Background

Chimeric antigen receptor T cell (CART) therapy against blood malignancies exhibits unique pharmacokinetic profiles in vivo. There are reports of CART kinetics against acute or chronic lymphoblastic leukemia (ALL/CLL), but reports on CD19-41BBζ CART against relapsed or refractory non-Hodgkin's lymphoma (r/r NHL) are rare. Here we present a summary of the cellular kinetics of CART cells from 35 patients of aggressive or indolent NHL, highlighting the potential correlations between CART kinetics and clinical outcomes

Methods

Thirty-five consecutive patients with r/r NHL from a CART trial (NCT03196830) were analyzed. CART cells were administrated at doses of 5-10x106/kg in 3 split doses following lymphodepleting chemotherapies of fludarabine and cyclophosphamide. CART kinetic profiles in peripheral blood were determined by CAR transgene copies measured by quantitative PCR at various time points, and correlation analyses between kinetic and clinical factors were performed. Kinetic parameters included peak CAR transgene (Cmax), time to reach Cmax (Tmax), and area under the curve at day 28 (AUC0-28d) and day 70 (AUC0-70d) describing short-term and long-term persistence, respectively. Clinical factors included responses, adverse events, and disease characteristics.

Results

Correlation analysis between kinetic and efficacy profiles revealed that responders exhibited significantly higher AUC0-70d (but not Cmax and AUC0-28d) compared with non-responders, indicating that AUC0-70d rather than Cmax and AUC0-28d was correlated with clinical efficacy. This result was partially inconsistent with previous reports of CD19-BBζ CART in ALL/CLL patients (Mueller KT, et al., 2017), suggesting that CART might exhibit distinct kinetic profiles in NHL compared with ALL/CLL. Additionally, AUC0-70d was 7-fold higher in patients who maintained in remission for > 6 months after prior therapies compared with those who did not, and 278% higher in patients receiving < 3 prior lines of therapies than those > 3 lines, both of which were correlated with better durations of remission (DOR) and/or overall survival (OS). These results indicated that longer remission periods after and fewer lines of prior therapies were favorable factors of long-term persistence of CART cells, which subsequently led to improved clinical efficacy. Again, Cmax or AUC0-28d did not show any differences between groups. No difference in Tmax were achieved between any two groups

Correlation analysis between kinetic and safety profiles revealed that patients with grade 3-4 cytokine release syndrome (CRS) had significantly higher Cmax than those with grade 1-2 CRS, indicating that CART expansion correlated with the severity of CRS. No correlation was observed between neurotoxicity and Cmax. Neither neurotoxicity nor CRS was correlated with Tmax, AUC0-28d, or AUC0-70d.

Lastly, transient drops in CAR copies early after infusions were observed in 78% patients (24/35). A transient drop usually happened around Day 5, divided the whole expansion phase into two peaks, and created a trough value of <10% of the earlier peak. Interestingly, AUC0-70 was significantly higher in patients with transient drops than in those without, while Cmax, Tmax and AUC0-28 were nearly identical. Moreover, 75% patients (18/24) with transient drops were responders, compared with only 36% patients (4/11) without transient drops. Patients with transient drops also exhibited better DOR. This unique kinetic feature tended to be a favorable factor of efficacy, leaving the mechanism to be further explored.

Conclusion

Overall, significantly higher levels of long-term persistence were observed in patients who successfully responded to CART therapy in r/r NHL patients, indicating that the kinetics of CART cells could be used as predictive factors in clinical practice. This is the first summary of CD19-41BBζ CART against r/r NHL emphasizing different kinetic profiles between NHL and ALL/CLL. These results also imply that means to enhance long-term persistence of CART cells may be potential strategies of to improve response rates.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution