Reversal of NK-cell exhaustion in advanced melanoma by Tim-3 blockade

Abstract

The immunoregulatory protein T-cell immunoglobulin- and mucin-domain-containing molecule-3 (Tim-3) mediates T-cell exhaustion and contributes to the suppression of immune responses in both viral infections and tumors. Tim-3 blockade reverses the exhausted phenotype of CD4+ and CD8+ T cells in several chronic diseases, including melanoma. Interestingly, natural killer (NK) cells constitutively express Tim-3; however, the role of Tim-3 in modulating the function of these innate effector cells remains unclear, particularly in human diseases. In this study, we compared the function of Tim-3 in NK cells from healthy donors and patients with metastatic melanoma. NK cells from the latter were functionally impaired/exhausted, and Tim-3 blockade reversed this exhausted phenotype. Moreover, Tim-3 expression levels were correlated with the stage of the disease and poor prognostic factors. These data indicate that Tim-3 can function as an NK-cell exhaustion marker in advanced melanoma and support the development of Tim-3-targeted therapies to restore antitumor immunity.

Figure 1. MD NK cells up-regulate inhibitory receptors and down-regulate activating receptors

(A) Graphs compare the expression of the inhibitory receptors KIR3DL1 (HD n=18; MD n=8) and KIR2DL3 (HD n=26, MD n=12) and the activating receptors NKG2D (HD n=25; MD n=12), NKp46 (HD n=20, MD n=11) and DNAM-1 (HD n=10; MD n=10), in peripheral blood NK cells purified from healthy donors and from melanoma patients. On the right panel, plots depicting the expression of KIR3DL1, KIR2DL3, NKG2D, NKp46 and DNAM-1 according to CD56 expression in NK cells from a representative healthy donor and a representative melanoma patient. All experiments were performed in duplicate.

Figure 2. MD NK cells are functionally impaired/exhausted

(A) Freshly purified NK cells (HD n=12; MD n=5) were stimulated with 200U/ml of IL-2. Expression of IL-2R (α chain), IFNγ production, and cytotoxicty were monitored every two days over 6 days (day 0, 2, 4 and 6) by flow cytometry. (B) The percentage of Lamp-1⁺ NK cells from healthy (n=30) and melanoma donors (n=12) after a cytotoxic assay is shown using K562 cells as target cells (upper left panel). The percentage of IFNγ⁺ NK cells from healthy (n=22) and melanoma donors (n=9) is shown after 4h stimulation with IL-12 (middle left panel). The percentage of proliferating NK cells from healthy (n=10) and melanoma donors (n=7) is shown after 6 days of culture in presence of 200U/ml of IL-2 (lower left panel). On the right panel of each graph, plots depicting the expression of Lamp-1, IFNγ, and CFSE in NK cells purified from a representative healthy donor and a representative melanoma patient are shown. (C) Graphs representing the MFI of T-bet and Eomes on NK cells purified from healthy (n=19) and melanoma donors (n=14). Representative plots are shown (Isotype control: black; HD: unfilled; MD: gray). All experiments were performed in duplicate.

Figure 3. Tim-3 is up-regulated in MD NK cells

(A) Graph comparing Tim-3 expression in NK cells from healthy donors (HD; n=45) and melanoma patients (MD; n=41). Represented as the percentage of Tim-3⁺ cells (left panel) and the MFI of the Tim-3⁺ population (right panel). (B) The graphs show the percentage (left panel) and MFI (right panel) of Tim-3⁺ NK cells from healthy donors (n=30) and patients with melanoma stage I (n=47), II (n=18) and III/IV (n=18). All experiments were performed in duplicate.

Figure 4. Tim-3 engagement inhibits NK cell functions

(A) The percentage of LAMP-1⁺ (n=8) and (B) the MFI of IFNγ⁺ cells (n=6) of NK cells from melanoma donors pre-incubated with IgG-coated beads or anti-Tim-3-coated beads for 2h prior to evaluating the cytotoxic function or IFNγ production. (C) Reverse-ADCC assay using FcR⁺ P815 cells. NK cells from melanoma or healthy donors were co-cultured with P815 cells and different antibodies were added to the reaction: anti-Tim-3, anti-CD94 (negative control) or anti-CD16 (positive control). Data were normalized to the values obtained for the condition: (A and B) with IgG-coated beads (100%); (C) with no antibody. All experiments were performed in duplicate.

Figure 5. NK cell exhaustion can be reversed by Tim-3 blockade

(A) Graph represents the expression of LAMP-1 (n=15; MFI) in NK cells from melanoma donors incubated with 10 or 20μg/ml of soluble Tim-3 blocking antibody 1h before the cytotoxic assay. (B) Plot shows the percentage of CFSE⁺7AAD⁺ K562 cells (% of killed K562 cells) in the presence of 10μg/ml of two different Tim-3 blocking antibodies (clone 2E2 or R&D #AF2365) 1h before the killing assay (n=9 MD). Graphs represent the expression of (C) IFNγ (n=12; MFI) and (D) the percentage of proliferating cells (n=10; %) in NK cells from melanoma donors incubated with 10 or 20μg/ml of soluble Tim-3 blocking antibody 1h before the functional assays. (A, C and D) Plots depicting the expression of LAMP-1, IFNγ production and proliferation with (right panel) and without (middle panel) Tim-3 blockade from a representative melanoma patient. Data were normalized to the values obtained for the condition without blocking antibody. All experiments were performed in duplicate.

Figure 6. Soluble Tim-3 blocking antibody induces internalization of Tim-3 and up-regulation of IL-2R (α and γ chains)

(A) Plot depicting the decrease of Tim-3 expression in the membrane of NK cells after 1h treatment with soluble Tim-3 blocking antibody. (B) Graph represents the MFI of NK cells positive for the PE-conjugated, anti-mouse IgG antibody, with or without permeabilization (n=6). (C) The previous experiment was repeated also including an isotype control. Graph shows the percentage of cells positive for the PE-conjugated, anti-mouse IgG antibody after 1h of incubation with antibodies (n=6). (D) The represents the MFI of Lamp-1⁺ NK cells from healthy donors (n=6) after a cytotoxic assay with K562 cells. (Left panel) NK cells were pre-incubated (2h) with anti-Tim-3-coated beads or IgG-coated beads. (Right panel) Soluble Tim-3 blocking antibody was added 1h before crosslinking with anti-Tim-3-coated beads as described previously. Data were normalized to the values obtained for the condition with beads alone. (E) The plots show the expression of IL-2R α and γ chains (% of positive cells) in NK cells untreated or after 1h of treatment with 10μg/ml of soluble Tim-3 blocking antibody (n=7). Data were normalized to the values obtained for the condition without blocking antibody. (F) The graph depicts the percentage of Lamp-1⁺ NK cells after two days of culture with 200U/mL of IL-2, untreated or treated with blocking antibody for α, β or β chain (n=5). All experiments were performed in duplicate.