ZNF397 Deficiency: A Key Driver of Lineage Plasticity in Prostate Cancer Therapy Resistance – Ping Mu

Andrea Miyahira: Hi everyone, I’m Andrea Miyahira here at the Prostate Cancer Foundation, and thank you for joining me today. With me is Dr. Ping Mu, an Associate Professor at UT Southwestern Medical Center. Dr. Mu is discussing his latest paper, “ZNF397 Deficiency Triggers TET2-Driven Lineage Plasticity in AR-Targeted Therapy Resistance in Prostate Cancer,” which was published in Cancer Discovery. Dr. Mu, thanks for joining me again.

Ping Mu: Thanks, Andrea, for inviting me to this great platform to share with you our latest work published in Cancer Discovery last month. Thank you very much for listening and thanks to UroToday for giving me this opportunity to share with you our newest work published in Cancer Discovery last month. As we all know, in the field of prostate cancer research, we can actually manage prostate cancer pretty well, especially the primary ones. Primary prostate cancer has a survival rate of 99%. However, the problem is that many of those tumors will eventually become resistant and become mCRPC and no longer respond to our standard of care, AR-targeted therapies. So, my lab is very interested in understanding why prostate cancer becomes resistant to our current AR-targeted therapies. In the past five years, one very new phenomenon called lineage plasticity has been more and more recognized as a mechanism leading to AR-therapy resistance. What exactly is lineage plasticity? It is that prostate cancer cells gain the ability to switch their lineage or identity, almost like a cancer cell could change its face to escape the therapy we designed to target its original lineage.

In the past five years, many key molecular drivers or genes have been connected to lineage plasticity, including JAK/STAT signaling, which is driven by the Janus kinase. This is one of the works we published two years ago in Nature Cancer. However, although many genomic or transcriptional molecules have been connected to lineage plasticity, as I showed here, this is partially the list. There is still one key question that has puzzled me for a long time. We know that lineage plasticity is a whole spectrum of turning down the transcriptome and AR-driven signals in a very short period of time. This process is often reversible. At the same time, cancer cells have to turn on a large spectrum of genes, which regulate a different lineage and become resistant.

So, we are here talking about turning off thousands of AR genes and turning on thousands of lineage plasticity genes in a very short period of time. How could cancer cells do that? For example, JAK/STAT are mostly upstream signals and many of those genomic alterations also happen upstream. How can they actually come down to the final step to turning this molecular switch to switch the cancer cell from an AR-driven program into a lineage plasticity program? It’s like apoptosis; no matter what happens at the top with hundreds of genes, eventually, you go to [inaudible 00:03:18]. But we don’t know what this final molecular switch is. We think in this recent work we find that it is a protein called ZNF397, which is a zinc finger protein. Interestingly, this protein or gene is very frequently depleted in many cancers. For example, about 10 to 25% of prostate cancers carry this ZNF397 depletion. There is much more depletion in metastatic mCRPC compared to primary prostate cancer.

Not only in prostate cancer, but also in many cancers with known lineage plasticity phenomena, including breast cancer and lung cancer, ZNF397 depletion is very frequent. This depletion is highly correlated with the clinical outcome of prostate cancer. For example, it is one of the most significant risk factors correlated with resistance in prostate cancer. Patients with more ZNF397 depletion develop resistance to AR therapies much quicker. We used many different in vitro and in vivo models to show that if we CRISPR delete ZNF397, we can cause very significant resistance. The tumor no longer responds to any AR-targeted therapies. But when we tried to understand the mechanism of this resistance in those tumors, this finding really blew us away.

We thought that tumors with ZNF397 depletion would have a very down-regulated AR downstream gene expression. Even without AR-targeted therapy treatment, those AR genes were already down. As we already know, prostate cancer cells are highly dependent on AR signaling for survival. But how can they grow tumors so effectively without AR signaling? What happens? So, we took all the tumors with ZNF397 depletion and checked their AR signaling with AR chip. We saw this: in wild-type, we can see AR binding to many of the canonical AR genes, but when ZNF397 was depleted, almost 40% of all the AR bindings in those tumors disappeared. This means AR cannot maintain these targets. This was confirmed in many different ways, and I will skip all those detailed results since we only have seven to eight minutes.

At the same time, we know those tumors turned off their AR signaling but they also turned on something different. We can see from RNAseq in those tumors that although they turned off all the AR genes, they turned on many of those lineage plasticity genes, including neuroendocrine and JAK/STAT genes. We have confirmed this in all the in vivo models, including patient-derived organoid models using Corning 3D Matrigel. More importantly, this is not just a turning up of gene suppression; it reflects the ability of those tumors to become more stem-like and more metastatic. This is just one of the in vivo data we actually confirmed, which is a classic assay to examine the stemness of prostate cancer tumors. We can see that with ZNF397 depletion, those tumors become much more stem-like and easily cause new tumors and resistant tumors.

What is the driver of all this plasticity? We did a CRISPR library screening and found that the top gene regulating this plasticity is a gene called TET2. What is TET2? TET2 is a very important epigenetic rewiring regulator, regulating methylation from 5mC to 5hmC. We see that TET2 is definitely required for this resistance and the lineage plasticity. If we take out TET2 by CRISPR, we can reverse resistance and lineage plasticity. But how does TET2 do this? I will just jump through all the detailed molecular mechanisms and tell you the conclusion. We found that ZNF397, very interestingly, binds to both AR and TET2 but plays exactly opposite functions. ZNF397 is a very important co-factor for AR activation, but on the other hand, it is a blocker for TET2 function. When ZNF397 is present, it blocks TET2 from binding to many of the sites.

To delve into more detail about the mechanism of how TET2 and ZNF regulate resistance and lineage plasticity, we collaborated with our friends Dr. Felix Fang and Martin and examined the patient data of their 5hmC profile. We can see two things. First, 5hmC is highly correlated with those lineage plasticity genes in mCRPC and small cell neuroendocrine cancers. They are highly added to those neuroendocrine and lineage plasticity gene loci. On the other hand, in adenocarcinoma of the prostate, they are highly on the luminal genes. We went one step further. We segregated patients based on ZNF levels, high versus low. We can see a strong correlation between ZNF levels and lineage plasticity. When you have low ZNF397, all 5mC are on these lineage plasticity genes. When you have high ZNF397, they are highly towards the AR-positive and luminal lineage.

More importantly, we used our in vitro and in vivo models to prove that this epigenetic rewiring is indeed driven by TET2. Once ZNF397 is deleted, you see all this adding of 5hmC [inaudible 00:09:13] genes, and when we delete TET2 on top of it, you reverse it. Knowing that ZNF loss and TET2 causes resistance, can we do anything to reverse it? The answer is yes. We use a TET2 inhibitor called Bobcat, which can largely slow down the resistant tumor growth and almost completely reverse the gain of lineage plasticity. We have confirmed this in many different models, including the patient-derived organoids. We can see that those organoids do not respond to enzalutamide, which is a standard of care AR-targeted therapy, but they highly respond to the combined therapy of enzalutamide plus Bobcat.

This is the take-home message. We think we finally found a very crucial downstream molecular switch of AR-driven luminal prostate cancer versus lineage plasticity prostate cancer. It’s called ZNF397. We show that both genetic and pharmaceutical inactivation of TET2 could eliminate therapy resistance and lineage plasticity, which also explains the paradoxical roles of TET2 and 5hmC epigenetic rewiring in prostate cancer. This work was just published in Cancer Discovery last month. Thank you for listening.

Andrea Miyahira: Well, thank you so much, Ping, for sharing that. So what is the frequency of prostate cancer with ZNF397 and TET2 alterations or activity? And have you evaluated patient samples before or after transition to NEPC or lineage plasticity for their levels and activity?

Ping Mu: Yes, that’s a very good question. As I briefly mentioned in my slide, it’s around 10 to 25% of patients who carry the ZNF397 depletion. Around 10% in primary and 25% in mCRPC. You can see a much higher ratio in resistant and metastatic prostate cancer, and TET2 patients normally carry 10% TET2 alterations. These two populations do not overlap. So you can see that’s a very substantial number of patients. In collaboration with Felix and Martin, we examined the correlation of patient epigenetic rewiring and lineage plasticity with ZNF397. You see a very high correlation between ZNF levels and lineage plasticity in those patient samples. The less ZNF you have, the more lineage plasticity you have, or the more neuroendocrine you become.

Andrea Miyahira: Okay, thank you. Does this study shed any insights into how lineage plastic prostate cancer could be reversed?

Ping Mu: Yes, that’s exactly the goal of this study. As you can see, I briefly showed in the data that we used CRISPR-based genomic editing and pharmaceutical agents to target TET2 to regulate epigenetic rewiring. With both of these approaches, we can reverse lineage plasticity and reverse the resistance.

Andrea Miyahira: Okay, thanks. And have you evaluated the relationships between TET2 and any other lineage plasticity drivers?

Ping Mu: That’s a great question, and we indeed examined the relationship of many known drivers of lineage plasticity. For example, JAK/STAT activation is one of the known key drivers of lineage plasticity, and we surprisingly saw that many of those JAK/STAT activation genes are highly correlated with the downregulation of ZNF397 and activation of TET2. So it makes it very likely that ZNF is a downstream molecular switch of JAK/STAT activation, which is how JAK/STAT activation could cause this lineage switch at the bottom.

Andrea Miyahira: And how do you envision targeting ZNF397 or TET2 clinically? Are there treatment combinations that might be most effective or strategies to exploit this ZNF397 TET2 molecular switch?

Ping Mu: Yes, that’s definitely what we want to do. As we show that pharmaceutically targeting TET2 is very efficient in reversing lineage plasticity and resistance. More importantly, ZNF levels, as I mentioned, are highly correlated with [inaudible 00:13:26] resistance. So this expression could be used as a predictive biomarker to pre-select those patients who are likely to develop resistance and are more likely to benefit from this new therapy.

Andrea Miyahira: Okay. Well, thank you so much, Dr. Mu, for sharing this, and I look forward to your next study.

Ping Mu: Thank you so much, Andrea, for giving me this opportunity to share this work.