Breakthrough Leukemia Treatment Backfires in a Rare Case

The groundbreaking treatment that genetically engineers a patient’s own cells to fight leukemia turned lethal in one patient, reversing his remission.

Comments: 18

  1. What are the statistics from Novartis's registry to monitor the over 400 patients who received Kymriah therapy?

  2. The Nature Medicine paper gives a pretty good explanation for why the researchers think that one leukemic cell survived. The primary reason is that the CAR gene that is inserted in the process was mistakenly inserted into a leukemic B cell instead of a cancer-fighting T cell. In a T cell, the CAR activates the cell to fight cancer. In the B cell, it likely doesn't do anything because the CAR wasn't designed to work in a B cell nor is a B cell naturally designed to fight cancer cells. What the CAR does do in a B cell is bind the CD19 molecules on the surface and mask it from detection by the cancer-fighting T cells which are also looking for CD19 to bind. In essence, the leukemic B cell became invisible to the immune system. Couple that with the fact that the now-invisible B cell likely went through an unintended selection process during the external in vitro cell culturing and outgrowth (cell culture conditions are likely much more harsh than in vivo conditions inside the body) you get an invisible cancerous B cell that is much more hardy and proliferative (not to mention possibly chemo-resistant due to earlier rounds of treatment). That's a recipe for an aggressive relapse and that's exactly what happened unfortunately. It's a series of unfortunate and unpredictable events.

  3. @SBR, I’m not a scientist but it appears you are. Is this ‘one rogue cell’ incident akin to ‘residual disease’, referring to the number of cancer cells remaining after traditional chemo/radiation/biologic treatment, w/wo ASCP or donor transplant? I’ve wondered if ‘residual disease’ cancer cells could be removed in vitro by CAR-T from stem cells collected via apheresis and BEFORE a patient’s ASCT (essentially, ‘sterilize’ stemcells of cancer cells before infusion). Also, because CAR-T treatment is only indicated in patients who have either already received traditional therapy and have relapsed, or are refractory, their residual disease cells have often been rendered resistant to further traditional treatment already. And, apparently, CAR cells aren’t affected by this resistance. Is this possible because traditional treatment and CAR-T involve different mechanisms of action? Finally, where might R&D go now to guarantee that only T-cells are engineered ex vitro then reintroduced in vivo, not B-cells?

  4. @Teresa Pierce very good questions! In many ways, this "one rogue cell" is very similar to residual disease in the sense that both are remnants of the tumor that resist treatment and expand after the cull. However, in this instance, the rogue cell attained its resistance as a direct result of the therapy protocol whereas in traditional "residual disease" the remnants already contain the resistance trait or gain resistant traits as a result of selective pressure due to the treatment. The latter is a biological process whereas the former is a technical flaw of the protocol which can be mitigated. My understanding of the stem cell collection protocols for autologous SCT is that unless the cancer cells themselves are stem cells (in a very undifferentiated state), the apheresis and subsequent cell sorting procedures are very good at removing the cancer cells from the stem cells. In other words, stem cells and cancer cells are in most cases different enough in their surface markers that they are effectively separated before an SCT.

  5. @ Teresa Pierce: as for your question about traditional treatments vs. CAR-T, yes. Traditional treatment regimes are often cytotoxic chemotherapeutics that target fast growing cells. Recent advances have led to antibody/biologic drugs that mark cancer cells as targets to help the immune system find and kill them in a more targeted and less generally toxic fashion (rituximab is a good example of this for B cell malignancies). These new drugs have obvious advantages compared with traditional chemotherapy and can often synergize with traditional chemotherapy to deliver more durable remission. However, the new drugs that help the immune system are not without drawbacks either. One of the big challenges is that many cancer patients do not have robust immune systems to begin with either due to disease burden or age. Immune systems that have fought cancer for a long time can also become "exhausted" (because remember that a prolonged immune response such as chronic inflammation is often bad in many ways). CAR-T treatment aims to get around this problem by directly taking the patient's immune T cells, reprogramming them to specifically target the cancer, and rejuvenating them to be more aggressive and less "exhausted" before reinfusing them back into the patient. These immune "super-soldiers" then eradicate the cancer. So in many ways, CAR-T is as radical of a treatment compared to previous generations of biologics such as rituximab as rituximab was to traditional chemotherapy.

  6. The breakthrough treatment didn't really backfire; it had an unintended and unanticipated consequence, which has deepened understanding of the process one more step. @ SBR, than you for summarizing the research.

  7. Every process ought to use root cause failure analysis. The up front cost is usually recouped in superior performance over time. For 475000 per treatment, one would expect no less.

  8. SBR, Feel free to post here any time! NY Times: Do yourself a favor, and offer SBR an Immunologic and Biological Sciences writing job...

  9. To further clarify the commenter SBR's excellent explanation: CD19 is a molecule that is produced by leukemia cells and displayed on its surface. A "CAR" (Chimeric Antigen Receptor) is an engineered molecule designed to trick T cells into attacking and killing cells displaying CD19, by hijacking its normal function. The way this is accomplished is as follows: Every normal T cell has a receptor (TCR) displayed on its surface that can recognize and bind to a target (for example, a pathogen). Part of the TCR is inside the T cell (connected by a "transmembrane" piece), and when the binding to a pathogen occurs, the inside part become activated and tells the T cell to attack and kill the pathogen. A "CAR" is a TCR-like molecule in which the target recognition piece of the TCR is replaced with an antibody that specifically binds to CD19. This molecule is then inserted into the patient's T cells, and these cells now making the CAR think that any cell displaying CD19 must be attacked. This can be very effective, but what if the CAR inadvertently gets into a leukemia cell? Then, the CD19-binding CAR is displayed on the leukemia cell's surface, but that's where CD19 is! So those CARs bind to the CD19, and block access of the CD19-recognizing CARs on the T cells. These leukemia cells therefore become invisible to the CAR-T cells, so, unfortunately and ironically, the therapeutic molecule actually defeats itself.

  10. @Steve Thank you - that was well explained.

  11. Could the University of Pennsylvania have tried to use a redesigned targeted form of CAR T-Cell immunotherapy to attack their mistakenly created immune cancer B-cells? Does the institution have any responsibility for correcting their mistakes? What are the patent rights of University of Pennsylvania and Novartis for a therapy that other labs could duplicate? Kymriah is not a drug, but a lab technique using the patient's own cells. Why allow them to charge $475,000 for a lab process that is based upon government funded research that is currently available to many other research labs? It's a new medical business model worthy of further ethical and economic debate.

  12. @Brighteyed To answer your questions: 1. Could the doctors have redesigned the CAR-T therapy? No not in any timely fashion. While CAR-T itself is really a class of therapy modality (akin to saying "antibiotics" or "NSAIDS"), the key component is the "CAR" (Chimeric Antigen Receptor) portion. I will refer you to Steve's excellent explanation below of what a CAR is. But the point is that each CAR-T therapy is based upon the target molecule and the CAR that's been developed to recognize it (kind of like a lock and key system). To "redesign" the CAR-T therapy would essentially mean to find a whole new target that's specific and safe enough to target as well as designing a whole new CAR that is specific enough for that particular target. Needless to say, that process is essentially 99% of the R&D for any new CAR-T therapy and simultaneously highly expensive, incredibly time consuming, and chock full of regulatory hurdles. This is also one of the reasons why we don't current have CAR-T therapies against every form of cancer known to medicine. 2. As to whether an institution has any responsibility to correct their mistakes. I won't get into an ethical debate. But I will point out that there are currently childhood cancers for which the standard chemotherapy treatment measurably increases risk of secondary malignancy later in life. Also, CAR-T is generally a treatment of last resort at the moment. For many patients on the trial, it literally may be either CAR-T or hospice care.

  13. Also, many cancer therapies have high failure rates. No therapy has 100% efficacy.

  14. "The treatment costs $475,000." The science is interesting, but this is horrifying.

  15. @The Poet McTeagle Yes, the price is horrifying. But I cannot begin to explain how difficult this procedure is to execute successfully. And If I had a half million dollars and spending it would cure my cancer, I'd spend it. Just having this option is equally as amazing as the price is horrifying. BTW, I just had an accident which then required emergency spine surgery to stabilize me (it worked). The total cost: $250,000 --all medical costs in America are astronomical!

  16. How much is a life worth? And a complete cure?

  17. Great questions that society has yet to grapple with.

  18. With the high price of this treatment, has this been approved in the EU? And if so, how much does it cost there?