Introduction

Among the various genomic abnormalities, allelic loss at human chromosome 8p21 is particularly frequent in all kinds of cancer and thus has received widespread attention in recent years1,2. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is a homotrimeric cytokine located at chromosome band 8p21. It has been suggested that TRAIL is a promising anticancer agent due to its critical regulatory role in apoptosis, a cell suicide mechanism with an important role in maintaining normal cell cycling and abrogating the unwanted or potentially threatening cells3,4. TRAIL binds to the TRAIL receptor 1 (TRAIL-R1), a gene also known as DR4 and TNFRSF10A5. TRAIL-R1 enables cell death and triggers apoptotic proteases to regulate apoptosis through inducing the oligomerization of intracellular death domains required for the apoptotic signal transduction and forming an extracellular cysteine-rich, ligand-binding domain6,7,8,9.

The polymorphic TRAIL-R1 encodes nearly 480 amino acids. Downregulation of TRAIL-R1 may accelerate tumor formation and progression. Previous work has reported a significant relevance of lowly expressed TRAIL-R1 to a variety of cancers and breast cancer cell lines9,10. The TRAIL-R1 mutation is a frequent event that has been associated with many types of human malignancy11,12. There are multiple well-characterized polymorphisms in the TRAIL-R1 gene, but the most extensively studied polymorphism has been the C > G substitution resulting in a threonine to arginine amino acid change in exon 4 (Thr209Arg, rs20575). Thr209Arg is of special interest in recent decade most likely due to the involvement in receptor ligand binding activity and stimulation of apoptotic pathways12. A great deal of attention has been directed to the testing of a hypothesis that Thr209Arg may modulate host susceptibility to cancer. However, the previous investigations, either in the form of genetic association study or meta-analysis, fail to provide compelling evidence13,14,15,16. The relatively small sample size may account a large part for the limited statistical power of these studies.

To determine whether Thr209Arg in the ectodomain of the TRAIL-R1 gene is independently associated with cancer, we conducted a meta-analysis where all usable data identified through several medicine-specific databases have been incorporated.

Materials and Methods

Search strategy, inclusion criteria and data extraction

Using the combinations of polymorphism, polymorphisms, variants, genotypes, TRAIL receptor 1, DR4 and cancer, we searched the PubMed database (http://www.ncbi.nlm.nih.gov/pubmed), in an effort to identify all relevant peer-reviewed literature published prior to February, 2014. To get additional usable data, we screened all articles assessing the effects of Thr209Arg on cancer and their references. A study was considered eligible if the association between Thr209Arg and cancer susceptibility was investigated, the study subjects were composed of cancer patients and well-matched healthy controls and count of Thr/Thr, Thr/Arg and Arg/Arg genotypes was clearly reported or information on genotype distribution was sufficiently provided in the research article. For the studies containing overlapped samples, the largest study with complete data was considered in the meta-analysis.

We collected data on authors, publication year, type of cancer, genotyping methods, study country, ethnicity of each population included and genotype frequency for each of the eligible studies. To maximize data accuracy, the information listed above was extracted independently by two investigators.

Statistical analysis

Based on genotypic and allelic data, we estimated cancer susceptibility (OR and 95% CI: odds ratio and 95% confidence interval) in relation to Thr209Arg for homozygous model, heterozygous model, allele frequency model, dominant model and recessive model, by applying a fixed or random effects meta-analysis. Stratified analysis was conducted by ethnicity, cancer type and Hardy-Weinberg equilibrium (HWE), to assess the association for each subgroup.

Heterogeneity across studies was checked by the χ2-based Q-test17, to determine whether the Mantel-Haenszel method (fixed effects model, FEM)18 or the DerSimonian and Laird method (random effects model, REM)19 was used to pool the data from the published studies. In case of absence of inter-study heterogeneity (P values >0.05), we chose the former method; otherwise, the latter method was applied for pooling purpose.

HWE was checked for the control group in each study using χ2 test20. Sensitivity analysis was performed by consecutively excluding every study to see whether the single data set had obvious influence on the combined ORs. The Egger regression test and Begg’s funnel plots were utilized to determine publication bias21.

Meta-analysis was performed using the software STATA 12.0 (Stata Corporation, College Station, TX, USA). A P value <0.05 was deemed statistically significant.

Results

Summary of study characteristics

We retrieved a total of 952 records matching pre-listed keywords. Title and abstract evaluation led to an elimination of 891 records. We then read the full text of all 61 articles and found 32 articles reported an association unrelated to the polymorphism being examined, 8 articles offered insufficient raw data, 4 articles were systematic reviews and 1 was case-only designed. After discarding all useless records, we at last included 16 articles13,14,15,22,23,24,25,26,27,28,29,30,31,32,33,34 (Fig. 1). Genotype and allele frequencies, along with main characteristics of the studies involved in this meta-analysis are detailed in Table 1. According to Table 1, breast cancer was the most studied cancer type, followed by lung cancer. Other cancers, such as cancers of bladder, hematological, gastric, ovarian, colorectal, liver and gallbladder were relatively less investigated and thereby merged into “other” category when performing meta-analysis. Caucasian and Asian ethnicities were all investigated, with Caucasian individuals outnumbering Asians. The genotype frequencies of Kuraoka et al. (2005) and Mittal et al. (2011) in control population were not in HWE, according to χ2 test.

Table 1 Characteristics of studies involved in this meta-analysis.
Full size table
Figure 1
figure 1

Flow diagram of study selection for meta-analysis.

Full size image

Meta-analysis

As shown in Table 2, there was no substantial inter-study heterogeneity and we hence selected the FEM for the calculation of pooled ORs. A fixed effects meta-analysis revealed that there was no overall association between Thr209Arg and cancer (homozygous model: OR 0.98, 95% CI 0.88–1.09; heterozygous model: OR 0.95, 95% CI 0.87–1.04; allele frequency model: OR 0.99, 95% CI 0.94–1.05; dominant model: OR 0.98, 95% CI 0.91–1.05; recessive model: OR 1.01, 95% CI 0.92–1.10, Fig. 2).

Table 2 Summary ORs (95% CI) for TRAIL-R1 Thr209Arg and cancer.
Full size table
Figure 2
figure 2

Meta-analysis using a fixed effects model for the association between cancer susceptibility and TRAIL-R1 Thr209Arg stratified by ethnicity (recessive model). OR: odds ratio; CI: confidence interval; I-squared: measure to quantify the degree of heterogeneity in meta-analyses.

Full size image

Similar results were seen when the data were stratified by ethnicity (Fig. 2), cancer type and HWE deviation (Table 2).

With the aid of sensitivity analysis, we found that the combined effects remained stable when excluding each study. Neither did we find any evidence of significant publication bias, by using the funnel plots and Egger’s test (the recessive model: P = 0.304, Fig. 3).

Figure 3
figure 3

Begg’s funnel plot of publication bias test (recessive model). Each point represents a separate study for the indicated association. Log (OR): natural logarithm of OR; horizontal line: mean effect size.

Full size image

Discussion

Apoptosis is a defence mechanism against the malignant progression of cancer. Resistance to apoptosis destroys the balance between cell death and growth, thus facilitating tumorigenesis. TRAIL-R1 is a transmembrane protein with a death domain essential for apoptotic regulation. Variations in this gene are proved detrimental, as these alternations suppress cell death and promote proliferation, two causes reported to account for increases in the likelihood of carcinogenesis35,36,37. A large body of research has focused on the role of TRAIL-R1 Thr209Arg polymorphism in predisposition to cancer. However, there is a lack of consistency in the reported results. Hazra et al. conducted a large-scale study linking Thr209Arg with bladder cancer, providing epidemiological data that Thr209Arg plays a major role in the development of bladder cancer22. A subsequent study of German samples reported a decreased susceptibility of hematological malignant diseases in relation to TRAIL-R1 polymorphic alleles23. Inconsistent with the former Germany study, Frank et al. genotyped 521 breast cancer cases and 1,100 control subjects and found an almost 4-fold increased susceptibility attributable to the carriage of 626Thr-683Ala haplotype, though Thr209Arg alone was not found to contribute towards incident breast cancer24. Similarly, the three most recent studies revealed substantially different findings, with Körner et al. and Rai et al. reporting 626Thr as an independent susceptibility factor for liver cancer13,15 and no associations between Thr209Arg and susceptibility to lung cancer, according to Taştemir-Korkmaz et al.14. The heterogeneity of the findings among investigations addressing the association of Thr209Arg polymorphism with cancer is biologically possible, as the etiology may vary widely due to the differences in cancer type. Another plausible explanation is related to the limited number of subjects in each published study. We here infer that the polymorphism being investigated may exert similar effects on all cancer types and a sufficiently large study is needed to test this inference.

To provide compelling evidence of the association between Thr209Arg and cancer susceptibility, we performed a meta-analysis on 4,673 cancer cases and 4,844 controls from a total of 16 publications. Overall analysis revealed that this polymorphism predisposed no host susceptibility to cancer. We then performed stratified analysis by ethnicity, cancer type and HWE deviation to estimate the association for each subgroup, failing to demonstrate any statistical evidence of a significant association related to Thr209Arg. Our observations are not in accordance with those reported in a previous study, in which the investigators included 2,941 cases and 3,358 controls and found a marginal association (OR = 0.77, 95% CI:0.65–0.91; OR =0.84, 95% CI:0.72–0.99)16. The null associations implicated in the current analysis where nine additional studies have been included highlights the importance of a sufficient sample in detecting the true polymorphism-cancer associations. Despite the wide discrepancy in sample size between the two meta-analyses, we cannot rule out one possibility that Thr209Arg is a low-penetrance polymorphism and its effects on overall cancer and specific subtypes merit further investigations.

It is reported that genetic alterations in the TRAIL-R1 gene lead to an impaired apoptotic mechanism, one of the prerequisites required for the development of cancer11,12. Functional studies have exhibited data on increased possibility to develop cancers of head and neck, lung and gastric attributed to the nucleotide substitution in ectodomain of TRAIL-R112. Several lines of evidence from candidate gene studies lend further support to the notion that TRAIL-R1 Thr209Arg represents an effect modifier for cancer. 209Arg/Arg genotype was shown to modulate bladder cancer susceptibility via mediating the capacity of receptor ligand complexes involved in apoptotic pathways23. In addition, the 209Thr allele modifies risk of breast cancer by regulating TRAIL binding efficiency24. According to these data, we hypothesize that Thr209Arg may confer host susceptibility to cancer. This hypothesis nevertheless remains to be tested.

We have to address the limitations of this analysis. First, most of the single studies included have a small number of individuals, making the total sample underpowered to detect the association between Thr209Arg and various types of cancer. Second, it is important to note that the vast majority of published studies employed samples of Caucasian ancestry, thus the estimation of cancer susceptibility in Asians may be derived by chance as a result of sample insufficiency. Third, similar to many meta-analyses, we categorized the populations into either Caucasian or Asian ethnicity, which may lead to overgeneralization in results. For example, although Thr209Arg does not modify cancer susceptibility in total Asian populations, but it may have effects on some specific populations, such as Chinese and Japanese. The above-mentioned shortcomings suggest the necessity of further studies.

To sum up, our meta-analysis indicated that TRAIL-R1 Thr209Arg polymorphism was not significantly associated with overall cancer susceptibility. Stratified analysis by ethnicity and cancer type yielded similar results. The present findings, along with those suggested in previous analyses, merit further investigation involving more cancer types and ethnic groups.

Additional Information

How to cite this article: Geng, P. et al. Genetic association between TRAIL-R1 Thr209Arg and cancer susceptibility. Sci. Rep. 5, 10382; doi: 10.1038/srep10382 (2015).