Transcript: Therapeutic Advances and Considerations for Treatment Selection in High-Risk AML

Dr Guenther Koehne: Good morning, welcome to Oncology Learning Network. My name is Dr Guenther Koehne. I'm the Deputy Director and Chief of Blood and Marrow Transplant and Hematologic Oncology at Miami Cancer Institute, part of Baptist Health South Florida.

I will be moderating today's discussion on immunotherapy for hematologic malignancies. We will focus on acute myeloid leukemia (AML).

Today we have great panelists who have extensive expertise in treating hematologic malignancies. I'm joined by Dr Richard Stone, Dr Robert Soiffer, Dr Siddhartha Mukherjee. I would like to ask all of you to introduce yourselves, starting with Dr Stone.

Dr Richard Stone: Hi, I'm Dr Richard Stone, Chief of Staff at Dana-Farber Cancer Institute and Director of the Leukemia Division of Medical Oncology Translational Research.

Dr Robert Soiffer: I'm Robert Soiffer also from Dana-Farber Cancer Institute, where I'm the Chief of Hematologic Malignancies and the Vice Chair of Medical Oncology.

Dr Siddhartha Mukherjee: I'm Siddhartha Mukherjee, I have the privilege of being trained by both Dr Stone and Dr Soiffer. I'm a professor at Columbia University where I am at the Division of Hematology and Oncology, but also part of the Myelodysplastic Center, which has been newly created. Thank you.

Dr Koehne: I wanted to have all 3 around me with a focus on AML. Let's get an update on treatment approaches for AML, with a focus on high-risk AML, or differentiating AML. As we know, not every leukemia in AML is the same. We have different approaches, Dr Stone.

Dr Stone: Absolutely. Well AML, like most cancers today are understood to be very heterogeneous and they're heterogeneous in terms of their host. They affect younger people, older people. In AML, it turns out that most leukemias that affect older people are rather resistant to chemotherapy. Even in younger people, we know through genetic and cytogenetic analysis, that likelihood of response can range from very high to very low. Likelihood of survival can range from high to low, depending on the individual type of leukemias. We certainly know how to identify poor prognosis leukemias, but we don't really know how to treat them effectively yet.

Dr Koehne: You are alluding to age of patients with AML, but we also have certain factors that, from the beginning, I believe, would allow us to risk stratify the treatment that you will be choosing for the patients.

Dr Stone: Absolutely. For the average adult with AML, we begin to think about several factors. The host factors, whether they've had a family history of blood problems or leukemias or mild dysplasia, whether the individual with the AML has a history of prior therapy for their cancers, or indeed for mild dysplastic syndrome. We're really focused mainly on the disease features, which are a combination of the cytogenetic and genetic findings subsumed into one of several prognostic algorithms.

The one in most common use would be the European LeukemiaNet Prognostic System, which divides patients with AML into favorable risk, which is about 15%; adverse risk which is about 15%; and the vast middle ground, which of course leaves about 70% of people.

To make a long story short, certain cytogenetic factors including abnormalities of the core binding factor apparatus called inversion (16) or (8;21) are in the favorable risk group. Patients with p53 mutations, complex karyotypes, everything, unfavorable risk group, and not to belabor the point that there's lots of fine tuning within that and beyond those features.

Dr Koehne: Thank you. That would do. Would you treat a patient that has been diagnosed with FLT3-positive AML differently than from the beginning that has been diagnosed core-binding factor AML?

Dr Stone: Absolutely. We would, of course, separate prognosis and prediction. My remarks a few minutes ago really designed to talk about prognosis but features within the prognostic algorithm can be used to predict which patients should receive which therapies.

You said FLT3-mutant patients. We know that about 30% of AML patients have a FLT3 mutation in their blast. At the moment, for those who can tolerate chemotherapy, the recommended therapy for that subset is standard chemotherapy plus midostaurin. In the future, that very likely will be other FLT3 inhibitors that are more specific and potent beyond midostaurin that will be used.

I think for the foreseeable future, for those who have a FLT3 mutation of either of the 2 subtypes—either the tyrosine kinase domain mutation which is less common, with a more common and more adverse internal tandem duplication or length mutation—those patients will receive chemotherapy plus a FLT3 inhibitor.

For older adults, and perhaps for many of the patients with adverse prognosis, we are now using non-intensive approach with a hypomethylating agent (HMA), usually azacitidine, plus a BCL2 inhibitor, namely venetoclax.

Indeed, the use of HMA plus venetoclax has revolutionized our approach to older adults with AML and other adults with AML who aren't good candidates for chemotherapy.

Dr Koehne: Thank you very much for that. One of the reasons I asked the question is we try to get patients that would benefit from an allogeneic stem cell transplantation (ASCT) into best remission and therefore try to increase the amount of patients or number of patients to undergo transplants.

Which then leads me to Dr Soiffer to ask a similar question. Would you transplant every patient that got diagnosed with AML or would you select certain patient populations, and also the timing of transplantation I wanted to address?

Dr Soiffer: We do not recommend transplantation for all patients with AML. There are a subset of patients, those that Dr Stone referred to a second ago, who actually do fairly well with chemotherapy, probably not as well as we'd really like, but there's no evidence that allogeneic transplant will offer them an additional benefit.

Those individuals with core binding factor mutations (8;21) translocations or inversion (16) we generally concentrate on using chemotherapy rather than allogeneic transplants.

There's also a subset of patients who have a mutated NPM1 which confers a relatively favorable prognosis, which in the absence of a concomitant FLT3 mutation, we again, would not offer transplant upfront. That is after they achieve a remission.

However, for the remainder of patients, those other individuals with so-called intermediate risk cytogenetics, or intermediate risk AML, and those with adverse risk AML, we do try to offer them a transplant in first remission, if they're able to achieve it.

These days, we're able to offer transplants to many, many individuals. We're really not limited by age. We do transplant patients as high as late 70s, if they're physiologically fit.

Dr Koehne: That's great. Follow up question. If a patient with core binding factor AML relapses, which is rare, but if it happened, would you consider at that point to bring the patient into an allogeneic transplant?

Dr Soiffer: Yes, absolutely. Individuals who have relapsed after initial induction chemotherapy—actually, no matter what their molecular or genomic characterization is--would be candidates for transplant. Actually, those individuals who core binding factor mutations, who go on to get transplant in second remission actually do relatively well compared to their counterpart.

Dr Stone: If I could just say one more thing about who gets transplants, I would even go one step farther.

What we really want to do for patients who have AML, let's get them to transplant, as you alluded to, in a deep, as remission as possible so called measurable or minimal residual disease (MRD) negative state. If a patient with core binding factor, AML who may have a "favorable prognosis", goes into clinical remission where we can't detect blast, but their blood counts recovered yet, they still have evidence of disease by either molecular or flow cytometry technologies.

We will recommend a transplant for those patients because their future with chemotherapy will not be rosy.

Dr Koehne: You're saying that if a patient with core binding factor positive AML has MRD, you would not hesitate to bring patient into transplant?

Dr Stone: That's correct. I wish I had a therapy—that maybe we'll hear about from Dr Murkherjee either in the future—of biospecific antibody or like they have in ALL (acute lymphoblastic leukemia) or a CAR T-cell that will reduce the MRD to less disease. They might be a better candidate for a transplant. At the moment, we don't have those tools, so we'll go right to transplant in such patients, if possible.

Dr Koehne: Dr Soiffer, one more. Molecular profile or cytogenetics can be sometimes confusing. You mentioned already NPM1 and FLT3 co-expression. If good risk expression is combined with bad risk expression, such as FLT3, would this convey or change your strategy?

Dr Soiffer:

That's a very important question. Generally we consider co-expression of FLT3 internal tandem duplication and NPM1 mutation that's not considered favorable and we would move ahead in transplant.

It gets a little more sophisticated though when you're looking at the amount of mutation there is with FLT3 and the percentage of mutation. There's some controversy about how bad is bad, but that is certainly an issue.

One of the things that we have seen with a transplant these days is we certainly know that while transplant can be curative for many patients, there are still patients who relapse after transplant. We have to understand why they've relapsed and there are a number of efforts being undertaking that Dr Mukherjee may talk about in a little bit, about post-transplant therapy, either with chemotherapy or immunotherapy or targeted therapy, to combine the effects of allogeneic transplant with these newer agents to try to improve overall outcome.

Dr Koehne: Thank you, Dr Soiffer. That's the ideal transition to talk about the work that Dr Mukherjee is trying to explain.

This Summit has a lot of focus on chimeric antigen specific T-cells as well. We heard a lot about the CD19 targeting, the BCMA targeting for multiple myeloma. Now, CAR T-cells for AML have been limited so far.

Dr Mukherjee would you explain why that's the case and how can we get out of this situation?

Dr Mukherjee: Let me begin by taking a step back, which is to say that when AML is at a genetic level, understood almost better than virtually any other cancer, gene sequencing has revolutionized our ability to make prognostic directions in AML as Dr Stone and Dr Soiffer pointed out.

The problem is that we still really fundamentally don't understand 2 things.

Number 1 is what these gene mutations or how these gene mutations are causing cancer. We might be surprised by this, but when we started doing sequencing of AML samples, perhaps some of us expected we'd find the typical oncogenes and tumor suppressor genes mutated. Instead, we found genes whose function is very mysterious. Genes that have to do with DNA methylation, the genes that have to do with splicing.

I would pause it, that despite probably 1000 papers published on this, as a field, we still don't understand why these gene mutations cause AML.

Number 2, we don't really have targeted therapies aside from the couple mentioned, FLT3-directed therapies, we don't have targeted therapies that target these particular mutations. That's one point.

The second point is that whenever we've looked quite deeply at hematopoietic stem cells and normal hematopoietic stem cells and normal myeloid progenitor cells and their myeloid progeny, we have tried very, very hard to find unique antigen determinants that are unique to the AML cells, but not present in normal myeloid or common myeloid progenitor or hematopoietic stem cells.

That effort, despite again, 20 years of trying to do this, has largely revealed nothing. Of course, there are mutations and those mutations create new antigens, but typically those are not amenable to targeted therapy.

What we come up with then is an approach, which has to take into consideration the fact that there are no unique antigens, as far as we can tell in the AML, which is not present in hematopoietic stem cells or myeloid progeny or myelodegenerative cells.

And therefore, as a consequence of that, every time you give any kind of therapy, you certainly can kill the AML, but concomitantly you have myelosuppression and myelotoxicity, which can be either moderate to extremely severe.

That brings up a central problem, which is that CAR-Ts are very good at killing, but they also persist. If they keep killing, then you might be lucky and kill all your AML, but you will also be unlucky because you'll kill the rest of your bone marrow and you can't survive without your bone marrow.

What people don't realize and often is confused is that you can do that with CD19 because you can live without CD19-positive normal B cells. In fact, a very good prognostic sign of success of CD19 therapy is normal B cell ablation.

A very bad prognostic sign in AML treatment would be normal hematopoietic ablation because you wouldn't get it back. If you really think about it, 20 years later, what we're really looking at is that fundamentally speaking, there has been one bright signal in this in entire universe.

That is that we know from transplantation that immunological therapy, however transplantation is a very crude form of immunological therapy, but it works. It has in patients with a certain risk, it can potentially cure patients.

There is a bright signal there, and that's the north star that we decided to follow. How do we do this?

Well, we took 2 or 3 approaches. One is we decided to take a gene agnostic approach. Let's not worry about because we don't have a lot of target therapies, FLT3 aside, we decided that let's ignore the genes and behave as if, let's kill the AML cells with a CAR-T or an antibody toxin therapy.

What's the problem with that? Well, the problem with that is that you'll eliminate or kill your normal hematopoietic stem cells, your normal myeloid progenitors and so forth.

Then we said, well, wait a second. What if we made CAR-T against AML, but we now remove using CRISPR that antigen from the donor normal hematopoietic stem cells? It's an upside down approach. I sometimes call it catching your right ear with your left hand because we are not targeting the AML cell, we're changing the normal cell.

We've done this very successfully with 1 antigen CD33, where we've done this now with 3 antigens, we've done this now in combination with 2 antigen combinations, we've done this with not just CRISPR, but with base editing, which is a much finer tool as it were. When you do that, at least in animal models, 2 things happen.

One is that you can kill AML because your normal donor cells lack the antigen, you can spare your normal donor cells and therefore you get repopulation of normal. What we call the therapeutic window, common term in oncology, the therapeutic window is infinite because the normal cells don't have the antigen. You can give them an antibody against it, you can give them a CAR-T against it, etc.

That's the approach that we've taken. It is gene agnostic. Whether you have JAK2, whether you have FS3B1, whether you have all these other mutations that Dr Soiffer and Dr Stone mentioned is antigen agnostic.

We are now in clinical trials and trying to test this approach. A variation of this approach is using bispecifics. We have designed also bispecifics against, in this case, CB33, or the other antigens that are relevant CLL-1.

These bispecifics bind to these on one hand, and then summon either T-cells or NK cells into the bone marrow, and again, kill the AML cells as long as you can remove the antigen from the normal cells.

Last quick point, we also had from the National Cancer Institute, a very good CAR-T against CD33. Now a very good CAR-T against CLL-1 another commonly found antigen.

The question there is whether you can use that without this CRISPR approach. In other words, can you ablate the bone marrow of all cells, including normal hematopoietic stem cells, then remove the CAR-T using various methods, a pill switch or chemotherapy or whatever, and then repopulate the bone marrow post the cleansing of the bone marrow with a donor transplant.

That is yet another approach that we can use potentially in AML.

Dr Koehne: Thank you so much. It's fascinating. In summary, at least with respect to CD33 down regulation by CRISPR technology, we can now down regulate CD33 on the normal hematopoietic stem cell. We can do a stem cell transplantation with CD33-negative stem cells, and then can come in post transplantation with specific CD33 targeting agents, such as CAR T-cells or bispecific antibodies, were the only cell that is left to be positive for CD33 is leukemia cell.

Dr Mukherjee: That's correct. The nature has given us a little bit of escape. A question often people ask is what would happen if you got a CD33 negative transplant, would that be even viable or not?

Well nature's given us a little bit of an escape because from the human databases, we know that there are several thousand humans who lack CD33. This is from the NOMAD database. There are several humans that lack some of these other antigens as well, probably because they play a minor role or some role in the elimination of pathogens.

We don't know exactly what role they play, but at least you can live without. In animal models right up to monkeys, we know that you can transplant these cells and they seem to survive, they seem to have normal functions. That's been an important side note to this conversation.

Dr Koehne: In other words CD33 expression on the hematopoietic stem cells is not needed.

Dr Mukherjee: It's not needed as far as we can tell.

Dr Koehne: I think this will be very exciting. Dr Stone, do you have anything to add?

Dr Stone: I think the technology and the theory that Dr Mukherjee has illustrated is indeed fascinating. Of course, having done AML for a long time, I remain hopeful but skeptical, or skeptical but hopeful, because I wonder if CD33 is the target that eliminating it in a AML cell, all the cells in the body that have CD33 will still leave some AML cells that may lack CD33 that will come back to haunt the patient in the future.

Dr Mukherjee: That's part of the reason we're concentrating on the dual targeted therapy. I'm always reminded of (this) very famous line: which is there are 3 zeros that separate 1 million from 1 billion. That is to say that in any cancer treatment, treating a person with 1 billion cells is very different from treating a patient with 1 million cells, because they have 10 to the power of 3 times the chances of acquiring resistant mutations. The trial is really designed in the context of minimal residual disease, where you have the least number of cells. That's one.

The second thing you're reminded about is that if you have one escape mechanism, you need to target the other. We are very, very actively doing dual targeting and there are various methods to do that. I'll talk about that later, but there are at least 2 antigens that we would target so that those nasty CD33 haunting cells have a second hit.

Virtually all cancer therapy really relies on 2 possible hits. We also note the fact that in, I would say virtually every AML trial, a single drug always causes resistance through various mechanisms. We are very intrigued and interested in doing at least 2 hits.

Dr Koehne: Well for CAR T-cell therapy or CAR T-cell targeting, we need a marker that can be targeted. Otherwise, it's not possible. There are certainly other indigence expressed on leukemia cells that are not accessible to CAR T-cells, but they are presented in a certain HLA-restricted fashion where the peptide derived from the protein—I'm thinking about Wilms tumor antigen protein—presented my own work on multiple myeloma cells and plasma cell leukemia yesterday about this.

But I also know that during my time at Memorial Sloan Kettering Cancer Center, we targeted WT1 [Wilms' tumor-1] on AML cells. But the limitation is we have one HMA-restricted T-cell that can be administered, but it was a factor. We had that and we added a WT1 vaccine into the whole paradigm of treatment.

There are different approaches and that in fact goes back then for donor lymphocytes, that can be modified specifically to recognize a peptide in the context of HMA. I don't think in the time of CAR T-cell therapies, we should wipe off all the other approaches that may be beneficial. CAR T-cell production is very labor intensive. So is the generation of antigen T-cells from the donor, but the mechanism and principles are very similar.

Dr Mukherjee: One thing we didn't talk about is of course, antibody toxins. Antibody toxins have had lots of problems, but those were, I would say, generation 1 antibody toxins. What we're seeing now from the general biotech field is a new interest in antibody toxins with newer toxins, better linkers, more stable, longer high half-life, the toxin doesn't come off, etc.

What's nice about them of course, is that they don't depend on the auto autonomous T-cells—that's one feature. The other feature, which I will briefly only mention this talk because a nice paper about this is coming out is that again, nature has given us a little bit a hopeful window, which is that you will realize that especially in MDS that transforms to AML, which of course is a very high-risk population. There are very common mutations in splicing factors SF3B1, we've published many papers on this.

We've described the mechanism of this. We know now the binding partner, SSRSF2 and many other genes like this.

Now, what is very interesting and I'm giving you a little sneak preview of what is to come. What is very interesting is that because they're splicing factors, these generate neopeptides. These neopeptides are unique to the Leukemic cell and not present in normal hematopoietic stem cells.

We have identified these neopeptides. We know which ones are loaded on common HLA alleles. We are now raising donor T-cells against these as well. That again is a little window of opportunity that will lead to another trial. I hope in which we can target because now finally there will be a unique tumor antigen.

Dr Koehne: This is all fascinating. If I can just summarize what we discussed. Number 1, AML has to be looked at in the context of gene modifications for future analyses and should be risk stratified and appropriated treated from the beginning.

We have with few exceptions, that's a co-binding factor positive AMLs, we would bring patients to allogeneic stem cell transplantation, which is the curative therapy at this point.

We are now at this juncture also working on post transplantation immunotherapies in various forms, whether it's CAR T-cells that's specifically targeting the AML cells or bispecific antibodies or donor derived antigen specific T-cells or combination thereof. In fact, I think we have lots of work to do. We have a great future ahead of us.

The final statement I'd like to say is that I'm the proudest man on the planet right now to be surrounded by the 3 of you and have this discussion on a podcast with Oncology Learning Network.

Thank you for all the contributions that you did today and for the success of my meeting here.