Abstract
Prostate cancer (PCa) is a clinically heterogeneous disease. Clinical manifestations range from indolent, slow-progressing tumours to aggressive tumours that invade and metastasise. Overdiagnosis and overtreatment of indolent tumours remain a problem within the PCa screening/diagnostic pathway. Nomograms based solely upon clinicopathological variables provide crude estimations of long-term oncological outcomes. Molecular biomarkers can interrogate tumour status, offering a more tailored approach to PCa diagnosis and management. This project aims to investigate DNA methylation and gene expression status in prostate tumour tissue to inform the development of molecular signatures predictive of PCa-associated adverse outcomes (clinical recurrence, metastatic disease, and disease-specific mortality), in a New Zealand cohort from Dunedin Hospital (n = 50).
We used next-generation sequencing methods, cell-free Reduced Representation Bisulfite Sequencing (cf-RRBS) and RNA sequencing (RNA-seq), to generate genome-scale DNA methylome and transcriptome profiles from formalin-fixed paraffin-embedded (FFPE) prostate biopsy tissue samples. We have identified differentially methylated regions (DMRs) and differentially expressed genes (DEGs). To identify our candidate signatures, we are modelling them against known prognostic variables (i.e. PSA, tumour stage, histopathological grade) and PCa-associated adverse outcomes.
In our current analysis between tumour (n = 33) and healthy (n = 17) biopsy tissue, we have identified 443 differentially methylated regions (DMRs) (ANOVA test p-value<0.05 & absolute methylation difference ≥20%) and 232 differentially expressed genes (DEGs) (Wald test false discovery rate-adjusted p-value<0.05 & absolute log2fold change ≥1.5). Concordant signatures across analyses have been categorised as differentially expressed, differentially methylated regions (DE-DMRs). In our initial analysis, we discovered patterns of promoter hypomethylation in genes associated with telomerase activity, growth signalling pathways, and hypermethylation in genes associated with hormone regulation and metabolism. Additionally, we discovered upregulation in well-characterised PCa-associated genes ERG and PCA3, alongside other novel candidate genes.
In the future, candidate signatures identified in this work could be incorporated into a tissue prognostic panel that integrates current predictive algorithms to assist clinical decision-making.