Activation of Cell-Intrinsic Signaling in CAR-T Cells via a Chimeric IL7R Domain

Abstract

Chimeric antigen receptor (CAR) T cells can effectively treat leukemias, but sustained antitumor responses can be hindered by a lack of CAR T-cell persistence. Cytotoxic effector T cells are short-lived, and establishment of CAR-T cells with memory to ensure immune surveillance is important. Memory T cells depend on cytokine support, with IL7 activation of the IL7 receptor (IL7R) being critical. However, IL7R surface expression is negatively regulated by exposure to IL7. We aimed to support CAR T-cell persistence by equipping CAR-T cells with a sustained IL7Rα signal. We engineered T cells to constitutively secrete IL7 or to express an anti-acute myeloid leukemia-targeted IL7Rα-chimeric cytokine receptor (CCR) and characterized the phenotype of these cell types. Canonical downstream signaling was activated in CCR-T cells with IL7R activation. When coexpressed with a cytotoxic CAR, functionality of both the CCR and CAR was maintained. We designed hybrid CAR-CCR and noted membrane proximity of the intracellular domains as vital for signaling. These data show cell-intrinsic cytokine support with canonical signaling, and functionality can be provided via expression of an IL7Rα domain whether independently expressed or incorporated into a cytotoxic CAR for use in anticancer therapy.

Significance: To improve the phenotype of tumor-directed T-cell therapy, we show that provision of cell-intrinsic IL7R-mediated signaling is preferable to activation of cells with exogenous IL7. We engineer this signaling via independent receptor engineering and incorporation into a CAR and validate maintained antigen-specific cytotoxic activity.

Figure 1

Human primary T cells can be engineered for stable IL7 pathway activation. A, Schematic of CCR and sIL7, as well as transgenes engineered into cells. B, Transduction efficiency of T cells engineered to express GFP only, IL7R + GFP, or to secrete IL7 + GFP was measured by detecting eGFP positivity and CCR expression with flow cytometry. C, Concentration of IL7 measured in the supernatant of sIL7-transduced T cells and control cells (G-o) in cytokine-free culture medium. D, Immunophenotype characterization of T-cell subsets present in healthy donor and transduced T-cell preactivation (CD4 and CD8) and on day 8 of T-cell expansion [CCR and sIL7 treated with IL7: +IL7 (G-o); G-o]. E, Ratio of STAT5 phosphorylation to total STAT5 expression measured with TR-FRET in indicated cell types following 24 hours of cytokine starvation and following stimulation with IL7 for 30 minutes. Comparison between untreated and IL7-treated cells. F, Percent cells positive for CD127 expression measured by flow cytometry. Statistical comparison is between cell type and control (G-o, no treatment). G, Representative histograms from flow cytometric measurement of %CD127⁺ cells of each type. Data representative of n = 3 to 5 independent donors. *, P < 0.05; **, P < 0.01; ***, P < 0.001. For DG, +IL7 indicates cells activated and expanded in the presence of IL7 prior to indicated analysis. H, MS proteomics analysis of IL7R signaling pathway downstream molecules for total protein and phosphosite abundance. Raw detected protein and phosphoprotein abundance were normalized to the control condition per donor and expressed as mean log₂ fold change of three donors. *Note STAT5A pY694 and STAT5B pY699 are indistinguishable due to sequence identity. G-o, GFP-only.

Figure 2

T cells with IL7 pathway activation expand and circulate in vivo. A, Schematic of experimental design. NSG mice were injected on day 0 with 10⁷ T cells expressing a ffLuc reporter. BLI was conducted biweekly, and peripheral blood collected on days 15 and 30. One group of mice was treated with ffLuc-only–modified T cells and intraperitoneal injections of IL7 daily from day 1 to 21. B, BLI monitoring T-cell expansion. C, Quantified radiance per mouse. Thick solid line, median BLI; dotted lines, individual values per mouse. D, Calculated doubling time of T-cell products with 95% confidence interval represented. E, T cells (CD3⁺GFP⁺) detected in peripheral blood on days 15 and 30 after injection. Thick solid line, median; dotted lines, individual values per mouse. Comparison between treatment cohorts on day 30 vs. no treatment: *, P < 0.05; ***, P < 0.001. F, Concentration of human IL7 in murine plasma detected with ELISA. For the intraperitoneally injected mice on day 15, blood was collected 1 hour after IL7 injection. (n = 5 mice per group). hIL7, human IL7; NSG, NOD/SCIDγ; NT, nontransduced; PB, peripheral blood.

Figure 3

CD123-targeted CAR and CCR can be coexpressed with maintained functionality. A, Schematic showing CAR and CCR structures. B, Transduction efficiency measured using flow cytometric staining of Halo and SNAP tags. n = 3 to 6 unique T-cell donors, unless noted no significant difference between groups. C, Measurement of phosphorylated STAT5 percentage in cytokine starved, engineered T cells measured with and without activation on the immobilized rhCD123 target. n = 3 to 4 unique T-cell donors, comparison of CCR + CAR⁺ and CAR⁺ to NT (black) and with stimulation (colored). D, Soluble IL2 and IFNγ measured in the supernatant following coculture of unmodified (NT), CCR⁺, CAR⁺, or CCR + CAR⁺ T cells with CD123-negative (K562) and CD123-positive (K562.CD123, MV-4-11, and Molm-13) targets. n = 3 to 6 unique T-cell donors; data represented as mean ± SD. Significance noted is in comparison to NT (black asterisks) and/or CCR⁺ cells (green asterisks) or as noted. NT vs. CCR⁺ comparison was nonsignificant in all instances. E, Bioluminescence-based cytotoxicity assays performed using K562, K562.CD123, MV-4-11, and Molm-13 stably expressing ffLuc; n = 3 to 5 donors. For B–E, *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Unless noted, comparisons were nonsignificant. NT, nontransduced.

Figure 4

CCR expression does not diminish CAR T-cell antitumor activity. A, Schematic of the xenograft model. On day 0, NSG mice were injected via tail vein with 1 × 10⁶ CD123⁺ MV-4-11.ffLuc cells. In treatment groups, 2 × 10⁶ T cells were administered on day 7. Cohorts: unmodified (NT), CD28.ζ CAR, and CD28.ζ CAR + IL7Rα CCR-T cells. B, Leukemia proliferation was monitored with BLI, and whole-animal radiance was recorded (photons/second/cm²/sr). Comparison of radiance in each group was performed following log transformation of each measured value. C, Kaplan–Meier survival analysis of MV-4-11 xenografts (n = 9 mice per condition; two independent experiments; median survival: NT: 45.5 days; CAR: 93.5 days; CAR + CCR: undefined; ***, P < 0.001; ****, P < 0.0001). NSG, NOD/SCIDγ; NT, nontransduced.

Figure 5

Hinge modifications do not improve antigen specificity of CD123-targeted chimeric IL7Rα. A, Amino acid sequences of hinge regions used in functional studies. B, Transduction efficiency measured using his-tagged rhCD123 binding of engineered receptors and flow cytometry. n = 3 unique T-cell donors. C, Quantification of phosphorylated STAT5 percentage in CCR-T cells measured with and without activation on the immobilized rhCD123 target. n = 3 unique T-cell donors, comparison between pSTAT5% in cytokine-starved cells with or without rhCD123 activation; *, P < 0.05. D, Representative Western blot testing IgG4 vs. IH STAT5 activation with and without immobilized rhCD123. NT, nontransduced.

Figure 6

Single-chain IL7Rα signaling stimulates STAT5 activation in CAR-T cells. A, Schema of the compared T-cell conditions. A: CAR and sIL7; B: CAR with the distal IL7Rα domain; C: CAR with the proximal IL7Rα domain. B, Levels of STAT5 phosphorylation before and after exposure to plate-bound recombinant antigen (rhCD123) for 30 minutes, measured with a TR-FRET assay. Comparison between conditions with and without rhCD123 stimulation. n = 3 donors, performed in duplicate. C, Relative ECAR and (D) relative OCR of T cells measured after 24-hour rhCD123 stimulation in cytokine-free media. n = 4 individual donors tested in five replicates per assay. For each donor, normalization of all values to the baseline (NT control) was performed. E, Cytotoxic activity measured by luminescence-based assay. CD123-negative K562 cell line and NT T cells included as controls. n = 4 T-cell donors, E:T. Comparison of cytotoxicity of engineered T cells against all CD123-expressing targets with that from NT T cells was very significant (P < 0.0001 in K562.CD123, MV-4-11, and Molm-13). Differences between engineered T cells noted only with MV-4-11 target. ****, P < 0.0001; **, P < 0.01. F, Long-term cytotoxicity measured with six 48-hour serial stimulations with indicated NLS eGFP-modified AML cell lines with initial plating at 1:1 E:T (n = 4 individual donors in technical triplicate each normalized to timepoint 0). G, Half-life decay of each target cell line calculated at each stimulation. Shaded bars, 95% confidence intervals for decay calculated at each timepoint. NT, nontransduced.

Figure 7

Single-chain proximal IL7Rα CAR-T cells have decreased in vivo antileukemia activity. A, Schematic of the xenograft model. On day 0, NSG mice were injected via tail vein with 1 × 10⁶ CD123⁺ MV-4-11.ffLuc cells. In treatment groups, 2 × 10⁶ CAR⁺ T cells were administered via tail vein on day 7. Cohorts: unmodified (NT, n = 2), CD28.ζ + sIL7 (n = 6), CD28.IL7R.ζ (n = 6), and IL7R.CD28.ζ (n = 5) T cells. Untreated mice (n = 3) served as controls. B, Representative mouse pictures of leukemia proliferation monitored with BLI. C, Radiance. Dotted lines, individual mice; solid line, median of each cohort. D, Kaplan–Meier survival analysis; comprehensive curve comparison; P < 0.0001; NT vs. CD28.ζ_sIL7, vs. CD28.IL7R.ζ, and vs. IL7R.CD28.ζ: all **, P < 0.01 individual comparisons. CD28.IL7R.ζ vs. IL7R.CD28.ζ: *, P < 0.01. E, AML and (F). T-cell counts detected in peripheral blood of mice and measured by flow cytometry. No significant differences found between groups. NSG, NOD/SCIDγ; NT, nontransduced.