Impact of Caspase-8 (CASP8) -652 6N Del and D302H Polymorphisms on Prostate Cancer in Different Ethnic Groups

Background: Despite evidence suggesting roles for caspase-8 (CASP8) -652 6N del and D302H polymorphisms in prostate cancer (PCa), the association of these polymorphisms with PCa risk remains inconclusive. Therefore, a meta-analysis was performed to more precisely estimate the association of CASP8 -652 6N del and D302H polymorphisms with PCa susceptibility. Materials and Methods: A comprehensive literature search was conducted to identify all case-control studies of CASP8 D302H and -652 6N del polymorphisms and PCa risk. Odds ratios (ORs) and 95% confidence intervals (CIs) were used to assess the strength of the association and the precision of the estimate, respectively. Results: Nine -625 6N del studies and 4 D302H studies were included. CASP8 -652 6N del and D302H polymorphisms were not significantly associated with PCa risk in the overall analyses. However, in the subgroup analysis stratified by ethnicity, -625 6N del was significantly associated with PCa risk in the East Asian and Indian populations under the recessive model. Furthermore, the subgroup analysis strongly suggested that D302H was associated with lower PCa risk in the Non-Indian population under the dominant model. Conclusions: In our meta-analysis, ethnic-specific differences were evident in the association of CASP8 -625 6N del and D302H polymorphisms with PCa risk.


Impact of Caspase-8 (CASP8) -652 6N Del and D302H Polymorphisms on Prostate Cancer in Different Ethnic Groups
Cheng-Dong Zhang 1,3& , Hong-Tao Li 1& , Kun Liu 2& , Zhi-Di Lin 3 , Qi-Liu Peng 5 , Xue Qin 5 , Min He 1 , Hua Wu 1 , Zeng-Nan Mo 4 , Xiao-Li Yang 1,3 * susceptibility has been rapidly increasing (Zheng et al., 2008), and SNPs associated with tumor cell apoptosis have received the most attention. Caspases, a family of proteases, mediate the regulation of apoptosis. One of the initiator caspases, CASP8 (also known as FLICE), has been reported to play an important role in apoptosis (Ho and Hawkins, 2005). Given the role of CASP8 in apoptosis, genetic variants of CASP8 that can alter its function are likely to affect apoptosis in cancer cells.
Recently, many studies have reported that 2 common SNPs within the CASP8 promoter gene, -652 6N del and D302H, are associated with many cancer types. The -652 6N del polymorphism (rs3834129) is a 6-nucleotide insertion-deletion substitution at nucleotides 657_652. The D302H polymorphism (rs1045485) is a G-C substitution at 1207 bp, which results in an aspartic acid to histidine exchange at position 302 in the amino acid sequence. Sun et al. (2007) first reported that the -652 6N del variant reduced the risk of developing many types of cancer including lung, esophageal, gastric, colorectal, cervical, and breast cancers (Sun et al., 2007). However, in subsequent studies, demonstration of the protective effect of -652 6N del failed to be replicated in several cancer types, including breast and prostate tumors (Cybulski. et al., 2007;Haiman et al., 2008). Interestingly, the variant genotype (del/del) of -652 6N del exhibited a trend of increased PCa risk in a North Indian population (George et al., 2010;Kesarwani et al., 2011). In contrast, the same variant was highly associated with reduced PCa risk in a Chinese population (Fu et al., 2011). In contrast to -652 6N del, the D302H H variant has demonstrated a consistent protective effect for breast carcinoma risk (Sergentanis and Economopoulos, 2010). However, the association of D302H with PCa risk has been inconclusive. Lubahn et al. and Meyer et al. reported that the D302H polymorphism may play a protective role against PCa (Lubahn et al., 2010;Meyer et al., 2013); whereas, in a North Indian population, the H variant allele was associated with increased PCa risk (George et al., 2010).
Inconsistencies in the association of CASP8 polymorphisms with PCa risk may be attributable to different ethnicities. Therefore, we performed a metaanalysis to further evaluate the association of D302H and -652 6N del with PCa risk.

Literature search
We performed a literature search using the following keywords "CASP8 [MeSH]", "prostate cancer [MeSH]", and "polymorphism" or "variation" or "mutation" in PubMed, Cochrane Library, and Web of Science databases. All eligible studies were published before July 1st, 2014. The reference lists of review and retrieved articles were hand searched at the same time. Abstracts, unpublished reports, and articles written in non-English languages were not considered. When overlapping data of the same patient population were included in more than 1 publication (sending email to the communication author to confirm this), only the most recent or complete study was used in the meta-analysis.

Inclusion and exclusion criteria
The following inclusion criteria were used to select literature for the meta-analysis: (1) information on the evaluation of CASP8 -652 6N del or CASP8 D302H polymorphisms and PCa susceptibility; (2) case-control studies; and (3) studies with sufficient genotype data to calculate odd ratios (ORs) with 95% confidence intervals (CIs). Major criteria for study exclusion were: (1) no controls; (2) reviews and duplication of the previous publication; and (3) no usable data reported.

Data extraction
Two of the authors independently extracted the study information from all eligible publications based on the inclusion criteria above. Any discrepancies regarding study inclusion were resolved by discussion, and a third party was involved when necessary. The following characteristics were recorded from each study: the name of the first author, publication year, ethnicity of the population, and genotype frequencies of -652 6N del and D302H polymorphisms in both PCa cases and controls.
Ethnic descent was categorized into the following groups: Caucasian, Indian, East Asian, African, and other. For case-control studies, data were extracted separately for each group whenever possible.

Statistical analysis
Crude ORs with the corresponding 95% CIs were used to detect the strength of the association of -652 6N del and D302H polymorphisms with PCa risk. For both CASP8 polymorphisms, the relationship between genotypes and PCa susceptibility was evaluated using co-dominant, recessive, over-dominant, and dominant genetic models. The relationship between alleles and PCa susceptibility was also examined using an allelic model. A Z-test was used to determine the significance of the overall OR, and P<0.05 was considered statistically significant. Betweenstudy heterogeneity and between-study inconsistency were evaluated using Cochran's Q statistic and estimating I 2 , respectively (Higgins et al., 2003). P<0.10 was considered statistically significant heterogeneity, and Ι 2 was used to qualify variation in OR attributable to heterogeneity. In studies without substantial heterogeneity (Q-test P≥0.10 or I 2 <50%), the fixed effects model (Mantel-Haenszel method) was used to estimate the overall OR (Mantel and Haenszel, 1959). The random effects model (DerSimonian and Laird method) was used in studies with substantial heterogeneity (DerSimonian and Laird, 1986). In addition to the comparison among all subjects, stratification analyses by ethnicity were performed to explore the reasons for heterogeneity. To adjust for multiple comparisons, we applied the Benjamini-Hochberg (BH) method (Benjamini and Hochberg, 1995) and stepdown Bonferroni method (Holm, 1979), which control for false discovery rate (FDR) and familywise error rate (FWE), respectively. One-way sensitivity analyses were performed to assess the stability of the results. To investigate the potential publication bias, both visual funnel plot and Egger's linear regression test were applied, and P<0.05 was considered statistically significant. All statistical analyses were performed mainly using STATA version 12.0 (Stata Corp, College Station, TX, USA) and R package version 3.1.0.

Eligible studies
Based on our inclusion criteria, 13 eligible independent studies in 7 reports were identified. Three of the reports contained data on different ethnic groups or variants (Haiman et al., 2008;George et al., 2010;Lubahn et al., 2010). Thirteen available studies with 5423 cases and 6994 controls were finally included in the meta-analysis. Among these 13 studies, 9 studies were eligible for the -652 6N del polymorphism (3955 PCa cases, 4241 controls) and 4 studies were eligible for the D302H polymorphism (1468 PCa cases, 2753 controls). The characteristics of the eligible studies are summarized in Table 1 and Table  2, respectively.

Meta-analyses
In the overall analysis, the -652 6N del polymorphism  Figure  2 and Table 3). The pooled meta-analysis results for the D302H polymorphism were similar to those for the -652 6N del polymorphism; the D302H polymorphism was not significantly associated with PCa risk in the dominant (DH+HH vs. DD), recessive (DD vs. HH+DH), heterogeneous co-dominant (DH vs. DD), or   Figure 4). The detailed data are presented in Table 4.

Sensitivity analysis
Because eligible studies were not available for sensitivity analysis of the D302H polymorphism, a oneway sensitivity analysis across all ethnic groups was performed to evaluate the stability of the -652 6N del polymorphism data. The results of this analysis showed     DOI:http://dx.doi.org/10.7314/APJCP.2014.15.18.7713 Two CASP8 Polymorphisms and Prostate Cancer that statistical significance was not changed when any single study was omitted (data not shown). The results of the sensitivity analysis demonstrated that the -652 6N del polymorphism data in this meta-analysis were relatively stable and credible.

Publication bias
In the -652 6N studies, both the Begg's funnel plot ( Figure 5) and Egger's test were performed to estimate publication bias. Because of the limited number of D302H studies, only Egger's test was performed. The results did not show any statistical evidence of publication bias (all P>0.05). Results of the publication bias test are shown in Table 3.

Discussion
To date, our study is the first meta-analysis to determine the association of CASP8 -652 6N del and D302H SNPs with PCa. The association of CASP8 -652 6N del and D302H SNPs with PCa risk was evaluated in 8196 and 4201 subjects, respectively. The lack of significant results in the pooled analysis suggested the absence of robust associations between the 2 SNPs and PCa. In the pooled analysis, both SNPs showed substantial heterogeneity in all the genetic models. The subgroup analyses based on ethnicity yielded more homogeneous genetic effects than the pooled analyses. The sensitivity analysis of the -652 6N del polymorphism indicated that our results were relatively robust.
Nowadays a number of studies have investigated association between CASP8 gene polymorphisms and risk of cancer. It is reported that CASP8 -652 6N del and D302H polymorphisms play a vital role in diversity tumors, such as lung cancer (Son et al., 2006), breast cancer (Cybulski. et al., 2007;Frank et al., 2008) and colorectal cancer (Wu et al., 2013). Sun et al. reported that the del allele of -652 6N del reduces T lymphocyte apoptosis by inhibiting the binding site for the transcriptional activator stimulatory protein 1 in the CASP8 promoter (Sun et al., 2007). This function of -652 6N del has been associated with increased immune surveillance (Sun et al., 2007). Therefore, individuals with the del allele may be less susceptible to tumorigenesis. However, different ethnicities may have distinct genetic backgrounds; therefore, tumor susceptibility can be influenced by ethnicity (Hirschhorn et al., 2002). The subgroup analyses revealed that the association of -652 6N del with PCa susceptibility was paradoxical among the different ethnicities. The subgroup analysis stratified by ethnicity indicated that -652 6N del was associated with increased PCa risk in subjects of Indian descent and with reduced PCa risk in subjects of East Asian descent. The -652 6N del polymorphism was not significantly associated with PCa susceptibility in subjects of Caucasian descent under any genetic model. With regard to D302H polymorphism, its functional effect is yet unknown. Between mouse and human, D302H map to an evolutionarily conserved region. This suggests that this variant can directly influence PCa rather than depending on an unknown linkage disequilibrium causative variant (Bethke et al., 2009). It has been suggested that the D302H polymorphism may affect the interaction of CASP8 with regulatory proteins (MacPherson et al., 2004;Bethke et al., 2009). Previous meta-analyses of the association of the D302H polymorphism with cancer risk have been inconclusive (Bethke et al., 2009;Sergentanis and Economopoulos, 2010). Meanwhile, Michailidou (Michailidou et al., 2013) and Turnbull (Turnbull et al., 2010) indicated that D302H variant in CASP8 may be a false positive association through recent GWAS studies. In the present meta-analysis, the protective effect of the D302H polymorphism for PCa risk was observed only in the Non-Indian population under over-dominant, heterogeneous co-dominant, and dominant models.
Given the complex etiology of cancer, PCa risk cannot be attributed to a single factor. Demographic patterns, lifestyle, hormonal influences, inflammation, Statin drugs, cholesterol, and genome wide variation play a role in prostate carcinogenesis. Considering the multitude of factors associated with PCa risk, the varied results of CASP8 polymorphisms and PCa risk among the different ethnicities in our meta-analysis is not surprising. It's worth noting that the rare allele of -652 6N del and D302H of CASP8 all have a trend of an increased risk of PCa in the Indian population. Meanwhile the rare allele of -652 6N del has a positive association with PCa in the Indian population, but a negative observation in the East Asian population. Previous studies on genetic association suggested that the characteristics of Indian population might be different from other Asian populations. A metaanalysis of p53 codon 72 polymorphism conducted by Zhou et al. indicated that Pro carrier showed significant associations with increased risk of cervical cancer among Indian populations, but not among Chinese, Japanese and Korean populations (Zhou et al., 2012). However, the study on KCNQ1 rs2237892 C"T gene polymorphism and type 2 diabetes mellitus in the Asian population demonstrated that a significant association was found in Chinese, Korean and Malaysia population, but not in Indian population (Li et al., 2014). The prevalence of prostate cancer and the allele frequencies differ across populations (Zhao et al., 2013), and some genes are specific in region and race, any conclusion concerning genetic effects should be carefully interpreted.
The limitations of our meta-analysis should be noted. First, although 9 -652 6N del studies were included in the meta-analysis, the subgroup analysis was limited to 3 studies after stratification by ethnicity. The limited number of D302H studies in the subgroup analysis may have increased the risk of bias in the meta-analysis. Second, given that PCa is a multifactorial disease, confounding factors, such as age, tobacco smoking, PSA levels, clinical aggressiveness, and Gleason score, may have affected the associations between the 2 CASP8 polymorphisms and PCa risk. Finally, our analysis was based on unadjusted estimates. Adjusted ORs for the above risk factors should also be pooled to provide exact summary estimates. We were unable to perform these analyses because of the lack of individual data.
In conclusion, our meta-analysis demonstrated that the association of CASP8 -652 6N del and D302H polymorphisms with PCa risk may be influenced by ethnicity. The association of the -652 6N del polymorphism with PCa susceptibility was different among subjects of Asian descent. -652 6N del polymorphism was positively associated with PCa risk in the Indian population and negatively associated with PCa risk in the East Asian population. D302H polymorphism was associated with reduced PCa risk in the Non-Indian population. Given the limitations of our meta-analysis, confirmation of our findings in larger multicenter studies is warranted, especially between the Indian and East Asian populations.