A rate-limiting step for circulating tumor cells to colonize distant organ sites is their ability to locate a microenvironmental niche that supports their survival and growth. This can be achieved by features intrinsic to the tumor cells and/or by the conditioning of a “premetastatic” niche. To determine if pulmonary inflammation promotes the latter, we initiated models for inflammatory asthma, hypersensitivity pneumonitis, or bleomycin-induced sterile inflammation before introducing tumor cells with low metastatic potential into the circulation. All types of inflammation increased the end-stage metastatic burden of the lungs 14 days after tumor cell inoculation without overtly affecting tumor extravasation. Instead, the number and size of early micrometastatic lesions found within the interstitial tissues 96 hours after tumor cell inoculation were increased in the inflamed lungs, coincident with increased tumor cell survival and the presence of nearby inflammation-induced monocyte-derived macrophages (MoDM; CD11b+CD11c+). Remarkably, the adoptive transfer of these MoDM was sufficient to increase lung metastasis in the absence of inflammation. These inflammation-induced MoDM secrete a number of growth factors and cytokines, one of which is hepatocyte growth factor (HGF), that augmented tumor cell survival under conditions of stress in vitro. Importantly, blocking HGF signaling with the cMET inhibitor capmatinib abolished inflammation-induced early micrometastatic lesion formation in vivo. These findings indicate that inflammation-induced MoDM and HGF in particular increase the efficiency of early metastatic colonization in the lung by locally preconditioning the microenvironment.

Implications:

Inflammation preconditions the distant site microenvironment to increase the metastatic potential of tumor cells that arrive there.

This article is featured in Highlights of This Issue, p. 1971

Metastasis is remarkably inefficient: The vast majority of tumor cells that disseminate via the circulatory system either do not extravasate from the vasculature—and are consequently cleared by the body—or they extravasate but do not survive at distant sites (1). In contrast, when prosurvival cues are present within a primed “pre-metastatic niche,” distant site lesions arise that are life-threatening (2, 3). The premetastatic microenvironment can be conditioned to provide such prosurvival cues by factors produced by the primary tumor at a distance. For example, soluble mediators secreted by primary tumors (e.g., lysyl oxidase) can cause stromal cells at distant organs to cross-link and deposit additional extracelluar matrix there. This preconditions the site in two ways. First, the highly cross-linked matrix itself supports the survival of newly seeded tumor cells on their arrival to the organ. Second, the disruption of tissue homeostasis leads to the infiltration and accumulation of immune cells that can also increase tumor cell survival in the niche (4–6). Although the specific cues that lead to immune cell infiltration in these cases are not fully understood, it has recently been shown that chemokines, including CCL2 and CCL3, made by the primary tumor can be involved. Such tumor-produced chemokines also help induce infiltrating monocytes to become “metastasis-associated macrophages” that participate in secondary lesion growth (7, 8).

Conditioning of the premetastatic microenvironment can also occur due to local factors that are produced entirely extrinsic to the tumor progression process. Environmental or genetic instigation of ECM deposition and stiffening, coincident with the accumulation of immune cell infiltrates, conditions premetastatic niches, including in the lung (9). Although inflammation is generally appreciated to be a driver of tumorigenesis at primary sites (10, 11), much less is known about how inflammation contributes to premetastatic niche conditioning at the distant site. Further, although microenvironments that support metastasis often involve an intimate association between immune and tumor cells, this has largely been studied after tumor cells have arrived at the distant site. Such studies have shown that monocyte-derived macrophages (MoDM) in particular are attracted to highly metastatic tumor cells at secondary sites where they secrete vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), and/or epidermal growth factor (EGF) to promote tumor cell survival and proliferation (12–14). Given that tumor-extrinsic MoDM infiltration occurs in virtually all types of inflammatory responses (e.g., Th1, Th2, and sterile inflammation), understanding if and how MoDM contribute to a prior conditioning of the premetastatic niche would seem paramount.

To begin to address this issue, we first established models of inflammation in the lung and subsequently delivered tumor cells into the circulation that normally form metastatic lesions very inefficiently there. We report that three very different models of experimental inflammation significantly increased the ability of low metastatic potential B16F0 melanoma cells to generate large, visible malignant lesions in the lung 14 days after intravenous inoculation. This inflammation-induced prometastatic effect was also evident early in the process; tumor cell survival was enhanced 96 hours after inoculation, and this prosurvival effect was associated with an increased number and enlargement of micrometastatic tumor cell clusters within the lung interstitium. Further, we determined that inflammation caused MoDM derived from inflammatory monocytes (Ly6Chi) to infiltrate the lungs where they associated with the micrometastatic B16F0 cell clusters. These inflammation-induced MoDM secreted factors that promoted the survival of B16F0 cells under stress conditions in vitro and their adoptive transfer specifically increased micrometastatic lesion formation in the absence of inflammation in vivo. One factor secreted by these MoDM is HGF whose cognate receptor, cMET, is expressed by B16F0 cells. Treatment of B16F0 cells with the cMET-specific inhibitor capmatinib (15) prevented the activation of the cMET tyrosine kinase on B16F0 cells in vitro and its instillation in the lung inhibited inflammation-induced B16F0 cell lung colonization in vivo. These findings indicate that a prior inflammatory conditioning of the lung microenvironment by HGF-producing MoDM contributes to a premetastatic niche that facilitates tumor cell survival within early micrometastases which affects the emergence of late-stage metastatic lesions.

Mice

C57BL/6J mice were purchased from The Jackson Laboratory, housed, and bred at the University of British Columbia (UBC). C57BL/6J transgenic mice ubiquitously expressing GFP driven by an α-actin-CMV hybrid promoter were provided by Dr. Fabio Rossi of the UBC. Females, 6–12 weeks of age, were used in the experiments unless explicitly noted. Animal experiments were conducted with protocols approved by the Animal Care Committee at UBC in accordance with the Canadian Council of Animal Care guidelines for ethical animal research.

Lung inflammation models and bronchoalveolar lavage

Asthma was induced as previously described (16) with minor modifications. Briefly, mice were sensitized intraperitoneally with 200 μL 0.2% chicken ovalbumin (OVA) adsorbed to 1 mg/mL Al(OH)3 in PBS (both from Sigma-Aldrich) on days 1 and 8. Mice were subsequently challenged intranasally on days 21, 22, 23, 25, and 27 with 50 μL of 2% OVA and lungs were harvested for histologic analysis 24 hours after the last intranasal administration. Hypersensitivity pneumonitis was induced as previously described (16) by intranasal injecting 50 μL of 4 mg/mL endotoxin-free Saccharopolyspora rectivirgula antigen prepared in-house three times a week for three weeks (days 1, 2, 3, 8, 9, 10, 15, 16, and 17), and lungs were harvested for analysis on day 21. Sterile inflammation and injury were induced by a single intratracheal instillation of 2.5 U/kg bleomycin sulfate (BLM, Cayman Chemicals) dissolved in 50 μL of PBS (17), and lungs were harvested for histologic analysis at day 7. As a control for each inflammation model, we also treated mice with PBS according to the times and routes described for each inflammatory method. In addition, at the completion of each protocol described above, mice were anesthetized at the times specified (see Fig. 1) and bronchoalveolar lavage was performed by three sequential introductions and aspirations of 1 mL PBS to recover immune cell infiltrates, which were subsequently analyzed by flow cytometry (see below).

Figure 1.

Inflammation increases the generation of macroscopic metastatic lesions in the lungs of mice inoculated with B16F0 cells. A, Schematic depicting the treatments to induce either inflammatory asthma via repeated intraperitoneal (IP) and intranasal (IN) administration of OVA, hypersensitivity pneumonitis via repeated IN administration of Saccharopolyspora rectivirgula antigen (SR), or sterile inflammation via a single intratracheal administration of bleomycin (BLM) in the lung. B, Representative lung images stained with Masson's trichrome for collagen in PBS-treated (control) and inflamed lungs as indicated. Scale bar, 1 cm. C, Total number of BAL cells in PBS (control) and inflamed lungs. D, Immune cells isolated by BAL from PBS-treated control and inflamed lungs. Immune cell populations were classified as either AM (SiglecF+CD11c+), lymphocytes (CD3+ or B220+), neutrophils (7/4+ or Ly6G+), or eosinophils (SiglecF+CD11c) and quantified as a percentage of the total CD45+ leukocytes isolated. Data are pooled from two experiments. Total m = 6 for each treatment. E and F, Pigmented, end-stage macrometastatic melanoma nodules in PBS and inflamed lungs at 14 days after intravenous inoculation of B16F0 cells. Images of representative lungs (E) and quantification of nodules (F). Data are pooled from two experiments. Significance indicated as **, P < 0.01; ***, P < 0.001; unpaired Student t test.

Figure 1.

Inflammation increases the generation of macroscopic metastatic lesions in the lungs of mice inoculated with B16F0 cells. A, Schematic depicting the treatments to induce either inflammatory asthma via repeated intraperitoneal (IP) and intranasal (IN) administration of OVA, hypersensitivity pneumonitis via repeated IN administration of Saccharopolyspora rectivirgula antigen (SR), or sterile inflammation via a single intratracheal administration of bleomycin (BLM) in the lung. B, Representative lung images stained with Masson's trichrome for collagen in PBS-treated (control) and inflamed lungs as indicated. Scale bar, 1 cm. C, Total number of BAL cells in PBS (control) and inflamed lungs. D, Immune cells isolated by BAL from PBS-treated control and inflamed lungs. Immune cell populations were classified as either AM (SiglecF+CD11c+), lymphocytes (CD3+ or B220+), neutrophils (7/4+ or Ly6G+), or eosinophils (SiglecF+CD11c) and quantified as a percentage of the total CD45+ leukocytes isolated. Data are pooled from two experiments. Total m = 6 for each treatment. E and F, Pigmented, end-stage macrometastatic melanoma nodules in PBS and inflamed lungs at 14 days after intravenous inoculation of B16F0 cells. Images of representative lungs (E) and quantification of nodules (F). Data are pooled from two experiments. Significance indicated as **, P < 0.01; ***, P < 0.001; unpaired Student t test.

Close modal

Cell lines and cell culture

B16 mouse melanoma cell lines F0 (ATCC CRL-6322), F1 (ATCC CRL-6323), and F10 (ATCC CRL-6475) were purchased from the ATCC. Cell lines were expanded immediately after purchase and frozen in liquid nitrogen (lN2). Cells were routinely cultured in high glucose DMEM (Sigma-Aldrich) supplemented with 8% FBS, utilized within 10 passages after thawing from lN2, and regularly tested for Mycoplasma prior to injection into animals. Low metastatic potential B16F0 cells stably expressing EGFP or mCherry were generated by transfection of the EGFP-N1 vector or mCherry-C1 vector (Takara Bio) using Lipofectamine 2000 (Invitrogen) followed by selection in 1.2 mg/mL G418 (Life Technologies) and two rounds of fluorescence-activated cell sorting to select stable pools of cells with high fluorescent protein expression (top 10%). In some experiments, B16F0 cells were maintained in normal B16F0 medium containing methotrexate (MTX; Cayman Chemicals) without or with z-vad-fmk (R&D Systems) or media conditioned by either MoDM or monocytes. After these treatments, apoptosis and cell death were quantified by flow cytometry using Annexin V binding (BD Biosciences) and DAPI uptake (Thermo Fisher Scientific).

Tracking B16F0 cell colonization of the lung

For metastatic lung colonization experiments, subconfluent B16F0 monolayers were detached and dissociated to single cells using 0.25% trypsin-EDTA (Sigma-Aldrich) followed by 2X washing in serum-free DMEM. Cells were injected into mice via the tail vein on day 28 in the asthma model, day 21 in the hypersensitivity pneumonitis model, or day 7 after BLM instillation. In select experiments, B16F0 cells were mixed 1:2 with MoDM (sorted from the inflamed lung) or monocytes (isolated from bone marrow using EasySep mouse monocyte isolation kit; STEMCELL Technologies) and injected into the tail vein of uninflamed mice. For late-stage macroscopic lesion assays, mice were injected with 2.5 × 105 B16F0 cells in 100 μL of DMEM and euthanized 14 days later for the counting of visible pigmented melanoma lesions. For early-stage microscopic lesion assays, 1 × 106 B16F0 cells expressing EGFP or labeled with the CellTracker Green CMFDA fluorescent dye (5-chloromethylfluorescein diacetate; Thermo Fisher Scientific) were injected into the tail vein of mice in 100 μL of DMEM with or without 5 mg/kg AlexaFluor647-dextran (Molecular Probes) to label the microvasculature. Depending on the assay, the animals were euthanized between 4 and 96 hours after tumor cell inoculation, and the lungs were processed for immunofluorescence or flow cytometry as described below.

Imaging and immunofluorescence

The right lung was dissected, fixed by immersion in 4% paraformaldehyde in PBS for 1–2 hours, and lung tissue was embedded in OCT or NEG-50 (Thermo Fisher Scientific), frozen, and sectioned at 10–12 μm thickness for immunofluorescence or 35 μm for extravasation experiments. Sections prepared for analysis of extravasation were immediately mounted and imaged. Sections prepared for immunofluorescence were refixed in −20°C acetone, dried, rehydrated in PBS, blocked in 10% normal goat serum (Sigma-Aldrich), and stained with primary antibodies against CD11c (N418; 1:100 dilution) and CD11b (M1/70; 1:100) followed by staining with fluorescently labeled, species-specific secondary antibodies (primary and secondary antibodies are from eBioscience/Thermo Fisher Scientific). Images were acquired on an Olympus FV1000 confocal microscope with 405, 488, 543, and 633 nm laser lines or on a Leica SP5 confocal microscope with 405, 488, 561, and 633 nm laser lines. Images were analyzed using the Fiji distribution package of ImageJ 1.51 (NIH).

RNA-seq and analysis

Total RNA from subconfluent B16F0, F1, and F10 cell monolayers was isolated using the QIAshredder followed by RNeasy kit (Qiagen). Sample integrity was tested on an Agilent Bioanalyzer 2100 RNA 6000 Nano chip (5067-1511), and samples with an RNA integrity number > 8 were used to prepare libraries following the standard protocol for the TruSeq Stranded mRNA library kit (Illumina). Paired-end (PE75) sequencing was then performed on an Illumina MiSeq, and the output generated bcl file was demultiplexed by bcl2fastq2. Demultiplexed read sequences were then aligned to the Mouse Genome mm10 reference sequence using TopHat splice junction mapper with Bowtie 2 (http://ccb.jhu.edu/software/tophat/index.shtml) aligner. Assembly and differential gene expression were estimated using Cufflinks (http://cole-trapnell-lab.github.io/cufflinks/). The complete B16 RNA-seq data set is deposited at the NCBI Gene-Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/; accession #GSE182970). An expression data set for macrophages including MoDM isolated from BLM-treated mice, previously generated by Misharin and colleagues (18), was reanalyzed.

Flow cytometry and tissue processing

Lungs from animals that had been inoculated with EGFP-expressing B16F0 cells 96 hours earlier (i.e., the early, micrometastatic lesion stage) were minced and enzymatically dissociated using 5 mL of 1 mg/mL collagenase IV, 0.05 mg/mL DNase I, 0.5 U/mL dispase (all from Worthington), 1 U/mL bovine hyaluronidase (Sigma-Aldrich), and 0.1 mg/mL liberase TM high (Roche) in RPMI-1640 at 37°C for 45 minutes. Alternatively, lungs from animals that had been inoculated with B16F0 cells 14 days earlier (i.e., the late, macrometastatic lesion stage) were minced and dissociated with collagenase IV and DNase I only. In both cases, dissociated tissues and cells were pushed through a 70-μm cell strainer, red blood cells were lysed with ACK lysis buffer, and the remaining material passed through a 35-μm strainer to generate a single-cell suspension. Isolated single cells were then incubated with 2.4G2 tissue culture supernatant to block Fc receptors and immuno-labeled for cell-surface antigens for 20 minutes at 4°C in flow cytometry buffer (PBS, 2% bovine serum albumin, 2 mmol/L EDTA). For MoDM isolation, a MACS magnetic depletion step (Miltenyi Biotech) using biotin-labeled TCRβ (H57-597), NK1.1 (PK136), Ly6G (1A8), B220 (12-0452-82) was conducted followed by fluorescence activated cell sorting for CD11b+CD11c+SiglecFloLy6GLy6CDAPI cells. For intracellular labeling, the transcription factor staining buffer set was used to fix, permeabilize, and label cells as per manufacturer's instructions (Thermo Fisher Scientific). The following antibodies against mouse antigens were used for flow cytometry: CD11c (N418), F4/80 (BM8), CD11b (M1/70), Gr1 (RB6-8C5), MHCII (M5/114.15.2), CD45 (30-F11), CD206 (C068C2), Ki67 (SolA15), Ly6C (HK1.4), Ly6G (1A8), SiglecF (E50-2440), PD-L1 (MIH5), B220 (RA3-6B2), Bcl-2 (10C4) from eBioscience/Thermo Fisher Scientific; 7/4 (Ly6B.2), CD3 (145-2C11) from Ablab (UBC); and GFP (FM264G) from BioLegend. APC-conjugated Arginase-1 sheep polyclonal antibody was purchased from R&D Systems. Zombie Aqua fixable viability kit was used to identify live cells (BioLegend). Flow cytometry was performed on a BD LSRII equipped with 405, 488, 561, and 633 nm laser lines or Attune NxT Flow cytometer (Thermo Fisher Scientific), and analyzed using FlowJo (FlowJo LLC). Data were analyzed in GraphPad Prism 6.0, two-tailed unpaired Student t test was used when comparing different biological samples, Welch correction was used when comparing between time points after inflammatory insult, or between inflammatory and control conditions due to the possibility of unequal variance introduced by the induction of inflammation (19).

Western blot and ELISA analysis

B16F0 cells were serum-starved overnight. Pretreatment with cMET kinase inhibitors SU11274 (20) or capmatinib (ref. 15; MedChemExpress; HY-12014 and HY-13404, respectively) was for 1 hour followed by stimulation with mouse recombinant HGF (PeproTech). Cells were rapidly washed with ice-cold PBS and lysed in RIPA lysis buffer supplemented with protease and phosphatase inhibitors (Roche). Proteins were separated by SDS-PAGE, transferred to PVDF membranes, blocked with 5% BSA, and probed with primary antibodies to phospho-MET (#3077; Cell Signaling Technology) or total cMET (#37-0100, Thermo Fisher Scientific). Following chemiluminescent detection, densitometry was performed using ImageJ software (NIH). To measure HGF levels in media conditioned by either MoDM or monocytes, a commercial ELISA kit (#MHG00, R&D Systems) was used.

cMET inhibitor treatment in vivo

A single dose of 2.5 U/kg BLM was instilled in mice via intratracheal injection to induce sterile inflammation. Seven days later, EGFP-expressing B16F0 cells were inoculated into the mice, and 25 μg/100 μL of capmatinib in 2% DMSO or 2% DMSO vehicle control alone were given to mice by intratracheal instillation for four consecutive days. Mice were euthanized one day after the last intratracheal instillation (i.e., 96 hours after tumor cell inoculation), and lungs were processed for flow cytometry as described above. Cell numbers were calculated using counting beads (123 count eBeads, eBioscience/Thermo Fisher Scientific) by flow cytometry.

Lung inflammation stimulates the formation of end-stage metastatic lesions after intravenous injection of B16F0 cells

To first establish if and what types of inflammation might contribute to priming of the premetastatic niche, we opted to use a mouse melanoma B16 cell subline, B16F0, that colonizes the lungs at very low efficiency (i.e., it has a low metastatic potential; ref. 21). Before introducing the melanoma cells into the circulation, we first subjected mice to divergent types of inflammation: ovalbumin-mediated allergic asthma (OVA), Saccharopolyspora rectivirgula antigen-mediated hypersensitivity pneumonitis (SR), or bleomycin-mediated sterile inflammation (BLM; Fig. 1A). Masson's trichrome staining for collagen deposition (Fig. 1B and Supplementary Fig. S1 for a higher power view) and cellular infiltrates (Fig. 1C) demonstrated considerable lung inflammation and damage in all three preconditioning approaches. Although the total number of cells obtained by bronchoalveolar lavage (BAL) increased over all models used (Fig. 1C), distinct profiles of infiltrating immune cells were generated in each condition (Fig. 1D). OVA treatment generated the largest proportions of eosinophils, which is consistent with the expected type 2 immune response (22). Intranasal SR treatment generated the largest proportions of lymphocytes, consistent with the reported hypersensitivity pneumonitis phenotype of lymphocytic alveolitis (23), a pathology driven by a Th1/17 response (24). BLM treatment—which causes DNA damage, death of the lung epithelia, and an associated sterile inflammation—showed recruitment of lymphocytes, eosinophils, and neutrophils (Fig. 1D; ref. 17).

Given the differences in the immune cell infiltrates in each of these models, it was striking that, compared with PBS-treated control mice, all three inflammatory treatments resulted in a significant increase in the number of end-stage macroscopic metastatic lesions (i.e., “lung nodules”) 14 days after the intravenous injection of 2.5 × 105 B16F0 cells (Fig. 1E and F). These data indicate that an inflammatory lung microenvironment promotes metastatic colonization and/or metastatic lesion expansion, regardless of the origin or type of inflammation.

The inflammatory microenvironment does not increase tumor cell extravasation

Inflammation could precondition metastatic colonization of the lung by stimulating the exit of tumor cells from the vasculature. Given that damage and infection of tissues activate the vascular endothelium, leading to the upregulation of adhesion ligands and an increased permeability (25), we suspected that inflammation might increase tumor cell extravasation. To assess this, we labeled B16F0 cells (CMFDA, green) and coinjected them with fluorescently-labeled high-molecular-weight dextran (Alexa Fluor 647, pseudocolored red) to mark the endothelium. After four hours, a post-injection time that is optimal for this assay when using B16 cells (26), the lungs were sectioned, imaged by confocal microscopy, and scored for tumor cells that were found within (colocalized, yellow) or outside (green) of the vasculature and within the lung interstitium (Fig. 2A). The acquisition of serial confocal optical sections followed by 3D reconstruction demonstrated tumor cells either within or extravasated out of clearly defined vascular structures (Supplementary Fig. S2). We found that none of the three models of lung inflammation increased the proportion of tumor cell extravasation compared with uninflamed lungs (Fig. 2B). We previously determined that extravasation of B16 melanoma cells occurs in capillary beds (26), whereas leukocytes extravasate from the post-capillary venules (27). Thus, the two cell types may utilize different mechanisms and adhesion cascades for their exit from the vasculature. Regardless, changes in tumor cell extravasation do not appear to be responsible for the increase in metastatic colonization in response to inflammation in these models.

Figure 2.

Lung inflammation does not enhance tumor cell extravasation. In vivo extravasation assay in mice treated with PBS (control) or in various inflamed models. One million B16F0 cells labeled with CMFDA (green) were intravenously injected together with AlexaFluor-647-labeled dextran (red) which labeled the microvasculature. After four hours, lungs were harvested, sectioned, and imaged as shown in A. Tumor cells within the microvasculature are shown in yellow due to colocalization with endothelium (indicated with white arrowheads). Tumor cells that extravasated into the lung interstitium are green (open arrowheads). B, The ratio of green/yellow cells per section was quantified as percent extravasation.

Figure 2.

Lung inflammation does not enhance tumor cell extravasation. In vivo extravasation assay in mice treated with PBS (control) or in various inflamed models. One million B16F0 cells labeled with CMFDA (green) were intravenously injected together with AlexaFluor-647-labeled dextran (red) which labeled the microvasculature. After four hours, lungs were harvested, sectioned, and imaged as shown in A. Tumor cells within the microvasculature are shown in yellow due to colocalization with endothelium (indicated with white arrowheads). Tumor cells that extravasated into the lung interstitium are green (open arrowheads). B, The ratio of green/yellow cells per section was quantified as percent extravasation.

Close modal

Inflammation promotes the formation of early-stage micrometastatic lesions

Given that inflammation did not affect tumor cell extravasation, we extended our analysis to 96 hours after inoculation to determine if there were any changes in early-stage micrometastatic lesion formation in the lung interstitium. These experiments were performed using B16F0 cells stably expressing EGFP (Fig. 3A). At this time point, both the number and the size of EGFP-positive microscopic metastatic tumor cell clusters were significantly increased in the lung interstitia of BLM-, SR-, and OVA-inflamed animals compared with uninflamed controls (Fig. 3B,D). BLM-mediated inflammation also increased the total number of EGFP-positive B16F0 cells that could be isolated from the lung 96 hours after tumor cell inoculation (Fig. 3E,G).

Figure 3.

Inflammation increases the number and size of early micrometastatic lesions and tumor cell survival. One million EGFP-expressing B16F0 cells were intravenously injected into PBS-treated control or inflamed mice. Lungs were harvested 96 hours later for frozen sectioning and flow cytometry. Experimental schematics and representative images for uninflamed (PBS) and inflamed (BLM) lung are shown in A. White circles indicate micrometastatic lesions formed by small clusters of EGFP-positive tumor cells. Scale bar, 2.5 mm. B, The numbers of micrometastatic lesions were quantified from at least three sections per lung. All three inflammation conditions (BLM, SR, and OVA) were normalized to the average of uninflamed PBS-treated control. m = 5. C, Representative confocal images of EGFP-positive micrometastatic tumor cell clusters from uninflamed PBS-treated control and inflamed BLM-treated lungs. Scale bar, 50 μm. D, The size of micrometastatic tumor cell clusters within each lesion, determined by measuring cluster area. A total of 28 clusters from 5 PBS-treated mice and 64 clusters from 5 BLM-treated mice were analyzed. E–G, Cell suspensions were generated from PBS- and BLM-treated lungs and analyzed for EGFP-positive tumors cells and CD45+ cells within the live gate (E). Total EGFP-positive tumor cell numbers per lung (F) and percent EGFP-positive cells within the live-cell population (G) were determined. m = 8. H–L, Flow-cytometric analysis of ex vivo EGFP-positive B16F0 cells from uninflamed (PBS) and inflamed (BLM) lungs. H, Representative flow cytometry plots of viability and Ki67 staining. I, Histogram overlay of Bcl-2 expression. Fluorescence minus one (FMO) serves as a negative staining control. Plots indicate percent of live from the entire tumor cell population (J), Ki67-expressing tumor cells within the live population (K), and mean fluorescence intensity (MFI) of Bcl-2 staining normalized to the average of uninflamed lungs (L). Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Figure 3.

Inflammation increases the number and size of early micrometastatic lesions and tumor cell survival. One million EGFP-expressing B16F0 cells were intravenously injected into PBS-treated control or inflamed mice. Lungs were harvested 96 hours later for frozen sectioning and flow cytometry. Experimental schematics and representative images for uninflamed (PBS) and inflamed (BLM) lung are shown in A. White circles indicate micrometastatic lesions formed by small clusters of EGFP-positive tumor cells. Scale bar, 2.5 mm. B, The numbers of micrometastatic lesions were quantified from at least three sections per lung. All three inflammation conditions (BLM, SR, and OVA) were normalized to the average of uninflamed PBS-treated control. m = 5. C, Representative confocal images of EGFP-positive micrometastatic tumor cell clusters from uninflamed PBS-treated control and inflamed BLM-treated lungs. Scale bar, 50 μm. D, The size of micrometastatic tumor cell clusters within each lesion, determined by measuring cluster area. A total of 28 clusters from 5 PBS-treated mice and 64 clusters from 5 BLM-treated mice were analyzed. E–G, Cell suspensions were generated from PBS- and BLM-treated lungs and analyzed for EGFP-positive tumors cells and CD45+ cells within the live gate (E). Total EGFP-positive tumor cell numbers per lung (F) and percent EGFP-positive cells within the live-cell population (G) were determined. m = 8. H–L, Flow-cytometric analysis of ex vivo EGFP-positive B16F0 cells from uninflamed (PBS) and inflamed (BLM) lungs. H, Representative flow cytometry plots of viability and Ki67 staining. I, Histogram overlay of Bcl-2 expression. Fluorescence minus one (FMO) serves as a negative staining control. Plots indicate percent of live from the entire tumor cell population (J), Ki67-expressing tumor cells within the live population (K), and mean fluorescence intensity (MFI) of Bcl-2 staining normalized to the average of uninflamed lungs (L). Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Close modal

EGFP-positive tumor cells isolated from lungs at the micrometastatic 96 hours timepoint were assessed by flow cytometry for their viability (live/dead), proliferative status (Ki67), and expression of the prosurvival protein Bcl-2 (Fig. 3H and I). Quantitative analysis of these data indicated that the percentage of live tumor cells was increased in the inflamed lungs (Fig. 3J). Although the percentage of Ki67-positive tumor cells did not increase in the inflamed lungs (Fig. 3K), the percentage of Bcl-2-positive tumor cells did (Fig. 3L). Taken together, these data suggest that inflammation promotes early micrometastatic lesion formation by facilitating the survival of B16F0 tumor cells in the lung interstitium.

Inflammation induces an influx of monocytes that differentiate into a distinct CD11b+CD11c+ MoDM population not present in the uninflamed lung

We sought to determine if a common feature of the divergent types of inflammation in each of the three models used above constituted a shared step in preconditioning the metastatic niche. As each model generated different proportions of neutrophils, eosinophils, and lymphocytes in the BAL, we focused our attention toward monocytes, as all forms of inflammation are known to cause an influx of these highly motile, differentiation-capable cells from the circulation into affected tissues (28). For operational reasons, BLM-inflamed lungs, which receive a single inflammatory insult as opposed to the multiple insults delivered in the SR and OVA models (see Fig. 1), were analyzed for myeloid populations that may contribute to the premetastatic niche. However, it is possible that SR and OVA models generate different myeloid populations.

We characterized the immune infiltrate from the inflamed lungs by flow cytometry (Fig. 4A) and determined that the peak infiltrate occurred at seven days after BLM treatment (Fig. 4B). We first distinguished nonneutrophil immune cells by gating on the CD45+Ly6G infiltrates. From this cell population, we analyzed the SiglecF+ and CD11c+ cells, which together mark alveolar macrophages (AM; Fig. 4A). Consistent with other inflammatory models (29, 30) and the data presented in Fig. 1D above, this AM population was depleted in BLM-treated lungs (Fig. 4A and C). Inflammatory Ly6Chi monocytes, on the other hand, infiltrated the lungs rapidly after BLM treatment (i.e., day 3) and subsequently slowly decreased over time in a manner similar to neutrophils (compare Fig. 4D and E).

Figure 4.

BLM treatment induces an influx of monocytes that differentiate into monocyte-derived macrophages, a distinct CD11b±CD11c± population in the inflamed lung. A–H, Flow-cytometric analysis of immune cell populations from the uninflamed (PBS) and inflamed (BLM) lungs at indicated time points. A, CD45+Ly6G cells were analyzed using a SiglecF versus CD11c plot (top) or CD11c versus CD11b plot (bottom). Three major immune cell populations, AM (SiglecF+CD11c+), eosinophils (SiglecF+CD11c), and recruited differentiating macrophages (CD11c+SiglecFlo) are shown. SiglecFlo cells were further analyzed on a F4/80 versus CD11c plot to show the formation of a MoDM (F4/80+CD11c+) population over time. B–F, Quantification of total cell numbers, and various immune cell populations. G, MFI of SiglecF staining in AM, eosinophils (Eos) from PBS-treated control (closed circles), and MoDM from BLM-treated (open circles) lungs as analyzed in A (top). H, Percent Ly6C expressing MoDM over the BLM treatment time course. I–J, Arginase-1 expression is mainly in the MoDM population. Mice were first treated with PBS (control) or BLM (inflamed lungs) and injected with B16F0 cells on day 7. Lungs were harvested 96 hours later for single-cell analysis of cell population expressing Arginase-1 by flow cytometry. I, Gating strategy. Ly6G cells were plotted on a Arginase-1 versus PD-L1 plot. Arginase-1 and PD-L1 double-positive population was further analyzed on a SiglecF versus CD11c plot to differentiate the various immune cell populations. J, Total lung cells were first gated for the MoDM population (CD45+Ly6GSiglecFloCD11c+) and percentage of Arginase-1–expressing cells were determined in PBS- and BLM-treated lungs without and with B16F0 cell injection. All data were pooled from two experiments with m = 5–6. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Figure 4.

BLM treatment induces an influx of monocytes that differentiate into monocyte-derived macrophages, a distinct CD11b±CD11c± population in the inflamed lung. A–H, Flow-cytometric analysis of immune cell populations from the uninflamed (PBS) and inflamed (BLM) lungs at indicated time points. A, CD45+Ly6G cells were analyzed using a SiglecF versus CD11c plot (top) or CD11c versus CD11b plot (bottom). Three major immune cell populations, AM (SiglecF+CD11c+), eosinophils (SiglecF+CD11c), and recruited differentiating macrophages (CD11c+SiglecFlo) are shown. SiglecFlo cells were further analyzed on a F4/80 versus CD11c plot to show the formation of a MoDM (F4/80+CD11c+) population over time. B–F, Quantification of total cell numbers, and various immune cell populations. G, MFI of SiglecF staining in AM, eosinophils (Eos) from PBS-treated control (closed circles), and MoDM from BLM-treated (open circles) lungs as analyzed in A (top). H, Percent Ly6C expressing MoDM over the BLM treatment time course. I–J, Arginase-1 expression is mainly in the MoDM population. Mice were first treated with PBS (control) or BLM (inflamed lungs) and injected with B16F0 cells on day 7. Lungs were harvested 96 hours later for single-cell analysis of cell population expressing Arginase-1 by flow cytometry. I, Gating strategy. Ly6G cells were plotted on a Arginase-1 versus PD-L1 plot. Arginase-1 and PD-L1 double-positive population was further analyzed on a SiglecF versus CD11c plot to differentiate the various immune cell populations. J, Total lung cells were first gated for the MoDM population (CD45+Ly6GSiglecFloCD11c+) and percentage of Arginase-1–expressing cells were determined in PBS- and BLM-treated lungs without and with B16F0 cell injection. All data were pooled from two experiments with m = 5–6. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Close modal

Additionally, at the peak of inflammation seven days after BLM treatment, a distinct population of F4/80+ CD11b+ CD11c+ macrophages appeared that acquired low levels of SiglecF over time (Fig. 4A, F, and G). The drop in Ly6C+ classic monocytes in the inflamed lungs 4–14 days after BLM treatment (Fig. 4E and H) coincided with the rise of the F4/80+CD11c+ macrophages (Fig. 4F) and the modest increase in expression of SiglecF (Fig. 4G). These changes are indicative of monocytes differentiating into MoDM—and eventually into AM as they gain higher levels of SiglecF—a process that has been documented previously in BLM- and LPS-mediated inflammation models (18, 29).

The population of inflammation-induced MoDM we identify here in the lung are F4/80+, Ly6G, Ly6C, CD11b+, CD11c+, and SiglecFlo (Fig. 4A). Arginase-1, a marker identified in alternatively activated M2 macrophages (31), also appeared in the BLM-induced immune infiltrates. This was prominent in a subset of cells expressing PD-L1, an inhibitory ligand for checkpoint blockage (32), and with the CD11c+, SiglecFlo, and MoDM phenotype (Fig. 4I). Arginase-1 and PD-L1 macrophages are found in the resolving phases of inflammation and can be expressed in tumor-associated macrophages (33). We found that the percentage of Arginase 1-positive MoDM was a result of the inflammatory preconditioning of the metastatic niche rather than by the tumor cells themselves (Fig. 4J), and this PD-L1, CD11c+, MoDM population was sustained throughout the growth of the metastatic nodules (see Fig. 5).

Figure 5.

MoDM correlate to tumor burden and are found in proximity to early-stage B16F0 micrometastatic clusters in BLM-inflamed lungs. A–D, Mice were treated with PBS (control) or BLM (inflamed lungs) and injected with EGFP-expressing B16F0 on day 7. A, Lungs were harvested 96 hours later for immunofluorescence staining with the following pseudocolors B16F0 (white), CD11b (green), CD11c (red), DAPI (blue). A 50-μm margin was drawn around the center of the EGFP-positive cell cluster (white dotted line). B, Single and merged channel fluorescent images of immunostained lung sections. Pseudocolors of various channels are as indicated; note that CD11b+CD11c+ cells appear yellow in the merged image due to dual staining. C, Quantification of CD11c+, CD11b+, or CD11c+CD11b+ cells within the 50-μm margin from the micrometastatic tumor cell cluster, normalized to cluster area. D, Quantification of the total number of CD11c+, CD11b+, and CD11c+CD11b+ cells either outside (nonmarginal; black bars) or inside (B16-marginal; open bars) the 50-μm margin surrounding micrometastatic tumor cell clusters. Measurements were pooled from two experiments of m = 6. A total of 28 and 56 micrometastatic lesions from PBS- and BLM-treated mice, respectively, were analyzed. At least three lung sections per animal were analyzed. E–I, Mice were treated with PBS (control) or BLM (inflamed lungs) and injected with B16F0 on day 7. Lungs were harvested 14 days later for flow-cytometric analysis. E, Using mCherry as the reporter for B16F0, tumor cells can primarily be identified as negative for CD45 and having a large cell size (FSC-Ahi). F, Representative flow cytometry plots of cells from either PBS-treated mice not injected with B16F0 cells (No B16), PBS-treated mice injected with B16F0 cells (PBS + B16), or BLM-treated mice injected with B16F0 cells (BLM + B16). The number and percentage of B16F0 cells recovered from the injected PBS-treated (solid circles) and BLM-treated (open circles) lungs are enumerated in G. H, Flow-cytometric analysis of cells from BLM-treated inflamed lungs harvested 14 days after B16F0 cell injection. The monocytes (blue) and MoDM (red) immune cell populations were further analyzed for PD-L1 and CD206 expression. I, Correlation of tumor and immune cell populations in lungs from BLM-treated mice. Spearman rank-order correlations are shown; significance from nonzero slope and the correlation coefficients are inset. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Figure 5.

MoDM correlate to tumor burden and are found in proximity to early-stage B16F0 micrometastatic clusters in BLM-inflamed lungs. A–D, Mice were treated with PBS (control) or BLM (inflamed lungs) and injected with EGFP-expressing B16F0 on day 7. A, Lungs were harvested 96 hours later for immunofluorescence staining with the following pseudocolors B16F0 (white), CD11b (green), CD11c (red), DAPI (blue). A 50-μm margin was drawn around the center of the EGFP-positive cell cluster (white dotted line). B, Single and merged channel fluorescent images of immunostained lung sections. Pseudocolors of various channels are as indicated; note that CD11b+CD11c+ cells appear yellow in the merged image due to dual staining. C, Quantification of CD11c+, CD11b+, or CD11c+CD11b+ cells within the 50-μm margin from the micrometastatic tumor cell cluster, normalized to cluster area. D, Quantification of the total number of CD11c+, CD11b+, and CD11c+CD11b+ cells either outside (nonmarginal; black bars) or inside (B16-marginal; open bars) the 50-μm margin surrounding micrometastatic tumor cell clusters. Measurements were pooled from two experiments of m = 6. A total of 28 and 56 micrometastatic lesions from PBS- and BLM-treated mice, respectively, were analyzed. At least three lung sections per animal were analyzed. E–I, Mice were treated with PBS (control) or BLM (inflamed lungs) and injected with B16F0 on day 7. Lungs were harvested 14 days later for flow-cytometric analysis. E, Using mCherry as the reporter for B16F0, tumor cells can primarily be identified as negative for CD45 and having a large cell size (FSC-Ahi). F, Representative flow cytometry plots of cells from either PBS-treated mice not injected with B16F0 cells (No B16), PBS-treated mice injected with B16F0 cells (PBS + B16), or BLM-treated mice injected with B16F0 cells (BLM + B16). The number and percentage of B16F0 cells recovered from the injected PBS-treated (solid circles) and BLM-treated (open circles) lungs are enumerated in G. H, Flow-cytometric analysis of cells from BLM-treated inflamed lungs harvested 14 days after B16F0 cell injection. The monocytes (blue) and MoDM (red) immune cell populations were further analyzed for PD-L1 and CD206 expression. I, Correlation of tumor and immune cell populations in lungs from BLM-treated mice. Spearman rank-order correlations are shown; significance from nonzero slope and the correlation coefficients are inset. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; unpaired Student t test with Welch correction.

Close modal

CD11b+CD11c+ MoDM localize to micrometastases and increase in numbers during macrometastatic lesion formation in inflamed lungs

For CD11b+CD11c+ MoDM to have an effect on the premetastatic niche, they would need to localize near the regions where micrometastases form. We therefore stained either uninflamed (PBS-treated) or inflamed (BLM-treated) lungs 96 hours after tumor cell injection with anti-CD11b and CD11c antibodies to determine the spatial relationship between this population and the micrometastatic lesions in situ. We found that single-positive CD11b+ and CD11c+ cells represented distinct populations in uninflamed lungs and that double-positive (CD11b+CD11c+) MoDM were greatly increased in inflamed tissue (Fig. 5A,C). Importantly, these CD11b+CD11c+ MoDM, and not singly positive CD11b+ or CD11c+ populations, were enriched in close proximity to the micrometastatic lesions in the inflamed lung (Fig. 5B and C). Interestingly, there was a similar increase in the total number of myeloid cells within the tumor margin in control lungs compared with the total number of myeloid cells outside the margin (Fig. 5D). This suggests that, rather than increasing the total myeloid cell numbers at micrometastatic sites, inflammatory BLM treatment instead skewed the composition of these cells to favor the inflammation-derived pulmonary MoDM.

We next determined if these MoDM persist as B16F0 cell numbers increase during the formation of end-stage macrometastatic lesions. Using mCherry-labeled B16F0 cells, we demonstrated that the tumor cells could be delineated from leukocytes given that they were CD45, FSC-Ahi (Fig. 5E). We then used this gating strategy to numerate the number of tumor cells under control (PBS) and inflamed (BLM) conditions (Fig. 5F). Consistent with BLM sterile inflammation increasing the number of pulmonary macrometastases at day 14 after tumor injection (see Fig. 1E and F), it also significantly increased the number and percentage of B16F0 cells (Fig. 5G). SiglecFlo, CD11c+, and F4/80+ MoDM remained present in the lung at this end-stage timepoint in the BLM-treated animals, and these cells uniformly expressed PD-L1 as well as CD206, both present on alternatively activated macrophages and AM (Fig. 5H). The percentage and number of MoDM were also positively correlated with the B16F0 tumor cell load whereas the number of neutrophils and monocytes were not (Fig. 5I). These data indicate that inflammation-induced MoDM selectively persist and increase in numbers as end-stage macrometastatic lesions emerge over time.

Inflammation-induced MoDM are sufficient to enhance the metastasis of injected B16F0 cells in the lung

The pulmonary MoDM in inflamed lungs that associate with early micrometastatic lesions and persist as macrometastatic lesions emerge could, on their own, functionally contribute to the metastatic progression. To test this, we isolated the CD11b+CD11c+ MoDM from the lungs of BLM-treated GFP+ mice and coinjected them along with B16F0 cells at a 2:1 ratio into the circulation of mice with uninflamed lungs. We then assessed metastasis 14 days later (Fig. 6A). These adoptive transfer experiments were revealing; the addition of MoDM increased nodule formation 5-fold from ∼5 tumors per lung to ∼25 (Fig. 6B). Importantly, this effect was specific to MoDM as the adoptive transfer of Ly6C+ monocytes did not enhance B16F0 metastasis (Fig. 6C). This indicates that these MoDM specifically are capable of conditioning the metastatic niche.

Figure 6.

Isolated MoDM increase metastasis in uninflamed lungs and their CM promotes B16F0 survival under stress conditions. A, Experimental scheme; MoDM were isolated from the lungs of GFP+ mice 11 days after BLM treatment (i.e., equivalent to the 96h time point for micrometastatic lesion formation) by fluorescent activated cell sorting. Half a million sorted MoDM or bone marrow monocytes (BM-Mono) were mixed with 2.5 × 105 B16F0 cells and intravenously injected into naïve wild-type (WT) mice, followed by assessment for macrometastatic lesion assays 14 days later. Alternatively, sorted MoDM or BM-Mono were cultured in vitro for 24–48 hours, and the conditioned media (CM) were harvested. B and C, Macroscopic quantification of lung melanoma nodules 14 days after injection of B16F0 cells alone (B16F0), B16F0 cells together with GFP-expressing MoDM (B16F0 + MoDM), or bone marrow monocytes (B16F0 + mono) in naïve mice. As a positive control, mice were instilled with BLM, followed by B16F0 cell inoculation (BLM B16F0) on day 7. D, Representative flow cytometry plots identifying GFP+CD45+ MoDM cells isolated from the lungs of animals 14 days after their coinjection into mice with B16F0 cells. E, MFI of cell-surface markers on host monocytes (mono) and AM normalized to levels present on the isolated MoDM. F–I, Apoptotic and cell death assay. F, Annexin V and DAPI staining of B16F0 cells maintained in either normal B16F0 culture medium (untreated) or the same medium containing the folate synthesis inhibitor MTX in the absence or presence of the apoptosis inhibitor z-vad-fmk (z-VAD). G, Staining was also assessed in cells treated with MTX in the presence of fresh culture media (media) or culture media mixed with medium conditioned by MoDM (MoDM-CM) at 25% v:v. H and I, Quantification of proportions of live (double-negative), Annexin V single-positive (AnnV+), Annexin V/DAPI double-positive (AnnV+DAPI+), and DAPI only (DAPI+) B16F0 cells upon in vitro culturing without or with MTX and either z-VAD (left in H), MoDM-CM (right in H) or culture media mixed with medium conditioned by monocytes (mono-CM) at 25% v:v, or z-VAD (in I). Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired Student t test with Welch correction.

Figure 6.

Isolated MoDM increase metastasis in uninflamed lungs and their CM promotes B16F0 survival under stress conditions. A, Experimental scheme; MoDM were isolated from the lungs of GFP+ mice 11 days after BLM treatment (i.e., equivalent to the 96h time point for micrometastatic lesion formation) by fluorescent activated cell sorting. Half a million sorted MoDM or bone marrow monocytes (BM-Mono) were mixed with 2.5 × 105 B16F0 cells and intravenously injected into naïve wild-type (WT) mice, followed by assessment for macrometastatic lesion assays 14 days later. Alternatively, sorted MoDM or BM-Mono were cultured in vitro for 24–48 hours, and the conditioned media (CM) were harvested. B and C, Macroscopic quantification of lung melanoma nodules 14 days after injection of B16F0 cells alone (B16F0), B16F0 cells together with GFP-expressing MoDM (B16F0 + MoDM), or bone marrow monocytes (B16F0 + mono) in naïve mice. As a positive control, mice were instilled with BLM, followed by B16F0 cell inoculation (BLM B16F0) on day 7. D, Representative flow cytometry plots identifying GFP+CD45+ MoDM cells isolated from the lungs of animals 14 days after their coinjection into mice with B16F0 cells. E, MFI of cell-surface markers on host monocytes (mono) and AM normalized to levels present on the isolated MoDM. F–I, Apoptotic and cell death assay. F, Annexin V and DAPI staining of B16F0 cells maintained in either normal B16F0 culture medium (untreated) or the same medium containing the folate synthesis inhibitor MTX in the absence or presence of the apoptosis inhibitor z-vad-fmk (z-VAD). G, Staining was also assessed in cells treated with MTX in the presence of fresh culture media (media) or culture media mixed with medium conditioned by MoDM (MoDM-CM) at 25% v:v. H and I, Quantification of proportions of live (double-negative), Annexin V single-positive (AnnV+), Annexin V/DAPI double-positive (AnnV+DAPI+), and DAPI only (DAPI+) B16F0 cells upon in vitro culturing without or with MTX and either z-VAD (left in H), MoDM-CM (right in H) or culture media mixed with medium conditioned by monocytes (mono-CM) at 25% v:v, or z-VAD (in I). Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired Student t test with Welch correction.

Close modal

To determine if these adoptively transferred MoDM persist in the emerging metastatic lesions within uninflamed lungs, we assessed the prevalence of GFP-expressing CD45+ cells (Fig. 6D) in the dissociated tumor nodules by flow cytometry. At the terminal day 14 time point, the tumors retained these GFP+ cells and they were found to be F4/80+, CD11b+, CD11c+, SiglecFlo, and PD-L1+, the precise phenotype of the MoDM that we previously identified in inflamed lungs (Fig. 6E, compare with Fig. 4).

CD11b+CD11c+ MoDM secrete prosurvival factors that support tumor cell survival under stress

Although the number of MoDM surrounding micrometastatic lesions was increased in inflamed lungs, these cells did not typically contact the tumor cells directly (see Fig. 5A and B). We therefore reasoned that soluble, protective factors produced by these MoDM likely diffuse radially to reach the tumor cells within the local microenvironment. The protection against stress that leads to apoptosis can be modeled in vitro by treating adherent monolayer cultures with the folate synthesis inhibitor MTX (Supplementary Fig. S3) or by culturing B16F0 cells in suspension on the antiadhesive polymer polyHEMA (Supplementary Fig. S4A–S4C). To determine if MoDM secreted survival factors for stressed tumor cells, we isolated and sorted CD11b+CD11c+ MoDM from inflamed lungs and generated MoDM-conditioned media (MoDM-CM) in vitro. We then added it to B16F0 monolayers in the presence of MTX or to suspended cells cultured in polyHEMA-coated plates to determine the extent of cell survival. The addition of MoDM-CM significantly increased the number of live cells in both stress assays (Fig. 6F,H; Supplementary Fig. S4D–S4F). MTX treatment decreased the number of live B16F0 cells (Annexin VDAPI) and concomitantly, increased the number of early apoptotic (Annexin V+), late apoptotic and dying (Annexin V+DAPI+), as well as dead B16F0 cells (DAPI+), and this was largely prevented by z-vad-fmk, a known apoptosis inhibitor (Fig. 6F and H; ref. 34). The media conditioned by MoDM increased cell viability from ∼45% to ∼65% when cells were stressed with MTX and decreased the number of Annexin V–positive cells (Fig. 6G and H). Notably, the survival effect of MoDM-CM was specific as monocyte-CM had no effect (Fig. 6I). These data suggest that the ability of MoDM to increase B16F0 lung colonization may be mediated by their ability to secrete soluble factors that facilitate tumor cell survival under conditions of stress.

CD11b+CD11c+ MoDM secrete HGF, which is a survival factor for B16F0 cells

To identify potential survival factors secreted by these inflammation-induced MoDM, we analyzed published RNA-seq data generated by Misharin and colleagues, who previously isolated and characterized BLM-induced MoDM from the lung using a strategy very similar to ours (18). Their RNA-seq data generated using cells isolated 14 days after BLM induction indicate that MoDM express the transcripts of a number of growth factor and cytokines, many at higher levels than those of either monocytes or AM (Fig. 7A). We also performed our own comparative RNA-seq analysis examining B16F0 cells as well as the more metastatic B16F1 and B16F10 sublines looking for the expression of known cognate receptors to the putative MoDM factors (Fig. 7B). In doing so, we identified the HGF receptor, cMET, thus identifying one potential factor–receptor pair. We then confirmed that B16F0 cells express cMET protein on their cell surface, noting that the more metastatic B16F10 cells express higher levels of the receptor (Fig. 7C and D) as predicted by the RNA-seq data. We also confirmed that the MoDM-CM contains secreted HGF by ELISA (Fig. 7E). In contrast, monocytes did not secrete HGF (Fig. 7E).

Figure 7.

HGF and cMET signaling in MoDM is critical in promoting inflammation-induced metastatic colonization of the lung. A, Transcript levels of various growth factors in MoDM isolated from inflamed lungs 14 days after BLM treatment (MoDM D14) were compared with levels in monocytes from inflamed lungs (mono D14), alveolar macrophages from inflamed lungs (AM D14), and alveolar macrophages isolated from uninflamed control lungs (AM D0) as previously determined using RNA-seq by Mishrin et al. (18). B, Transcript levels of growth factors identified by Mishrin et al. (18) in MoDM from inflamed lungs were compared with the transcript levels of known cognate receptors in B16F0 (F0), B16F1 (F1) and B16F10 (F10) tumor cells as determined by RNA-seq. C, Cell-surface expression of cMET in RAW 264.7 murine macrophages (negative control) as well as B16F0 and B16F10 cells was assessed by flow cytometry (dotted line represents the isotype antibody staining control) and quantified by MFI from three pooled experiments in D. E, HGF protein levels in culture media (media), media conditioned by MoDM isolated from inflamed lungs (MoDM-CM), and media conditioned by bone marrow monocytes (mono-CM) from naïve mice were determined by ELISA. ND, not detected. F, cMET phosphorylation (phosphoMET) following serum starvation in B16F0 cells was assessed in the absence (−) and presence of HGF (+: 0.2 ng/mL; ++: 20 ng/mL) without (−) or with (+) the cMET kinase inhibitors SU11274 (5 μmol/L) or capmatinib (5 nmol/L) over the indicated time course. Total cMET is shown as the loading control. G, The ability of either media conditioned by MoDM (MoDM-CM; left) or control media (media right) to affect B16F0 cell viability (live), early apoptosis (AnnV+), late apoptosis (AnnV+DAPI+), or death (DAPI+) after MTX treatment in the absence or presence of 5 nmol/L cMET kinase inhibitor, capmatinib, was quantified. Data were pooled from two experiments for each condition. H, Representative flow cytometry plots of EGFP-expressing B16F0 cells and CD45-positive cells from lungs of uninflamed mice or inflamed (BLM) mice treated with cMET kinase inhibitor capmatinib or vehicle (control). Tumor and immune cell numbers recovered from the lung of each group are enumerated in I and J, respectively. Both females and males (11–14 weeks of age) were used. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired Student t test with Welch correction. NS, no significant differences.

Figure 7.

HGF and cMET signaling in MoDM is critical in promoting inflammation-induced metastatic colonization of the lung. A, Transcript levels of various growth factors in MoDM isolated from inflamed lungs 14 days after BLM treatment (MoDM D14) were compared with levels in monocytes from inflamed lungs (mono D14), alveolar macrophages from inflamed lungs (AM D14), and alveolar macrophages isolated from uninflamed control lungs (AM D0) as previously determined using RNA-seq by Mishrin et al. (18). B, Transcript levels of growth factors identified by Mishrin et al. (18) in MoDM from inflamed lungs were compared with the transcript levels of known cognate receptors in B16F0 (F0), B16F1 (F1) and B16F10 (F10) tumor cells as determined by RNA-seq. C, Cell-surface expression of cMET in RAW 264.7 murine macrophages (negative control) as well as B16F0 and B16F10 cells was assessed by flow cytometry (dotted line represents the isotype antibody staining control) and quantified by MFI from three pooled experiments in D. E, HGF protein levels in culture media (media), media conditioned by MoDM isolated from inflamed lungs (MoDM-CM), and media conditioned by bone marrow monocytes (mono-CM) from naïve mice were determined by ELISA. ND, not detected. F, cMET phosphorylation (phosphoMET) following serum starvation in B16F0 cells was assessed in the absence (−) and presence of HGF (+: 0.2 ng/mL; ++: 20 ng/mL) without (−) or with (+) the cMET kinase inhibitors SU11274 (5 μmol/L) or capmatinib (5 nmol/L) over the indicated time course. Total cMET is shown as the loading control. G, The ability of either media conditioned by MoDM (MoDM-CM; left) or control media (media right) to affect B16F0 cell viability (live), early apoptosis (AnnV+), late apoptosis (AnnV+DAPI+), or death (DAPI+) after MTX treatment in the absence or presence of 5 nmol/L cMET kinase inhibitor, capmatinib, was quantified. Data were pooled from two experiments for each condition. H, Representative flow cytometry plots of EGFP-expressing B16F0 cells and CD45-positive cells from lungs of uninflamed mice or inflamed (BLM) mice treated with cMET kinase inhibitor capmatinib or vehicle (control). Tumor and immune cell numbers recovered from the lung of each group are enumerated in I and J, respectively. Both females and males (11–14 weeks of age) were used. Significance indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; unpaired Student t test with Welch correction. NS, no significant differences.

Close modal

To determine if the HGF–cMET axis was functional in the B16F0 cells, we treated them with recombinant HGF (PeproTech). This increased the tyrosine phosphorylation of the cMET receptor, a measure of its activation, and this was prevented by addition of the cMET kinase inhibitors SU11274 and capmatinib (Fig. 7F). Functionally, capmatinib was able to partially blunt the prosurvival effects of MoDM-CM in the MTX assay (Fig. 7G). These data provide the proof-of-principle that specific soluble factors secreted by the CD11b+CD11c+MoDM, one of which is HGF, promote the survival of low metastatic potential B16F0 cells under stress conditions and may do so within the inflamed lung.

Inflammation-induced early lung colonization by B16F0 cells is prevented by the pharmacologic inhibition of HGF–cMET signaling

To determine the in vivo role of HGF in promoting inflammation-induced early metastatic colonization in the lung, the cMET inhibitor capmatinib, which we had shown was effective in blocking HGF signaling in vitro, was instilled intratracheally into the lungs at the time that EGFP-expressing B16F0 cells were intravenously injected into day 7 BLM-inflamed lungs and daily for three more days. Mouse lungs were then analyzed at day 11 after BLM (i.e., 96 hours after tumor cell inoculation) for the effect on early-stage lung colonization by measuring the number of EGFP-B16F0 tumor cells by flow cytometry. Strikingly, the cMET inhibitor abolished the inflammation-induced increase in early metastasis of B16F0 cells in the lung as the number of tumor cells decreased to the levels observed in the uninflamed mice (Fig. 7H and I). This was not due to capmatinib-reducing inflammation, as the BLM-induced increase in CD45+ immune cells was similar in both capmatinb and vehicle treated mice (Fig. 7J). Additionally, this effect was not sex dependent as capmatinib was equally effective in both male and female mice. These data illustrate the importance of the HGF–cMET axis in promoting inflammation-induced early metastatic lung colonization in vivo.

We have demonstrated that inflammation of diverse origins promotes the metastatic colonization of circulating B16F0 cells in the lung. Although an inflammatory microenvironment is known to have gross impacts on the vasculature, we did not find that inflamed lungs showed a preference for the extravasation of circulating B16F0 tumor cells. Previous studies had demonstrated that the adhesion of B16 melanoma cells in the lung is largely mechanical (35, 36), supporting the notion that the upregulation of adhesion ligands (e.g., ICAM and selectins) by inflammatory stimuli are unlikely to influence the arrest of these cells. Although extravasation is a rate-limiting and targetable step in the complete metastatic cascade (37), it was not facilitated by an inflamed vasculature under the carefully circumscribed conditions used here. The rolling–adhesion cascade for circulating tumor cells is presumably quite different depending on the organ of metastasis. Therefore, we cannot exclude the possibility that the inflamed vasculature helps populate the premetastatic niche in other organs or when other tumor cell types are used, particularly in spontaneous metastasis models.

Metastasis is a multistep process. In this study, we deliberately chose to focus on the impact of lung inflammation on the final steps of this process. Thus, we used various models of lung inflammation together with the intravenous delivery of tumor cells. This experimental pseudometastasis model allowed us to examine metastatic lung colonization in isolation. As such, we did not address how the systemic effects of lung inflammation affect the earlier stages of the metastatic process, including how tumor cells escape from the primary site lesion and enter the circulation. Notably, a recent paper has shown that asthma-related allergic lung inflammation in a mouse model can promote primary breast tumor cells to metastasize to the lung via a chemokine-dependent mechanism (38). Therefore, in the future it will be important to determine the relative contribution of premetastatic niche conditioning at the distant site to the overall metastatic process using spontaneous models of metastasis.

Rather than increasing cell proliferation per se, inflammation increased B16F0 melanoma cell survival in the lung interstitium, and this was associated with an upregulation of the apoptosis suppressor Bcl-2 in the early micrometastatic lesions that formed there. Bcl-2 is widely expressed in melanoma clinically (39), and it is known to promote lung metastasis experimentally (40, 41). As the initial seeding of secondary organs is associated with increased tumor cell stress, inflammation may be shifting the balance to increased tumor cell survival rather than proliferation during the early formation of micrometastatic lesions. This effect could decrease the effectiveness of cytotoxic antineoplastics and highlights the need to study targeted therapies under these conditions where applicable.

Lung inflammation primed an early micrometastatic niche by stimulating the recruitment of monocytes that become protumorigenic macrophages (i.e., MoDM). In order for tumor cells to grow at secondary sites, they need to evade immune responses while coaxing innate immune cells to support their survival (3). Importantly, the MoDM that we characterized as F4/80+SiglecFloCD11b+ and CD11c+ adopted an intermediate phenotype between monocytes and resident AM, and they localized to the vicinity of early-stage micrometastatic lesions formed by B16F0 cells, which normally have a very low malignant potential in the absence of inflammation. These MoDM were also present within the end-stage macrometastatic tumor nodules formed by B16F0 cells after 14 days. Adoptive transfer experiments showed that CD11b+CD11c+ MoDM alone were sufficient to facilitate an increase in metastatic lesion formation within uninflamed lungs that was similar to that observed in inflamed lungs. This effect was specific to MoDM as the adoptive transfer of monocytes did not increase metastasis.

Misharin and colleagues previously demonstrated that inflammation-induced SiglecFlo MoDM are generated in BLM-inflamed lungs and drive the fibrotic, wound-healing process (18). In this study, we did not formally exclude the possibility that fibrotic tissue also contributes to the premetastatic niche in these lungs as the MoDM, which themselves are not major contributors to ECM deposition (18), enhanced metastasis on their own. Given that CD11b+CD11c+ MoDM surrounded early-stage B16F0 tumor cell clusters as they were forming micrometastatic niches in the lung interstitium, we instead reasoned that MoDM may be secreting soluble factors that act in a paracrine fashion to support tumor cell survival there.

MoDM-CM partially reversed the apoptotic cell death of B16F0 cells induced by MTX-induced stress in vitro. RNA-seq data implicated the HGF–cMET axis as one potential factor that contributes to this effect in vivo, which we confirmed experimentally. In noninflammatory conditions, CD11b+F4/80+ macrophages facilitate metastatic breast cancer lesion formation, at least in part, by secreting the chemokine CCL-2 (8). And once the tumor is established, the tumor cell–dependent secretion of growth factors such as VEGFα and EGF by macrophages of monocyte lineage has been documented in both the lung and the peritoneum (6, 14). Although these likely represent additional factors that contribute to established tumor growth at distant sites, our data provide evidence for the HGF–cMET axis in, at the very least, the initial metastatic colonization of the inflamed lung. Interestingly, we found that cMET receptor expression is progressively upregulated in the B16 F1 and F10 melanoma sublines that are increasingly more successful in colonizing the lung than the low malignant potential B16F0 subline. Such data rather elegantly demonstrate that the HGF–cMET axis can be selected for during metastasis by tumor cell—intrinsic changes, as has clearly been shown in the case of B16F10 cells (42), or by extrinsic production of HGF by, for example, inflammation-induced MoDM.

During BLM-induced lung inflammation, MoDM had much higher levels of HGF mRNA than monocytes or AM (18). However, the ability of tumor-associated macrophages to produce HGF can be inducible and not restricted to inflammation-induced MoDM. For example, Kitamura and colleagues recently demonstrated that macrophages associated with highly malignant breast cancer cells secrete HGF, and this secretion further increases the metastatic load of these tumor cells in uninflamed lungs (12). In the inflammation process, MoDM provide a wound-healing function, and HGF is a multifunctional cytokine that orchestrates tissue regeneration under these conditions by stimulating the mitogenesis of epithelial cells, the angiogenesis of endothelial cells, and the chemotaxis of lymphocytes (43). Thus, by producing growth factors such as HGF to facilitate repair, inflammatory MoDM may inadvertently promote metastatic seeding regardless of the intrinsic metastatic potential of the tumor cells that arrive at the inflammatory site. To wit, we showed pharmacologic inhibition of the HGF–cMET signaling axis within the inflamed lung abolished the inflammation-induced increase in lung colonization by low malignant potential B16F0 cells. This strongly implicates both MoDM and HGF in creating a supportive premetastatic niche in the inflamed lungs. However, understanding the precise nature of the contribution of HGF in this regard will require the conditional ablation of HGF in the MoDM.

We found that the late (i.e., 14 day), end-stage macrometastatic lesions that arose within inflamed lungs could be digested and sorted to distinguish B16F0 tumor cells from stromal and immune cells. This quantitative approach revealed the continued presence of SiglecFlo, CD11c+, and F4/80+ MoDM in the lung at this late time point. These MoDM uniformly expressed PD-L1 and CD206, two molecules expressed on alternatively activated macrophages as well as AM. In addition, the percentage and numbers of MoDM were positively correlated with the end-stage tumor cell load. It is remarkable that these inflammation-induced MoDM that have characteristics of alternatively activated macrophages and secrete growth factors that promote wound healing become lesion-associated early in the establishment of the micrometastatic lesion formation and persist throughout the macrometastatic growth of the lesion. It also suggests an immunosuppressive function for the MoDM could blunt the clearance of metastatic cells by T cells. Interestingly, in the noninflamed lung, intravital imaging has revealed that an interaction of macrophages of monocyte lineage with high malignant potential B16F10 cells displaces CD103+ dendritic cells to dampen antitumor immunity in the micrometastatic niche (44). In contrast, a separate population of Ly6Clo monocytes in the blood (i.e., patrolling monocytes) is capable of inducing killing of the same highly metastatic B16F10 cells before they extravasate and enter the lung interstitium (45). Both of these populations are derived from classic inflammatory Ly6Chi monocytes, indicating that their differentiation and location vis-à-vis metastasizing tumor cells will be critical in determining their impact on the survival and, ultimately, the colonizing ability of newly arrived tumor cells during the earliest stages of the process. Thus, modulating the fate and position of monocytes to manipulate the subsequent function of monocyte-derived macrophage populations will likely have profound effects on the efficiency of metastatic take, survival, and growth in distant site organs.

Y.-H. Huang reports grants from Canadian Institutes of Health Research and grants from Natural Sciences and Engineering Research Council during the conduct of the study. M.R. Gold reports grants from Canadian Institutes of Health Research and Cancer Research Society of Canada during the conduct of the study; grants from Zymeworks, Inc. outside the submitted work. P. Johnson reports grants from Canadian Institutes of Health Research and Natural Sciences and Engineering Research Council during the conduct of the study. C.D. Roskelley reports grants from Canadian Institutes of Health Research and Cancer Research Society during the conduct of the study. No disclosures were reported by the other authors.

A.A. Arif: Conceptualization, data curation, formal analysis, investigation, methodology, writing–original draft, project administration. Y.-H. Huang: Formal analysis, investigation, methodology, writing–original draft, writing–review and editing. S.A. Freeman: Conceptualization, investigation, methodology, writing–review and editing. J. Atif: Investigation, methodology. P. Dean: Data curation, formal analysis, investigation, methodology. J.C.Y. Lai: Formal analysis, investigation, methodology. M.-R. Blanchet: Resources, methodology. K.C. Wiegand: Investigation, methodology. K.M. McNagny: Resources, methodology. T.M. Underhill: Investigation, methodology. M.R. Gold: Conceptualization, formal analysis, writing–review and editing. P. Johnson: Conceptualization, formal analysis, investigation, writing–original draft, writing–review and editing. C.D. Roskelley: Conceptualization, formal analysis, funding acquisition, writing–original draft, writing–review and editing.

We thank Kenneth Harder and Kevin Bennewith for thoughtful discussions as well as Megan Gilmour and Tate Goodman for technical assistance. We also recognize the contributions of Andy Johnson and Justin Wong from the UBC Flow Cytometry Core, the staff at the Life Sciences Institute Imaging Core, and the animal care technicians at the UBC Modified Barrier Facility. This work was supported by Canadian Institute of Health Research (CIHR) grant PJT-148663 to M.R. Gold, P. Johnson, and C.D. Roskelley, Cancer Research Society grant to M.R. Gold and C.D. Roskelley, Natural Sciences and Engineering Research Council (NSERC) grant to P. Johnson and CIHR grant PJT-148915 to T.M. Underhill. A.A. Arif acknowledges fellowship support from CIHR (CGS-M) and NSERC (CGSD-3). J. Atif acknowledges support from NSERC.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Suzuki
M
,
Mose
ES
,
Montel
V
,
Tarin
D
. 
Dormant cancer cells retrieved from metastasis-free organs regain tumorigenic and metastatic potency
.
Am J Pathol
2006
;
169
:
673
81
.
2.
Joyce
JA
,
Pollard
JW
. 
Microenvironmental regulation of metastasis
.
Nat Rev Cancer
2009
;
9
:
239
52
.
3.
Quail
DF
,
Joyce
JA
. 
Microenvironmental regulation of tumor progression and metastasis
.
Nat Med
2013
;
19
:
1423
37
.
4.
Hoshino
A
,
Costa-Silva
B
,
Shen
TL
,
Rodrigues
G
,
Hashimoto
A
,
Mark
MT
, et al
Tumour exosome integrins determine organotropic metastasis
.
Nature
2015
;
527
:
329
35
.
5.
Erler
JT
,
Bennewith
KL
,
Cox
TR
,
Lang
G
,
Bird
D
,
Koong
A
, et al
Hypoxia-induced lysyl oxidase is a critical mediator of bone marrow cell recruitment to form the premetastatic niche
.
Cancer Cell
2009
;
15
:
35
44
.
6.
Kaplan
RN
,
Riba
RD
,
Zacharoulis
S
,
Bramley
AH
,
Vincent
L
,
Costa
C
, et al
VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche
.
Nature
2005
;
438
:
820
7
.
7.
Kitamura
T
,
Qian
BZ
,
Soong
D
,
Cassetta
L
,
Noy
R
,
Sugano
G
, et al
CCL2-induced chemokine cascade promotes breast cancer metastasis by enhancing retention of metastasis-associated macrophages
.
J Exp Med
2015
;
212
:
1043
59
.
8.
Qian
B
,
Deng
Y
,
Im
JH
,
Muschel
RJ
,
Zou
Y
,
Li
J
, et al
A distinct macrophage population mediates metastatic breast cancer cell extravasation, establishment and growth
.
PLoS One
2009
;
4
:
e6562
.
9.
Cox
TR
,
Bird
D
,
Baker
AM
,
Barker
HE
,
Ho
MW
,
Lang
G
, et al
LOX-mediated collagen crosslinking is responsible for fibrosis-enhanced metastasis
.
Cancer Res
2013
;
73
:
1721
32
.
10.
Hanahan
D
,
Weinberg
RA
. 
Hallmarks of cancer: the next generation
.
Cell
2011
;
144
:
646
74
.
11.
Coussens
LM
,
Werb
Z
. 
Inflammation and cancer
.
Nature
2002
;
420
:
860
7
.
12.
Kitamura
T
,
Kato
Y
,
Brownlie
D
,
Soong
DYH
,
Sugano
G
,
Kippen
N
, et al
Mammary tumor cells with high metastatic potential are hypersensitive to macrophage-derived HGF
.
Cancer Immunol Res
2019
;
7
:
2052
64
.
13.
Qian
BZ
,
Zhang
H
,
Li
J
,
He
T
,
Yeo
EJ
,
Soong
DY
, et al
FLT1 signaling in metastasis-associated macrophages activates an inflammatory signature that promotes breast cancer metastasis
.
J Exp Med
2015
;
212
:
1433
48
.
14.
Yin
M
,
Li
X
,
Tan
S
,
Zhou
HJ
,
Ji
W
,
Bellone
S
, et al
Tumor-associated macrophages drive spheroid formation during early transcoelomic metastasis of ovarian cancer
.
J Clin Invest
2016
;
126
:
4157
73
.
15.
Hara
T
,
Kimura
A
,
Miyazaki
T
,
Tanaka
H
,
Morimoto
M
,
Nakai
K
, et al
Cabozantinib inhibits AXL- and MET-dependent cancer cell migration induced by growth-arrest-specific 6 and hepatocyte growth factor
.
Biochem Biophys Rep
2020
;
21
:
100726
.
16.
Gold
MJ
,
Antignano
F
,
Halim
TY
,
Hirota
JA
,
Blanchet
MR
,
Zaph
C
, et al
Group 2 innate lymphoid cells facilitate sensitization to local, but not systemic, TH2-inducing allergen exposures
.
J Allergy Clin Immunol
2014
;
133
:
1142
8
.
17.
Moeller
A
,
Ask
K
,
Warburton
D
,
Gauldie
J
,
Kolb
M
. 
The bleomycin animal model: a useful tool to investigate treatment options for idiopathic pulmonary fibrosis?
Int J Biochem Cell Biol
2008
;
40
:
362
82
.
18.
Misharin
AV
,
Morales-Nebreda
L
,
Reyfman
PA
,
Cuda
CM
,
Walter
JM
,
McQuattie-Pimentel
AC
, et al
Monocyte-derived alveolar macrophages drive lung fibrosis and persist in the lung over the life span
.
J Exp Med
2017
;
214
:
2387
404
.
19.
McDonald
JH
.
Handbook of biological statistics
. 3rd ed.,
Baltimore, MD
:
Sparky House Publishing
; 
2014
.
20.
Zhang
R
,
Qi
F
,
Shao
S
,
Li
G
,
Feng
Y
. 
Human colorectal cancer-derived carcinoma associated fibroblasts promote CD44-mediated adhesion of colorectal cancer cells to endothelial cells by secretion of HGF
.
Cancer Cell Int
2019
;
19
:
192
.
21.
Hart
IR
. 
The selection and characterization of an invasive variant of the B16 melanoma
.
Am J Pathol
1979
;
97
:
587
600
.
22.
Daubeuf
F
,
Frossard
N
. 
Eosinophils and the ovalbumin mouse model of asthma
.
Methods Mol Biol
2014
;
1178
:
283
93
.
23.
Blanchet
MR
,
Bennett
JL
,
Gold
MJ
,
Levantini
E
,
Tenen
DG
,
Girard
M
, et al
CD34 is required for dendritic cell trafficking and pathology in murine hypersensitivity pneumonitis
.
Am J Respir Crit Care Med
2011
;
184
:
687
98
.
24.
Simonian
PL
,
Roark
CL
,
Wehrmann
F
,
Lanham
AK
,
Diaz del Valle
F
,
Born
WK
, et al
Th17-polarized immune response in a murine model of hypersensitivity pneumonitis and lung fibrosis
.
J Immunol
2009
;
182
:
657
65
.
25.
Vestweber
D
. 
How leukocytes cross the vascular endothelium
.
Nat Rev Immunol
2015
;
15
:
692
704
.
26.
Freeman
SA
,
McLeod
SJ
,
Dukowski
J
,
Austin
P
,
Lee
CC
,
Millen-Martin
B
, et al
Preventing the activation or cycling of the Rap1 GTPase alters adhesion and cytoskeletal dynamics and blocks metastatic melanoma cell extravasation into the lungs
.
Cancer Res
2010
;
70
:
4590
601
.
27.
Ley
K
,
Laudanna
C
,
Cybulsky
MI
,
Nourshargh
S
. 
Getting to the site of inflammation: the leukocyte adhesion cascade updated
.
Nat Rev Immunol
2007
;
7
:
678
89
.
28.
Medzhitov
R
. 
Origin and physiological roles of inflammation
.
Nature
2008
;
454
:
428
35
.
29.
Dong
Y
,
Poon
GFT
,
Arif
AA
,
Lee-Sayer
SSM
,
Dosanjh
M
,
Johnson
P
. 
The survival of fetal and bone marrow monocyte-derived alveolar macrophages is promoted by CD44 and its interaction with hyaluronan
.
Mucosal Immunol
2018
;
11
:
601
14
.
30.
Ghoneim
HE
,
Thomas
PG
,
McCullers
JA
. 
Depletion of alveolar macrophages during influenza infection facilitates bacterial superinfections
.
J Immunol
2013
;
191
:
1250
9
.
31.
Mantovani
A
,
Sica
A
,
Locati
M
. 
Macrophage polarization comes of age
.
Immunity
2005
;
23
:
344
6
.
32.
Zou
W
,
Wolchok
JD
,
Chen
L
. 
PD-L1 (B7-H1) and PD-1 pathway blockade for cancer therapy: mechanisms, response biomarkers, and combinations
.
Sci Transl Med
2016
;
8
:
328rv4
.
33.
Cassetta
L
,
Pollard
JW
. 
Targeting macrophages: therapeutic approaches in cancer
.
Nat Rev Drug Discovery
2018
;
17
:
887
904
.
34.
Slee
EA
,
Zhu
H
,
Chow
SC
,
MacFarlane
M
,
Nicholson
DW
,
Cohen
GM
. 
Benzyloxycarbonyl-Val-Ala-Asp (OMe) fluoromethylketone (Z-VAD.FMK) inhibits apoptosis by blocking the processing of CPP32
.
Biochem J
1996
;
315
:
21
4
.
35.
Chambers
AF
,
Naumov
GN
,
Vantyghem
SA
,
Tuck
AB
. 
Molecular biology of breast cancer metastasis. Clinical implications of experimental studies on metastatic inefficiency
.
Breast Cancer Res
2000
;
2
:
400
7
.
36.
Chambers
AF
,
Groom
AC
,
MacDonald
IC
. 
Dissemination and growth of cancer cells in metastatic sites
.
Nat Rev Cancer
2002
;
2
:
563
72
.
37.
McDowell
SAC
,
Quail
DF
. 
Immunological regulation of vascular inflammation during cancer metastasis
.
Front Immunol
2019
;
10
:
1984
.
38.
Bekaert
S
,
Rocks
N
,
Vanwinge
C
,
Noel
A
,
Cataldo
D
. 
Asthma-related inflammation promotes lung metastasis of breast cancer cells through CCL11-CCR3 pathway
.
Respir Res
2021
;
22
:
61
.
39.
Selzer
E
,
Schlagbauer-Wadl
H
,
Okamoto
I
,
Pehamberger
H
,
Potter
R
,
Jansen
B
. 
Expression of Bcl-2 family members in human melanocytes, in melanoma metastases and in melanoma cell lines
.
Melanoma Res
1998
;
8
:
197
203
.
40.
Del Bufalo
D
,
Biroccio
A
,
Leonetti
C
,
Zupi
G
. 
Bcl-2 overexpression enhances the metastatic potential of a human breast cancer line
.
FASEB J
1997
;
11
:
947
53
.
41.
Choi
J
,
Choi
K
,
Benveniste
EN
,
Rho
SB
,
Hong
YS
,
Lee
JH
, et al
Bcl-2 promotes invasion and lung metastasis by inducing matrix metalloproteinase-2
.
Cancer Res
2005
;
65
:
5554
60
.
42.
Qian
F
,
Engst
S
,
Yamaguchi
K
,
Yu
P
,
Won
KA
,
Mock
L
, et al
Inhibition of tumor cell growth, invasion, and metastasis by EXEL-2880 (XL880, GSK1363089), a novel inhibitor of HGF and VEGF receptor tyrosine kinases
.
Cancer Res
2009
;
69
:
8009
16
.
43.
Kato
T
. 
Biological roles of hepatocyte growth factor-Met signaling from genetically modified animals
.
Biomed Rep
2017
;
7
:
495
503
.
44.
Headley
MB
,
Bins
A
,
Nip
A
,
Roberts
EW
,
Looney
MR
,
Gerard
A
, et al
Visualization of immediate immune responses to pioneer metastatic cells in the lung
.
Nature
2016
;
531
:
513
7
.
45.
Hanna
RN
,
Cekic
C
,
Sag
D
,
Tacke
R
,
Thomas
GD
,
Nowyhed
H
, et al
Patrolling monocytes control tumor metastasis to the lung
.
Science
2015
;
350
:
985
90
.