An "off-the-shelf" CD2 universal CAR-T therapy for T-cell malignancies

Abstract

T-cell malignancies are associated with frequent relapse and high morbidity, which is partly due to the lack of effective or targeted treatment options. To broaden the use of CAR-T cells in pan T-cell malignancies, we developed an allogeneic "universal" CD2-targeting CAR-T cell (UCART2), in which the CD2 antigen is deleted to prevent fratricide, and the T-cell receptor is removed to prevent GvHD. UCART2 demonstrated efficacy against T-ALL and CTCL and prolonged the survival of tumor-engrafted NSG mice in vivo. To evaluate the impact of CD2 on CAR-T function, we generated CD19 CAR-T cells (UCART19) with or without CD2 deletion, single-cell secretome analysis revealed that CD2 deletion in UCART19 reduced frequencies of the effector cytokines (Granzyme-B and IFN-γ). We also observed that UCART19ΔCD2 had reduced anti-tumor efficacy compared to UCART19 in a CD19+NALM6 xenograft model. Of note is that the reduced efficacy resulting from CD2 deletion was reversed when combined with rhIL-7-hyFc, a long-acting recombinant human interleukin-7. Treatment with rhIL-7-hyFc prolonged UCART2 persistence and increased survival in both the tumor re-challenge model and primary patient T-ALL model in vivo. Together, these data suggest that allogeneic fratricide-resistant UCART2, in combination with rhIL-7-hyFc, could be a suitable approach for treating T-cell malignancies.

Fig. 1. UCART2 kills T-ALL and CTCL in vitro.

a Schematic of UCART2 design. b Schematic of CAR2-28ζ-34 and CAR19-28ζ-34 CAR constructs. c Timeline of UCART2 production with CRISPR/Cas9 gene editing. T cells were cultured in Xcyte media supplemented with 50 U/mL IL-2 and 10 ng/ml IL-15 and activated with anti-CD3/CD28 beads (bead to cell ratio 3:1) for two days followed by T-cell transduction of either CD2 or CD19 CAR construct. Transduced T cells were expanded for 9 days, followed by CD3+ depletion and hCD34 enrichment. Efficiencies of multiplex CRISPR/Cas9 gene editing of CD2 and TRAC were assessed by d. flow cytometry (CD2 vs. TRAC), or e, f. targeted deep sequencing of CD2 or TRAC. % NHEJ was determined as a percentage of sequencing reads with indels relative to WT cells. g In vitro killing assay. UCART2 or UCART19 cells were cultured with ⁵¹Cr-labeled HH, Jurkat, or Molt-3 cells at various E: T ratios for 4 h. Specific lysis was calculated based on released ⁵¹Cr in the culture medium.

Fig. 2. In vivo efficacy of UCART2 in a xenogeneic model of CTCL.

a Schema of the xenogeneic mouse model of CTCL. NSG mice were injected with 5 × 10⁵ HH^CBR-GFP cells on day -5, then infused with 2 × 10⁶ UCART2 or UCART19 on day 0. b, c tumor burden was assessed with BLI weekly (n = 5 per group). d Kaplan–Meier survival curve of mice treated UCART19, UCART2, or untreated control. Median survival: untreated mice (26 days), UCART19 treated mice (27 days), UCART2 treated (no death at the end of the experiment on day 65, p < = 0.0001).

Fig. 3. CD2 deletion affects the efficacy of UCART19.

a Schema of the generation of UCART19 (TRAC-CD2+) and UCART19ΔCD2 (TRAC-CD2-). b in vitro killing efficacy of UCART19 and U CART19ΔCD2. UCART19 and UCART19ΔCD2 were cultured with Nalm6 ^CBRGFP CD19+ targets at various E: T ratios for 16 h and target-specific killing was measured by luciferase activity. c Single-cell cytokine analysis using the IsoCode assay. UCART19 and UCART19ΔCD2 were incubated with B-cell lymphoma cell line Ramos at an E: T 1:2 for 20 h prior to loading purified CAR-T populations (CD19 depleted and CD34-affinity purified) onto the IsoCode chip. d Polyfunctionality and e Polyfunctional strength index (PSI) of the CD4 or CD8 subpopulation from UCART19 (CD2 +) or UCART19ΔCD2 (CD2-). f The signal strength of key cytokines driving polyfunctionality (GMZB, IFN-γ, MIP-1α, MIP-1β, Perforin, TNF-α, and TNF-β) was not affected by CD2 deletion. g The frequency of cells secreting Granzyme (Gzmb), IFN-γ, MIP1α, MIP1-β, Perforin, TNF-α, and TNF-β in the CD4 or CD8 subpopulation from UCART19 or UCART19ΔCD2. h Schema of the xenogeneic mouse model of CD19 + B cell acute lymphoblastic leukemia: 1 × 10⁶ NALM6^CBR-GFP cells were I.V. inoculated into NSG mice on day -5, followed by infusion of 1 × 10⁶ UCART19 or UCART19ΔCD2 on day 0 (n = 10 per group). i Tumor burden as determined by BLI. j Kaplan–Meier survival curve. Median survival: untreated mice, 26 days, mice treated with UCART19 ΔCD2, 45.5 days, all mice treated with UCART19 were alive at day +65.

Fig. 4. In vivo efficacy of UCART2 in combination with rhIL-7-hyFc.

a Schema of the xenogeneic mouse model of CTCL. NSG mice were injected with 5 × 10⁵ HH^CBR-GFP cells on day -4, then infused with 1 × 10⁶ UCART2 or UCART19 on day 0. 10 mg/kg of rhIL-7-hyFc was delivered subcutaneously on day 1, day 15, and day 29. b tumor burden was measured weekly by BLI. c Kaplan–Meier survival curve of mice treated with UCART19, UCART2, UCART19+rhIL-7-hyFc, and UCART2+rhIL-7-hyFc. Median survival: untreated mice, mice treated with NT-17 alone, and mice treated with UCART19 were 25 days, mice treated with UCART19 + NT-17 was 26.8 days, mice treated with UCART2 alone was 45.4 days, mice treated with UCART2+rhIL-7-hyFc were all alive at 200 days when the experiment was ended. d Tumor burden as determined by BLI. e Flow cytometry analysis was performed bi-weekly to assess circulating CAR-T cells present per µl of peripheral blood. CAR-T population determined by hCD45+, 7aad-, hCD34+ cells. P values < 0.05 considered significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. d BLI images normalized to a color gradient scale.

Fig. 5. rhIL-7-hyFc prolongs UCART2 persistence in vivo and overcomes tumor re-challenge.

a Schema of the CTCL re-challenge mouse model. NSG mice were injected with 5 × 10⁵ HH^CBR-GFP cells on day -4, then infused with 1 × 10⁶ UCART2 or UCART19 on day 0. 10 mg/kg of rhIL-7-hyFc was delivered subcutaneously on day +1, day +15, and day +29. Surviving mice were re-challenged with 5 × 10⁵ HH^CBR-GFP cells on day +83. 10 mg/kg of rhIL-7-hyFc was delivered subcutaneously on days +84, +98, and 112. b Kaplan–Meier survival curve. Mice receiving UCART. UCART19 Vs. UCART2, Median survival 30 days Vs. 58 days p = 0.0027. UCART19+rhIL-7-hyFc Vs. UCART2+rhIL-7-hyFc p = 0.0023 30 days Vs. 200+ days. c Tumor burden as determined by BLI imaging. d Normalized BLI images.

Fig. 6. UCART2, in combination with rhIL-7-hyFc, kills primary patient T-ALL in vivo.

a Schema of the patient-derived T-ALL xenograft model. NSG mice were injected with 5 × 10⁵ DCFI-15 cells on day -12, then infused with 1 × 10⁶ UCART2 or UCART19 on day 0. 10 mg/kg of rhIL-7-hyFc was delivered subcutaneously on day +1, day +15, and day +29. b Kaplan–Meier survival curve Median survival: untreated mice 26 days, rhIL-7-hyFc alone - 20 days, UCART19 26 days, UCART19+rhIL-7-hyFc - 21 days, UCART2 - 54 days, mice treated with UCART2+rhIL-7-hyFc 300+ days. UCART19 Vs. UCART2 p = 0.0019, UCART19+rhIL-7-hyFc Vs. UCART2+rhIL-7-hyFc, p = 0.0031. c Flow cytometry analysis was performed by-weekly to assess circulating tumor cells present per µl in peripheral blood. Tumor population was determined as hCD45 +, 7-AAD-, hCD34-, hCD2+ cells. d. Flow cytometry analysis was performed bi-weekly to assess circulating CAR-T cells present per µl in peripheral blood. CAR-T population determined by hCD45 +, 7-AAD-, hCD34 +, hCD2- cells.