Background
Two anti-CD19 chimeric antigen receptor (CAR) T cell (CD19 CAR T cell) products, tisagenlecleucel and axicabtagene, were the first US Food and Drug Administration (FDA)-approved gene-modified cell therapies for any indication [
1,
2]. Despite the great successes with CAR T cell therapy in leukemia that have been published previously by our group and others [
3‐
7], up to 60% of relapses after CD19 CAR T cell therapy are characterized by CD19 antigen loss, which involves several different mechanisms [
8‐
10]. Hence, the development of improvements in CAR design to target antigen loss represents a critical need. One approach to overcoming antigen loss following CAR T cell therapy is to simultaneously target more than one antigen on cancer cells, an approach that is compelling for B-ALL, given that anti-CD22 CAR T cells have also demonstrated clinical efficacy [
11].
Some studies have indicated that compared with single-antigen targeting, dual-antigen or multiantigen targeting by CARs may result in synergistic responses in solid tumors, optimizing response rates to therapy [
12,
13] and preventing antigen escape [
14,
15]. Preclinical data supporting the multitargeted approaches include tandem anti-CD19-CD20 CAR constructs; combinatorial anti-CD19 and anti-CD123 strategies, targeting both CD19 and CD22; and several clinical trials utilizing strategies are underway.
Here, we report the results of using autologous CD19/CD22 CAR T cells in six adult patients with chemotherapy-resistant or refractory ALL. All patients went into remission with no minimal residual disease (MRD), showing the efficacy of the bispecific CD19/CD22 CAR T cells, and remission was accompanied by robust expansion of the CAR T cells in vivo in these patients who were previously considered to have refractory and incurable disease.
Discussion
Adult patients with relapsed B-ALL have a dismal outcome with current therapies, especially patients who relapse within 6 months of their CR1. After the CR1, which is relatively easy to achieve, it becomes very difficult to achieve a second complete remission (CR2) when the patient relapses again. In adult patients, CR2 rates are lower than 50% [
27].
Despite unprecedented successes in early phase trials of anti-CD19 CAR T cells for the treatment of relapsed/refractory B-ALL, relapses with CD19 loss occurs in a large fraction of patients [
28]. Some studies [
29,
30] have indicated that simultaneous targeting of multiple antigens may reduce the likelihood of single target loss-induced relapse, similar to the well-established paradigm of multiagent combination chemotherapy regimens. Here, we report that autologous bispecific CAR T cells targeting both CD19 and CD22 mediated MRD-negative CRs in six adult patients with recurrent B-ALL.
FACS analysis of bone marrow from the three patients with disease relapse following bispecific CD19/CD22 CAR T cell therapy indicated two major patterns: relapse of antigen-positive tumor cells and relapse associated with antigen loss. ALL antigen-positive relapse is often associated with limited CAR T cell persistence [
5]. Patients 1 and 4 had loss of CAR T cell persistence at the time of relapse. Notably, in vitro experiments indicate that relapsed leukemia cells retain CD19 expression and sensitivity to bispecific CD19/CD22 CAR T cells ex vivo. These results suggest that relapse with CD19+ cells represents a potential opportunity for additional infusions of CAR T cells; however, repeated treatment strategies to treat CD19+ relapse occurring with the loss of CAR T cell persistence have unfortunately produced limited responses. Some groups have reported poor outcomes following reinfusion into patients with B cell non-Hodgkin lymphoma, but the use of an intensified lymphodepletion regimen containing fludarabine in addition to cyclophosphamide improves the reinfusion response and enhances initial CAR T cell expansion and persistence [
31]. This concept of intensified lymphodepletion has similarly been used in reinfusion strategies for other CAR-based therapies and has produced improved clinical outcomes [
32]. In addition, recipient anti-CAR immune responses that might lead to the loss of CAR T cell have been demonstrated against CAR T cells in clinical trials [
18,
33]. A solution to potentially reduce immunogenicity of CAR binding domains is to use human instead of murine sequences.
Although some studies [
34] have indicated that dual targeting is more effective in preventing antigen escape than sequential anti-CD19 CAR T cell and anti-CD22 CAR T cell infusions, we observed the unfortunate emergence of both CD19
− and CD22
dim blast cells in one patient treated with bispecific CD19/CD22 CAR T cells on this protocol. Thus far, established mechanisms leading to loss of CD19 expression include alternative splicing [
8], acquired mutations [
9], and lineage switch [
10]. Additionally, trogocytosis, whereby CAR T cells themselves acquire CD19 on the cell surface leading to both a reduction in the target antigen level on the tumor cell surface and more importantly CAR T cell death via fratricide, may lead to diminished responses or disease relapse [
34]. In contrast to the relapsing mechanism in most resistant leukemias that occurs after CD19 CAR T cell therapy, relapse after CD22 CAR T cell therapy typically occurs with sustained, but diminished, cell surface CD22 expression on leukemia cells [
11], thus implicating post-transcriptional mechanisms in this biology. The emergence of CD19
− and CD22
dim escape variants has been reported in patients sequentially infused with anti-CD19 CAR T cells and anti-CD22 CAR T cells [
11,
35]. It is possible that coinfusing CAR T cells targeting CD123 or CD38, in addition to those targeting CD19 and CD22, may diminish the likelihood of this event. Some studies indicated that sequential infusion of CAR19/22 T cell is a promising approach to reduce antigen loss relapse in CD19/CD22-directed therapy [
36]. In brief, this case provides proof-of-concept data on the potential for target antigen loss in leukemia following bispecific targeted immunotherapy and, within the growing field of novel targeted immunotherapies, highlights the importance of tumor evaluation throughout the course of treatment. We propose that reengineering CARs for multispecificity may diminish the likelihood of tumor escape through antigen loss. Proof of principle for the bioactivity of a bispecific CAR T cells in humans is demonstrated here.
For adult patients who achieve CR after CAR T cell therapy, treatment choices are very limited. Allo-HSCT may be a reasonable option. However, it is controversial whether allo-HSCT improves the outcomes of patients who achieve MRD-negative CR after CAR T cell therapy. A previous study showed that no significant differences in event-free survival or OS were found between patients who underwent transplantation and those who did not [
20]. However, Pan et al. reported that CAR T cell therapy bridging to allo-HSCT could significantly improve the clinical outcomes of refractory/relapsed B-ALL patients [
37]. In our study, similar to CD19 CAR T cell therapy, bispecific CD19/CD22 CAR T cell therapy provided a temporal window for patients otherwise ineligible or eligible under very suboptimal conditions (MRD+) to bridge to potentially life-saving allo-HSCT.
The primary CD19/CD22 CAR T cell-attributable toxicity observed was one instance of grade 1 and 2 CRS, which was similar to macrophage activation syndrome [
5,
38,
39]. Hyperferritinemia is a characteristic of these conditions. We observed substantial elevations in the ferritin level during the first month after infusion in our patients. In addition, we observed no evidence of neurotoxic effects, even in patient 5 with CNS leukemia, which is consistent with a previous report that observed follow-up therapy with anti-CD22 CAR T cells [
11]. The adverse events of CRS and neurotoxicity are closely related to marked increases in the serum levels of cytokines produced directly by rapidly proliferating CAR T cells after encountering CD19+ tumor cells or normal B cells or indirectly by myeloid cells activated by the CAR T cells [
40‐
43]. In contrast to previous studies showing marked increases in serum cytokine concentrations after CD19 CAR T cell therapy [
1,
3‐
5], our study found that serum cytokine and immune modulator levels were moderately increased after cell infusion in all treated patients. The results of in vitro and in vivo studies indicate that incorporating both CD19 and CD22 scFv sequences is critical for cytokine production in CAR T cells. Several studies have suggested important roles for the scFv framework region sequences in CAR T cell activation and function [
29,
30]. This study provides an avenue for CAR design with additional scFv sequences to improve the safety and efficacy of CAR T cell therapy.
The patterns of expansion and persistence of CD19/CD22 CAR T cells with a 4-1BB costimulatory endodomain are similar to those observed for anti-CD19 CAR T cells incorporating the 4-1BB endodomain [
4,
5]. CD19/CD22 CAR T cells also migrated efficiently to the CSF, raising the prospect that CD19/CD22 CAR T cells can prevent or treat CNS leukemia. Our results show the efficacy of CD19/CD22 CAR T cells in adults with acute leukemia.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.