Expert Insight: Clinical development behind a novel chimeric antigen receptor T-cell therapy - on demand webinar

Distinguished speaker Dr. Bruce Levine addresses some of the current healthcare challenges that impact the development of new therapies

22 Nov 2019

Dr. Bruce Levine, Barbara and Edward Netter Professor in Cancer Gene Therapy at the University of Pennsylvania and Founding Director of CVPF

Some of the latest advancements in gene delivery technology are highlighted in a SelectScience webinar presented by Dr. Bruce Levine, Barbara and Edward Netter Professor in Cancer Gene Therapy, University of Pennsylvania, and Founding Director of CVPF. This includes the creation of engineered T-cells for developing chimeric antigen receptor (CAR)-T cell therapies and other novel therapeutic approaches such as T cells resistant to HIV infection. 

This webinar is now available to watch on demand >> 

In particular, Levine discusses the clinical development of a unique CAR-T cell therapy and addresses the current healthcare challenges that impact the development of such a treatment.

Watch this on-demand webinar to learn the following:

  • The mechanism of action of CAR-T cells
  • The clinical development from bench to bedside of a unique cell therapy
  • The vein-to-vein complexities and healthcare challenges of developing a paradigm shift in therapies


Here are the highlights from the Q&A discussion at the end of the webinar:

Q: Do PR or NR patients benefit from repeated CAR-T infusions?

BL: It depends. In our pediatric acute lymphoid leukemia patients, the clinicians have been using B-cell aplasia as a surrogate marker for the persistence of the CAR-T cells and, because of the kinetics data which I showed earlier in the presentation, we know what the average aplasia and the area under the curve of the CAR-T cells should be. If they see early disappearance of the CAR-T cells using this marker, then in some of our clinical trials we have the option to give an additional dose or a supplemental dose. In many trials that's not the case, so we just give one infusion. Now, there are other avenues, as I did present. With lymphoma, we have a trial open where patients roll over into a clinical trial where they can receive checkpoint antibody therapy. So, we're just learning the kinetics of the T cells and how this is reflected in B-cell aplasia, and we'll be learning more as the data comes in from our solid tumor clinical trials and those of collaborators, other institutes and other industry players in the field.

Q: These therapies are very personalized and very expensive. Do you see in the future the possibility of a more off-the-shelf solution?

BL: People often ask, are we going to transition from autologous CAR-T cells to allogeneic CAR-T cells? Are they going to replace autologous? I don't see that happening anytime soon. It's still very, very early in the development of off-the-shelf allogeneic CAR-T cells and CAR-NK cells. For the most part, when those off-the-shelf CAR products are used, the patients are transitioning to a definitive therapy, most often a stem cell transplant. Now, there are improvements being made to the off-the-shelf products using HLA insertion and a knockout of MHC class II and class I, but we'll have to see how long those cells can possibly last.

So, I think the off-the-shelf products are more likely to be used in the setting of minimal residual disease or very low disease burden patients. I don't see how they could be effective in high disease burden patients. But, I will say that there absolutely is a role for these off-the-shelf products because we have a number of patients, particularly those who have had their disease for many years and who have undergone multiple rounds of chemotherapy, from whom we cannot collect enough cells, or from whom we are not able to generate a clinical dose. This also applies to very young children or infants with ALL (acute lymphoblastic leukemia). It can be difficult to generate an autologous CAR-T dose. So, in some, I do see a co-existing role for off-the-shelf and autologous products and look forward to more clinical data.

Q: What is your take on the use of Dynabeads versus the CliniMACS beads?

BL: I'll state upfront that I have a conflict of interest here in that the development of the Dynabeads was my postdoctoral research project. So, I believe in them and a lot of other people do as well. They generate a very potent activation signal and are very efficient in growing T cells. The CliniMACS beads work as well and certainly if you're using the Prodigy system, it’s a requirement to use these. There have been very few studies comparing them head to head, but in our center, we're continuing to use the Dynabeads.

Q: Do you use a Prodigy system or open system of expanding cells?

BL: There are many groups that are using the Prodigy for generation of CAR-T cells in clinical trials. We, in Philadelphia, are using a combination of gas permeable culture bags and the WAVE Bioreactor. We've looked at a number of open systems, and other systems, and the Prodigy, but currently we're using this system of a static gas permeable bag and the WAVE Bioreactor.

Q:  We've talked a little bit about CRISPR during the presentation. A question has come in asking for more detail on how the T cells were genetically modified.

BL: The way that we're delivering the genetic material for the CAR is with a lentivirus, and we've been using lentivirus for about the past 15 years. We are the first center in the world to do a lentiviral clinical trial. Prior to that, we did use a retrovirus for the CD4 Zeta clinical trials in collaboration with Cell Genesys. As far as CRISPR, we're delivering ribonuclear proteins by electroporation. There are a number of other viral vector-free methods in development beyond electroporation, but I'm not aware that any have made it to clinical trials.

Q: There are some clinical trials using irradiated NK-92 cells, what's your opinion on the possibility of using irradiated T cell lines to generate off-the-shelf T cell therapy?

BL: That's a very interesting question. There are groups that are looking at using viral-specific T cell lines and inserting the CARs in those T-cells as they may receive stimulation and restimulation by multiple methods. As for NK-92 cells, I don't know the results of groups that have been looking at those for inserting CARs, but it may be that there's another question embedded in there related to using T cells as antigen-presenting cells, and we've done that in the lab. We're looking at using CD19 expressing cell lines as vaccine boosters. The group at Fred Hutch is doing that. Finally, we have used not a T cell line, but a K562 cell line that has been genetically modified to serve as artificial antigen-presenting cells for T cells. These would be in lieu of the Dynabeads or the Miltenyi beads, and they work quite well and have been used in clinical trials, not of CAR-T cells but of regulatory T cells, by the group at the University of Minnesota.

Q: You mentioned that there are certain characteristics that one would have thought were related to clinical efficiency, when they're not, e.g., percent viability. In your clinical trials, would you eliminate these QC spec measurements, and has the FDA been receptive to these kinds of changes based on clinical outcomes?

Do we know that a viability of 59% won't induce a complete clinical response in patients?

Dr. Bruce Levine

University of Pennsylvania and Founding Director of CVPF

BL: I would not eliminate viability or the other tests as release criteria. I present the data to be provocative in thinking about what is acceptable when we have a drug that is dividing. The viability is a marker of quality, but remember that each of these products is a lot size of one. Do we know that a viability of 59% won't induce a complete clinical response in patients? Well, I just showed you data indicating it does. But if we have a product where one would expect, under ideal conditions, to have a final product viability of 90 or 95, and let's say the viability comes in at 65 or 70, then that can point to issues in the process, formulation or the storage of the cells. So, I see viability and the other release criteria as important and useful, but we have to remember that we're dealing, pharmacokinetically, with a drug that is of many different characteristics than standard drugs. Where we have the bar set, I don't know, but we'll continue to collect the data on the viability and hope to have that set at the appropriate level.

Q: What are the criteria for assay validation for final product release tests? Are there any lab or company-specific?

BL: The criteria is what the FDA or your regulatory agency will accept. What they will ask is to show the testing under conditions that are expected and unexpected, and the minimum and the maximum concentration of an analyte of the cells in various buffers. So, you need to be able to collect enough data to be confident that under field conditions, meaning under manufacturing patient material, that the release test is going to be giving you valid answers. I can tell you that we have evaluated a number of assays that can present problems and can be very sensitive. Therefore, it's important to communicate with the agency through the INTERACT Program or through a pre-IND meeting about your release tests and the data that you have to support those tests.

Q: Do you think patient stratification using immune profiling is important in CAR-T cell therapy?

BL: While in pediatric acute lymphoid leukemia we have a complete response rate in the order of 83% to 90%, in lymphoma, for Kymriah and Yescarta and others, the overall response rate is lower (around 45% to 50%) and it's about that for chronic lymphoid leukemia as well. We have done a series of biomarker studies to show that a particular phenotype, a central memory stem cell, in CD8 cells is predictive of a good clinical outcome. Now, that leads to the question: We have an expensive therapy, we have a method of biomarker analysis that could predict, not with 100% certainty, but with some indication in cells collected from a patient, what product is likely to induce a clinical response and what might not induce a clinical response? 

So it is, I think, foreseeable that we will be implementing some biomarker criteria and some eligibility criteria. We'd like to do that in a clinical trial in CLL (chronic lymphocytic leukaemia), for example, and to look at the stratification of response when one looks at these biomarkers and what we believe is going to be a response. It's both a very interesting scientific question, but it is also a health policy question and an ethical question. So would you want to collect cells from patients if you had indications it's fairly unlikely, and I'll just throw out a number of, say, 20%, that they would respond as opposed to another patient where you have biomarker analysis that would predict a, let's say, 70% or 80%, response? 

Q: Can we have your opinion on the idea of using CAR-T cells against more than one target? For instance, infusing the patient with CAR-T against both CD33 as well as another cell, 30CD.

BL: I mentioned a couple of trials. We're doing one in myeloma targeting BCMA and targeting CD19, where there is some evidence CD19 is present on the myeloma stem cell. We and others, Stanford included, are targeting both CD19 and CD22 in leukemia to prevent antigen escape, and, relevant to the question on CD33, that is a special case because what we're planning to do in the CD33 trial is to create a CD33 negative stem cell population using CRISPR. In the cells that mature from them, those stem cells would be resistant to the CAR-T 33 cells that would be infused to the patient to kill acute myelogenous leukemia. So, in that particular trial, we don't see a combinational approach, at least not at first, but the point is valid not only in hematologic malignancies but also in solid tumors to either target two antigens, or perhaps a targeted tumor antigen, and a tissue-specific marker to enable CAR activation with a switch.

Q: Would you mind giving an overview on paradigm shift, if possible.

BL: Well, that's an open-ended question. So, the way I'll address that is that we have cells and tissues as drugs. We have engineered cells that can be thought of as drug factories. There is preclinical work showing that CAR-T cells can also secrete proteins, cytokines, and bispecific antibodies. Oncolytic viruses can be engineered to produce bispecific antibodies. So, in thinking about how we develop these therapies, of course, it is challenging. I think what's not appreciated is that the regulations, and often regulators in some countries, don't have experience with these types of therapies. It's a multifaceted effort to develop this field, not only scientific and tactical, but with regulation, education, and training as well.

Learn more about CAR-T cells as a model for clinical development: Watch this webinar on demand >>