Abstract
Prostate cancer (PCa) is the most commonly registered cancer and the third most common cause of cancer-related death in male New Zealanders. PCa-related death primarily occurs due to metastasis. Androgen deprivation therapy (ADT) is the most predominant treatment for metastatic PCa. While ADT is initially useful, PCa cells can develop resistance to ADT and restore androgen receptor (AR) signalling with low androgen levels, referred to as castration- resistant prostate cancer (CRPC). Second-generation AR-targeted therapies have been developed to treat CRPC. However, some CRPC tumours can develop resistance via lineage switching, transdifferentiating from prostate adenocarcinoma (PRAD) luminal epithelial cells to neuroendocrine-like cells. About 20%–25% of patients with CRPC can relapse and develop tumours with such neuroendocrine features, called neuroendocrine prostate cancer (NEPC). NEPC is one of the most lethal subtypes of PCa.The current approaches to clinical diagnosis of NEPC have apparent limitations. The currently available treatment options for NEPC only have short-response durations. Therefore, a better understanding of the molecular mechanisms underlying the transdifferentiation from PRAD to NEPC is clearly needed.
NEPCs have few genomic aberrations, suggesting that epigenetic alterations likely drive the neuroendocrine transdifferentiation process. DNA methylation is the best-characterised epigenetic alteration in cancers. Increasing evidence suggests that aberrant DNA methylation patterns are implicated in NEPC development. Nevertheless, very few genome-scale DNA methylation analyses using matched PRAD cells and NEPC-like cells over an extended period of time have been conducted. Therefore, we aimed to generate an in vitro model of neuroendocrine transdifferentiation from PCa LNCaP cells (a model of PRAD cells) to NEPC- like cells, and then identify regions of differential methylation associated with neuroendocrine transdifferentiation.
We induced neuroendocrine transdifferentiation of PCa LNCaP cells into NEPC-like cells by culturing them in androgen-depleted media over a period of 28 days. We found that PCa LNCaP cells started showing a neuronal-like morphology after five days of androgen deprivation (AD) treatment, primarily characterised by the emergence of neurite-like extensions in cells. We also found that from Day 5 to Day 13, the average log-transformed length of neurite-like extensions in AD group cells were significantly longer than those of their respective Control group cells. Moreover, we found that three out of four well-established neuroendocrine marker genes were consistently significantly upregulated in PCa LNCaP cells after six to 24 days of AD treatment. It was also worth noting that 28 days into treatment, both Control group and AD group LNCaP cells seem to acquire a neuroendocrine-like phenotype. Hence qualitative and quantitative analyses suggested that we had successfully transdifferentiated PCa LNCaP cells into NEPC-like cells. Henceforth, we will refer to Control group cells as PRAD cells and AD group cells as NEPC-like cells.
We then generated genome-scale DNA methylation profiles for PRAD cells and NEPC-like cells during neuroendocrine transdifferentiation using reduced representation bisulfite sequencing. This analysis identified 16 significantly differentially methylated fragments (DMFs) between PRAD cells and NEPC-like cells. Apart from the DMF located within an intron of RPS6KA2, no other DMFs have been previously reported in NEPC-like cells. We also identified four DMFs with associated genes previously implicated in cancer development and progression. DMFs within the gene body of mitogen-activated protein kinase kinase 3 (MAP2K3) and FERM domain containing 4A (FRMD4A) were hypermethylated in NEPC-like cells. MAP2K3 and FRMD4A may act as oncogenes in PCa. In contrast, DMFs within the gene body of ribosomal protein S6 kinase A2 (RPS6KA2) and deuterosome assembly protein 1 (DEUP1) were hypomethylated in NEPC-like cells. RPS6KA2 and DEUP1 may act as tumour suppressor genes in PCa.
The information generated throughout this project provides new insights into regions of differential methylation associated with neuroendocrine transdifferentiation from PRAD to NEPC. Future RNA sequencing data analysis may establish the associations between the differential DNA methylation patterns and gene expression changes. These findings may serve as a foundation for further mechanistic investigations into how DNA methylation changes drive gene expression changes relevant to NEPC development. It is hoped that site-specific DNA methylation signatures can eventually be identified, which can be used as diagnostic and prognostic biomarkers for NEPC.