Send Orders of Reprints at
[email protected] Current Pharmacogenomics and Personalized Medicine, 2013, 11, 000-000
1
Original Research Article
Androgen Pathway Related Gene Variants and Prostate Cancer Association in Auckland Men Nishi Karunasinghea, Katja Langeb, Dug Yeo Hanb, Megan Goudied, Shuotun Zhua, Alice H. Wanga, Karen Bishopa, Lynnette R. Fergusona,b,c,* and Jonathan G. Mastersd a
Auckland Cancer Society Research Centre, Faculty of Medical and Health Sciences (FM&HS), The University of Auckland, Auckland, New Zealand; bDiscipline of Nutrition, FM&HS, The University of Auckland, Auckland, New Zealand; cNutrigenomics New Zealand, New Zealand; dUrology Department, Auckland Hospital, New Zealand Abstract: Multiple single nucleotide polymorphisms (SNPs) associated with prostate cancer (PC) have been reported in statistically robust studies in the past but these require an in-depth mechanistic understanding with respect to the biological pathways leading to disease. The current study was carried out to examine the PC and benign urology disease risk associations with lifestyle, demographic and genetic factors in a group of men between the ages 40-81 years from Auckland, New Zealand. The data presented herein support a significant positive association of PC risk with tobacco smoking, and a negative association with alcohol intake. The BMI was not associated with the disease risk. The SRD5A2 rs632148 G allele was associated with PC compared to those with benign urology disease, after adjustments were made for the confounding variables. The Gleason score as well as disease aggressiveness of the PC group showed no association with lifestyle, demographic factors or the SNPs studied. The levels of prostate-specific antigen (PSA) significantly increased with age, smoking status and BMI, and decreased with alcohol consumption. The AKR1C3 rs12529 G allele was significantly associated with lower PSA levels in PC and benign urology disease groups compared to healthy controls. The G allele of the SRD5A2 rs632148 SNP has shown a significant interaction with PSA and a higher Gleason score outcome. Taken together, these findings show the utility of these gene variants and patient lifestyle history, together with the diagnostic serum PSA levels, to collectively enhance the understanding of the clinical-pathological variables of PC. Such information will support the selection of more personalised treatment options for this disease greatly impacting public health.
Keywords: Androgen pathway related gene variants, Gleason score, personalised medicine, prostate cancer, PSA. 1. INTRODUCTION Prostate cancer (PC) is the most common non-skin cancer among men in the western world with Australia and New Zealand grouped as showing the highest estimated agestandardised rate for this disease [1, 2]. Apart from age, ethnicity and family history, very little is known about its etiology. Unfavourable changes related to reproductive, hormonal, metabolic, dietary and behavioural factors have been attributed to a general increase in cancer risk worldwide [3]. The modulation of the function of genes by such factors could result in various genetic risk associations with cancer. Studies with twins and close relatives confirm that there is a heritable factor for PC disease risk [4]. Using multi-stage genome-wide association studies, various susceptibility loci have been identified that account for approximately 25% of familial risk in PC [5, 6]. Multiple single nucleotide polymorphisms (SNPs) associated with PC have been recorded in high powered studies, although these
*Address correspondence to this author at the Discipline of Nutrition, Faculty of Medical and Health Sciences, The University of Auckland, Auckland, New Zealand; Tel: +64 9 9236372; Fax: +64 9 3035962; E-mail:
[email protected]
1875-6921/13 $58.00+.00
need replication as well as require an understanding of their pathway associations leading to the outcome of the disease [7, 8]. Among the widely recorded gene variants associated with increased PC disease risk are the androgen receptor (AR), steroid5 reductase type 2 (SRD5A2), cytochrome P450, family 17, subfamily A, polypeptide 1 (CYP17A)1 and cytochrome P450, family 3, subfamily A (CYP3A) genes [9]. Androgens when bound to AR cause transcriptional regulation of genes including kallikrein-related peptidase 3 (KLK3 or PSA) [9, 10]. Although the risks of PC with AR gene variants are widely recorded there is contradictory reporting of its link with the disease [11-13]. The Ala49Thr (G/A) variant rs9282858 of SRD5A2 has also reported contradictory linkage results with PC with the minor variant leading to significantly lower concentration of 3-androstenediol glucuronide a metabolite of dihydrotestosterone (DHT) [14-17]. The expression of aldo-keto reductase family 1, member C3 (AKR1C3) is up regulated in localized and aggressive PC, and is found to be associated with angiogenesis [18]. The AKR1C3 rs12529 C>G variant has been previously associated with bladder cancer risk, with the minor G allele showing protection [19]. The beta-microseminoprotein (MSMB) gene is considered as a tumour suppressor [20-22] © 2013 Bentham Science Publishers
2 Current Pharmacogenomics and Personalized Medicine, 2013, Vol. 11, No. 1
and lower levels of prostatic secretory protein-94 (PSP94) produced by this gene are associated with aggressive PC [23]. Pomerantz et al. has shown that PSP94 levels determine the level of nuclear receptor co-activator 4 (NCOA4) coded by an adjacent gene on chromosome 10q11 [24] . NCOA4 is reported to transactivate the androgen receptor in hormone refractory PC [25]. The variant rs10993994, located at the telomeric end of the chromosome 10q11, is in linkage disequilibrium with the MSMB gene [24] and is associated with PC risk [26-29]. The 5-- reductase enzyme coded by SRD5A2 gene converts testosterone to the more biologically active DHT [30, 31]. The gene polymorphism rs523349 (V89L) is considered to modulate the level of testosterone [32, 33]. The SRD5A2 rs632148 is in linkage disequilibrium with the rs523349 (V89L) SNP that has previously been associated with ovarian cancer [30]. CYP17 is involved in the biosynthesis of testosterone through the regulation of both 17- hydroxylase and 17,20-lyase [34]. Douglas et al. have shown that Cyp17 variants are reported among men in families with increased PC risk [34]. Here we present a set of androgen pathway related gene polymorphisms evaluated against PC risk and disease aggressiveness in a cohort of Auckland men. The data have been adjusted for various lifestyle and demographic confounding factors. The polymorphisms analysed here are SRD5A2 rs632148, CYP17A1 rs743572, AKR1C3 rs12529 and MSMB associated SNP rs10993994 (hereafter referred to as MSMB rs10993994). 2. MATERIAL AND METHODS 2.1. Study Population Our patient and control database consisted of 885 subjects within the age range of 40-81 years and selfreported a European ancestry. The patient group consisted of 363 with confirmed malignancies and 86 with negative biopsies for PC and therefore considered as having benign prostate disease. PC patients were selected from the databases of Auckland Regional Urology facility covering Auckland District Health Board, Counties Manukau District Health Board, Waitemata District Health Board and private practices within Auckland region and were invited for this study. The patients with benign urology disease were also selected from Auckland District Health Board and Counties Manukau District Health Board databases and were invited for the study. The patient recruitment with informed consent took place through the Department of Urology, Auckland Hospital (ethics ref NTY/05/06/037). The control group consisted of a total of 436 participants that have self-reported as having no history of cancers except for skin cancers or any urological problem. They are part of the volunteers recruited with informed consent for the selenium (Se) supplementation trial carried out by the Discipline of Nutrition, University of Auckland (ethics ref: NTY/06/07/060). 2.2. Collection of Demographic, Lifestyle and Clinical Data At the beginning of the study, both patients and controls completed a demographic and lifestyle questionnaire. The reported data was transferred to a central study database.
Karunasinghe et al.
Those reporting present or past tobacco smoking practice were considered as smokers. The patient clinical data were extracted from the hospital databases and were also transferred to the central study database. The age considered in this analysis for the patient cohort was that of their age at diagnosis, while that of the controls was the age at recruitment. The patients reporting a Gleason score of 7(3+4) were considered as having non aggressive disease while >7(4+3) were considered as having aggressive disease based on Khoddami et al. [35]. 2.3. Blood Collection and Processing At the beginning of the study, blood samples from each volunteer were collected in each of an EDTA, heparin and plain Vacutainer® tube from Becton Dickinson. An aliquot of the EDTA sample was subsequently used for DNA extraction. Total genomic DNA was extracted from blood with the QIAamp DNA Blood Mini Kit (Qiagen) according to the manufacturer’s instructions, using a fully automated procedure on the QIAcube. 2.4. SNP Genotyping of candidate Genes The TaqMan® SNP Genotyping Assay from Applied Biosystems was used for the SNP genotyping of the panel of genes selected for this study. The primers for the assays were obtained either as pre-designed from Applied Biosystems or custom-made through the Assay-by-Design service by ABI. The SRD5A2 rs632148 variant which is in linkage with rs523349 [30] was used in the current study as the latter variant cannot be genotyped on a Taqman platform. Samples were assayed along with no-template and HapMap controls and run on the AB 7900HT Fast Real-Time PCR System using the conditions: 10 min 95 °C enzyme activation followed by 40 cycles at 92 °C for 15 s and 60 °C for 1 min (annealing/extension). The allelic discrimination results were determined after amplification by performing an endpoint read. 2.5. Measurement of Serum PSA Level The patient total PSA level at diagnosis was obtained from their clinical records. For the controls, the total PSA level was measured from serum aliquots at LabPlus, Auckland hospital. Total PSA was measured by electrochemiluminescence immunoassay (Roche Cat. #. 04641655 190) on a Roche Modular E170 anaylser (Roche Diagnostics, NZ). Total assay imprecision was 3.2% at a level of 1.12 ng/mL, 3.7% at 4.61 ng/mL, and 2.7% at 27.5 ng/mL. 2.6. Statistical Analysis The three outcomes of interest, recorded pathology, Gleason score (of malignant patients), and PSA level (ng/ml), were fitted for association with four androgen related SNPs in this study. Serum PSA data was not normally distributed, therefore the data were log transformed for analysis; the estimated actual measures were determined by utilizing the exponential (anti-log) functions. Four explanatory variables (smoking status, alcohol consumption, age and BMI) were tested with the outcomes of interest. Smoking status, alcohol consumption, and age were significantly associated with the recorded pathology (Table 1)
Androgen Pathway Gene Variants and Prostate Cancer
Current Pharmacogenomics and Personalized Medicine, 2013, Vol. 11, No. 1 3
and all of four explanatory variables (smoking status, alcohol consumption, age and BMI) were significantly associated with serum PSA (Table 6). These were used for adjustment of data for further analyses. The additive model was used for linearity of the genotype-phenotype relationship and each SNP was coded 0, 1, and 2 for each tested allele [36]. The PSA and SNPs interaction on aggressive PC was also tested. SAS (V9.2 SAS Institute., Cary, NC, USA) and R (R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, http://www.R-project.org) were used for statistical analyses.
significantly different between the controls and patients. 87% of the control cohort has recorded alcohol consumption compared to 75% for the malignant and 66% for the benign urology cohorts. Smoking status was significantly different among patients and controls with only 38.5% of the controls recording themselves as smokers. Among malignant and benign patients the frequency of tobacco smoking was 55.4% and 45.4% respectively. The PC patient cohort consisted of a total of 257 (72.6%) with non-aggressive disease and 97 (27.4%) with aggressive disease. The aggressiveness of PC showed no significant difference between age, BMI, smoking status or alcohol intake (data not shown). 3.2. Gene Variants and Pathology
3. RESULTS 3.1. Pathology, Age, BMI and Lifestyle The variation of age, BMI, alcohol consumption and smoking status are compared in Table 1. Although patient and control groups were selected to represent the age range 40-81 years, the patient cohorts were unavoidably and significantly older than the controls. There was no significant difference between the BMI among patients and controls. However, the alcohol consumption was Table 1.
Comparison of the explanatory variables between healthy, benign and malignant cohorts.
Age (years)
BMI
Alcohol consumption
Smoking status (ever + current)
Table 2.
Table 2 gives the tested allele variation with pathology. A risk assessment on malignancies and benign urology disease in association with the SNP genotypes under consideration is given in Table 3. AKR1C3 rs12529 G allele shows an increased risk of malignancies compared to the controls. However, this association was not significant (p=0.08) after adjustments were made for the confounding variables. SRD5A2 rs632148 G allele showed a significant association with malignancies compared to benign patients after the data were adjusted for confounding factors.
N
Mean (SE)
Range
Estimate (95% CI)
p
Benign
86
66.5 (0.83)
41-81
9.07 (6.96-11.2)
1.25E-16
Malignant
363
66.4 (0.41)
45-81
9.04 (7.77-10.3)
