Structural Optimization of Dual-Target CD22/CD19 CAR-T and Its Anti-Tumor Effects on B Cell Malignancies

Relapse to CD19 chimeric antigen receptor-modified (CAR) T cell therapy, which associated with diminished antigen density and presence of isoforms, were emerging and precluded the widespread application of this novel therapy. As such, CARs target dual antigens, such as CD19/CD22 bispecific CAR may become a feasible and promising strategy to reduce immune escape caused by tumor clonal evolution and antigen loss. Here, we designed a novel CD19/CD22 CAR, whose scFvs were derived from HI19α and HI22 clone generated in our institution, and then screened the most efficient construct into clinical application.

Considering the impact of different linkers and sequence of chains in scFvs on CAR efficiency, we constructed 4 types of TanCAR constructs and 2 types of LoopCAR constructs. Of the 6 CAR constructs, TanCAR 2 (CD22VL-CD22VH-CD19VL-CD19VH, Tan22-19 CAR) and LoopCAR 2 (CD22VL-CD19VL-CD19VH-CD22VH, Loop22×19 CAR) were validated with higher infection efficiency and killing effect on tumor cell lines.

To investigate the impact of CD19/CD22 CAR on CD19 and/or CD22 positive cells, we constructed CD19^KO (CD19 knock out) or CD22^KO (CD22 knock out) Namalwa cell lines through CRISPR-Cas9 technology. When co-cultured with Namalwa, Namalwa-CD19^KO and Namalwa-CD22^KO cells, compared with CD19 CAR-T or CD22 CAR-T which could only kill its corresponding target positive cells, Tan22-19 CAR-T and Loop22×19 CAR-T had significant cytotoxicity on CD19 and/or CD22 positive cells, indicating a dual-targeting efficacy.

To further explore the difference of Tan22-19 CAR-T and Loop 22×19 CAR-T under continuous exposure to antigen, re-challenge experiment was carried out. It was found that compared with Tan22-19 CAR-T, Loop22×19 CAR-T proliferated faster, eliminated tumor cells more significanly, and had a higher proportion of central memory T cells (T_CM), indicating that Loop22×19 CAR-T might have a more persisting function. In the simulated tumor recurrence mouse model, single CD19 or CD22 CAR-T could only partially eliminate tumor cells, while Loop22×19 CAR-T significantly reduced the recurrence caused by antigen loss and prolonged the survival of mice compared with that of Tan22-19 CAR-T and single-target CAR-T.

Then, Loop22×19 CAR-T was finally selected for evaluation of its clinical therapeutic effect. A total of 6 patients with relapsed/refractory B-cell acute lymphoblastic leukemia (R/R B-ALL) received Loop22×19 CAR-T cells therapy, in which 2 patients had undergone allogeneic hematopoietic stem cell transplantation (allo-HSCT) before receiving CAR-T therapy. The median age of the 6 patients was 18 years (range: 8-52 years), and the median percentage of blast cells in BM prior to lymphodepletion therapy was 18.1% (range, 0.03%-70.6%). Of the 6 patients, 5 received 3×10⁶/kg CAR T-cells infusion, and the dose of the other one was 1×10⁶/kg. Encouragingly, all 6 (100%) patients achieved MRD-negative remission within 32 days after CAR-T infusion. In addition, one BCR-ABL positive patient achieved molecular CR. These results demonstrate that Loop22×19 CAR-T therapy results in a deep and rapid response for patients with R/R B-ALL. Furthermore, after a median follow-up of 11.2 months, both the median overall survival (OS) and relapse-free survival (RFS) have not been reached. In entire cohort, 4 patients keeps remission persistantly, including 2 patients with TP53 mutation who bridged transplantation after CAR-T treatment. The remaining 2 patients experienced relapse, including 1 patient, who underwent allo-HSCT before CAR-T therapy, died 10 months after CAR-T infusion. As such, the one-year OS and RFS rate of the 6 patients were 77.8% (95% CI, 16.6%-96.5%) and 48.0% (95% CI, 0.04%-84.9%), respectively.

Although CRS was detected in all 6 patients, no severe CRS (grade ≥3) or CAR-T- related neurotoxicity was observed after Loop22×19 CAR-T infusion. For the hematological adverse events, all 6 patients developed reversible grade 3-4 neutropenia.

In conclusion, our study optimized bispecific CD19/22 CAR-T by constructing Tan and Loop CAR structures with different linkers and sequences of scFvs. Preclinical study demonstrated Loop structure 22×19 CAR-T exhibited more robust proliferation and anti-leukemia effect than Tan22-19. Besides, Loop22×19 CAR-T showed durable efficacy and reliable safety for B-ALL patients, and was expected to become a new strategy for reducing B-ALL recurrence.

Wang:AstraZeneca: Consultancy; Abbvie: Consultancy, Membership on an entity's Board of Directors or advisory committees.