Modified T cells offer a viable step toward glioblastoma treatment: A discussion with Dr. Marcela Maus

Background
There’s no curative treatment for glioblastoma. Few treatment advancements have even been made in the past 25 years. Median survival is still less than two years for patients with newly-diagnosed glioblastoma.

However, immunotherapy appears to be a promising treatment. All that's needed is a breakthrough. A new study, published in the journal Science Translational Medicine, has provided at least a step–an important step–in that direction. The results of this study indicate that CAR (chimeric antigen receptor) T cell immunotherapy is a viable option for treating glioblastoma, even in refractory patients.

In this interview, senior author Marcela Maus, MD, PhD, discusses how this investigative therapy takes a different tack than other cancer immunotherapy research. She also describes how (and how well) the therapy works, and explains the current and future hurdles for this experimental therapy to become an accepted treatment.

Marcela Maus, MD, PhD, is Assistant Professor of Medicine at Harvard Medical School and Director of Cellular Immunotherapy at Massachusetts General Hospital Cancer Center in Boston.

MDLinx: This is the first such study of its kind. What makes it unique?

Dr. Maus: There has been other research that studied CAR T cells in brain tumors, but this is the first test of genetically-modified autologous CAR T cells that targeted a specific mutation in brain tumors called EGFRvIII (epidermal growth factor receptor variant III). EGFRvIII is expressed in about 30% of patients with newly diagnosed glioblastoma, so it's meaningful for a substantial number of patients.

One of the things that made our study unique is that we had patients who had surgery after their CAR T cell infusion, so we were able to ask some scientific questions about whether the CAR T cells were getting to the brain tumor, and in what quantities, and whether the cells were having any effect on the brain tumor at the pathologic level.

Because this was the first time that we tested this particular CAR T cell product, we started off fairly conservatively in that we gave a single dose intravenously. We did show that it was safe and feasible to manufacture these cells for patients, even though they had poor prognosis and had prior chemotherapy and radiation.

MDLinx: Have other studies found that CAR T cells are not safe?

Dr. Maus: It runs the gamut. For leukemia, CAR T cells have been incredibly effective. They are associated with toxicity but it's been generally manageable. There have been some other studies where genetically-modified T cells have not been safe because they've targeted other life-sustaining tissues such as the lungs or the heart. But we didn't see anything like that with this study.

In many other CAR T cell trials in solid tumors, the main issue has been that the T cells haven't been potent enough to mediate a very strong anti-tumor effect.

MDLinx: In your study, what showed you that CAR T cells had an effect–that they indeed reached the brain tumors?

Dr. Maus: The cells are "marked"because they do have this CAR sequence in them that isn't naturally present in the human body. In the cohort of 10 patients that we treated, seven of the patients had surgery on their brain tumor after the CAR T cells were infused. This gave us an opportunity to look for the CAR sequences in the tumor specimens from those seven patients. We were able to show that the CAR T cells did get into the brain tumor and expand there.

MDLinx: About how long did the effect of the infusion last?

Dr. Maus: We can't draw strong conclusions on the exact kinetics because each patient had surgery only once, so it gave us just a snapshot in these seven different patients. But with that snapshot, coupled with the data that we have from how long the CAR T cells were in the blood, it seems that the T cells were at the highest levels in the brain tumor anywhere from six days to two weeks after infusion, and were barely detectable after two months.

MDLinx: Is that long enough to provide a meaningful effect?

Dr. Maus: Well, nobody knows. We know that in leukemia, having T cells stay longer is a good thing in terms of having durable remission. What we saw in the brain tumors was that some of the patients had decreased expression of the EGFRvIII target, which indicates some direct anti-target activity. In those cases, you could say that the CAR T cells did the job they were trained for.

But sometimes you need to have the T cells continue to do that job longer and deeper. For instance, we had one patient where we couldn't really find the CAR T cells in the brain tumor even though the surgery was done only a month afterward. The CAR T cells were at a very low level in the blood of that patient, too. We presume that the CAR T cells have to get to a certain level in the blood before they can get to the brain in substantial numbers, and only when they get into the brain tumor can they proliferate in response.

As for the participants in this trial, we do have a patient who is still alive for close to two years now, which is unusual for glioblastoma and particularly for patients like her.

MDLinx: What were the other results?

Dr. Maus: One of the challenges with assessment is that we follow brain tumor by MRI. On MRI, inflammation and immune reactions look very similar to tumor, so it's very difficult to know exactly how much is tumor and how much is inflammation from immune therapy. So we learned a little bit about the imaging changes that can occur with these CAR T cells.

In terms of other results, we found that the tumor adapted to the CAR T cells. The CAR T cells lost the EGFRvIII target that they were going after, which indicates activity against the tumor, but the tumor also mounted an immunosuppressive response. Tumors that didn't have high levels of immune checkpoints all of sudden had high levels of immune checkpoints, and started making the proteins that turn off T cells. So it seems that the tumors didn't just sit idly by while this immune cell infiltrate was happening; they actually responded to it.

MDLinx: Was that response expected?

Dr. Maus: Not really; at least I didn't expect it. I don't think it had ever been shown before with CAR T cells in solid tumors, or in any other kind of tumor.

MDLinx: What do you think could be done to overcome that immune response?

Dr. Maus: A couple of things. There are now a lot of checkpoint blockade antibodies like anti-PD1 and anti-PDL1 that specifically address this immunosuppressive environment that's present in tumors. Also, as I said, we were fairly conservative in giving only one dose but one could consider giving more doses, and giving it in combination with checkpoint blockade antibodies to change that immunosuppressive environment. One could also potentially administer the CAR T cells directly into the tumor itself so that we could get higher numbers into the tumors.

Even though the T cells were able to get into the tumor, what we don't know is how many you need to be there in order to have a potent antitumor effect.

MDLinx: What's the next step in this line of research?

Dr. Maus: This study was done at the University of Pennsylvania. I've since moved to Massachusetts General Hospital. In my laboratory now, we're going back to the bench. We're going to design CAR T cells that are resistant to this immunosuppressive environment and that can address some of the antigen escapes. We're making CAR T cells that will target two antigens on glioblastoma instead of just one, and we're modeling the effect of the immunosuppressive Treg cells on CAR T cells directed to glioblastoma.

About Dr. Maus: Marcela Maus, MD, PhD, is an Assistant Professor of Medicine at Harvard Medical School and Director of Cellular Immunotherapy at Massachusetts General Hospital Cancer Center, in Boston, MA.

This study was supported by an alliance between the University of Pennsylvania and Novartis for the development of CAR T cells for cancer.