Margolimab Generates Hope for Patients With Myelodysplastic Syndromes, AML

July 8, 2020
Danielle Ternyila

In an interview with Targeted Oncology, Naval G. Daver, MD, discussed the safety and efficacy findings of margolimab as treatment of patients with TP53-positive AML and high-risk MDS.

Margolimab is an investigational first-in-class CD47-direct antibody that is now being evaluated in patients with acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). In particular, the agent is demonstrating promising responses in patients with TP53-mutant AML and high-risk MDS, according to findings from the phase 1b study presented during the European Hematology Association Virtual Congress.

In this study, margolimab was combined with azacytidine (Vidaza) for the treatment of these patient populations. CD47 lies on the surface of macrophages, which are often suppressed or exhausted in patients. By blocking CD47 with margolimab, the macrophages are not exhausted, and cancer cells can signal to the macrophages to attack the cancer cells.

The study drug has been evaluated across a number of tumor types, but preclinical data have suggested that the agent may prolong survival and improve responses in combination with azacitidine. This study in AML and MDS, among other clinical trials evaluating margolimab in lymphomas and solid tumors, remains ongoing.

In an interview with Targeted Oncology, Naval G. Daver, MD, associate professor in the Department of Leukemia at The University of Texas MD Anderson Cancer Center, discussed the safety and efficacy findings of margolimab as treatment of patients with TP53-positive AML and high-risk MDS.

TARGETED ONCOLOGY: Could you introduce margolimab and the rationale for evaluating this agent among patients with MDS and AML?

Daver: Margolimab is an anti-CD47 antibody. CD47 is an important immune checkpoint on the surface of macrophages in patients with AML. Macrophages are a very important part of anti-tumor responses, and leukemia is able to shut down macrophages by using the linkage between CD47 and SIRP alpha, which causes macrophage suppression and exhaustion. By blocking CD47 with an antibody, we're removing the exhaustion from macrophages, and this results in the macrophages being able to attack cancer cells. This method should technically and theoretically work across many tumor types, but most of the preclinical data that had been published by the Stanford group was focused in patients with AML, and they did show that in these patients with AML, the margolimab, especially in combination with hypomethylating agent azacitidine, was extremely effective and synergistic in multiple immunodeficient mouse models showing very prolonged survival after only 14 days of dosing. This is the background data that has led to the current studies of margolimab, the CD47 antibody in AML and MDS, as well as many studies actually ongoing in lymphomas and in solid tumors.

TARGETED ONCOLOGY: Regarding the data that was presented at EHA, could you talk about the methods, the methods that were implemented in this study, the design of the trial and what patients were included?

Daver: Regarding the data that we presented at EHA, we had 2 oral presentations on margolimab. One covered the AML population, the other covered the MDS population. This was done under 1 study. This was a phase 1b study that looked at the combination of azacitidine with margolimab. The rationale for that specific combination was that azacitidine upregulates certain signals that signal the macrophages to do more potent phagocytic activity. This was shown very elegantly preclinically by a number of groups, and so in the data that we presented, we initially showed the data with the MDS.

There was a total of 39 patients who were treated with the standard dose of azacitidine 75 mg/m2 day 1 through 7. Margolimab has an interesting dosing regimen; we initially give it twice a week for the first 4 weeks, and we do an inter-patient dose-escalation. The reason for that dose escalation or dose ramp-up is to try to mitigate on target anemia, which is probably the only significant toxicity that we see. CD47 is expressed heavily on the surface of red cells, especially older red blood cells, and when we do this kind of stepwise approach, we see that we're able to get clearance of the older red cells and the body then compensates by producing younger red blood cells. Usually around a 12 to 14, once we get to the full dose, we no longer see the anemia, and that's been a very consistent finding. Usually in the first 2 weeks, we see anemia, we manage it, and then we don't see it further.

We had 39 patients presented on the MDS cohort, and then we had 29 on the AML cohort. The good thing about the safety was that we had 0 early mortality 60-day mortality, which is a high-risk MDS and AML population is quite positive. We had only 1 patient who discontinued treatment that was on the AML cohort, and that was because of an infusion reaction. He had significant tremors, actually got better within 24 hours, but the patient and physician decided they didn't want to continue. All in all, I would say this is a very favorable safety profile similar to what we've seen with azacitidine alone. We did not see any immune toxicities. We did not see significant neutropenia, thrombocytopenia, and in fact, we see improvement of neutropenia and thrombocytopenia in about 6 to 8 weeks. So overall, it's quite safe.

TARGETED ONCOLOGY: What efficacy have we seen with this agent?

Daver: Of the 2 presentations that we had, we first spoke to the MDS one, so in MDS cohort, we had 39 evaluable patients at the time of the data cutoff. These were all intermediate high, very high IPSS risk for MDS criteria, and in fact, a large proportion of them, about 65%, had adverse cytogenetics per MDS criteria. So, in general, this is quite a high-risk group of the population. The median age was around 70 years of age.

What we saw was that the overall response rate was 91%. This is very encouraging and very high for such a high-risk MDS population, and when we look at the true CR rate, which is a CR with full count recovery, kind of a gold standard response, this was around 42% to 45%. What was interesting is that as we continue treatment in those patients who were able to get beyond 6 months, which was kind of majority of the patients, the true CR rate went up to about 56%. What we're seeing is that with continued exposure, there was better adjustment count recovery, and we were getting more and more patients to a CR, which is really the ideal goal of treatment.

Also, what was interesting is we did have cytogenetic assessment, and those patients who had baseline invaluable cytogenetic abnormalities, about 50% of them had complete recovery inside of genetics. So how does this compare? Historical phase 2/phase 3 studies with azacitidine alone have shown CR rates of about 15% to 20%, and aggregate response rates of about 30 to 35%, so the 45% CR and the 80% to 90% aggregate responses are very encouraging. Small numbers and the follow up is around 6 to 8 months, so it's still early, but we did show survival. The 6-month projected survival was 90% or higher, and even at 1-year follow-up there seem to be a very high survival probability. The median is obviously not been reached, either for survival or duration of response.

In general, this is quite encouraging and very effective so far in the data we've seen in this high-risk MDS. The second population was an AML population. This was predominantly older patients who are considered not suitable for cytotoxic induction chemotherapy, so these are patients we usually give HMA-based approaches. We initially allowed all patients who were older and unsuitable, but over time we started seeing the responses were especially encouraging in the adverse cytogenetic TP53. The protocol was actually amended about 5 or 6 months ago to focus only on TP53-mutated AML who were older and unsuitable for induction, and there were 2 reasons for that. One is we were seeing encouraging activity and safety, and the second is there are other combinations such as azacitidine and venetoclax (Venclexta), which are showing very good activity in most patients, but unfortunately, not really improving the outcome in TP53 and some other high-risk subsets.

In our study, we had 29 patients, and of those, 72% actually had what we consider poor risk cytogenetics and almost half of the patients had a TP53 mutation. The median age was 74, and we had a patient all the way up to 89 years of age, so this is definitely a population we will not routinely consider intensive induction.

The overall response rate here was 64%, which included a CR/CRi of about 56% with some few additional PRs. That's very encouraging because again, azacitidine single agent has been studied in multiple phase 2 and 3 trials AML. The CR rate is around 15% to 18%. We had a CR rate of around 45%, and the CR/ CR with incomplete hematologic response (CRi) rate is around 30% to 35%. Ours was about 55% to 60%, so definitely quite promising. However, what's especially interesting was in the TP53-positive AML, we had small numbers, so I think we need more data to be very confident with the findings, but of the 12 TP53 older AML patients, we had a CR/Ci rate of 75%, and duration of response had not been reached with a median follow up of about 8 and a half months. If we compare this to what is probably our most effective other therapy, which is azacitidine/venetoclax, the CR/CRi rate is 45% to 50%, and the duration of response is around 4 and a half to 5 months.

We think this is looking quite encouraging. We also had some nice correlative analysis looking at TP53 variant allele frequency, as well as cytogenetic remissions, and we see that majority of the patients had a TP53 variant allele frequency clearance, and the median TP53 before starting treatment was around 40% to 45%, so these people had a good amount of TP53, probably 1 of the main drivers of the disease.

The study is now going to be focusing only on TP53-positive AML. We continue to enroll and hopefully by ASH, we will have an update with many more patients. Many more have already been enrolled, and the data has not yet been analyzed. The 2 paths for registration at this time are in MDS and there is consideration based on the current study to expand it to 900 patients, and that may be considered for a single-arm registration in the higher-risk MDS with a subsequent confirmatory phase 3, which is actually just started in MDS. In AML, there are plans to look at randomized potential confirmatory studies in the TP53 AML, so we think this will be quite exciting drug to follow in the next 1 to 2 years.

TARGETED ONCOLOGY: What would you say are the clinical implications or the potential clinical implications of this combination for these two disease subtypes?

Daver: These are in the clinical trial setting, phase 2 and moving to phase 3, so at our institution, MD Anderson, we are fortunate to have access to most of the new agents, and we specifically think that especially in the TP53, AML, and MDS, probably in all high-risk MDS but especially in the TP53 AML, we think margolimab could play a very, very important role. One because that's an area of unmet need; venetoclax has been fantastic across the board, but in TP53, it has not improved outcomes like we were hoping. The tolerability is very, very good. This is predominantly an outpatient regimen, so we think that for TP53 AML, our focus has been margolimab, but there's another drug as well called APR, which has shown good activity in TP53. We really try to enroll patients on to 1 of these, and I think in the community if they know that they have an AML with TP53 or a high suspicion, those would be really good patients to send for clinical trials for either margolimab or APR-based approaches.

In MDS, I think the light is kind of closer at the end of the tunnel. We think that this could potentially be considered based on what we see emerging as a safety efficacy as an approval strategy. If that happens, of course, this could be a new frontline for MDS across the board. We still need more data and maturity, but I think those are the 2 kind of focuses.

In the end, you know, I think this is good as tablets, but the interest will be can you then do triplets. We have a study that will open very soon in the next few weeks at MD Anderson of azacitidine, venetoclax with margolimab, and we think that may be the future just like in multiple myeloma where we're able to combine 2 or 3 drugs. Of course, we have to find the ideal safety, dosing, the schedule, and the efficacy, but I think that is where we hope to be in 2 to 3 years, with ideal triplet combinations that really improve long term survival, even in this high risk, TP53, and other populations.