Sangeeta Goswami: Good morning, everyone. Thank you Ashish for the kind introduction, and I'm grateful to you and the UroToday team for the invitation to speak today. Today, as you mentioned, I would like to focus on KDM6A and its role as a potential biomarker for therapy stratification in bladder cancer. When we look at the genomic landscape of muscle-invasive bladder cancer, we can see that out of the top five mutations, three mutations are in the epigenetic regulators, namely KMT2D, KDM6A and ARID1A. And they constitute close to around 50 to 60% of this population. So what my group is working on is identifying the therapeutic relevance of these mutations and how they modulate the therapeutic perturbations. And today, I would like to focus on KDM6A. KDM6A is a histone demethylase that plays an important role in bladder tumorigenesis. And this is very well studied. In patients with advanced bladder cancer, around 26% of patients harbor this mutation and this percentage is much higher in non-invasive bladder cancer, where it can go up to around 70%. So what does KDM6A actually do? It is a histone demethylase. So in condensed chromatin where genes are trimethylated in the H3K27 methylation mark, KDM6A catalyzes their demethylation and it opens the chromatin for gene expression. So what we focused on, on the knowledge gap that how does a recurrently mutated epigenetic regulator shape therapeutic response? To do that, we looked at multiple big retrospective data sets. And one thing we have noted consistently is that patients harboring KDM6A mutation have lower overall survival compared to patients who don't carry the KDM6A mutation to cisplatin-based chemotherapy.
On the other hand, interestingly, patients carrying this mutation respond better to anti-PD-1, PD-L1-based therapies in advanced bladder cancer. And to validate this clinical observation, we prepared CRISPR knockout KDM6A murine cell lines. And when we injected in the syngeneic bladder cancer model, we noted a very similar phenotype that mice harboring KDM6A knockout tumors do not respond to cisplatin as their wild-type counterparts. And similar to the human observation, we saw that they actually do respond better to immune checkpoint therapy compared to their wild-type counterpart, concluding that inactivating mutations in KDM6A potentially drive resistance to standard cisplatin-based chemotherapy. But it confers sensitivity to PD-1 immunotherapy. The next question we asked, what are the potential mechanisms of cisplatin resistance? To do that we looked at the retrospective WGS data from human bladder cancer, as well as our human KDM6A knockout cell lines. We performed whole genome sequencing. As you could see the data, we did further analysis. And in this analysis what we found was that in the absence of KDM6A, the tumor makes more ecDNA. There's extrachromosomal DNA that does not follow and have the regulation that the normal transcription does. And therefore we see increased copy number of the genes carried by those extrachromosomal DNA. And in the absence of KDM6A, we saw ecDNA carrying TP63,
Claudin-4 and other genes that are known to confer resistance to cisplatin. So our hypothesis is that in the absence of KDM6A, the tumor cells are producing more ecDNA carrying those chemoresistance genes that lead to resistance to cisplatin-based chemotherapy. Then we also tried to understand why KDM6A-mutated patients are responding better to immunotherapy. And one thing we noted in our retrospective analysis is that both in IMvigor and TCGA bladder as well as CheckMate-275 patients with KDM6A mutation had a higher tumor mutation burden. Then in our murine and human cell lines, we did ChIP sequencing, RNA sequencing. And one thing that came out is that although there is no concurrent mutation of the MSI genes such as EXO1, MSH6, MSH2, but in the absence of KDM6A, their transcription is less because KDM6A is important in transcribing these genes. So when you have an inactivating mutation, you have lower protein of these genes as you could also see in the immunohistochemistry on the left. So what we propose is that KDM6A loss reduces histone H3K27 trimethylation. And that downregulates genes such as EXO1, MSH2, MSH6, leading to genomic instability and higher tumor mutation burden. But interestingly, there is another layer of regulation. When we looked at this tumor both in the human as well as in our cell lines and mouse model, what we have consistently noted is that it regulates the histone lactylation and the lactate production. So we saw that there is a decrease in some key glycolytic genes like HK2, LDHA which leads to reduced lactate production in the tumor microenvironment of KDM6A-mutated tumors. And concurrently we saw that there are decreased regulatory T-cells, which are suppressive T-cells conferring resistance to immunotherapy. So when we tried to understand and link this phenomenon, we found that indeed the lactate in the tumor microenvironment is important in transcribing some of the key genes in regulatory T cells like PD1, FOXP3, and TGF-beta.
And that regulates the fitness and function of T cells. So to summarize this, in the absence of KDM6A, we see less lactate in the microenvironment, the lactate which normally regulates the transcription of key suppressive genes. And because the T-reg lose their suppressive identity, therefore we see an increased response to anti-PD-1 therapy. So finally, to put everything together, we feel that this KDM6A paradox, it's a double-edged sword in bladder cancer treatment. And we can utilize this to stratify patients because this sword is pretty blunt because when it comes to chemotherapy, it reduces the responsiveness to cisplatin-based chemotherapy. But with the genomic instability, higher tumor mutation burden and by regulating the microenvironment, these mutations improve survival to immunotherapy. So finally, I would like to thank everyone who was involved in this work, including our collaborators, all the funding agencies, and last but not least, our patients. Thank you and I'll be happy to take any questions now.
Ashish Kamat: Thanks so much, Sangeeta. That was a really elegant summary of really elegant work and it's hard always to distill such detailed, basic translational work into the way you did it. So thanks again for taking the time. You recognize the double-edged sword. And a question is, in the therapeutic world, since KDM6A+ drives resistance by ecDNA, and again it's genomic flexibility, do you think that there's a real risk that these patients and these tumors will evolve immune-escape clones that bypass the TMB advantage? Any thoughts there?
Sangeeta Goswami: Yeah, this is a great question, Ashish. It can because ecDNA, and that is one thing we know at least based on the current paradigm, that ecDNA also carries genes to regulate responsiveness to immune checkpoint therapy. And most of the time we see that it can account for resistance as well. But what I feel is that is context-dependent, just two ways to think about it. ecDNA can also be a reservoir for new antigen production. Although it can carry the genes for resistance to cisplatin as well as immunotherapy, but can it also lead to more new antigen production? So in that case, making it more sensitive to immune checkpoint therapy. But then as we evolve clonally, the bladder cancer evolves, there is always a risk that this high ecDNA can actually tweak the response to different kinds of therapeutic perturbations.
Ashish Kamat: Yeah, no, absolutely. And then again, I'm sure you've looked into this since the loss impairs Treg fitness by reducing lactate, is there potential to enhance the efficacy of PD-1 inhibitors with LDH and any thought there in the clinic?
Sangeeta Goswami: We recently also published another paper in Nature Immunology on histone lactylation and its role on CD8 T-cell function. And putting both the studies together, yes, there could be potential impact on LDH inhibitor. However, these are very promiscuous, these metabolic inhibitors because there are different other cell types, they require this endogenous lactate in terms of transcribing. So I do not think the lactate LDH inhibitor per se would be a good clinical target because of its global impact on the tumor as well as other immune microenvironment cells. We have to get more specific targets, I would say, or maybe epigenetic or other downstream metabolic target rather than LDH inhibitor because of its global impact on other cell types as well.
Ashish Kamat: Great. And then I'm sure you've thought about this and it's the million-dollar question or the billion-dollar question that everybody asks nowadays. With us moving away from pure cisplatin-based chemo, for example, you have GEM/CIS/NIVO or GEM/CIS/DURVA and now we have EV-Pembro. How is this going to factor into how you would use this in the clinic moving forward?
Sangeeta Goswami: So one question that this paper hasn't addressed and we are currently working on it, how KDM6A mutation might impact enfortumab? Does it impact NECTIN expression? Or what are the retrospective analyses we could do to look at its impact on EV? What I would say, once we know that that would help us in guiding for example, neoadjuvant therapy as you said, like GEM/CIS/DURVA versus EV-Pembro because both are available to us at this point of time and how do we sequence not only in the advanced metastatic setting but even in the perioperative setting. So then one thing for us now, like the next step what we are trying to understand is that okay, what is a net positive and net negative when we combine GEM/CIS/DURVA in a KDM6A-mutated patient? It should not respond to cisplatin as well, but it should respond better with immunotherapy. So what is the net positive, net negative? That is one thing we are trying to do in a murine model. On the other hand, if we feel that it is also improving response to EV by regulating NECTIN expression, which is purely a speculation at this point in time because this work is ongoing. So can we combine? Then if it does, then EV-Pembro becomes a natural choice for this cohort of patients, even in the neoadjuvant setting. So there's ongoing work, but I hope that with all the pieces together, it would help us to stratify patients in different settings in bladder cancer.
Ashish Kamat: Yeah, so I mean that's a good segue for me first off, thank you for taking the time. But also saying well, we'll have you back when you have that report and that publication. So looking forward to it. And thanks, Sangeeta.
Sangeeta Goswami: Thank you so much, Ashish. And UroToday Group, thank you.