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.

Thank you for the submission of your manuscript to EMBO Molecular Medicine. We have now heard back from the three Reviewers whom we asked to evaluate your manuscript.
Although the Reviewers agree on the potential interest of the manuscript, the issues raised are of a fundamental nature. I will not dwell into much detail, but I would like to highlight the main points.
You will see that Reviewer #1 laments the descriptive nature of the work presented and the lack of mechanistic insight. Reviewer #2, while more positive, does find insufficient experimental support for the main conclusions. Reviewer #3 finds that essential novelty is limited by an earlier study, but also that the potential differences (and advances) with respect to the previous study are not convincingly supported.
In conclusion, while publication of the paper cannot be considered at this stage, given the potential interest of your findings and after internal discussion, we have decided to give you the opportunity to address the criticisms. It will be important that you conclusively show that it is lymphnode NK and not circulating NK that mediate DC activation (reviewer #2 suggests a number of experimental approaches in this respect). You should also discuss, but preferably show, using another model or assay why you find that DC-NK interactions are not short-lived, at variance with previous findings. Indeed, your current discussion appears somewhat dismissive of earlier studies. Finally, a significant effort should be made to gain further mechanistic insight, especially with reference to the role of IL-18 as pointed out by both Reviewers #1 and 3. concept is novel and interesting. However, I feel there are some concerns that need to be addressed to corroborate this principle.
1) It is unclear whether all peripheral NK cells are CD62L positive and whether neutralization of CD62L entirely blocks the recruitment of peripheral NK cells into the lymph nodes. The authors should analyze CD62L KO mice and see whether NK cells are depleted in lymph nodes in steady state And after LPS injection.
3) Fig 3a. needs to be graphed with matched +/-FTY time points in order to evaluate whether administration of FTY influences NK cell function since it appears FTY does influence acquisition of IFN-g in contrast to the authors claims that it does not. 4) Fig 3,b I am confused as to whether egress/function are really affected here. What % of NK cells and IFNg+NK cells are being presented? A % of total cells, or % of just NK cells? The % might be affected by a lack of leukocyte egress from the lymph node if FTY blocks. Therefore, it might be better to display absolute No¹s of NK and IFNg+NK -is this possible? If not, please explain. Fig 3c,d -difficult to ascertain whether the reduced tumor vol. and %IFN-g+ NK cells in tumors is due to reduced recruitment or activation. 6) Fig. 4. Cells recruited to the lymph node are blocked with anti-CD62L, but the authors do not find that the % IFNg+ NK cells is significantly increased even though there is definitely an increased trend. The authors need more samples to gain statistical significance before they can conclude there is an increase otherwise they cannot claim that resident NK cells are being activated, especially since tumor vol is decreased. 7) It¹s easy to say lymph node resident NK are preferentially activated if you¹ve included an anti-CD62L blocking mAb to block cellular recruitment to the lymph node. Why can¹t NK cells that are freshly recruited to lymph nodes also subsequently become activated? In which case -what is the point of this experiment? -I may be wrong, but I¹m not sure there is any evaluation of whether NK cells can reach lymph nodes in this experiment because it is not clear whether the IFN-g+ NK cells present in the skin after LPS administration are resident or recruited or whether they have migrated to resident lymph nodes and that is why there is no difference in % IFN-g+ NK cells -can the authors please clarify what is happening here? The techniques are great, but more statistical analysis required for the in vivo imaging, esp since the author have different conclusion versus previous study. Fairly novel in that very few studies of DC-NK interaction in vivo, but idea of NK-DC crosstalk in LN has been around for a decade. System is contrived an unlikely pertains the physiology of tumor immunity or infection. Not sure it is relevant to immune therapy.

5)
Referee #3 (Remarks): In the presented manuscript, the authors provide some evidence that NK cells become activated by DCs in the lymph node to exert their antitumoral functions after stimulation with lps. The results showing a direct in vivo interaction between NK cells and DCs in the lymph node are interesting and new. The authors have modified and improved their 2-photon method, resulting in convincing data that lps induces increased interactions between adoptively transferred NK cells and DCs. The authors show nicely that the lymph node is required for endogenous NK cell activation by lps stimulation. The kinetic experiments indicate that LN activated NK cells travel to the tumor to induce antitumor immunity. The authors argue that sustained interaction between DC and NK cell in the LN leads to IL-18 induced NK cell IFN production but these data are weak.
Overall, the data are convincing and the authors interpret the data well. The finding that lps induces NK cell IFN production in the LN is not novel. In fact, Martin-Fontecha and Sallusto published these results in seminal studies in nature Immunology 2004, but as discussed in the manuscript introduction, these results are likely due to recruitment of NK cells into the LN, and not due to LN resident NK cells. The authors posit that the novelty of their findings is that resident NK cells are activated by lps in their model system. This somewhat contradicts the Sallusto findings and could be due to differences in the model system. There is a wealth of literature discussing the role of NK cell -DC crosstalk. The authors results add to this concept by showing direct in vivo interactions between NK and DC in the LN. In my opinion this is the most novel aspect of the data, although the results contradict a previous study that showed that NK-DC interactions were short-lived. The authors discuss this difference. These two apparent contradictions to the literature highlight how context-dependent the authors' results are. One way to address this issue is to use another model system, although this could be beyond the scope of the study.

Specific comments
In the methods, lps was injected sc in some exps and iv in others. This could be better defined in the figures and figure legends. • The authors should not use "CTL" to abbreviate "control". It is used to abbreviate cytotoxic T lymphocytes.
• Are T cells already active in the cancer model used? The protective function of LPS could also be lost in the DC-depleted mice because of loss of T cell function Fig 2: • In b, please correct "untreaded" into "untreated". • Please provide a scale bar for the microscopic pictures • How many tumors were assessed? • What is the relevance of this figure? Fig: 3 -interesting data -need to discuss that effect of FTY on tumor growth could be due to blocking ag-specific T cell migration -cannot conclude that NK cell antitumor activity is due to their IFN production in the tumor. It is possible that NK cells make IFN to polarize to Th1 in the LN, and the T cells are required for tumor rejection. • Anti-L-Selectin is supposed to inhibit the migration of NK cells into the lymph nodes. Here, the % of NK cells is increased. This is probably due to the antibody inhibiting the ingress of T cells, which increases the percentage of NK cells. The authors should show the data for total NK cells. Only this would show that the antibody actually inhibited the ingress of NK cells from blood into lymph nodes.
• In fact, the authors should show the effect of lps on total LN cellularity since previous studies show that LN   • The authors should show that LPS-stimulated DCs indeed secrete IL-18 in their hands. If they don't show that, they cannot say that these cells are insufficient to induce IFN release by NK cells. Maybe the release of IL-18 by DCs did just not work in their experiment. • This experiment is insufficient to make the conclusion "IL-18 is the contact-dependent signal required in DC-NK cell interactions". So far, the authors have only shown that it is sufficient. The contact-dependence is in no way shown, b suggests both (NK cells can still be activated even though they don't come in direct contact with DCs, as long as DCs are stimulated with LPS). The authors should perform additional experiments. • In fact, I would interpret the data to show that lps+DC induces a soluble factor that can activate NK cell IFNg in the presence of rIL-18 given to the NK cell.  We thank this referee for his very insightful and helpful comments. The issues raised helped us to better focus the manuscript and to better introduce in the text the single experiments proposed.

The study analyzes the effect of LPS injections on the anti-tumor activity of NK cells. Tumor infiltration by lymph node-resident NK cells is enhanced by LPS injections leading to slower tumor growth. The authors observe an interaction of NK cells with dendritic cells. In addition, a stimulatory activity of IL-18 is documented.
The authors make a couple of potential interesting observations. The underlying molecular mechanisms should be worked out. Answer: We thank the reviewer for this comment concerning the scope of the work. To better focus on the mechanism we have better introduced in the text the data previously published by our group (pages 12, 13) and we have added new original data (new figures 3, 4, 7a). extensively addresses the question of the origin of IL-18 necessary for NK cell activation in LPSmediated inflammatory conditions. In particular, starting form the observation that TLR4-deficient but not MyD88-deficient NK cells could be activated by LPS-stimulated DCs we hypothesized that either IL-1β or IL-18 were required for the DC-mediated NK cell activation since these two cytokines require MyD88 for signaling. We therefore blocked IL-1β and IL-18 in DC-NK cell cocoltures, and we observed that only blocking IL-18 using soluble IL-18 binding protein we could block NK cell activation. We then used IL-18Rdeficient NK cells and IL-18-deficient DCs and in both cases we could observe an inhibition of NK cell activation. Finally, and of great importance for this reviewer comment, we generated chimeric mice in which only DCs were deprived of the capacity to produce IL-18 in vivo. Also here we observed a significant reduction in NK cell activation in vivo following LPS administration. From all of these experiments we concluded that DC-derived IL-18 was required for NK cell activation. This point has been better introduced and clarified in the text (page 13).
Also, this manuscript extends the above-described original observation by showing that IL-18 requires contact between DCs and NK cells to be active. This is highly unusual for secreted factors, and it indicates that the local concentration of IL-18 drops sharply as the distance between DCs and NK cells increases, to the point that IL-18 concentration is no longer biologically active when DCs and NK cells are not in close proximity. Therefore, despite its soluble nature, IL-18 acts as a contact-dependent signal.

2.
The LPS is interesting but it has to shown whether this is a direct or indirect effect. On which cells does LPS act. Which genes are stimulated and which mediators are released? Answer: This point has also been extensively addressed in our previous work (Zanoni I et al. Cell Rep. 2013 26:1235-49) where we definitively demonstrate that LPS acts on DCs (TLR4-deficient NK cells are activated by TLR4+ DCs as efficiently as TLR4-NK cells) and not on NK cells. We also demonstrate that LPS elicits from DCs the production of IL-2, IL-18 and IFNb, three cytokines necessary and sufficient to activate NK cells. This point has been better clarified in the text (pages 12-13).

The authors treat mice with an anti-CD62L antibody to block the entry of NK cells to lymph nodes through HEVs. What else is the anti-CD62L antibody doing in vivo? This should be explained and discussed.
Answer: The anti-L selectin CD62L antibody is commonly used in vivo to block the passage of lymphocytes through HEVs and their entry into secondary lymphoid organs. With the purpose to block NK cell entering in lymph node has been used for instance by Fontecha et al. Nevertheless the possibility that the antibody has other activities in vivo cannot be totally excluded. For this reason, a second approach has been used to strengthen our results on the activation of NK cells present in the lymph node at the moment of stimulus administration. To investigate whether LPS favors the recruitment of NK cells for their activation, CFSE-labeled NK cells have been injected iv in mice at the time of LPS injection. Adoptively transferred cells represent circulating cells that can be recruited in a preferential way at the inflamed lymph node.
Number and activation of labeled NK cells were then evaluated 5 hours after LPS administration in the draining and the contralateral lymph nodes. The prediction was that if NK cells were actively recruited by LPS at the inflamed lymph node for their activation, a larger number of CFSE+ cells should be found at the draining lymph node compared to the contralateral one. We observed that labeled NK cells injected at the time of LPS administration reached the draining and the contralateral lymph nodes with equal efficiency, suggesting homeostatic turnover rather than active recruitment (new figure 4 C). This excludes that LPS favors NK cell recruitment early after administration. Also, fractions of IFN-γ + NK cells observed in the CFSE positive and negative populations were comparable in the LPS draining lymph node ( Figure 4D). This again supports the prediction that NK cells are not recruited to the inflamed lymph node to be activated, otherwise most of the CFSE+ cells should have been IFN-γ+ at the peak of NK cell activation (5 hours after LPS administration). These experiments support the scenario that NK cells resident in lymph nodes at the moment of stimulus administration are the cells preferentially undergoing activation.

Have the plasma levels of cytokines been measured?
Answer: Since the stimulation with LPS is local rather than systemic, the plasma levels of cytokines would be of limited information.

If IL-15 can act in an autocrine fashion, does this necessarily exclude an effect of this cytokine on NK cells? This should be explained and discussed.
Answer: In our previous work (Zanoni I et al. Cell Rep. 2013Rep. 26:1235 we showed that both cis and trans presented IL-15 are required for optimal NK cell activation. Nevertheless NK cell activation can occur also if IL-15 is only cis-or trans-presented. Our observation has been then confirmed in human studies (Mattiola I et al J Immunol. 2015, 195(6):2818-28.). Since NK cells can produce and cis-present IL-15, the immune-synapse between DC and NK cells is not required to deliver the signal of IL-15 but only IL-18. This point has been clarified on pag 13. Mechanistically it binds four out five sphingosine-1-phosphate "S1P" receptors, namely, S1P1, S1P3, S1P4, and S1P5; S1P1 with higher efficiency. In particular fingolimod-phosphate initially activates lymphocyte S1P1 receptor subsequently induces S1P1 down-regulation that prevents lymphocyte egress from lymphoid tissues (M. Matloubian, C. G. Lo, G. Cinamon et al., "Lymphocyte egress from thymus and peripheral lymphoid organs is dependent on S1P receptor 1," Nature, vol. 427, no. 6972, pp. [355][356][357][358][359][360]2004). NK cell egress from lymph nodes depends on S1P5 and S1P1 ( 7a).

The attached movies need to be better explained.
Answer: The legends of the attached movies are included in the supporting information.

Referee 2
We thank this referee for his very insightful and helpful comments. The issues raised helped focus our manuscript and we hope that our follow up experiments have addressed all the major concerns.

The manuscript addresses a fundamentally important question of where NK cell become activated in vivo, which is an important question for stimulating anti-tumor responses. The study proposes that lymph node NK cells and not circulating NK cells are responsible for anti-tumor responses following local administration of adjuvant, such as LPS. This concept is novel and interesting.
However, I feel there are some concerns that need to be addressed to corroborate this principle. We thank the referee for the positive comment on our work The prediction was that if NK cells were actively recruited by LPS at the inflamed lymph node for their activation, a larger number of CFSE+ cells should have been found at the draining lymph node compared to the contralateral one. We observed that labeled NK cells injected at the time of LPS administration reached the draining and the contralateral lymph nodes with equal efficiency, suggesting regular homeostatic turnover rather than active recruitment (new figure 4 C). This excludes that LPS favors NK cell recruitment early after administration and that a preferential recruitment is necessary for NK cell activation. Fractions of IFN-γ + NK cells observed in the CFSE positive and negative populations were comparable in the LPS draining lymph node (new figure  4D). This again supports the prediction that NK cells are not recruited to the inflamed lymph node to be activated, otherwise most of the CFSE+ cells should have been IFN-γ+ at the peak of NK cell activation (5 hours after LPS administration). These experiments support the scenario that lymph node resident NK cells at the moment of stimulus administration are the cells that preferentially undergo activation.

2) Fig 3a. needs to be graphed with matched +/-FTY time points in order to evaluate whether administration of FTY influences NK cell function since it appears FTY does influence acquisition of IFN-g in contrast to the authors claims that it does not.
Answer: The graph has been completely reorganized. Absolute numbers are shown in the new figures 3, 4. Fig 3, The numbers refer to one draining lymph node.

5) Fig 3c,d difficult to ascertain whether the reduced tumor vol. and %IFN-g+ NK cells in tumors is due to reduced recruitment or activation.
Answer: As shown in figure 3a the presence of FTY does not affect NK cell activation but partially affect the egress of activated cells from lymph node. We assume therefore that there is a reduced recruitment at the tumor site. Fig. 4. Cells recruited to the lymph node are blocked with anti-CD62L, but the authors do not find that the % IFNg+ NK cells is significantly increased even though there is definitely an increased trend. The authors need more samples to gain statistical significance before they can conclude there is an increase otherwise they cannot claim that resident NK cells are being activated, especially since tumor vol is decreased. Answer: Figure 4 has been significantly modified. Absolute numbers of NK cells present in the lymph node in animals treated or not with the anti-CD62L antibody are now shown. The data presented support the notion that NK cells continuously recirculate among secondary lymphoid organs and blocking NK cell entry in the lymph node for 24 hours strongly reduces the total number of NK cells present in the lymph node at the steady state. The absolute number analysis indicates that the number of activated NK cells at the draining lymph node is only minimally affected if the recruitment of blood born NK cell is blocked. The notion that resident NK cells are activated is supported by using a second, independent approach as described above (see point 1).

7) It's easy to say lymph node resident NK are preferentially activated if you've included an anti-CD62L blocking mAb to block cellular recruitment to the lymph node. Why can't NK cells that are freshly recruited to lymph nodes also subsequently become activated? In which case what is the point of this experiment?
Answer: The recruitment, over the regular homeostatic turnover, of NK cells from blood induced by LPS takes more than 12 hours and occurs after the egress of early activated NK cells, as shown in the new figure 3b and 4a. The peak of NK cell activation is seen at 5 hours. Consistently, CD11c+ cells-derived cytokines required for NK cell activation (IL-2, IL-18 and IFNbeta) are produced at early time points after stimulation with LPS (from 2 to 6 hours, see Zanoni I et al. Cell Rep. 2013Rep. 26:1235. Therefore, the NK cells activated in the lymph node are preferentially those already present in the lymph node at the time of LPS administration. It is also worth noting that LPS is a transient stimulus and quickly locally inactivated by de-acylation (Lu M, Munford RS (2011) The transport and inactivation kinetics of bacterial lipopolysaccharide influence its immunological potency in vivo. J Immunol 187: 3314-3320). Therefore the effect of LPS cannot persist over time. Fig 4c,d should come before c,d since this explains the above query. Answer: there must be a typo in the request that prevented us from being able to address this point. We are open to accommodate the request once the typo is fixed. Answer: The cells that we delete in the lymph node by diphteria toxin administration are CD11c high CD11b low/-(see the image below). CD169 + macrophages are identified as CD11b + CD11c low . The authors of the cited work call CD169 + CD11c high cells dendritic cells. Therefore, we are mostly depleting dendritic cells, although we cannot exclude that a few macrophages get deleted as well. As shown in the figure below, diphteria toxin administration depletes mostly CD11c high cells.

Overall, the data are convincing and the authors interpret the data well. The finding that lps induces NK cell IFN production in the LN is not novel. In fact, Martin-Fontecha and Sallusto published these results in seminal studies in nature Immunology 2004, but as discussed in the manuscript introduction, these results are likely due to recruitment of NK cells into the LN, and not due to LN resident NK cells. The authors posit that the novelty of their findings is that resident NK cells are activated by lps in their model system. This somewhat contradicts the Sallusto findings and could be due to differences in the model system. There is a wealth of literature discussing the role of NK cell -DC crosstalk. The authors results add to this concept by showing direct in vivo interactions between NK and DC in the LN. In my opinion this is the most novel aspect of the data, although the results contradict a previous study that showed that NK-DC interactions
were short-lived. The authors discuss this difference. These two apparent contradictions to the literature highlight how contextdependent the authors' results are. One way to address this issue is to use another model system, although this could be beyond the scope of the study. +diphteria*toxin* CD11b* 1,19%* 0,11%* 0,21%* 0,03%* 0,82%* 0,09%* control*

In the methods, lps was injected sc in some exps and iv in others. This could be better defined in the figures and figure legends.
Answer: We thank the reviewer for this comment. LPS has always been injected sc. We have erroneously stated in the Mat and Met section that LPS has been injected iv. This error has been corrected.  . In b, please correct "untreaded" into "untreated".
Answer: This typos has been corrected .

Please provide a scale bar for the microscopic pictures
Answer: The scale bar has been provided . How many tumors were assessed? . What is the relevance of this figure? Answer: We have assessed three different tumors from three different mice. We observed that tumors were significantly less vascularized when NK cells had been activated. Therefore, we think that the anti-tumor effect of NK cells is most likely related to neovascularization.

Fig: 3 -interesting data -need to discuss that effect of FTY on tumor growth could be due to blocking ag-specific T cell migration
Answer: The anti-tumor effect observed following LPS administration is T cell independent as discussed above (Figure 1c in the revised manuscript).

-cannot conclude that NK cell antitumor activity is due to their IFN production in the tumor. It is possible that NK cells make IFN to polarize to Th1 in the LN, and the T cells are required for tumor rejection.
Answer: The anti-tumor effect observed following LPS administration is T cell independent as discussed above (Fig 1C in the revised manuscript). Figure 4 has been significantly changed and absolute numbers of NK cells present in the lymph node in animals treated or not with the anti-CD62L antibody are now shown (new Figure 4a).

Fig 4: . Anti-L-Selectin is supposed to inhibit the migration of NK cells into the lymph nodes. Here, the % of NK cells is increased. This is probably due to the antibody inhibiting the ingress of T cells, which increases the percentage of NK cells. The authors should show the data for total NK cells. Only this would show that the antibody actually inhibited the ingress of NK cells from blood into lymph nodes. . In fact, the authors should show the effect of lps on total LN cellularity since previous studies show that LN size increased after adjuvant injection. . Authors should show that anti-CD62L does NOT affect IFNg+ NK cell in TUMOR to confirm that ingress in LN is not important
Total lymph node cellularity is strongly reduced in the presence of CD62L antibody (from a mean of 500,000 cells to a mean of 100,000) and does not increase following LPS administration. Absolute NK cell numbers are also strongly reduced. The data presented support the notion that NK cells continuously recirculate through secondary lymphoid organs and blocking NK cell entry in the lymph node for 24 hours strongly reduces the total number of NK cells present in the lymph node at the steady state. The absolute number analysis indicates that the number of activated NK cells at the draining lymph node is only minimally affected if the recruitment of blood borne NK cell is blocked. See also supplementary figure 1.
To strengthen our results on the activation of lymph node resident NK cells, we followed a second, independent strategy. CFSE-labeled NK cells were adoptively transferred at the same time of LPS administration. The presence of labeled NK cells and their activation were then evaluated 5 hours after LPS administration in the draining and the contralateral lymph nodes. The prediction was that if NK cells were actively recruited by LPS at the inflamed lymph node for their activation, a larger number of CFSE+ cells should have been found at the draining lymph node compared to the contralateral one. We observed that labeled NK cells injected at the time of LPS administration reached the draining and the contralateral lymph nodes with equal efficiency, suggesting homeostatic turnover rather than active recruitment (new figure 4 C). This excludes that LPS favors NK cell recruitment early after administration in order to induce their activation. The percentage of CFSE+IFN-γ+ NK cells observed in the LPS draining lymph node was comparable to that of CFSEnegative IFN-γ+ (new figure 4D). This again supports the prediction that NK cells are not recruited to the inflamed lymph node to be activated, otherwise most of the CFSE+ cells should have been IFN-γ+ at the peak of NK cell activation (5 hours after LPS administration). These experiments support the scenario that lymph node resident NK cells at the time of stimulus administration are the cells that preferentially undergo activation.
The presence of IFNγ+ NK cells inside the tumor in mice treated with anti-CD62L has been added to the new figure 4F.

Fig 5 If I understand correctly, the authors show examples of 3 different types of DC-NK interactions after lps injection. Why is there heterogeneous interactions? Is this dependent on the timing of the lps?
Answer: The plots in this Figure show the evolution of the four parameters used in our algorithm to identify the DC-NK cells interactions. Our reasoning is that during the time course these parameters change because of the motion of the NK cells (mainly) in a random (no-interaction case) or coordinated (when there is an interaction) way. The evolution of the parameters shown in the plots is not due to the timing of LPS, which was injected 2-4 hours before these measurements. Its action should be at a regime at the time of the observations reported in Fig.4. We apologize if our explanation in the original manuscript was not sufficiently clear. The additional materials that we have now added to the revised manuscript, discussing in detail the assignment of interactions, should clarify the meaning of the plots in Fig.4. Answer: Fig.6 indeed reports one of the major results of the manuscript, namely that the fraction of the long lasting (assumed to be > 900s) interactions increases to 12% upon LPS treatment (panel b). At the same time the average speed is not significantly changing in the two cases (panel c).
The speed reported in panel B is the average speed. The two parameters plotted in Fig.6, panels B and C, are to be considered the output of the algorithm for the detection of cell contacts. In fact the duration is the direct output of the algorithm, which is based on three parameters: instantaneous speed, cell-cell distance and confinement ratio. The speed reported in Fig.6, panel C is the average speed over the whole trajectory and is used to show that the overall motion, observed over long times (much larger than 900 s) is not sensibly affected by the interactions. The four parameters in Fig.5 are the three parameters used for the interaction algorithm and a monitor parameter T on T off . This parameter (green up-triangles) is defined as T on T off = -1 if Dist(t i ) < d = 25 µm and T on T off = -1 if Dist(t i ) ≥ d and indicate the putative interactions according to a more simplified algorithm based on the DC-NK cell distance alone. The difference in the two algorithms is remarkable and may explain at least partially some of the differences that we observe here with respect to literature results. We apologize for the lack of definition of this parameter in the original Ms, and thank the Reviewer for having pointing this out for us. In particular, starting form the observation that TLR4-deficient but not MyD88-deficient NK cells could be activated by LPS-stimulated DCs we hypothesized that either IL-1b or IL-18 were required for the DC-mediated NK cell activation since these two cytokine require MyD88 for signaling. We therefore blocked IL-1b and IL-18 in DC-NK cell cocoltures, and we observed that only blocking IL-18 using soluble IL-18 binding protein we could block NK cell activation. We then used IL-18Rdeficient NK cells and IL-18-deficient DCs and in both cases we could observe an inhibition of NK cell activation. Finally we generated chimeric mice in which only DCs were deprived of the capacity to produce IL-18 in vivo and again we could observe a significant reduction in NK cell activation in vivo following LPS administration. From all of these experiments we concluded that DC-derived IL-18 was required for NK cell activation. The other two DC-derived cytokines we have found indispensable for NK cell activation in the presence of LPS are IL-2 and IFNβ. IFNβ, in turn, is required to elicit IL-15 and IL-15Ra production from both DCs and NK cells. IL-15 induces NK cell activation via cis-and transpresentation. This point has been better clarified in the revised version of the manuscript (page 12). In the present manuscript we have added an experiment showing that IL-18 is secreted at the contact site between DC-NK cells (new figure 7A).

Minor comments
The legends in Figures 3c and 4c   Thank you for the submission of your revised manuscript to EMBO Molecular Medicine. We have now received the enclosed reports from the referees that were asked to re-assess it. As you will see the reviewers are now globally supportive and I am pleased to inform you that we will be able to accept your manuscript pending the following final amendments: 1) Reviewer 3 would like you to clarify a few remaining points. Please deal with these appropriately, especially with respect to the question on significance. Provided you do so carefully, I will make an editorial decision on your next final version.
2) As per our Author Guidelines, the description of all reported data that includes statistical testing must state the name of the statistical test used to generate error bars and P values, the number (n) of independent experiments underlying each data point (not replicate measures of one sample), and the actual P value for each test (not merely 'significant' or 'P < 0.05').
3) We encourage the publication of source data, particularly for electrophoretic gels and blots, with the aim of making primary data more accessible and transparent to the reader. Would you be willing to provide a PDF file per figure that contains the original, uncropped and unprocessed scans of all or at least the key gels used in the manuscript? The PDF files should be labeled with the appropriate figure/panel number, and should have molecular weight markers; further annotation may be useful but is not essential. The PDF files will be published online with the article as supplementary "Source Data" files. If you have any questions regarding this just contact me.
4) The manuscript must include a statement in the Materials and Methods identifying the institutional and/or licensing committee approving the experiments, including any relevant details (like how many animals were used, of which gender, at what age, which strains, if genetically modified, on which background, housing details, etc). We encourage authors to follow the ARRIVE guidelines for reporting studies involving animals. Please see the EQUATOR website for details: http://www.equator-network.org/reporting-guidelines/improving-bioscience-research-reporting-thearrive-guidelines-for-reporting-animal-research/. I note that you have provided most details in the author checklist, but it is also important that you also integrate the manuscript with additional information.
5) Please adjust the figure callouts in the article according to our guidelines (http://embomolmed.embopress.org/authorguide) and refer to movies or videos consistently and correctly. Furthermore, the appendix file is missing the Table of Contents and features some erroneous figure labels; it also needs to be submitted as a PDF file.
6) The movie legends need to be removed from the Appendix file. Please provide each movie legend as a read-me file uploaded together with the movie-file as zip file -per each movie please! 7) Please remove all red lettering from the manuscript and the appendix as it is no longer needed.
8) I have slightly edited the Abstract section (please see attached manuscript) and suggested an alternative title. Please accept/modify as appropriate using the attached version. I should add that I find the parts of the "problem" paragraph in your "The Paper Explained" section to be more compelling and efficacious than the Abstract introductory part. You might consider modifying the Abstract a little in that respect.
Please submit your revised manuscript within two weeks. I look forward to seeing a revised form of your manuscript as soon as possible.
***** Reviewer's comments ***** Referee #1 (Comments on Novelty/Model System): The authors have convinced with their revision Referee #1 (Remarks): The study analyzes the effect of LPS injections on the anti-tumor activity of NK cells.  Fig 2b "untreaded" needs to be changed (pointed out in first review, but still not done!) 3. Disagree with interp of Fig 4B: in fact, the number of activated NK cells in the draining lymph nodes is strongly reduced (about half) after CD62L antibody treatment, indicating that the migration of circulating NK cells into the LNs does have some effect. Are these results significant? They look significant but the stars were left out.
2nd Revision -authors' response 09 June 2016 We are pleased that the reviewers were satisfied with our revision. We have edited the manuscript and figures to further clarify any possible remaining confusing point and to meet your suggestions. 2. Fig 2b "untreaded" needs to be changed (pointed out in first review, but still not done!) Untreated has been removed and substituted with NT, similarly to other figures. Fig 4B: in fact, the number of activated NK cells in the draining lymph nodes is strongly reduced (about half) after CD62L antibody treatment, indicating that the migration of circulating NK cells into the LNs does have some effect. Are these results significant? They look significant but the stars were left out.

Disagree with interp of
We thank the reviewer for this comment that helped us to further clarify this important point of the manuscript. Mice treated with anti-CD62L have a strongly reduced number of both total NK cells and IFN-γ+ NK cells in steady state conditions. When mice are treated with LPS the absolute numbers of NK cells that become IFN-γ+ are very similar in mice treated or not with anti-CD62L (as depicted in the figure below).
These numbers are calculated by subtracting to the total number of IFN-γ+ NK cells the numbers of basally activated cells. Therefore, although the total number of activated cells appears to be higher in mice that did not receive the anti-CD62L antibody, this is only due to the higher number of basally activated cells. The numbers of newly activated cells after LPS treatment is only minimally affected by the antibody treatment (see figure). This minimal, non statistically significant, difference in the numbers of newly activated cells is likely due to the regular homeostatic recerculation that is not present in anti-CD62L treated mice.
We have introduced a further clarification in the text to stress this point (page 7, lines 3-10).
2) As per our Author Guidelines, the description of all reported data that includes statistical testing must state the name of the statistical test used to generate error bars and P values, the number (n) of independent experiments underlying each data point (not replicate measures of one sample), and the actual P value for each test (not merely 'significant' or 'P < 0.05').
The actual p values have been added in each figure and the numbers of mice used added in each legend. All the statistical analyses have been performed according to the paragraph "Statistical analysis" in the Material and Methods section.
3) We encourage the publication of source data, particularly for electrophoretic gels and blots, with the aim of making primary data more accessible and transparent to the reader. Would you be willing to provide a PDF file per figure that contains the original, uncropped and unprocessed scans of all or at least the key gels used in the manuscript? The PDF files should be labeled with the appropriate figure/panel number, and should have molecular weight markers; further annotation may be useful but is not essential. The PDF files will be published online with the article as supplementary "Source Data" files. If you have any questions regarding this just contact me.
Electrophoretic gels and blots are not present in the manuscript. http://www.equator-network.org/reportingguidelines/improving-bioscience-research-reporting-the-arrive-guidelines-for-reportinganimal-research/. I note that you have provided most details in the author checklist, but it is also important that you also integrate the manuscript with additional information.

4) The manuscript must include a statement in the
The required additional information concerning the studies involving mice has been added in the Material and Methods section (Mice paragraph).

5) Please adjust the figure callouts in the article according to our guidelines
(http://embomolmed.embopress.org/authorguide) and refer to movies or videos consistently and correctly. Furthermore, the appendix file is missing the Table of Contents and features some erroneous figure labels; it also needs to be submitted as a PDF file.
The figure callout has been changed. The Table of Contents has been added to the appendix file and figure labels have been corrected.
6) The movie legends need to be removed from the Appendix file. Please provide each movie legend as a read-me file uploaded together with the movie-file as zip file -per each movie please!
A zip file has been generated for each movie including the movie and its legend.

7)
Please remove all red lettering from the manuscript and the appendix as it is no longer needed.
Red lettering has been removed.

8) I have slightly edited the Abstract section (please see attached manuscript) and
suggested an alternative title. Please accept/modify as appropriate using the attached version. I should add that I find the parts of the "problem" paragraph in your "The Paper Explained" section to be more compelling and efficacious than the Abstract introductory part. You might consider modifying the Abstract a little in that respect.
Title and abstract have been slightly changed according to your suggestions.