Despite Challenges, Targeted Cellular Therapies Continue to Be Explored in AML

John F. DiPersio, MD, PhD

Ongoing research in acute myeloid leukemia (AML) is surrounding a number of investigational therapies, including dual affinity re-targeting (DART), bi-specific T-cell engagers (BiTEs), CAR T-cell therapies, and antibody-drug conjugates (ADCs). However, challenges include finding improved targets and avoiding cytokine release syndrome (CRS), explained John F. DiPersio, MD, PhD.

“There is a false presumption that it is easy to develop a CAR, DART, BiTE, or ADC to a specific myeloid antigen and it should work and kill all the leukemia. It doesn't work that way. There are sorts of problems in AML that we don't see in acute lymphocytic leukemia (ALL),” said DiPersio, the deputy director of the Siteman Cancer Center, and a professor in the Department of Medicine, Oncology Division, at Washington University School of Medicine in St. Louis.

However, one ADC has become available in the paradigm. Gemtuzumab ozogamicin (Mylotarg) received FDA approval for the treatment of patients with newly diagnosed CD33-positive AML in September 2017.

In an interview with OncLive during the 2019 SOHO Annual Meeting, DiPersio discussed the challenges in research surrounding targeted therapies for AML, including DARTs, BiTEs, ADCs, and CARs.

OncLive: What ongoing research is being conducted for these forms of treatment?

DiPersio: There are a number of people, laboratories, and pharmaceutical companies interested in targeting myeloid malignancies, specifically AML. It's a very tough job because these drugs are shared by mature myeloid cells; therefore, you are not just targeting the immature cells, but the mature cells, as well. There are a huge number of those [mature] cells and it creates this large “sink” you have to get past to get to the malignant cells. The maturing cells are the cells that make all of these bad cytokines and chemokines that are responsible for CRS toxicity. In the context of bispecific therapy, DARTs, or CARs, it is very difficult to avoid significant CRS. There is a lot of expression of the targets and nonmalignant cells in this issue of CRS.

What data have been reported thus far with these types of agents?

There have been very little data on CARs, DARTs, BiTEs, or ADCs. However, there are several ADCs that have been developed. Gemtuzumab ozogamicin is approved by the FDA and there are a number of others that are in development targeting myeloid antigens, such as CD33 or CD123. Those are plugging along slowly. There are a number of physicians interested in CARs, BITEs, and DARTs. The response rates have been modest, but nothing like we saw with obinutuzumab (Gazyva) or CARs for ALL.

Why is that? I would argue that there are issues with the choice of the targets; they may not be ideal. Additionally, the targets express widely in the other cells—in maturing myeloid cells, especially monocytes, which are the cells that create the chemokines and cytokines that cause CRS. For patients who have relapsed AML and are going to undergo transplant, a modest response rate might be okay, but we need improvements. People are trying to improve that [through different methods] and increasing the dose of these [therapies] to get past these barriers. One way is to use checkpoint inhibitors to make them work better, but that will also exacerbate the CRS. The other is to figure out some better way to mitigate CRS so you can advance the dose of the drugs. There are a lot of thoughts about how to do that.

Are any of these investigational agents being explored in combination with currently available regimens?

Right now, most of the work is single-agent therapy to establish efficacy and toxicity. Once that gets established, then physicians will start to use different drugs. For instance, some physicians feel like they've “hit the wall,” meaning they have a low response rate and a fair amount of CRS. There have been recent efforts to add other drugs that block CRS and enhance the activity of these bispecifics to see if that can be overcome.

What are the next steps for this research?

We need to keep on searching for better targets. We also need to look for [agents] that can target two antigens rather than one. When you target 2 antigens, you make the effect much more specific and you can eliminate all those other cells that are problematic. That has been incredibly difficult to do for many reasons. The problem is you have to find something that will only work when you target two antigens at the same time; AML is a hard disease to go after with targeted therapy.

How do you determine which treatment is best for each patient?

Right now, most of these treatments are being broadly applied to anyone with AML, even in patients who don't express the target. There are early-stage studies [designed to teach us] what happens to patients who have no expression, low expression, and high expression of [a specific] antigen. Researchers have been trying to look at the correlation between response rates and expression of an antigen in AML. Eventually, it may be that only those patients who express high levels of the target antigen will be the appropriate candidates for trials or treatments in the future.

Where could these treatments fall in terms of sequencing?

Since none have been approved, they are all being used in the context of relapsed/refractory disease as single-agent therapy to evaluate toxicity and get an efficacy signal. Like everything else, once you establish efficacy and safety, then you move closer to the frontline setting.

What are the key takeaways for physicians to know about these treatments?

The targets are really problematic and are expressed on a lot of normal cells that are very abundant, and can produce all of these cytokines and chemokines for CRS. The cells have a suppressive effect on immunotherapy, especially cellular immunotherapies such as CARs, BiTEs, and DARTS. There so many problems with targeting AML that we haven't even scratched the surface of yet. Because of the effectiveness of these agents in ALL, people think we should be able to find the same benefit when we target AML with bispecifics. Because of all of those things, it has been a very difficult problem.