the third most common cancer in males, and the second in females, worldwide, with approximately 221,313 new cases and 110,486 deaths annually in China (http://globocan.iarc.fr/factsheets/cancers/colorectal.asp).

Studies suggest that environmental factors such as low fiber diet, low physical activity, obesity, smoking and alcohol consumption are risk factors for colorectal cancer. Risk of colorectal cancer is also higher in individuals with inflammatory bowel disease and Crohn’s disease, which are related to immune responses (Cooper et al. 2010; Baumgart and Sandborn 2012). The immune system plays a key role in suppressing tumor growth, and the incidence of cancer would be much greater if not for the ability of the immune system to identify and eliminate nascent tumor cells (Peggs et al. 2009). During the development of cancer, the innate and adaptive immune responses are carefully orchestrated through soluble and membrane-bound regulators, resulting in the deployment of the most suitable effectors for controlling tumor growth (Dranoff 2005). The most significant anti-tumor response is mediated by T lymphocytes and natural killer (NK) cells (Kuznetsov et al. 1994). The generation and maintenance of immune responses are controlled by both co-stimulatory and co-inhibitory signaling through T cell co-receptors (Peggs et al. 2009). Cytotoxic T lymphocyte antigen-4 (CTLA-4) and programmed death 1 (PD-1) are the most well-established inhibitory members of the Ig superfamily. The B and T lymphocyte attenuator (BTLA) is the most recently described inhibitory member of this superfamily (Watanabe et al. 2003a). Meanwhile, the variants within this gene family may affect the function and risk of cancer (Wang et al. 2007b). Considering the importance of co-inhibitory molecules in cancer immunology, we hypothesized that genetic variation in co-inhibitory molecule genes may associate with CRC susceptibility. To test this hypothesis, we selected ten tagging SNPs in the CTLA-4, BTLA and PD-1 genes to study the associations between selected co-inhibitory molecule gene polymorphisms and the risk of colorectal cancer in Chinese Han population of Heilongjiang province.

Materials and methods

Subjects

The study subjects consisted of 601 new sporadic colorectal cancer patients. These patients came from the Department of Abdominal Surgery of Cancer Hospital of Harbin Medical University and the Department of General Surgery of The First and Second Affiliated Hospital of Harbin Medical University. The CRC patients were diagnosed via pathology and ranged from 25 to 89 years old (mean age at diagnosis 60.5 ± 11.2 years). The controls consisted of 627 patients without cancer or a history of personal malignancy and autoimmune disorders and other related diseases. The controls came from the Department of Orthopedics and Department of ophthalmology of The Second Affiliated Hospital of Harbin Medical University with ages ranging from 21 to 86 (mean age at sampling 57.1 ± 11.2 years). Both patients and controls originated from the Heilongjiang Province of China from July 2004 to May 2011. Blood samples were obtained from subjects after they had provided written informed consent.

Ethics statement

This study is approved by the ethics committee of Harbin Medical University. Informed written consent was obtained from all participants.

Data collection

All subjects completed the Epidemiology Questionnaire via face-to-face interviews within 3–5 days after their surgery. The questionnaire included information about demographics, body height and weight, medical history, smoking, drinking, dietary habits and other lifestyle factors over the prior year. Twelve food intake items were surveyed: roughage, vegetables, fruits, meats, fishery product, eggs, milk, soybean, pungent food, fried food, pickled food and overnight food. The patients’ pathological and clinical information were obtained from their medical records.

Selection of tagging SNP

The tagging SNPs in the study were ascertained from the HCB/CHB cohort via the HapMap consortium database (http://hapmap.ncbi.nlm.nih.gov/index.html.en) and were analyzed with HaploView using pairwise tagging between SNPs (with a minimum LD of r 2 > 0.8). SNPs with a minor allele frequency (MAF) <5 % in the HCB/CHB cohort were excluded. The selected 10 SNPs are listed in Table 1.

Table 1.

Selected tagging SNPs, the primers and condition of PCR–RFLP and TaqMan assays

Genotyping methods Gene SNP Primers Sequence 5′-3′ Condition of PCR Restriction enzyme# Length of PCR product (bp) Length of digested fragments (bp)
PCR–RFLP CTLA-4 rs3087243 Forward ATTCAGTATCTGGTGGAGT 95 °C 5 min—(95 °C 30 s, 56 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min HpyCH4 IV 291 176 + 115
Reverse ATGCCTGTGATAGTTGAG
rs231775 Forward CTAAACCCACGGCTTCCTT 95 °C 5 min—(95 °C 30 s, 58 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min ApekI 406 262 + 144
Reverse CACTGCCTTTGACTGCTGA
rs231777 Forward GCTCTTCAGAGACTGACACC 95 °C 5 min—(95 °C 30 s, 61 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min DdeI 125 95 + 30
Reverse CCTACTTCATACAAACTACATGG
BTLA rs1844089 Forward AGGCATCGCATCAATAGC 95 °C 5 min—(95 °C 30 s, 54 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min PstI 310 183 + 127
Reverse CAAGGCATACATCCCAAT
rs2705535 Forward TCTGTCTCTCAAATCTTCC 95 °C 5 min—(95 °C 30 s, 54 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min BstZ17I 256 187 + 69
Reverse ACCCTAACCTCATGTCAC
PD-1 rs10204525 Forward TCAGAAGAGCTCCTGGCTGT 95 °C 5 min—(95 °C 30 s, 60 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min NdeI 422 320 + 102
Reverse GGGGAACGCCTGTACCTT
rs2227982 Forward GGGCTTGGTCATTTCTTATC 95 °C 5 min—(95 °C 30 s, 58 °C 30 s, 72 °C 30 s) 32 cycles—72 °C 5 min DrdI 444 306 + 138
Reverse ATGAGGTGCCCATTCCGCTA
TaqMan assays BTLA rs9288953 TaqMans – 95 °C 10 min (95 °C 15 s, 60 °C 1 min) 40 cycles –
PD-1 rs6710479 TaqMans – 95 °C 10 min (95 °C 15 s, 60 °C 1 min) 40 cycles –
rs7421861 TaqMans – 95 °C 10 min (95 °C 15 s, 60 °C 1 min) 40 cycles –
Open in a new tab

DNA extraction and genotyping

Genomic DNA was extracted from 2 ml frozen peripheral whole blood using a DNA Extraction Kit (Qiagen, Germany) according to the manufacturer’s protocol. The rs3087243, rs231775, rs231777, rs1844089, rs2705535, rs10204525 and rs2227982 genotypes were determined by polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP). The polymorphic region was amplified by PCR with a GeneAmp 9700® PCR System (ABI, America), in a 25-µl reaction solution containing 1.0 µl genomic DNA, 2.5 µl 10 × PCR buffer, 2.0 µl 0.2 mM dNTPs, 5 U Taq DNA polymerase (Takara, Japan) and 0.6 µmol of each primer (Sheng Gong, China). PCR products were digested overnight with restriction enzymes (New England BioLabs, Beverly, USA) according to the manufacturer’s protocol and analyzed by 2 % agarose gel electrophoresis. Ten samples of each SNP selected at random were verified by direct sequencing. The average successful genotype concordance proportion of PCR–RFLP and sequencing was 98.0 %. Rs9288953, rs7421861 and rs6710479 were genotyped using the TaqMan SNP genotyping assays (Applied Biosystems, Foster City, CA, USA) because restriction enzymes were unavailable. The assays were run on a Roche LightCycler® 480 (Roche Applied Science, Switzerland) and evaluated according to manufacturer’s instructions. The primers for genotyping, PCR programs, restriction enzymes and the length of digested fragments are shown in Table 1.

Statistical analysis

Hardy–Weinberg equilibrium (HWE) was checked by comparing observed to expected genotype frequencies with a Chi-square test. Associations of specified SNPs with CRC were evaluated using odds ratios (OR) with 95 % confidence intervals (CI) by Statistic Analysis System 9.2 (SAS Institute, Cary, NC, USA). Interactions between SNPs and the environment were studied by crossover analysis and multivariate logistic regression in order to understand both additive and multiplicative interactions (Hosmer and Lemeshow 1992; Hallqvist et al. 1996; García-Martínez et al. 2008). Haplotype analysis was performed by estimating the haplotype frequencies using the online SHEsis package. The MDR software package (www.multifactordimensionalityreduction.org) was used to detect gene–gene interactions. The best disease-predicting MDR model was identified on the basis of interacted-genotypes carrying different sets of risk alleles using the gene counting method. P values ≤ 0.05 were considered significant. For multiple comparisons, significance levels were adjusted via Bonferroni correction, where P values are multiplied by the number of comparisons.

Results

Characteristics of study subject

In total, 601 CRC cases and 627 controls were recruited in the case–control study. The characteristics of the study subjects are summarized in Table 2. There were significant differences between the two groups with respect to age, body mass index (BMI) and occupation. Therefore, these variables could be considered confounding factors and adjusted for when analyzing associations between selected SNPs and CRC risk and the SNP–environment interactions.

Table 2.

Demographic and baseline clinical characteristics of colorectal cancer patients and controls

Characteristics Number of controls (%) (n = 627) Number of cases (%) (n = 601) P value
Age (years)
 <50 158 (25.20) 96 (15.97) <0.0001
 50~ 220 (35.09) 187 (31.11)
 60~ 155 (24.72) 172 (28.62)
 70~ 94 (14.99) 146 (24.29)
Gender
 Male 362 (57.74) 353 (58.73) 0.7224
 Female 265 (42.26) 248 (41.27)
Occupational
 Mental worker 169 (26.95) 228 (38.45) <0.0001
 Manual worker 291 (46.41) 255 (43.00)
 Combined 167 (26.63) 110 (18.55)
BMI (kg/m2)
 ≤18.50 36 (5.82) 44 (7.39) 0.0033
 18.5–23.00 200 (32.31) 240 (40.34)
 ≥23.00 383 (61.87) 311 (52.27)
Smoking
 No 328 (52.31) 344 (57.91) 0.0493
 Yes 299 (47.69) 250 (42.09)
Alcohol
 No 276 (56.67) 239 (59.90) 0.3328
 Yes 211 (43.33) 160 (40.10)
Family history of cancer
 No 527 (84.46) 475 (81.62) 0.1886
 Yes 97 (15.54) 107 (18.38)
NSAIDs use
 No 506 (92.17) 459 (94.25) 0.1854
 Yes 43 (7.83) 28 (5.75)
Tumor site
 Colon 233 (38.77)
 Rectal 368 (61.23)
Dukes stagea
 A–B 315 (58.55)
 C–D 223 (41.45)
General classification of tumorb
 Protrude type 243 (65.15)
 Other types 130 (34.85)
Histological classification of tumorb
 Adenocarcinoma 293 (75.91)
 Other types 93 (24.09)
Degree of differentiationb
 Low 62 (16.06)
 Medium 299 (77.46)
 High 25 (6.48)
CEA level (ng/ml)b
 0–5 165 (42.75)
 ≥5 221 (57.25)
CA19-9 level (u/ml)b
 0–37 286 (74.09)
 ≥37 100 (25.91)
Chemotherapyb
 Yes 155 (40.47)
 No 228 (59.53)
Anastomat on surgeryb
 Yes 278 (76.37)
 No 86 (23.63)
Open in a new tab

a63 subjects have missing data

bThe data only for patients followed up

Association between selected SNPs and colorectal cancer

All genotypes were consistent with Hardy–Weinberg equilibrium (P > 0.05). The frequencies of CTLA-4, BTLA and PD-1 genotypes are shown in Table 3. The frequency of the PD-1/rs7421861 CT genotype was significantly higher than in the wild-type genotype (TT genotype), in CRC patients as compared to controls (OR 1.314, CI 1.012–1.706). The frequency of the BTLA/rs2705535 TT genotype was significantly higher in rectal cancer patients than in controls when using CC genotype as reference (OR 1.819, CI 1.093–3.027). CTLA-4/rs231777 was significantly associated with the risk of CRC in a dominant genetic model (OR 1.299, CI 1.003–1.682). Both BTLA/rs2705535 (OR 1.841, 95 % CI 1.116–3.036) and BTLA/rs9288953 (OR 0.701, 95 % CI 0.514–0.956) were significantly associated with increased risk of rectal cancer in a dominant model. There were no statistical associations between any other tagging SNPs and colon cancer.

Table 3.

Associations between CTLA-4, BTLA and PD-1 polymorphisms and total CRC and colon and rectum cancer separately

Genotype Controls No. (%) Total patients Colon cancer Rectal cancer
No. (%) OR (95 % CI)a P No. (%) OR (95 % CI)a P No. (%) OR (95 % CI)a P
rs231777
 TT 9 (1.45) 9 (1.51) Ref 3 (1.51) Ref 6 (1.65) Ref
 CT 194 (31.29) 154 (25.84) 0.968 (0.431–2.171) 0.9366 54 (27.14) 1.207 (0.328–4.440) 0.7769 94 (25.90) 1.006 (0.390–2.592) 0.9902
 CC 417 (67.26) 433 (72.65) 1.269 (0.576–2.796) 0.5545 142 (71.36) 1.534 (0.427–5.508) 0.5120 263 (72.45) 1.282 (0.508–3.236) 0.5986
 Dominant model 1.299 (1.003–1.682) 0.0471 1.267 (0.873–1.839) 0.2126 1.249 (0.929–1.681) 0.1412
 Recessive model 0.942 (0.341–2.607) 0.9091 1.084 (0.218–5.389) 0.9213 0.766 (0.258–2.276) 0.6313
rs9288953
 CC 128 (20.51) 127 (21.34) Ref 44 (22.22) Ref 79 (21.88) Ref
 CT 310 (49.68) 319 (53.61) 0.984 (0.727–1.333) 0.9184 99 (50.00) 0.909 (0.589–1.402) 0.6656 200 (55.40) 0.975 (0.692–1.374) 0.8868
 TT 186 (29.81) 149 (25.04) 0.774 (0.551–1.088) 0.1401 55 (27.78) 0.805 (0.497–1.303) 0.3774 82 (22.71) 0.689 (0.465–1.021) 0.0633
 Dominant model 0.783 (0.601–1.020) 0.0696 0.860 (0.590–1.255) 0.4346 0.701 (0.514–0.956) 0.0248
 Recessive model 0.9050.679–1.206) 0.4947 0.869 (0.577–1.307) 0.5000 0.867 (0.626–1.201) 0.3918
rs2705535
 CC 395 (63.40) 374 (62.54) Ref 130 (65.33) Ref 223 (61.43) Ref
 CT 193 (30.98) 173 (28.93) 0.979 (0.754–1.270) 0.8728 53 (26.63) 0.927 (0.633–1.357) 0.6951 105 (28.93) 0.956 (0.708–1.291) 0.7674
 TT 35 (5.62) 51 (8.53) 1.519 (0.949–2.431) 0.0814 16 (8.04) 1.324 (0.673–2.607) 0.4162 35 (9.64) 1.819 (1.093–3.027) 0.0213
 Dominant model 1.525 (0.960–2.423) 0.0739 1.351 (0.693–2.634) 0.3775 1.841 (1.116–3.036) 0.0168
 Recessive model 1.059 (0.831–1.351) 0.6425 0.986 (0.692–1.405) 0.9368 1.084 (0.822–1.431) 0.5682
rs7421861
 TT 440 (70.97) 395 (66.28) Ref 133 (66.83) Ref 241 (66.57) Ref
 CT 163 (26.29) 187 (31.38) 1.314 (1.012–1.706) 0.0407 60 (30.15) 1.322 (0.909–1.922) 0.1440 114 (31.49) 1.294 (0.961–1.742) 0.0891
 CC 17 (2.74) 14 (2.35) 0.677 (0.312–1.469) 0.3232 6 (3.02) 0.953 (0.350–2.594) 0.9254 7 (1.93) 0.616 (0.236–1.608) 0.3224
 Dominant model 0.624 (0.289–1.349) 0.2304 0.881 (0.326–2.382) 0.8035 0.571 (0.220–1.485) 0.2504
 Recessive model 1.247 (0.968–1.608) 0.0877 1.280 (0.891–1.838) 0.1812 1.226 (0.917–1.637) 0.1687
Open in a new tab

aAdjusted for age, occupational, BMI, family history of cancer

Haplotype and colorectal cancer

Haplotypes with frequencies ≥1 % in each gene are shown in Table 4. No significant differences in the frequencies of these haplotypes between colorectal cancer patients and controls were observed.

Table 4.

Associations between haplotypes of co-inhibitory molecule gene and CRC risk

Haplotypes Cases No. (%) Controls No. (%) OR (95 % CI) P value
CTLA-4
 A-C-A 143.54 (12.1) 176.61 (14.2) 0.832 (0.656–1.055) 0.128
 A-C-G 34.10 (2.9) 21.26 (1.7) – –
 A-T-G 142.22 (12.0) 174.88 (14.1) 0.833 (0.656–1.057) 0.132
 G-C-A 39.53 (3.3) 30.62 (2.5) 1.368 (0.848–2.209) 0.198
 G-C-G 799.83 (67.3) 799.50 (64.5) 1.175 (0.984–1.403) 0.074
 G-T-G 23.84 (2.0) 17.36 (1.4) – –
BTLA
 C-C-C 240.31 (20.2) 232.96 (18.7) 1.098 (0.898–1.344) 0.362
 C-C-T 82.80 (7.0) 87.55 (7.0) 0.987 (0.723–1.349) 0.935
 C-T-C 15.99 (1.3) 21.76 (1.7) – –
 C-T-T 233.90 (19.7) 222.73 (17.9) 1.122 (0.915–1.377) 0.269
 T-C-C 585.20 (49.3) 650.15 (52.2) 0.878 (0.746–1.033) 0.116
 T-T-C 20.50 (1.7) 17.13 (1.4) – –
PD-1
 A-C-T-A 270.49 (22.8) 302.18 (24.4) 0.879 (0.727–1.062) 0.180
 A-T-T-A 554.66 (46.8) 554.04 (44.7) 1.029 (0.874–1.212) 0.733
 G-C-C-G 190.43 (16.1) 174.34 (14.1) 1.129 (0.902–1.412) 0.289
 G-C-T-A 54.74 (4.6) 55.88 (4.5) 0.993 (0.677–1.454) 0.970
 G-C-T-G 69.20 (5.8) 68.72 (5.5) 1.022 (0.724–1.442) 0.901
CTLA-4 and PD-1
 A-C-A-A-T-T-A 71.93 (6.1) 88.22 (7.1) 0.821 (0.593–1.136) 0.234
 A-T-G-A-C-T-A 32.48 (2.8) 41.67 (3.4) 0.791 (0.495–1.262) 0.324
 A-T-G-A-T-T-A 60.51 (5.1) 90.10 (7.3) 0.666 (0.474–0.935) 0.018
 G-C-G-A-C-T-A 175.42 (14.9) 201.94 (16.3) 0.864 (0.689–1.083) 0.204
 G-C-G-A-T-T-A 391.31 (33.2) 354.83 (28.7) 1.221 (1.015–1.468) 0.034
 G-C-G-G-C-C-G 122.45 (10.4) 107.57 (8.7) 1.190 (0.903–1.568) 0.216
 G-C-G-G-C-T-A 43.46 (3.7) 45.35 (3.7) 0.980 (0.640–1.501) 0.926
 G-C-G-G-C-T-G 43.56 (3.7) 33.69 (2.7) 1.339 (0.846–2.120) 0.210
Open in a new tab

CTLA-4 and PD-1 are located on the same chromosome, and a haplotype was constructed by combining the SNPs those in these two genes. Haplotype G-C-G-A-T-T-A was the most frequent haplotype observed in both patients and controls (33.2 and 28.7 %, respectively) and was associated with the risk of CRC (OR 1.221, CI 1.015–1.468), while haplotype A-T-G-A-T-T-A was statistically associated with reduced CRC risk (OR 0.666, CI 0.474–0.935, P = 0.018).

Gene–gene interaction and colorectal cancer

We used multifactor dimensionality reduction (MDR) algorithms to detect the top potential interactions between the analyzed SNPs in relation to colorectal cancer risk. As shown in Table 5, analysis of the immune-related genes with MDR yielded the top ranking SNP combination: rs2227982, rs9288953 and rs6710479 (testing balance accuracy 0.51, cross-validation consistency 6/10). There was no significant gene–gene interaction associated with risk of CRC (Fig. 1).

Table 5.

Predication of colorectal cancer risk factors in MDR analysis

Best model Testing bal acca CV consistencyb Sign test (P value)
rs231777 0.47 5/10 0.9950
rs3087243, rs9288953 0.47 2/10 0.9950
rs2227982, rs9288953, rs6710479 0.51 6/10 0.7040
Rs2227982, rs10204525, rs9288953, rs6710479 0.50 3/10 0.8710
Open in a new tab

aTesting bal acc: testing balance accuracy

bCV consistency: cross-validation consistency

Fig. 1.

Fig. 1

Open in a new tab

Entropy-based interaction graph. The percentage of entropy removed by each single nucleotide polymorphism (SNP) is shown in the boxes. The percentage of entropy removed by the two-way interactions between SNPs is shown by each connection. Positive entropy values indicate synergic interaction, while negative entropy values indicate redundancy

Interaction between gene and environmental factors on the risk of CRC

We observed a synergistic statistical interaction between PD-1/rs2227982 CT + TT genotypes and high (average intake > once per week) seafood intake on the risk of CRC (ORi 4.06, 95 % CI 1.47–11.23, P = 0.007). Similarly, the CTLA-4/rs231777 CT + CC genotypes interacted with high pungent food intake (average intake times per week >3) (ORi 1.63, CI 1.00–2.65). For the CTLA-4/rs3087243 AG + AA genotypes, we observed antagonistic interactions with elevated soybeans and pork intake and alcohol drinking that were associated with risk of CRC (ORi 0.59, CI 0.35–0.99; ORi 0.61, CI 0.39–0.97; ORi 0.54, CI 0.32–0.94, respectively). Analogously, BTLA/rs1844089 interacted with high pork intake (ORi 0.61, 95 % CI 0.42–0.90), PD-1/rs7421861 with high beef and lamb intake (ORi 0.54, 95 % CI 0.30–0.98) and PD-1/rs6710479 with high barbecue intake (ORi 0.51, 95 % CI 0.26–0.99) (Table 6).

Table 6.

Combined and interactive effect between SNPs and environmental factors on CRC risk

SNP Genotypes Environmental factors
rs2227982 Seafood (average times/week)
≤1 >1 Interaction
OReg (95 % CI) ORi (95 % CI) p value
CC Ref. 0.10 (0.01–0.81)
CT + TT 1.21 (0.91–1.61) 2.75 (1.27–5.95) 4.06 (1.47–11.23) 0.007*
rs3087243 Soybeans (average times/week)
≤1 >1 Interaction
OReg (95 % CI) ORi (95 % CI) p value
GG Ref. 1.44 (1.03–2.03)
AG + AA 1.18 (0.71–1.96) 1.03 (0.70–1.52) 0.59 (0.35–0.99) 0.046*
rs3087243 Pork
<250 g/week ≥250 g/week Interaction
OReg (95 % CI) ORi (95 % CI) p value
GG Ref. 1.84 (1.38–2.47)
AG + AA 1.04 (0.74–1.46) 1.21 (0.82–1.78) 0.61 (0.39–0.97) 0.037*
rs3087243 Alcohol intake
No Yes Interaction
OReg (95 % CI) ORi (95 % CI) p value
GG Ref. 1.14 (0.80–1.62)
AG + AA 1.30 (0.87–1.95) 0.69 (0.43–1.10) 0.54 (0.32–0.94) 0.028*
rs1844089 Pork
<250 g/week ≥250 g/week Interaction
CC Ref. 2.22 (1.59–3.10)
CT + TT 1.40 (1.02–1.94) 1.60 (1.14–2.24) 0.61 (0.42–0.90) 0.013*
rs231777 Pungent food (average times/week)
≤3 >3 Interaction
TT Ref. 0.15 (0.02–1.36)
CT + CC 0.46 (0.11–1.86) 0.37 (0.09–1.50) 1.63 (1.00–2.65) 0.048*
rs7421861 beef and lamb
<250 g/week ≥250 g/week Interaction
TT Ref. 0.90 (0.61–1.32)
CT + CC 1.42 (1.07–1.88) 0.62 (0.35–1.08) 0.54 (0.30–0.98) 0.043*
rs6710479 barbecue (average times/week)
<1 ≥1 Interaction
AA Ref. 4.76 (2.81–8.07)
AG + GG 1.21 (0.88–1.67) 2.74 (1.52–4.93) 0.51 (0.26–0.99) 0.047*
Open in a new tab

*OReg, ORi were adjusted  for age, occupational, BMI, family history of cancer. Significant results (p < 0.05)

Discussion

In the present hospital-based case–control study, we investigated the association between ten selected tagging SNPs of the co-inhibitory molecule gene and the risk of sporadic colorectal cancer in Northeast China. Our results showed that the rs231777CC genotype increased CRC risk 29.9 % as compared to individuals with the TT + CT genotype. No significant associations between any polymorphism and colon cancer were observed.

Human CTLA-4 gene (Gene ID: 1493, MIM number: 123890) is located on the long arm of chromosome 2q33, is approximately 6.2 kb in length and consists of 4 exons. CTLA-4 is involved in establishing and maintaining peripheral T cell tolerance, which controls T cell activation and reactivity through the apoptotic pathway (Brunet et al. 1987; Gribben et al. 1995). It negatively regulates the activation of T cells through CD28 (Kallinich et al. 2007). CTLA-4 transmits inhibitory signals to attenuate T cell activation by competing for B7 ligands with its homologue CD28 (van der Merwe et al. 1997; Ostrov et al. 2000). In addition, CTLA-4 can inhibit TCR signaling by direct interaction with the TCR complex (Lee et al. 1998), acting as an intracellular phosphatase. Recently, CTLA-4 has been suggested as a candidate gene for many autoimmune diseases and malignant tumors, such as Graves’ disease, celiac disease, autoimmune hypothyroidism, multiple sclerosis, rheumatoid arthritis, breast cancer and colorectal cancer (Scalapino and Daikh 2008; Qi et al. 2009; Wang et al. 2007a). To our knowledge, our result is the first to indicate an association between CTLA-4/rs231777 and CRC. However, further studies of the functional role that CTLA-4/231777 may play in CRC are needed. In addition, research has shown that CTLA-4/rs231775 is associated with the risk of CRC in Chinese populations (Qi et al. 2009), although our research did not support the same conclusion.

Our study revealed that the PD-1/rs7421861 CT genotype may confer 31.4 % higher risk of CRC when compared to the wild-type genotype. However, the C allele did not have a dose–response relationship with the risk of CRC, and thus this relationship may be a chance finding. PD-1 exerts distinct independent effects during different phases of T cell responses, including regulating the threshold for T cell activation, down-regulating T cell proliferation, and inducing apoptosis in activated T cells (Hua et al. 2011). The extracellular domain of PD-1 is an immunoglobulin-like structure and the cytoplasmic tail contains an immunoreceptor tyrosine-based inhibitory motif (ITIM). Interactions between PD-1 and its ligands PD-L1 (B7-H1; CD274) or PD-L2 (B7-DC; CD273) can activate the cytoplasmic ITIM of PD-1 and induce the inhibitory signal to attenuate T lymphocyte activation and proliferation, to suppress cytokine secretion and to induce T cell apoptosis as well as maintain peripheral tolerance (Freeman et al. 2000; Latchman et al. 2001; Nishimura and Honjo 2001; Droeser et al. 2013). Moreover, research shows that PD-1.5 C/T (rs2227981) is associated with colon cancer in Iranians (Mojtahedi et al. 2012).

The present study is first to probe the association between BTLA SNPs and CRC. We found that the rs9288953 TT genotype decreased rectal cancer risk 29.9 % when compared to the CC + CT genotypes. Conversely, the rs2705535 TT genotype increased rectal cancer risk 84.1 % when compared to the CC + CT genotype. Additionally, the risk of rectal cancer in carriers of the TT genotype is 1.819-fold greater than that of the CC genotype. BTLA was identified as an inhibitory coreceptor expressed at various levels in a wide range of hematopoietic cells including Th1 cells, B cells, CD8+ T cells, NKT cells, NK cells, macrophages and dendritic cells (Watanabe et al. 2003b; Hurchla et al. 2005; Murphy et al. 2006; Iwata et al. 2010). The inhibitory signal of BTLA appears to be mediated by the recruitment of SHP-1 and SHP-2 to two immunoreceptor tyrosine inhibitory motifs (ITIMs) in the cytoplasmic region of BTLA (Watanabe et al. 2003a). We hereby propose that rs2705535 may influence splicing transforms (Jonsson et al. 1992; Majewski and Ott 2002). We found no evidence regarding the role of rs9288953.

We found that rs231777, rs7421861, rs9288953 and rs2705535 were associated with CRC. The current study did not determine the impact of the polymorphisms on baseline T cell activation in the patient blood samples. However, in future studies, we will determine whether SNPs associated with CRC downregulate the activation of T cells which would impede T cell surveillance mechanisms. We applied human splicing finder (HSF) software to predict the effect of SNPs on splicing site signals (Desmet et al. 2009). The results show that rs231777 could activate three new splice sites both in potential splicing sites and splicing enhancer motifs, while rs9288953 could activate six new splice sites in splicing enhancer motifs and break one splicing sites in silencer motifs. Thus, rs231777 and rs9288953 may enhance the splicing signal and strengthen the expression of CTLA-4 and BTLA. However, this mechanism requires further study.

MDR is a model-free, nonparametric approach that can detect higher-order interactions even in a small population by reducing the dimensionality of multi-locus information to identify the polymorphisms or factors associated with an increased risk of disease (Motsinger and Ritchie 2006). However, in our study, we failed to obtain any meaningful results from gene–gene interaction analysis by means of MDR. This suggests that the immune molecules we studied played entirely co-inhibitory roles in immune response. However, the three co-inhibitory molecules have different mechanisms as previously mentioned. We should therefore study specific co-inhibitory molecules and their corresponding ligands to find gene–gene interactions associated with risk of colorectal cancer.

Because SNPs in the same gene are not inherited randomly, but as combination of alleles, we used a haplotype-based approach to simultaneously analyzing inherited patterns of the ten SNPs. Haplotypes composed of the three to four SNPs did not enable us to find significant effects on and modification of CRC risk. The G-C-G-A-T-T-A haplotype, which combined CTLA4 and PD-1 SNPs, was associated with a 22.1 % higher risk than that of all the other haplotypes combined.

The interaction of environmental and genetic factors may result in overwhelming diseases, especially in chronic diseases (Cheng and Lee 2012; Marcus et al. 2000; Han et al. 2004). In the study of gene–environment interaction, we observed statistically significant interactions between rs3087243 and rs231777 in CTLA-4 and alcohol intake as well as several significant combinations with soybeans, pork and pungent food. Meanwhile, rs2227982, rs7421861 and rs6710479 in PD-1 had interactions with seafood, beef and lamb and barbecue. Until now, related molecular mechanism of the interaction between these SNPs and environmental factor has not been reported, although some studies have found molecular mechanisms for environmental risk factors for colorectal cancer such as high alcohol consumption (Gao et al. 2013; Brooks and Theruvathu 2005; Hoek and Pastorino 2002), soy food intake (Yang et al. 2009; Lechner et al. 2005; Burns and Korach 2012; Jassam et al. 2005; Martineti et al. 2005), red meat and high-fat dairy products (Yusof et al. 2012; Kesse et al. 2006). Given that our study is carried out in northeast China, local residents seldom eat seafood. Interactions between seafood and SNPs need to be further studied.

This is the first systematic study to describe co-inhibitory tag-SNPs and colorectal cancer occurrence and some limitations should be taken into consideration. First, selection bias cannot be discounted because this is a hospital-based case–control study. However, this bias will not have a substantial impact on the results in gene–disease association studies. Moreover, recall bias in the survey of dietary factors is inevitable, but might be minimized by investigating new cases. The sample size in this study was not large enough and had insufficient power to detect very weak associations between polymorphisms and CRC risk. Assuming the risk genotype frequency ranged from 0.2 to 0.4, the power to detect an OR of 1.4 at a two-sided α = 0.05 was less than 85 % (from 68.5 to 82.4 %). Finally, limited by the crossover analysis, we combined ex- and never drinkers or smokers when analyzing the interaction between gene–environment, and if non-drinkers or smokers included ex-drinkers who gave up drinking due to ill health, the odds ratios of current drinkers might be underestimated.

In conclusion, our study showed that the representative genetic variants rs231777 in CTLA-4 and rs7421861 in PD-1 may serve as candidate markers for susceptibility to CRC in a Chinese population. The variants of rs2705535 and rs9288953 in BTLA may also relate to risk of rectal cancer. Additionally, there was evidence suggesting that the association between rs3087243, rs231777, rs2227982, rs1844089, rs7421861 and rs6710479 and CRC risk may be modified by environmental factors such as seafood, soybeans, pork, beef, lamb, alcohol and pungent food intake. Studies with larger sample and functional evaluation are needed to confirm our findings.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (DOC 763 kb) (763KB, doc)

Acknowledgments

This work was supported by grants from Natural Science Foundation of China (Grant No. 30972539) and Specialized Research Fund for the Technical Innovation Talented Person of Harbin (Grant No. 2011RFXXS045).

Conflict of interest

None.

Contributor Information

Yashuang Zhao, Phone: 86-451-87502823, Email: zhao_yashuang@263.net.

Binbin Cui, Email: hydcui_binbin@163.com.

References

  1. Baumgart DC, Sandborn WJ (2012) Crohn’s disease. Lancet 380(9853):1590–1605 [DOI] [PubMed] [Google Scholar]
  2. Brooks PJ, Theruvathu JA (2005) DNA adducts from acetaldehyde: implications for alcohol-related carcinogenesis. Alcohol 35(3):187–193 [DOI] [PubMed] [Google Scholar]
  3. Brunet JF, Denizot F, Luciani MF, Roux-Dosseto M, Suzan M, Mattei MG, Golstein P (1987) A new member of the immunoglobulin superfamily–CTLA-4. Nature 328(6127):267–270. doi:10.1038/328267a0 [DOI] [PubMed] [Google Scholar]
  4. Burns KA, Korach KS (2012) Estrogen receptors and human disease: an update. Arch Toxicol 86(10):1491–1504. doi:10.1007/s00204-012-0868-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Cheng KF, Lee JY (2012) Assessing the joint effect of population stratification and sample selection in studies of gene–gene (environment) interactions. BMC Genet 13:5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cooper K, Squires H, Carroll C, Papaioannou D, Booth A, Logan RF, Maguire C, Hind D, Tappenden P (2010) Chemoprevention of colorectal cancer: systematic review and economic evaluation. Health Technol Assess 14(32):1–206. doi:10.3310/hta14320 [DOI] [PubMed] [Google Scholar]
  7. Desmet FO, Hamroun D, Lalande M, Collod-Beroud G, Claustres M, Beroud C (2009) Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 37(9):e67 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dranoff G (2005) CTLA-4 blockade: unveiling immune regulation. J Clin Oncol 23(4):662–664 [DOI] [PubMed] [Google Scholar]
  9. Droeser RA, Hirt C, Viehl CT, Frey DM, Nebiker C, Huber X, Zlobec I, Eppenberger-Castori S, Tzankov A, Rosso R, Zuber M, Muraro MG, Amicarella F, Cremonesi E, Heberer M, Iezzi G, Lugli A, Terracciano L, Sconocchia G, Oertli D, Spagnoli GC, Tornillo L (2013) Clinical impact of programmed cell death ligand 1 expression in colorectal cancer. Eur J Cancer 49(9):2233–2242 [DOI] [PubMed] [Google Scholar]
  10. Freeman GJ, Long AJ, Iwai Y, Bourque K, Chernova T, Nishimura H, Fitz LJ, Malenkovich N, Okazaki T, Byrne MC, Horton HF, Fouser L, Carter L, Ling V, Bowman MR, Carreno BM, Collins M, Wood CR, Honjo T (2000) Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J Exp Med 192(7):1027–1034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gao CM, Ding JH, Li SP, Liu YT, Cao HX, Wu JZ, Tang JH, Tajima K (2013) Polymorphisms in XRCC1 Gene, Alcohol drinking, and Risk of Colorectal Cancer: a case–control Study in Jiangsu Province of China. Asian Pac J Cancer Prev 14(11):6613–6618 [DOI] [PubMed] [Google Scholar]
  12. García-Martínez C, Herrera F, Molina D, Sánchez AM (2008) Global and local real-coded genetic algorithms based on parent-centric crossover operators. Eur J Oper Res 185(3):1088–1113 [Google Scholar]
  13. Gribben JG, Freeman GJ, Boussiotis VA, Rennert P, Jellis CL, Greenfield E, Barber M, Restivo VA Jr, Ke X, Gray GS et al (1995) CTLA4 mediates antigen-specific apoptosis of human T cells. Proc Natl Acad Sci U S A 92(3):811–815 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hallqvist JAA, Diderichsen F, Reuterwall C (1996) How to evaluate interaction between causes: a review of practices in cardiovascular epidemiology. J Intern Med 239(5):377–382 [DOI] [PubMed] [Google Scholar]
  15. Han J, Hankinson SE, Colditz GA, Hunter DJ (2004) Genetic variation in XRCC1, sun exposure, and risk of skin cancer. Br J Cancer 91(8):1604–1609 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hoek JB, Pastorino JG (2002) Ethanol, oxidative stress, and cytokine-induced liver cell injury. Alcohol 27(1):63–68 [DOI] [PubMed] [Google Scholar]
  17. Hosmer DW, Lemeshow S (1992) Confidence interval estimation of interaction. Epidemiology 3(5):452–456 [DOI] [PubMed] [Google Scholar]
  18. Hua Z, Li D, Xiang G, Xu F, Jie G, Fu Z, Jie Z, Da P (2011) PD-1 polymorphisms are associated with sporadic breast cancer in Chinese Han population of Northeast China. Breast Cancer Res Treat 129(1):195–201. doi:10.1007/s10549-011-1440-3 [DOI] [PubMed] [Google Scholar]
  19. Hurchla MA, Sedy JR, Gavrieli M, Drake CG, Murphy TL, Murphy KM (2005) B and T lymphocyte attenuator exhibits structural and expression polymorphisms and is highly Induced in anergic CD4+ T cells. J Immunol 174(6):3377–3385 [DOI] [PubMed] [Google Scholar]
  20. Iwata A, Watanabe N, Oya Y, Owada T, Ikeda K, Suto A, Kagami S, Hirose K, Kanari H, Kawashima S, Nakayama T, Taniguchi M, Iwamoto I, Nakajima H (2010) Protective roles of B and T lymphocyte attenuator in NKT cell-mediated experimental hepatitis. J Immunol 184(1):127–133 [DOI] [PubMed] [Google Scholar]
  21. Jassam N, Bell SM, Speirs V, Quirke P (2005) Loss of expression of oestrogen receptor beta in colon cancer and its association with Dukes’ staging. Oncol Rep 14(1):17–21 [PubMed] [Google Scholar]
  22. Jonsson JJ, Foresman MD, Wilson N, McIvor RS (1992) Intron requirement for expression of the human purine nucleoside phosphorylase gene. Nucleic Acids Res 20(12):3191–3198 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kallinich T, Beier KC, Wahn U, Stock P, Hamelmann E (2007) T-cell co-stimulatory molecules: their role in allergic immune reactions. Eur Respir J 29(6):1246–1255 [DOI] [PubMed] [Google Scholar]
  24. Kesse E, Clavel-Chapelon F, Boutron-Ruault MC (2006) Dietary patterns and risk of colorectal tumors: a cohort of French women of the National Education System (E3N). Am J Epidemiol 164(11):1085–1093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kuznetsov VA, Makalkin IA, Taylor MA, Perelson AS (1994) Nonlinear dynamics of immunogenic tumors: parameter estimation and global bifurcation analysis. Bull Math Biol 56(2):295–321 [DOI] [PubMed] [Google Scholar]
  26. Latchman Y, Wood CR, Chernova T, Chaudhary D, Borde M, Chernova I, Iwai Y, Long AJ, Brown JA, Nunes R, Greenfield EA, Bourque K, Boussiotis VA, Carter LL, Carreno BM, Malenkovich N, Nishimura H, Okazaki T, Honjo T, Sharpe AH, Freeman GJ (2001) PD-L2 is a second ligand for PD-1 and inhibits T cell activation. Nat Immunol 2(3):261–268. doi:10.1038/85330 [DOI] [PubMed] [Google Scholar]
  27. Lechner D, Kallay E, Cross HS (2005) Phytoestrogens and colorectal cancer prevention. Vitam Horm 70:169–198 [DOI] [PubMed] [Google Scholar]
  28. Lee KM, Chuang E, Griffin M, Khattri R, Hong DK, Zhang W, Straus D, Samelson LE, Thompson CB, Bluestone JA (1998) Molecular basis of T cell inactivation by CTLA-4. Science 282(5397):2263–2266 [DOI] [PubMed] [Google Scholar]
  29. Majewski J, Ott J (2002) Distribution and characterization of regulatory elements in the human genome. Genome Res 12(12):1827–1836. doi:10.1101/gr.606402 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Marcus PM, Hayes RB, Vineis P, Garcia-Closas M, Caporaso NE, Autrup H, Branch RA, Brockmoller J, Ishizaki T, Karakaya AE, Ladero JM, Mommsen S, Okkels H, Romkes M, Roots I, Rothman N (2000) Cigarette smoking, N-acetyltransferase 2 acetylation status, and bladder cancer risk: a case-series meta-analysis of a gene-environment interaction. Cancer Epidemiol Biomarkers Prev 9(5):461–467 [PubMed] [Google Scholar]
  31. Martineti V, Picariello L, Tognarini I, Carbonell Sala S, Gozzini A, Azzari C, Mavilia C, Tanini A, Falchetti A, Fiorelli G, Tonelli F, Brandi ML (2005) ERbeta is a potent inhibitor of cell proliferation in the HCT8 human colon cancer cell line through regulation of cell cycle components. Endocr Relat Cancer 12(2):455–469 [DOI] [PubMed] [Google Scholar]
  32. Mojtahedi Z, Mohmedi M, Rahimifar S, Erfani N, Hosseini SV, Ghaderi A (2012) Programmed death-1 gene polymorphism (PD-1.5 C/T) is associated with colon cancer. Gene 508(2):229–232 [DOI] [PubMed] [Google Scholar]
  33. Motsinger AA, Ritchie MD (2006) Multifactor dimensionality reduction: an analysis strategy for modelling and detecting gene-gene interactions in human genetics and pharmacogenomics studies. Hum Genomics 2(5):318–328 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Murphy KM, Nelson CA, Sedy JR (2006) Balancing co-stimulation and inhibition with BTLA and HVEM. Nat Rev Immunol 6(9):671–681 [DOI] [PubMed] [Google Scholar]
  35. Nishimura H, Honjo T (2001) PD-1: an inhibitory immunoreceptor involved in peripheral tolerance. Trends Immunol 22(5):265–268 [DOI] [PubMed] [Google Scholar]
  36. Ostrov DA, Shi W, Schwartz JC, Almo SC, Nathenson SG (2000) Structure of murine CTLA-4 and its role in modulating T cell responsiveness. Science 290(5492):816–819 [DOI] [PubMed] [Google Scholar]
  37. Peggs KS, Quezada SA, Allison JP (2009) Cancer immunotherapy: co-stimulatory agonists and co-inhibitory antagonists. Clin Exp Immunol 157(1):9–19. doi:10.1111/j.1365-2249.2009.03912.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Qi P, Ruan CP, Wang H, Zhou FG, Xu XY, Gu X, Zhao YP, Dou TH, Gao CF (2009) CTLA-4 +49A> G polymorphism is associated with the risk but not with the progression of colorectal cancer in Chinese. Int J Colorectal Dis 25(1):39–45. doi:10.1007/s00384-009-0806-z [DOI] [PubMed] [Google Scholar]
  39. Scalapino KJ, Daikh DI (2008) CTLA-4: a key regulatory point in the control of autoimmune disease. Immunol Rev 223:143–155 [DOI] [PubMed] [Google Scholar]
  40. van der Merwe PA, Bodian D, Daenke S, Linsley P, Davis SJ (1997) CD80 (B7-1) Binds Both CD28 and CTLA-4 with a low affinity and very fast kinetics. J Exp Med 185:393–403 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Wang L, Li D, Fu Z, Li H, Jiang W (2007a) Association of CTLA-4 gene polymorphisms with sporadic breast cancer in Chinese Han population. BMC Cancer 7:173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Wang L, Li D, Fu Z, Li H, Jiang W, Li D (2007b) Association of CTLA-4 gene polymorphisms with sporadic breast cancer in Chinese Han population. BMC Cancer 7(1):173. doi:10.1186/1471-2407-7-173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Watanabe N, Gavrieli M, Sedy JR, Yang J, Fallarino F, Loftin SK, Hurchla MA, Zimmerman N, Sim J, Zang X, Murphy TL, Russell JH, Allison JP, Murphy KM (2003a) BTLA is a lymphocyte inhibitory receptor with similarities to CTLA-4 and PD-1. Nat Immunol 4(7):670–679 Epub 2003 Jun 2008 [DOI] [PubMed] [Google Scholar]
  44. Watanabe N, Gavrieli M, Sedy JR, Yang J, Fallarino F, Loftin SK, Hurchla MA, Zimmerman N, Sim J, Zang X, Murphy TL, Russell JH, Allison JP, Murphy KM (2003b) BTLA is a lymphocyte inhibitory receptor with similarities to CTLA-4 and PD-1. Nat Immunol 4(7):670–679 [DOI] [PubMed] [Google Scholar]
  45. Yang G, Shu XO, Li H, Chow WH, Cai H, Zhang X, Gao YT, Zheng W (2009) Prospective cohort study of soy food intake and colorectal cancer risk in women. Am J Clin Nutr 89(2):577–583 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Yusof AS, Isa ZM, Shah SA (2012) Dietary patterns and risk of colorectal cancer: a systematic review of cohort studies (2000-2011). Asian Pac J Cancer Prev 13(9):4713–4717 [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary material 1 (DOC 763 kb) (763KB, doc)

Articles from Journal of Cancer Research and Clinical Oncology are provided here courtesy of Springer

RESOURCES