Gamma Secretase Inhibitor Improves Responses to BCMA CAR T Cells in Myeloma

In an interview with Targeted Oncology, Andrew J. Cowan, MD, discussed the findings from the first-in-human clinical trial evaluating the combination of a GSI and BCMA CAR T cells in patients with heavily pretreated multiple myeloma. He highlighted the next steps for this research and how he sees CAR T-cell therapy evolving over the coming years.

Andrew J. Cowan, MD

B-cell maturation antigen (BCMA) chimeric antigen receptor (CAR) T-cell therapy induced promising responses in patients with multiple myeloma, but most patients tend to relapse over time, said Andrew J. Cowan, MD, assistant professor of Medicine, Division of Medical Oncology, University of Washington School of Medicine; physician at Seattle Cancer Care Alliance; and assistant member in the Clinical Research Division, Fred Hutchinson Cancer Research Center. According to a presentation at the 2019 American Society of Hematology (ASH) Annual Meeting, a phase I clinical trial (NCT03502577) is evaluating the combination of a gamma-secretase inhibitor (GSI) therapy with BCMA CAR T cells to improve response in heavily pretreated patients.

BCMA is widely expressed on the surface of plasma cells in patients with multiple myeloma, which makes it a reliable target for this patient population. BCMA CAR T cells have induced high response rates initially, but the durability of these responses remains a challenge.

Preclinical data suggest that GSIs can increase BCMA density on the surface of plasma cells. BCMA CAR T cell responses could improve with an increased target density. This was the rationale for combining the use of GSIs with BCMA CAR T cells in patients with multiple myeloma in the phase I first-in-human trial presented at ASH.

The trial, which included heavily pretreated patients with multiple myeloma, led to promising responses. The best overall response rate (ORR) was 100% among 6 evaluable patients who were treated with the GSI JSMD194 and BCMA CAR T cells. This included 5 very good partial responses (PRs) and 1 PR. Five patients achieved minimal residual disease negativity by flow cytometry. No patients relapsed as of the data cutoff date of July 15, 2019.

“When we use [the GSI] in combination with BCMA CAR T cells, we have seen promising results with high response rates at very low dose levels, as well as very durable responses,” Cowan said. “The findings are encouraging that this could be a way to make BCMA CAR T cells better.”

In an interview withTargeted Oncology, Cowan discussed the findings from the first-in-human clinical trial evaluating the combination of a GSI and BCMA CAR T cells in patients with heavily pretreated multiple myeloma. He highlighted the next steps for this research and how he sees CAR T-cell therapy evolving over the coming years.

TARGETED ONCOLOGY: Could you provide some background to why you evaluated GSI in patients with multiple myeloma?

Cowan:One of the more common targets for CAR T-cell therapy in multiple myeloma over the past few years has been BCMA. The reason that this is employed as a target is because it is widely expressed on plasma cells but is not expressed in other tissues, which are the typical characteristics of a good target.

We have recently learned 1 thing about BCMA from a publication inNatureCommunicationsa few years ago. They showed that a cell membrane protein called gamma secretase, which is found in plasma cells, is responsible for cleaning BCMA from the surface of plasma cells and releasing it into the blood as soluble BCMA, which is emerging as a biomarker of multiple myeloma disease burden. Potentially, that would be 1 way that plasma cells could have reduced BCMA expression.

In a recent publication through our work inBloodin November 2019, we discussed the use of a class of medications called GSIs, in which the story behind is very interesting. GSIs were developed for Alzheimer’s dementia. They were studied in Alzheimer’s in many studies. The GSIs didn’t pan out there, but it was also studied in cancer. It was incidental that we discovered we could use these and increase BCMA expression on the surface of plasma cells and cell cultures. That led to some xenograft models, where we showed it worked in that scenario as well.

The next logical step forward was to see if we could increase the target density. One thing we know about CAR T cells is they work better when there is more target density. If we could use GSIs to increase BCMA on the surface of plasma cells, it could result in deeper and longer responses when treating patients. We showed that in the mouse models that are in the paper, and that led to the development of the phase I study we presented at ASH this year.

TARGETED ONCOLOGY: How was the trial designed?

Cowan:This was a phase I study, and we employed fully-human BCMA CAR. It uses a 4-1BB costimulatory domain, and similar to our other CAR, we use an equal ratio of CD4 to CD8 T cells, which we feel creates a composition that is more reproducible amongst patients when they receive it.

Eligibility for this study meant that you had to be resistant to more than 4 cycles of a proteasome inhibitor or an immunomodulatory drug or have relapsed after transplant. They had to have more than 10% of plasma cells in a bone marrow biopsy. Those patients who were eligible underwent leukopoiesis, and then they underwent a brief run-in with the GSI. That drug is called JSMD194, an oral GSI that has been studied extensively in Alzheimer’s and inhuman cancer patients. The dose we used for the study was 25 mg 3 times a week.

When the patients originally undergo leukopoiesis, they undergo a run-in with GSI to determine the effects of the GSI on BCMA expression and soluble BCMA. We perform a bone marrow aspirate on day 5 of the run-in, and we compared BCMA expression on that aspirate to the BCMA expression on a screening bone marrow biopsy. The way we determine expression is by flow cytometry that we do here at the UW pathology lab. We also looked at the impact on soluble BMCA.

During this period of time, patients are undergoing cell manufacturing; then once the cells are available for release, we treat with lymphodepleting chemotherapy using cyclophosphamide and fludarabine. Patients get CAR T cell infusion on day 0, and on the same day, they start the GSI at 25 mg 3 times a week for a total of 3 weeks.

TARGETED ONCOLOGY: What were the findings you presented at ASH?

Cowan:We have treated 10 patients so far, and the patients we have treated have been very heavily pretreated. The median number of prior lines of therapy was 10. One patient had been treated with 23 lines of prior lines of therapy. Almost all patients had an autologous stem cell transplant (SCT). One patient had a prior allogenic SCT (alloSCT). Nearly all patients had a high-risk feature at the time of entry to the study. The highest dose level we have reached so far is 300 x 106.

As this was a phase I study, the primary goal was to look at safety and determine the optimal dose. We have not determined the optimal dose yet, but we do have a readout on the safety so far. With respect to safety, cytokine release syndrome (CRS) was fairly common in our study. Actually, 100% of the patients had CRS, but it was most commonly grade 1/2. Neurologic toxicity was also common, but in the vast majority of cases, it resolved with 1 to 2 doses of dexamethasone. No patient had any tumor lysis syndrome. We did have 1 DLT on the study, and that was a patient who ultimately passed away at day 33 after treatment in the setting of a disseminated fungal infection, grade 4 CRS, and neurologic toxicity. That patient was the one I mentioned that was heavily pretreated with multiple prior lines of therapy. The DLT did motivate some changes in the protocol, including addition of performance status and eligibility criteria which we previously had not included.

One of the exciting things about this study was looking at the impact of the GSI on bone marrow plasma cell BCMA expression. That is the run-in I mentioned. When we compared the screening to the post-GSI run-in, we found that all patients increased their BCMA on the surface of plasma cells. The median BCMA expression after GSI was 99%. We also looked at a measurement called antibody binding capacity, which measures the total number of BCMA binding sites. That increased dramatically by a median of 20-fold to a maximum reported value of 57-fold. These findings confirmed our preclinical work that we saw in our translational studies. We could do the same thing with GSI in humans.

In respect to efficacy, these are early results, but the results are promising. We have a 100% ORR in patients treated. All but 1 patient had no evidence of abnormal plasma cells in the bone marrow by flow cytometry at day 28. Even in the lowest dose level, which was 50 x 106, every patient had more than at least a PR, and 1 patient had a complete response. With a median of 20 weeks of follow-up, no patients have relapsed to date.

These are early findings so more follow-up is needed, but the findings we have seen so far are encouraging. We are seeing deep responses. The responses appear to have been durable so far. We have seen rapid responses, just like we have seen in our other BCMA trials. The key thing will be to see if the responses are more durable with this combination because as most are aware, response rates are high with CAR T cells. The distinguishing factor will be are the responses more long-lasting. That remains to be seen, and we need to treat more patients and get more follow-up time to know whether that is the case.

TARGETED ONCOLOGY: What are the next steps for this research?

Cowan:Everybody is looking for ways to improve CAR T cells, whether it is using checkpoint inhibitors or using different dual combination targets. I think everybody is looking for a way to make the CAR T cell survival a little bit better. The thing people are disappointed about is that even though a high fraction of patients respond to CAR T cells, almost all patients relapse unlike CD19 CAR T cells where we see a proportion of patients, around 30% to 40%, who have long-term durable responses. We are just not seeing that with multiple myeloma.

The next steps for this are to treat more patients once we get that optimal dose to expand the number of patients we are treating. Currently we are treating 19 patients on this study, but expanding beyond that would be excellent. Ideally, having some kind of phase II randomized study comparing BCMA CAR T cells to BMCA CAR T cells with a GSI will get prospective comparison of these 2 approaches. I think those are the critical next steps, having more patients and more studies as we develop it further.

TARGETED ONCOLOGY: What is the main take home message from these data?

Cowan:GSIs do increase BCMA expression in patients with multiple myeloma. When we use this in combination with BCMA CAR T cells, we have seen promising results with high response rates at very low dose levels, as well as very durable responses. The findings are encouraging that this could be a way to make BCMA CAR T cells better.

TARGETED ONCOLOGY: What are your thoughts on how you have seen the role of CAR T cells evolve in this patient population over the last few years?

Cowan:One thing I always tell patients is that this is very early. The analogy is when alloSCT was first done in the 1980s to now. We have seen dramatic improvements in how we manage toxicities and how we improve long-term outcomes. This is the beginning [for CAR T cells].

Some examples of how we are improving are by using a fully-human BCMA CAR. While that might seem trivial, it is important because there is potential for reduced immunogenicity or reduced potential for the immune system to eliminate the CARs. If you think about it, if there is a mouse antibody and there are some CAR T cells are used, that has a greater potential for immunogenicity. Using a fully-human CAR might be better.

Another promising area is using checkpoint inhibitors with BCMA CAR T cells. I also found the data in lymphoma for obinutuzumab (Gazya), a CD20 bispecific antibody to be exciting. We saw patients respond who have been treated with prior CAR T cells. People are talking a lot about whether it’s going to be CAR T cells or bispecific antibodies, but I think we should leverage the capabilities of both. I don’t think they have to be mutually exclusive. I think the combination approach would be a logical step forward.

There are a lot of exciting ways to move the field forward. The key will be trying all these approaches and not being restricted to just saying we have BCMA CAR T cells, they work like this, and that’s it. I think we have to improve on it, and I hope our GSI study turns out to be one way we can do that.


Cowan AJ, Pont M, Sather BD, et al. Efficacy and Safety of Fully Human Bcma CAR T Cells in Combination with a Gamma Secretase Inhibitor to Increase Bcma Surface Expression in Patients with Relapsed or Refractory Multiple Myeloma. Presented at: 2019 ASH Annual Meeting and Expo; December 7-10, 2019; Orlando, FL. Poster 204.