Genetic variants in NF-κB binding sites are associated with clinical outcomes in
prostate cancer patients
Shu-Pin Huanga,b, Victor C. Linc,d , Yung-Chin Leea,b, Chia-Cheng Yue,f,g, Chao-Yuan Huangh , Ta-Yuan Changi, Hong-Zin Leej, Shin-Hun Juangj, Te-Ling Luj, and Bo-Ying Baoj,k,＊
aDepartment of Urology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan
bDepartment of Urology, Faculty of Medicine, College of Medicine, Kaohsiung Medical University,
Kaohsiung, Taiwan
cDepartment of Urology, E-Da Hospital, Kaohsiung, Taiwan
dDepartment of Healthcare Administration, I-Shou University, Kaohsiung, Taiwan
eDivision of Urology, Department of Surgery, Kaohsiung Veterans General Hospital, Kaohsiung,
Taiwan
fDepartment of Urology, School of Medicine, National Yang-Ming University, Taipei, Taiwan gDepartment of Pharmacy, Tajen University, Pingtung, Taiwan
hDepartment of Urology, National Taiwan University Hospital, Taipei, Taiwan
iDepartment of Occupational Safety and Health, China Medical University, Taichung, Taiwan jDepartment of Pharmacy, China Medical University, Taichung, Taiwan
kSex Hormone Research Center, China Medical University Hospital, Taichung, Taiwan
S.-P. Huang and B.-Y. Bao contributed equally to this work.
*Corresponding Author:
Bo-Ying Bao, Department of Pharmacy, China Medical University, 91 Hsueh-Shih Road, Taichung 40402, Taiwan. Phone: 886-4-22053366 ext. 5126; Fax: 886-4-22031075; E-mail: [email protected]
Key words: Prostate cancer; single-nucleotide polymorphism; NF-κB; outcomes
Abbreviations: SNP, single-nucleotide polymorphism; RP, radical prostatectomy, ADT, androgen-deprivation therapy; PCSM, prostate cancer-specific mortality; ACM, all-cause mortality; PSA, prostate-specific antigen; OR, odds ratio; 95% CI, 95% confidence interval; HR, hazard ratio
Abstract
NF-κB transcription factors have been suggested to be involved in prostate cancer progression. Activated NF-κB translocates to the nucleus, binds to NF-κB binding sites and regulates target gene expression, leading to the given physiological response. It was hypothesized that the sequence variants in NF-κB binding sites might affect prostate cancer progression. We systematically evaluated 15 single-nucleotide polymorphisms (SNPs) within NF-κB binding sites those were predicted using a genome-wide database in a cohort of 1,024 prostate cancer patients. Associations of these SNPs with prostate cancer characteristics and clinical outcomes after radical prostatectomy (RP) for localized disease, and after androgen-deprivation therapy (ADT) for advanced disease were assessed by Kaplan-Meier analysis and Cox regression model. We found that PSMD7 rs2387084 and
MYCN rs1429409 were significantly related to earlier onset of prostate cancer and advanced clinical
stage, respectively. No SNPs significantly associated with disease recurrence after RP. Four and three SNPs were notably associated with prostate cancer-specific mortality (PCSM) and all-cause mortality (ACM), respectively, after ADT. LSAMP rs13088089, CCL17 rs223899, PSMD7 rs2387084 and MON1B rs284924 remained the significant predictors for PCSM while PSMD7 rs2387084 remained a significant predictor for ACM in multivariate models including clinical predictors. Moreover, we also noted that there were strong effects of the combined genotype on PCSM and patients with a greater number of unfavorable genotypes had a shorter time to PCSM during ADT (P for trend < 0.001). It was concluded that SNPs inside NF-κB binding sites might be useful to improve outcome prediction in prostate cancer patients.
1. Introduction
Prostate cancer, the most commonly diagnosed non-skin cancer, had become the second leading cause of cancer deaths in the United States in 2011.1 While treatment options for organ confined
prostate cancer includes radical prostatectomy (RP) and radiation therapy, the common management for advanced prostate cancer is androgen-deprivation therapy (ADT). Unfortunately, the recurrence rate in patients after ADT was considerably high even though ADT has positive effectiveness in initial treatment. In order to improve the efficacy prediction and to clarify the pathways governing prostate cancer progression, identification of prognostic molecular biomarkers for disease progression after treatment modalities is necessary.
The transcription factor NF-κB, a crucial regulator of immune responses, inflammation and developmental processes, has recently been proposed as a potential target for therapeutic intervention in prostate cancer. Upon stimulation, the IκB becomes phosphorylated and subsequently degrades by the proteasome. Released NF-κB translocates into the nucleus where it binds to the cognate binding sites of its target genes. The activated NF-κB communicates with the general transcription apparatus to regulate target gene expression and to exert its physiological functions ultimately. Nuclear localization of NF-κB has been detected in prostate cancer cells and the degree of nuclear NF-κB correlated closely with the tumor grade.4 In addition, patients with elevated NF-κB also have a worse
prognosis,5 suggesting the significant role of NF-κB signaling in prostate cancer progression.
Given that the sequence variants within NF-κB binding sites might affect the NF-κB binding and result in altered target gene expression/function, we systematically performed a genome-wide search for single-nucleotide polymorphisms (SNPs) within putative NF-κB binding sites and evaluated their associations with prostate cancer characteristics as well as clinical outcomes after RP and ADT.
2. Patients and Methods
A total of 1,024 patients who were diagnosed as having prostate cancer between 1995 and 2009, and had blood collected for research purposes, was recruited from Kaohsiung Medical University Hospital, Kaohsiung Veterans General Hospital and National Taiwan University Hospital with informed written consent (Table 1). Their clinical data and outcomes were obtained from the clinical records and pathological reports. This study was approved by the Institutional Review Board of the three recruitment hospitals.
In this cohort, 321 patients with localized prostate cancer who underwent RP as initial treatment and 605 patients with advanced prostate cancer who underwent ADT were followed prospectively and have been described previously.6-12 In the RP subcohort, prostate-specific antigen (PSA)
recurrence was defined as two consecutive PSA measurements of >0.2 ng/mL at an interval of >3 months,13 and the date of recurrence was determined as the PSA level of >0.2 ng/mL at the first
follow-up. No PSA recurrence was defined as PSA persistently <0.2 ng/mL during the post-operative follow-up period. One hundred nineteen (37.1%) patients experienced PSA recurrence during the mean follow-up of 38.5 months (range, 1-120 months). High preoperative PSA level (>20 ng/mL), high pathologic Gleason score (8-10), locally advanced pathologic stage (T3/T4/N1), and positive surgical margin were significantly associated with shorter time to PSA recurrence after RP (data not shown).
In the ADT subcohort, disease progression was defined as a serial rise in PSA, at least two rises in PSA (>1 week apart), and greater than the PSA nadir.14 Initiation of secondary hormone treatment for
rising PSA was also considered as a progression event. Time to progression was determined as the interval in months between the initiation of ADT and progression. Prostate cancer-specific mortality (PCSM) and the all-cause mortality (ACM) were defined as the intervals from the initiation of ADT to death from prostate cancer and to death from any cause, respectively. Overall, 416 patients progressed, 141 died, and 98 died from prostate cancer during a mean follow-up of 39 months (range, 3-125 months). Gleason Scores 8-10, metastatic stage of the disease, higher PSA nadir, and shorter time to PSA nadir were significantly associated with shorter time to progression, PCSM, and
ACM. Age at diagnosis was only associated with ACM, and PSA level at ADT initiation was associated with shorter time to PCSM and ACM, but not time to progression (data not shown). Written informed consent was obtained from each participant, and this study was approved by the Institutional Review Board of the three recruitment hospitals.
2.2. SNP selection and genotyping
Since transcription factors are known to regulate different genes in varied cellular contexts,15 we
used a genome-wide in silico prediction database of transcription factor binding sites instead of the chromatin immunoprecipitation data. We used PReMod (genomequebec.mcgill.ca/PReMod), a genome-wide cis-regulatory module prediction database, to identify computationally all putative NF-κB binding sites in the human genome.16 PReMod used TRANSFAC version 7.2 position weight
matrices (PWMs) to score the putative transcription factor binding sites based on how faithfully the binding site in human and its orthologs, in mouse and rat, match the PWM. In addition, the prediction algorithm of PReMod exploits the observation that many known cis-regulatory modules often contain clusters of phylogenetically conserved and repeated transcription factors binding sites,17
suggesting it is more reliable than other algorithms. The PReMod algorithm predicted that a total of 5,175 sites within the human genome were bound by the NF-κB (canonical NF-κB PWM: M00054; consensus: GGGAMTTYCC).18 We further identified SNPs within putative NF-κB binding sites by
comparing NF-κB binding sites with HapMap SNPs CHB (Han Chinese in Beijing, China) data in the UCSC table browser (NCBI35/hg17). SNPs with a minor allele frequency less of than 0.100 in the HapMap CHB population were excluded, thus leaving 19 SNPs in putative NF-κB binding sites for subsequent analysis.
Genomic DNA was extracted from peripheral blood using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA) and stored at -80°C for this study. Genotyping was performed using Sequenom iPLEX matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry technology at the National Center for Genome Medicine, Academia Sinica, Taiwan as
described previously.6 The average genotype call rate for these SNPs was 96.2% and the average
concordance rate was 99.8% among 9 blind duplicated quality control samples. Any SNP that did not conform to the Hardy-Weinberg equilibrium (P < 0.005) or fell below a genotyping call rate of 90% was removed (n = 4). Finally, a total of 15 SNPs were selected for further statistical analysis.
2.3. Statistical analysis
The odds ratios (ORs) and their 95% confidence intervals (95% CIs) were estimated using logistic regression to investigate the associations of genotypes with age (≤70 versus >70 years), Gleason score (>7 versus ≤7) and clinical stage at diagnosis (T3/T4/N1/M1 versus T1/T2). The linear regression was used to determine the effects of genotypes on PSA level at diagnosis. Because the function of the SNPs remains unknown, the Kaplan-Meier analysis with log-rank test under different genetic models, including dominant, recessive and trend models, was used to first assess the associations of SNPs with PSA recurrence after RP, disease progression, PCSM and ACM after ADT. Tests for trend were performed by modeling the number of minor alleles (0, 1 or 2) as a continuous variable. Cox proportional hazards regression was conducted on each SNP as an isolated covariate, with adjustment for clinicopathologic variables, as previously described. In the RP subcohort, the multiple explanatory variables included the known prognostic factors including age, PSA level, pathologic Gleason score, pathologic stage and surgical margin. In the ADT subcohort, the multiple explanatory variables included the known prognostic factors including age, PSA at ADT initiation, biopsy Gleason score, clinical stage, PSA nadir, time to PSA nadir and treatment modality. Statistical Package for the Social Sciences (SPSS) version 19.0.0 (IBM, Armonk, NY, USA) was used for other statistical analyses. A two-sided P value of <0.05 was considered statistically significant. The false-discovery rates (q values) using the R q value package (genomics.princeton.edu/storeylab/qvalue/) were calculated to determine the degree to which the tests for association were tend to false-positives.21
3. Results
3.1. Association between SNPs in NF-κB binding sites and clinical characteristics of prostate cancer
Association of SNPs in NF-κB binding sites with age, biopsy Gleason score, clinical stage and PSA at diagnosis were shown in Supplementary Table 1. Under a dominant, recessive or additive model, PSMD7 rs2387084 minor allele G was found to increase the risk of developing prostate cancer at or before 70 years (P ≤ 0.044, Table 2). We also classified prostate cancer patients as either having evidence of extraprostatic or metastatic disease (T3/T4/N1/M1) or localized disease (T1/T2) at diagnosis. Compared with the MYCN rs1429409 GG genotype, patients carrying one or two minor allele A had an OR of 1.33 (95% CI 1.02-1.72, P = 0.034). The mean level of PSA was significantly higher in patients with CCL17 rs223899 AA genotype (mean 500.2, 95% CI 42.1-958.2, P = 0.008) than in those with CC or CA genotypes (mean 196.7, 95% CI 147.3-246.1). However, there was no statistically association between SNPs in NF-κB binding sites and biopsy Gleason score.
3.2. Association between SNPs in NF-κB binding sites and PSA recurrence after RP for clinically localized prostate cancer
The SNPs in NF-κB binding sites were not closely related to PSA recurrence in localized prostate cancer patients receiving RP (Supplementary Table 2).
3.3. Association between SNPs in NF-κB binding sites and clinical outcomes after ADT for advanced prostate cancer
Associations of SNPs in NF-κB binding sites with disease progression, PCSM and ACM in the prostate cancer patients treated by ADT were reported in Supplementary Table 3. No significant association existed between SNPs in NF-κB binding sites and disease progression, but 4 and 3 SNPs had statistical associations with time to PCSM and ACM, respectively. LSAMP rs13088089, CCL17 rs223899, PSMD7 rs2387084 and MON1B rs284924 were associated with PCSM during ADT (P ≤ 0.047) with a q value less than 0.453 (Table 3). To assess the predictive effects of these SNPs beyond
the clinical factors to influence PCSM, we performed a multivariate analysis, adjusting for age, PSA at ADT initiation, biopsy Gleason score, clinical stage, PSA nadir, time to PSA nadir, and treatment modality. After adjusting for the clinical factors, these 4 SNPs persisted significant association with PCSM during ADT (P ≤ 0.037). A strong gene-dosage effect on PCSM during ADT was also observed when these 4 SNPs were analyzed in combination (log-rank P < 0.001, Table 3 and Fig. 1A left). The time to PCSM decreased as the number of unfavorable genotypes increased, and the combined genotype remained as a significant predictor after adjusting for clinical factors (P for trend < 0.001, Table 3). The model based on the clinical factors plus the 4 SNPs of interest was significantly improved over the model with clinical factors only, as indicated by the likelihood ratio test (chi-square 76.9, df 4, P < 0.001).
PSMD7 rs2387084, MON1B rs284924 and WWOX rs386497 had significant effects on ACM (P ≤
0.043), with a q value less than 0.615 (Table 4). After adjusting for clinical factors, however, only
PSMD7 rs2387084 remained a significant predictor for time to ACM in patients receiving ADT
[hazard ratio (HR) 1.78, 95% CI 1.14-2.76, P = 0.011]. Kaplan-Meier survival curves and log-rank test showed that the PSMD7 rs2387084 TG/GG genotypes were significantly associated with poorer overall survival when compared to the TT genotype (P = 0.043, Fig. 1B left). The model including
PSMD7 rs2387084 fitted significantly better than that without the SNP (likelihood ratio chi-square
119.1, df 1, P < 0.001).
Since patients with distant metastasis are considered high risk, a substratification of patients according to the clinical staging was performed to further evaluate the clinical relevance of these SNPs. The combined genotypes also had significant effects on PCSM in patients with or without distant metastasis (P ≤ 0.047, Fig. 1A middle and right). This additional information might lead to better risk prediction, and support that SNPs inside NF-κB binding sites might be independent predictors of PCSM following ADT along with the current prognostic markers.
On the basis of the well-documented role of NF-κB in cancer development, this study systematically evaluated the associations of SNPs inside genome-wide predicted NF-κB binding sites with the characteristics and clinical outcomes of prostate cancer. Our results revealed that several common inherited variants in MYCN, PSMD7, LSAMP, CCL17 and MON1B alone or in combination showed association with clinical outcomes for prostate cancer patients.
The present study showed that MYCN rs1429409 was significantly related to advanced clinical stage. rs1429409 is an intronic SNP in N-myc proto-oncogene (MYCN) that is a transcription factor in the MYC family involved in nervous system development and has been linked to an aggressive subtype of neuroendocrine prostate cancer.22
Significant association was also noted between PSMD7 rs2387084 and earlier onset of prostate cancer, PCSM and ACM after ADT for advanced prostate cancer in our study. rs2387084 is located upstream of 26S proteasome non-ATPase regulatory subunit 7 (PSMD7). The 26S proteasome complex consisting of a 20S core and a 19S regulator, is a non-lysosomal proteolytic machine in eukaryotes, and has been implicated in the cancer progression through deregulation of a variety of oncoproteins, transcription factors, cell cycle specific cyclins, cyclin-dependent kinase inhibitors and other key regulatory cellular proteins.23 PSMD7 gene encoding a component of the 19S regulatory
particle has been found to be over-expressed in breast cancer.24
LSAMP rs13088089, CCL17 rs223899 and MON1B rs284924 had significant association with
PCSM following ADT after adjusting for all clinical predictors. rs13088089 is intronic to limbic system-associated membrane protein (LSAMP) that encodes a neuronal surface glycoprotein found in the limbic system and might function as a putative tumor suppressor. LSAMP has been found to be frequently deleted, down-regulated or epigenetically silenced in osteosarcoma, brain tumors and renal cell carcinoma,25-27 and it also has been reported to significant associate with recurrence of acute
myeloid leukemia.28 rs223899 is located in the promoter region of C-C motif chemokine 17 (CCL17).
The chemokines play important roles in activation and trafficking of T cells, as well as in infiltration and metastasis of tumor cells. It has been shown that T cells could migrate to tumors under the
interaction between chemokines and their receptors to create a favorable environment for tumor growth. Furthermore, a strong, inverse correlation between T cell content and survival in patients with ovarian carcinoma has been noted.29 ADT has also been reported to induce CCL17 production
in the prostate and to increase the number of T cell infiltration.30 rs284924 is in upstream region of
vacuolar fusion protein MON1 homolog B (MON1B). Recently, it has been suggested that MON1B was involved in vesicular trafficking, maturation of apoptotic-cell containing phagosome and autophagy system.
Based on the systematical evaluation of sequence variants in NF-κB binding sites in this research, we lighted up the potential pathways that might influence the clinical outcomes of prostate cancer, e.g., MYCN and LSAMP in neuroendocrine differentiation, PSMD7 in proteasome regulation, CCL17 in tumor microenvironment and MON1B in autophagy. Future studies may be dedicated to clarifying how these sequence variants in NF-κB binding sites affect the gene expression and whether its alteration contributes to outcomes of prostate cancer. In addition, we also found no SNP associated simultaneously with both clinical outcomes after RP and after ADT, suggesting that there seemed to be different biological pathways. The distinct risk factors underlying the progression of the localized and advanced prostate cancer also need further investigation.
The strengths of our study included detailed clinical information and comprehensive evaluation of the common genetic variants in genome-wide predicted binding sites, while the limitations included a hypothesis generating research and the results constrained by multiple comparisons. In addition, our findings from the homogeneous Chinese Han population in this study might less applicable to other ethnic groups. Therefore, investigation of these SNPs in independent populations and their biologic mechanisms underlying the associations are needed.
In conclusion, we identified several plausible candidate SNPs in NF-κB binding sites that were significantly associated with the clinical outcomes in patients with prostate cancer. We also noted significant cumulative effects of multiple SNPs on the predicting prognosis in patients receiving ADT. The identification of novel genetic markers for prostate cancer might be helpful not only for
understanding the biology of prostate carcinogenesis, but also for identifying individuals at high risk by integrated with the known clinical factors.
Role of the Funding Source
This work was funded by the National Science Council (NSC), Taiwan (NSC-98-2320-B-039-019-MY3, NSC-99-2314-B-037-018-MY3 and NSC-100-2314-B-039-009-MY3), and Kaohsiung Medical University Hospital (KMUH100-0R42). The study funders had no involvement in the study design, data analysis or manuscript preparation.
Conflict of interest statement
None declared.
Acknowledgements
We gratefully acknowledge the help from Chao-Shih Chen in data collection and the technical support from the National Center for Genome Medicine, NSC, Taiwan.
References
1. Siegel R, Ward E, Brawley O, et al. Cancer statistics, 2011: the impact of eliminating
socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212-36. 2. Hayden MS, Ghosh S. Signaling to NF-kappaB. Genes Dev 2004;18:2195-224.
3. Oeckinghaus A, Ghosh S. The NF-kappaB family of transcription factors and its regulation.
Cold Spring Harb Perspect Biol 2009;1:a000034.
4. Fradet V, Lessard L, Begin LR, et al. Nuclear factor-kappaB nuclear localization is
predictive of biochemical recurrence in patients with positive margin prostate cancer. Clin Cancer Res 2004;10:8460-4.
5. Shukla S, MacLennan GT, Fu P, et al. Nuclear factor-kappaB/p65 (Rel A) is constitutively
activated in human prostate adenocarcinoma and correlates with disease progression. Neoplasia 2004;6:390-400.
6. Huang SP, Huang LC, Ting WC, et al. Prognostic significance of prostate cancer
susceptibility variants on prostate-specific antigen recurrence after radical prostatectomy. Cancer Epidemiol Biomarkers Prev 2009;18:3068-74.
7. Bao BY, Lin VC, Huang SH, et al. Clinical significance of tumor necrosis factor receptor
superfamily member 11b polymorphism in prostate cancer. Ann Surg Oncol 2010;17:1675-81. 8. Huang SP, Ting WC, Chen LM, et al. Association analysis of Wnt pathway genes on
prostate-specific antigen recurrence after radical prostatectomy. Ann Surg Oncol 2010;17:312-22. 9. Bao BY, Pao JB, Huang CN, et al. Polymorphisms inside microRNAs and microRNA
target sites predict clinical outcomes in prostate cancer patients receiving androgen-deprivation therapy. Clin Cancer Res 2011;17:928-36.
10. Bao BY, Pao JB, Huang CN, et al. Significant associations of prostate cancer susceptibility
variants with survival in patients treated with androgen-deprivation therapy. Int J Cancer 2012;130:876-84.
11. Huang CN, Huang SP, Pao JB, et al. Genetic polymorphisms in androgen receptor-binding
sites predict survival in prostate cancer patients receiving androgen-deprivation therapy. Ann Oncol 2012;23:707-13.
12. Huang CN, Huang SP, Pao JB, et al. Genetic polymorphisms in oestrogen receptor-binding
sites affect clinical outcomes in patients with prostate cancer receiving androgen-deprivation therapy. J Intern Med 2012;271:499-509.
13. Freedland SJ, Sutter ME, Dorey F, et al. Defining the ideal cutpoint for determining PSA
14. Ross RW, Oh WK, Xie W, et al. Inherited variation in the androgen pathway is associated
with the efficacy of androgen-deprivation therapy in men with prostate cancer. J Clin Oncol 2008;26:842-7.
15. Zeitlinger J, Simon I, Harbison CT, et al. Program-specific distribution of a transcription
factor dependent on partner transcription factor and MAPK signaling. Cell 2003;113:395-404.
16. Ferretti V, Poitras C, Bergeron D, et al. PReMod: a database of genome-wide mammalian
cis-regulatory module predictions. Nucleic Acids Res 2007;35:D122-6.
17. Bulyk ML. Computational prediction of transcription-factor binding site locations. Genome
Biol 2003;5:201.
18. Roche PJ, Hoare SA, Parker MG. A consensus DNA-binding site for the androgen
receptor. Mol Endocrinol 1992;6:2229-35.
19. Karolchik D, Hinrichs AS, Furey TS, et al. The UCSC Table Browser data retrieval tool.
Nucleic Acids Res 2004;32:D493-6.
20. Thorisson GA, Smith AV, Krishnan L, et al. The International HapMap Project Web site.
Genome Res 2005;15:1592-3.
21. Storey JD, Tibshirani R. Statistical significance for genomewide studies. Proc Natl Acad
Sci U S A 2003;100:9440-5.
22. Beltran H, Rickman DS, Park K, et al. Molecular characterization of neuroendocrine
prostate cancer and identification of new drug targets. Cancer Discov 2011;1:487-95.
23. Frankland-Searby S, Bhaumik SR. The 26S proteasome complex: an attractive target for
cancer therapy. Biochim Biophys Acta 2012;1825:64-76.
24. Thompson HG, Harris JW, Brody JP. Post-translationally modified S12, absent in
transformed breast epithelial cells, is not associated with the 26S proteasome and is induced by proteasome inhibitor. Int J Cancer 2004;111:338-47.
25. Chen J, Lui WO, Vos MD, et al. The t(1;3) breakpoint-spanning genes LSAMP and
26. Kresse SH, Ohnstad HO, Paulsen EB, et al. LSAMP, a novel candidate tumor suppressor
gene in human osteosarcomas, identified by array comparative genomic hybridization. Genes Chromosomes Cancer 2009;48:679-93.
27. Reed JE, Dunn JR, du Plessis DG, et al. Expression of cellular adhesion molecule
'OPCML' is down-regulated in gliomas and other brain tumours. Neuropathol Appl Neurobiol 2007;33:77-85.
28. Kuhn MW, Radtke I, Bullinger L, et al. High-resolution genomic profiling of adult and
pediatric core-binding factor acute myeloid leukemia reveals new recurrent genomic alterations. Blood 2012;119:e67-75.
29. Curiel TJ, Coukos G, Zou L, et al. Specific recruitment of regulatory T cells in ovarian
carcinoma fosters immune privilege and predicts reduced survival. Nat Med 2004;10:942-9.
30. Sorrentino C, Musiani P, Pompa P, et al. Androgen deprivation boosts prostatic infiltration
of cytotoxic and regulatory T lymphocytes and has no effect on disease-free survival in prostate cancer patients. Clin Cancer Res 2011;17:1571-81.
31. Behrends C, Sowa ME, Gygi SP, et al. Network organization of the human autophagy
system. Nature 2010;466:68-76.
32. Kinchen JM, Ravichandran KS. Identification of two evolutionarily conserved genes
Figure legend
Fig. 1. Kaplan-Meier curves of (A) time to PCSM during ADT for patients with 0, 1, 2, 3, or 4 unfavorable genotypes at the 4 genetic loci of interest, and (B) time to ACM during ADT stratified by genotypes at PSMD7 rs2387084 in all patients (left), in patients without distant metastasis (middle) or in patients with distant metastasis (right). Numbers in parentheses indicate the number of patients. Subtotals do not sum to the total number of patients due to missing data for metastatic status.
Table 1
Clinical characteristics of the study cohort. Characteristic Patients, n 1,024 Age at diagnosis Median, y (IQR) 71 (65-77) PSA at diagnosis Median, ng/mL (IQR) 21.5 (10.1-74.3) Biopsy Gleason score at diagnosis, n (%)
<7 437 (43.4)
7 300 (29.8)
>7 270 (26.8)
Clinical stage at diagnosis, n (%)
T1/T2 509 (50.2) T3/T4/N1 240 (23.7) M1 265 (26.1) RP subcohort Patients, n 321 Age at diagnosis Median, y (IQR) 66 (61-70) PSA at diagnosis Median, ng/mL (IQR) 11.4 (7.4-19.7) Pathologic Gleason score, n (%)
<7 125 (39.8) 7 145 (46.2) >7 44 (14.0) Pathologic stage, n (%) T1/T2 201 (63.6) T3/T4/N1 115 (36.4) M1 0 (0.0) Surgical margin, n (%) Negative 193 (69.2) Positive 86 (30.8) Disease progression No PSA recurrence 202 (62.9) PSA recurrence 119 (37.1) ADT subcohort Patients, n 605 Age at diagnosis Median, y (IQR) 73 (67-78)
PSA at ADT initiation
Median, ng/mL (IQR) 34.5 (11.0-129.0) Biopsy Gleason score at diagnosis, n (%)
<7 195 (32.9)
7 184 (31.1)
>7 213 (36.0)
Clinical stage at diagnosis, n (%)
T1/T2 187 (31.1)
T3/T4/N1 191 (31.8)
PSA nadir
Median, ng/mL (IQR) 0.18 (0.01-1.35) Time to PSA nadir
Median, months (IQR) 10 (5-17)
Treatment modality
ADT as primary treatment 330 (54.8) ADT for post RP PSA failure 72 (12.0) ADT for post RT PSA failure 18 (3.0) Neoadjuvant/adjuvant ADT with RT 127 (21.1)
Others 55 (9.1) Disease progression No ADT failure 187 (31.0) ADT failure 416 (69.0) PCSM Alive 506 (83.8) Dead of disease 98 (16.2) ACM Alive 463 (76.7)
Dead of any cause 141 (23.3)
Abbreviations: IQR, interquartile range; PSA, prostate-specific antigen; RP, radical prostatectomy;
ADT, androgen-deprivation therapy; PCSM, prostate cancer-specific mortality; ACM, all-cause mortality.
Table 2
The genotyping frequencies and the associations of genotypes with the clinical characteristics.
Gene SNP Genotype a Age at diagnosis, n (%)
>70 a≤70 OR (95% CI) bP q PSMD 7 rs2387084 TT 160 (32.3) 116 (26.2) 1.00 TG 236 (47.6) 211 (47.7) 1.23 (0.91-1.67) 0.175 GG 100 (20.2) 115 (26.0) 1.59 (1.11-2.27) 0.012 TG/GG vs TT 1.34 (1.01-1.78) 0.044 0.503 GG vs TT/TG 1.39 (1.03-1.89) 0.034 0.280 Trend 1.26 (1.05-1.51) 0.012 0.161
Gene SNP Genotype a Clinical stage, n (%)
T1/T2 aT3/T4/N1/M1 OR (95% CI) bP q MYC N rs1429409 GG 215 (46.0) 180 (39.1) 1.00 GA 188 (40.3) 213 (46.3) 1.35 (1.02-1.79) 0.033 AA 64 (13.7) 67 (14.6) 1.25 (0.84-1.86) 0.268 GA/AA vs GG 1.33 (1.02-1.72) 0.034 0.510 AA vs GG/GA 1.07 (0.74-1.55) 0.707 0.967 Trend 1.17 (0.98-1.41) 0.090 0.509
Gene SNP Genotype a PSA at diagnosis, ng/mL
CCL1 7 rs223899 CC 404 (44.2) 237.6 (157.4-317.7) CA 372 (40.7) 152.3 (96.9-207.6) 0.338 AA 138 (15.1) 500.2 (42.1-958.2) 0.032 CA/AA vs CC 246.4 (116.5-376.4) 0.915 0.974 AA vs CC/CA 500.2 (42.1-958.2) 0.008 0.232
Abbreviations: SNP, single nucleotide polymorphism; OR, odds ratio; 95% CI, 95% confidence interval; PSA, prostate-specific antigen.
aColumn subtotals do not sum up to a total number of 1,024 patients due to missing genetic or clinical data. bP values were calculated using the logistic regression.
cP values were calculated using the linear regression.
Table 3
The genotyping frequencies and the associations of genotypes with PCSM during ADT.
Gene SNP Genotype aNo. of patients aNo. of events 5-year survival rate (%) bP q HR (95% CI) cP
LSAMP rs13088089 AA 421 66 78.7 1.00 AC 130 27 69.4 1.63 (1.01-2.61) 0.044 CC 12 4 62.5 1.71 (0.61-4.82) 0.311 AC/CC vs AA 0.059 0.221 1.64 (1.05-2.56) 0.031 CC vs AA/AC 0.187 0.552 1.54 (0.55-4.32) 0.409 Trend 0.040 0.200 1.46 (1.02-2.09) 0.037 CCL17 rs223899 CC 257 50 72.2 1.00 CA 228 37 78.1 0.60 (0.38-0.94) 0.025 AA 82 9 84.4 0.54 (0.26-1.12) 0.097 CA/AA vs CC 0.055 0.221 0.58 (0.38-0.89) 0.013 AA vs CC/CA 0.124 0.457 0.70 (0.35-1.40) 0.307 Trend 0.034 0.200 0.68 (0.49-0.95) 0.022 PSMD7 rs2387084 TT 169 21 81.4 1.00 TG 279 53 73.7 2.02 (1.16-3.53) 0.014 GG 110 21 75.3 1.76 (0.91-3.39) 0.093 TG/GG vs TT 0.047 0.221 1.94 (1.13-3.32) 0.016 GG vs TT/TG 0.451 0.832 1.08 (0.65-1.79) 0.768 Trend 0.082 0.308 1.30 (0.97-1.75) 0.083 MON1 B rs284924 GG 353 70 72.3 1.00 GT 191 25 82.4 0.65 (0.40-1.04) 0.073
TT 35 1 95.2 0.19 (0.03-1.41) 0.105
GT/TT vs GG 0.009 0.135 0.59 (0.37-0.95) 0.030
TT vs GG/GT 0.035 0.453 0.22 (0.03-1.63) 0.139
Trend 0.003 0.045 0.59 (0.39-0.91) 0.016
dNo. of unfavorable genotypes present
0 30 2 92.5 <0.001 1.00 1 118 14 84.2 6.26 (0.81-48.2) 0.078 2 225 38 74.8 8.66 (1.17-64.3) 0.035 3 143 31 70.1 11.4 (1.53-85.0) 0.018 4 28 9 54.9 25.5 (3.06-212) 0.003 Trend 1.60 (1.27-2.02) <0.001
Abbreviations: ADT, androgen-deprivation therapy; HR, hazard ratio; 95% CI, 95% confidence interval; PSA, prostate-specific antigen.
aColumn subtotals do not sum up to 605 patients and to 98 PCSM in the ADT subcohort due to missing genetic or clinical data. bP values were calculated using the log-rank test.
cHRs were adjusted for age, PSA at ADT initiation, biopsy Gleason score, clinical stage, PSA nadir, time to PSA nadir and treatment modality. dUnfavorable genotypes refer to AC/CC in LSAMP rs13088089, CC in CCL17 rs223899, TG/GG in PSMD7 rs2387084 and GG in MON1B rs284924.
Table 4
The genotyping frequencies and the associations of genotypes with ACM during ADT.
Gene SNP Genotype aNo. of patients aNo. of events 5-year survival rate (%) bP q HR (95% CI) cP
PSMD7 rs2387084 TT 169 31 76.0 1.00 TG 279 74 67.6 1.89 (1.20-2.98) 0.006 GG 110 27 70.3 1.52 (0.87-2.65) 0.145 TG/GG vs TT 0.043 0.615 1.78 (1.14-2.76) 0.011 GG vs TT/TG 0.694 0.859 0.98 (0.63-1.53) 0.928 Trend 0.119 0.506 1.23 (0.96-1.59) 0.105 MON1 B rs284924 GG 353 90 68.2 1.00 GT 191 40 74.4 0.80 (0.54-1.19) 0.264 TT 35 4 89.6 0.55 (0.20-1.55) 0.260 GT/TT vs GG 0.082 0.615 0.77 (0.52-1.13) 0.176 TT vs GG/GT 0.106 0.620 0.60 (0.22-1.66) 0.328 Trend 0.042 0.429 0.78 (0.56-1.08) 0.136 WWOX rs386497 TT 195 45 72.2 1.00 TC 237 50 73.4 0.73 (0.48-1.12) 0.147 CC 106 33 61.5 1.18 (0.73-1.90) 0.498 TC/CC vs TT 0.963 0.963 0.86 (0.59-1.26) 0.430 CC vs TT/TC 0.020 0.300 1.41 (0.93-2.14) 0.107 Trend 0.198 0.506 1.05 (0.82-1.36) 0.696
Abbreviations: ACM, all-cause mortality; ADT, androgen-deprivation therapy; HR, hazard ratio; 95% CI, 95% confidence interval; PSA,
aColumn subtotals do not sum up to 605 patients and to 141 ACM in the ADT subcohort due to missing genetic or clinical data. bP values were calculated using the log-rank test.
cHRs were adjusted for age, PSA at ADT initiation, biopsy Gleason score, clinical stage, PSA nadir, time to PSA nadir and treatment modality.
