Read the Full Video Transcript

Pedro Barata: Hello. I'm Dr. Pedro Barata, GU oncologist at University Hospital Seidman Cancer Center in Cleveland. I'm very, very happy to be joined by Dr. Yousef Zakharia, a colleague and friend and kidney cancer expert. He's a full professor of medicine at Iowa University. He's also the director of the phase one and the co-leader of the gene malignancy there at the Holden Comprehensive Cancer Center. Such a pleasure to have you here, Yousef.

Yousef Zakharia: Hi, Pedro. Thank you for the kind invitation. It's always a pleasure to talk to you.

Pedro Barata: Absolutely. We're very happy to get a minute to chat, and I would like to focus on your great work that you and the Big Ten important consortium did. It is actually a proof of concept, phase one, two trial combining two therapies, right?

Yousef Zakharia: Correct.

Pedro Barata: I would like us to tell... By the way, you presented that data. We've seen that data being presented, also published in Nature Communications, quite a prestigious journal. Very important study, provocative concept here. I would love for you to share with me and the audience here how that concept was born and what the main results of the study were?

Yousef Zakharia: Yeah. Sure. Happy to do so. Thank you again, Pedro. It's always nice to talk to you. Indeed, this clinical trial is a multicenter collaboration through the Big 10 Cancer Consortium. It's combining durvalumab, plus guadecitabine as a hypomethylating agent in metastatic kidney cancer, whether treatment-naive or immune checkpoint inhibitor-naive or immune checkpoint inhibitor-refractory. I wanted to give a shout-out to all my co-investigators listed here on this, mainly Dr. Ajjai Alva from the University of Michigan, who was the power behind the delivery of this clinical trial. As a way of background, as we know, DNA hypomethylation is associated with tumor progression, higher stage grade, and can carry poor prognosis in different malignancies. Hypomethylating agents have been utilized extensively in hematologic malignancy, but not so much in solid tumors, although there are multiple efforts to try to deliver it in solid malignancies.

We do have multiple clinical trials to suggest the synergistic effect between hypomethylating agents and checkpoint inhibitors. More so, few studies suggested that the hypomethylating agents can increase or lead to higher levels of chemokines CXCL9, 10, and can result in reversal of immune evasion and can help in tumor regression, basically. As we know, guadecitabine is one of the hypomethylating agents, and that was utilized in combination with durvalumab and type PD-L1 in our patient or in our clinical trial. This is a study design for the clinical trial. It's a smaller study pilot proof of concept phase one, two. Two cohorts after a quick false phase one escalation, de-escalation. Basically, we enrolled into two cohorts. Cohort one, which was about 36 patients, and those were checkpoint inhibitor-naive patients. They were allowed to be either first-line therapy or second-line therapy as long as they did not see a checkpoint inhibitor in the past.

The cohort two, which was a smaller cohort, was in a checkpoint inhibitor-refractory setting. Up to two prior lines of therapy, checkpoint inhibitors should have been one of them. Initially, quick dose escalation, de-escalation, phase one, only six subjects, followed by the dose expansion. Basically, what we learned in the dose escalation was starting at 60 milligrams per meter squared and then day one through five. Basically, Monday through Friday, and then we add durvalumab the following Monday. Basically, all happening in an outpatient setting, basically. The first dose at 60 milligrams per meter squared, we did encounter a DLT, dose-limiting toxicity with significant grade three, four neutropenia. Basically, we decided to go with 45 milligrams per meter squared of guadecitabine as a recommended phase two. This was the dose that was tested with the standard of care, durvalumab at 1500 milligrams IV.

Obviously, we were observing our patients closely every eight weeks with CT scan and treatment was continued until disease progression or unacceptable toxicity. This is a top-line baseline characteristic, if you will. The median age and the distribution male to female do mimic what we know already in metastatic kidney cancer here. All our patients were intermediate and poor risk characteristics per IMDC criteria. Majority of them were clear cell RCC, above 60%, and some of them have some mixed histology sarcomatoid differentiation in about 20% of cases. We will discuss shortly the toxicity profile encountered in the dose expansion, but as I mentioned in the dose escalation, neutropenia was the major dose-limiting toxicity here.

From an efficacy standpoint, basically we did encounter some interesting signals here. Just to rewind a little bit, our goal or primary endpoint of the clinical trial was objective response rate in cohort one in patients with checkpoint inhibitor-naive therapy, where the goal was to improve it to 45%, basically. Now, bear in mind, this trial was conducted through COVID and in an era where even IO/TKI was not widely utilized at that point. The primary endpoint was not met, basically, so the ORR of our patient population was 22%, so it was not met.

However, we did encounter a few interesting observations here in our patients who have... Or in the cohort one, basically, the median PFS was 14 months. This was in 36 patients in cohort one, which was quite interesting to see. We did have a lot of patients who had stability of their disease for a prolonged period of time. You will encounter those necrotic type of responses, and we always talk about the limitation of resist in kidney cancer where you are achieving the necrosis, but yet you are not shrinking the mass. That can impact the overall response, but yet the patient can have a durability of their response technically counted as stable disease. Really, the PFS signal that we encountered here of 14 months was quite interesting to see in this relatively smaller patient population with 36 patients.

In refractory or checkpoint inhibitor-refractory patients, which was about 16 patients, the PFS signal was not as strong. It was only about like two and a half, three months median PFS, but we did encounter stability of disease in about 60% of those patients. Some of those responses, as you can see here in the swimmer plot, the gray by the way represents the cohort one checkpoint inhibitor-naive and the purple is cohort two. Many of the cohort one did have ongoing responses lasting beyond one year. The median or the overall survival percentage at one and two years were about 95 and 89 respectively in the cohort one patient population. Quite an intriguing signal to see.

Now, in looking at the adverse events that we encountered on this clinical trial, really nothing life-threatening per se. We did encounter quite significant neutropenia, which grade three, four was encountered in about 38% of patients. No severe life-threatening infection as a result of this. A lot of times, GMC-SF took care of this. We did encounter asymptomatic elevated lipase in about 10% of patients, but really the rest of the side effects, as you can see here, were fairly tolerable and did not seem to be very life-threatening per se.

We did do extensive work on biomarker work looking at peripheral blood. We did have actually a goal to do baseline and on-treatment biopsies. However, unfortunately, this study was conducted through COVID and a lot of those research procedures in different hospitals were shut down, basically, as you remember, yearly, COVID. All our biomarker work was related to peripheral blood and this is a flow cytometry studying the relationship between the immune cell phenotype and the type of response we have seen. Interestingly so, we have seen an inverse relationship between MDSC, myeloid-derived suppressor cells, which we know are immune-suppressive cells along with a Treg here in Figure L. Inverse relationship between response and the level of these cells, which is not surprising because we know these types of cells are immunosuppressants.

Lower levels, basically, are expected to be seen in patients with a good response in this situation with, as you can see here, in chart A and L, an inverse relationship between Treg and responders. The other observation we have made is higher levels of interferon-gamma in CD8 T-cells and TNF alpha in different CD48 and innate lymphocytes. Higher expression of those inflammatory markers in responders compared to non-responders. Again, this is nice to see that patients who have... It goes well with what we know, with the higher interferon-gamma and TNF alpha, you are expecting better responses. We did look at the transcription factor, the RARgamma T in CD8 T cells, which can be markers of pro-inflammatory basically. Same thing, we observed higher expression in those patients who had benefit whether with stability of disease or PR. Those were quite interesting observations to see on peripheral blood flow cytometry.

Then, we looked at the relationship between the immune cell phenotype and immune-mediated toxicities. Per CTCAE, basically we define them as no toxicity at all, mild toxicity, or severe toxicity, which was like a grade three and above, basically. Interestingly, we did encounter in patients who have severe toxicity, higher levels of IL-2 in the innate lymphocytes, but not in CD4 and CD8, and then that was at baseline. Then, in those patients who have severe toxicity, this level of IL-2 was going up with therapy at cycle two, day eight compared to baseline, but that was not the case in patients who did not experience significant side effects or had only mild side effects. Same thing when it comes to interferon-gamma with treatment. Basically, mainly interferon in the CD8 T cells, the level of interferon-gamma was going up in patients who achieved severe toxicity as opposed to the ones who had no toxicity or mild toxicity.

Next, we looked at the methylation, guadecitabine hypomethylating agents. We are looking at what is it doing at the level of methylation and the surrogate for that, we used line one which refers basically to repetitive elements of DNA forming around 17% of the genome that can be a surrogate for global DNA methylation. We studied that level at cycle one, day one and after treatment and cycle two, day eight. We encountered a significant decrease in methylation before and after treatment as you can see here, alluded to in the orange color. That was an interesting observation and reassuring to show that guadecitabine is doing what it's supposed to do and we are able to measure that in peripheral blood.

Next, we looked at the chemokine levels. As we discussed earlier, the preclinical data suggested hypomethylating agents can increase different chemokine levels, CXCL, 10, and 11. We did study that in peripheral blood and compared it between day one and cycle two, day eight after two months' worth of treatment. We have observed quite a dramatic increase in different levels of CXCL9, 10, and 11. That correlated with better response in a smaller number of patients. Obviously, that was an early observation that... We observed that in patients who had higher levels of CXCL 9, 10, and 11. It did correlate with better progression-free survival, if you will. That needs to be studied in a larger patient population.

In summary, we thought the combination of guadecitabine and durvalumab is quite interesting, and the style of toxicity profile we have seen is acceptable, if you will. For us in oncology, we deal a lot with quite toxic medication or drugs. Outside that grade three, four neutropenia, which seemed to be common and mitigated with GM-CSF, we thought this combination was a relatively acceptable toxicity profile. While it did not meet its primary endpoint of improving ORR to 45%, our pre-specified goal, the observation of prolonged PFS in those patients who are checkpoint inhibitor-naive was quite interesting to see and we wish to see larger studies with this combination to validate our findings. With that, I would like to thank you and I would be happy to take any questions.

Pedro Barata: Yeah. Thank you so much. This is great. Great summary, my friend. I should highlight that, in fact, for folks who want to go over the details, the paper is available and we will provide the link for the paper again, published in Nature Communications. I think folks can get access to that data. Right? Let me ask you really quick... You were very thorough. Thank you for that. Just a question or two, I noticed... I agree with you. Perhaps responses were coming in line with immunotherapy by itself, right? I usually quote 25, 30% or so. We have great data with each different checkpoint, that durvalumab, but I think I agree with you in your comments regarding response rate. As you said, it's interesting to see the PFS perhaps better than we would think. Right? Particularly, when we look at cohort one, I think you have around 20% of people with non-clear cell histologies. If you look at the signal specifically for clear cell or whether or not that was actually looking better and if that's something you would... In your further research, would focus on?

Yousef Zakharia: Right. Right. Great question. Thank you for asking. Actually, I should have clarified that it was all clear cell, but it allowed mixed histology, meaning sarcomatoid or rhabdoid. We did not enroll patients who had non-clear cell like papillary or chromophobe or other tumors were not included.

Pedro Barata: Got it.

Yousef Zakharia: It was all clear cell RCC.

Pedro Barata: Got it. Clear cell competent and what was the dominant in the sample? Is that something you found?

Yousef Zakharia: Correct. Correct. Clear cell was the dominant, but we did allow sarcomatoid differentiation or rhabdoid.

Pedro Barata: Right. Okay. Got you. Is it my understanding too, Yousef, that the safety that you see is mainly driven by your [inaudible 00:18:09]? Is that special, the neutropenia-

Yousef Zakharia: Correct.

Pedro Barata: ... which I believe is something we tend to see?

Yousef Zakharia: Correct. Correct. That's a common finding actually with hypomethylating agents, the neutropenia. It is specific to that and we do not believe it's related to immunotherapy. Correct.

Pedro Barata: Got you. What is your sense... Because we do expect a little bit of a tail of the curve, right? With checkpoint by itself. Is that what you see with long follow-up for these patients we enroll in the study?

Yousef Zakharia: Indeed. Yeah. We had a lot of patients who achieved their best response as stable disease and kept going for a prolonged period of time beyond two years with stability of their disease. We did have a tail of that curve that we did not present, but we are hoping to... Hopefully, in a future publication, we can present a longer follow-up.

Pedro Barata: The long-term follow-up. Yeah. No, that's super important. Yousef, before I let you go, let me just ask you one final question. I would think that if you want to leverage this further, I would predict maybe it would be a path forward to, number one, think of a different checkpoint like a PD-1 inhibitor that might be already available in the RCC arena. Just a reminder for those who are listening to us, that durva has not been able thus far to be available as standard of care. Also, I'm wondering if there's any preclinical data combining this hypomethylated agent with a different target approach. I don't know your thoughts about that, if you look into that at a preclinical level. I understand DLTs might be problematic because some of the target therapies can also give you a little bit of that. I'm wondering your thoughts, have you seen any preclinical data justifying a different combination approach beyond a checkpoint? What are your thoughts?

Yousef Zakharia: That's a very interesting question, actually. I'm not aware of preclinical data combining TKI, per se, plus hypomethylating agents. There was previously a study combining in melanoma actually with BRAF inhibitors, but I'm not aware of TKI. As a fast-forward, as you mentioned, basically maybe I would like to see it as a triplet combination with IO/TKI, for example, and see if that helps basically more durability of response by reversing or boosting the immune reaction or the immune response from the checkpoint inhibitor. I do not expect overlapping toxicity per se, so I wouldn't think that we might run into a major toxicity issue similar to the COSMIC trial with a triplet combination with Ipi, Nivo and Cabo. If we think with a triple combination strategy, I would say, these three drugs have distinct toxicity profiles. I doubt there will be overlapping toxicity. That would be an interesting thing to study moving forward. I would hope somebody would be interested in taking this study or this data to the next level and combine it in a triple combination, some sponsor.

Pedro Barata: Got you. Well, yeah. No. I agree with you. Again, congratulations. It's not easy to get data published at a very prestigious journal, right? It's a lot of effort for this investigating shared study. Kudos to the Big Ten Consortium as well, because you're able to put that together. It's been a pleasure to chat with you and learn from this very, very interesting and novel concept. Thank you for sitting down with us and walking us through your data.

Yousef Zakharia: Yeah. Thank you so much again for the opportunity to present our data here. Indeed, it was a great collaborative effort through the Big Ten, through many friends and colleagues you saw listed on the first slide here. It has been a pleasure collaborating with them. More importantly, actually collaborating with the basic scientists doing all this translational work for us because that's eventually how we move the needle forward with all these types of collaborations.

Pedro Barata: Yeah. No. Well said. Amazing. All right. Well, thank you so much.

Yousef Zakharia: Thank you.

Pedro Barata: I'll see you soon.

Yousef Zakharia: Thank you. Thank you. Have a great day.

Pedro Barata: Bye.