Animal studies were approved by Massachusetts General Hospital Institutional Animal Care and Use Committee (IACUC).

C57BL/6 wild-type (WT) mice (Charles River, Wilmington, MA, USA, strain code: 207) and C57BL/6 Ncr1^iCre,ROSA^mT-mG (bred in-house) were used in the tumor experiments. All mice were housed under specific pathogen-free conditions, given water and food ad libitum, in the animal facility at Massachusetts General Hospital in accordance with animal care regulations. All mice were closely monitored by the authors, facility technicians, and an independent veterinarian when necessary. All procedures were performed according to the protocols approved by the Institutional Animal Care and Use Committee (IACUC) at Massachusetts General Hospital.

WT and B2m^-/- B16-F10 (B16) cell lines were maintained in RPMI 1640 (Life Technologies, catalog no. 21870076) media supplemented with 10% FBS and 1% P/S and cultured at 37°C/5% CO₂.

All metastatic melanoma studies were performed in eight to twelve-week-old C57BL/6 female WT or Ncr1^iCre,ROSA^mT-mG mice. Mice were intravenously injected with 2x10⁵ WT or B2m^-/- B16 cells (in RPMI medium) via the tail vein or retroorbital vein. Control mice were injected with an equivalent volume of RPMI medium via the tail vein. Mice were euthanized 14 days post-tumor injection to collect the organs. For NK and T cell depletion, mice were intraperitoneally (I.P.) injected with 500 µg/mouse of IgG isotype control (Southern Biotech, Birmingham, AL, USA, catalog no. 0107-01), depleting anti-NK1.1 (clone PK136, BioXcell, West Lebanon, NH, USA, catalog no. BE0036), or depleting anti-CD4 (clone GK1.5, BioXcell, catalog no. BE0003-1) plus depleting anti-CD8 (clone YTS 169.4, BioXcell, catalog no. BE0117) antibodies two days before the tumor cell injection and 250 µg/mouse every other day starting the day of the tumor cell injection.

Mice were anesthetized using ketamine/xylazine and 1 µg of monoclonal CD45-BV605 (clone 30-F11, BioLegend, San Diego, CA, catalog no. 103155) was injected by retroorbital injection for three minutes to label circulating CD45⁺ cells. Peripheral blood was collected by retroorbital bleeding and, following euthanasia, lungs and liver were collected for further processing and analysis. Red blood cells in peripheral blood were lysed using red blood cell lysis buffer (RBC lysis buffer 10X, Biolegend, catalog no. 420301). After washing with 1x DPBS/2% FCS/5 mM EDTA, 5x10⁶ cells were prepared for flow cytometry staining.

For paraffin embedding, lungs and livers were fixed in 4% Paraformaldehyde overnight at 4°C (PFA, Sigma Aldrich, catalog no. P6148) and collected for histological analysis. Subsequently, tissues were dehydrated in ethanol, processed, and embedded in paraffin according to standard histology processes. For paraffin embedding of Ncr1^icre,Rosa^mT-mG, a different fixation protocol was used to maintain the NKp46-GFP and Tdt-Tomato-ROSA fluorescence. Lungs from Ncr1^icre,Rosa^mT-mG mice were collected and fixed in pre-chilled 95% ethanol overnight at 4°C. Tissues were subsequently dehydrated in 100% ethanol, cleared in 100% xylene, and embedded in paraffin according to standard histology processes. Blinded quantification of the metastatic lung foci was performed on macroscopic images and the size of metastatic foci was measured using the ImageJ software.

For flow analysis, lungs were chopped with scissors into ~1 mm pieces and were incubated into a 15 mL tube with 10 mL digestion buffer (RPMI 1640 (Life Technologies, catalog no. 21870076), 200 U/mL Collagenase IV (Worthington Biochemical, Lakewood, NJ, USA)) for 2 hours at 37°C with shaking. Liver lobes were collected carefully from the peritoneum of the mice without taking the gall bladder. Digested lungs and liver lobes were mashed with a plunger through the 70 µM strainer which was washed with 1x DPBS/2% FCS/5 mM EDTA by centrifuging at 300 g for 5 minutes at 4°C. Following centrifugation and aspiration of supernatant, the lungs and liver cell pellets were resuspended in 4ml 40% Percoll (Healthcare Biosciences, Uppsala, Sweden, catalog no. 17-0891-01) and gently load onto 2.5ml 70% Percoll. Samples were centrifuged at 2400 rpm for 20 minutes at 4°C without brakes and leukocytes were separated. Following washing the leukocytes with 1x DPBS/2% FCS/5 mM EDTA, the pellet was resuspended in 1x DPBS/2% FCS/5 mM EDTA buffer for flow cytometry staining.

Single-cell suspensions from all samples were prepared by straining through a 70 µm filter. Cells were pre-incubated for 5 minutes at room temperature with CD16/CD32 masking antibody (clone 2.4G2, Tonbo Bioscience, San Diego, CA, catalog no. 70-0161-U500) to prevent non-specific antibody binding. Cells were stained with Zombie-NIR™ fixable viability dye (Biolegend, catalog no. 423106) to differentiate live/dead cells. Then, cells were stained in 1x DPBS/2% FCS/5 mM EDTA with the appropriate surface antibodies ( Table S1 ) for 30 minutes at 4^°C, washed, and analyzed by flow cytometry. For intracellular staining, cells were fixed and permeabilized using a True-Nuclear™ transcription factor kit (Biolegend, catalog no. 424401) according to the manufacturer’s protocol. Fixed cells were stained in 1x permeabilization buffer with the appropriate intracellular antibodies ( Table S1 ) for 60 minutes at room temperature, washed, and analyzed by flow cytometry within a week. Cells were assayed on a BD LSR Fortessa X-20 flow cytometer (BD Bioscience, Billerica, MA, USA), and data were analyzed using FlowJo software Version 10 (Tree Star, Ashland, OR, USA). NK, CD4⁺, and CD8⁺ T cells were identified as CD3^-NK1.1⁺NKp46⁺, NK1.1^-CD3⁺CD4⁺, NK1.1^-CD3⁺CD8⁺, respectively.

For IF of paraffin-embedded tissues, sections of 5 µm were rehydrated and permeabilized with 1x DPBS supplemented with 0.2% Triton X-100 (Thermo Fisher Scientific, catalog no. BP151) for 5 minutes. Antigen retrieval was then performed using a Cuisinart pressure cooker for 20 minutes at high pressure in an antigen unmasking solution (Vector Laboratories, Burlingame, CA, catalog no. H-3300). Slides were then washed three times for three minutes each in 1x DPBS supplemented with 0.1% Tween 20 (Sigma-Aldrich, catalog no. P1379). Sections were blocked, stained, and mounted as described above. Tissue sections were stained overnight at 4°C with primary antibodies ( Table S1 ). The following day, slides were washed as above and incubated for two hours at RT with secondary antibodies conjugated to fluorochromes ( Table S1 ). After washing as above, sections were incubated with a 1:4000 dilution of DAPI (Invitrogen, Carlsbad, CA, catalog no. D3571) in 1x TBS for 5 minutes at RT, washed, air dried and coverslips mounted with Prolong Gold Antifade Reagent (Invitrogen, catalog no. P36930). Five or six randomly selected fields of view at 200x total magnification were obtained for each section using a Zeiss Axio Scan (Zeiss, Oberkochen, Germany). Blinded manual counting of NKp46⁺, CD4⁺, and CD8⁺ T cells was performed using ZEN Blue Software (Zeiss, Oberkochen, Germany). Whole slide imaging was performed using a NanoZoomer S60 Digital slide scanner (Hamamatsu, Japan) and analyzed with NDP-view2 software (Hamamatsu).

Bar graphs show mean values + standard deviation (SD). The numbers of mice per group used in each experiment are annotated in the corresponding figure legend as n. Graphs and statistical analysis were performed using GraphPad Prism 8 (La Jolla, CA, USA). All tumor quantifications were performed blindly. One-way ANOVA and Two-tailed Mann-Whitney U tests were used for the comparisons. A P value of less than 0.05 was considered significant.

To investigate the role of NK and T cells in the surveillance of cancer metastasis, we employed the B16 metastatic melanoma model together with antibody-mediated depletion of NK and T cells. B16 melanoma cell line with endogenous expression of MHC-I (WT B16) and with loss of MHC-I through B2m gene deletion (B2m^-/- B16) were utilized. Removal of MHC-I, a ligand for inhibitory Ly49 receptors on mouse NK cells, is an established approach to induce NK cell activation (24). However, loss of MHC-I did not significantly alter number or size of metastatic foci in the lungs, which may be because B16 cells endogenously express very low levels of MHC-I ( Figures 1A–C , Figure S1 ) (25).

To identify the anti-tumor role of NK and T cells, mice were treated with NK and CD4/CD8 T cell-depleting antibodies to eliminate NK and T cells in the metastatic melanoma model, respectively. For control, mice were treated with IgG isotype antibody, which did not affect NK and T cells. Compared to IgG isotype-treated mice, NK cell depletion resulted in a significant increase in the numbers of metastatic foci in the lungs on day 14 following tail vein injections of WT and B2m^-/- B16 cells ( Figures 1A , C ). Surprisingly, NK cell depletion did not alter the size of WT and B2m^-/- B16 metastatic foci in the lungs compared to the IgG treatment ( Figures 1B, C ). In contrast, the depletion of CD4⁺ and CD8⁺ T cells did not affect the number of B2m^-/- B16 lung metastasis but significantly increased the size of metastatic foci compared to IgG treatment ( Figures 1A –C ). These findings indicate that NK cells are required to block the seeding of metastatic B16 melanoma cells in the lung, while T cells are not prominent players in this process but rather control the growth of the metastatic foci formed in the lung. Interestingly, there was a slight but significant increase in the number of lung metastasis of B2m^-/- B16 compared to that of WT B16 melanoma cells in NK cell-depleted groups, indicating that T cells may utilize the MHC-I axis to play a minor compensatory role in limiting the number of lung metastasis in the absence of NK cells ( Figures 1A, C ).

Consistently, the depletion of NK plus T cells worsened the metastatic burden of B2m^-/- B16 melanoma cells in the lungs compared to the NK cell depletion, which likely indicates the cooperative role of NK cell-mediated restraint of B16 lung seeding and T cell-mediated control of metastatic foci growth in the lungs ( Figure 1A ). Notably, the combined depletion of NK and T cells resulted in an increased number of B2m^-/- B16 metastatic foci that were large and fused on day 14, preventing the accurate quantification of each metastatic focus ( Figure 1C ).

Despite the pivotal role of NK cells in limiting the B16 melanoma lung metastasis, we did not detect NK cells infiltrating the metastatic foci in the lungs. However, there were infrequent NK cells within the parenchyma of the lung away from the metastatic foci on day 14 post-tail vein cancer cell injection ( Figure 1D ). This finding suggested that NK cells cleared the metastatic cancer cells outside the lung to prevent their pulmonary seeding. To examine whether NK cell clearance of melanoma cells took place in the circulation, we changed the route of melanoma cell delivery from the tail vein to retroorbital vein injection to avoid the retrograde blood flow in the liver (26). Interestingly, the retroorbital vein injection of B2m^-/- B16 melanoma cells led to a massive increase in lung metastasis ( Figures 1E, F ). This finding strongly suggests that the liver is a prominent site where productive interactions between NK cells and metastatic cancer cells occur to control pulmonary melanoma cell seeding. Accordingly, occasional liver metastasis of intravenously injected B2m^-/- B16 cells was noted only in the mice depleted of NK cells. In addition, the combined depletion of NK plus T cells in the mice that received B2m^-/- B16 cells via tail vein injection led to the emergence of extrapulmonary metastasis in various organs such as the liver, kidney, stomach, ovaries, and uterine horns ( Figure S2 ). These findings further highlight the cooperative role of NK and T cells in controlling extrapulmonary metastasis of melanoma in various organs.

Based on the pivotal role of T cells in curbing melanoma metastatic foci growth in the lungs, we assessed the infiltration of CD4⁺ and CD8⁺ T cells in the lungs on day 14 following the tail vein injection of WT and B2m^-/- B16 melanoma. Although the metastatic lung foci at the endpoint were largely devoid of NK cells, we found prominent infiltration of CD4⁺ and CD8⁺ T cells into the WT B16 metastatic foci in the lungs ( Figures 2A, B ). NK cell depletion significantly reduced T cell infiltration into WT B16 metastatic foci ( Figures 2A, B ). Although CD4⁺ and CD8⁺ T cells had lower infiltration into B2m^-/- B16 metastatic foci compared to WT B16 metastatic foci, they infiltrated B2m^-/- B16 metastatic foci in an NK cell-dependent manner ( Figures 2C, D ). These findings demonstrate that NK cells, despite their absence in the metastatic foci, play a prominent role in promoting T cell recruitment into the metastatic foci in the lung.

We next characterized NK and T cells in mice on day 14 post-tail vein injection of WT and B2m^-/- B16 melanoma cells. Immune cells from peripheral blood (PB), lungs, and liver were collected to quantify NK, CD4⁺, and CD8⁺ T cells proportions among total CD45⁺ immune cells in the organs ( Figure S3A ). For specific evaluation of the immune cells infiltrating into the lung tissue, fluorescently labeled anti-CD45 antibody was intravenously administered to mice three minutes before organ harvest. This enabled efficient tagging of all the immune cells in the vasculature while the tissue-resident/emigrated immune cells remained unlabeled (17). CD45-stained immune cells from the vasculature are termed CD45 IV⁺ (IV stands for intravenous) while tissue-resident/emigrated immune cells which remained unstained after intravenously injection of CD45 antibody are termed CD45 IV^-. Following the organ harvest, immune cells were stained ex vivo by distinctly labeled CD45 antibody to stain all immune cells. As expected, 100% of the ex vivo CD45-stained immune cells from peripheral blood (PB) were found to be positive for intravenously injected CD45 antibody (CD45 IV⁺), while a proportion of ex vivo CD45-stained lung immune cells were negative for intravenously injected CD45 antibody (CD45 IV^-) ( Figure S3B ). Compared to the mice that were not injected with cancer cells (Ф), WT and B2m^-/- B16 melanoma cell administration through the tail vein led to a significant increase in the proportions of NK cells in the liver ( Figure 3A ). The percentage of NK cells among total CD45⁺ immune cells in PB and lungs remained unaltered between melanoma cell-injected and control mice ( Figure 3A ). There was no significant difference in the NK cell proportions in the organs between mice that received WT versus B2m^-/- B16 melanoma cells ( Figure 3A ). Quantifying IV^neg NK cells in the lungs revealed a slight but significant enrichment of lung-infiltrating NK cells in WT B16 melanoma cell-injected mice, which further increased in mice injected with B2m^-/- B16 melanoma cells ( Figure 3B ). On the other hand, T cell quantification showed significant enrichment of total and IV^neg proportions of CD4⁺ and CD8⁺ T cells in the lungs of mice that had received WT and B2m^-/- B16 tumors ( Figures 3C –F ). In contrast, T cell proportions were not affected by melanoma cells in the liver ( Figures 3C, E ). Of note, among IV^neg immune effector populations in the lungs, T cells, especially CD4⁺ T cells, showed the most robust increase in response to WT and B2m^-/- B16 melanoma challenge compared to the untreated control mice ( Figures 3B, D, F ). Thus, these data strongly indicate that tumor burden in the lungs promotes modest recruitment of circulatory NK cells to the lungs, which are largely excluded from the metastatic foci in the lungs. Although NK cells lack a direct ability to curb the growth of metastatic foci in the lungs, they regulate CD4⁺ and CD8⁺ T cell recruitment to the lungs, which infiltrate and control the in situ metastatic tumor growth in the lungs.

Next, we assessed lung-recruited NK and T cells on day 14 post-tail vein injection of WT and B2m^-/- B16 melanoma cells. Intravenous administration of fluorescently labeled anti-CD45 antibody allowed us to differentiate the lung-resident/recruited IV^neg NK cells from those in the lung vasculature marked as IV^pos ( Figure S4A ). IV^neg and IV^pos NK cells were quantified for the expression of functional markers ( Figures S4B–E ). Likewise, IV^neg and IV^pos CD3⁺ T cells were separated by CD45 IV labeling followed by gating for CD4⁺ and CD8⁺ T cells and quantification of functional markers ( Figures S4F–L ). In tumor-free mice, most IV^pos NK cells in the circulation were positive for granzyme B and perforin, while only a minor fraction of IV^neg NK cells in the tumor-free lungs showed granzyme B/perforin expression ( Figure 4A ). Intravenous WT and B2m^-/- B16 melanoma challenge resulted in a significant increase in the proportions of granzyme B/perforin positive IV^neg NK cells in the lungs ( Figure 4A ). Tissue residency of NK cells was defined by the mutually exclusive expression of integrins CD49b versus CD49a (27). The majority of IV^pos circulating NK cells in the lung were CD49b⁺CD49a^- conventional NK (cNK) cells in WT, B2m^-/- B16 melanoma-challenged and untreated mice ( Figure S5A ). In contrast, among IV^neg lung-infiltrating NK cells, we found significantly increased cNK cells and decreased CD49b^-CD49a⁺ tissue-resident NK (trNK) cells proportions upon WT and B2m^-/- B16 melanoma challenge ( Figure 4B ).

The relative expression and surface density of integrin CD11b (Mac-1) and CD27 distinguish NK cells in four different maturation and functional stages. CD11b^-CD27^- (double negative, DN) and CD11b^-CD27⁺ NK cells are closely related and developmentally immature precursor-like NK cells that give rise to more mature and cytotoxic CD11b⁺CD27⁺ (double positive, DP) and CD11b⁺CD27^- populations (28). In control mice, almost all IV^pos NK cells in the lungs were mature and functionally potent NK cells with DP and CD11b⁺CD27^- phenotype ( Figure S5B ). Tail vein injections of WT and B2m^-/- B16 melanoma cells did not alter the CD11b⁺ mature NK cell dominance, but it significantly decreased IV^pos DP population and likely led to their conversion into CD11b⁺CD27^- NK cells in the circulation ( Figure S5B ). In the absence of WT or B2m^-/- B16 melanoma cells, almost 50% of IV^neg NK cells in the lungs showed an immature precursor-like phenotype with a lack of CD11b marker expression ( Figure 4C ). WT and B2m^-/- B16 melanoma challenge significantly reduced the IV^neg proportions of DN and CD11b^-CD27⁺ precursors and increased functionally dynamic and mature NK cell populations comprising DP and CD11b⁺CD27^- NK cells ( Figure 4C ). TNFα expression in IV^pos circulating and IV^neg lung-infiltrating NK cells was significantly enhanced in the lungs of mice that received WT and B2m^-/- B16 melanoma cells ( Figures 4D , S5C ). On the other hand, IFNγ expression was undetectable irrespective of the tumor challenge in both IV^pos and IV^neg NK cell populations in the lungs ( Figures 4D , S5C ). In addition to circulating NK cells, liver NK cells in mice that received WT and B2m^-/- B16 melanoma were activated compared to the control mice on day 14-post melanoma challenge ( Figures S5D, E ).

We found that PD1⁺ IV^pos CD4⁺ T cells remained comparable between the control and WT and B2m^-/- B16 melanoma-challenged mice ( Figure S6A , “CD4”). In contrast, PD1 expression on IV^neg CD4⁺ T cells from the lungs was significantly elevated in WT and B2m^-/- B16 melanoma-challenged mice ( Figure 5A , “CD4”). In WT and B2m^-/- B16 tumor-bearing lungs, around 40-50% IV^neg CD8⁺ T cells expressed PD1 compared to only 10-20% IV^pos CD8⁺ T cells in the circulation ( Figures 5A , S6A , “CD8”). Upon lung infiltration, IV^neg CD4⁺ and CD8⁺ T cells significantly upregulated TNFα expression in WT and B2m^-/- B16 melanoma-challenged mice compared to the controls ( Figure 5B ). In circulation, only IV^pos CD8⁺ T cells from mice that received WT B16 melanoma cells significantly upregulated TNFα expression ( Figure S6B ). In contrast to NK cells, IV^pos and IV^neg T cells showed robust expression of IFNγ in the lungs of WT and B2m^-/- B16 melanoma-challenged mice compared to control animals ( Figures 5C , S6C ).