Affronti, H. C. | bioRxiv Science

bioRxiv2020

Reduced NCOR2 expression accelerates androgen deprivation therapy failure in prostate cancer

NCOR2 is frequently and significantly mutated in late stage androgen deprivation therapy resistant prostate cancer (ADT-RPCa). NCOR2 has been characterized as a transcriptional corepressor and has mechanistic links to DNA methylation, but its global functions and overall contributions to PCa progression remain enigmatic. In the current study, we utilize immunohistochemical staining of samples from over 700 PCa patients and reveal associations of reduced NCOR2 expression with correlates of aggressive primary PCa and recurrence in patients who received adjuvant androgen deprivation therapy. We mapped the dihydrotestosterone (DHT) dependent and independent effects of NCOR2 on the transcriptome, cistrome and DNA methylome in androgen sensitive (AS) and ADT-RPCa cells using the isogenic LNCaP and LNCaP-C4-2 (C4-2) cell models. Transcriptional profiling identified androgen dependent and independent regulatory roles of NCOR2, the latter of which was enhanced in the ADT-RPCa state and included neuronal differentiation. Interestingly, reduced expression of NCOR2 resulted in a striking global DNA hypermethylation pattern that significantly enriched at enhancer regions. ChIP-seq revealed that NCOR2 was more clearly associated with promoters in AS LNCaP cells, which was modestly enhanced by DHT treatment. However, in ADT-RPCa C4-2 cells, the NCOR2 cistrome was larger and more distal. Motif analyses and integration of large-scale public cistrome data revealed strong enrichment for FOXA1 in mediating NCOR2 binding, and included additional factors such as AR, E2F, TET2, MED1 and MBD2. Utilizing the CWR22 xenograft model, we demonstrate a direct role for NCOR2 in PCa progression as reduced NCOR2 expression attenuated the impact of ADT, and significantly accelerated recurrence of disease. Transcriptomic analyses in recurrent CWR22 tumors indicated NCOR2-dependent gene expression profiles during ADT that were enriched for neuroendocrine associated genes and also associated with worse survival in human patients with ADT-RPCa. DNA methylation profiles in CWR22 tumors with reduced NCOR2 expression recapitulated the hypermethylation observed in vitro, and further revealed that hypermethylation patterns are commonly associated with ADT-RPCa disease, which was also confirmed in human samples. These studies reveal robust roles for NCOR2 in regulating the PCa transcriptome and epigenome and underscore recent mutational studies linking NCOR2 loss of function to PCa disease progression.Competing Interest StatementThe authors have declared no competing interest.View Full Text

cancer biology↗

bioRxiv2018

Pan-cancer analysis of transcriptional metabolic dysregulation using The Cancer Genome Atlas

Understanding the levels of metabolic dysregulation in different disease settings is vital for the safe and effective incorporation of metabolism-targeted therapeutics in the clinic. Using transcriptomic data from 10,704 tumor and normal samples from The Cancer Genome Atlas, across 26 disease sites, we developed a novel bioinformatics pipeline that distinguishes tumor from normal tissues, based on differential gene expression for 114 metabolic pathways. This pathway dysregulation was confirmed in separate patient populations, further demonstrating the robustness of this approach. A bootstrapping simulation was then applied to assess whether these alterations were biologically meaningful, rather than expected by chance. We provide distinct examples of the types of analysis that can be accomplished with this tool to understand cancer specific metabolic dysregulation, highlighting novel pathways of interest in both common and rare disease sites. Utilizing a pathway mapping approach to understand patterns of metabolic flux, differential drug sensitivity, can accurately be predicted. Further, the identification of Master Metabolic Transcriptional Regulators, whose expression was highly correlated with pathway gene expression, explains why metabolic differences exist in different disease sites. We demonstrate these also have the ability to segregate patient populations and predict responders to different metabolism-targeted therapeutics.

bioinformatics↗