Q&A With Naval Daver, MD: Expert Insight on CD47 as a Therapeutic Target
Targeted OncologyTM (TO): What are the treatment goals and challenges of patients with high-risk myleodysplastic syndrome?
NAVAL DAVER, MD: There are 2 primary goals for MDS treatment. The first is to reduce comorbidities, which are usually in the form of low blood counts. Most patients with MDS have cytopenia, such as decreased white blood cell count, platelets, or hemoglobin, and may need transfusions. One of our goals is to help patients become transfusion independent and improve their white blood cell count. This will help improve patients’ immune systems and reduce infections, which tend to be major complications in MDS.
The second goal is to biologically modify the disease and slow down its progression. If MDS is left untreated, especially the higher risk MDS, over time it will transform into acute myeloid leukemia (AML). AML is a very life-threatening condition that is difficult to treat. Therefore, a lot of the treatments we use in MDS are geared towards controlling it, putting into a remission, and either significantly delaying transformation into AML by many years or avoiding it completely, if possible, which will significantly improve the overall survival (OS) of the patient.
TO: What are the available systemic therapy options for patients with MDS? How do you choose which therapy is appropriate for a given patient?
DAVER: Epigenetic therapy, or hypomethylating agents (HMAs), are the backbone of treatment for MDS. Currently, there are 2 that have been approved: azacitidine and decitabine. MD Anderson did a lot of work with both of these agents and we’re actually the center that helped support the approval of these agents in MDS and AML.
The overall response rate with single-agent azacitidine or decitabine in higher-risk AML is about 35% and the median OS ranges between 15 and 20 months. Since these drugs have been approved, there have been a lot of efforts to see what agents could be added to them to further improve both the response rate as well as OS. That has led to a number of combinations of HMAs with targeted agents like venetoclax, immune-activating drugs, like magrolimab or sabatolimab, or new approaches, such as NEDD8 inhibitor pevonedistat.
In at least 2 studies of these combinations, we’re now seeing the response rates pretty much doubling from 35% with single-agent HMAs to about 75% to 85% with combinations. We’re waiting for phase 3 data from these different agents to come through, and hopefully, 1 or more of these will become a new frontline standard of care for the patients with high-risk, high-blast MDS.
TO: Historically, the development of novel therapeutics to treat MDS has been challenging. What has driven the increase in phase 3 studies of treatments with higher-risk MDS in the past year?
DAVER: I think one of the big factors has been that we have now identified drugs in AML that have been proven effective and received (or are close to receiving) FDA approval. I think that a trickledown effect is now benefiting MDS.
MDS is often a precursor to AML, so biologically, there are a number of similarities. As with other types of cancer, like AML and lung cancer, it has taken many years to understand the genomic and immune landscape of MDS. Over the past 15 to 20 years, we have learned a lot about important targets, such as CD47, BCL2, and TIM3. This has led to the development of drugs that are more effective, and more importantly, much safer. Some of the previous agents that we used in MDS in combinations were not as safe and were associated with myelosuppression or toxicities. The median age at diagnosis is 68 to 70 years and the marrow dysfunction is often present for many years, which is why tolerability is a very critical factor for patients with MDS, maybe even more so than with AML. Some therapies, especially those used in the past prior to this new generation of targeted and immune therapies caused very prolonged cytopenia or considerably increase the risk of infections. Even if a drug is effective in MDS, if it is not well tolerated, eventually, it may not show much benefit. In MDS, we really do need to strike a balance between the efficacy and the tolerability.
The newer generation of agents seem to be much more tolerable as well as much more effective and specific to MDS, compared with drugs that are used more generally for leukemia or solid tumors. These developments have really benefited the field of MDS, leading to many phase 3 trials. Hope-fully, the outcomes of these trials will be positive as well.
TO: What is the clinical significance of the CD47-SIRPα signaling pathway to the development and progression of MDS?
DAVER: CD47-SIRPα is a major immune pathway in the innate immune system and it functions by shutting down the macrophage function. CD47 is expressed on the surface of
the tumor cells and SIRPα on the surface of macrophages. By blocking this interaction, we reactivate or remove exhaustion from macrophages. Macrophages are very critical cells to fight against tumors or any foreign antigen. This is basically the mechanism of action in anti-CD47 drugs in a nutshell.
There have been studies looking at expression of CD47 on the surface of both leukemia, AML blasts, and MDS blasts; these studies have found very high levels of CD47 expression (about 98%-100%) on these tumors. We know that CD47 plays a role in the innate immune escape of MDS and AML, which is why the initial studies of therapies targeting the CD47 blockade were in MDS and AML.
The key here is to activate macrophages, ideally in combination with therapeutic agents that will simultaneously increase the pro “eat-me” signals. By blocking the CD47 interaction, we remove the “don’t eat me” signal on macrophages so they remain primed in the activated stage. To get macrophages to home in and attack the AML cells, we need a secondary signal produced by the AML cell called the “eat-me” signal. HMAs (azacitidine) very potently increase that “eat-me” signal on the surface of tumor cells, which creates a very nice synergistic dual combination of a drug that blocks the “don’t eat-me” negative signal and a drug that increases the “eat-me” signal. I think that’s we’re seeing these very high remission rates.
Overall, the combination of azacitidine with the anti-CD47 agent magrolimab is well tolerated. We don’t see a significant degree of neutropenia, thrombocytopenia, or cytopenia with this combination. We don’t see much nausea, diarrhea, hair loss, mucositis, or other tolerability issues, so overall, this combination is very exciting.
TO: What types of biomarker testing does your institution do for patients with MDS?
DAVER: For all new patients with MDS, we do a baseline morphologic assessment, cytogenetic profiling, fluorescence in situ hybridization cytogenic probes for specific MDS-associated cytogenetic aberrations, and then a next generation sequencing with what we call our 81-gene pan-el. This looks at the common genes mutated in MDS. We use that information to select frontline therapies because we do know that there are particular molecular mutations or chromosome subsets that may benefit from specific targeted therapies or combinations.
For example, if a patient has deletion 5q MDS, they may benefit from lenalidomide, either a single agent or combination. If they have IDH1 or IDH2 or FLT3 mutations, they may benefit from addition of a targeted IDH or FLT3 inhibitor to the backbone of azacitidine or decitabine. If they have a
TP53 mutation, then maybe drugs that block CD47-SIRPα in combination with azacitidine could be a very good option.
I think that’s very critical to know the baseline molecular cytogenetic profile to help select the optimal therapy for patients with MDS.
TO: Based on data from preclinical and clinical studies, which anti-CD47 agents have shown promise as targeted therapies? What are some key questions that still need to be answered?
DAVER: Overall, I’m excited by the clinical data as well as the preclinical rationale that was published in many high impact papers of the use of the anti-CD47-SIRPα-blocking drugs in MDS and AML. The one that is the most advanced in clinical development is an agent called magrolimab. There are studies investigating magrolimab in the frontline setting in patients with higher-risk MDS, included a multi-national randomized phase 3 study comparing magrolimab in combination with azacitidine versus azacitidine alone.
There are other anti-CD47s are now entering clinical development: ALX, lemzoparlimab, and TTI-622 or Trillium. These are all in very early stages of development and are just starting their trials, so we don’t really have a good sense yet of differences in efficacy or safety, or whether certain molecular groups will respond better to one vs the other in terms in terms of duration of response or survival. One of the things we do watch closely for with these drugs is early on-set anemia because CD47 is highly expressed on red blood cells. We do see an initial clearance of the red blood cells when we give a CD47 antibody; in most cases, it’s mild to moderate anemia, but we have seen a couple of cases of more significant anemia. None of those led to critical clinic events, but it is important to monitor patients often and early, especially during the first 4 to 7 days of treatment, so anemia associated with an anti-CD47 can be treated.
TO: How is magrolimab different from other CD47 antibodies that have been studied in the past? Do you see a potential role for magrolimab as a monotherapy or as part of a combination regimen?
DAVER: One of the differences is that magrolimab has less toxicities than previously studied CD47 antibodies, which had a lot more gastrointestinal toxicity. We have not seen this to occur in patients treated with the magrolimab.
The only issue as is early onset anemia because of the high CD47 expression on red blood cells, which causes them to get cleared out by the spleen and lymphatic system with the CD47 antibody after the first or second dose. With some close monitoring and transfusion support, this adverse effect has been very manageable and not a major issue.
The data that we have seen thus far in both MDS and AML shows that magrolimab works really well in combination with other drugs that increase the pro “eat-me” signal, including azacitidine and decitabine. I think it will be very interesting to see how well magrolimab works in combination with in-tensive chemotherapy, as well as with drugs like venetoclax, which kill cells rapidly and also profoundly increase the pro “eat-me” signal. Therefore, it stands to reason that the com-bination of magrolimab with those drugs could also be quite effective and should be evaluated in clinical trials.
TO: How do you see the role of anti-CD47 therapies evolving over the next 5 years?
DAVER: I think that, initially, we’re going to look at anti-CD47 therapies the front-line space, but they could potentially play a role as maintenance therapy after intensive chemotherapy or after stem cell transplantation. They could be combined with other immune-activating agents, especially in some very difficult types of AML and MDS, such as mutated TP53, that don’t respond to traditional and molecular therapies, [ie], venetoclax, cytotoxic chemo-therapy. I can definitely see a future where we have combinations of anti-CD47 agents with other immune-activating pathway drugs that target TIM3, CD200, or other immune pathways as part of full, immune-activating approach.
The next step from there would be to potentially combine anti-CD47 agents with chimeric antigen receptor (CAR) T cells or NK-cell therapy approaches. We are looking at immune checkpoint combinations with different CAR T cells to enhance the T-cell function and longevity, but it’s also known that the innate immune system exhaustion could reduce the efficacy or the durability of activity with CAR-T or CAR-NK approaches. I think whether we could use drugs like magrolimab, either in combination or as conditioning therapy with different CAR T-cell or CAR NK-cell therapies in lymphoma, leukemia, and myeloma would also be of high interest. I believe there could be a huge role for anti-CD47 agents in a multitude of different disease settings.
