Read the Full Video Transcript

Andrea Miyahira: Hi, everyone. Thanks for joining us today. I'm Dr. Andrea Miyahira at the Prostate Cancer Foundation. Today I'm joined by Dr. Amina Zoubeidi, a professor at the University of British Columbia. She'll be discussing her group's research on prostate cancer lineage plasticity and will highlight her recent paper, "ASCL1 is activated downstream of the ROR2/CREB signaling pathway to support lineage plasticity in prostate cancer," that was published in Cell Reports. Thanks so much for joining me today, Dr. Zoubeidi.

Amina Zoubeidi: Thank you so much, Andrea. Thank you for having me, and thank you for highlighting my recent paper in this forum. It is my great pleasure today to share our recent paper published in Cell Reports, as Andrea already mentioned. This one is very specific to what we understand about how ASCL1 is activated. Before starting, I would like to give you a synopsis of what lineage plasticity is, and where we fit in and what necessarily regulates lineage plasticity in prostate cancer post hormone therapy.

To start with, I would like to tell you that there is a non-genetic mechanism, still underappreciated, playing a role in therapy resistance. And what that means specifically for lineage plasticity is that, for example, when we talk about non-genetic mechanisms, we can think about Darwinian fitness or a hierarchical model or a small population of cancer cells capable of resisting through clonal evolution and expansion following treatment.


But what interests us more is this adaptive theory. In the adaptive theory, when you have the effect of lineage plasticity, it means that this process is not new. It's not only related to neuroendocrine but we studied it very well in the 2000s, specifically when we studied metastatic events about epithelial to mesenchymal transition, and that process also holds for lineage plasticity in prostate cancer and the development, for example, of aggressive neuroendocrine prostate cancer.


This adaptive theory suggests that resistance is dynamic and reversible and can translate to any epigenetic and transcriptional control to support lineage switching. What makes this adaptive theory interesting is the fact that it can be dynamic and reversible. It mainly refines the sweet spots where we can perhaps use combination therapy to reverse the treatment or even block this lineage plasticity and aggressive disease.


Based on this theory and work that has been previously published by other colleagues in the field, we went on to examine what hormone therapy is doing. And what it is doing in these adaptive or trans-differentiation states? We looked at the chromatin landscape and how we did that was, we know that hormone therapy can induce changes based on different studies, and we were interested to see what is in the open chromatin or which part of the chromatin transcription factors can bind to. For that, we took cells, specifically adenocarcinoma cells, and treated them at different times with hormone therapy. Here we used maximum androgen blockade including enzalutamide. What we observed is that in adenocarcinoma, for instance, we have androgen receptor response elements at the top that are highly enriched. With treatment, we start seeing, for example, androgen response elements going down, and we start finding regions where transcription factors like ASCL1 are getting enriched at the top of this motif analysis.


But what was very interesting is that after only 10 days of treatment with hormone therapy, we can see the same quality in the landscape as we see in neuroendocrine prostate cancer, suggesting to us that perhaps if we examine how these factors are working, we can lead to a better understanding of neuroendocrine prostate cancer. This will also give us an opportunity to identify factors or potential targets but also markers of the development of this lineage plasticity, which in the future we can work on to block. To make the story short, we previously published a paper showing how ASCL1, a neural transcription factor that is open in the chromatin, leads to the upregulation of ASCL1 when hormone therapy is applied. It can play an important role in this lineage plasticity and also in the effects of endocrine differentiation.


What we found interesting is that ASCL1 can regulate this lineage plasticity but also, when targeted, can reverse it to adenocarcinoma via its ability to activate a factor called UHRF1, which acts as a gatekeeper for AMPK. When we have this gatekeeper blocking AMPK, we're also blocking the lineage of adenocarcinoma. Once we target ASCL1 in this situation, it's important to note that this transcription factor is really small and very hard to target with a small molecule inhibitor. What we found is that when we knock it down, we affect, for example, the activity of AMPK, which makes EZH2 remain in the cytoplasm. This shows us that there is beautiful crosstalk between ASCL1 and the epigenome to support lineage plasticity.


But think about it; now we have the opening of the chromatin and activation of certain transcription factors, but how can we sustain this activity? A transcription factor is expressed, but we also need upstream signaling to activate them and maintain the sustainability of activation of this neural transcription factor to support lineage plasticity. We started looking at what the potential transcription factors could be.


With a lot of findings and examination, we found, for example, this receptor tyrosine kinase ROR2 to be upstream of ASCL1. In this paper, I'm going to delve into how we looked at it and how we found it. For instance, the first thing I want to share with you about ROR2 is that it is a member of the non-canonical WNT signaling pathway. It is well-known that the WNT signaling we talk about in prostate cancer is mostly the canonical one; this is a non-canonical one. It is essential for embryonic development, activates AKT, MAPK, and PKC signaling, and drives tumor development and progression as well as metastasis, invasion, and therapy resistance. It is important to highlight that it is also upregulated, for example, in ovarian cancer but also in metastatic squamous cell carcinoma.


We found that ROR2 can be induced by hormone therapy. This is not only in cell lines but also in patients treated with neoadjuvant hormone therapy, especially when treated for three months with enzalutamide. Here we have patients who are high-risk and treated in a neoadjuvant setting, and ROR2 comes up as one of the top 10 receptor tyrosine kinases to be upregulated. When we apply algorithms, we find that this receptor tyrosine kinase comes to the top and is also upregulated in neuroendocrine versus adenocarcinoma, suggesting to us that it might have a role in neuroendocrine cancer. Specifically, we find it's more about ROR2 rather than WNT5A. I would like to highlight here that ROR2 is very specific to this AR-negative and could also be neuroendocrine-negative, but all fall under lineage plasticity. When we look at WNT5A, we find it's scattered; it doesn't constitute a subtype and is not very specific to only lineage plasticity but we see it throughout CRPC.


This was when we were thinking that perhaps the presence of ROR2 alone can make it work alone without WNT5A. However, when we look and combine ROR2 and WNT5A, we find that they can induce lineage plasticity when they are together. Basically, what I want to say here is that if we overexpress WNT5A alone in CRPC cells but do not express ROR2, we're not inducing lineage plasticity. But once the receptor is there, we can appreciate how WNT5A can enhance this lineage plasticity. This means that there is a need for this receptor for lineage plasticity to occur, and in the absence of ROR2, WNT5A doesn't assume this function, which definitely supports the data we see in patients. Again, overexpressing ROR2 alone induces lineage plasticity.


But how does this help me? When we took an unbiased approach, we found that ROR2 can regulate lineage plasticity and also the hallmarks of lineage plasticity, including EMT and also epigenetic factors including EZH2 and the regulation of methylation, along with neuronal stem cell pathways that can be activated. This data is not specific to cells but also in comparison, we find it to be true; activated genes and pathways are similar to what we see in the transition from adenocarcinoma to NEPC. But now when we look at NEPC patients, for instance, or patients with lineage plasticity and we stratify them based on high versus low levels of ROR2, we find that those with high ROR2 are activated with ASCL1. Not only with ASCL1 but also with EZH2 and other neuronal differentiation targets.


Looking at that tells us that there is the potential for ROR2 to activate downstream factors like ASCL1 as well as EZH2. We went on to examine what those downstream factors could be and specifically looked at transcription factors. ASCL1 came to the top as one of the targets of ROR2. To validate our study, the overall expression of ROR2 induces a nice upward migration of ASCL1 when it's lost in adenocarcinoma and when it's lost in neuroendocrine, completely upregulating the expression of ASCL1, not only at the expression level but also at its activity level, which is normal to see. When we target ROR2, we find that it has the same transcriptomic profile as ASCL1, which is quite staggering to see, suggesting to us that the ROR2 pathway in the context of NEPC is working downstream of ASCL1. The fact that we see that we're obliterating all ASCL1 signaling pathways or ASCL1 activity only by affecting ROR2 downstream was really interesting to us.


Why? Because, as I told you, ASCL1 is a small transcription factor that is very hard to target; it has a very tight interaction with DNA. The fact that we have something upstream can help us to target the upstream factor and target this subtype of neuroendocrine cancer that is enriched with ASCL1. Yet again, we went on to find what is definitely the downstream mechanism of ROR2, and we find that ROR2 activates ASCL1 specifically downstream of the MAPK pathway. Until recently, this MAPK pathway was also shown to be upregulated in NEPC compared to adenocarcinoma in two different datasets and correlated very well with ROR2 as well as with ASCL1. We went on to ask, "How is that working? And what is the factor that can lead to the activation of ASCL1 upstream?" We used the Cistrome toolkit and looked for potential transcription factors and, looking for the regulator potential score, we found that CREB could be the potential transcription factor that can regulate ASCL1 downstream of the MAPK pathway.


To confirm, downregulation of CREB leads to the downregulation of ASCL1. Again, this brings us back to this beautiful signaling pathway, suggesting that we have ROR2 that activates MAPK, and MAPK, we know, can activate CREB phosphorylation, and that can lead to the activation of ASCL1. To make the story short here, without going into details, we found that we have some CREB consensus motifs on ASCL1, and these consensus motifs are very important for the activity of ASCL1 downstream of ROR2 and downstream of ERK. And if you look at how it's binding, we find that we have a nice binding of phosphorylated CREB to ASCL1 that is very important for its acetylation, and this acetylation leads to the activation of ASCL1. In conclusion, we discovered that ROR2, a receptor tyrosine kinase and orphan receptor tyrosine kinase by itself, can activate ERK and phosphorylate CREB, which leads to the acetylation of ASCL1 and its activity, leading to lineage plasticity and neuroendocrine cancer.


We propose that targeting ROR2 in lineage plasticity would be a viable target and would allow work now on ROR2, for example, in other cancers that we might be inspired by, and using their inhibitors in the future. And thank you. Finally, I would like to thank the Prostate Cancer Foundation for investing in me early in my career, but most importantly, for investing in my research program and in lineage plasticity, which has helped me understand how lineage plasticity is an important mechanism of treatment resistance and the development of aggressive disease. We hope that this work will shed light on a potential pathway to be targeted in the future by potential drug conjugates as well. Thank you so much.


Andrea Miyahira: Thank you so much for sharing that with us, Dr. Zoubeidi. So just some questions. Is the ROR2 pathway a survival pathway, or is it only a lineage switch pathway? And do you expect ROR inhibition will kill tumor cells? Does it have a cytotoxic effect, or is it more about blocking the transition or maintenance of the NEPC phenotype?

Amina Zoubeidi: Well, thank you so much; this is a very important question. I think that we're talking about context specificity. The upregulation of ROR2, what we see with hormone therapy, is important to induce lineage plasticity. Definitely, this is important, but what is interesting is that targeting ROR2, perhaps in combination with hormone therapy, will be very important. The second thing we know is that ROR2 is also an androgen receptor-regulated gene, and combining those aspects, we know that hormone therapy, in which we know that the androgen receptor binds to ROR2 and inhibits it, when we use hormone therapy, we lift these inhibitory effects, and it becomes activated. Using ROR2 in combination with ASCL1, for example, with hormone therapy like enzalutamide or others, would be very important to induce lineage plasticity. However, by itself, it is important to block lineage plasticity, for example.

Andrea Miyahira: Thank you so much. So I know you're talking about ROR2 as being the therapeutic target, but of all the other things in the pathway, including ASCL1, is ROR2 the best target for this pathway, or are there others that might be promising?

Amina Zoubeidi: Well, I think that we cannot just come and say this one is the best thing; in the era of receptor tyrosine kinases, as you know, you could have an addiction to a receptor to be the best target. But the question for us here is, we don't have any treatment for ASCL1. Now we know that there is some DLR3, but how about the others that may respond to DLR3 or may not respond to DLR3? This could be a potential target, and I think that we should not exclude other members of the pathway. Perhaps it's going to be more about combination therapy, targeting now with P300, perhaps with ROR2 together, will have more effects because when we talk about CREB and P300 and the CBP complex, perhaps targeting that, but what about the side effects? That will be very interesting to look at. ROR2 also could have potential for detection rather than therapeutic. And I think that using some ADC targeting will inform us better about which one will be the best target. The big challenge with transcription factors, specifically small transcription factors, is where to bind and how to find a small molecule that can get there, something that can get to the nucleus, and that's what will be very interesting to target these pathways from different angles.

Andrea Miyahira: Thank you. WNT pathway alterations are seen in about 20% of CRPC, but this is a non-canonical WNT pathway.

Amina Zoubeidi: Yes.

Andrea Miyahira: So do we see somatic alterations in ROR2 or other non-canonical WNT pathway members in CRPC or NEPC?

Amina Zoubeidi: Well, we didn't find any somatic alterations in ROR2 in CRPC or NEPC, but what we did find is that there is some amplification of the ROR2 gene in about 4% of cases in the TCGA dataset. When we get to the metastatic setting, about 2% of metastatic CRPC cases show a 10% gene amplification of ROR2. This is very important to look at. Now, when you have neuroendocrine or what we call more lineage plasticity, it is a whole group that represents this resistance in androgen receptor non-driven cancers, and you find a subset that has ROR2. It is just beginning to see the potential of targeting ROR2. But definitely, it is a non-canonical pathway. The ligand, WNT5A, for example, is everywhere; it's not specific to that phenotype or specific to ROR2. You can have WNT5A without ROR2 overexpression, and you can have ROR2 overexpression without WNT5A. The combination of the two is powerful; however, it is not specific to a particular subtype. In respect to ROR2, it is mostly when you have androgen receptor-negative or inactivated cancers.

Andrea Miyahira: Okay. Thank you so much for sharing this really exciting paper with us, Dr. Zoubeidi. I look forward to your next publications and learning more from you.

Amina Zoubeidi: Thank you so much. And again, I would like to thank the Prostate Cancer Foundation for this opportunity and for all the support over the years. Thank you again.