Antitumor NK activation induced by the Toll-like receptor 3-TICAM-1 (TRIF) pathway in myeloid dendritic cells

Abstract

Myeloid dendritic cells (mDCs) recognize and respond to polyI:C, an analog of dsRNA, by endosomal Toll-like receptor (TLR) 3 and cytoplasmic receptors. Natural killer (NK) cells are activated in vivo by the administration of polyI:C to mice and in vivo are reciprocally activated by mDCs, although the molecular mechanisms are as yet undetermined. Here, we show that the TLR adaptor TICAM-1 (TRIF) participates in mDC-derived antitumor NK activation. In a syngeneic mouse tumor implant model (C57BL/6 vs. B16 melanoma with low H-2 expresser), i.p. administration of polyI:C led to the retardation of tumor growth, an effect relied on by NK activation. This NK-dependent tumor regression did not occur in TICAM-1(-/-) or IFNAR(-/-) mice, whereas a normal NK antitumor response was induced in PKR(-/-), MyD88(-/-), IFN-beta(-/-), and wild-type mice. IFNAR was a prerequisite for the induction of IFN-alpha/beta and TLR3. The lack of TICAM-1 did not affect IFN production but resulted in unresponsiveness to IL-12 production, mDC maturation, and polyI:C-mediated NK-antitumor activity. This NK activation required NK-mDC contact but not IL-12 function in in vivo transwell analysis. Implanted tumor growth in IFNAR(-/-) mice was retarded by adoptively transferring polyI:C-treated TICACM-1-positive mDCs but not TICAM-1(-/-) mDCs. Thus, TICAM-1 in mDCs critically facilitated mDC-NK contact and activation of antitumor NK, resulting in the regression of low MHC-expressing tumors.

Fig. 1.

PolyI:C induces NK-mediated MHC class I-negative tumor regression. (A) Establishment of tumor-implantation model for evaluation of polyI:C antitumor activity. PolyI:C (250 μg i.p. injected twice a week) caused antitumor effect on B16D8 cells (SI Fig. 8) implanted into C57BL/6 mice. Arrow indicates the start point of polyI:C treatment (tumor average size >0.8 cm³). (B and C) NK is an effector for poly(I:C)-mediated antitumor activity. Mice were challenged with B16D8 cells and i.p. injected with anti-NK1.1 (B) or anti-CD8β (C) ascites according to the schedule of polyI:C treatment (see Materials and Methods). Antibody and polyI:C treatment was repeated twice a week from day 10. One of the two similar experiments is shown.

Fig. 2.

Antitumor NK cells isolated from polyI:C-injected C57BL/6 mice. (A) In vitro B16 cytotoxicity by splenic NK cells of polyI:C-treated mice. C57BL/6 mice were administered with polyI:C (or control saline only) i.p. After 18 h, NK cells were isolated with MACS-negative selection beads and NK cytotoxicity against B16 cells was measured by ⁵¹Cr release assay. (B) YAC-1 cytotoxicity of splenocytes was tested as in A. YAC-1 is known to be targets for NK. (C) NK activation depends on TICAM-1 and IFNAR. YAC-1 cytotoxicity of splenocytes from gene-disrupted mice stimulated with polyI:C. Indicated gene-disrupted mice were treated with polyI:C. After 18 h, splenocytes were prepared and their cytotoxicity against YAC-1 was measured by ⁵¹Cr release assay. NK activation by polyI:C was impaired in splenocytes derived from IFNAR^−/− or TICAM-1^−/− mice. Effector-to-target cell ratio (E/T ratio) = 100. One of the three independent experiments is shown.

Fig. 3.

Antitumor activity of PolyI:C depends on TICAM-1 and IFNAR in vivo. Antitumor effect of polyI:C on various KO mice were evaluated by using in vivo mouse tumor implant model. B16 tumor cells were inoculated on day 0. Arrow indicates the start point of polyI:C administration. Each point represents tumor size average ± SE (n = 4–6). The disrupted genes in mice are shown over the graphs.

Fig. 4.

TICAM-1 in mDCs is essential for antitumor activity of polyI:C. (A) Adoptive transfer of TICAM-1-transduced mDCs confers tumor growth suppression in mice. The lentiviral system was used for gene transfection into mDCs (SI Fig. 12). mDCs were prepared from bone marrow cells and transfected with TICAM-1 (TICAM-1 DC) or empty vector (vector DC). Wild-type mice were implanted with B16 cells on day 0. On day 12, 16, and 19, the mice with tumor burden were i.p. injected with TICAM-1-expressing mDCs (1 × 10⁶ cells). Retardation of tumor growth was measured in the mice of control, with vector-containing mDCs or with TICAM-1-expressing mDCs. One of the three independent experiments is shown. (B) B16 killing by NK cells cocultured with mDCs. TICAM-1-expressing mDCs (TICAM-1 DC) and vector DC were prepared as in A. Poly I:C-stimulated mDCs (polyI:C DC) were prepared by incubation of mDCs with poly I:C for 4 h. The mDCs were cocultured with NK cells (DC:NK = 1:2) for 24 h. NK cytotoxicity against B16 was measured by ⁵¹Cr release assay. (C) TICAM-1 in mDCs is required for polyI:C-mediated tumor regression. mDCs were prepared from wild-type and TICAM-1 KO mice. mDCs (3 × 10⁶ cells) either from wild-type (WT DC) or TICAM-1 KO mice (TICAM-1 KO DC) and polyI:C (250 μg) were injected into the peritoneal cavity of IFNAR^−/− mice, which had the tumor burden. Growth retardation of implanted tumor in response to polyI:C was measured in the mDC-injected mice. The arrows indicate the time points at which the mDCs were administered.

Fig. 5.

Production of IFN-α and IL12p40 in response to poly I:C stimulation in vivo and in vitro. The concentrations of IFN-α and IL12p40 in cultured sup (A and B) of mDCs and in mouse serum (C and D) were determined by ELISA. mDCs were prepared from each KO mice and stimulated with 50 μg of polyI:C. After 24 h, culture supernatants were collected and the levels of cytokines (A and B) were measured. In other experiments, each indicated KO mouse was i.p. injected with polyI:C. After 16 h, blood was directly drawn from heart and clotted to collect serum (C and D). It is notable that IFNAR^−/− mice failed to produce IFN-α in vivo and in vitro, but TICAM-1^−/− mice retained IFN-producing capacity. Gray bars, controls with no stimulation; white bars, cells/mice stimulated with polyI:C. Figure represents one of four experiments.

Fig. 6.

mDCs activate NK cells via cell-cell contact, which depends on the TICAM-1 pathway. mDCs prepared from wild-type or TICAM-1 KO mice were incubated with poly I:C for 24 h, and NK cells were added to the culture at an mDC/NK ratio of 1:2. After 24 h, NK cells were incubated with B16 cells for 5 h at the indicated effector-to-target cell ratio (E/T ratio). Cytotoxicity of NK against B16 cells was examined by ⁵¹Cr release assay. (A and B) NK cells were cocultured with mDCs in the presence of 5 μg/ml anti-IL12 antibody (A) or in the transwell system (B). (C) NK cells cocultured with mDCs derived from TICAM-1 KO mice.