Read the Full Video Transcript

Andrea Miyahira: Hi, I am Andrea Miyahira here at the Prostate Cancer Foundation. Today with me is Dr. Tanya Stoyanova, an assistant professor at UCLA. Dr. Stoyanova will discuss her group's most recent paper, "UCHL1 is a potential molecular indicator and therapeutic target for neuroendocrine carcinomas." This was published in Cell Reports Medicine. Dr. Stoyanova, thank you for joining us and I look forward to learning about this paper.

Tanya Stoyanova: Thank you so much, Andrea, for the kind introduction. It's really my pleasure to be here today and to have this opportunity to share our work with you. As Andrea mentioned, this was a study that was published just a couple of months ago where we identified this very interesting molecule as a potential new biomarker and therapeutic target for neuroendocrine prostate cancer as well as for other cancers with neuroendocrine phenotypes such as neuroendocrine carcinomas and neuroblastoma.

So neuroendocrine prostate cancer is rarely seen when it's at the stage of localized disease. However, we see this rapid increase of neuroendocrine prostate cancer in patients that develop castration-resistant prostate cancer and have undergone multiple rounds of anti-androgen therapies. So this subtype of cancer has a phenotype that is characterized by the expression of neuroendocrine markers and very often with also the loss of AR. So it is believed that this neuroendocrine prostate cancer arises through trans-differentiation from adenocarcinoma, which first goes through a mixed variant, and then it can eventually go through epigenetic reprogramming and other mechanisms into these small cell neuroendocrine prostate cancer phenotypes, which are characterized a lot by the loss of androgen receptor and neuroendocrine markers.

It can go back and forth, if you will, through this differentiation and trans-differentiation. So we set to identify new potential therapeutic targets for neuroendocrine prostate cancer because a lot of neuroendocrine prostate cancer is characterized by the loss of AR. Now these patients do not respond to anti-androgen therapies and there are very limited treatment options. So we set to identify new potential druggable targets for neuroendocrine prostate cancer. And what we did, we compared adenocarcinoma model cell line xenograft models to neuroendocrine prostate cancer xenograft models, and we performed proteomics. We started looking at proteins that are specifically elevated in neuroendocrine prostate cancer. We went one by one to look for druggable targets, any molecules that have experimental compounds or FDA-approved compounds for other conditions.

So this proteomics was published in 2020 and identified quite a few targets. One of them that sparked our interest is a molecule that wasn't that well studied and has a very interesting function in protein stability. This is UCHL1. So what it does, it has a dual function, so it can serve as a hydrolase to remove ubiquitin from proteins, which makes them more stable, or it can serve as a ligase where it can actually now add ubiquitin to proteins and lead to their degradation.

So how it chooses its substrate, which direction to go, is not known. However, the molecule has a couple of experimental compounds that were not tested in the settings of neuroendocrine carcinomas or neuroendocrine prostate cancer. So we first went ahead and validated these findings in patient samples. Here we compared patients' prostate benign tissues, localized prostate cancer, adenocarcinoma, castration-resistant prostate cancer, and neuroendocrine prostate cancer. You could see by this brown staining of immunohistochemical analysis that we see significant enrichment of UCHL1, specifically in neuroendocrine prostate cancer. Studies here from UCLA and other groups also have shown that there is this convergence and common transcriptional program and expression profiling between cancers with neuroendocrine phenotype that is shown here: small cell lung cancer, small cell prostate cancer, as well as other cancers such as neuroblastoma, for example.

So we went ahead and tested if this UCHL1 is specifically overexpressed in neuroendocrine prostate cancer or if it's also applicable to other tumors or malignancies with neuroendocrine phenotype. Here we compared non-neuroendocrine tumors versus neuroendocrine tumors. Those are from the GI, pancreas, and other organs. You can see that UCHL1 is specifically elevated in neuroendocrine tumors. It was also very high in patients with neuroblastoma. This is a childhood cancer that has a high expression of neuroendocrine markers.

We also saw high expression of UCHL1 in lung carcinoid and small cell lung cancer that both have this neuroendocrine phenotype but not in non-small cell lung cancer that lacks this neuroendocrine phenotype. So UCHL1 seems to be very specifically expressed in neuroendocrine prostate cancer as well as other malignancies with neuroendocrine phenotype as well as neuroblastoma. So another interesting characteristic of the molecule, it has been shown to be potentially able to be secreted outside of the cells.
So we were very interested to see if we can detect it outside of cells and if it can serve as a potentially minimally invasive biomarker to reflect the status of UCHL1 in tissues. So here we took media from cells. The red cells are cells with neuroendocrine phenotype, neuroendocrine prostate cancer, neuroblastoma, and some small cell lung cancer, and we can see that the UCHL1 is indeed secreted from the cells in the media.

We developed ELISA as a more high-throughput assay to detect UCHL1 and we were able to recapitulate these findings and be able to capture it in the media. We now went ahead and tested the serum. Those are the serum from mice bearing either adenocarcinoma, castration-resistant prostate cancer, or neuroendocrine prostate cancer. UCHL1 was specifically detected in the serum of mice with neuroendocrine prostate cancer.

The same when we looked at small cell lung cancer and also in neuroblastoma. This is serum from mice bearing these tumors, but we don't see it in the serum from non-neuroendocrine tumors such as non-small cell lung cancer. We went ahead and tested this in patients. We compared the serum from patients with metastatic prostate cancer and adenocarcinoma and indeed we saw enrichment of UCHL1 in the serum of patients with metastatic prostate cancer. Presumably, 20 to 30% of them would have a neuroendocrine disease.

So we see the same in small cell lung cancer. We do see it in patient serum but not in patients with non-small cell lung cancer. So this was exciting and we very much look forward to expanding the cohort in these studies. So we went ahead and asked if UCHL1 actually regulates tumor growth in neuroendocrine prostate cancer and other neuroendocrine carcinomas. Here is an example of neuroendocrine prostate cancer. We introduced gene deletion via CRISPR-Cas9 knockout, and we could see that when we knock out UCHL1, now those tumors grow significantly slower.

We also saw a decrease in neuroendocrine markers, synaptophysin, Chromogranin A, and CD-56. And when we overexpressed UCHL1 in an adenocarcinoma cell line, we saw an increase in tumor growth as well as an increase in these neuroendocrine markers, three neuroendocrine markers. So we also tested the effect of UCHL1 loss in metastatic colonization. Here, we labeled cells with luciferase and RFP. Some of the cells had wild type, their neuroendocrine prostate cancer model, and others had gene deletion of UCHL1.

So once we injected these cells in the heart, we could monitor metastatic colonization via bioluminescence or fluorescent imaging of organs. And what we observed was that when we introduced deletion of UCHL1, there was a very significant decrease in overall metastatic colonization by overall bioluminescence as well as when we analyzed each individual organ.

So we went ahead and tested the commercially available inhibitors of UCHL1. There are two of them; here, I'm showing you one example we tested in patient-derived xenograft models of neuroendocrine prostate cancer and cell line models of neuroendocrine prostate cancer. You can see a significant delay in tumor growth over time when we introduced inhibition of UCHL1. Similar results were observed when we tested the effect of UCHL1 inhibitors on metastatic colonization. You can see a decrease in overall metastatic colonization, and we also saw a significant decrease in the number and size of metastatic nodules at many different sites. Here I'm showing you the liver, because the liver is associated with worse outcomes as well as a neuroendocrine phenotype. You can see much smaller nodules and much fewer nodules upon UCHL1 inhibition.

Cisplatin, a platinum-based agent, is used for the treatment of patients with neuroendocrine prostate cancer as well as small cell lung cancer. So we went ahead and tested if the combination of cisplatin and a UCHL1 inhibitor could enhance therapeutic efficacy. Indeed, we saw that when we treated mice with neuroendocrine prostate cancer, patient-derived xenografts, the combination of UCHL1 inhibition and cisplatin had greater therapeutic efficacy over time. We also saw no toxicity, as measured by body weight. We also measured liver enzyme panels, liver histology, and kidney function. We saw no side effects from these combined treatments. So this was exciting to see.

So we further delineated the molecular mechanism of how UCHL1 regulates tumor growth in neuroendocrine prostate cancer, neuroendocrine carcinomas, and neuroblastoma. What is the mechanism of its function and how does it potentially regulate this neuroendocrine phenotype? Because it's involved in protein stability, we did proteomics upon modulation of UCHL1. We knocked down UCHL1 in the context of neuroendocrine prostate cancer and looked at changes in proteins that decreased. Of course, UCHL1 was the top target that was knocked down.

One interesting class of proteins that we identified as decreasing upon knockdown of UCHL1 was the nuclear pore complex proteins. This was very interesting for us to see, as there was a very good study in 2018 showing that these nuclear pore protein POM-121 actually regulates the import of transcriptional factors in castration-resistant prostate cancer.

So we went back and looked at the levels of POM-121 in TMPO. This was also identified in our proteomics. This is a nuclear envelope protein and a nuclear pore complex protein. So indeed, when we knocked out UCHL1, we saw a significant decrease in POM-121, TMPO, as well as neuroendocrine markers. We tested the stability of POM-121, and indeed, when we introduced the loss of UCHL1 and inhibited protein synthesis by cycloheximide, we saw that POM-121 half-life or stability significantly decreased.

So it seems that UCHL1 stabilizes POM-121, and when there is a loss of UCHL1, it is indeed destabilized. We tested the levels of E2F1 and cMYC in the cytoplasm and nucleus upon loss of UCHL1 or inhibition of UCHL1. And indeed, we saw a significant decrease in the nuclear import of E2F1 and cMYC upon loss of UCHL1 or inhibition of UCHL1. So we had this model that UCHL1 can actually remove ubiquitin from POM-121, stabilizing it on the nuclear envelope, allowing these two transcriptional factors to enter the nucleus and really drive the growth of this neuroendocrine phenotype and potentially the neuroendocrine phenotype itself.

Another interesting observation was that UCHL1 can at the same time degrade P53. P53 loss is characteristic of neuroendocrine prostate cancer. So what is interesting is that we found that UCHL1 binds P53. When we overexpressed UCHL1, we saw a significant decrease in P53 half-life, so it gets degraded much faster. If we overexpressed a mutant of UCHL1 that doesn't have the ability to bind and have the ligase activity, then we saw that P53 stability is very similar to the wild type.

We looked at ubiquitination since UCHL1 is known to serve as a ligase. And indeed, when we overexpressed UCHL1, we saw a significant increase in the ubiquitination of P53. So another part of the mechanistic function of UCHL1 is its ligase function, which actually targets P53 for degradation.

So this is just a summary of our findings. I did share most of the work on neuroendocrine prostate cancer. As I mentioned, this is applicable to neuroendocrine carcinomas, small cell lung cancer, and neuroblastoma. In the study, we reported across these three different models that UCHL1 is a promising plasma or serum biomarker. We've seen it in both. It's also a tissue biomarker for neuroendocrine carcinomas and neuroblastoma. In terms of mechanistic insights, it can serve as a hydrolase and remove ubiquitin from POM-121, which stabilizes it and allows transcriptional factors to enter, and at the same time, it serves as a ligase to attach ubiquitin to P53, leading to its degradation.

And very excitingly, we saw that this UCHL1 inhibitor actually has therapeutic efficacy in these neuroendocrine carcinomas, and when combined with cisplatin, we can enhance the therapeutic efficacy. I would like to thank my lab, of course, for doing the hard work on this study. This study was led by Laura Liu in my lab. Of course, our funding sources and all our collaborators, without whom this work would not have been possible. I would be delighted to take any questions or hear any suggestions. Thank you.

Andrea Miyahira: Thank you so much, Dr. Stoyanova, for sharing this study. Did your team evaluate whether UCHL1 impacts epigenetic mechanisms that are involved in lineage plasticity or neuroendocrine progression?

Tanya Stoyanova: We have not. We would be very interested in doing that. We haven't looked at epigenetic changes. The reason is that from the proteomic analysis that we performed, we didn't see changes in epigenetic factors per se. That doesn't mean that there are no epigenetic changes; it's something that we would like to explore in the future. The only reason we focused on the nuclear pore complex is because that was really the top pathway that was altered based on our proteomic analysis. But I agree that it would be very interesting to see, upon modulation of UCHL1, if there are any epigenetic changes, and indeed if this can be potentially reversed as UCHL1 is being modulated.

Andrea Miyahira: Thank you. Do you know whether UCHL1-driven mechanisms are reversible? For instance, can blocking UCHL1 reverse to adenocarcinoma and reinstate sensitivity to AR-targeted therapy?

Tanya Stoyanova: Yes. So we see the reversal of neuroendocrine markers and AR potentially coming back at a later time point when we knock out UCHL1. With the inhibitor, we do see the reversal of neuroendocrine markers about three weeks post-treatment. We haven't tested sequential therapies; if you will, can we treat with a UCHL1 inhibitor and then come back with AR to see the response? But that's something that we would like to explore. But it is reversible, yes.

Andrea Miyahira: Okay, thank you. Given that NEPC is heterogeneous and there are no single markers that are shared across all cases, do you see if UCHL1 increases our understanding of different NEP subtypes, for instance? Or are there any other NEPC markers or drivers that either overlap with or are mutually exclusive with UCHL1?

Tanya Stoyanova: So that's very interesting because if we look at the neuroendocrine prostate cancer tissues that we analyzed, about 60% of them have very high and in a way homogeneous levels of UCHL1. But we have some patients that are neuroendocrine prostate cancer, but they don't have high levels of UCHL1. We have not yet figured out why that is and what the difference is between these two patients. We would be very interested in doing that because in prostate cancer, we see that.

In neuroblastoma, for example, 95% of the patients are positive, so there it's very clear that these are the positive cases. In small cell lung cancer, we also see many more patients with UCHL1. In prostate cancer, it's 60 to 40, so this will be very interesting to explore what it correlates to and what the differences are between these cases of neuroendocrine prostate cancer that actually have a small cell neuroendocrine phenotype. Those are from Duke University and they were assigned small cell neuroendocrine prostate cancer, but yet UCHL1 is high in only 60%. So this would be a very interesting study to conduct. We haven't done it yet, but it's a great question and we should really explore further what the difference is between these two patient groups.

Andrea Miyahira: Thank you. And do you have any translational plans for these findings, and what are the next steps for your study?

Tanya Stoyanova: So the next steps we are excited about involve two different avenues. One is to further explore the utility of UCHL1 as a minimally invasive biomarker. So we are working now with other collaborators to see if we can get patients with confirmed neuroendocrine prostate cancer to really expand the cohort. This is one aspect. The other aspect is that we are working on developing a new compound that's more potent and more effective, hopefully, than the two commercial compounds that we have already tested. So we do have some lead compounds and we are very excited about them. So developing a new compound as well as testing it for use as a biomarker.

Andrea Miyahira: Okay. Well, thank you. This is all so exciting, and congratulations again on this paper.

Tanya Stoyanova: Thank you very much, Andrea. Thank you for having me here with you today.