Antitumor T-cell Homeostatic Activation Is Uncoupled from Homeostatic Inhibition by Checkpoint Blockade

Homeostatic activation is coupled to homeostatic inhibition in a murine BMT model. BALB/c splenocytes were intravenously injected into the tails of congenic, syngeneic recipients that were treated with 9 Gy total-body irradiation (BMT; ) or nonirradiated, naïve mouse (No Rx; ). Seven days after transfer, splenocytes of recipient mice were harvested, and the expression of (A) homeostatic proliferation and activation markers Ki-67, CD44, and IL2Rβ and (B) checkpoint receptors CTLA4 and PD-1 on CD45.1⁺CD8⁺ T cells was assessed by flow cytometry. Unpaired t test was performed. Double checkpoint blockade treatment + transfer into nonirradiated recipient mice = antibody only (dCB only; ); double checkpoint blockade treatment + transfer into irradiated recipient mice = immunotransplant (IT; ). Additionally, BALB/c donors were treated with anti-CTLA4/anti–PD-1 checkpoint blockade antibodies. After treatment, donor splenocytes and bone marrow were intravenously injected into the tails of congenic recipients that had been irradiated (IT) or not (dCB only). Recipients were also treated with double checkpoint blockade. Seven days after transfer, recipient splenocytes were magnet-sorted and enriched for CD8⁺ T cells and then cultured with 500 ng of plate-bound anti-CD3ε and 20 μg PD-L1/IgG and 2 μg/mL of soluble anti-CD28 for 12 or 24 hours. Activation of CD45.1⁺CD8⁺ T cells was assessed by intracellular IFNγ and IL2 by flow cytometry (C; left). n = 6–8 mice per group from two independent experiments. Unpaired t test was performed. Similarly, magnet-sorted CD8⁺ T cell–enriched splenocytes were cultured with 300 ng of plate-bound anti-CD3ε and plate-bound 500 ng of CD80/IgG for 24 hours. Activation of CD45.1⁺CD8⁺ T cells was assessed by extracellular CD69 (C; right). n = 4–6 mice per group from two independent experiments. One-way ANOVA and multiple comparisons post-test were performed compared with BMT⁺CD80. Anti-GFP CD8⁺ T-cell donor mice (background F1) were treated with double checkpoint blockade antibodies and transferred to congenic WT recipient mice that had or had not been irradiated. Seven days after transfer, splenocytes were harvested and stimulated with a variant of A20 lymphoma cell line that expressed GFP. D, Intracellular IFNγ production was assessed by flow cytometry. Representative data are shown. n = 6–8 mice from two independent experiments. One-way ANOVA and multiple comparisons post-test performed compared with IT. IT was compared with “dCB only” and “No Rx”; ***, P < 0.001. Data, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

IT induces tumor-specific immune responses, reverses T-cell exhaustion, and treats lymphoid and solid tumors. Donor mice were inoculated with the murine B-cell lymphoma A20 cell line for 7 days and then treated with anti-CTLA4/anti–PD-1 checkpoint blockade. After treatment, splenocytes and bone marrow were transferred to TBI syngeneic recipient mice (IT) that bore the same A20 tumor; recipient mice were also treated with dCB. As controls, untreated splenocytes were transferred to irradiated recipients (BMT). Further, tumor-bearing mice were treated with double checkpoint blockade alone (dCB only) or given no treatment (No Rx). A, Tumor growth curve and survival curve of the mice from each group. n = 11 mice per group. Two-way ANOVA was performed and compared with IT. Mice from each group were bled 11 days after transfer or treatment start. Lymphocytes were isolated and restimulated in the presence of irradiated A20 tumor cells in vitro for 48 hours. Activation of CD8⁺ T cells was assessed by intracellular IFNγ and extracellular CD44 by flow cytometry (B). n = 6 mice per group from two independent experiments. One-way ANOVA and multiple comparisons post-test were performed compared with IT + A20. IT + A20 was compared with dCB only + A20 and No Rx + A20; ****, P < 0.0001. Tumors from No Rx (left) and IT (right) were harvested 7 days after transfer, and tumor infiltrate was assessed by immunofluorescence (C; top). The activation status of tumor-infiltrating cells was assessed by mass cytometry (CyTOF). C, Bottom, normalized number of tumor-infiltrating PD-1^hiCD8⁺T cells. Fold change = (percent log ratio)/(average percent log ratio of No Rx). One-way ANOVA and multiple comparisons post-test were performed compared with IT. D, Activation status of PD-1^hi CD8⁺T cells is shown. Fold change = (fold change)/(average fold change of No Rx). One-way ANOVA and multiple comparisons post-test were performed compared with IT. Mice bearing murine melanoma B16, lung carcinoma KLN205, and T-cell lymphoma EL4 tumors were treated as previously described. NK, natural killer. E, Graph showing tumor growth curve and survival curve. n = 8–12 mice per group. Two-way ANOVA and multiple comparisons posttest were performed compared with IT. Data, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

IT antitumor immunity is CD8-, IFNγ-, and lymphodepletion-dependent. All therapies (No Rx, dCB only, BMT, and IT) were performed as previously described with the exception that, at the time of transfer, all recipient mice were not tumor bearing. All recipient mice were tumor challenged with A20 2 days post-transfer or post–treatment initiation. A, Tumor growth curve and survival curve of the mice from each group. n = 9–11 mice per group. Two-way ANOVA was performed and compared with IT. Key effectors of IT antitumor immunity were assessed by treating IT-treated recipients with depleting CD8 (IT + αCD8), CD4 (IT + αCD4), and IFNγ (IT + αIFNγ) antibodies before and after splenocyte transfer. B, Graph showing tumor growth curve and survival curve. n = 9–12 mice per group. Two-way ANOVA was performed and multiple comparisons post-test was performed compared with IT + IgG control. Additionally, IT mice were treated as described, but the donor lacked the A20 tumor (donor no tumor + IT), or the recipient mouse was not irradiated (IT+ 0 Gy) or was lymphodepleted with chemotherapy (fludarabine/cyclcophosphamide) instead of TBI (IT + Flu/Cy). C, Graph showing tumor growth curve and survival curve. n = 8–12 mice per group. Two-way ANOVA and multiple comparisons post-test were performed compared with IT + Flu/Cy. After IT-treated mice had cleared the A20 tumor, recipient mice were rechallenged with the same A20 tumor cells or 4T1 breast tumor cells. D, Graph showing tumor growth curve and survival curve. n = 6–8 mice per group. Two-way ANOVA and multiple comparisons post-test was performed compared with IT + A20. Data, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

γ_c Cytokine signaling is amplified by IT and necessary for antitumor immunity. Serum was isolated from mice of each therapy group, and (A) cytokine levels were assessed by bead-based multiplex assay using a Luminex platform. One-way ANOVA was performed and compared with IT. n = 6–9 mice per group from two independent experiments. BALB/c donor mice were treated with anti-CTLA4/anti–PD-1 checkpoint blockade antibodies, and splenocytes were harvested and intravenously injected into congenic recipient mice that had received total-body irradiation (modeling IT). As controls, anti-CTLA4/anti–PD-1 checkpoint blockade–treated congenic splenocytes were also intravenously injected into recipients that had not received TBI (modeling dCB only); additionally, splenocytes of untreated donor mice were transferred into recipients that had not received TBI (modeling No Rx) or TBI recipients (modeling BMT). Three days after transfer, recipient splenocytes were harvested, and (B) the expression of γ_c receptor subunits was assessed by flow cytometry. One-way ANOVA was performed and compared with IT. n = 8 mice per group from two independent experiments. Splenocytes from each group were stimulated in vitro with 100 ng of recombinant IL2, IL7, or IL15 for 20 minutes. Activation and downstream signaling was measured by (C) nuclear phospho-STAT5 (Y694) levels assessed by flow cytometry. One-way ANOVA and multiple comparisons post-test were performed compared with IT. n = 8 mice per group from two independent experiments. Splenocytes from each group were labeled with CFSE and treated in vitro with soluble 100 ng of anti-CD3e and 100 ng of recombinant IL2, IL7, or IL15 for 48 hours. Proliferation was assessed by CFSE dilution (D). One-way ANOVA was performed and multiple comparisons post-test was performed compared with IT. n = 6 mice per group from two independent experiments. GFP-specific CD8 donor mice were treated with anti-CTLA4/anti–PD-1 double checkpoint blockade antibodies and intravenously injected into congenic recipients that had been irradiated. Splenocytes were harvested 7 days after transfer and stimulated in vitro with GFP-expressing A20 cells in the presence of DMSO, JAK3-specific inhibitor tofacitinib (iJAK3), or a mix of 100 ng recombinant IL2, IL7, and IL15. Activation was assessed by intracellular IFNγ production (E). n = 7–8 mice per group from two independent experiments. Donor mice whose lymphocytes had been engineered to lack IL7Rα (B6 IL7Rα KO IT) or IL15Rα (B6.129 IL15Rα KO), or WT mice were inoculated with T-cell lymphoma EL4 tumor and then treated with anti-CTLA4/anti–PD-1 checkpoint blockade antibodies. After treatment, splenocytes and bone marrow were transferred to TBI syngeneic recipient mice that bore the same EL4 tumor. F, Graph showing tumor growth and survival. Two-way ANOVA and multiple comparisons post-test were performed compared with IT. KO, knockout. Data, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

BMT increases γ_c receptor and signaling on donor T cells in lymphoma patients. A, The expression of γ_c receptor subunits was assessed by flow cytometry on patients' CD8⁺ T cells PRE and POST transplant. n = 8 patients. Paired t test was performed. PRE and POST PBMCs were stimulated in vitro with recombinant IL2, IL7, and IL15. Activation and downstream signaling were determined by (B) nuclear phospho-STAT5 (Y694) levels, and proliferative capability was assessed by (C) Ki-67 levels gated on CD8⁺ T cells. n = 8 patients. Paired t test was performed. The activation status of PRE and POST PBMCs was assessed by CyTOF. D, Graph (on left) showing the expression of several markers on patients' CD8⁺ T cells as measured by CyTOF. For comparison, the activation status of splenocytes in No Rx- and BMT-treated mice were assessed by CyTOF (on right). N = 4 patients. Paired t test was performed. N = 3 mice per group. Unpaired t test was performed. Fold change of PRE * = (Raw fold change of POST)/(Raw fold change of PRE per person). Fold change of No Rx * = (Fold change of BMT)/(Average fold change of No Rx). *, Except for CLTA4, PD-1, and CD27. Fold change = [(Count of BMT CD8⁺ T cells positive)/(Total CD8⁺ T-cell per person or per mouse)]/[(Count of PRE or No Rx CD8⁺ T cells positive)/(Total CD8⁺ T-cell per person or per mouse)]. Data, mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.