Molecular profiling of radical prostatectomy tissue from patients with no sign of progression identifiesERGas the strongest independent predictor of recurrence

Abstract

Background: As a major cause of morbidity and mortality among men, prostate cancer is a heterogenous disease, with a vast heterogeneity in the biology of the disease and in clinical outcome. While it often runs an indolent course, local progression or metastasis may eventually develop, even among patients considered “low risk” at diagnosis. Therefore, biomarkers that can discriminate aggressive from indolent disease at an early stage would greatly benefit patients. We hypothesized that tissue specimens from early stage prostate cancers may harbor predictive signatures for disease progression.Methods: We used a cohort of radical prostatectomy patients with longitudinal follow-up, who had tumors with low grade and stage that revealed no signs of future disease progression at surgery. During the follow-up period, some patients either remained indolent (non-BCR) or progressed to biochemical recurrence (BCR). Total RNA was extracted from tumor, and adjacent normal epithelium of formalin-fixed-paraffin-embedded (FFPE) specimens. Differential gene expression in tumors, and in tumor versus normal tissues between BCR and non-BCR patients were analyzed by NanoString using a customized CodeSet of 151 probes.Results: After controlling for false discovery rates, we identified a panel of eight genes (ERG, GGT1, HDAC1, KLK2, MYO6, PLA2G7, BICD1 and CACNAID) that distinguished BCR from non-BCR patients. We found a clear association of ERG expression with non-BCR, which was further corroborated by quantitative RT-PCR and immunohistochemistry assays.Conclusions: Our results identified ERG as the strongest predictor for BCR and showed that potential prognostic prostate cancer biomarkers can be identified from FFPE tumor specimens.

Reduced KLK2 expression is a strong and independent predictor of poor prognosis in ERG‐negative prostate cancer

“…While KLK2 expression in IHC was typically at least weak to moderate in the cytoplasm of normal prostate epithelium, it was reduced or lost in a significant fraction of prostate cancers. This is in line with analyses of published transcriptome data describing lower mRNA expression in prostate cancer than in normal prostatic tissue (reviewed in 11,12 showing associations between low AR signaling and adverse patient outcome 30 and a comparably strong link of reduced expression of the KLK2 regulated PSA to unfavorable prognosis for our prostate cancer cohort. 28 These observations fit with results of Tremblay et al, 10 who also described associations of reduced KLK2 expression with unfavorable prostate cancer phenotype in their 70 cancers.…”

“…One study on 70 prostate cancers reported an association between KLK2 downregulation and high tumor grade. 10 Other studies described a reduced KLK2 messenger RNA (mRNA) expression in prostate cancer as compared with normal prostate tissue (reviewed in 11,12 ). To clarify the clinical relevance of KLK2 protein expression in prostate cancer tissue, we took advantage of our pre-existing tissue microarray (TMA) containing more than 12 000 prostate cancer specimens and analyzed KLK2 expression by immunohistochemistry (IHC).…”

Proteomic Tissue-Based Classifier for Early Prediction of Prostate Cancer Progression

“…The initial list of candidate markers included 151 genes that were selected based on the following criteria [17]: (1) have significant differential gene expression in prostate tumor versus normal comparison; (2) are regulated by androgen; (3) are associated with prognosis of prostate cancer; (4) are associated with the ETS family of transcription factors; (5) are commonly rearranged in prostate cancer; (6) are involved in prostate cancer cell invasion; (7) are associated with multiple malignancies; (8) or can distinguish prostate epithelial from stromal cells (Table S1). These 151 genes were incorporated into a NanoString Code Set, which was used to identify differential mRNA expression in a series of PCa samples across clinical outcomes.…”

“…These 151 genes were incorporated into a NanoString Code Set, which was used to identify differential mRNA expression in a series of PCa samples across clinical outcomes. Based on the significance in differential mRNA expression [17] and the likelihood of detecting them at protein level, the original list of 151 was down-selected to 52 candidates for targeted proteomics measurements using PRISM-SRM.…”

Development of a Novel Immune-Related Prognostic Signature for Prostate Cancer

“…In addition, PCa patients are often excluded from molecular investigations and randomized clinical trials for PCa. Therefore, there is still lack of solid data for further investigation the molecular pro ling of PCa, which may unearth novel targets for diagnosis, treatment, and prognosis of PCa [9][10][11].…”