Okay, good morning, everyone. My name is Daina Graybosch. My team and I cover immuno-oncology here at Leerink Partners. And I, I guess I should say, and cell therapy directed at oncology and autoimmune diseases-
That's right.
Based on where we're gonna start here with Nkarta. I'm really excited to host both Paul Hastings and David Shook. We're gonna jump right into the fireside chat because I counted the pages. This was the most pages of interview guide of any company I'm hosting.
Wow.
So there's no way we're getting through all of it, but we're gonna hit on the most interesting topics. If you don't know Nkarta, I'll just say they have donor-derived allogeneic CAR NK for both oncology and now autoimmune indications. And, given, that we're new in autoimmune, we're gonna start there, and if we have time, we will get to oncology. So thank you both. And so let's start with autoimmune. I'm just gonna start with the value prop.
Sure.
Compared to an auto CAR T, where we've seen the most data for a cell therapy in autoimmune, what's the value proposition for an allo CAR NK product in your product specifically?
You go first, I'll sweep behind. How's that?
Put most simply, we have an accessibility and safety advantage. The accessibility of being off-the-shelf, obviously, I think is there, eliminates not only really burdensome infrastructure at sites to deal with apheresis and manufacture, but also spares the patient that experience and delay. So clearly, accessibility from the allogeneic piece of it is there. In terms of safety, NK cells, you know, have a long history of having a more quote-unquote, "clean safety profile" on the expansion-related toxicities of CRS and ICANS, though that doesn't seem to be as much of an issue in autoimmune disease because of the, you know, likely because of the B-cell load. But the real safety advantage that we see is in the lymphodepletion regimen.
NK cells are immediately active, don't require expansion in vivo, and our NK cells are cytokine independent, engineered with a membrane-bound form of IL-15, and so don't require the biologic effects, I won't call them benefits, of fludarabine. And fludarabine in this is a genotoxic agent that the risks of are sort of forgiven in oncology. Long-term risks of MDS and AML, risk of neurotoxicity, long-term cytopenia, etc. Those kind of risks are taken on with patients with sort of immediately life-threatening conditions like oncology. In autoimmune disease, we feel that the lymphodepletion regimen is a risk that has been overlooked. And so we think that off-the-shelf, no manufacturing infrastructure, patient burden with less chemotherapy, we think there's a real, real advantage there.
2 follow-up questions.
Okay.
The first one is, can't somebody with a T cell just follow? For instance, Obsidian, a TIL company-
Mm-hmm.
is engineering a membrane-bound IL-15 into their T-cell product.
Mm-hmm.
We're gonna see clinical data of that ACR. So is this just a time advantage, and do you expect any T-cell products to be able to follow a similar cytokine-only strategy if they do a similar engineering?
Hard to say what the future will come. You know, I think that there will be, if that's the case, you know, there will still be T cells, they will still deal with expansion-related toxicities, and so, and will require that sort of mechanism of action. Whether or not they can take our approach and, you know, overcome their inadequacies that way, remains to be seen. We do think in the current market, the availability of a cytokine-only approach is a feature of a cytokine-independent NK cell. But again, I think once those, once data come out and something else, I could...
Yeah, if I could just add on to the 'cause I always like to add the patient perspective here, Daina. I think that if you're a patient with lupus or other autoimmune diseases, you wanna have the easiest chance of not having to go into a medical center, not having to get followed up for cytokine release syndrome and other things that might happen with long-term expansion, not have to go through the logistics of in the autologous cells case, the autologous manufacturing and what have you. And so longer term, this is a disease area that's made for NK cells and vice versa. So it's always gonna be a T cell. So if we do get the same kind of efficacy with a shorter acting cell, with a more convenience, we think we have a major advantage.
Since we're the first in the space, we'd like to keep our lead and expand into other areas in autoimmune disease. We do think that it's if we have the same efficacy with a better safety profile, that's a major advantage for us long term, as well as mid to short term.
And I'll highlight the IL-15, you know, to you know, the obsidian question. We know that the IL-15 has different effects on at different doses, etc. IL-15 as a membrane-bound form provides a sort of relatively stable tonic signal, rather than the sort of high-dose surge that comes with clearing a cytokine sink. Again, I think you can you can apply a biologic principle and say IL-15 is IL-15. It's not the same at various doses and various cells. So, you know, even in K cells, the higher dose IL-15 does something different than a lower dose IL-15. And so I don't think it's as easy as sort of strapping on that edit, if you will.
Said the man who was on the pad for membrane-bound IL-15.
Yes.
Let me ask about the flu point, and this is a question I've been getting increasing, which is: How do we know that in these lupus patients, that aren't cancer patients, that getting a short-dose flu or Cy/F lu is actually gonna be genotoxic and, and lead to MDS AML? The question is, is all the data we know of the negative impacts of flu coming from cancer patients that are maybe being exposed to longer durations and have other attributes of their treatment or disease that's leading to the MDS AML risk?
Yeah, I wouldn't see risks of chemotherapy as sort of binary, that you either have them or have not. I don't think there is a safe level of, it's like lead in drinking water. Like, you at some point you make a cutoff and say that this above this is unacceptable, but below it doesn't necessarily mean it's safe. So I disagree that a small amount of fludarabine is, you know, fine. You know, fludarabine can cause really idiosyncratic toxicities. You know, I have the advantage of being a practicing oncologist. I've seen patients get fludarabine for the very first time and have, you know, life-threatening neurotoxicity from it.
There are patients that have received autologous CAR-T, that have only seen fludarabine as a part of their conditioning for autologous CAR-T and develop MDS later. Now, you're right, as on background of prior therapy, but the risk is there. And, and especially the other thing that we're also not forgiving is that genotoxic stress is cumulative over time. Many of these patients are in their twenties or younger, and exposure we know that patients, pediatric cancer doctor, we know patients get exposed to chemotherapy early in their life, have a much higher risk of secondary cancers later on in life, and that is cumulative. I would not discount that risk.
And we do know these patients now are exposed to multiple doses of Cy. So if there's a need, not there is right now, but if there's ultimately a need for the option of retreatment, you can retreat with Cy and NK cells without the fludarabine risk.
Do we know the long-term risk of multiple doses of Cy in these patients?
We do. So cyclophosphamide in autoimmune disease is typically dosed at this dose, 1 gram per meter squared, monthly, for six doses is sort of the, quote, unquote, "NIH dosing schedule" that's been used as a comparator arm in many studies. So we do know the cumulative risk, that comes with that. And, you know, but again, that's typically on multiple, multiple doses.
What is the risk?
So, the biggest one that gets brought up is fertility. So right now, the cyclophosphamide has fallen out of favor in lupus and lupus nephritis because of its fertility risk, because this is a disease of young women, and prioritizing fertility is important. And so the study that actually got MMF as standard of care was comparing head-to-head cyclophosphamide and MMF, and MMF was not any better. It was just as good at most outcomes, and it spared fertility. And so that was the sort of the main reason why our and the rheumatology group in general has moved away from cyclophosphamide, though it's still in wide use.
Development-wise, I wonder if you could talk about why, why you're starting in lupus nephritis versus a broader set of lupus patients or a basket of autoimmune conditions like we've seen some other companies do?
Yeah. So lupus nephritis, one, has pretty discrete outcome measures, so things like renal function and proteinuria, etc., that one we can demonstrate a clear benefit in. Some lupus studies, large lupus studies have been that address a larger lupus population, have been foiled by disease activity indices, particularly with large response in the placebo group. And so having things that we could track biological outcomes were important. And it is a manifestation of severe lupus, and we felt that demonstrating that we can have activity in a target tissue, it would have broad applicability and we work pretty closely with the Lupus Research Alliance and the Lupus Therapeutics Network. And the investigators told us, "Listen, if it works in lupus nephritis, it'll work in lupus.
But if it works in lupus, it may not work in nephritis. And so if you could demonstrate activity in nephritis, then you would already have... You won't even need to do another study. People will just accept that it will, it will work."... Now, the other question is interesting in terms of more autoimmune diseases, and see other companies have multiple INDs and multiple indications. And I think that's where we have a real advantage because as you get into the less morbid diseases like non-nephritis lupus, where patients are being treated outpatient, as you know, with steroids or other more mild immune suppressants. Those patients will have an even higher level of burden to get past to take on the risk of fludarabine. Right?
If you're asking a patient, you know, 'cause again, high-dose cyclophosphamide has been used in a variety of autoimmune conditions. And as you know, if this revolution is really gonna happen, that cell therapy is really gonna transform the treatment of autoimmune disease more broadly. It's gonna happen by getting into patient groups that have a lower sort of morbidity or differential morbidity index. 'Cause right now, lupus nephritis, life-threatening dialysis, super sick patients, and making decision, "Hey, do you have this life-threatening renal failure? Okay, fine, I'll take chemotherapy." But if the question is, you know, you can pick up your MMF at Walgreens, or I can hospitalize you with chemotherapy, it's like, I think I'll take the other.
We are looking at other autoimmune conditions, and we do think we have a similar, if not broader, advantage in those because of the LD regimen. There's more to come on that.
Yeah, lupus nephritis will just be the first-
Yeah.
of many, we're hoping.
Let's talk about your biologic rationale. I wonder what you would point to as the strongest preclinical support for the potential of your CAR NK in autoimmune disease.
One intrinsic to our platform is the PK, so these cells are immediately active. They're never more active than they are in day zero. And T cells are, you know, they're infused with a subtherapeutic dose, and they expand to fit the cells they have to kill. NK cells come in immediately active. There's no need for... You don't, you know, hand the drug over to the body and let it activate. You control dosing, control intensity, control toxicity. And so it does have sort of a more of a biologic rather than a cell therapy approach. In terms of other biologic, I think what we can show and have shown, NK cells are pound for pound better than T cells at killing CD19-bearing targets.
Mm-hmm.
The E to T ratios overlap, and NK cells reliably, and we published it at ASH a couple of years back, that NK cells kill cells better than T cells, CD19-bearing cells do. And even as CD19 down-regulates, now in the case of oncology, that's immune evasion, but in the case of autoimmune disease, that's B-cell ontogeny. So as cells go through, you know, develop from plasmablast to, say, long-term plasma cells, and CD19 is more of a grade, it's not a binary on and off. And so immune subsets that express maybe more dim amounts of CD19, we know NK cells are better at finding those and killing those cells. So NK cells kill with the help of a CAR, not exclusively because of a CAR.
If a CAR T, you know, you put 1 million CAR T cells into a dish with one cancer cell, that's not bearing that target, it's gonna ignore it. NK cells have the ability to recognize an abnormal target independent of the antigen. So, that's where I see the NK cell advantage, is that they can come in active and kill a broader range of targets.
Let me ask on the CD19 dim. Is that because those experiments preclinically are in tumor cells, so they have these other stress receptors that NK cells also recognize? Or are you taking advantage of your allo in your allo mismatch to get to the dim?
Yeah.
So will you see variability depending on how well you match or mismatch your patient's B cells?
No, I don't. I think the preclinical stuff has been in tumor models, you're right. In terms of, I do think there will be an allo advantage in terms of the ability to recognize target cells. But in general, NK cells spare normal, you know, normal white blood cells. And it's really how much CD19 it takes to trigger that killing.
Mm-hmm.
So if it's bearing the target, it's a matter of it will see the CD19—if it's a CD19-positive cell, it should kill it and know that that's its target, whereas a T cell may have a higher threshold in which to activate that CAR.
Got it. Let's talk about some of the peripheral B-cell kinetics, 'cause that's what everybody's looking at right now because that's what we have. And particularly, you've published and others have published and are talking about B-cell kinetics in oncology.
Mm-hmm.
What makes you believe that the B-cell kinetics in oncology are gonna translate to a clinical response and durability in autoimmune disease? Meaning, what you observe there is-
Yeah.
Actually a signal.
Yes. So, this will leverage B or NK cell kinetics in biology because we're killing those B cells in addition to the cancer cells, right? So, in addition, these B cells or the CAR NK cells will go in and kill anything that's CD19-bearing. In the case of T cells, that group means a very small expansion peak, relatively mild CRS, and, you know, really no ICANS to speak of. It's because there's not a lot of antigen around, and so you don't get a lot of drug, the cells respond. So if you have, again, just pulling numbers out of there, 100,000 B cells, the T cells will multiply to kill those 100,000 B cells, right? If you have the 100,000 B cells, we're still giving 1 billion NK cells right away.
So the E to T ratio actually favors no cancerous cells in addition. So I look at that, and I look at our B cell kinetics and say, "Yeah, we—these CAR NK cells have killed all these tumor cells and killed the B cells." If you infuse them in the presence of only B cells, the E to T ratio is going to be massively higher in the non-oncology setting. And there will not be that sort of sink to go around-
Mm-hmm.
and kill other targets. So I expect it to only be better than what we saw in oncology. One could look at it and say, "Yeah, well," because NK cells do get depleted, right? They kill, and then they go away. So if you don't have those other targets around, it just artificially increases our E to T ratio.
It's interesting. It's almost like from a small molecule, you're like, greater than 100% receptor occupancy.
Correct.
Where they, T cells, by definition, will be right at 100%.
1 to 1, right. Exactly.
Will always be 1 to 1, and you can exceed that.
Yeah.
It's a really interesting way to think about it. When I, when I look at the oncology data, and I think the latest publication with the CD19 CAR NK from MD Anderson-
Mm-hmm.
had in their supplement, every single patient in the B-cell depletion.
Yeah.
What was really notable about that is, which I know, but it really showed it, which is all these patients had low normal B cells at start.
Start.
Are you confident at the actual normal B-cell level, or do you expect the lupus nephritis patients to also have low B cells at start? You're gonna have a high enough effector at your doses going in.
We are confident in part because, as you said, we won't have the other B cells around, and so we are coming in with that sort of greater than 100% occupancy. And again, one of the advantages of coming in with cyclophosphamide ahead of time is there's going to be some degree of B-cell depletion or at least reduced proliferation there. So there's an opportunity to kill those cells-
Mm.
In addition. So the bar won't be quite as high as the normal complement of B cells, but even if there were, we're confident that the math holds up.
Let's talk about the peripheral or the pharmacodynamic markers that you'll measure in your autoimmune trials. So there's arguments from the academic lab that this is about depth and successful immunologic reset-
Mm-hmm.
as measured by depleting your B-cell compartment and then having the naive B cells reconstitute faster.
Mm-hmm.
Do you believe that that's gonna end up being the most important surrogate marker for durable benefit, or are you thinking about other markers?
I think that's an important correlational one. I think what-- I think with what has happened is people have looked at the data and kind of reverse engineered what that-
Yeah
Milieu looks like and said, "This is what you need to do in order for it to be successful." So it stands to reason that if you deplete at the level of CD19, that you get this naive recovery. So I would expect that to happen. Will that be, you know, can you get a response without that reset? Probably. If you could get a, you know... What everybody proposes is happening is having a clonal group of B cells that are secreting autoantibody.
Yeah.
You've got to somehow target it and suppress those cells. I would argue you don't actually need to eliminate them, but you need to suppress them low enough that the native immune system can control it, because these autoantibodies happen kind of in waves anyway. Lupus and other disease, autoimmune diseases come in flares and remissions, and the immune system itself helps, you know, Tregs and others help to control the amount of inflammation. So I, I see this as adequate suppression of this-
Mm
... this, pathogenic clone or, or group of B cells. I don't think we think this is truly clonal; it's probably polyclonal. And I think adequate suppression of the B cells so far seems to be sufficient. You know, how long is, how long is long enough? How deep is deep enough? I think all we've seen is what's been seen, and that's been enough. So we know the B cells are recovering as soon as 45, 50 days in the Schett data. Is that long enough for long-term durable? Appears to be. Is it too long? I mean, what's to say that it just doesn't need to be suppressed for one day, for-
Yeah
... two weeks, etc.? So I do think duration and depth are TBD, but I think we'll expect to see that immune reconstitution with a more naive B-cell subset, in that sort of immune reset or clonal depletion.
The other marker, I think, that I pick up on reading the rituximab and obinutuzumab data, and I think everybody asks about this too, is where in the body these clonal or polyclonal B cells that are diseased are hanging out with some evidence that they're hanging out in the kidney-
Mm-hmm
... in aggregates. And as you think about NK cells versus T cells, are there differences in tracking that you expect for CAR NK and CAR T? And are you gonna be able to measure where these cells are going, to provide sort of pharmacodynamic evidence for hitting those clonal cells out of the periphery? Does it matter to do that?
Yeah, I mean, I think patients will respond, and I think that will be evidence that they trafficked. It'll be challenging to either biopsy or track cells that are transient by design. We've used response data in our lymphoma program to show that cells are getting to the kidney, to the marrow, to the blood, and secondary lymphoid organs. We've depleted those malignant clones, and malignant B cells are a pretty good marker or proxy for normal B cell trafficking. And I think as you highlighted before, like, tumor masses are pretty good proxies for, you know, clusters of
Mm.
-of B cells. So we don't expect trafficking to be a limitation. As I mentioned, NK cells will traffic to anywhere in the body. They, that's sort of their job, is to go everywhere that other cells go and make sure that malignant clones aren't there. They probably traffic in different ways than T cells do, but I certainly don't expect trafficking to be a limitation. It has not yet been in oncology, and malignant cells can go anywhere.
Okay, let's talk, 4 minutes, 2 questions, one biology, one data. The biology is you're going after CD19?
Correct.
What do you think about CD20, BCMA, bispecifics, and how are you thinking about that, you know, looking forward beyond this initial trial?
Okay, I'll try and keep it brief. 19 appears to be the right target. No reason to believe that you need anything beyond that. 19-positive plasmablasts are known to be the autoantibody-secreting population in SLE in particular, and so far, targeting 19 has led to long-term durable responses. The addition of, say, CD20. One, CD20 doesn't appear to be the right target. Obinutuzumab-
Mm.
-Rituximab have some benefit, don't appear to have real durable, effective benefit. I haven't seen all the obinutuzumab data, but-
Yeah.
CD20 doesn't appear to be the right target. Plasmablasts are generally CD20 negative, CD19 positive. Their argument is it's not really negative, it's dim, and if you have a high enough avidity, you can get them.
Yeah.
I think that story for CD20 depletion in these autoimmune diseases has been written. It's not effective or not ideal. Adding on something further along the line, hitting the long-term plasmablast, the BCMA-positive cells, I think is likely to add immune suppression.
Mm.
It's likely to take away vaccine immunity and infection immunity that patients in the SLE data have spared. So I think it's gonna be unnecessary immune suppression, to be honest. It doesn't appear, at least in lupus nephritis and now in scleroderma and others, that it doesn't appear that needs to be necessary. I think it's a good argument if you have a bispecific that targets BCMA, but I just-- it just doesn't appear like it's gonna be needed. Yeah, is that-
There are-
Give me another two minutes.
There are some autoantibodies that aren't reduced in the Schett.
Mm-hmm.
People have argued those are the ones coming from the plasma cells. And so I guess you don't-
I would argue that they still have disease control.
It didn't matter.
Yes-
So the sample too small, meaning are there some patients where a long-lived plasma cell may be contributing to disease, and we're seeing the heterogeneity of a small sample in the Schett?
That may, that may be true, and I think if, you know, we're, we're following the data like everyone else. I will say that the long-term, those plasma cells, BCMA-positive cells tend to be CD19 dim and not negative-
Mm.
Similar to the plasmablast being CD20 dim and not negative. And so, again, why I highlighted this increased sensitivity of NK cells, I do think that we'll be able to cut a wider swath around the plasmablast, both the CD19, so the CD19 dim, BCMA positive, CD20 negative cells, I do think we'll be able to increase the hit with NKX019 without adding in another target.
Interesting. Okay, so the last question is, what data shall we expect and when? I think it seems like enrolling these trials has been quite difficult across the industry.
For autologous T cells, yes.
Yeah, for autologous. So what should we expect from you guys in terms of when patients and how fast you could start to accrue patients?
He's giving me the easy one. Yes, so we've guided to, we're gonna dose our first patient in the first half of this year. We haven't guided to when we'll present data. We believe that enrollment for us is going to be relatively straightforward, not as complicated as what others are seeing. And so when we've got a patient or two under our belt, we'll be able to give guidance on when to expect to see more data and what additional indications we'll-
Do you have the 28-day span between your first patient's dose?
We do
For the first three?
Yeah. So FDA released industry guidance for the development of CAR T cell products earlier this year in January, and essentially have codified the stagger, so I don't think anybody's getting around it.
Once we get that, and then we should come to you and be like: "How fast is your enrollment going?
We're looking forward to having this discussion where we've dosed a couple of patients, where we can have a patient discussion. It would be fun.
Perfect. Okay, well, that's it. So thank you very much.
Thank you.
We appreciate it.
Appreciate the opportunity.
In the 10 minutes, while you're running off, some people can grab you and ask a question if they have a follow-up.
We'd love to.
Thank you, everyone.
Thank you.