Harnessing 'Immune Power'

CAR T Cell Therapy: 'It's the Real World Now'

Caron A. Jacobson, MD, MMSc; Marcela V. Maus, MD, PhD; David Miklos, MD, PhD


January 24, 2019

Caron A. Jacobson, MD, MMSc: Hello. I'm Caron Jacobson, medical director of the Immune Effector Cell Therapy Program at the Dana-Farber Cancer Institute. Welcome to Medscape Oncology Insights, coming to you from the 2018 American Society of Hematology (ASH) annual meeting.

We will be discussing some of the highlights of chimeric antigen receptor (CAR) T-cell research presented here, as well as developments in the field more broadly. Joining me in this discussion is Marcela Maus, director of the Cellular Immunotherapy Program at Massachusetts General Hospital, and David Miklos, clinical director of the Cancer Cell Therapy Program at Stanford University. Welcome, Marcela and David.

Some things we are seeing here at the meeting have been updates of old data from pivotal studies of the CD19 CAR T-cell therapies in lymphoma. We also got to see a first glimpse of what these therapies look like in the real world with two series presented earlier this morning. Marcela, you were at that session. What did you think about the abstracts?

CAR T-Cell Therapy in the 'Real World'

Marcela V. Maus, MD, PhD: It was really exciting to see the progress from clinical studies to real-world use. Two sessions talked about real-world use. One was a series of 17 centers where they treated over 270 patients with CAR T cells. This was all axicabtagene ciloleucel (Yescarta®).[1] It was nice to see the centers working together. We noticed, which is not surprising, that some patients were a little bit older, a little bit sicker in terms of lower performance status, and also had transformed disease.

It was interesting that response rates stayed about the same as what had been observed in ZUMA-1.[2,3] Toxicity rates were also quite similar. There was a little bit more use of toxicity intervention, such as with tocilizumab and steroids, and I think that as the field is getting more mature, people are starting to use these especially because the response rates are not really affected. That was encouraging to see.

You presented the second one, which was great. This was a different series of centers that also treated more than the number of patients who were treated on the original ZUMA-1. Again, there were similar responses. You did some correlative analyses, which were helpful. Maybe you can tell us more about it.

The fact that two series now show that you can use axicabtagene ciloleucel in a real-world setting with broader eligibility and get similar results is really encouraging.

Jacobson: Yes, it was a slightly smaller series than the first group.[4] What was really remarkable was that when each group did univariate analysis on some of the variables that might predict for outcome and that may also predict for toxicity, we saw some similar findings. The same very few variables came out of this analysis. It was nothing about how high-risk the disease was. Worsened performance status and tumor burden were two things that predicted for worse outcome in both groups. It was interesting that rates of intensive care unit admissions, tocilizumab utilization, and steroid utilization were very similar across the groups, indicating that we may be approaching a uniformity of management for these patients.

The fact that two series now show that you can use axicabtagene ciloleucel in a real-world setting with broader eligibility and get similar results is really encouraging.

One of the outstanding questions across the different clinical trials is with the differential use of bridging chemotherapy. We tried to address that in our series because 40% of our patients did get bridging therapy. It did not seem to improve outcomes or toxicity the way we might have expected it would.

David Miklos, MD, PhD: That is a hard confounder to understand, of course, because as clinicians we felt the need to give bridging therapy. We are obviously selecting for patients who might have been sicker. I don't think we actually got the answer on bridging therapy across, even with the other study talking about bridging therapy at the Fred Hutchinson Cancer Research Center and the University of Washington analysis.[5] It's very difficult to understand what bridging therapy is doing there.

Jacobson: Yes, it's provocative, it's thought-provoking. We need a formal study that is prospective in nature that treats patients universally across the board to really understand the impact.

Strategies to Improve Response

Jacobson: We realize from the studies that there is an unmet need, because 60% of patients either relapse after initial response or don't respond to this therapy. We are seeing at this meeting a number of different strategies to try to improve responses in patients. One of them is targeting antigen loss using dual-antigen CAR T cells. I know you have an abstract.

Miklos: Our organization back at Stanford is interested in this. I think it goes back to Shannon Maude and the acute lymphoblastic leukemia (ALL) experience showing that although the patients initially had 3-month 80% complete response (CR) rates in ALL, at 12 months the event-free survival in ALL was only 50%. Eighteen out of the 19 [patients] had lost the expression of CD19. Shannon Maude published this in a beautiful New England Journal of Medicine paper.[6]

Last year, Dr Neelapu from MD Anderson [Cancer Center] presented some of the data from ZUMA-1[7] that we all participated in by providing biopsies at the time of progression. In that analysis of 21 patients, he showed seven of the 21 by immunohistochemistry staining to have lost or significantly lost CD19 expression, suggesting loss of the target. We are talking about maybe eight amino acids at that epitope being recognized by the FFC 63 antibody. That is enough, of course, for the CAR then to lose function.

My group has been interested in carrying that forward. We'll be presenting a poster with Dr Jean Oak, our pathologist, where we studied 16 of the patients who were treated with the real-world Yescarta® study since January of last year that had progression any time after Yescarta®. We put a big emphasis on rebiopsying those samples, showing that six of the 12 that we biopsied were negative for CD19.[8]

If you looked back to what was there before we biopsied, there were low[-expression] patients, low being less than 100%. Once you have that escape cohort of large aggressive lymphomas that are not expressing a target, it's very easy to see how resistance could develop.

What do you do with that? I guess you could choose a different target. Dr Lung-Ji Chang and colleagues[9] took a very similar approach by looking at pretreatment biopsies to see whether a number of different antigens, including CD20, CD22, CD30, CD38, and CD70, were widely expressed or not. They did two cohorts, roughly 18 patients each, with one cohort getting just CD19 and the other getting the double bispecific [CAR T -cell therapy]. We have not seen the exact structure of that bispecific, but we look forward to it. Just reading their conclusion, it says that the bispecifics are clearly better for clinical response. They cited six CRs, four partial responses (PRs), and one case of progressive disease (PD) in that [group getting the double] going forward 3 months and beyond. That is provocative and exciting.

I know of three centers in America, at this point, that are doing multiapproach targeting. Nirav Shah's group at Medical College of Wisconsin has been doing a double [, bispecific CAR T-cell therapy] that is targeting CD19/CD20.[10] We certainly know that Mike Jensen's group and the Stanford Alliance out of pediatrics at the University of Washington have been doing multicistronic targets of CAR T cells.[11] Our own group with the director at our center, Crystal Mackall, is working off the work that Terry Fry had begun at the National Cancer Institute. We have a single polypeptide that is targeting both CD19 and CD22.[12] It's going through a dose-escalation phase 1 [study], and we are about to complete that. My own colleague, Nasheed Hossain, will present data showing that we are safely escalating, that the CARs are persisting.

Because I've only treated nine patients so far and we are moving to the final dose where we will expand, perhaps most provocative is that when patients did have progression, both in our group as well as in the Wisconsin group, nobody has demonstrated CD19 loss yet.

Maybe we can take the antigen evasion question off the table by hitting it with a double.

Jacobson: Targeting more—an HIV approach.

Miklos: Zidovudine (AZT) was always a good drug; it just failed by itself.

Armored CARs

Jacobson: We are also starting to see the first data for some of the armored CARs, which people have been excited about here at this meeting. Can you tell us a little bit about those, Marcela?

Maus: The concept of an armored CAR is to enable the T cell to have more than one function. Not just be able to kill a single target, but maybe get some other signals to either "immunize" it from the tumor microenvironment or just potentiate it a little bit to make it last longer.

We're starting to see some of the first clinical data now. This has been in process in terms of scientific development for a while. Jae Park and Michel Sadelain are presenting their first data with their armored CD19 CAR.[13] This is a CD 19-28z CAR that has a second transgene that encodes for 4-1BB ligand. It actually costimulates the T cell both in an autocrine way and in a paracrine way.

They have treated several patients with non-Hodgkin lymphoma; transformed follicular lymphoma; and chronic lymphocytic leukemia (CLL), including Richter transformation, and they are starting to see that the response rate is higher than the same CAR without the 4-1BB ligand was in previous trials. Second of all, there has not been any increased toxicity: the abstract reports no increase in cytokine release syndrome (CRS), no increase in neurotoxicity.

Encouragingly, the persistence of those CARs is now out to more than 100 days in their abstract. That's a significant improvement over what the persistence was in CLL and in some of these other diseases when they did not have the 4-1BB ligand. Perhaps this is a way of making some of those patients respond who initially might not be a responder to a standard standalone 19-28z CAR. It would not overcome antigen loss necessarily, but it gets at some of the issues of T-cell fitness and perhaps loss of functionality.

Jacobson: It may get over some antigen loss, though, if you have some epitope spreading within the T-cell infiltrate in the tumor targeting different tumor epitopes.

Miklos: It's important to have multiple strategies to improve the CAR. We can't just rely on any one.

Jacobson: One thing I found striking about the real-world experience was that the things associated with outcome were very different from the things that were associated with toxicity. It's unlike transplant, where it's been very hard to disassociate toxicity from outcome. We have learned that there is no reason you need to have toxicity to have good outcomes. Some of these approaches to improve safety alone are great, but if they can also improve outcome at the same time, that would be phenomenal.

Miklos: Perhaps the simplest of those would be age. For years, we have avoided transplant in the elderly. We say it is going to be too toxic, too difficult. In this session, we saw the presentation from Dr Neelapu and his group at MD Anderson where they went over age as a simple cutoff for receiving axicabtagene ciloleucel.[14] The cutoff was 65 years of age. They went through the usual patient characteristics.

They had 20 patients over the age of 65, and roughly twice that number that were below 65 being compared for outcomes. Rates of CRS were comparable. Rates of neurologic toxicity were exactly the same. [This is similar to] what we have seen across ZUMA and the real-world experience. Grade 3 neurotoxicity of 30% is the bugaboo of the system right now. Isn't it remarkable? These were people up to the age of 83 getting a therapy and surviving.

I think I only saw three patients dying among the 500 that were presented today, during nonrelapsed mortality on these studies. I think the [key] is to understand that this is immunotherapy, and it has different toxicities that are going to be amenable to older patients.

Let's make one more point, I'm bookended here. Women did better than men did.

Jacobson: Both from the science side and from the patient side.

Miklos: Isn't that better? The women recipients of this immunotherapy statistically are having better outcomes. Older people are having tremendous outcomes. There is a lot to study to understand this, but it's wonderful, and it gets to how we want to make sure we don't underserve our patients. Please get your patients to the centers that have this experience, and try to adopt this therapy if you don't have a center near you.

Please get your patients to the centers that have this experience, and try to adopt this therapy if you don't have a center near you.

Combination Studies

Jacobson: I totally agree. The other thing I'll say about the persistence that you noted with the 4-11B ligand armored CAR is that there are some combination studies looking at combining the conventional CAR T cells with other agents. I have a poster on Monday night looking at atezolizumab in combination with Yescarta®.[15] We show not only a greater expansion of CAR T cells both by area under the curve (AUC) and by peak CAR T cells, but also a greater persistence out.

Other combinations with PD-1 agents are being presented here, but a very intriguing abstract is being presented on the combination of ibrutinib in patients with CLL.

Miklos: The University of Pennsylvania group came forward at the American Society of Clinical Oncology [meeting] maybe a year and half ago with the first announcement that they had treated patients with CLL. Up until now, response rate has been roughly 50%, and we have always wondered why. Dr Saar Gill and the rest of the University of Pennsylvania group have taken that 4-1BB construct and added on the use of ibrutinib, a standard therapy for patients with CLL before CAR T-cell [therapy], leaving the patients on the therapy and extending that treatment during the management of the CAR T recovery.[16]

It's quite dramatic. They treated 19 patients, 18 of the 19 remain alive, and 15 of the 19, or 78%, had MRD (measurable residual disease) undetectable disease by 3 months after the CAR-T therapy. This is dramatically different from what people have seen in CLL so far. Why? Maybe you want to speak to why.

Maus: Ibrutinib has dual effects on Bruton tyrosine kinase (BTK) and interleukin-2–inducible kinase (ITK), and my lab was involved in some of the preclinical work leading up to this study. It really does seem to change the T-cell fitness, which is, I think, going to be a really interesting parameter. We are seeing a lot of different ways to overcome [resistance] in some of the few patients who don't respond at all. You can increase T-cell fitness perhaps by giving ibrutinib or ITK small-molecule drugs. You might be able to also improve it by checkpoint blockade with anti–PD-1. You might be able to engineer the T cells to become resistant to these molecules or to be able to receive additional costimulation.

The field is growing tremendously and rapidly. That room was packed today. And I was just moved by the group's collective breath when Loretta showed a PET scan of a patient with diffuse disease who was able, after an infusion of CAR T cells, to completely clear that PET scan.

Miklos: With an ejection fraction of 45%. Age above 65.

Maus: Pulmonary embolism, poor performance status.

Jacobson: As a lymphoma doctor, that was one of the largest disease burdens I've ever seen. And it was a single infusion. That was totally amazing.

Miklos: It was a complete response. Wow.

Jacobson: We are going to learn a lot about [it]. Right now, we have just been infusing a hodgepodge of T cells. We don't know anything about the phenotype of the T cells. We learned from your lab and from others at Penn about T-cell fitness, and some of the characteristics from responding patients in terms of their pretreatment T cells and the T cells in the product. Improving the T-cell composition of the product, either in vivo before you collect the T cells or even in the laboratory, [is a possibility]. Bluebird Bio, Inc., is presenting a study using a PI3 kinase inhibitor to try to do this themselves.[17]

Maus: I think getting those postinfusion biopsies of the tumor [is important] so we can understand the mechanisms of resistance that can occur. You showed two mechanisms in your study.

Jacobson: [For] the only two we looked at, we saw two different mechanisms.

Maus: You learn so much from each of them. Maybe not right now, but eventually I think being able to make treatment decisions about what therapy should come next is going to be part of that.

Miklos: That is true. Hopefully, we will be able to try to take some of this correlative science and apply it to patients in real time to decrease the risks that they do still face. Again, I want to emphasize [that] nonrelapsed mortality is less than 2% today in the real world. Nobody got graft-versus-host disease. This is a different therapy from allotransplant. The immune power has been harnessed without the off-target toxicities. It's really the real world now.

The immune power has been harnessed without the off-target toxicities. It's really the real world now.

What Next?

Jacobson: It's amazing. Was anything not presented at this meeting that you think is important and coming down the pipeline that we might see next year at ASH?

Miklos: Where I practice, there was a lot of clamoring about where are all the clinical trials? We did the initial phase 2 studies, then we all have an experience now of doing commercial work over the last year. I just want to make the point that there are a whole slew of new opportunities for patients to participate in what I would call the next generation of clinical trials. [Trials of] drugs where constructs are already available and [they are being] augmented, and other trials looking at bispecifics. It's going to be an exciting next couple of years.

Jacobson: Maybe some of the gated CARs and the CARs that actually use physiologic T-cell receptors signaling to try to modulate activity could be very interesting too.

Miklos: It remains unanswered. What is the best T cell to transduce? This is an important study that is going to come out of patient experience and those repeat biopsies. That is going to be the Holy Grail too.

Maus: What percentage of patients with lymphoma do you think are being offered this therapy?

Miklos: We estimate that there are 14,000 patients with relapsed refractory diffuse large cell lymphoma nationally. We know for a fact that less than 1000 patients have been treated thus far. Far less than 10% are getting the therapy that is curing lives. I hope that changes. That is my mission this year: Get the word out. That is why we are doing this.

Jacobson: That is right. Marcela and David, thank you for joining me. This has really been a fascinating discussion. Thank you, the audience, for joining us. This is Caron Jacobson at ASH 2018 in San Diego.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.