CAR T Cell Therapy for Acute Lymphoblastic Leukemia Recommended by FDA
Because CAR T cells interact directly with cancer cell surface antigens, they appear to be limited to the risky use of self-antigens or to rare universal neo-antigens
A US Food and Drug Administration (FDA) panel recently recommended the approval of CTL019 chimeric antigen receptor (CAR) T cell therapy for acute lymphoblastic leukemia (ALL). This first-ever recommendation for a CAR T therapy was based on evidence that the treatment's clinical benefits outweighed considerable safety concerns.1,2 The population eligible for this therapy is restricted to children through young adults with relapsed or refractory B cell ALL, among whom the survival rate can be less than 30%.3
CAR T cells have engineered receptors comprised of a cell-surface component that binds to a tumor antigen and is attached to one or more internal signal initiators. CTL019 CAR T cells are designed to identify ALL B cells by the CD19 cell surface protein also expressed on normal B cells.
By design, CAR T cells circumvent major histocompatibility complex (MHC) presentation and thus the pathways to this mechanism of immunological activation. The genetically modified receptor aboard these cytotoxic T cells binds directly with a tumor antigen to initiate cell killing. This approach necessitates CAR T cell interaction with a specific surface molecule, which will often be a self-antigen found on normal and cancer cells.
In childhood ALL that does not respond to standard therapy, the risk of CAR T killing normal cells must be weighed against the probability of death from the cancer itself. For ALL an alternative therapy is life-long immune suppression following hematopoietic stem cell transplant (HSCT). The potential for destruction of normal B cell populations by CAR T cells may therefore be considered an acceptable risk.
It is, however, important to recognize that years of experience titrating immune suppressive drugs to prevent graft-versus-host disease after HSCT may minimize the risk to the surviving patient, whereas, at this time, CAR T cells cannot be withdrawn, and deleterious effects of B cell depletion may differ from patient to patient.
CAR T cells are sometimes referred to as a “living drug” because their action directs a therapeutic molecule against a cellular target. From an evolutionary perspective this term is apt because, like an antimicrobial drug favoring the evolution of antimicrobial resistance, CAR T cells can favor evolution of resistance through selective pressure on an antigen. This evolutionary effect has already been noted in response to CTL019: subpopulations of tumor cells altered CD19 have emerged during treatment, resulting in patient relapse.4
There are additional concerns regarding CTL019 CAR T cell therapy in particular and engineered T cell therapies in general. In the relatively few CAR T trials, there have been numerous deaths directly attributable to the therapy. In one CTL019 CAR T cell study, furthermore, severe to life-threatening cases of cytokine release syndrome occurred in almost half of the patients.5 Other adverse events can include lethal T cell cross-reactivity and cerebral edema.
Despite early positive results, a phase 2 trial using an alternative CD19-targeted CAR T cell therapy was discontinued after multiple deaths. Trade secret protection restricts public access to information that might clarify why the trial was unsuccessful.6
CAR T cells are representative of the engineered arm of adoptive cell therapy (ACT). The other arm, which uses natural variation in T cell receptors, also requires that T cells be obtained from the patient, but then selects and expands the subpopulations of T cells that interact with tumor antigens ex vivo. These cells are then returned to the patient as in CAR T therapy.7 This non-engineered arm uses the versatility of the immune system, which has been molded by natural selection to kill cells with foreign and abnormal antigens and has a low avidity for self-antigens. The selective amplification of the normal repertoire can produce a T cell population capable of interacting with multiple tumor antigens.
In the patient, these already-existing T cells have, by definition, failed at tumor destruction — a fact that may lend more credence to engineering receptors than warranted, as the failure may not be due to the activity of the receptor but because of a suppressive tumor microenvironment. Checkpoint inhibitors have been used to augment ACT, and careful analyses of these results may help parse out the relative value of ACT approaches.
The promise of CAR T cells is attractive to researchers because of, rather than in spite of, the need to use self-antigens like CD19 given their presence in virtually the entire population. Patients that receive CTL019 therapy will have their normal B cell population depleted and therefore require immunoglobulin therapy. This is not a trivial effect of therapy, as B cells are essential for protection from infection and immunoglobulin replacement therapy is not always well-tolerated.8
Because CAR T cells interact directly with cancer cell surface antigens, they appear to be limited to the risky use of self-antigens or to rare universal neo-antigens. On the other hand, CAR T cells are not subject to problems with isolating or expanding subpopulations of patient-derived T cells capable of tumor-killing, a process that is sometimes unsuccessful and protracted in a clinical context — where time is a scarce resource.
Evolution has had a tremendous head start on generating effective and safe cytotoxic T cells. Efforts to improve engineered T cell components include multi-antigen binding and suicide switches which, while promising, may solve one problem and introduce others. Developing therapies that apply the versatile sophistication of the immune system to fight cancer is likely the best long-term strategy.
The FDA recommendation to move forward with ALL CAR T cell therapy for relapsed or refractory disease appears appropriate. Nonetheless, engineering a self-antigen-targeted T cell, which necessitates immunoglobulin replacement therapy, makes sense only in the context of a likely cancer-related death due to a paucity of alternative therapies. Targeting self-antigens is a potential weakness in CAR designs that might constrain it to a final-effort therapy as opposed to a universal cancer treatment.
Ultimately, we are concerned that enthusiasm for the CAR T cell approach will redirect rather than augment other possibly safer, more effective, evolutionarily informed approaches to cancer therapy.
- Maude SL, Frey N, Shaw PA, et al. Chimeric antigen receptor T cells for sustained remissions in leukemia. N Engl J Med. 2014;371:1507-17. doi: 10.1056/NEJMoa1407222
- Novartis CAR-T cell therapy CTL019 unanimously (10-0) recommended for approval by FDA advisory committee to treat pediatric, young adult r/r B-cell ALL [news release]. Basel, Switzerland: Novartis; July 13, 2017. https://www.novartis.com/news/media-releases/novartis-car-t-cell-therapy-ctl019-unanimously-10-0-recommended-approval-fda. Accessed August 2017.
- Bhojwani D, Pui CH. Relapsed childhood acute lymphoblastic leukaemia. Lancet Oncol. 2013;14(6):e205-17. doi: 10.1016/S1470-2045(12)70580-6
- Ruella M, Maus MV. Catch me if you can: leukemia escape after CD19-directed T cell immunotherapies. Comput Struct Biotechnol J. 2016;14:357-62. doi: 10.1016/j.csbj.2016.09.003
- Fitzgerald JC, Weiss SL, Maude SL, et al. Cytokine release syndrome after chimeric antigen receptor T cell therapy for acute lymphoblastic leukemia. Crit Care Med. 2017;45(2):e124-31. doi: 10.1097/CCM.0000000000002053
- Hey SP, and Kesselheim AS. 2016. The FDA, Juno Therapeutics, and the ethical imperative of transparency. BMJ. 2016;354:i4435. doi: 10.1136/bmj.i4435
- Rosenberg SA, Restifo NP. Adoptive cell transfer as personalized immunotherapy for human cancer. Science. 2015;348(6230):62-8. doi: 10.1126/science.aaa4967
- Cherin P, Marie I, Michallet M, et al. 2016. Management of adverse events in the treatment of patients with immunoglobulin therapy: a review of evidence. Autoimmun Rev. 2016;15(1):71-81. doi: 10.1016/j.autrev.2015.09.002