If at first you don’t succeed: Testing outcomes of second CAR T-cell infusions

From the Turtle group and collaborators, Clinical Research Division

Chimeric antigen receptor-engineered T cell (CAR T cell) therapy is a strategy for re-targeting patients’ own immune cells against their cancers. CAR T cells equipped with chimeric receptors that recognize the CD19 antigen, widely expressed across B cell malignancies, achieve high response rates in patients with acute lymphoblastic leukemia (ALL), non-Hodgkin lymphoma (NHL), and chronic lymphocytic leukemia (CLL) that have previously resisted or relapsed against other therapies. However, some patients do not enjoy lasting benefits. Faced with very few alternative treatment options at this stage, one possible strategy is to perform a second infusion of CAR T cells. Dr. Jordan Gauthier, assistant professor and member of the Turtle group in the Fred Hutch Clinical Research Division, and colleagues investigated the feasibility and efficacy of second administrations of CAR T-cell therapy to improve outcomes in ALL, NHL, and CLL patients. Their findings were recently published in Blood.

Reasons for failure of initial CAR T cell infusions (CART1) to provide durable responses are not completely understood and could potentially have multiple underlying mechanisms. For example, T cells harvested from patients for genetic modification may not be highly functional due to various patient or disease-related factors, cancers may employ immunosuppressive mechanisms against engineered CAR T cells, tumor cells may evade CAR T-cell responses by losing expression of the CD19 target antigen, and/or a patient’s own immune system may reject the therapeutic CAR T cells due to recognition of non-self portions of the engineered chimeric receptor. Several of these causes for initial treatment failure could portend low likelihoods of success for second infusions of CAR T cells (CART2). Furthermore, CAR T cell therapy itself carries risks of adverse events, including severe treatment-related toxicities. Thus, the authors  set out to test whether CART2 would be tolerated and improve outcomes for patients with relapsed and refractory B cell malignancies. Here, the Turtle group report that second doses of CAR T cells achieved complete responses in approximately 1 in 5 patients, with low incidences of severe toxicities, in ALL, NHL, and CLL patients. Further, they shed light on several factors associated with beneficial responses to CART2, providing important guidance for the design of future CAR T cell trials.

Higher rates of durable responses were observed in NHL and CLL after an increased CART2 dose and CyFlu pre-CART1
Image Provided by Dr. Gauthier

This study followed the outcomes after CART2 of 44 patients with heavily-pretreated ALL, NHL, and CLL cancers. Thirty-four percent  of patients showed no response to CART1, 50% experienced initial complete or partial responses to CART1 but later relapsed or progressed, and 16% experienced partial responses that were deemed clinically insufficient after CART1. Prior to CART2, expression of the CD19 target antigen remained at normal levels on tumor cells in 28 patients, decreased in 3 patients, and was not evaluated in 13 patients. Patient groups also differed based on conditioning lymphodepletion regimens (which aid the engraftment of therapeutic T cells) and relative CAR T cell dose sizes given in each infusion (some patients received the same numbers of CAR T cells for both infusions, while others were given ten times more cells in CART2 than CART1).

Despite often increased numbers of cells administered in CART2, the authors observed comparable adverse event rates between the first and second doses of CAR T cells, including similar rates of cytokine release syndrome (CRS; 7 and 9%, respectively), a collection of symptoms associated with increased levels of inflammation in the body, and neurotoxicity (14 and 11%, respectively). All of these events were reversible, sometimes with mitigating treatments, and no therapy-related deaths were observed. Encouragingly, complete responses (CR) were achieved in 21% of ALL patients, 19% of NHL patients, and 21% of CLL patients . Overall response rates (ORR) were 21%, 52%, and 33%, respectively. The median durations of response (DOR) were 4 months in ALL, 6 months in NHL, and 33 months in CLL. Amongst responders, 2-year overall survival (OS) was 36% compared to 19% in non-responders. Thus, second infusions of CAR T-cell therapies can benefit significant fractions of patients with relapsed or refractory B cell malignancies, without increased risks associated with adverse events.

To begin to investigate mechanisms behind outcomes in this study, Gauthier et al. collected samples from patients for more in-depth analyses. By flow cytometry, they were able to detect CAR T cells in the blood of 42/44 patients after CART2 (compared to only 19/38 of patients with quantitative PCR [qPCR]-detectable CAR T cells prior to CART2). The magnitude of CAR T cell expansion after infusion was strongly associated with clinical responses, and CAR T cells persisted longer in patients exhibiting CR and partial responses (PR) compared to those with stable or progressive disease (SD/PD). In patients from whom samples could be collected at the time of PD, 1/20 showed loss of CD19 target antigen expression by the tumor cells. These results suggest that insufficient CAR T cell expansion and persistence may be limiting factors for therapeutic efficacy, and loss of antigen expression may be a mechanism of tumor escape in some patients.

Finally, multivariable regression analysis was performed to uncover factors associated with outcomes of therapy. From this analysis, the conditioning lymphodepletion regimen emerged as a highly relevant variable in this study. Patients who received cyclophosphamide with fludarabine (Cy-Flu) to deplete lymphoid cells prior to CART1 were more likely to exhibit enhanced CAR T cell expansion and persistence after CART2, and more likely to show therapeutic responses after CART2 infusion.They also had a lower risk of disease progression or death after CART2 than patients who received cyclophosphamide-based lymphodepletion regimens not including fludarabine (non-Cy-Flu).  The Turtle group  also reported correlations between receiving higher CART2 dose and persistence, and better patient outcomes. Pursuing this further, they found that NHL and CLL patients who received increased CART2 doses and Cy-Flu lymphodepletion had significantly more CR/PR and longer progression-free survival (PFS) than those receiving equivalent doses of CART1 and CART2 or who received non-Cy-Flu lymphodepletion.

Interestingly, patients receiving Cy-Flu, and patients receiving higher CART2 doses amongst those receiving Cy-Flu, were less likely to exhibit signs of immune rejection of CAR T cells, supporting the hypothesis that these protocols may help mitigate anti-CAR immune rejection. The lead author commented, “Our data suggest anti-CAR immune responses, directed at murine sequences located in the single chain variable fragment (scFv) of the CAR, could be responsible for the decreased efficacy of second CAR T cell infusions. We are currently investigating the use of CAR T cells engineered with a fully human scFv in a phase I/II clinical trial, with the aim of mitigating anti-CAR immune responses, and increasing efficacy after prior CAR T cell failure.” He added, “Further studies are needed to clarify the biological effects of lymphodepletion and increased cell dose on immune rejection and on the tumor microenvironment”.

These findings shed light on important variables for the success of repetitive CAR T cell infusions on outcomes for patients for B cell leukemias. “Our work is the first to identify actionable factors associated with the outcomes of second CD19 CAR T cell infusions for patients with relapsed or refractory B cell malignancies”, said Dr. Gauthier.  “Our data will inform the design of future CAR T-cell trials in patients who have previously failed CAR T-cell immunotherapy.”


This work was supported by the National Institutes of Health, National Cancer Institute, National Institute of Diabetes and Digestive and Kidney Diseases, and National Heart, Lung, and Blood Institute; the Life Science Discovery Fund; the Bezos family; and Juno Therapeutics.

UW/Fred Hutch Cancer Consortium members Merav Bar (Fred Hutch), Ryan Cassaday (UW), Aude Chapuis (Fred Hutch), Andrew Cowan (UW), Damian Green (Fred Hutch), Hans-Peter Kiem (Fred Hutch), Filippo Milano (Fred Hutch), Mazyar Shadman (Fred Hutch), Brian Till (Fred Hutch), Stanley Riddell (Fred Hutch), David Maloney (Fred Hutch), and Cameron Turtle (Fred Hutch) contributed to this work.

Gauthier J, Bezerra ED, Hirayama AV, Fiorenza S, Sheih A, Chou CK, Kimble EL, Pender BS, Hawkins RM, Vakil A, Phi TD, Steinmetz RN, Jamieson AW, Bar M, Cassaday R, Chapuis A, Cowan AJ, Green DJ, Kiem HP, Milano F, Shadman M, Till B, Riddell SR, Maloney DG, Turtle CJ. Factors associated with outcomes after a second CD19-targeted CAR T-cell infusion for refractory B cell malignancies. Blood. 2020 Sep 23:blood.2020006770. doi: 10.1182/blood.2020006770. Epub ahead of print.