Tanya Dorff: You did a great job introducing the constructs, the modalities. So now we're going to dive into how different antigens have manifested with different modalities in the clinic. So we have a variety of targets. I'll just talk about a few of them and I won't touch on CAR-T, although I'm excited to say our PSMA CAR-T paper finally just got published in clinical cancer research, which was a relatively large study and pretty interesting experience. So we can talk a little bit about PSMA in that context.

But when we think about the actual targets on the cell surface that we can use ADC or radioconjugate or T-cell engagers, one major question is how much expression is needed. We still don't know, and it may be very different among the different modalities. We heard yesterday about bystander effect with radioconjugates. For immune activation, how much is needed to get that immune system going to generate what we call epitope or antigen spread, where now the dying cancer cells release other proteins that the immune system can go after so that it can see cells that maybe don't have the cell surface target that we were originally going after.

We also don't know how to best measure expression to even say how much needs to be there for a treatment to be effective. In a lot of the studies that I'll talk about, immunohistochemistry from metastatic biopsies is still being used. And we know there are a lot of insufficiencies with this and we really need to develop blood-based markers like CTC or mRNA. PSMA has the advantage of imaging, but even there we know there can be heterogeneity in a tumor that is PSMA positive on a PET scan. So there can be varying levels of actual expression of that protein in that setting. So doctor, you did a great job setting the stage for the different ways we can design antibody-drug conjugates. And this is a nice summary of some of the different targets, but let's dive into the actual experience. So one of the early PSMA-targeted antibody-drug conjugates with an MMAE payload showed that you could have some effect, but it was really limited by the ability to deliver the cytotoxic payload.

So neuropathy became a dose-limiting toxicity because of the payload, not because of the target and febrile neutropenia. So really chemo-type side effects. And it turned out that that chemistry, that linker really mattered. If you're not actually targeting your cytotoxic payload to specific cells in the tumor microenvironment, then you're going to get these broader systemic effects just like standard chemotherapy. CD46 is an interesting target because it's expressed both in adenocarcinomas and more neuroendocrine-transformed prostate cancer. And Rahul Agarwal, who will speak next, did present a few years ago, an ADC, again with an MMAE payload targeting this very interesting cell surface protein. And again, you see a lot of the side effects that are related, not so much to the target, but to the payload in this experience. What I also thought was interesting here is that the responses were seen, whether objective or PSA across a very broad spectrum of immunohistochemistry age scores from 10 all the way to 300.

So if you can get a response with a lower level of expression, clearly that speaks to the fact that the payload is being a little bit promiscuous and not being very specifically delivered. B7-H3 is very commonly expressed in prostate cancer. Jim Allison's group published this almost 20 years ago. BioNTech had this ADC switching to a topoisomerase payload. Now we didn't know MMAE's microtubule that made sense for prostate cancer, which we know is microtubule-sensitive. What about topoisomerase? But clearly we see from this experience that prostate cancer cells can die when faced with topoisomerase payload because we had a high response rate over 40% and more than two thirds still progression-free at six months. So this is highly active, definitely could compete with something like docetaxel. And the side effect profile is a bit better for our prostate cancer patients, I would say. You don't have that peripheral neuropathy that's very vexing, especially in a taxane-pretreated population, but you do have those other kinds of chemoside effects, fatigue, some GI, and of course, bone marrow effects.

So this is a list, but not a comprehensive list of ADCs that are in development. I do want to call attention to dual-targeting because I think this is really going to be important given the heterogeneity of antigen expression in prostate cancer, given the fact that we're targeting PSMA with radioconjugate. So ABV969 is targeting both PSMA and STEAP1 with a topoisomerase payload. I think we're going to see that this is highly effective. Hopefully we'll be able to present some data later this year from the phase-one dose-escalation, but I think it may have some advantages over the single-targeting, like the STEAP1 alone or the PSMA alone, but we know that these drugs have been very active in other cancers and they'll certainly beat taxane chemotherapy and prostate cancer is my prediction.

So now looking at T-cell engagers, starting out with PSMA, that was the obvious first target. And I worked a lot with this AMG-160. You can see a very exciting waterfall plot, lots of PSAs going down. However, as we dose escalated, it became clear that there was just too much cytokine release syndrome and some of it was Grade 3, although not usually life-threatening. A lot of times that was just the LFTs going up very transiently that made it into the grade three category, but the burden of toxicity was simply too much. And then really the final flaw was the anti-drug antibodies that were neutralizing. So even though we had these responses, they were relatively short-lived. And so unfortunately, AMG-160 is not going to be moving forward. It seems like PSMA is just not the best target potentially for T-cell engagers. At the higher doses, not only did we see the cytokine release, but we started to see neurologic toxicities, ocular and hearing, which you don't see with bispecifics and CAR-Ts targeting other antigens.

Now, I don't think it's a factor of the antigen directly being on those tissues. It theoretically is not there, and yet we don't see it with other targets. So this is a real intersection of both the antigen and the construct playing a role in whether this is clinically feasible. Interestingly, the J&J PSMA CD3 bispecific showed very limited activity until they went to subQ. So potentially antidrug antibodies playing a role, I'm not sure, but subQ administration has been looked at with some of these T-cell engagers, both to reduce anti-drug antibodies, although it hasn't always been successful in doing that, and also potentially slowing that initial release of immune activation to kind of take the sting out of that initial cytokine release syndrome. So subQ was not great when we looked at it in the STEAP1 experience with AMG-509 cell, uritamig. This is a two-to-one STEAP1 with CD3 T-cell engager.

So STEAP1 six transmembrane epithelial antigen of the prostate, very highly expressed in prostate cancer. Very strong response rates, as you see on the top. Here, we did not expect a certain toxicity based on where STEAP1 is expressed, and yet we see these musculoskeletal inflammatory events. So these are not necessarily muscle cells, bone cells that have STEAP1. So we're trying to figure out what's going on. Is this a factor of just immune activation? Are there some kinds of cells in the tissues nearby that are causing this? But very wide manifestations from jaw, tongue swelling, muscle swelling, myositis, bone pains at a site of tumor, not at a site of tumor. So we've had to learn a lot. They are manageable and this drug has gone into phase-three. We can do mostly outpatient dosing because the cytokine release syndrome is much less than what we saw with AMG-160.

And yet sometimes patients do end up in the hospital with that first dose because the cytokine release syndrome does happen in that first cycle. But the phase-three trial, I'm excited to say a two-to-one randomization against cabazitaxel or AR pathway inhibitor. And I think most patients actually ended up getting cabazitaxel. So this will be a real head-to-head, is almost done accruing. The last slot is about to be filled. And so hopefully this is something that might make it all the way from phase one into our clinics in a relatively near-term, which will be really exciting for patients. Patients jumped on this study. They want immunotherapy. This is a little bit tough to give, but very effective. Paceritamig is exciting to patients for different reasons. So AMG-509, I would say because of the efficacy, paceritamig is exciting because of the lack of toxicity.

So this is J&J's HK2 targeted bispecific where we saw very low rates of cytokine release, and it can be dosed every six weeks, which is much more patient friendly than a lot of the other T-cell engagers. The efficacy might be a little lower, but this might still be advantageous depending which situations we use it in. And they've moved it into phase-three as docetaxel plus minus paceritamig, as well as a late line versus placebo. So we heard about the mask PSMA. I think that's the only way PSMA bispecifics will make it into the clinic. Other targets, Dr. Agarwal will talk about DLL3. So these are some of the issues that we've learned through our experience with the T-cell engagers and modality as well as antigen really matter. The next frontier will, of course, be how to combine and sequence these with radioligands and our other treatments. Thank you.