Recently published in the New England Journal of Medicine, researchers from Stanford University found that comprehensive genetic characterization of T-cell lymphomas after CAR T-cell therapy remains essential for understanding the risk of potential tumors with this therapy. Our own Dr. Todd Druley sat down with two of the publication’s authors, Dr. Ash Alizadeh and Dr. Mark Hamilton, to discuss their findings. Interview has been lightly edited for clarity.
Todd Druley, MD, PhD: Welcome everyone. I’m excited to be joined today by 2 really outstanding physician scientists, Mark Hamilton, MD, PhD. He’s a Heme-onc and Cellular Therapy fellow at Stanford University, and Professor of Medicine at Stanford, Ash Alizadeh, leader of the Cancer Genomics program. We’re here to discuss their recent New England Journal of Medicine article, “Risk of Second Tumors and T-cell Lymphoma after CAR T-cell Therapy” that was recently released on June 13, 2024. Very exciting reports. And we wanted to learn more from them about what drove this particular study and how they were able to use the Mission Bio Tapestri platform to further inform the conclusions of the study.
Thank you both for joining.
Ash Alizadeh, MD, PhD: I’ll give a big picture overview and then have Mark dig into the details. This study was kind of prompted by the announcement by the FDA around Thanksgiving of last year, of the risk of second cancers after CAR T-cell therapy– in particular, T-cell lymphomas, that can arise after administration of CAR T-cell therapies. At the time there were 6 approved CAR T-cell therapies when this announcement was made. This made us think about this problem in terms of how prevalent this problem was and had us examine our experience at Stanford since 2016, looking to see how many patients we’ve given various cell therapies to and how often we saw second cancers.
In looking at over 700 patients and almost 800 cell therapy infusions, we found a group of 25 patients that developed second cancers. And among those 25 patients, we really focused on a single patient – one patient with a T-cell lymphoma who unfortunately succumbed to that T-cell lymphoma within 60 days of her treatment for a B-cell lymphoma.
So Mark led this study and studied [the patient’s] case at a very broad and deep genetic and genomic level including by the Tapestri Mission Bio platform to look for evidence of how the CAR T-cells may have participated in the formation of this secondary T-cell malignancy.
Mark Hamilton, MD, PhD: I was fortunate to be joined by my colleagues, Takeshi Sugio and Troy Noordenbos, under the lead of Dr. Alizadeh and Dr. Miklos for this study. As Dr. Alizadeh said, the major question was the frequency of these tumors, especially secondary hematologic tumors, and how they come about.
The major tumor studied was a T-cell lymphoma with the concern being that CAR T-cell vectors could cause an insertional mutagenesis and lead directly to a T-cell lymphoma forming.
So it was very important for us to understand the origin of this specific tumor, how it came about, and why we were seeing it in this patient, and whether we were finding the CAR T-cell vector embedded within the tumor.
Todd Druley: Thank you for that overview. Doctor Alizadeh mentioned that the FDA has labeled CAR T-cells with this particular warning of secondary malignancies. Just in more broad terms, do you feel like this has set the CAR T-cell therapy industry back? Has this been something patients have been concerned about? I think it will lead into the conclusions of your study, and whether or not the concerns are founded, or maybe a little bit over presumptuous. I’m just curious of your opinions on that.
Ash Alizadeh: Because it got so much publicity and press in being an FDA announcement and now a “box label” for these therapies, it’s often the topic of a conversation with patients, and even referring providers, and to educate about the magnitude of the risk.
Our study and several others now kind of give a better sense of the magnitude of second cancer risk overall after CAR T-cell therapy. With most of the studies, capturing that risk to be relatively small, and comparable to risks in the context of other therapies like autologous stem cell transplantation as the historic standard. For in the same setting for many of these patients, and so given the cure their curative potential, we generally still substantially find more evidence for benefit as opposed to risk and advise the patients to get cell therapies in the right indication.
That’s not to say that there isn’t risk. There is risk and that risk seems to manifest in several different ways. From our study, it looks like the CAR T-cells don’t have to directly cause the T-cell lymphoma. They can do so indirectly. But some other studies have shown the effect to be direct and for distinguishing between these one key instrument that was critical for us was cell-level, gene-level, vector-level analysis with Tapestri.
Todd Druley: That leads us to the next question. So, Dr. Hamilton, you obviously use the entire portfolio of different technologies to analyze these cases, and this one particular index case specifically. How did you and the team come up with this portfolio of technologies? And how did that guide your conclusions?
Mark Hamilton: Stanford and Dr. Miklos’ and Dr. Alizadeh’s lab are very fortunate to have a broad array of technologies that we’ve been working for several years now to implement in the CAR T-cell space, and really also just focused on the question of kind of secondary cytopenias after CAR T-cells, and so we were fortunate to have a large portfolio of technologies available to us, and it was important to us to show, to ask the question how we should look for these secondary tumors, and how they should be investigated in the broadest sense possible. And we leveraged multiple methods in this one patient to perform a very deep analysis.
So we ultimately looked on the RNA, protein, and DNA level for the axi-cel vector using multiple techniques, including: capture-based sequencing, which is a DNA level assay that can tile across the axi-cel vector; simple qPCR, which is also a DNA based assay; single cell RNA sequencing, which allows us to have single cell resolution to directly identify the tumor in a population of cells; flow cytometry, which is a protein based assay, allowing us to look for expression of the axi-cel protein a more single cell resolution as well, but with lower fidelity than some of the more technically complex single cell assays. And then the Mission Bio [Tapestri] instrument which allows us to distinguish the tumor based on cell surface markers. And again, as a DNA based assay tiled across the array.
We were able to take all of these technologies together and see that at the protein, RNA, and DNA level using multiple techniques that would be very unlikely to miss any fragmented insertion of this vector or cryptic insertion. We did not see axi-cel in this T-cell lymphoma at all.
This provided both clear evidence that axi-cell was not directly contributing by vector-mediated mutagenesis, but also provided a blueprint of strategies and how you could look for vectors in using multiple techniques and future tumors should they arise. I think the second important thing is, these technologies don’t only capture the axi-cel vector but also many other genes.
And so, using these capture based sequencing technologies initially, we were able to distinguish the original large cell lymphoma from the T-cell lymphoma showing definitively that there were 2 different cancers rather than one cancer that had undergone lineage switching.
We were able to show that there were 2 originating, or several originating mutations and genes called DNMT3A and TET2 that appear to be present in both tumors.
And then, importantly, using the Tapestri analysis, we could actually locate these clonal hematopoiesis mutations that tend to exist in patients that are older or have pre-existing DNA damage. We could show that those were likely existing in the hematopoietic stem cells based on the fact that they were existent throughout multiple lineages including myeloid and T-cells in this patient, and we could show that these were directly contained within the tumor suggesting an underlying susceptibility for this patient to both tumors. And so [the patient] was just likely very susceptible to developing these malignancies, unfortunately.
Todd Druley: Thank you. Is it accurate for me to say that you would conclude that the clonal hematopoietic mutations contributed to the transformation of the T-cell lymphoma? Is that an accurate statement?
Mark Hamilton: Yes, I believe that that’s accurate. They were likely an underlying susceptibility that contributed and the tumor is clearly very clearly derived from a set of these mutations really 3, 2, and the gene a gene called TET2 and one in a gene called DNMT3A
Todd Druley: And both of this patient’s tumors were Epstein–Barr virus (EBV) positive. Is there any speculation as to the role of EBV in the transformation of the T-cell lymphoma?
Mark Hamilton: Yeah, so that’s accurate. And notably this patient had underlying autoimmune disease and patients with underlying immune suppression, which she had been on for many years, can be susceptible to EBV activation, causing secondary lymphomas, including both B- and T-cell. Her B-cell lymphoma was also EBV-positive and so this viral reactivation, probably within these already mutated cells led to a diffuse large B-cell lymphoma.
Notably there is a known immune suppression associated with CAR T-cells. The major cause of non-relapse mortality in our CAR T-cell patients is not these secondary tumors, it is infections. So it’s possible that the CAR T-cells contributed to an immune suppressive environment, or the lympho-depleting chemotherapy, or just all of her chemotherapies that she’d had prior to this. They led to EBV reactivation inside the T-cells, and the same kind of mutated subset of cells that led to proliferation, overgrowth and subsequent additional mutation in this T-cell lymphoma. And so we believe EBV was active in both tumors and was a cause of this happening.
Ash Alizadeh: I guess the big picture regarding the original question, it’s when you don’t find something, it’s difficult to be definitive, because absence of evidence and evidence of absence are not totally the same thing. So the way we approached that conundrum was to use a whole suite of methods to be sure that we’re as broad and deep as possible. So as Mark mentioned, first starting at the protein level is the vector expressed on the T-cells. Second, is the RNA seen, when we have RNA data, third is there DNA level evidence. And to do that, both at bulk, level and single-cell level with sensitive methods and then also in cell free nucleic acids. The collection of evidence is pretty strong to show that there is no direct role for this vector, in integrating into this tumor. Unlike the other tumors, where the vector has been found.
And then, with regards to “what made this tumor tick?” In the absence of individual things that we could perturb and mess with, it’s very hard to be definitive about it, but the circumstantial evidence really points to the things Mark was talking about – the [clonal hematopoiesis of indeterminate potential] CHIP mutations, the shared EBV infection by both tumors, the shared CHIP mutations and the immunosuppression both in the B-cell lymphoma, and then immunosuppression in the context of the CAR T therapy for the T-cell lymphoma.
Exactly which of those 3 things is the most important? What’s the stepwise order of events to make a T-cell tumor? That’s much harder to answer, but I think, as Mark said, this study sets up a blueprint for how you would use technologies like this to study future patients. We should learn from every single such patient, as rare as a phenomenon as this is, to read the tea leaves of what happened to avoid this risk for future patients– to intercept such lesions at the earliest time point that we can so we don’t have lethal outcomes like we did for this patient.
Todd Druley: Is there a more systematic way to either screen or do surveillance for patients like this that would take some of these things into account? Perhaps at some point there’s only so much you can do, but I’d still like to know if these things are evolving in the background. Do you think single cell technology specifically can have a role in that sort of pre-treatment assessment, or surveillance, or perhaps both?
Ash Alizadeh: For us, this patient, and future patients where we know about certain features that we identified here, and it seems like are being found in the other studies of these T-cell cancers – viral infection, clonal hematopoiesis, and immunosuppression. We knew about these beforehand. For a future patient of mine that comes to the clinic who has these shared features? It’s going to make my antennas go up and think about this risk now. I don’t think I’m going to ever suggest that we not give CAR T therapies when those risks are there but to be hyper vigilant about the risk.
Now at Stanford we routinely evaluate for CAR dynamics as part of the research studies that we’re doing, but that’s not routinely being done. Those involve flow cytometry based characterizations of cell expansion through the course of therapy. If the CAR T-cells were to persist and expand in ways that we didn’t expect that might make you worried about a CAR T driven cancer that’s caused by the CAR T-cell.
But what about cases like our case here? I think the things that we saw here were the dramatic expansion of [clonal hematopoiesis] CH in the liquid biopsy studies, the dramatic expansion of EBV. But these happen so quickly that it’s hard for me to imagine a workflow today, even with the cutting edge tools that we have of routine monitoring with real turnaround times to leverage Tapestri or other methods for routine use.
I guess you know more rapid clinical grade assays that are executed as lab developed tests and clear labs and things like that could be very valuable for CAR dynamics and CH dynamics but they’re only emerging now.
Mark Hamilton: And both us and others have also kind of looked at some of these pre-existing, clonal hematopoiesis mutations, and they can be quite common. I think an important question will be, which ones tend to become malignant or cause problems. But it may be infeasible to track that many mutations, if they are as common as some of these studies are showing. And you know importantly, in a very aggressive cancer, like large cell lymphoma the risk of a secondary tumor is very low relative to the risk of the aggressive lymphoma progressing especially using a curative therapy.
It is important to note that as CAR T-cells come out into newer areas with non malignant diseases, including autoimmunity, being one that I think many people are excited about, some of these risk/benefit ratio may rebalance some and you may take into consideration these pre-existing susceptibilities if you have other available therapies that may not perturb as much the immune environment, or cause some of these risks. But I don’t think that we really know those patient populations well enough. There’s not enough patients yet to study it thoroughly, but it will be an important topic moving forward that will be informed, of course, by our studies and aggressive diseases, where we have much more experience.
Todd Druley: Excellent. So I’ll just wrap up and my final question would be beyond Stanford. What do you think could be the impact of this particular study on CAR T development and treatment and surveillance in general across the broader community. Think there’s a message there for other centers as well?
Mark Hamilton: You know, we’re discussing with other centers now, and it’s important that our study is just a single patient. Though it was a very large cohort of patients, and it gives us some idea of the rarity of this. I think that other centers are also quite attuned to these risks and are looking in their own patient populations.
There was a companion paper out of Dana Farber that came along with our study showing an indolent T-cell lymphoma that was cell to cell positive published at the same time. And so I think that our study informs this conversation, and certainly, you know, we would engage with the greater community regarding these cancers, and how to go about studying them on a multi- institutional or collaborative level so that we can really understand these risks better as a scientific whole rather than just a single institution.
Ash Alizadeh: I think that’s right. I think trying to apply the same set of tools that we described here to other cases to understand them, I think, would be quite valuable. As a practical consideration, I think some of the things we did here may not be equally feasible at other centers if they did not viably preserve second cancers, it’s hard to do [single-cell sequencing with] Tapestri on them, right?
So we would encourage our colleagues in the community to really work diligently to build the research protocols and consent and bank specimens from unfortunate tumors that arise, secondary tumors, so that they could be studied with high precision using the tools that we and others I’m sure will be talking about. We didn’t find the CAR T-cells in these tumors. But for those that do, where did the CAR T-cells integrate? What are the nearby genes? Do allogeneic products need to be vetted for not having had such integrations? Do autologous products need to be screened for such integrations before there? These are all open questions over time as we learn more to refine cell therapies, and I’m sure all the companies are thinking about these questions.
Todd Druley: Absolutely. Well, that’s all my questions for today, unless either of you gentlemen have anything else to add. I wanted to thank you for your time. Thank you for your work, and thank you for continuing to pioneer the field of cellular therapies.
Ash Alizadeh: Thank you. Also for collaborating in really helping us analyze these data with your tools and helping us interpret the results.
Todd Druley: Pleasure to work with you and your team.
