Scientist Stories: Saar Gill, CAR-T Cell Therapies

There are a few minor differences, but the major difference that we should be aware of is the co-stimulatory molecule that is within the CAR polypeptide. The one— the product that is approved for pediatric ALL by Novartis is co-stimulated by CD137, and that's a co-stimulatory molecule that leads to, I think, a pretty distinct and well-recognized set of behaviors by the T cells, whereas the CAR design, the CAR molecule that is part of the Kite Gilead product that is approved for lymphoma, is co-stimulated by CD28, which leads to distinct other behaviors behaviors, and those behaviors are both pharmacokinetic as well as related to toxicity. I think it's pretty well established, even clinically, that the CD28 co-stimulated products expand— the T cells expand more rapidly and probably have a shorter persistence. Comparatively, the 4-1BB co-stimulated product expands very well indeed, but less abruptly than the CD28 product and has longer persistence. And that very much relates to the toxicity. So the cytokine release syndrome that is seen after the CD28 co-stimulated lymphoma product occurs earlier, actually, than does the cytokine release syndrome that occurs after infusion of the ALL-approved product. The on-target toxicity, in my opinion, is the single biggest limitation when it comes to toxicity, and that's really an important hurdle. The other ones are, I think, less of an important hurdle, but are probably even more— people are more aware of them, and they are cytokine release syndrome and neurotoxicity. So cytokine release syndrome, I think we know reasonably well what the risk factors for developing severe cytokine release syndrome. And we have approaches that can help with some, I would say, confidence to mitigate severe cytokine release syndrome. I think that with education and experience, any center, particularly specialized centers with the right sort of with the right specialty expertise can infuse CAR T cells, should be able to infuse CAR T cells safely. So cytokine release syndrome, although it has certainly been and continues to be fatal in some patients, I think is less of a problem. Neurotoxicity is more of a problem in my opinion because we don't really, with the exception of some recent papers, we don't really understand what causes neurotoxicity. It certainly appears to be associated with cytokine release syndrome, And most importantly, we don't have a good way of reversing it once it occurs. And so to me, the most concerning side effect at the moment is neurotoxicity. I think the logistical hurdles are not to be underestimated. We are and will shortly see what those are when it comes to real-world as opposed to clinical trial settings. As I said at the start of our conversation, we and others are just starting to treat patients on commercial, with commercial product. And so I do think that some of those will be stumbling blocks. What we've already learned from clinical trials is that since these are not really off-the-shelf products by any sense, the ability to deliver this product to a patient in a timely manner at a point in time when they need it and yet are not too sick to undergo the treatment safely is is, I think, it's a considerable hurdle and requires a lot of clinical know-how and finesse and some amount of flexibility on the part of the, particularly the institution and the patient and the treating physician. So I think that's a major challenge. Another major challenge is something that we also touched on before, which is the costs of the technology. Having said that, As probably with all technologies, and I don't think I'm being too optimistic in saying that, the cost, if this is successful and there is a very high uptake of this technology, the cost will come down. We know where the costs are, we know where the fixed costs are and where the relative costs are, and I can only imagine that with increased uptake of this technology, there will be approaches that will, narrow down those costs over time. In DLBCL, the current, the most recent advancements are that the Axicel or Yescarta was approved by the FDA late last year, and that has been really incredible in terms of firstly, a rapid transition from a clinical trial to FDA approval and hence clinical practice in terms of a drug. But the other really interesting and I think important aspect of that, as for the other FDA-approved CAR T-cell product, is that we and other centers have had to learn individually, actually, how to How to commercialize it. How to, I guess I should say, how to deliver this newly available commercial product. It is not a drug. It is not a, it is nothing like anything that we or anyone else has done before in terms of applying this new product and way of treatment to patients. Still very early days in terms of the number of patients around the country who've been treated or about to be treated with commercial CAR T-cell for DLBCL. In fact, we've just infused our first patient, commercial patient, a few days ago. So that's been really interesting in terms of the practice of medicine, in terms of how to arrange for patients to come, get through the hoops of insurance, and then be treated. And of course, There are other aspects to that which relate to the hospitals or the institution's ability to bill for that. And after all, the logistics of the whole procedure have to be ironed out and will be ironed out in due course. So we have learned and are learning a lot about how you can make this newfangled therapeutic a reality for patients coming from across the country. So that's the first thing. The second thing about DLBCL and CAR T cells, to answer your question, is there is a difference, it would seem, from the way CAR T cells behave in lymphoma compared to how they behave in ALL and maybe even in CLL, all of which, of course, express CD19. They're all the same product, be it wherever it comes from, from whatever institution or whatever company. What seems to happen with lymphoma is that the complete response rates are lower, the overall response rates, and particularly the complete response rates, which of course are what counts, are lower than they are in ALL. They are around 50%, whereas in ALL they're around 90%. The, the, and so the, that's interesting. The other interesting thing that is becoming apparent, I think, is that persistence perhaps matters less. So persistence is, we think, critical for the long-term disease-free survival in ALL, relapse-free survival in ALL, whereas in lymphoma, it appears that you can generate profound responses and then the T cells fail to persist or no longer persist, and yet the patient's response persists. With the sort of relatively limited follow-up that we have to date. And so maybe the tumor biology is a little different. Maybe you need only relatively short persistence to generate a profound sterile immunity against lymphoma, whereas maybe that is not the case in the leukemias. I think time will tell. Maybe to be a little provocative, one might think That's kind of what we see with, and have seen for a long time with chemotherapy. When you treat lymphoma with a relatively short course of chemotherapy, we're able to generate complete responses in the majority of patients that first present with lymphoma, and we don't need maintenance treatment, at least not with diffuse large B-cell lymphoma. Whereas with ALL, we have known for many years that we do need maintenance treatment. So maybe that speaks a little bit to the difference in tumor biology inherent to the disease that is also surprisingly, I think, translating to what you get with CAR T cells. Myeloma, I think, I mean, the results are more preliminary than they are in diffuse large B-cell lymphoma, obviously. One thing that unites diffuse large B-cell lymphoma, ALL, CLL, and myeloma is that the malignant cell is probably, we can probably live without the normal counterpart of the malignant cell. So the malignant cell being a plasma cell in multiple myeloma, its normal functions predominantly we think are production of antibody that we need for the generation of a normal immune response. It so happens that we have a pretty effective way of replacing antibodies with intravenous immunoglobulin infusions. Lots of people get those. In fact, people with myeloma and some other Other B-cell malignancies are dependent on IVIG because they fail to produce normal amounts. The main target in myeloma that people have been working on so far is BCMA. There are others. BCMA is the most advanced. There are several different centers and companies that are working on BCMA-specific CAR T cells. The preliminary— the results that have been published and presented in abstract form as well are reasonably— I would I wouldn't say they're quite as exciting yet as for ALL, at least if you look at the average response across several different centers and several different products, I would say the responses aren't quite as high as one saw for ALL, but they seem to be better than we see for, maybe the same or a little better than what we see for diffuse large B-cell lymphoma, and so I think that myeloma has to be the next the next major player when it comes to CAR T cells. If you can survive a protracted loss of CD33-specific normal cells, then you can probably eradicate malignant myeloid cells such as AML or MDS. However, we know that you cannot survive a a protracted loss of CD33-positive cells because those are your neutrophils, your monocytes, your myeloid progenitors, and maybe even earlier progenitors too. So I think AML, exactly the same approach that has been successful in ALL, in my opinion, is probably not going to translate exactly to acute myeloid leukemia. And I think that's why While there is no dearth of preclinical publications showing that you can target AML quite successfully using anti-CT33 or CD123 or FLT3 CAR-T cells, I think that the translation to the clinic is going to be more cumbersome and more complicated. There are two recent approaches that I think have a lot of merit. One is, if you'll bear with me for a moment, so one is putting a CAR on a CAR T cell, the antigen to that CAR being a pan-T cell antigen. So in theory, what that should cause is the CAR T cells to kill one another by virtue of recognizing that antigen on themselves or their associated T cells, a process that we call fratricide. And so a way to deal with that is to actually, again, with genetic, with gene editing, to knock out that antigen on those gene-edited T cells. And that has been done, for example, with CD7, such that you put a CD7-specific CAR on the T cells, but at the same time you knock out CD7 from those T cells. And it turns out that CD7 isn't crucial for the function. So in that regard, you're eliminating the antigen that is shared between the CAR T cells and the malignant T cells. And so the CAR T cells are no longer subject to fratricide. They're no longer subject to killing themselves or each other. And so that has merit. However, what it will cause is the CAR T cells to kill not only the malignant T cells in the patient, but also the residual normal T cells in the patient. And really what that means is that we would expect quite a significant T cell lymphopenia for for as long as those CAR T cells are active, which I frankly am concerned about because we know that T cells are critical for protecting us from viral and some fungal infections as well. So that's one approach, which is interesting. The other approach is probably even more interesting and relates, however, to being able to target a particular molecule, a particular portion of the T cell receptor itself. So this is, in my view, a very elegant approach. Approximately half of our T cell receptors on our T cells have one particular variant of the T cell receptor, of part of the chain extracellularly of the T cell receptor, beta chain extracellularly, and half of the T cells have a different variant.. And there are antibodies to those different variants, and those antibodies have been put into CARs. And this group from London was able to show that you can very specifically kill malignant T cells that have a T cell receptor that belongs to one or another of those isotypes. And this is very much similar to the the light chains that we have on our B cells that we all are familiar with having kappa or lambda, and that we don't necessarily think you need both in order to— both kappa and lambda in order to have a B cell response. So the corollary of that in the T cell also exists, unbeknownst to me and I think a lot of other people before the publication of this really elegant data. Having said that, there are two problems with that approach. The first one is that we don't know for sure that we can live with just half our complement of T cells, speaking in terms of a T cell repertoire. And the other problem is that T-ALLs don't generally express a T cell receptor, so you're really only being able to— this latter approach that I've been discussing allows you to target really only the more mature T cell lymphomas. Nonetheless, that's a great area of need. Those are very difficult diseases to treat and to cure. And so I think that approach really has merit as well. It's a very attractive concept. I think it's very early in the lifecycle of universal CAR-T cells to tell whether they will or will not be as active as autologous CAR T cells or even allogeneic CAR T cells made from the donor of a patient who has previously undergone an allogeneic stem cell transplant. I think we all need to be very aware that the gene editing to remove the TCR does nothing to remove the MHC class I and II molecules from those universal CAR T cells, and therefore they're very subject to being rejected by the host. And that would be expected to limit their persistence, and that in turn is, we think, important, particularly for acute leukemias. And so I think we need to think, we need to be very aware of the cons as well as the pros of universal CAR T cells and to consider that as attractive as they might be from a logistical and economic perspective and a commercial perspective, they do have their limitations too. I think in 5 years we will be routinely treating patients with B-cell malignancies and with and with plasma cell malignancies with CAR T cells. I suspect that because of the logistics, it is probably going to remain, certainly in the next 5 years, a province of specialist institutions, probably ones that are currently allogeneic stem cell transplant institutions because of the parallel expertise that is required for that. And personally, I think that is how it should remain. Though I think that we'll see over time how the uptake of this occurs across the country and the world. I think that in the next 5 years, we will still be doing clinical trials and moving the needle pretty slowly in the setting of solid tumors because of the problems that I mentioned, including the fact the presence or, in fact, lack of tumor-specific antigens.