LAUSR

Dear Editor, CD19 chimeric antigen receptor T cell (CAR-T) therapy has achieved high response rates in patients with relapsed/ refractory acute lymphoblastic leukemia (R/R ALL). However, it was reported that approximately 50% of patients who achieved complete remission (CR) eventually relapsed in 1 year. Therefore, rational prophylaxis against the escape of CD19-negative tumors is to generate T cells capable of recognizing multiple antigens. In addition to CD19, CD22 is another member of the B cell antigen family that has been validated as a successful target for B cell leukemias. Hence, we presented a CD19 and CD22 bispecific CAR, which is under investigation in clinical studies. CARs were constructed as described in previous reports (Supplementary Fig. 1). Notably, similar to its CAR-T signaling domain counterpart, CAR structure has a significant impact on mediating transduction, cytokine production, and cytotoxicity, which correspondingly impacts clinical outcomes. Hence, this article retrospectively analyzed in vitro characteristics and in vivo kinetic differences between CD19 and CD19/CD22 bispecific CAR-T. All patients enrolled in the CD19 CAR-T clinical trials (ChiCTR-ORN-16008948, n= 35) and CD19/22 CAR-T clinical trials (ChiCTR1800015575, n= 15) from 1 July 2015 to 30 April 2020 were selected for this retrospective series. None of the patients had previously been treated with any cell immunotherapy or Bite CD19 and CD22 antibody-drug conjugates. Other baseline characteristics are summarized in Supplementary Table 1. Before 1 October 2017, patients had not continuously observed the subtype of their cells by flow cytometry in the early stage. Hence, eight patients received CD19 CAR-T therapy, and 15 patients received CD19/CD22 CAR-T therapy later and were analyzed for the transduction and differentiation characteristics of the two types of CAR-T cells (Supplementary Table 2). The average transduction rate of CD19 CAR-T cells was higher than that of the CD19/CD22 CAR-T cells (45.5% vs 35.9%, P= 0.038), which shows that the CAR sequence did affect the initial transduction rate because the CD19/CD22 CAR had a larger size (Supplementary Fig. 2a). Moreover, during the cultivation process, there was no difference in the cell growth rate between the two groups. Furthermore, we analyzed the cell differentiation phenotype of CAR-T cells between the two groups. The median ratio of CD4+ to CD8+ CAR-T cells was 0.92 (range, 0.57–2.70) and 0.52 (range, 0.07–3.45), respectively. A slight but statistically significant increase in the CD4+ naïve T cell (Tn) population was consistently observed in cultured CD19/CD22 CAR-T cells compared with CD19 CAR-T cells (30.6% ± 29.0 vs 57.5% ± 20.9, P= 0.024), and fewer CD8+ effector T cells were observed in the CD19/CD22 CAR-T group (0.24% vs 0.03%, P= 0.029, Supplementary Fig. 2b). The peripheral blood samples of patients were continuously tested using flow cytometry. On average, the CD19 CAR-T group expanded >1% on day 4 (range, 1–6) and the CD19/CD22 CAR-T group expanded on day 3 (range, 2–10) after infusion. They achieved the peak number on day 10.5