Helper activity of natural killer cells during the dendritic cell-mediated induction of melanoma-specific cytotoxic T cells

Abstract

Natural killer (NK) cells have been shown to mediate important immunoregulatory "helper" functions in addition to their cytolytic activity. In particular, NK cells are capable of preventing maturation-related dendritic cell (DC) "exhaustion," inducing the development of "type-1 polarized" mature DCs (DC1) with an enhanced ability to produce interleukin (IL)-12p70, a factor essential for type-1 immunity and effective anticancer responses. Here we show that the NK cell-mediated type-1 polarization of DCs can be applied in the context of patients with advanced cancer to enhance the efficacy of DCs in inducing tumor-specific cytotoxic T lymphocytes. NK cells isolated from patients with late-stage (stage III and IV) melanoma responded with high interferon-γ production and the induction of type-1-polarized DCs on exposure to defined combinations of stimulatory agents, including interferon-α and IL-18. The resulting DCs showed strongly-enhanced IL-12p70 production on subsequent T-cell interaction compared with immature DCs (average of 19-fold enhancement) and nonpolarized IL-1β/TNF-α/IL-6/PGE(2)-matured "standard" DCs (average of 215-fold enhancement). Additional inclusion of polyinosinic: polycytidylic acid during NK-DC cocultures optimized the expression of CD80, CD86, CD40, and HLA-DR on the resulting (NK)DC1, increased their CCR7-mediated migratory responsiveness to the lymph node-associated chemokine CCL21, and further enhanced their IL-12-producing capacity. When compared in vitro with immature DCs and nonpolarized standard DCs, (NK)DC1 were superior in inducing functional melanoma-specific cytotoxic T lymphocytes capable of recognizing multiple melanoma-associated antigens and killing melanoma cells. These results indicate that the helper function of NK cells can be used in clinical settings to improve the effectiveness of DC-based cancer vaccines.

Figure 1. Two-signal activation requirement for IFNγ production by NK cells isolated from late-stage melanoma patients

Negatively-isolated NK cells (1×10⁵ cells/well) were incubated for 24 h in the presence of the indicated combinations of activating factors. Supernatants were subsequently assayed by ELISA for the presence of IFNγ. A, NK cell production of IFNγ in response to stimulation with IFNα (1000 IU/ml) and/or IL-18 (1 µg/ml) (top); IFNα and/or exposure to NK-sensitive K562 leukemia tumor cells (middle); or IFNα and/or exposure to antibody (R24)-opsonized, nominally NK-resistant FEM-X melanoma cells (bottom). The data shown represents one of six independent experiments, which all yielded similar results. Data recorded as the mean (± SD) of triplicate cultures. ***p<0.001 compared to all groups. B, Comparison of IFNγ production by NK cells derived from six melanoma patients or six healthy donors in response to stimulation with IFNα and IL-18. Data recorded as the mean of triplicate cultures for each patient or healthy donor. ns: p>0.05. C, Comparison of IFNγ production by unstimulated or IFNα/IL-18-stimulated NK cells isolated from individual melanoma patients. Data is presented as the mean of triplicate cultures for each patient (total of 6 patients). *p<0.05.

Figure 2. Two-signal-activated NK cells from late-stage melanoma patients stably induce DCs with an enhanced capacity to produce IL-12p70

Previously isolated and cyropreserved NK cells (1.5×10⁵ cells) were added to autologous day 6 DCs (2–3×10⁵ cells/well) in the presence of IFNα (1000 IU/ml) and IL-18 (1 µg/ml). After 48 h, the DCs were harvested, plated (2×10⁴ cells/well), and exposed to J558-CD40L to induce IL-12p70 production. IL-12p70 concentrations in 24 h supernatants were determined by ELISA. A, IL-12p70 production by untreated immature DCs (iDCs), DCs treated with the standard cytokine maturation cocktail of TNFα/IL-1β/IL-6/PGE₂ (sDCs), or DCs treated with IL-18/IFNα with or without autologous NK cells. Data recorded as the mean (± SD) of triplicate cultures. Data shown was obtained from one representative experiment of five performed, all yielding similar results. ***p<0.001 compared to all groups. B, IL-12p70 production by DCs treated with the standard cytokine cocktail (sDCs) or autologous NK cells with IL-18/IFNα in direct or transwell-separated co-cultures. Data recorded as the mean (± SD) of triplicate cultures. Data from one representative experiment of two performed, both of which yielded similar results. ***p<0.001 compared to all groups, **p<0.01 compared to sDC group. C, IL-12p70 production by untreated immature DCs (iDCs), DCs treated with the standard cytokine maturation cocktail (sDCs), or DCs treated with autologous NK cells and IL-18/IFNα. Data recorded as the mean of triplicate cultures for each patient (total of 5 patients). **p<0.01. D, IL-12p70 production by differentially-matured DCs stimulated with CD40L directly after harvesting (top) or after an additional 24 h of culture in the absence of maturation factors (bottom). Data presented as the mean (± SD) of triplicate cultures for each patient. Data from one representative experiment of three performed, all of which yielded similar results. ***p<0.001 compared to all groups.

Figure 3. Inclusion of poly-I:C in NK-DC co-cultures results in _NKDC1s with optimal surface expression of T cell-activating molecules and CCR7 and optimal ability to produce IL-12p70

Surface expression (open histograms) of CD80, CD86, CD40, HLA-DR, and CCR7 on untreated immature DCs (iDCs), DCs treated with the standard cytokine maturation cocktail (sDCs), or DCs treated with autologous NK cells and IL-18/IFNα with or without poly-I:C. Surface expression was analyzed directly after DC harvesting or after an additional 24 h of culture in the absence of maturation factors. Gray histograms represent isotype controls. Inset numbers represent fold MFI increase over isotype controls. Right: The corresponding IL-12p70 production after J558-CD40L-stimulation. Data from one representative experiment of three performed, all of which yielded similar results. ***p<0.001 for _NKDC1s (NK/IL-18/IFNα/poly-I:C DCs) compared to iDCs, sDCs, and NK/IL-18/IFNα DCs, or for NK/IL-18/IFNα DCs compared to iDCs and sDCs.

Figure 4. _NKDC1s are efficient inducers of melanoma-specific CTLs

Immature (i)DCs, sDCs, and _NKDC1s from HLA-A2⁺ stage III and stage IV melanoma patients were pulsed with MHC Class I-restricted melanoma-associated peptides and used to sensitize autologous CD8⁺ T cells. CTLs were assayed on day 24 of culture. A, Frequencies of IFNγ-producing CD8⁺ T cells responsive to T2 cells loaded with individual peptides, as determined by ELISPOT assay. Data recorded as the mean (± SD) of triplicate cultures. Data shown is from one representative experiment of three performed. ***p<0.001, ns: p>0.05. B, Flow cytometric analysis showing percentage of tetramer-positive MART-1-specific CD8⁺ T cells generated through in vitro stimulation with melanoma peptide-pulsed, differentially-activated DCs. Inset numbers represent percentages of CD8⁺ MART-1⁺ cells. Results from one representative experiment of three performed. C, Frequencies of IFNγ-producing CD8⁺ T cells responsive to the relevant (HLA-A2⁺) and irrelevant (HLA-A2⁻) target melanoma cell lines FEM-X and MEL-397, respectively, as determined by ELISPOT assay. Blockade with the W6/32 pan-MHC Class I-neutralizing antibody was used to demonstrate the MHC Class I-dependence of the T cell recognition. Data recorded as the mean (± SD) of triplicate cultures. Data shown is from one representative experiment of three performed. ***p<0.001, ns: p>0.05. D, Antigen-specific cytotoxic activity of CTLs induced by _NKDC1s, iDCs, or sDCs against FEM-X (HLA-A2⁺) and MEL-397 (HLA-A2⁻) melanoma cell lines, as determined by standard 4 h ⁵¹Cr-release assay. Data recorded as the mean (± SD) of triplicate cultures. Similar data were obtained in two additional experiments. ***p<0.001, **p<0.01 compared to all groups.