Trivalent CAR T-Cell Design Can Mitigate Antigen Escape Seen With CD19-Targeted Therapy

Trivalent CAR T-Cell Design Can Mitigate Antigen Escape Seen With CD19-Targeted Therapy

A novel approach to CAR T-cell therapy involving a trivalent design targeting CD19, CD20, and CD22 can effectively target primary acute lymphoblastic leukemia (ALL) cells, and could potentially mitigate CD19-negative relapse, according to a new laboratory study (Abstract 121).

CD19-targeted CAR T-cell therapy has proven effective in certain hematologic malignancies, said Kristin Fousek, MS, a predoctoral fellow at Baylor College of Medicine, who presented results of the new study during Oral Abstract Session B on CAR T-cell therapy on January 26. That success led to the recent approvals by the U.S. Food and Drug Administration for CD19 CAR T-cell therapy for the treatment of relapsed/refractory ALL (tisagenlecleucel) and for relapsed/refractory diffuse large B-cell lymphoma (axicabtagene ciloleucel).1,2

Ms. Kristen Fousek presents Abstract 121 during Oral Abstract Session B.
“As more and more patients are treated with this therapy, and as we have longer and longer follow-up, we’re beginning to see that a subset of patients are experiencing disease relapse,” and 40% do so with antigen-negative disease, Ms. Fousek said.

Primary ALL cells, however, express targetable levels of other chimeric antigen receptors, CD20 and CD22. First, Ms. Fousek and colleagues conducted an analysis of three samples of ALL; all three ubiquitously expressed CD19, whereas the other two CARs, CD20 and CD22, were expressed more variably. Still, CD22 was expressed in most B-ALL, and CD20 has high levels of expression on more mature ALL or lymphomas, Ms. Fousek said.

The investigators hypothesized that targeting CD19, CD20, and CD22 at the same time might overcome CD19 antigen escape and result in enhanced tumor efficacy.

They created a CAR T-cell product that expresses the three separate receptors on the surface of a single T cell. To do so, they used a single DNA vector that had each CAR-encoding segment included, separated by self-cleaving viral 2A peptides. This allowed for a near-equal expression of the three CARs on the so-called “TriCAR.”

Using the same three primary ALL samples that were analyzed for CAR expression, they then tested the efficacy of the new TriCAR approach. They found that the TriCAR T cells were effective against all three, as measured by rates of lysis.

They also created an experiment where it is possible to assess the ability of the CAR T cells to effectively kill targets in a single-cell assay. They could measure how long the T cell takes to find the target cell, the time to two cells remain conjugated, and the time to cell death by measuring the annexin V taken up by the media.

This experiment showed that the TriCAR T cells are, as Ms. Fousek said, “better serial killers.” They were more efficient than CD19 CAR T cells at killing target cells, as well as subsequent targets, with a shorter time to cell death.

Dr. Malcolm K. Brenner discusses Abstract 121 during Oral Abstract Session B.
The researchers then examined how effective this therapy might be in the absence of CD19 in order to model how it might work in cases of CD19-negative disease. They used CRISPR technology to knock out CD19 in a primary ALL cell line; they noted that when CD19 was knocked out, there was also a decrease in CD20 and CD22 expression. They also used two primary relapse samples to test TriCAR’s efficacy from patients who received CD19 CAR T-cell therapy and subsequently relapsed with CD19-negative disease; those cells maintained targetable levels of CD22.

The first indication that TriCAR T cells may work in this setting is that, although CD19 CAR T cells were only able to produce IFN-γ as a response in CD19-positive ALL cells, TriCAR T cells can produce this cytokine response in a challenge to all three CD19-negative ALL samples. The same was true for TNF-α.

The cytokine response, of course, is not sufficient. But a subsequent measure of lysis showed that TriCAR T cells do in fact kill primary CD19-negative variants of ALL, which Ms. Fousek called the most important result of these experiments.

An outstanding question is whether this approach will work in vivo. Ms. Fousek has studies ongoing in mice and in other primary ALL models.

Another question is whether antigen escape would still be observed. “I don’t think we yet know whether three is a sufficient number, or what the magic number may be in number of antigens to target, or if at a certain point you no longer increase efficacy by increasing the antigen number,” Ms. Fousek said. “Potentially, we may need to consider combining with additional modalities of therapy if this is the case.”

Malcolm K. Brenner, MD, PhD, of Baylor College of Medicine Texas Children’s Hospital, was the discussant for the abstract, and he said that there are theoretically a number of approaches to the problem of antigen loss. One could use the approach Ms. Fousek described, with different CARs on the same cell, or one could put different CARs on different cells (which he called a “CAR pool” approach), or one could use a single CAR but with two or more specificities (a tandem CAR T cell). Along with further fine-tuning of each of those approaches that might help limit toxicity and improve efficacy, he said, is likely further development of this entire field of therapy.

­–Dave Levitan