Read the Full Video Transcript

Andrea Miyahira: Hi. I'm Andrea Miyahira at the Prostate Cancer Foundation. I'd like to welcome Dr. Balaji Venkadakrishnan of Dana-Farber Cancer Institute. We'll be discussing his latest paper, Lineage-specific canonical and non-canonical activity of EZH2 in advanced prostate cancer subtypes, that was published in Nature Communications. Dr. Venkadakrishnan, thanks for joining us.

Varadha Balaji Venkadakrishnan: Thanks for inviting me to present our recent publication on lineage-specific activity of EZH2 in advanced prostate cancer subtypes. To provide some context, up to 15% of treatment resistance in prostate cancer is due to tumors undergoing histologic transformation from prostate adenocarcinoma to Neuroendocrine Prostate Cancer or NEPC, a process termed as lineage plasticity.

NEPC is associated with poorer prognosis due to the loss of antigen receptor expression. And currently, there are no approved targeted therapies for this subtype. Biologically, NEPC is driven by epigenetic reprogramming, including the dysregulation of the histone modifier and lineage specification factor, EZH2.

EZH2 is part of a repressive complex that brings about trimethylation of histone 3 lysine 27 residue, which is associated with silencing of genes during lineage specification. For example, during embryonic development, EZH2's job is to silence lung-related genes in brain development and brain-related genes in lung development, and so on.

While EZH2's role in prostate cancer has been extensively studied, its canonical repressive function in lineage plasticity has been primarily examined in models transitioning from adenocarcinoma to NEPC, wherein EZH2 inhibitors can reverse lineage and re-express the antigen receptor. However, its role in terminally differentiated NEPC remains less understood.

In other words, if EZH2 has fully silenced androgen receptor-related genes to allow for a mature NEPC phenotype, would EZH2 inhibition still work in this context? Additionally, the non-canonical coactivator function of EZH2, which is associated with the AR transcriptional machinery, is poorly understood in the AR-negative setting of NEPC. These are the two key questions our study aims to address.

This research has significant clinical implications, as understanding any differential activity of EZH2 in NEPC versus prostate adenocarcinoma could reveal novel therapeutic targets. Moreover, several EZH2 inhibitors are currently in clinical trials for prostate cancer, but they are being tested in biomarker unselected populations that include both prostate adenocarcinoma and NEPC. And understanding the lineage-specific activity of EZH2 will aid in developing biomarkers for patient selection.

Our findings indicate that patient-derived NEPC preclinical models show a modest response to EZH2 inhibitors and do not reverse lineage reprogramming. We hypothesize that this may be because EZH2 may have different lineage-specific targets in adenocarcinoma compared to NEPC. H3K27 trimethylation CUT&Tag on rapid autopsy samples distinguishes NEPC samples from adenocarcinoma, with neuroendocrine lineage genes bearing high levels of H3K27 trimethylation in prostate adenocarcinoma, and luminal AR-related genes marked with this repressive label in NEPC.

So what happens because of this distinct epigenetic profile? In preclinical models, we found that neuroendocrine lineage genes may harbor bivalent promoters carrying both the repressive K27 trimethylation mark and the activating H3K4 trimethylation. This bivalency allows them to be induced after EZH2 inhibition, which may facilitate maintenance or even forward differentiation of neuroendocrine prostate cancer state. We have validated this in experiments with the genetic silencing of EZH2 as well as rescue of EZH2.

To summarize our preclinical data, we show that EZH2 targets are distinct in NEPC as compared to prostate adenocarcinoma, and encompass bivalent promoters that may enable forward differentiation upon further induction after EZH2 inhibition. In the context of EZH2's noncanonical coactivator function, we know that in prostate adenocarcinoma, its inhibition leads to downregulation of cell cycle-related genes. But in the absence of antigen receptor in NEPC, EZH2 no longer regulates cell cycle-related genes, suggesting that cell cycle inhibition could be explored as a therapeutic strategy to compensate for the lack of response to EZH2 inhibition.

I'll stop here and thank people who have supported this project. Very grateful for strong support from my mentor, Dr. Himisha Beltran, and fruitful discussion and collaborations within the Beltran Lab, who are represented here as cluster of genes. Big thanks to funding support, especially PCF and collaborators within the institute and in Weill Cornell. And a special thanks to UroToday for hosting me on this digital platform. Thank you.

Andrea Miyahira: Thank you so much for sharing this interesting study. So one question is—is it the neuroendocrine phenotype that makes NEPC targeting difficult or is it the ongoing plasticity and lack of terminal differentiation? And in that regard, does it seem possible to push cells into a terminally differentiated any state and then target it on any target?

Varadha Balaji Venkadakrishnan: I think several factors may contribute to the challenge of targeting NEPC—loss of AR expression or dependence on AR signaling are some, along with heterogeneity and several pathways driving stemness and differentiation, also tumor suppressor loss and high proliferative index. But the question can be pushed tumors to a more mature NEPC state and targeted.

In this study, we show that NEPC cell surface targets, like DLL3, also harbor bivalency. And they're induced upon EZH2 inhibition. We'll have to see if that translates to patients to make tumors homogeneously express any neuroendocrine target proteins and go after them with a combinatorial approach.

Andrea Miyahira: OK. So then would you conclude that EZH2 inhibitors push cells to differentiate toward whatever state they are nearest? And is there any biomarker to determine what that might be?

Varadha Balaji Venkadakrishnan: Yeah, that's interesting as well. So in this study, we treat prostate adenocarcinoma models with EZH2 inhibitors, and it results in further induction of AR and AR target genes. This is consistent with previous studies, where they showed that prostate adenocarcinoma models can become more sensitive to androgen receptor pathway inhibitors after blockade of EZH2.

So overall, it seems like EZH2 inhibition pushes cells to differentiate to whatever cells is nearest. And as for gene signatures that define cell states, we have a rudimentary gene set in this paper that we refer to as lineage-specific targets of EZH2 inhibition. We observed bivalency for luminal-related genes, such as NKX3.1 in prostate adenocarcinoma, and neuroendocrine lineage genes, such as ASCL1 in NEPC, both of which are induced after EZH2 inhibition in their respective phenotypes.

But I think this needs to be further refined with more preclinical models for it to translate to patients. And also, I just want to add that I strongly believe epigenetic profiling of patients with liquid biopsies will inform underlying lineage states of the disease.

Our lab has collaborated with Francesca in developing the NEMO panel, which is a DNA methylation-based assay, and Matt Freedman and Sylvan Baca have developed plasma ChIP-Seq for H3K4 trimethylation. And also, David Takeda recently published on using H3K4 dimethylation, which marks both promoters and enhancers. And all of these are going to be way more informative, I think.

Andrea Miyahira: OK. Thanks. And one of the genes that was bivalent in your study was the antigen presentation gene of the MHC class I. How do you think this relates to tumor immunogenicity versus immune escape of prostate cancer cells and their differentiation or plasticity states?

Varadha Balaji Venkadakrishnan: So MHC class I genes have conserved bivalency across different tissue types. And we observe that EZH2 inhibition dramatically induces MHC expression in NEPC as compared to prostate adenocarcinoma, suggesting a potential role for EZH2 in immune evasion. This has also been reported by Mark Dawson's group in lung adenocarcinoma transforming to small cell lung cancer, with small cell lung cancer acquiring bivalency in MHC and becoming repressed upon lineage transformation.

So whether we can combine EZH2 inhibition with immunotherapy or checkpoint blockade needs to be further explored. There is some incoming evidence in prostate cancer, such as Leigh Ellis's recent work. And the ongoing clinical trial with the Valemetostat and CTLA-4 antibody at M.D. Anderson will also be informative.

Andrea Miyahira: OK. Thanks. And do you think EZH2 inhibition is promising in prostate cancer? And what would be the optimal therapeutic strategy?

Varadha Balaji Venkadakrishnan: Yes. I think EZH2 inhibition holds a lot of promise in prostate cancer. The NEPC models that we use in this study may represent only a small subset of terminally differentiated state that are not responding to EZH2 inhibition. But what we see clinically may be a spectrum of lineage states.

Perhaps tumors that still rely on AR signaling and those that are in initial stages of lineage plasticity may benefit the most when EZH2 inhibition is combined with antigen receptor pathway inhibitors. And I look forward to data from the MEVPRO trial that combines EZH2 inhibition with enzalutamide for mCRPC.

Andrea Miyahira: Thanks. And what are your next steps?

Varadha Balaji Venkadakrishnan: From a biological standpoint, I'm investigating candidates that are derepressed by differential EZH2 activity as potential drivers of NEPC progression, with a goal of defining them as therapeutic targets.

In another project, in collaboration with David Reichman's lab at Weill Cornell, we're investigating a potential crosstalk between EZH2 and DNA methyltransferases, and how that might mediate lineage plasticity.

From a therapeutic perspective, in collaboration with Myles Brown, we're evaluating how EZH2 inhibition could be combined with PARP inhibition in advanced prostate cancers. Atish Choudhury is leading a trial combining tazemetostat and talazoparib. And we are working on complementary preclinical data.

Andrea Miyahira: Thank you. Well, that's a lot. So I look forward to the next time we'll hear from you. Thanks for sharing with us today.

Varadha Balaji Venkadakrishnan: And thanks for having me.