Validation of the novel susceptibility loci for prostate cancer in a Chinese population
- Authors:
- Published online on: December 13, 2017 https://doi.org/10.3892/ol.2017.7602
- Pages: 2567-2573
Abstract
Introduction
Prostate cancer (PCa) is the second most common cancer and one of the leading causes of mortality among males worldwide by 2012 (1). The incidence of PCa in China is considered reduced compared with that in Western countries; however, it has been progressively increasing over the past 30 years (2).
Genetic susceptibility to PCa has been well established and almost 100 common risk loci have been identified by genome wide association studies (GWAS) among European, African-American, Japanese and Chinese populations (3,4). However, only 10 of these loci were initially identified from GWAS in Japanese and Chinese populations. Among previous evaluation and validation studies, a part of the loci was revealed to be associated with PCa risk in Chinese population (5). Since these risk-associated single nucleotide polymorphisms (SNPs) exhibited a cumulative effect on PCa risk, the genetic risk scores (GRS) derived from PCa risk-associated SNPs were able to evaluate an individual's risk of PCa. The GRS based on the Chinese population is established and demonstrated to be a significant predictor of biopsy outcome in previous studies (5–9).
With an increasing sample size used in GWAS through combined data, a meta-analysis of a multi-ethnic population, which included 87,040 individuals, identified 23 new susceptibility loci for PCa (including 15 in European, 7 in multiethnic and 1 in the early onset analysis) (10). These PCa risk-associated alleles exhibited decreased effects with odds ratios (ORs) ranging between 1.06 and 1.14 (10). However, since the Chinese population was not included in the study, the effects of these 23 novel risk variants in individuals of Chinese descent remains unknown.
The objective of the present study was to evaluate the 23 newly identified susceptibility loci for PCa in a Chinese population and assess whether any validated loci contributed to the GRS in predicting the risk of PCa in a Chinese population.
Materials and methods
Population
The baseline characteristics of the present study subjects were summarized (Table I). A total of 2,425 subjects including 1,417 patients with PCa and 1,008 controls were recruited in the present study. All patients were part of the China PCa consortium from the southeast of China (11–13) recruited during January 2010 and December 2011, from which data were obtained. All cases were pathologically diagnosed with primary PCa and all the controls were recruited from the community or selected from subjects who had undergone routine physical examination in local hospitals. Written informed consent was obtained from subjects for their participation in the present study and a blood sample was taken from each subject at the time of recruitment for DNA extraction. The present study was reviewed and approved by the Institutional Review Board of every participating institution.
Genotyping and quality control
DNA samples were genotyped in the Center for Cancer Genomics at Wake Forest University (Winston-Salem, NC, USA) using the Illumina HumanOmniExpress BeadChips (Illumina, Inc., San Diego, CA, USA), which included 731,458 SNPs. For PCa risk-associated SNPs that were not included in the GWAS array, imputation was performed using IMPUTE 2.2.2 based on the combined data of the 1,000 Genomes project and HapMap3 data (14). A posterior probability of >0.9 was applied to call imputed genotypes. Imputed SNPs were excluded if they exhibited: i) A call rate <95%; ii) a minor allele frequency <0.05; or iii) P<1×10−3 in a Hardy-Weinberg equilibrium test in controls, as previously described (13).
Assessment of genetic risk
A GRS was calculated for each subject based on genotypes of the SNPs and weighted by their ORs and risk allele frequency, as described previously (15). GRS was calculated as
GRS=∏i=1nORigiWiwhere gi is the genotype of SNP i for an individual (0, homozygous of non-risk allele; 1, heterozygous; 2 homozygous of risk allele). ORi is the OR of SNP i estimated from external study (16), Wi is the average population risk of SNP i, calculated as Wi = fi2ORi2 + 2fi(1-fi) ORi + (1-fi)2, where fi is the risk allele frequency of SNP i based on the 1,000 Genome Project of the CHB (Han Chinese in Bejing, China) population (17). Therefore, a GRS value of 1.0 represents a population average risk.
Statistical analysis
A logistic regression model was used to analyze the association of each SNP with PCa risk, assuming an additive genetic model, which was implemented in PLINK version 1.07 (18). ORs and 95% confident intervals (CIs) were estimated from logistic regression analysis with adjustment for age and the highest eigen value. Student's t-tests were used to analyze the differences in means of normally distributed variables between 2 groups. For variables that were not normally distributed, 2 tests were performed: i) A nonparametric method using the Wilcoxon rank sum test and ii) Student's t-tests for different means between 2 groups following log-transformation. Differences in binary variables were investigated using χ2 tests. Area under the receiver operating characteristic curve (AUC) was used to evaluate the performance of GRS in discriminating between 2 groups of subjects. The difference between two AUCs was determined using Delong's test (19). P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using SPSS 19.0 (SPSS; IBM Corporation, Armonk, NY, USA).
Results
SNPs and PCa risk
The present study evaluated 19 newly reported SNPs. Among the 19 SNPs, only 1 (rs7153648 at 14q23) was associated with PCa risk in the China PCa cohort (OR=1.206, P<0.05). The direction of the effect was consistent with the previous multiethnic meta-analysis (10). The other 18 SNPs that had previously demonstrated genome-wide significance in European ancestry meta-analysis and multiethnic meta-analysis (Table II) either were not associated with PCa risk or did not demonstrate the same magnitude of effect in the Chinese population investigated in the present study.
SNPs and PCa aggressiveness
The association between the 19 SNPs and PCa aggressiveness was also investigated (cases with a Gleason score ≥7; Table III). The results did not demonstrate a significant association between rs7153648 and PCa aggressiveness, whereas rs636291 at 1p23 was significantly associated with PCa aggressiveness (OR=1.123, P<0.05).
SNPs, GRS and PCa
GRS was calculated using rs7153648 and 29 previously implicated SNPs (10). The mean GRS based on the 30 SNPs (GRS30) was significantly increased in patients with PCa compared with that among non-PCa individuals (1.439 vs. 0.961, P=7.44×10−41; Table IV). As reported in a previous study, it would be more efficient and reliable to calculate GRS using race-specific disease-associated SNPs that demonstrated genome-wide significance (20). Therefore, in the present study, GRS was also calculated based on the 9 strongest SNPs previously reported in individuals of Asian descent (GRS9; Table V) (16). The mean GRS based on 9 SNPs was 1.26 in patients with PCa and 0.99 in non-PCa controls (P=3.71×10−28).
Following adjustment for age (Table IV), GRS9 and GRS30 remained significantly associated with PCa (all P<0.01). The OR of the GRS30 for the prediction of PCa risk was 2.25 (95% CI, 1.976–2.598; P=2.97×10−31), decreased compared with that of GRS9 (OR=2.468; 95% CI, 2.053–2.967; P=6.9×10−22), although no significant differences were identified. When comparing the predictive ability of the GRS9 and GRS30, the AUCs were similar (0.792 for GRS9 vs. 0.7994 for GRS30, P=0.138).
Discussion
Genetic susceptibility is a major risk factor for PCa and is estimated to account for 42% of variation in the disease (21). In the past few years, GWAS and meta-analysis of combined data have identified 99 genomic variants associated with PCa in multiple populations of European, African-American, Japanese, Latino and Chinese ancestry (10). In the present study, 23 novel susceptibility loci detected in European ancestry or multi-ethnic analysis were investigated and their association in a Chinese population was evaluated.
Of the 19 SNPs evaluated in the present study, only 1 was identified to be associated with PCa. The estimate of risk of this SNP in the Chinese population was similar to that in European and multi-ethnic populations (10). Despite reaching genome-wide significance in European or multi-ethnic populations, the other 18 loci were not identified to be significant in the population of the present study. The discrepancy may be explained in multiple ways. First, since the effects of 18 SNPs (not including rs636291 at 1p36) were relatively low, with ORs ranging between 1.06 and 1.13, the present study may not possess the power to identify the small effects of these SNPs. This was also one of the reasons why rs7153648 did not reach a significant level following Bonferroni correction (P=0.05/19). Second, the risk allele frequencies in European and Chinese ancestry differed between SNPs evaluated (Table II); this difference may also influence the detection of significant effects of these SNPs in populations of Chinese ancestry. Finally, besides the different genetic backgrounds between European ancestry (or other populations) and Chinese ancestry, environmental factors, dietary-habit and other non-genetic factors may also affect the penetrance of these alleles, which may result in the difference of risk profiles.
When evaluating the association between the 19 SNPs and aggressive PCa (Gleason score, ≥7), the results demonstrated that rs636291 at 1p36 reached a significant level (P<0.05). This SNP reached genome-wide significance in early onset disease in European ancestry (10); however, a similar analysis could not be performed in the present study due to the lack of cases (only 34 patients with PCa were diagnosed <55 years of age). Nevertheless, this result may indicate that this risk variant was associated with more advanced PCa and should be further validated in an independent study.
In the comparison of the two GRS-based risk models, the results revealed that the performance was approximately the same between the two models. This may be attributed to the fact that certain risk variants were not strongly associated with PCa and others conferred a decreased effect to the risk of PCa in Chinese population compared with that in European whites. In a previous study, the plateau effect of PCa risk-associated SNPs was evaluated in predicting PCa in a Chinese population and it was identified that the predictive performance increased when the top 13 highest impact PCa risk-associated SNPs were included in the GRS (9). The results were similar in the present study; therefore, this may indicate that further SNPs weakly associated with PCa may not improve the predictive performance of GRS for PCa. Therefore, GRS only including the strongest SNPs may be appropriate while balancing the predictive performance and economic benefit.
In the present study, the variant rs7153648 at 14q23 that we demonstrated to be associated with PCa is located in the intergenic region of SIX homeobox 1. The regional information of the confirmed SNP (rs7153648) was presented (Fig. 1). In the LocusZoom plots of this loci, multiple SNPs located upstream of rs7153648 demonstrated marked association (P<0.01) but a weak correlation (dark blue circles), which may suggest the presence of multiple potential independent association signals. Variant rs636291 at 1p36, which was associated with early-onset PCa in European ancestry and was identified to be associated with aggressive PCa in Chinese ancestry in the current study, is located in intron 2 of peroxisomal biogenesis factor 14 and is associated with a variant (rs616488) reported in a GWAS of breast cancer (22).
There were multiple limitations to the present study. First, only 19 SNPs, rather than 23 of the novel identified loci, were genotyped or imputed due to 4 SNPs not being included in the GWAS panel and failing to impute using the CHB population of the 1,000 Genome project. Among the 4 SNPs, rs80130819 at 12q13 was not polymorphous in the CHB population, while the remaining 3 were polymorphous in the CHB population. Second, due to the open nature of the China PCa cohort, the clinical characterization of the cases was not consistent between distinct hospitals (e.g., Gleason score diagnosis in the present study), which limited further analysis of clinical phenotypes.
To conclude, by evaluating 19 PCa risk-associated SNPs identified in a large meta-analysis of GWAS from a European and multiethnic population, the results of the present study identified 1 SNP that was associated with PCa risk and 1 that was associated with aggressive PCa in a Chinese population. However, the validated small-effect SNP and other SNPs that weakly associated with PCa are not likely to improve the predictive ability of existing GRS in Chinese populations.
Acknowledgements
The present study was supported by the National Natural Science Foundation of China (grant nos. 81202269 and 81402339), the Clinical Science and Technology Innovation Project of Shanghai Shen Kang Hospital Development Center (grant no. SHDC12015105) and Scientific Research Project supported by Huashan Hospital, Fudan University (grant no. 2016QD079).
References
Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J and Jemal A: Global cancer statistics, 2012. CA Cancer J Clin. 65:87–108. 2015. View Article : Google Scholar : PubMed/NCBI | |
Liu M, Wang JY, Zhang YG, Zhu SC, Lu ZH and Wan B: Detection of urological and male genital tumors diagnosed in Beijing Hospital 1995–2004. Zhonghua Yi Xue Za Zhi. 87:2423–2425. 2007.(In Chinese). PubMed/NCBI | |
Xu J, Sun J and Zheng SL: Prostate cancer risk-associated genetic markers and their potential clinical utility. Asian J Androl. 15:314–322. 2013. View Article : Google Scholar : PubMed/NCBI | |
Eeles R, Goh C, Castro E, Bancroft E, Guy M, Al Olama AA, Easton D and Kote-Jarai Z: The genetic epidemiology of prostate cancer and its clinical implications. Nat Rev Urol. 11:18–31. 2014. View Article : Google Scholar : PubMed/NCBI | |
Na R, Liu F, Zhang P, Ye D, Xu C, Shao Q, Qi J, Wang X, Chen Z, Wang M, et al: Evaluation of reported prostate cancer risk-associated SNPs from genome-wide association studies of various racial populations in Chinese men. Prostate. 73:1623–1635. 2013.PubMed/NCBI | |
Aly M, Wiklund F, Xu J, Isaacs WB, Eklund M, D'Amato M, Adolfsson J and Grönberg H: Polygenic risk score improves prostate cancer risk prediction: Results from the Stockholm-1 cohort study. Eur Urol. 60:21–28. 2011. View Article : Google Scholar : PubMed/NCBI | |
Kader AK, Sun J, Reck BH, Newcombe PJ, Kim ST, Hsu FC, D'Agostino RB Jr, Tao S, Zhang Z, Turner AR, et al: Potential impact of adding genetic markers to clinical parameters in predicting prostate biopsy outcomes in men following an initial negative biopsy: Findings from the REDUCE trial. Eur Urol. 62:953–961. 2012. View Article : Google Scholar : PubMed/NCBI | |
Jiang H, Liu F, Wang Z, Na R, Zhang L, Wu Y, Zheng J, Lin X, Jiang D, Sun J, et al: Prediction of prostate cancer from prostate biopsy in Chinese men using a genetic score derived from 24 prostate cancer risk-associated SNPs. Prostate. 73:1651–1659. 2013.PubMed/NCBI | |
Ren S, Xu J, Zhou T, Jiang H, Chen H, Liu F, Na R, Zhang L, Wu Y, Sun J, et al: Plateau effect of prostate cancer risk-associated SNPs in discriminating prostate biopsy outcomes. Prostate. 73:1824–1835. 2013. View Article : Google Scholar : PubMed/NCBI | |
Al Olama AA, Kote-Jarai Z, Berndt SI, Conti DV, Schumacher F, Han Y, Benlloch S, Hazelett DJ, Wang Z, Saunders E, et al: A meta-analysis of 87,040 individuals identifies 23 new susceptibility loci for prostate cancer. Nature Genetics. 46:1103–1109. 2014. View Article : Google Scholar : PubMed/NCBI | |
Liu F, Hsing AW, Wang X, Shao Q, Qi J, Ye Y, Wang Z, Chen H, Gao X, Wang G, et al: Systematic confirmation study of reported prostate cancer risk-associated single nucleotide polymorphisms in Chinese men. Cancer Sci. 102:1916–1920. 2011. View Article : Google Scholar : PubMed/NCBI | |
Wang M, Liu F, Hsing AW, Wang X, Shao Q, Qi J, Ye Y, Wang Z, Chen H, Gao X, et al: Replication and cumulative effects of GWAS-identified genetic variations for prostate cancer in Asians: A case-control study in the ChinaPCa consortium. Carcinogenesis. 33:356–360. 2012. View Article : Google Scholar : PubMed/NCBI | |
Xu J, Mo Z, Ye D, Wang M, Liu F, Jin G, Xu C, Wang X, Shao Q, Chen Z, et al: Genome-wide association study in Chinese men identifies two new prostate cancer risk loci at 9q31.2 and 19q13.4. Nat Genet. 44:1231–1235. 2012. View Article : Google Scholar : PubMed/NCBI | |
Marchini J, Howie B, Myers S, McVean G and Donnelly P: A new multipoint method for genome-wide association studies by imputation of genotypes. Nat Genet. 39:906–913. 2007. View Article : Google Scholar : PubMed/NCBI | |
Wu Y, Zhang N, Li K, Chen H, Lin X, Yu Y, Gou Y, Hou J, Jiang D, Na R, et al: Genetic scores based on risk-associated single nucleotide polymorphisms (SNPs) can reveal inherited risk of renal cell carcinoma. Oncotarget. 7:18631–18637. 2016. View Article : Google Scholar : PubMed/NCBI | |
Wang M, Takahashi A, Liu F, Ye D, Ding Q, Qin C, Yin C, Zhang Z, Matsuda K, Kubo M, et al: Large-scale association analysis in Asians identifies new susceptibility loci for prostate cancer. Nat Commun. 6:84692015. View Article : Google Scholar : PubMed/NCBI | |
1000 Genomes Project Consortium, . Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA and Abecasis GR: A global reference for human genetic variation. Nature. 526:68–74. 2015. View Article : Google Scholar : PubMed/NCBI | |
Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ and Sham PC: PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 81:559–575. 2007. View Article : Google Scholar : PubMed/NCBI | |
DeLong ER, DeLong DM and Clarke-Pearson DL: Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics. 44:837–845. 1988. View Article : Google Scholar : PubMed/NCBI | |
Na R, Ye D, Qi J, Liu F, Lin X, Helfand BT, Brendler CB, Conran C, Gong J, Wu Y, et al: Race-specific genetic risk score is more accurate than nonrace-specific genetic risk score for predicting prostate cancer and high-grade diseases. Asian J Androl. 18:525–529. 2016. View Article : Google Scholar : PubMed/NCBI | |
Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A and Hemminki K: Environmental and heritable factors in the causation of cancer-analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med. 343:78–85. 2000. View Article : Google Scholar : PubMed/NCBI | |
Michailidou K, Hall P, Gonzalez-Neira A, Ghoussaini M, Dennis J, Milne RL, Schmidt MK, Chang-Claude J, Bojesen SE, Bolla MK, et al: Large-scale genotyping identifies 41 new loci associated with breast cancer risk. Nat Genet. 45:353–361. 2013. View Article : Google Scholar : PubMed/NCBI |