Krishna V. Komanduri, MD: I want to move on, and we kind of alluded to this, in fact, David, you just pointed it out, that this should really obviously be done in specialized centers. And one of the reasons that this really needs to be done in specialized centers is that although often manageable, there are some very serious toxicities that can emerge with all of these therapies. Some that are unique to individual targets, but some that are across T cell immunotherapies in general. So, maybe Michael, if you could try to talk to us about some of those toxicities that are common across these platforms.

Michael Pulsipher, MD: Sure. Let me talk about 4 different specific toxicities. First is the toxicity that most people have heard about with CARs and that is cytokine release syndrome. When T cells encounter the antigen on the tumor with the CAR construct in place, they multiply incredibly. They go up sometimes more than 5 logs and have release of cytokines as they’re rapidly expanding because of the engagement of these co-stimulants for molecules that David talked about earlier.

The body is bathed in cytokine. It’s acting like it has sepsis. And because of that, patients can have extremely high fevers. They can become hypotensive, they have significant capillary leak. So, we actually have to manage their hypotension a little bit differently than we usually do, flooding them less with fluid and giving them more inotropes to support their blood pressure. They may need to be intubated. They may need to be dialyzed, but it’s important to keep in mind that most of them don’t get that sick. That’s a small fraction, somewhere between 15% and 25% who may end up in the ICU very sick. But the vast majority of them will have fevers, and we, in fact, want them to have fevers in a way. We know that patients who have cytokine release syndrome have a better chance of having a good response and having a long-term expression of their CARs.

The second real T toxicity to mention is neurotoxicity. We know that CAR T cells, as all T cells do, migrate into the CNS. And they have activity that leads to neurologic changes. Some patients have been known to have seizures. Confusion is very common. Aphasia, which is somewhat disturbing, is where the patients can stop talking for a few days, which, of course, bothers their family members quite a bit. Mania. Patients becoming agitated and upset also occasionally occurs. Most neurotoxicity is reversible and patients go back to normal, but it does take several days. And we’re just beginning to understand why neurotoxicity can occur.

The third thing to mention is, especially for CD19- and CD22-targeted therapies, is B-cell aplasia. If you’re targeting the B-cell antigens on the cancer, you’re also destroying the B cells. And in situations where we have long-term CAR persistence, they’re no longer expressing B cells and, therefore, not producing immune globulin. And this is very important. Many begin infusion of immune globulin, and that’s standard in pediatric practice to actually put patients on immune globulin to avoid infections afterwards. And this can go on for years.

One final very important toxicity to mention is coagulopathy, where patients can have a very rapid drop in fibrinogen. They can have coagulopathy that can lead to bleeding complications. So, we need to watch that carefully. They look like they’re having hemophagocytosis and HLH-like state. So, they get very sick and we need to pay close attention. Fortunately, most of these toxicities can be reversed with IL-6 inhibitors, steroids, and other specialized treatments. The centers who do CAR T cells understand how to administer them.

Stephen J. Schuster, MD: I’m going to add to that with regard to the A- and B-cell aplasia and hypogammaglobulinemia. I think the pediatric experience is different. We did not routinely put our patients, who weren’t already on immunoglobulin, on immunoglobulin and we saw that even in patients who were in durable remissions for years, they begin to recover B cells while staying in remission at about 6 months. And by 12 months, 1 of the 3 immunoglobulin classes was generally increasing and even to normal. And so, over time, it seems that the adults can recover their immune system. And there also seems to be a population of long-lived plasma cells that can carry these adults through this period of hypogammaglobulinemia.

Michael Pulsipher, MD: Yes, I think that’s going to vary by which CAR you use and which patient you treat.

Stephen J. Schuster, MD: Yes, age and status of the immune system.

Michael Pulsipher, MD: Yes.

David Maloney, MD, PhD: I’d just like to add about the neurotoxicity. This is really one of the most vexing issues the field is facing right not and it’s really poorly understood. We have a recent paper where we postulate that it’s actually due to endothelial cell damage and that this leads to the breakdown of the blood brain barrier. It’s actually cytokines causing the problem, it’s not the actual T cells and the CNS. We’ve actually treated quite a few people with CNS-positive malignancies and they don’t get worse neurotoxicity even if they’re clearing their CNS malignancy. So, we think it’s probably cytokines, but it’s the most frustrating thing to treat as well because none of our treatments appear to have anything to do with reversing it. It seems to reverse on its own in almost all cases, but we really don’t know the right treatment and we need breakthroughs in that area.

Michael Pulsipher, MD: Well, one thing to add to that, David, that must be said and it’s very important, CAR T cells are very effective at treating a disease in the CNS, especially for leukemia lymphomas, because it does cross and have activity, and patients with significant CNS disease have gone into good remission. So, we want it to be active in the CNS.

Krishna V. Komanduri, MD: Let me point out a couple of things. One is it’s interesting that there’s really at least no clear evidence of CD19 expression in the brain and yet we have these toxicities. There are some very interesting, for example, primate models by some of your colleagues in Seattle that are starting to elucidate these mechanisms.

The other thing I would point out is that there are some biomarkers that seem to be associated with the onset of these things, but we can’t reliably yet predict who will and won’t develop these toxicities. I know I’ve seen patients who are actually almost in coma-like states that were fully reversible. That adds, Michael, as you pointed out earlier, a very terrifying effect and often we, along with family members with some of these early patients, are really learning about this first hand, and these are not trivial toxicities.

David Maloney, MD, PhD: And for the clinicians, I think it’s important that the timing is a bit different in these things. Usually, not universally, CRS, the cytokine release syndrome, happens 2 to 3 days before neurotoxicity, and then neurotoxicity can happen thereafter. They can be concurrent and there are cases where no CRS and neurotoxicity. So, you have to be alert for all of these.

Stephen J. Schuster, MD: On another practical note, if you’re treating people as an outpatient with CAR therapy, there needs to be a partner with them that understands the potential complications and what to bring to medical attention. We do treat some of these people as outpatients but only if there’s an educated partner that stays with them around the clock.

Michael Pulsipher, MD: And one important point to make about CNS toxicity, the one thing we do that has been associated with it is the higher the level of CRS that a patient has, the more likely they are to have CNS toxicities. With grade 4 CRS, we have seen these toxicities in about 65% of patients, whereas grade 1 CRS, maybe only 25% of patients will have CNS toxicities.

Transcript Edited for Clarity