CD22-Targeted Chimeric Antigen Receptor (CAR) T Cells Containing The 4-1BB Costimulatory Domain Demonstrate Enhanced Persistence and Superior Efficacy Against B-Cell Precursor Acute Lymphoblastic Leukemia (ALL) Compared To Those Containing CD28

Abstract B cell precursor acute lymphoblastic leukemia (BCP-ALL) remains a leading cause of death from childhood cancers despite survival rates exceeding 80%. Antibody-based CAR-engineered T cells can recognize and eliminate tumors by binding directly to a surface antigen independent from MHC restriction. CAR immunotherapy against BCP-ALL has demonstrated impressive responses and sustained remission in clinical trials targeting CD19. However, some patients receiving the CD19 CAR T cells relapse with a CD19 negative leukemia. Thus, additional CAR targets are needed. CD22 is a Siglec family lectin consisting of 7 extracellular Ig domains that is expressed on the cell surface from the pre-B cell stage of development through mature B cells and is expressed on most B cell hematologic malignancies. We previously generated a second-generation (CD3-Zeta + CD28 costimulatory domain) anti-CD22 CAR derived from a membrane proximal epitope binding scFv (m971-28z) with potent activity in vivo (Haso W et al, Blood 2013). In clinical trials T cells expressing CD19-targeted CAR with 4-1BB costimulatory domains on CD19 CARs show prolonged persistence. To improve long-term persistence of the CD22 CAR, we re-engineered our CAR vector to include a 4-1BB signaling domain (m971-BBz). In vitro data using m971-BBz improved proliferation and expansion compared to m971-28z especially when lower concentrations of IL2 were included in the culture media. When no IL2 was added to the media only the 4-1BB containing CAR expanded. No difference in killing was detected in in vitro cytotoxicity assays. We next evaluated anti-tumor activity and persistence in the NSG mouse model engrafted with the NALM6-GL cell line on day 0 and treated with CAR T cells on day 3 to directly compare the efficacy of m971-28z and m971-BBz modified T cells activated with either OKT3 or anti-CD3/CD28 beads. m971-BBz outperformed m971-28z in terms of in vivo anti-tumor activity and long-term persistence. This effect was only detected when anti-CD3/CD28 beads were used for T cell expansion. OKT3-activated cells failed to persist and demonstrated inferior antitumor activity compared to bead-expanded T cells irrespective of the costimulatory domain and despite a higher percentage of CD8 T cells with significantly better cytotoxicity in vitro. Interestingly, early peripheral blood numbers of CAR T cells in recipients of bead-expanded products demonstrated a predominance of CD4+CAR T cells consistent with preinfusion CD4/CD8 ratios. At later time points this ratio decreased with a predominance of CD8+CAR T cells. In mice receiving m971-28z CAR the CD8+CAR T cells failed to persist resulting in leukemic relapse. Furthermore, direct comparison to the CD19 CAR (FMC63-BBz) in vivo showed that the anti-CD22 CAR (m971-BBz) has equivalent activity. We conclude that anti-CD3/CD28 bead-activated T cells modified to express an anti-CD22 CAR with a 4-1BB costimulatory domain demonstrates potent antitumor activity with long-term leukemic control and offers a promising therapeutic option for pediatric ALL. Disclosures: No relevant conflicts of interest to declare.