The first 2 authors contributed equally to this work.
The functional cytotoxic T lymphocyte–associated Protein 4 49G-to-A genetic variant and risk of pancreatic cancer
Article first published online: 22 FEB 2012
Copyright © 2012 American Cancer Society
Volume 118, Issue 19, pages 4681–4686, 1 October 2012
How to Cite
Yang, M., Sun, T., Zhou, Y., Wang, L., Liu, L., Zhang, X., Tang, X., Zhou, M., Kuang, P., Tan, W., Li, H., Yuan, Q. and Yu, D. (2012), The functional cytotoxic T lymphocyte–associated Protein 4 49G-to-A genetic variant and risk of pancreatic cancer. Cancer, 118: 4681–4686. doi: 10.1002/cncr.27455
- Issue published online: 19 SEP 2012
- Article first published online: 22 FEB 2012
- Manuscript Accepted: 21 DEC 2011
- Manuscript Revised: 18 NOV 2011
- Manuscript Received: 24 AUG 2011
- pancreatic cancer;
- single-nucleotide polymorphism;
- association study
Antitumor T lymphocytes play an essential part in immune surveillance of cancer cells. Cytotoxic T lymphocyte–associated Protein 4 (CTLA-4) is a negative regulator of T cell activation and proliferation and therefore influences immune surveillance of carcinogenesis of pancreas. Thus, this study examined the association between functional CTLA-4 49G-to-A (49G>A) single-nucleotide polymorphism and pancreatic cancer risk.
Genotypes were determined in 368 patients with pancreatic cancer and 926 controls, and odds ratios (ORs) and 95% confidence intervals (CIs) were estimated by logistic regression.
A significant increased risk of pancreatic cancer was found to be associated with the CTLA-4 49G>A single-nucleotide polymorphism. Compared with noncarriers, the OR of developing pancreatic cancer for CTLA-4 49 GA or AA carriers was 1.75 (95% CI = 1.34-2.30, P = 4.83 × 10−5) or 2.54 (95% CI = 1.67-3.87, P = 1.36 × 10−5), respectively. In stratified analyses, the association was more pronounced in GA and AA carriers aged ≤60 years (OR = 3.10, 95% CI = 2.15-4.47, Pinteraction = .002), smokers with GA and AA genotypes (OR = 3.92, 95% CI = 2.39-6.43, Pinteraction = .037), and drinkers with GA and AA genotypes (OR = 4.55, 95% CI = 2.65-7.82, Pinteraction = .042), compared with GG carriers. Moreover, a supermultiplicative interaction between the CTLA-4 49AA genotype and smoking plus drinking was also evident in intensifying risk of pancreatic cancer (Pinteraction = 5.64 × 10−12).
These results suggest that CTLA-4 49G>A polymorphism is involved in susceptibility to developing pancreatic cancer, alone and in a gene–environment interaction manner. Cancer 2012. © 2012 American Cancer Society.
Cytotoxic T lymphocyte–associated Protein 4 (CTLA-4, also known as CD152) is a member of the immunoglobulin superfamily and is expressed on activated T cells.1, 2 Similar to CD28 (another T cell costimulatory molecule), CTLA-4 binds to B7.1 and B7.2 on antigen-presenting cells. However, the affinity of their interactions is much higher than those between CD28 and B7 molecules.3, 4 Through CTLA-4–B7 interaction, CTLA-4 can negatively regulate function and proliferation of T cells through depressing the expression of both interleukin-2 and interleukin-2 receptor and arresting T cells at the G1 phase.5 In addition, CTLA-4 can also induce FAS-independent apoptosis of activated T cells,6 which may further inhibit immune function of T lymphocytes. However, it is well known that antitumor T cells play a pivotal role in immune surveillance of cancer cells.7 In tumor-transplanted mice, injection with antibodies that block CTLA-4 function enhanced T cell activation,8 rejected several different types of tumors, and had long-lasting antitumor immunity,9 indicating the importance of CTLA-4 in carcinogenesis.
Pancreatic cancer, a highly lethal disease, is one of the leading causes of cancer-related death worldwide.10, 11 Because of the asymptomatic onset of pancreatic cancer, most patients already have metastatic or locally advanced diseases when being diagnosed, which results in poor prognoses. Although smoking, diabetes mellitus history, and perhaps alcohol drinking are risk factors for pancreatic carcinogenesis,12-15 only a fraction of exposed individuals develop pancreatic cancer in their lifetime. This phenomenon suggests that genetic susceptibility factors also play a part in cellular malignant transformation of pancreas. Preclinical and clinical trials have also demonstrated that CTLA-4 blockade may be a promising immunotherapeutic approach to treat patients with several advanced malignancies, including pancreatic cancer.16-19 Taken together, these evidences indicate that CTLA-4 may be a key molecule in oncogenesis of pancreatic cells.
In a previous study, we identified a 49G>A single-nucleotide polymorphism (SNP) in the coding region of CTLA-4, which causes an amino acid change from Ala17 to Thr17 in the leader sequence of CTLA-4 protein.20 Functional assays revealed that the CTLA-4 encoded by the CTLA-4 49A allele has enhanced affinity to bind B7.1 molecule and could increase inhibition of T cell activation and proliferation compared with that encoded by the CTLA-4 49G allele in vitro and in an ex vivo model. More importantly, the CTLA-4 49A allele appears to be associated with increased risk of multiple types of malignancies, including lung cancer, esophageal cancer, gastric cardia cancer, and breast cancer in a Chinese population. In addition, several meta-analyses demonstrated that the CTLA-4 49A allele is associated with an increased risk of cancer compared with the 49G allele, especially in Caucasians and Chinese.21-23 Until now, the potential role of this functional genetic variant in development of pancreatic cancer is still largely unknown. In view of the importance of CTLA-4 in pancreatic cancer, we performed a case–control study in a Chinese population to evaluate the association of this functional CTLA-4 49G>A polymorphism with risk of developing pancreatic cancer. In the stratified analysis, we also examined if there were gene–environment interactions between this genetic variant and sex, age, smoking, drinking, or diabetes mellitus history.
MATERIALS AND METHODS
This study consisted of 368 incident patients who had pancreatic cancer and 926 healthy cancer-free control individuals. All subjects were ethnic Han Chinese. Patients were recruited between June 2001 and May 2007 at the Peking Union Hospital and Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China. All patients with confirmed pancreatic ductal adenocarcinoma were enrolled with a response rate of 94%, and there was no sex and age restriction. The detailed diagnosis of patients was described previously.24, 25 Control subjects were cancer-free individuals, and they were selected from a community cancer-screening program for early detection of cancer conducted in the same regions during the same time period as the patients were collected. These control subjects were randomly selected from a pool of 2800 individuals on the basis of physical examination. The selection criteria included no individual history of cancer and frequency matching to cases by sex and age (±5 years). At recruitment, informed consent was obtained from each subject, and each participant was then interviewed to collect detailed information on demographic characteristics such as sex and age and related risk factors such as cigarette smoking, diabetes mellitus history, and alcohol drinking. This study was approved by the institutional review boards of the respective medical centers.
Genotypes of CTLA-4 49G>A polymorphism were determined by polymerase chain reaction–based restriction fragment length polymorphism assays as described previously.20 All subjects were successfully genotyped. To ensure quality control, genotyping was performed without knowledge of case/control status of the subjects. A 15% random sample of cases and controls was genotyped twice by different persons, and the reproducibility was 100%.
The Pearson chi-square test was used to examine the differences in demographic variables, smoking status, drinking status, diabetes mellitus history, and genotype distributions of CTLA-4 49G>A polymorphism between pancreatic cancer patients and controls. Associations between genotypes and risk of the development of pancreatic cancer were estimated by odds ratios (ORs) and their 95% confidence intervals (CIs) computed using an unconditional logistic regression model. Smokers were considered current smokers if they smoked up to 1 year before the date of cancer diagnosis or if they smoked up to 1 year before the date of the interview for control subjects. Information was collected on the number of cigarettes smoked per day, the age at which the subjects started smoking, and the age at which exsmokers stopped smoking. Subjects who never smoked or smoked less than 1 year before the date of cancer diagnosis for case patients or the date of interview for control subjects were defined as nonsmokers. The number of pack-years smoked was determined as an indication of the cumulative cigarette dose level: pack-years = (cigarettes per day/20) × (years smoked). Light and heavy smokers were categorized by using the 50th percentile pack-year value of the control subjects as the cutpoints (ie, ≤20 and >20 pack-years). Participants were classified as drinkers if they drank at least twice a week and continuously for at least 1 year during their lifetime; otherwise, they were defined as nondrinkers. All ORs were adjusted for age, sex, smoking status, drinking status, and diabetes mellitus history, where appropriate. A P value <.05 was used as the criterion of statistical significance, and all statistical tests were 2-sided. We tested the null hypotheses of multiplicative gene–environment interaction and evaluated departures from multiplicative interaction models25, 26 by including main effect variables and their product terms in the logistic regression model. All analyses were performed with Statistical Analysis System (version 9.0; SAS Institute, Cary, NC) and SPSS software package (version 16.0; SPSS Inc, Chicago, Ill).
No statistically significant differences were found between patients with pancreatic cancer and control subjects in terms of median age, sex distribution, and alcohol drinking status, suggesting that the frequency matching was adequate (Table 1). However, smokers were overrepresented in patients compared with controls (36.96% vs 26.89%, P < .001) although light or heavy smokers who smoked ≤20 or >20 pack-years were not significantly different (P = .578). In addition, history of diabetes mellitus was significantly higher in pancreatic cancer cases than in controls (17.12% vs 7.45%, respectively; P < .001).
|Variable||Patients (n = 368)||Controls (n = 926)||Pa|
|No. (%)||No. (%)|
|≤50||92 (25.00)||202 (21.81)|
|51-60||99 (26.90)||274 (29.59)|
|61-70||110 (29.89)||271 (29.27)|
|>70||67 (18.21)||179 (19.33)|
|Male||243 (66.03)||575 (62.10)|
|Female||125 (33.97)||351 (37.90)|
|Nonsmoker||232 (63.04)||677 (73.11)|
|Smoker||136 (36.96)||249 (26.89)|
|≤20||73 (53.68)||141 (56.63)|
|>20||63 (46.32)||108 (43.37)|
|No||259 (70.38)||659 (71.17)|
|Yes||109 (29.62)||267 (28.83)|
|Diabetes mellitus history||<.001|
|No||305 (82.88)||857 (92.55)|
|Yes||63 (17.12)||69 (7.45)|
Association Between CTLA-4 49G>A Polymorphism and Pancreatic Cancer Risk
Allele frequencies and genotype distributions of CTLA-4 49G>A variant in patients and controls are shown in Table 2. The allele frequencies for CTLA-4 49A were 0.28 in controls and 0.38 in patients. All observed genotype frequencies in both controls and patients conform to Hardy-Weinberg equilibrium. Distributions of these CTLA-4 49G>A variant genotypes were then compared among patients and controls. Frequencies of CTLA-4 49GG, GA, and AA genotypes among patients differed significantly from those among controls (chi-square = 25.00, P = 3.37 × 10−6, df = 2).
|Genotype||Patients (n = 368)||Controls (n = 926)||ORa (95% CI)||P|
|No. (%)||No. (%)|
|GG||140 (38.04)||482 (52.05)||1.00 (Reference)|
|GA||178 (48.37)||374 (40.39)||1.75 (1.34-2.30)||4.83 × 10−5|
|AA||50 (13.59)||70 (7.56)||2.54 (1.67-3.87)||1.36 × 10−5|
|A allele frequency||0.38||0.28|
Unconditional logistic regression analysis was used to estimate associations between genotypes of CTLA-4 49G>A variant and risk of pancreatic cancer (Table 2). The CTLA-4 49A allele was shown to be a risk allele; subjects having the AA or GA genotype had an OR of 2.54 (95% CI = 1.67-3.87, P = 1.36 × 10−5) or 1.75 (95% CI = 1.34-2.30, P = 4.83 × 10−5) for developing pancreatic cancer, respectively, compared with subjects having the GG genotype. Adjustment for sex, age, smoking, alcohol drinking, and diabetes mellitus history did not significantly affect respective ORs.
Stratified Analyses of Association Between CTLA-4 49G>A Polymorphism and Pancreatic Cancer Risk
The risk of pancreatic cancer associated with the CTLA-4 genotypes was further examined by stratifying for age, sex, smoking status, smoking level, alcohol drinking, and diabetes mellitus history (Table 3). A significantly increased risk of pancreatic cancer associated with the CTLA-4 49GA or AA genotype compared with the GG genotype was observed for the group aged 60 years or younger (OR = 3.10; 95% CI = 2.15-4.47; P = 1.50 × 10−9). However, this SNP was not associated with pancreatic cancer risk in the group aged older than 60 years (OR = 1.12; 95% CI = 0.78-1.61, P = .552). There was a significantly multiplicative gene–age interaction (Pinteraction = .002). Compared with the GG genotype, an increased risk of pancreatic cancer was only associated with 49GA or AA genotype for the male group (OR = 2.19; 95% CI = 1.58-3.03; P = 2.51 × 10−6), but not in female subjects (OR = 1.38; 95% CI = 0.87-2.21; P = .174). However, the gene–sex interaction was not statistically significant (Pinteraction = .109).
|GGa||GA+AAa||ORb (95% CI)||P|
|≤60||73/299||118/177||3.10 (2.15-4.47)||1.50 × 10−9|
|Male||91/314||152/261||2.19 (1.58-3.03)||2.51 × 10−6|
|Smoker||49/168||87/81||3.92 (2.39-6.43)||6.28 × 10−8|
|≤20||25/91||48/50||3.65 (1.90-7.02)||9.88 × 10−5|
|>20||24/77||39/31||5.17 (2.27-11.77)||9.30 × 10−5|
|Yes||36/178||73/89||4.55 (2.65-7.82)||4.29 × 10−8|
|Diabetes mellitus history||.411|
|No||120/446||185/411||1.75 (1.33-2.30)||5.73 × 10−5|
Because smoking, alcohol drinking, and diabetes mellitus history are predisposing factors for pancreatic cancer,12-15 we then examined whether a gene–environment interaction existed between the CTLA-4 49G>A polymorphism and these risk factors (Table 3). Elevated risk of the cancer associated with variant CTLA-4 genotypes was observed among both smokers and nonsmokers, both drinkers and nondrinkers, or both subjects with or without diabetes mellitus history, suggesting that this polymorphism is an independent risk factor. In smokers, compared with the CTLA-4 GG carriers, GA and AA carriers shown a 3.92-fold increased risk of developing pancreatic cancer (95% CI = 2.39-6.43, P = 6.28 × 10−8). However, only a 1.41-fold increased risk (95% CI = 1.03-1.93, P = .034) for nonsmoker individuals with GA and AA genotypes was observed compared with GG carriers. There was a statistically significant interaction between CTLA-4 GA and AA genotypes and smoking (Pinteraction = .037). Interestingly, this association was even more pronounced in heavy smokers (OR = 5.17; 95% CI = 2.27-11.77) compared with light smokers (OR = 3.65; 95% CI = 1.90-7.02). When subjects were stratified into drinkers and nondrinkers, compared with GG carriers, higher increased risk was found for GA and AA carriers who drink alcohol (OR = 4.55; 95% CI = 2.65-7.82; P = 4.29 × 10−8) than risk for GA and AA carriers who do not drink (OR = 1.42; 95% CI = 1.04-1.94; P = .027). A multiplicative gene–drinking interaction was also found with Pinteraction = .042. For diabetes mellitus history, OR of GA and AA genotypes (OR = 2.65, 95% CI = 1.23-5.75) was also higher in individuals with history of diabetes mellitus when compared with subjects without history of diabetes mellitus (OR = 1.75, 95% CI = 1.33-2.30). However, the interaction between diabetes mellitus history and CTLA4 genotype was not statistically significant (Pinteraction = .411).
Supermultiplicative Interaction Between CTLA-4 49G>A Polymorphism and Smoking Plus Drinking
Because a significantly increased risk of developing pancreatic cancer in either smokers or drinkers was observed, we then examined whether there is an interaction between CTLA-4 49G>A genotypes and smoking plus drinking (Table 4). Among individuals who were smokers and drinkers, participants with the CTLA-4 49AA genotype had an OR of 7.61 (95% CI = 1.92-30.14), which is 2.06-fold as high as the product (ie, 2.76 × 1.34 = 3.70) of the ORs for CTLA-4 49GG carriers who were smokers and drinkers and 49AA carriers who did not smoke and drink, suggesting a supermultiplicative interaction between the polymorphism and smoking plus drinking (Pinteraction = 5.64 × 10−12).
|Genotype||Nonsmoker and Nondrinker||Smoker and Drinker|
|Cases/Controls||ORa (95% CI)||Cases/Controls||ORa (95% CI)|
|GG||68/252||1.00 (Reference)||13/116||2.76 (1.38-5.54)|
|GA||81/265||1.10 (0.75-1.61)||23/46||1.42 (0.76-2.66)|
|AA||19/58||1.34 (0.74-2.45)||9/3||7.61 (1.92-30.14)b|
In this study, we examined whether the functional CTLA-4 49G>A genetic polymorphism is associated with risk of developing pancreatic cancer. On the basis of analysis of 368 patients and 926 control subjects, we demonstrated that the functional polymorphism has a significant impact on interindividual susceptibility to pancreatic cancer in a Han Chinese population. This genetic variant modifies the risk of pancreatic cancer not only by itself but also in a gene–environment interaction manner. To summarize, CTLA-4 49A allele is associated with significantly elevated risk for the development of pancreatic cancer compared with the 49G allele. This association is more pronounced in individuals aged 60 years or younger and males. In addition, the risk CTLA-4 49A allele genotypes appear to have multiplicative interaction with smoking or drinking in intensifying pancreatic cancer risk. This joint effect is more obvious in subjects who both smoke and drink.
The observations in this study are biologically plausible. Enhanced T cell activation and proliferation may prevent malignant cells to escape from cytotoxic T lymphocyte killing and therefore may inhibit cancer development.20 The CTLA-4 49G>A SNP weakens the binding affinity of CTLA-4 to B7.1, leading to attenuated CTLA-4–triggered inhibition of T cell activation and proliferation.20 This may be an underlying mechanism that contributes to less effective immune surveillance and increased susceptibility to the development of multiple types of common cancers among individuals carrying the CTLA-4 49A allele.20-23 The current study has extended our understanding to pancreatic cancer and further supports our hypothesis that the functional CTLA-4 49G>A polymorphism, which may influence immune status of an individual, can modify susceptibility to cancer.
We also observed a multiplicative interaction between the CTLA-4 49A allele and smoking or drinking. Such an interaction is within expectations, because smoking and drinking (especially high doses of alcohol consumption) are established risk factors for pancreatic cancer12, 13 and has a destructive effect on human immune responses.27, 28 It has been reported that cigarette smoking29 or chronic alcohol drinking30 might result in decreased T cell proliferation or suppression of T lymphocyte activities. Consequently, smoking or drinking and carrying the CTLA-4 49A allele simultaneously may lower immune surveillance to malignant cells, which could increase an individual's risk of developing pancreatic cancer. From this point of view, supermultiplicative interaction between CTLA-4 genotypes and the habits of smoking plus drinking, resulting in intensified risk of pancreatic cancer, is also biologically plausible.
In summary, to our knowledge, this study is the first to demonstrate that the functional CTLA-4 49G>A polymorphism is associated with risk of pancreatic cancer. The association manifests as a multiplicative gene–environment interaction between this polymorphism and smoking or drinking. These results are consistent with our initial findings in previous studies, further supporting the hypothesis that naturally occurring variants in T lymphocyte immune surveillance genes modify cancer susceptibility.
We thank Yuying Liu for assistance in recruiting the subjects and for technical support.
Supported by the Beijing Nova Program (grant 2010B013 to M. Yang), the Fundamental Research Funds for the Central Universities (grant ZZ1234 to M. Yang), National Natural Science Foundation of China (grant 81000872 to D. Yu), and the PhD Programs Foundation of the Chinese Ministry of Education (grant 20101106120011 to D. Yu).
CONFLICT OF INTEREST DISCLOSURE
The authors made no disclosure.