CD22-targeted chimeric antigen receptor-modified T cells for children and adults with relapse of B-cell acute lymphoblastic leukemia after CD19-directed immunotherapy

Abstract

Background: Relapse of B-cell acute lymphoblastic leukemia (B-ALL) with CD19-antigen loss after CD19-targeted chimeric antigen receptor (CAR) T-cell therapy has a dismal prognosis. Novel immunotherapeutic strategies for this patient population are urgently needed.

Methods: We tested a novel, fully human anti-CD22/4-1BB CAR T-cell construct, CART22-65s, in parallel phase I studies for pediatric and adult B-ALL. After lymphodepletion, CART22-65s was infused using a 3-day fractionated dosing scheme, allowing for omission of the second and third doses in cases of early cytokine release syndrome (CRS).

Results: Twenty-two patients, all with relapse after prior CD19-directed immunotherapy, were enrolled. Of 19 infused patients (pediatric, n=17; adult, n=2), 14 (74%) achieved a complete remission (CR), including 4 of 6 (67%) patients refractory to prior inotuzumab. Five of 14 patients in a CR proceeded to consolidative hematopoietic cell transplantation (HCT). With a median follow-up of 38 months, the 12-month relapse-free survival rate was 38.4% (95% CI 19.3% to 76.5%) and overall survival rate was 52.6% (95% CI 34.3% to 80.6%). Two patients received additional CART22-65s treatments for subsequent CD22-positive relapses; one achieved another CR. All CRS (n=17, 89%) and neurotoxicity (n=4, 21%) events after initial infusion were grades 1-2. The only grade 3 CRS/neurotoxicity and the only high-grade immune effector cell-associated hemophagocytic lymphohistocytosis-like syndrome occurred in the retreatment setting. In vivo cellular kinetic data revealed robust CART22-65s proliferation by quantitative PCR peaking at a median of 20 days postinfusion, with the cells persisting out to month 42 in one patient who achieved a long-term remission with CART22-65s alone.

Conclusions: The favorable safety profile and high remission rates in exceedingly refractory B-ALL support the continued development of CART22-65s but also highlight the need to use the product in combination with HCT or other novel strategies.

Trial registration numbers: NCT02650414 and NCT03620058.

Keywords: Chimeric antigen receptor - CAR; Cytokine release syndrome; Immunotherapy; Leukemia; Relapse.

Figure 1. CONSORT diagram. Screening, enrollment, and treatment on the pediatric and adult CART22-65s trials. CAR, chimeric antigen receptor; CART22-65s, CD22-directed CAR T cell product. CONSORT, Consolidated Standards of Reporting Trials.

Figure 2. EFS, RFS, OS, and clinical courses of individual patients. (A) EFS, calculated from the date of CART22-65s infusion to the earliest date of the following events: no response (n=5) or morphologic relapse after achieving a CR (n=8). Data were censored for any alternative therapy, except for hematopoietic cell transplant, during remission. (B) RFS, calculated from the date of response assessment to the date of relapse among the patients who achieved a CR. Data were censored for alternative therapy during remission, excluding hematopoietic cell transplant. (C) OS, defined as the time from CART22-65s infusion to date of death from any cause or date of last follow-up. 14 deaths occurred during the follow-up period, all from progressive disease (D). Swimmer plots illustrating clinical courses of individual patients in the pediatric (n=17) and adult (n=2) cohorts from time of infusion to last follow-up. Data shown include duration of remission and times to initial response, hematopoietic cell transplant or other alternative therapy in remission, relapse denoted by CD22 relapse immunophenotype, detection of NGS-MRD, and CART22-65s retreatment. CR, complete remission; EFS, event-free survival; MRD+, measurable residual disease as measured by multiparameter flow cytometry; NGS-MRD, MRD as measured by next generation sequencing; OS, overall survival; RFS, relapse-free survival.

Figure 3. CART22-65s expansion. (A) Measurements of CART22-65s gene-modified T cells in peripheral blood as assessed by qPCR assay for all infused patients (n=19) during the first month after CAR infusion. Blue lines represent patients who achieved a CR (n=14); red lines represent patients with NR (n=5). (B) Peak expansion was not significantly different between patients with a CR (blue) vs NR (red) to CART22-65s (p=1.0, as derived from the Wilcoxon rank sum test). (C) Peak expansion was also similar by CRS grade (p=0.21, as derived from the Kruskal-Wallis test). (D) We identified linear correlates of CAR expansion measured by qPCR at any timepoint (n=95 data points); the top 10 linear correlates are shown. (E) SIT1 was the protein most associated with CAR expansion (Pearson R=0.75, p<2.2e−16). Patients with CR are shown in blue; NR is shown in red. (F) Measurements of CAR transgene by qPCR in the CSF at day 28 in patients with CR (blue) or NR (red), demonstrating CART22-65s was detectable CAR in 12 of 14 (86%) CSF samples. (G) In patients in whom their prior CD19 CAR construct was detected in their manufactured CART22-65s product (n=4), measurements of that CD19 CAR transgene during the first month after CART22-65s infusion. (H) Measurements of CART22-65s transgene in peripheral blood by qPCR following retreatment in the two patients retreated for subsequent CD22+ relapses. For all qPCR curves, the horizontal line at 25 copies/µg DNA represents the lower limit of quantification of the assay. CFS, cerebrospinal fluid; CR, complete response; CRS, cytokine release syndrome; NR, non-response; qPCR, quantitative real-time PCR.

Figure 4. Persistence of CART22-65s. (A) Measurements of CART22-65s gene-modified T cells in peripheral blood as assessed by quantitative real-time PCR assay for all infused patients (n=19) from CAR infusion to date of the first event postinfusion event: non-response, relapse, hematopoietic cell transplant or other alternative therapy, or last follow-up. Data were censored at the month 12 time point. The horizontal line at 25 copies/µg DNA represents the lower limit of quantification of this assay. (B) Measurements of CART22-65s gene-modified T cells in peripheral blood as assessed by quantitative real-time PCR assay through month 42 for the patient in the pediatric cohort who remains in a long-term remission without consolidative hematopoietic transplant. The horizontal line at 25 copies/µg DNA represents the lower limit of quantification of this assay. CAR, chimeric antigen receptor.

Figure 5. Cytokine responses to CART22-65s. (A) NPX for the cytokines CD163, CSF3, TNF, IL10, IL2, IL2RA, IFNG, CXCL9, IL4, IL5, IL7, and IL8 is shown over time. Lines are colored by grade of CRS. (B) Differentially expressed proteins between patients who had a CR (n=13) vs NR (n=4) at day 21, the median time to peak CART22-65s expansion. IL-6 (C) and FLT3 (D) were the two most differentially expressed proteins higher in NR than CR. Notably, FLT3-LG, the ligand of FLT3, was the most highly differentially expressed protein in CR (E). Cadherin-1 (CDH1) was more highly expressed in CR than NR over time (F). All cytokine data are from the pediatric cohort only. CR, complete response; CRS, cytokine release syndrome; NPX, normalized protein expression; NR, no response.