CD22-directed CAR T-cell therapy induces complete remissions in CD19-directed CAR–refractory large B-cell lymphoma

Outcomes of patients with large B-cell lymphoma (LBCL) that has relapsed after or is refractory (R/R) to chimeric antigen receptor T-cell therapy targeting CD19 (CAR19) remain poor, with <25% responding to subsequent therapies, and a median overall survival (OS) of 3.6 months.¹  Targeting of alternative antigens represents an important therapeutic strategy for patients who relapse after CAR19, including those with CD19 loss or downregulation.^1,2  CD22 is expressed on LBCL cells at levels similar to those found on nonmalignant B-cells^3,4  and is an effective target for CAR T-cell therapy (CAR22) in children with R/R acute lymphoblastic leukemia.^5,6  No CD22-directed therapy is currently approved for use in LBCL, and only 2 cases of CD22-directed CAR T-cell therapy against LBCL have been described.^7,8  We report the safety and interim efficacy of the initial dosing cohort of a phase 1 dose-escalation study of CD22.BB.z-CAR in adults with LBCL that relapsed after CAR19.

Eligible patients had to have received at least 2 prior lines of therapy for LBCL, with measurable disease that demonstrated any level of CD22 expression. For patients who received prior CAR T-cell therapy, CAR⁺ cells had to represent <5% of peripheral blood CD3⁺ lymphocytes. Lymphodepletion with fludarabine 30 mg/m² and cyclophosphamide 500 mg/m² preceded infusion of 1 × 10⁶ CD22.BB.z-CAR⁺ cells per kilogram (Figure 1A). CAR T-cell–related toxicity was graded according to American Society for Transplantation and Cellular Therapy consensus criteria⁹  and was managed via institutional protocol. Common Terminology Criteria for Adverse Events (v5.0) and Lugano criteria¹⁰  were used to grade adverse events and responses, respectively. The trial protocol received institutional review board approval. Patients provided informed consent in accordance with the Declaration of Helsinki. Products were manufactured in the automated, closed-system Miltenyi CliniMACS Prodigy device. CD4/CD8 T-cell–enriched peripheral blood mononuclear cells were transduced with a lentiviral vector containing a single-cistron–encoded CD22.BB.z-CAR cell comprising a fully humanized CD22 scFv (m971), a CD8 transmembrane domain, a 4-1BB costimulatory domain, and a CD3ζ activation domain (supplemental Figure 1, available on the Blood Web site). Products were phenotypically similar and were predominantly CD4⁺ T cells (supplemental Figure 2). Details regarding patient eligibility, product manufacturing, toxicity management, and in vivo correlative studies are included in the supplemental Methods.

Three patients with R/R LBCL, in whom the disease had progressed after multiple treatments, including CAR19 therapy, were enrolled and treated as the first 3 consecutive patients in this clinical study of CAR22. Patient and disease characteristics at the time of enrollment are shown in supplemental Table 1. None had achieved a durable remission with any prior treatment (supplemental Table 2), including CAR19, after which patients 1 and 2 (P1 and P2) had achieved only a partial response (PR). All patients had progressive disease by 3 months after CAR19 infusion (supplemental Figure 3A). Patient 3 (P3) had progressed after CAR19 and bispecific CD19- and CD20-targeted CAR (CAR20.19; supplemental Figure 3B).¹¹  All patients had biopsy-proven expression of surface CD22 detected by flow cytometric cell sorting (FACS) and immunohistochemistry, whereas CD19 expression was downregulated or lost in 2 patients (P2 and P3; Figure 1C; supplemental Figure 4). P3 had surface CD19 downregulation after CAR19, followed by complete loss of CD19 after CAR20.19, whereas CD22 remained preserved throughout (Figure 1D).

No nonhematologic treatment–emergent adverse events of grade ≥3 occurred (supplemental Table 3). Grade 1 cytokine release syndrome (CRS) occurred in all patients within 24 hours after CAR22 infusion. P1 and P3 developed grade 2 CRS on days 11 and 7 after infusion, respectively; each received tocilizumab and dexamethasone with subsequent resolution. No patient demonstrated evidence of immune effector cell–associated neurotoxicity syndrome. All patients developed grade ≥3 neutropenia, thrombocytopenia, and anemia after lymphodepletion chemotherapy and CAR22 infusion, and recovery to grade ≤2 cytopenias did not occur until nearly 4 months after infusion (supplemental Figure 5).

All patients achieved a complete response (CR) as the best response after CAR22. Progressive reduction in the size of all index lesions was observed over the course of follow-up (Figure 2A; supplemental Table 3). Both patients with high tumor burden (P1 and P2) had initial PR at day 28, which subsequently improved to CR at 6 and 3 months, respectively (Figure 2B). At a mean follow-up of 7.8 months (range, 6-9.3), all patients remained in CR. Similarly, circulating tumor DNA (ctDNA) was undetectable (P1 and P3) or reduced by 99% (P2) from baseline at 6 months after infusion, demonstrating persistent CAR22 activity (Figure 2C).

Serial flow cytometric evaluation of peripheral blood (CAR-FACS) demonstrated a 100- to 400-fold expansion, with peak CAR⁺ T-cell levels occurring at ∼14 days. Circulating CAR⁺ T cells persisted for at least 3 months in all patients and were detectable by quantitative polymerase chain reaction (PCR) at the time of last assessment up to 6 months (Figure 2D-E). P1 and P3, who had received prior CAR19 therapy at our institution, experienced an 11.7- and 55.9-fold higher peak CAR⁺ T-cell level with CAR22 vs CAR19 therapy. Plasma levels of 48 cytokines were assessed over the first 28 days after infusion (supplemental Figure 7). In both patients who developed grade 2 CRS (P1 and P3), interleukin-6 and interleukin-1 receptor antagonist levels reached 7.7- to 12.1-fold and 3.7- to 4.6-fold higher concentrations before the event.

We report the first experience of autologous CAR T cells targeting CD22 for R/R LBCL and demonstrate ongoing CR in the first 3 consecutive patients treated. CAR22 induced clinical and molecular CRs despite prior failure of autologous CAR19 therapy.

Currently, data regarding retreatment with CAR T-cell therapies are limited. In patients with LBCL who relapse after CAR19, reinfusion with an identical or derivative CAR19 product has not produced a durable response.¹²  This report provides additional evidence that targeting CD22 can mediate meaningful clinical activity in LBCL, even in the post-CAR19 relapse setting.¹³ 

CD22 antigen density has been shown to be an important factor that modulates the persistence and efficacy of CAR22⁺ cells in B-acute lymphoblastic leukemia.^6,14  In 2 of our patients, CD22 expression was maintained at high levels, despite downregulation or loss of CD19 in response to the selective pressure of CAR19 therapy. The current ongoing clinical study will help inform the effect of CD22 antigen density on the response to CAR22 therapy for LBCL.

In all patients, CAR22 demonstrated robust expansion and persistence, similar to levels reported in CAR T-cell–naive patients who received 41BB-based CAR19.^15,16  The corresponding rapid and pleiotropic changes in infusion levels of inflammatory and effector cytokines mimicked patterns that have been shown to be associated with improved antitumor activity.^17,18  The absence of severe immune effector cell–associated neurotoxicity syndrome, or CRS, is noteworthy, given that analogous patterns of CAR⁺ T-cell expansion and inflammatory response are often associated with these toxicities in patients who undergo CAR19 therapy.^19-21 

The use of CD4/CD8-enriched apheresis material with the CliniMACS Prodigy platform resulted in products with a predominance of CD4⁺ T cells in all patients, in line with previous reports on a CD22.BB.z-CAR.⁶  However, in vivo CAR⁺ T-cell expansion was largely CD8⁺ T cells. Despite a minority of T cells in the product that exhibited an early memory phenotype (both central memory and naïve memory), which has been associated with greater persistence in vivo, we observed detectable CAR⁺ cells beyond 6 months after infusion.^22,23  Studies are ongoing to examine the clonal dynamics of CAR22 in vivo and better understand the contribution of these populations to response and toxicity.

In summary, we provide initial evidence that demonstrates the safety and antitumor activity of CAR T-cell therapy targeting CD22 in patients with R/R LBCL, establishing CD22 as a potential target for CAR T-cell therapy in LBCL. Further accrual at a higher dose level in this phase 1 dose-escalation study is ongoing and will explore the role of this therapy in patients in whom prior CAR T-cell therapies have failed.

The authors thank the staff and faculty of the Stanford University Blood and Marrow Transplantation Program and the Stanford Center for Cancer Cell Therapy for their tireless work in caring for the patients involved in this study.

This work was supported by National Institutes of Health (NIH), National Cancer Institute (NCI) grants 2P01CA049605-29A1 (C.L.M., R.S.N., D.B.M.), 5P30CA124435 (C.L.M.), and K08CA248968 (M.J.F.) and by the Virginia and D. K. Ludwig Fund for Cancer Research. C.L.M is a member of the Parker Institute for Cancer Immunotherapy, which supports the Stanford University Cancer Immunotherapy Program.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Contribution: L.M., M.J.F., C.L.M., and D.B.M. designed the study; J.H.B., L.M., and M.J.F. interpreted the data; J.H.B., S.P., J.T., L.M., and M.J.F. analyzed the data and created the figures; J.H.B., L.M., M.J.F., D.B.M., and C.L.M. drafted the manuscript; and all authors contributed to data collection, writing and revision of the manuscript, and approval of the final version.

Conflict-of-interest disclosure: P.S. has received research support from Kite Pharma-Gilead. A.R.R. has received research support from Pharmacyclics/AbbVie, served on 1-time ad hoc scientific advisory boards of Nohla Therapeutics and Kaleido, and was an expert witness for the U.S. Department of Justice. His brother works for Johnson & Johnson. C.L.M. has consulted for Lyell, Neoimmune Tech, Nektar, and Apricity; has received royalties from NIH and Juno Therapeutics for CD22-CAR; holds equity in Lyell and Apricity; and has received research support from Lyell. D.B.M. has consulted for Kite Pharma-Gilead, Juno Therapeutics-Celgene, Novartis, Janssen, and Pharmacyclics and has received research support from Kite Pharma-Gilead, Allogene, Pharmacyclics, Miltenyi Biotec, and Adaptive Biotechnologies. S.S. has consulted for Janssen. C.M., A.J., and I.K. are full-time employees of Adaptive Biotechnologies. S.A.F. has consulted for Lonza PerMed, Gradalis, Obsidian, and Samsara BioCapital; L.M. has received research support from Adaptive Biotechnologies and Servier Laboratories and has consulted for Amgen and Pfizer. The remaining authors declare no competing financial interests.

Correspondence: Lori Muffly, Stanford University, 300 Pasteur Dr, H0144, Stanford, CA 94305; e-mail: lmuffly@stanford.edu.