Page 84 - Read Online

Page 6 of 12 Lorenzin et al. J Transl Genet Genom 2019;3:5. I https://doi.org/10.20517/jtgg.2019.01

in a fraction of CRPCs under AR-targeted drugs pressure. Since the introduction of second generation
AR-directed therapies in the clinical management of CRPC patients, a significant increase in tumors with
a continuum of neuroendocrine markers and morphological variants was observed [80-83] in up to 20% of
resistant tumors [84,85] . These tumors demonstrate epithelial plasticity and transition towards a predominantly
[86]
AR-low neuroendocrine phenotype. Elegant work from Zou et al. applied lineage tracing in vivo in
genetically engineered mouse models resistant to abiraterone and developing tumors resembling treatment-
induced neuroendocrine disease in CRPC patients; the study provided quantitative evidence that both focal
[86]
and overt neuroendocrine tumor regions arise via transdifferentiation from adenocarcinoma cells . A
model of divergent clonal evolution in CRPC from adenocarcinoma to neuroendocrine phenotype was also
supported by the largest genomic and molecular study in human samples (metastatic biopsies) designed to
study the etiology and molecular basis for this treatment-resistant cell state. Where the genomic component
of the study recognized a substantial genomic overlap between CRPC neuroendocrine and CRPC
adenocarcinomas tumors, albeit with few exceptions (enrichment for TP53 and RB1 loss and depletion of
AR amplification), the genome-wide methylation profiles detected marked epigenetic differences, suggesting
epigenetic modifiers as possible players in the state transformation or maintenance; the transcriptomic
component of the study further highlighted the dysregulation of specific pathways including neuronal and
[87]
stem cell programs and epithelial-mesenchymal transition .
In 2017, two studies contributed novel insights into the mechanistic understanding of cell plasticity
of prostate cancer cells resistant to androgen deprivation therapies [88,89] . Altogether, they show that
downregulation of TP53 and RB1 leads to increased expression of genes regulated by E2F transcription
factor, including EZH2 and SOX2; that SOX2 induction contributes to AR pathway independence and
lineage reprogramming, and that EZH2 inhibition leads to increase in AR and restoration of responsiveness
to enzalutamide. Additionally, SOX2 activity is controlled by a master regulator of neural differentiation,
[90]
BRN2, which is inhibited by AR and therefore active in AR-null tumors . More recently, in the attempt
to look for shared vulnerabilities of small cell neuroendocrine carcinomas that arise from epithelial
cancers under therapies inhibiting critical oncogenic pathways, Park and colleagues utilized human tissue
[91]
transformation assays and tested the functional role of known main drivers . Starting from primary basal
epithelial cells isolated from the prostates of human donors and a leave-one-out approach, they observed that
inactivation of both TP53 and RB1 is required to reprogram the transcriptional profile and the chromatin
accessibility of normal prostate epithelial cells to small cell prostate cancer; and that the same driver events
can initiate small cell lung cancer from human normal lung epithelial cells.

[92]
Building on the initial observation that AURKA and MYCN are amplified and co-overexpressed in
[93]
neuroendocrine prostate cancer, Dardenne et al. , using MYCN transgenic mouse and organoid models,
showed that N-MYC overexpression leads to the development of poorly differentiated, invasive prostate
cancer, similar to human neuroendocrine tumors. They observed abrogation of androgen receptor signaling
and induction of Polycomb Repressive Complex 2 signaling. Inhibition of Aurora-A resulted in decreased
steady-state levels of N-MYC protein and of its target genes and in decreased cell viability, altogether
suggesting exploitation of the mutual dependence of N-MYC and Aurora-A to revert their oncogenic
functions in neuroendocrine disease.

Profiling of mCRPCs patients identified a double negative population of tumors (DNPCs) that shows a
lack of both the AR program and of neuroendocrine markers [85,94] . Interestingly, the proportion of DNPCs
increases in cohorts of patients collected after the approval of more effective AR antagonists such as
enzalutamide and abiraterone, suggesting a shift in the tumor phenotype probably linked to the use of these
[85]
second generation antiandrogens . In vitro work showed that these double negative cells rely on FGF and
MAPK activity for growth and targeting of this AR bypass pathway can be exploited for treatment of AR-
[85]
null tumors .