Prolonged contact with dendritic cells turns lymph node-resident NK cells into anti-tumor effectors

Abstract

Natural killer (NK) cells are critical players against tumors. The outcome of anti-tumor vaccination protocols depends on the efficiency of NK-cell activation, and efforts are constantly made to manipulate them for immunotherapeutic approaches. Thus, a better understanding of NK-cell activation dynamics is needed. NK-cell interactions with accessory cells and trafficking between secondary lymphoid organs and tumoral tissues remain poorly characterized. Here, we show that upon triggering innate immunity with lipopolysaccharide (LPS), NK cells are transiently activated, leave the lymph node, and infiltrate the tumor, delaying its growth. Interestingly, NK cells are not actively recruited at the draining lymph node early after LPS administration, but continue their regular homeostatic turnover. Therefore, NK cells resident in the lymph node at the time of LPS administration become activated and exert anti-tumor functions. NK-cell activation correlates with the establishment of prolonged interactions with dendritic cells (DCs) in lymph nodes, as observed by two-photon microscopy. Close DC and NK-cell contacts are essential for the localized delivery of DC-derived IL-18 to NK cells, a strict requirement in NK-cell activation.

Keywords: dendritic cells; immunosurveillance; innate immunity; natural killer cells; two‐photon microscopy.

© 2016 The Authors. Published under the terms of the CC BY 4.0 license.

Figures

Figure 1. Anti‐tumor effector functions of NK cells after LPS administration

1. Tumor volume at the indicated time points after CT26 cell transplant in mice treated or not with LPS, n (number of animals per group) = 6.

2. Tumor volume in mice depleted of DCs or NK cells before tumor cell transplant and LPS administration, n (number of animals per group) = 5.

3. Tumor volume in SCID mice treated or not with LPS, n (number of animals per group) = 5.

4. Tumor volume at the indicated time point after repetitive LPS injections as depicted in the bottom scheme, n (number of animals per group) = 6.

5. Percent of NK cells within the tumor in mice treated or not with LPS and depleted or not of DCs, n (number of animals per group) ≥ 4.

Data information: NT, control; DT, diphtheria toxin‐treated animals to deplete DCs; LPS, mice treated with LPS. Error bars depict SEM. Statistical significance was determined with a two‐tailed t‐test. NT, untreated.

Figure 2. Effect of LPS on tumor vascularization

1. Explanted tumors at Day 12 after repetitive injections of LPS as depicted in the scheme at the bottom of the panel.

2. Detection of blood vessels by immunohistochemistry using anti‐CD31 antibody (brown staining) in frozen sections of tumors from mice treated or not with LPS as in the scheme in (A). Nuclei are stained with hematoxylin. Two representative sections are shown, magnification 20×. Left‐panel scale bars, 500 μm; right‐panel scale bars, 100 μm. Quantification of vessel mean dimensions and area coverage by vessels is also shown. Three sections from three tumors were analyzed. Error bars depict SEM. Statistical significance was determined with a two‐tailed t‐test. NT, untreated.

Figure 3. Anti‐tumor effector NK cells are generated in the lymph node

1. A, B

   Absolute numbers of IFN‐γ+ and total NK cells in draining lymph nodes at the indicated time points before and after LPS administration in mice treated or not with FTY720 (25 μg/mouse) to reduce the ingress of lymphocytes in the efferent lymphatics.

2. C

   Tumor volume in mice treated or not with LPS and/or FTY720 according to the scheme.

3. D

   Percent of IFN‐γ+ NK cells within the tumor in mice treated or not with LPS and with FTY720.

Data information: n (number of animals per group) = 4. Error bars depict SEM. FTY: FTY720. Statistical significance was determined with a two‐tailed t‐test. NT, untreated.

Figure 4. Blocking NK‐cell recruitment at the draining lymph node does not interfere with the generation of effector NK cells

1. Absolute numbers of NK cells in one of the four draining lymph nodes in mice treated or not with the anti‐CD62L antibody to inhibit NK‐cell ingress in the lymph node from blood. Where indicated, mice were challenged with LPS.

2. Absolute numbers of IFN‐γ+ NK cells in one of the four draining lymph nodes before and after LPS administration in mice in which NK‐cell entry in the lymph node was blocked or not by anti‐CD62L treatment.

3. (Left panel) Absolute numbers of endogenous and adoptively transferred (CFSE+) NK cells in the draining and contralateral lymph node 5 h after LPS administration; (right panel) percent of endogenous and adoptively transferred NK cells on total lymphocytes in the draining and contralateral lymph node 5 h after LPS administration.

4. Flow cytometry analysis of LPS‐draining and contralateral lymph node from mice that received CFSE‐labeled cells at the moment of LPS administration. Lymph node cells were stained with anti‐CD3, anti‐DX5, and anti‐IFN‐γ antibodies, and the analysis was performed on gated CD3−DX5high cells. One representative of four independent experiments is shown.

5. Tumor volume in mice treated or not with LPS; where indicated, mice received anti‐CD62L antibody 12 h before LPS treatment.

6. Percent of IFN‐γ+ NK cells within the tumor in mice treated or not with LPS and with the anti‐CD62L antibody.

Data information: (A, B, C, E, F) n (number of animals per group) = 4. Error bars depict SEM. Statistical significance was determined with a two‐tailed t‐test. NT, untreated.

Figure 5. DC–NK‐cell interactions in lymph nodes

1. A

   Representative image of an NK cell (red) interacting with four DCs (green).

2. B–D

   Plot of the confinement ratio (CR, black squares), the normalized NK speed (red circles), and the digitalized NK–DC distance, DistD(ti) (green triangles), as a function of time for the NK cell reported in the image. Panels (B–D) correspond respectively to the interaction with the DC1, DC2, and DC4 displayed in (A). For reference, also the normalized distance Dist(ti) is reported in the plots (blue down‐triangles). The time periods in which all the three conditions discussed in the text are fulfilled are marked with red striped rectangles in the plots. The green up‐triangles indicate the parameter TonToff defined as TonToff = −1 if Dist(ti) < d = 25 μm and TonToff = −1 if Dist(ti) ≥ d and indicate the putative interactions according to a more simplified algorithm based on the DC–NK‐cell distance alone.

Figure 6. DCs and NK cells establish prolonged interactions in lymph nodes upon LPS injection

1. NK and DC trajectories as they are tracked by Volocity software on the 4D volumes collected in lymph nodes on the TPE microscope. The sequence of images shows the interaction between an NK cell (red) and a DC (green). In the top panels, the NK cell moves according to a directional random motion, while in the lower left and middle panels, the NK cell seems to recognize the DC and to go back. In the right lower panel, the stable contact DC–NK cell is clearly visible.

2. Time duration of the DC–NK contacts in steady state (−LPS) and inflammatory (+LPS) conditions. Bars indicate the mean of the distribution. The percent of cells showing long interaction times (≥ 15 min) is shown.

3. NK‐cell 3D velocities. 3D velocities of NK cells taking contacts with DCs before (black) and 5 h (green) after LPS treatment.

Figure 7. IL‐18 is the contact‐dependent signal required in DC–NK‐cell interactions

1. Differential interference contrast images (left panels) and confocal microscopy analysis of IL‐18 (right panels) in a DC/NK conjugate after 2 h of coculture in the presence of LPS and in unstimulated cells (NT). Representative images from 3 experiments are shown, magnification 63×. Scale bars, 4.36 μm.

2. FN‐γ release by NK cells cultured in the presence of supernatants recovered from untreated DCs (untreated) or DCs treated O/N with LPS (LPS). Where indicated, IL‐18 was added to the cultures.

3. Cell contact‐dependent activation of NK cells by DCs depends on IL‐18. Unstimulated or LPS‐activated DCs and NK cells were cocultured in the same wells (DC + NK + LPS) or separated by a porous membrane [(DC + LPS)/NK]. Where indicated, IL‐18 was added to the transwells [(DCs + LPS)/(NK + IL‐18)]. NK cells alone were also cultured in the presence of the three selected cytokines: IL‐2 and IFN‐β in the upper chamber and IL‐18 in the lower chamber. IFN‐γ in the supernatant was then measured by ELISA after 8 h of coculture.

Data information: (B, C) n (number of independent experiments) = 3. Error bars depict SEM.