Molecular Profiling of Prostate Cancer Patients

Abstract

In the UK, more than 30 000 men are diagnosed annually with prostate cancer (PCa) and about 10 000 men die from it each year. Although several molecular markers have been associated with prostate cancer development and/ or progression, only few of them are used in diagnostic pathology. The current standard tests include serum PSA test, digital rectal examination and histology of prostate biopsy. Recently the PCA-3 molecular test was approved in the European Union, and it is now used in many laboratories. But these tests are not sufficient to molecularly characterise the behaviour of prostate cancer in many patients. Through extensive literature review, a panel of sixteen molecular markers were selected for further evaluation in prostate cancer cases. They included KLK2, KLK3, MCM2, MCM5, TP53, Bcl-2, CD44, CDH1, AURKA, AURKB, and AURKC; ESR , ESR , AR, FASN, TMPRSS2: ERG, and TMPRSS2:ETV1. The aim was to examine the link between development/progression of prostate cancer and the production of diagnostic/prognostic biomarkers. An in vitro model consisting of PC-3 and PNTIA, MDA PCa 2b prostate cell lines were used to investigate the influence of steroid hormones on these biomarkers using molecular and proteomic techniques. All the three cell lines expressed AR, ESR , ESR and PSA at mRNA and protein levels. The AR expressed in PC-3 and MDA PCA 2b cells was 60 kDa while the PNT1A expressed a 90 kDa AR protein. The ESR was over-expressed in the MDA PCA 2b cells, and was also significantly up-regulated by 17 oestradiol treatment. At a concentration of 4.92 and 33.96μM 17 oestradiol inhibited the growth of 10 to 50% of PNT1A cell line and increased the doubling time three folds. Although the PC-3 cells expressed AR, it was still androgen insensitive and could not produce PSA in culture supernatants. AR and PSA were up-regulated in PNT1A cells in response to testosterone and dihydrotesterone treatment but were reduced in response to 17 oestradiol and Hydrocortisone treatment. All the molecular markers except the TMPRSS2: ERG and TMPRSS2:ETV1 were expressed in the cell lines. The MCM2 and MCM5 were not differentially expressed in response to hormonal treatment. However, the Aurora kinases A, B and C were up-regulated in response to steroid modulation. The KLK2 was only up-regulated by the androgens. Three candidate control genes: ABL1, GUS and G6PD were also evaluated in the cell lines and clinical samples; the ABL1 gene emerged as the most stably expressed house keeping gene and was subsequently used in the normalization of real time PCR assays (RQ-PCR). Analysis of the sixteen biomarkers in prostate tissues and exfoliated urine cells of benign, prostate cancer and non-involved cases (n = 228) showed that seven of the molecular markers were significantly strongly associated with prostate cancer progression (P<0.05). The Aurora kinases A and B were consistently significantly over-expressed in prostate cancer cases. The CD44 was also over-expressed in prostate cancer, and was associated with Gleason score. The TMPRSS2 fusion genes were detected in 15.6% of the prostate cancer cases. The TP53 was also over-expressed in prostate cancer, and significantly associated with tumour grade. The ESR was over-expressed in prostate cancer, and was significantly associated with high tumour grade. This implied a proliferative role for the ESR in prostate cancer progression, because the ESR was not differentially expressed among the sample groups. Concomitantly, the AR was also over-expressed in same pattern with ESR . The combination of these biomarkers: AR, ESR , CD44, TP53, TMPRSS2 fusion genes, AURKA and AURKB could molecularly characterise most prostate cancers. Therefore 2 sets of pentaplex RQ-PCR assays including ABL1 for normalization would provide a cost-effective, flexibly high throughput assay for molecular grading of tissue sections in diagnostic pathology. In addition to the gene expression studies, the genetic variation in KLK2 gene was further investigated by direct DNA sequencing, pyrosequencing and TaqMan allelic discrimination assay. Two SNPs in the gene were found significantly associated with prostate diseases. The T/T allele of rs198977 predicted the presence of prostate cancer at biopsy and was associated with high tumour grade. The A/A variant of rs2664155 was also significantly associated with the presence of benign nodular hyperplasia. The combination of gene expression and genetic variation using real time PCR applications would provide an accurate, reproducible and cheap method for molecular profiling of prostate cancer patients. An exploratory study of organic volatiles in urine of one prostate cancer patient and eight BPH patients using thermal desorption GC-MS showed that Ethanethiol, Dimethyl sulfide, Propyn-1-ol acetate, Nitro-2-propanone, pentane, Hydrazine and Nitrous oxide were differentially over-expressed in the prostate cancer patient compared to the benign cases. Further studies would be required to rule out possible contamination and drug metabolites.