Dendritic cell editing by activated natural killer cells results in a more protective cancer-specific immune response

Abstract

Over the last decade, several studies have extensively reported that activated natural killer (NK) cells can kill autologous immature dendritic cells (DCs) in vitro, whereas they spare fully activated DCs. This led to the proposal that activated NK cells might select a more immunogenic subset of DCs during a protective immune response. However, there is no demonstration that autologous DC killing by NK cells is an event occurring in vivo and, consequently, the functional relevance of this killing remains elusive. Here we report that a significant decrease of CD11c(+) DCs was observed in draining lymph nodes of mice inoculated with MHC-devoid cells as NK cell targets able to induce NK cell activation. This in vivo DC editing by NK cells was perforin-dependent and it was functionally relevant, since residual lymph node DCs displayed an improved capability to induce T cell proliferation. In addition, in a model of anti-cancer vaccination, the administration of MHC-devoid cells together with tumor cells increased the number of tumor-specific CTLs and resulted in a significant increase in survival of mice upon challenge with a lethal dose of tumor cells. Depletion of NK cells or the use of perforin knockout mice strongly decreased the tumor-specific CTL expansion and its protective role against tumor cell challenge. As a whole, our data support the hypothesis that NK cell-mediated DC killing takes place in vivo and is able to promote expansion of cancer-specific CTLs. Our results also indicate that cancer vaccines could be improved by strategies aimed at activating NK cells.

Figure 1. Subcutaneous administration of MHC-negative cells results in an NK cell-dependent decrease of CD11c^bright DCs in the draining lymph nodes.

A: representative analyses of mononuclear cells isolated from either draining or controlateral (Control LN) lymph nodes of mice depleted or not of NK cells (Draining LN + anti-asialo GM1). B: NK cells are efficiently depleted in mice upon administration of anti-asialo GM1 mAbs. C: the percentage (left) and the absolute number (right) of DCs among mononuclear cells isolated from lymph nodes. Bars represents mean values and SD of five independent experiments (three mice per group). ** =  p < 0,001; * =  p < 0,005.

Figure 2. Peripheral activation of NK cells results in the selection of lymph node DCs with higher T cell activating capabilites.

Draining lymph node DCs of mice injected subcutaneously with MHC^negcells were sorted and cultured in the presence of allogeneic splenocytes previously labeled with CFSE. A: Splenocytes were cultured alone (No stimulus) or with 10% of highly purified lymph node DCs from mice NK cell depleted (10% DC draining LN + anti-asialo GM1) or control (10% DC draining LN) before administration of MHC-negative target cells. CFSE dilution of the splenocytes at 6 days of culture is shown. NK cell depletion compromises the induction of proliferation by LN DCs. Data are representative of four independent experiments summarized in panel B: ▪ = 10% DC draining LN; ▴ = 10% DC draining LN + anti-asialo GM1. * =  p<0,02.

Figure 3. DC editing by NK cells promotes antigen specific CTL expansion.

A: Mice were inoculated with immunogenic TS/A cells, TS/A cells mixed with MHC-devoid cells (YAC-1) or YAC-1 cells alone. In a group of mice, anti-asialo GM1 mAbs were administered i.p. 48 h before administration of cell vaccines to deplete NK cells. After 21 days, splenocytes were restimulated with either TS/A cells, the TS/A MHC class I-restricted immunodominant peptide AH1 or with YAC-1 cells and the frequencies of IFNγ-producing cells were determined by ELISPOT assay. A significant increase of antigen-specific IFNγ-producing cells was detectable in mice vaccinated with TS/A mixed with YAC-1 cells (TS/A + YAC-1). The increase in antigen-specific CTL was abrogated when mice had been depleted of NK cells before vaccination (TS/A + YAC-1 + anti asialo GM1). B: Similar experiments were performed in perforin KO mice (pfn ^−/−) and in parallel with wild type (WT) mice. A significant lower number of TS/A tumor-specific CTL was induced by vaccination in perforin KO mice when compared to wild type mice. Bars represent mean values and SEM of results obtained in three independent experiments (three mice per group). ** =  p < 0,001; * =  p < 0,005.

Figure 4. DC editing by NK cells generates a more protective immune response during cancer cell vaccination.

Mice were vaccinated with immunogenic tumor cells alone (TS/A), tumor cells mixed with MHC-negative ells (TS/A + YAC-1) or MHC-negative cells alone (YAC-1) (five mice per group). In a group of mice, anti-asialo GM1 mAbs were administered i.p. 48 h before administration of cell vaccines to deplete NK cells. After 21 days, mice were challenged with a lethal dose of tumor cells and monitored for tumor growth twice a week for two months. Mice vaccinated with tumor cells mixed with MHC-negative cells (TS/A + YAC-1) displayed a delay in tumor growth and a striking survival to TS/A cancer cell challenge. This protective effect was abrogated when mice were depleted of NK cells (TS/A + YAC-1 + anti-asialo GM1).