CD19 CAR T-cell (CAR-19) therapy have elicited great clinical responses in B-cell malignancies. However, CD19-negative relapse has emerged as a major challenge for the long-term disease control post CAR-19 therapy and conferred a dismal outcome to these pts. Recently, co-targeting of CD19/CD20, CD19/CD22 or CD19/CD123 has been proposed as a strategic approach to overcome antigen loss relapse after CAR-19 therapy. However, the infusion scheme has not yet been defined, the efficacy to avoid antigen escape has not been tested, and its toxicity remains to be assessed. Furthermore, the impact of genetic abnormalities on the outcome post CAR T-cell therapy has not been fully elucidated.

We conducted an open-label, single-center and single-arm pilot study of sequential infusion of third generation CAR T-cell "cocktail" comprising both anti-CD22 and anti-CD19 CAR T-cells, which was registered with Chinese Clinical Trial Registry (ChiCTR, number ChiCTR-OPN-16008526).

Between March 2016 and January 2018, a total of 89 eligibility pts completed CAR-22/19 "cocktail" infusion and were included for analysis. Of 89 pts (median age 36 years; range, 9-71), 51 had refractory/relapsed (R/R) B-cell acute lymphoblastic leukemia (B-ALL) and 38 had R/R B-cell non-Hodgkin lymphoma (B-NHL). The cutoff date for data collection was April 30, 2018, with a median follow-up of 7.6 months (mos) for B-ALL and 5.4 mos for B-NHL.

CAR-22 and CAR-19 cells were given separately in 2 divided doses (range, 1 to 4 for CAR-22, and 1 to 5 for CAR-19) with a dose-escalation infusion. Eighty-one pts firstly received CAR-22 infusion, and then CAR-19 infusion, while 8 pts received CAR-19 first and CAR-22 later. B-ALL pts received 2.73±1.19×106/kg CAR-22 and 2.61±1.45×106/kg CAR-19; B-NHL pts received instead higher doses, i.e. 5.28±2.44 ×106/kg CAR-22 and 5.14±2.06 ×106/kg CAR-19.

Of the 50 evaluable B-ALL pts, 48 (96.0%) achieved complete remission (CR)/or CR with incomplete count recovery (CRi) at the day 30 assessment. By sensitive multi-parameter flow cytometry (MFC) assay, 47 (94.0%) achieved MRD-negative CR/CRi. At a median follow-up of 7.6 mos (range, 1.3 to 22.2), the progression-free survival (PFS) of all B-ALL pts was 12.0 mos and the overall survival (OS) was not reached. Of the 13 pts with Ph+ B-ALL, 6 of whom had T315I mutation, 11 (84.6%) achieved MRD-negative CR/CRi. The PFS was 15.4 months, and the OS was not reached at a median follow-up of 13.5 months (range, 3 to 22.5). Of the 6 pts with MLL translocations (mainly MLL-AF6), 66.7% were progression-free and the median PFS was not reached. However, among pts with Ph-like ALL, the PFS was only 4.6 months. A total of 23 pts relapsed, however, antigen loss of CD19 and CD22 was not detected by MFC. Forty-Seven pts (92.2%) experienced cytokine release syndrome (CRS) and 11 (21.6%) had severe CRS (≥ grade 3). Except for 1 pt, all the severe cases were reversible. Seven pts (13.7%) developed neurotoxicity.

At a minimum follow-up of 3 months, the overall response rate (ORR) and CR rate in 36 evaluable B-NHL pts was 72.2% and 50%, respectively. At the cutoff date for data collection, 50.5% of all pts remained progression-free. A total of 8 pts (26.7%) had relapsed with a median PFS of 2.6 months. Re-biopsy and immunophenotyping was performed in 3 relapsed pts and antigen loss of CD19 or CD22 was not detected. At a median follow-up of 5.4 months (range, 3.0 to 16.2), both PFS and OS of all pts were not reached. Of the 9 pts with MYC translocation, 8 had a response and 7 maintained their responses with a median follow-up of 10.1 months, including all the 4 pts with "double-hit" lymphoma (concurrently carrying MYC and BCL2 or BCL6 translocations). Of the 10 pts with del(17p) or TP53 mutation, 8 (80.0%) had a clinical response. With a median follow-up of 5.3 months, the PFS was 3.3 months, and 45.0% of these pts were progression-free. All the B-NHL pts experienced CRS, but only 21.1% had severe (grade 3 or 4), but reversible CRS and 13.2% developed reversible neurotoxicity.

In summary, our results indicated that infusion of third generation CD22/19 CAR T-cell "cocktail" is feasible and safe for pts with B-cell malignancies. Co-targeting to CD19/CD22 is a promising approach for overcoming antigen escape relapse of CAR-19 therapy, but this remain to be fully tested. The treatment response predictive value of distinct genetic subtypes for CAR T-cell therapy should be considered for selection of pts for personalized immunotherapy.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract