ONECUT2 is a novel target for treatment of castration-resistant prostate cancer

KEYWORDS:

• AR-independent prostate cancer
• castration-resistant prostate cancer
• neuroendocrine prostate cancer
• ONECUT2
• prostate cancer

1. Introduction

Inhibition of androgen receptor (AR) signaling is initially effective in treating metastatic prostate cancers (mCRPC). However, these treatments are not curative, and the emergence of castration-resistance followed by eventual progression to aggressive and lethal phenotypes remains a large challenge. Both AR, its associated key players and their pathways are of great interest as potential targets in drug discovery. The recent Food and Drug Administration (FDA) approval of AR inhibitors apalutamide and darolutamide for the treatment of non-metastatic CRPC patients supports that AR remains a viable therapeutic target. However, overcoming resistance will require new targets for treatment and combinations of treatments. Recently, two independent studies by Rotinen et al. [1] and Guo et al. [2] identified and experimentally validated an atypical homeobox transcription factor ONECUT2 as a targetable master regulator of lethal prostate cancer.

It is well observed that a subset of mCRPC tumors show treatment-emergent neuroendocrine prostate cancer (CRPC-NE). This ‘AR-indifferent’ cell state is thought to arise through divergent clonal evolution as a mechanism for treatment resistance [3,4]. It is clear that AR is no longer active, but the transcriptional regulators responsible for the AR inactivity and transdifferentiation to CRPC-NE have not been fully elucidated. In this editorial, we highlight the critical observations showing the role of ONECUT2 in mCRPC and comment on the challenges ahead for the development of ONECUT2-targeted therapeutics.

2. ONECUT2 is a key regulator of mCRPC

Rotinen et al. and Guo et al. used bioinformatics models to identify the transcription factor ONECUT2 as a master regulator for AR-independent growth [1,2]. They discovered that: ONECUT2 mRNA and protein expression increase with disease progression, although there is considerable patient-to-patient variability [1,2]; there is an inverse relationship between nuclear AR and ONECUT2 staining in high grade prostate cancer samples [1]; and, there is a 26% overlap between ONECUT2 and AR binding sites, with overexpression of ONECUT2 resulting in the repression of approximately half of the AR-targeted genes analyzed [1].

In addition to repressing a subset of AR targets, ONECUT2 can directly repress AR and FOXA1, likely inducing neuroendocrine differentiation [1]. In mCRPC tumor samples and cancer cell lines, high ONECUT2 expression was correlated with the upregulation of neuroendocrine and neuronal differentiation genes [1,2]. ONECUT2 is activated when REST, a known repressor of neuroendocrine differentiation [5] is depleted, and ONECUT2 upregulates PEG10, a known master regulator of neuroendocrine differentiation [1,6].

However, not all cells from mCRPC tumor samples with high nuclear ONECUT2 expression expressed neuroendocrine markers, despite being AR-negative [1]. It is known that a subset of AR-independent tumors can be AR-negative and not express neuroendocrine markers [7], and these findings warrant further studies of ONECUT2 in such tumors.

ONECUT2 is a promising novel target for the treatment of AR-indifferent or AR-inactive prostate cancers, as depletion of ONECUT2 inhibits tumor growth and metastasis in tumor xenograft models [1,2]. A key contribution from Rotinen et al. is the identification of CSRM617 as a small molecular inhibitor targeting the ONECUT2-HOX domain, and the demonstration of its efficacy in inhibiting tumor growth and metastasis [1]. Expression of PEG10 was reduced by CSRM617 suggestive of effective targeting [1]. These studies provide a strong foundation to pursue drug development targeting ONECUT2 for the treatment of mCRPC.

3. Expert opinion

AR is a significant oncoprotein for prostate cancer and mCRPC, and the primary focus of novel drug development. Yet, a remaining challenge is the emergence of a lethal AR-inactive cell state where next-generation AR-targeted agents are ineffective and current toxic chemotherapies show limited benefit [8].

Disease progression is compatible with reduced AR activity, as increased ONECUT2 function promotes transdifferentiation and lethal malignancy [1]. Thus, ONECUT2 but not AR is the driver and therapeutic target for a subset of mCRPCs. Rotinen et al. clearly showed that ONECUT2 is a druggable target, identify a promising ONECUT2 inhibitor CSRM617 and provide a strong rationale for further research on ONECUT2.

However, several key questions remain. Firstly, the high patient-to-patient variability in ONECUT2 mRNA and protein expression is a concern. Will ONECUT2 inhibition only be effective in patients with the highest ONECUT2 expression? The efficacy of CSRM617, and the lack of adverse effects appears very promising, but further preclinical testing with additional patient-derived xenograft and organoid models is needed, as only one prostate cancer xenograft line was treated with one drug dosage, and with a treatment duration limited to 3 weeks [1]. Additionally, predictive biomarkers for ONECUT2 targeted therapy are needed, and could be identified from the many ONECUT2 downstream targets [1] .

A second question is, when should patients be treated with ONECUT2 inhibitors, and how should ONECUT2 inhibition be combined with current treatments? Since ONECUT2 promotes neuroendocrine differentiation, should patients be treated with ONECUT2 inhibitors prior to developing resistance to AR-targeted agents? The presence of ONECUT2 positive cells in hormone naïve patients suggests that this could be beneficial [1]. It is interesting that the majority of ONECUT2 DNA binding sites are not at AR binding sites, as it points to many AR-independent functions of ONECUT2. Given the many ONECUT2 targets, a better understanding of the mechanism for ONECUT2 will be important.

Could ONECUT2 inhibition be combined with other treatments? ONECUT2 regulates hypoxia signaling by activating SMAD3 and modulating HIF1α chromatin-binding [2], but whether HIF1α and ONECUT2 inhibitors could act synergistically remains to be determined. It is possible that ONECUT2 could be a therapeutic target for non-prostate tumors, as findings from a recent study suggest that ONECUT2 promotes transdifferentiation and disease progression in certain lung cancers [9].

A major challenge for developing effective treatments for mCRPC is to understand the intratumor heterogeneity and tumor evolution of prostate cancers. Rotinen et al. detected a preexisting population of ONECUT2 expressing cells in hormone naïve prostate cancer [1]. In CRPC, AR-high and AR-low populations can exist within individual patient tumors [10]. This tumor heterogeneity may explain why paradoxically, high ONECUT2 expression was associated with biochemical recurrence, though ONECUT2 activation suppresses AR activity [1]. This suggests that ONECUT2 could serve as a prognostic biomarker in early-stage prostate cancers, but further studies are needed. On the therapeutic front, it is reasonable to consider combination therapies for heterogeneous tumors by targeting both AR signaling and ONECUT2.

In summary, ONECUT2 is a therapeutic target and its inhibition provides a promising avenue for treating a subset of AR-independent mCRPCs. Further preclinical studies on determining the prostate cancer types most likely to respond to ONECUT2 inhibition, understanding ONECUT2-driven castration resistance mechanisms, identifying predictive biomarkers, and understanding the role of ONECUT2 in tumor heterogeneity and tumor evolution will be important. AR-independent prostate cancers will continue to increase as more effective androgen-pathway inhibitors are used at earlier disease stages, underscoring the need for further research on ONECUT2 and the development of new treatments for AR-independent mCRPC.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose

Additional information

Funding

This work was supported by NIH grant R00 CA194287 awarded to M Shibata.