The clinical behavior of prostate cancer is highly heterogeneous, with most patients diagnosed with localized disease that successfully responds to surgery or radiotherapy or that can be followed by active surveillance. However, a fraction of men will relapse after initial treatment and eventually progress to an aggressive resistant form with metastasis spreading and high mortality, a state referred to as castration resistant prostate cancer (CRPC). The technological advances in next generation sequencing have enabled the deep genomic and epigenomic characterization of both the hormone naïve and CRPC states, leading to the definition of molecular subclasses of prostate cancer that could inform the clinicians on therapeutic strategies. These studies also shed light on the mechanisms driving resistance to therapy. CRPCs adapt to androgen receptor (AR) signaling impairment - which follows first-line therapies as androgen deprivation or AR targeting - by restoring the nuclear receptor signaling by means of multiple mechanisms. Alternatively, tumor cells might become resistant to targeted therapies by exploiting lineage plasticity and activating alternative pathways. This review will discuss the main mechanisms leading to the emergence of resistance to therapy in prostate cancer patients in the context of genomic and molecular features of CRPC and on their causal role in the development of resistance.
Prostate cancer (PCa) is the most common malignancy diagnosed among men in the US and is a major cause of death with 164,690 new estimated cases in 2018 and 29,430 expected deaths[
At disease initiation, PCa relies on the androgen receptor (AR) signaling pathway for growth. Hence, the mainstay of therapy for PCa is represented by androgen deprivation therapy (ADT) as initially described by landmark work of Charles Huggins and colleagues in 1941[
The genomics of PCa has been more challenging to study with respect to other solid tumor types due to multiple reasons, including intra-patient tumor heterogeneity and prevalence of structural genomic changes[
Primary prostate cancers are usually characterized by a diploid genome, low mutational burden, genomic rearrangements mainly involving the
Schematics of the main mechanisms linked to the development of resistance to androgen deprivation therapy and androgen receptor (AR)-targeted therapies in prostate cancer. Genomic rearrangements and mutations in the
AR is a nuclear hormone receptor that upon activation by androgens [testosterone and 5-alpha-dihydrotestosterone (DHT)] translocates into the nucleus and binds to specific regulatory regions [AR responsive elements (ARE); in promoters and enhancers], regulating specific target genes and promoting prostate cell proliferation and survival.
Functional studies consistently demonstrated that the majority of CRPCs are still dependent on AR signaling. The mechanisms and mutations that sustain AR signaling even at low concentrations of androgens are multiple and mainly involve the
AR point mutations are detected in 10%-20% of advanced prostate cancers upon development of treatment resistance but rarely before endocrine treatment[
Recent reports based on the analysis of cell free DNA from CRPC patient’s plasma associate AR aberrations (both amplification and mutations) with worse outcome and with resistance to abiraterone and/or enzalutamide. These observations highlight the possibility of using plasma-based AR status as biomarker for therapeutic management selection[
Studies leveraging liquid biopsy to obtain prospectively collected patients’ circulating tumor DNA also allowed for the evaluation of tumor clone dynamics and demonstrated the temporal association of AR point mutations with clinical progression. Emergence of AR-L702H mutation was observed in patients receiving exogenous glucocorticoid in combination with second generation AR antagonists (enzalutamide or abiraterone), whereas H875Y and T878A AR mutations were detected at progression on abiraterone and prednisolone[
Expression of AR splicing variants (AR-Vs) has emerged as an additional mechanism driving resistance to first and second generation AR-targeted therapies. These alternative AR-Vs lack the C-terminal ligand-binding domain via truncation or exon skipping while retain the amino-terminal transactivation and DNA-binding domains[
In prostate cancer cell-based models, the high expression of AR-Vs compared to AR full length (AR-FL) is associated with complex genomic rearrangements within the AR locus as exemplified by the 22Rv1 cell line harboring AR gene structural rearrangements and expressing high levels of AR-V7 relative to AR-FL[
The prognostic value of AR-Vs detection, and of AR-V7 in particular, is still unclear. AR-Vs can be detected in normal prostate tissue and early stage tumors, yet their expression seems to increase in CRPCs suggesting an evolutionary advantage for more aggressive disease. AR-V7 is significantly over-expressed in antiandrogen resistant tumors[
A direct link for AR-V7-induced resistance to enzalutamide was provided by the restoration of sensitivity to antiandrogens in 22Rv1 cells depleted of AR-V7[
Other mechanisms of resistance to ADT and AR-targeted therapies relying on AR and/or its regulated targets involve (1) maintenance of physiological intratumoral androgen levels, (2) dysregulated expression and/or mutation of AR cofactors, and (3) cross-regulation of AR targets by other nuclear receptors.
Following castration the levels of testosterone in circulation drop by > 90%. However the adrenal androgens dehydroepiandrosterone (DHEA) and androstenedione can serve as substrates for testosterone and DHT synthesis in the prostate replenishing the pool of androgens needed by the tumor cells[
Beside mutations and structural rearrangements affecting the AR locus itself, aberrations of other genes modulating the AR signaling pathway have been observed in metastatic disease[
Preclinical studies and indirect evidence support the hypothesis that other nuclear receptors [e.g., glucocorticoid receptor (GR), progesterone receptor (PGR) and mineralcorticoid receptor] could substitute AR in gene regulation given the shared structure and the high homology of the DNA-binding domain[
Lineage plasticity is recognized as one possible mechanism to evade targeted therapy where cancer cells transition towards an alternative cell lineage that is not dependent on the target. This mechanism is active 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[
In 2017, two studies contributed novel insights into the mechanistic understanding of cell plasticity of prostate cancer cells resistant to androgen deprivation therapies[
Building on the initial observation that
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[
The observation that a portion of prostate cancer samples overexpresses
An alternative pathway that could sustain growth of prostate cancers is the WTN signaling pathway. Indeed, similar to colon cancers, mutations in
Potent therapies often trigger resistance mechanisms that present or are facilitated by distinct genomic or molecular settings. Whereas functional studies are needed to reveal the mechanistic features of drug resistance, the role of large-scale genomic studies on contemporary cohorts of patients can nominate structural events associated to resistance that can eventually turn into biomarkers for treatment options. Recent work has also suggested that complex structural events possibly involving non-coding areas of the genome can contribute to our knowledge of disease progression.
The application of new therapeutic strategies to treat CRPC patients will further lead to the emergence of alternative mechanisms of resistance different from those identified so far. PARP and immune checkpoint inhibition in CRPC patients harboring defect in DNA repair pathways, such as homologous recombination and mismatch repair, and
Wrote this review: Lorenzin F, Demichelis F
Not applicable.
This work was supported by the European Research Council ERC Consolidator Grant (648670) to Demichelis F.
Both authors declared that there are no conflicts of interest.
Not applicable.
Not applicable.
© The Author(s) 2019.