Discovery of candidate long non-coding RNAs as biomarkers and treatment targets for prostate cancer

Chandler, Stephen; Gangadharannambiar, Priyadarsini; Crea, Francesco; Hirst, Mark; Cox, Angela and Rigas, Sushila (2019). Discovery of candidate long non-coding RNAs as biomarkers and treatment targets for prostate cancer. In: NEPC 2019. Treatment-emergent neuroendocrine prostate cancer: clinical relevance and molecular pathogenesis., 25 Jun 2019, The Open University, Milton Keynes, UK.

Abstract

The role of long non-coding RNAs (lncRNAs) in prostate cancer is largely unknown, and discovery of consistent biomarkers for cancer subtype diagnosis and novel treatment targets is needed, particularly for more aggressive tumours such as neuroendocrine prostate cancers (NEPCs). To identify lncRNA signatures in Gleason group subtypes, we used R scripting with Empirical Bayes Modified t-test to compare lncRNA expression levels of 98 normal samples to 262 prostate cancer tissue samples with up to 3,750 lncRNAs from 3 Caucasian microarray datasets in the Gene Expression Omnibus (NCBI). Across the Gleason group subtypes we identified 61-287 significantly (adjusted Benjamini-Hochberg FDR, P=1.09E-13-0.049) upregulated (log2 fold change=0.19-0.78) lncRNAs. The most significant and highest upregulated lnRNAs were identified as candidate biomarker/novel treatment target signatures. Consistently across all subtypes, one of the lncRNAs (log2 fold change=0.44-0.77) was significantly correlated with reduced overall survival (combined HR=10.44, P=0.026) using SurvExpress (omicX) and was identified as a candidate generic biomarker/treatment target for prostate cancer regardless of subtype. In Gleason groups 2-5, eight of the lncRNAs (log2 fold change=0.29-0.99) were correlated with reduced relapse free survival (HR=2.09-14.29, P=0.002-0.05). We also constructed weighted gene co-expression network images using Cytoscape 3.7.1 (NIGMS) and identified corresponding cell pathways using The Gene Ontology knowledgebase. We determined that across all subtypes of prostate cancer, the lncRNAs (with co-expressed protein coding genes) were operating mainly in immune response pathways. Next, we will analyse the functions of the candidate lncRNAs and propose a mechanism of action. We also hope to validate these findings using siRNA assays. The discovery of these lncRNAs together with our candidate protein coding genes could be developed as biomarkers (like the PAM50 genes in breast cancer) for the first time in prostate cancer and may be potential novel treatment targets.

Viewing alternatives

Item Actions

Export

You can export this page using these formats