Investigators Focus on Reducing On-Target, Off-Tumor Toxicities With CAR T-Cell Therapy in Solid Tumors

As investigations of CAR T-cell therapies are starting to break ground in a variety of solid tumors, finding ways to ensure on-target, off-tumor toxicities are avoided is a focal point of research, according to Renier Brentjens, MD, PhD.

“[CAR T-cell therapy is] something that’s near and dear to my heart. Cancer research is exciting and for those who doubt that we’ll get this technology to apply to solid tumors, they’re wrong. We will. It may take a bit longer than our patients want it to, but I have a great deal of confidence that in the next 5 years, these technologies will be used for patients with even more difficult to treat metastatic solid tumor malignancies,” Brentjens said.

Early research has already shown promise in the solid tumor landscape as data from arm A of a phase 1 trial (NCT04196413) showed that patients with H3K27M-mutated diffuse midline gliomas who received CAR-modified T cells targeting GD2 (n = 11) did not experience dose-limiting toxicities at the first dose level.1 However, 3 patients experienced dose-limiting cytokine release syndrome (CRS) at dose level 2. Regarding efficacy, 4 patients experienced major volumetric tumor reductions and 3 had smaller reductions; this included 1 patient who had a complete response which was ongoing for over 30 months following enrollment. Additionally, all patients had tumor inflammation-associated neurotoxicity, which was managed safely with intensive monitoring and care.

Furthermore, Memorial Sloan Kettering Cancer Center reported that a patient with mesothelioma experienced a response following chemotherapy and treatment with CAR T-cell therapy.2 The patient received pembrolizumab (Keytruda) as the last phase of therapy and is doing well 5 years after treatment.

In an interview with OncLive®, Brentjens detailed the solid tumor types his laboratory is investigating CAR T-cell therapy in and ways to mitigate on-target, off-tumor toxicities. Brentjens is chair of the Department of Medicine, the Katherine Anne Gioia Endowed Chair in Cancer Medicine, deputy director, and a professor of oncology at Roswell Park Comprehensive Cancer Center in Buffalo, New York. He is also a professor of medicine at the Jacobs School of Medicine and Biomedical Sciences at the State University of New York at Buffalo.

OncLive: What measures could minimize the risk of on-target, off-tumor toxicity during CAR T-cell therapy development for solid tumors?

Brentjens: I believe we were the first to recognize CRS and there are algorithms already in place for how to limit that toxicity. Now, the on-target, off-tumor toxicity is always a bit more worrisome because as much as we think that the targets we go after are safe targets, there aren’t a lot of CD19-type targets out there. You always run the risk, with maybe a handful of exceptions, that the protein that you’re going after isn’t only expressed on the tumor but is expressed on normal tissues elsewhere. There are preclinical things that you can do to try to mitigate that. You can take your binder and see if it stains on normal tissue microarrays. The first and foremost thing is you want to make sure that [the protein] isn’t expressed on normal tissues at an appreciable level, and you can reassure yourself to a degree by screening the antibody that you used to make your CAR T-cell against normal tissues and hope that there’s no binding. Sometimes the targets that you’re going after are cytoplasmic and not on the surface of normal tissue, so that may be one way of getting around that.

But we have seen more in depth that sometimes there’s low level of expression on neural tissues and you always need to be very cautious about that. The ways around that type of toxicity are 2-fold. The first is, fundamentally, you don’t want these CAR T cells to persist for a very long time because sometimes you’re going after a target that is well overexpressed on the tumor but is only expressed on very low levels in the normal tissue. [Therefore], initially, your CAR T cells are going to find the tumor [and] eradicate the tumor, but then if they stick around too long, they’re likely going to find this low-level expression on normal tissues. This may not be a universally shared [opinion] in the field, but the CAR T cells should be around for maybe 28 days, and then you would hope that they go away.

Sometimes we find reassurance if the target we’re going after [has] already been [examined in] a clinical trial with a bispecific [antibody]—if it’s well tolerated there, then you’re more confident that the CAR T cell will be well tolerated. Clinical trials need to be designed [where] you start at very low doses of T cells, so that if there is unforeseen on-target, off-tumor toxicity, you can treat [the patients] with steroids and chemotherapies to abrogate that response.

There’s another group that looked at putting suicide genes into the T cells, which allows you to turn on or turn off the CAR T cell. There are [also] gated strategies where 2 targets have to be present for the T cell to get activated; those 2 targets are only expressed on the tumors and either target is only expressed on normal tissue, so that the T cell won’t engage those normal tissues. That looks good on paper, but I can’t say that I’ve seen that work in real life. People are looking at natural killer cells, which tend to be shorter lived, so they don’t persist long term.

Is there any solid tumor type that CAR T-cell therapy could be effective in?

If you look hard, you can find suitable targets for any cancer type. Nowadays, with protein maps and RNA encyclopedias, you can find potential targets on virtually any cancer cell. Given the fact that now, which didn’t exist when we started out, they have phage display libraries, finding binders to any target that you want is much easier and can be done more rapidly. If you would have asked me 20 years ago if there is a favorable solid tumor, [I would’ve said] if you had access to an antibody that binds carcinoembryonic antigen or binds EGFR, then those would be the favorable ones to go after. Because the tumor microenvironment is different from tumor type to tumor type, we don’t quite yet know which armoring approach or which combination approach would be more favorable for one type of tumor vs another.

Right now, we’re going after cancers actively in the laboratory such as small cell lung cancer, sarcomas, and pancreatic cancer. There’s a list, but that list is not dictated by which tumor we think is most likely to respond; the list is in part dictated by the ones that are hardest to treat. We also have an active program in ovarian cancer. We did not select those cancers because we thought that they would be more likely to be treatable [with] CAR T cells, [we selected by] what we had access to, which targets popped out at us, and then we adapted to the cancer type itself. There may be other people in the field who may disagree with that. I suppose [as there are] hot tumors and cold tumors, [and hot tumors] have more immune cell infiltration, you could make an argument that hot tumors are more likely to respond to CAR T-cell therapy. But I’m not sure that will ultimately be born out with whatever the first solid tumor is that you can reliably cure with CAR T cells.

We’re going after a broad array of solid tumors, and some of the targets we’re going after are expressed on different types of tumors as well. We try to minimize the work and maximize the patient population that could benefit, and it’s dictated by which ones we have targets for, but it’s much less dictated by the tumor type itself.

Is there anything else you’d like to highlight regarding this area of research?

In this day and age, it should be noted that this technology, like many novel technologies, came out of academic institutions where innovation takes place. Even though this is technology that’s marketed by industry, it wasn’t developed by industry, it was developed by academia. In an era where funding is becoming more difficult to get, people need to realize that if we’re going to continue to make strides forward in these types of promising therapies, it’s going to come out of academia. This technology 100% came out of academia, it [has been] National Institutes of Health sponsored and funded, and would not have been possible without academia and without support for academia.

References

1. Monje M, Mahdi J, Majzner R, et al. Intravenous and intracranial GD2-CAR T cells for H3K27M+ diffuse midline gliomas. Nature. 2025;637(8046):708-715. doi:10.1038/s41586-024-08171-9
2. Stallard J. How MSK is teaching CAR T cells to attack solid tumors. Memorial Sloan Kettering Cancer Center. September 3, 2024. Accessed March 5, 2025. https://www.mskcc.org/news/car-t-breakthrough#:~:text=3%20Future%20CAR%20T%20Cell,to%20be%20harnessed%20more%20safely