Genetically modified CD7-targeting allogeneic CAR-T cell therapy with enhanced efficacy for relapsed/refractory CD7-positive hematological malignancies: a phase I clinical study

Abstract

Chimeric antigen receptor (CAR)-T cell therapy against T cell malignancies faces major challenges including fratricide between CAR-T cells and product contamination from the blasts. Allogeneic CAR-T cells, generated from healthy donor T cells, can provide ready-to-use, blast-free therapeutic products, but their application could be complicated by graft-versus-host disease (GvHD) and host rejection. Here we developed healthy donor-derived, CD7-targeting CAR-T cells (RD13-01) with genetic modifications to resist fratricide, GvHD and allogeneic rejection, as well as to potentiate antitumor function. A phase I clinical trial (NCT04538599) was conducted with twelve patients recruited (eleven with T cell leukemia/lymphoma, and one with CD7-expressing acute myeloid leukemia). All patients achieved pre-set end points and eleven proceeded to efficacy evaluation. No dose-limiting toxicity, GvHD, immune effector cell-associated neurotoxicity or severe cytokine release syndrome (grade ≥ 3) were observed. 28 days post infusion, 81.8% of patients (9/11) showed objective responses and the complete response rate was 63.6% (7/11, including the patient with AML). 3 of the responding patients were bridged to allogeneic hematopoietic stem cell transplantation. With a median follow-up of 10.5 months, 4 patients remained in complete remission. Cytomegalovirus (CMV) and/or Epstein-Barr virus (EBV) reactivation was observed in several patients, and one died from EBV-associated diffuse large B-cell lymphoma (DLBCL). Expansion of CD7-negative normal T cells was detected post infusion. In summary, we present the first report of a Phase I clinical trial using healthy donor-derived CD7-targeting allogeneic CAR-T cells to treat CD7⁺ hematological malignancies. Our results demonstrated the encouraging safety and efficacy profiles of the RD13-01 allogeneic CAR-T cells for CD7⁺ tumors.

Fig. 1. Construction of rejection-resistant “off-the-shelf” CD7-targeting CAR-T cells.

a CAR expression and efficiency of TCR/CD3 and CD7 disruption on CAR-dKO T cells, measured by flow cytometry. b, c Tumor control ability of CAR-dKO T cells evaluated in NCG mice bearing Jurkat tumors. Mice were injected with 1 × 10⁶ Jurkat tumor cells intravenously (i.v.) and treated with a single injection of 10 × 10⁶ non-transduced (NT) T cells (n = 4) or CAR-dKO T cells (n = 8) 72 h after tumor inoculation. Bioluminescent images obtained at the indicated time points after tumor injection (b). Kaplan–Meier curves of mice treated with NT or CAR-dKO T cells (c), with P value calculated using a Log-rank test. d Top, composition of CD7⁺ and CD7⁻ subsets in CD4⁺/CD8⁺ T cells from healthy donors (n = 5) and patients with T-ALL (n = 14). Bottom, elimination of HLA-II expression by disrupting RFX5, generating CAR-tKO T cells. e Expansion of allogeneic CD4⁺ T cells in the presence of CAR-dKO or CAR-tKO T cells in a mixed lymphocyte reaction (MLR) assay. f E-cadherin expression on CAR-tKO and CAR-NKi-tKO T cells. g NK cell cytotoxicity against CAR-tKO and CAR-NKi-tKO cells was detected using a luciferase-based killing assay at different NK:CAR-T ratios after 24 h of co-culture (n = 3). h Illustration of CAR-T cell development from “CAR-dKO” to “CAR-tKO” to “CAR-NKi-tKO”. Data are presented as means ± SD; All comparisons (except for Kaplan–Meier curves) were determined using student’s t-tests; *P < 0.05; **P < 0.01; ns not significant.

Fig. 2. Incorporation of γc improves CAR-T cell function.

a Schematic illustration of γc-tethered CAR design. b Secretion of IL-2 by CAR-NKi-tKO and CAR-g-NKi-tKO T cells following 8 h of stimulation using Jurkat tumor cells (n = 3). c Cytotoxicity of CAR-NKi-tKO and CAR-g-NKi-tKO T cells against Jurkat cells at indicated Effector: Target (CAR-T: Tumor) ratios after 8 h of co-culture. d Secretion of cytotoxic molecules in CD4⁺ and CD8⁺ subsets of CAR-NKi-tKO and CAR-g-NKi-tKO T cells following 8 h of stimulation using Jurkat tumor cells (n = 3). e NCG mice were injected with 1 × 10⁶ Jurkat tumor cells (i.v.) and treated with a single injection of 10 × 10⁶ NT, CAR-NKi-tKO or CAR-g-NKi-tKO T cells. Left, bioluminescent images obtained at the indicated time points after tumor injection; right, Kaplan–Meier curves of mice treated with NT, CAR-NKi-tKO or CAR-g-NKi-tKO T cells, with P value calculated using a Log-rank test. f Ex vivo expansion of CAR-NKi-tKO or CAR-g-NKi-tKO T cells during production. g Expression of TIM3, LAG3, and PD-1 on CAR-NKi-tKO or CAR-g-NKi-tKO T cells at indicated time points during production (n = 3). h Illustration of RD13-01 CAR-T cell degisn. Data are presented as means ± SD; All comparisons (except for Kaplan–Meier curves) were determined using student’s t-tests; *P < 0.05; **P < 0.01; ns not significant.

Fig. 3. Clinical procedure of treatment with RD13-01 CAR-T cells.

a CONSORT diagram of the clinical trial. AML, acute myeloid leukemia; CNSL, central neural system lymphoma; FCE, fludarabine cyclophosphamide and etoposide. b Diagram of clinical treatment protocol: the day of RD13-01 infusion was set as day 0; the lympho-depleting condition regimen consisted of fludarabine (30 mg/m²), cyclophosphamide (300 mg/m²) and etoposide (100 mg/day) on days ‒7 to ‒3.

Fig. 4. Safety and efficacy profiles of RD13-01 CAR-T cells.

a Graph of incidences of CRS and ICANS in all patients treated. b Serum concentrations of IFN-γ, IL-6 and IL-10 in patients who developed CRS after RD13-01 infusion. Samples were taken 2 h after CAR-T infusion (“before CRS”), and at the times of initiation, peak and restoration of CRS-related symptoms. Data are presented as means ± SD. c Swimmer plot (n = 12) showing patient responses. Each bar represents an individual patient. Responses were determined on day 28 (arrowhead) and are indicated by different colors (blue, CR; green, PR; orange, PD; gray, NE). Bars with solid arrows represent patients in an ongoing follow-up. d CD7⁺ malignant cells in bone marrow measured using flow cytometry (left) and microscopic images (100× and 1000×, bar indicates 10 μm) of bone marrow samples (right) of Patient 5, who achieved a complete response at day 28 after RD13-01 infusion. e Computed tomography scans showed a complete response of extramedullary leukemia in Patient 9. Pre-treatment (baseline) tumor lesion was indicated by the red arrow.

Fig. 5. RD13-01 expansion and endogenous T cell dynamics.

a, b Flow cytometry (a) and qPCR (b) detections of CAR-T cells in the peripheral blood of patients who achieved responses (Responders) or no response (Non-responders) at different time points post CAR-T infusion. c Stacked pie charts showing subset compositions of normal lymphocytes from responders. Plotted are median values. DNT, double negative T cell. d Fold changes of different types of endogenous lymphocytes at indicated time points compared to day 1 post-infusion. Plotted are median values for responders. e, f Surface antigen expression on normal lymphocytes (defined as the CD45-bright population) from Patient 5 at different times points. CD3 and CD7 expression on normal lymphocytes before CAR-T cell infusion (e); CD3, CD7 and CAR expression on normal lymphocytes on day 15 and 42 after infusion (f). CAR was gated from the CD3⁻ population, and CD4/CD8 were gated from the CD3⁺ population. g MLR assay was performed by mixing PBMCs from Patient 5 (day 42 post infusion) with naïve T cells from the same donor origin as RD13-01 CAR-T cells, or naïve T cells from an unrelated donor (PBMC: naive T cell = 4:1), for 72 h. Percentages of patient-originated CD7⁻CD8⁺ T cells expressing various activation markers (OX40, CD137, CD25, CD40, CD69) were determined by flow cytometry to indicate allogeneic T cell activation against the CAR-T cell donor. Data are presented as means ± SD. All comparisons were calculated using Student’s t-tests. *P < 0.05; **P < 0.01; ns not significant.