The ability of a cancer cell to migrate through the dense extracellular matrix within and surrounding the solid tumor is a critical determinant of metastasis. Macrophages enhance invasion and metastasis in the tumor microenvironment, but the basis for their effects is not fully understood. Using a microfluidic 3D cell migration assay, we found that the presence of macrophages enhanced the speed and persistence of cancer cell migration through a 3D extracellular matrix in a matrix metalloproteinases (MMP)-dependent fashion. Mechanistic investigations revealed that macrophage-released TNFα and TGFβ1 mediated the observed behaviors by two distinct pathways. These factors synergistically enhanced migration persistence through a synergistic induction of NF-κB–dependent MMP1 expression in cancer cells. In contrast, macrophage-released TGFβ1 enhanced migration speed primarily by inducing MT1-MMP expression. Taken together, our results reveal new insights into how macrophages enhance cancer cell metastasis, and they identify TNFα and TGFβ1 dual blockade as an antimetastatic strategy in solid tumors. Cancer Res; 77(2); 279–90. ©2016 AACR.

Cancer cells are surrounded by a complex tumor microenvironment consisting of extracellular matrix (ECM), tumor-associated stromal cells, and a myriad of signaling molecules (1), which can significantly influence tumor growth and metastasis (2). ECM in the tumor microenvironment acts as a barrier to metastasis, and cancer cells have enhanced capabilities to navigate through the dense 3D collagen ECM surrounding the tumor (3). To migrate through the ECM, cancer cells employ proteases, such as matrix metalloproteinases (MMP), to degrade ECM, kinases to assist in forming protrusions, and integrins to adhere to the matrix to enable movement (4). Indeed, the activities and/or expressions of these molecules have been shown to be elevated in cancer cells (5–7).

Macrophages, one of the most abundant stromal cell types in the tumor microenvironment, are key promoters of tumor metastasis (8). Various clinical data have revealed that the infiltration of macrophages in tumor tissues correlates with poor prognosis in cases of breast cancer, prostate cancer, and melanoma (9, 10). Moreover, in vivo and in vitro studies have shown that macrophages enhance cancer cell intravasation (11, 12) and invasion through various signaling pathways (13, 14). However, many of these in vitro migration studies were performed on 2D tissue culture substrates, which fail to capture the 3D microenvironment present in vivo. In addition, the majority of these studies were carried out using transwell assays, which only yield an endpoint readout of cell behaviors (15) and thereby provide little information on how macrophages affect different aspects of cancer cell migration, such as how fast or how persistently the cancer cell moves. These distinct features of migration (speed vs. persistence) describe cell migration dynamics, and they can be quantified using metrics such as total speed and directedness. Total speed (the total distance that cell migrated divided by the migration time) defines how fast a cancer cell migrates. In contrast, directedness (the ratio of cell displacement to the total distance that the cell traveled) measures the persistence of the cell movement (15, 16).

Recently, it has become increasingly clear that both migration speed and persistence determine the metastatic potential of a cancer cell (17), and stimuli that increase both of these factors can greatly enhance metastasis. More importantly, speed and persistence can be modulated independently of one another by a single stimulus. For example, inhibiting integrin has been shown to decrease cancer cell migration speed but has no effect on persistence (17). On the other hand, interstitial flow can increase cancer cell migration persistence without altering the speed of migration (18). Lending further complexity, a stimulus can affect speed and persistence of migration differently when cells are cultured on 2D substrates compared with in 3D matrix. Specifically, EGF has been shown to increase cancer cell migration speed and decrease persistence when cells are migrating on 2D surfaces. However, this same growth factor enhances both cancer cell migration speed and persistence when cells are cultured in 3D ECM (19). Collectively, these results highlight the importance of characterizing how a stimulus affects different aspects of cancer cell migration (i.e., speed and persistence) in 3D ECM to gain a detailed and quantitative understanding of metastasis. However, to our knowledge, the effects of macrophages on the dynamics (speed and persistence) of cancer cell migration in 3D ECM have not been explored.

In the current study, we employed a microfluidic 3D migration assay to examine how macrophages affect different aspects (speed and persistence) of cancer cell migration. This microfluidic assay allows us to perform real-time high-resolution imaging of cancer cells migrating in 3D collagen I ECM in the presence of macrophages, which recapitulates key aspects of their interactions in the primary tumor sites in vivo. By tracking the movement of cancer cells within the 3D ECM, we can evaluate the effects of macrophages on the dynamics of cancer cell migration in a more physiologically relevant environment than 2D in vitro assays. In addition, this microfluidic assay is better suited for the detailed mechanistic study of macrophage-assisted cancer cell migration than in vivo assays (such as intravital imaging), as it is easier to operate and offers a tightly controlled experimental environment. Using this microfluidic assay, we show that macrophages release TNFα and TGFβ1 that increase both migration speed (total speed) and persistence (directedness) of cancer cells in 3D ECM. Interestingly, macrophage-released TNFα and TGFβ1 were found to promote cancer cell migration speed and persistence through two different mechanisms. Specifically, macrophages enhance cancer cell migration speed mainly through TGFβ1-induced MT1-MMP expression in cancer cells. In comparison, macrophage-released TNFα and TGFβ1 synergistically enhance cancer cell migration persistence via NF-κB–dependent MMP1 expression. These results demonstrate, for the first time, that speed and persistence of cancer cell migration in 3D can be modulated by macrophages via different pathways, which strongly suggests that both of these pathways need to be targeted to effectively mitigate macrophage-induced metastasis.

Cell culture and reagents

MDA-MB-231 human breast carcinoma cells expressing GFP (MDA231) were kindly provided by Dr. Frank Gertler (Massachusetts Institute of Technology, Cambridge, MA). PC3 human prostate carcinoma cells (PC3), MDA-MB-435S human melanoma cells (MDA435), and Raw 264.7 mouse macrophages (Raw) were obtained from ATCC. MDA231, MDA435, and Raw cells were cultured in DMEM. PC3 were cultured in RPMI. All media were supplemented with 10% FBS and 100 U/mL penicillin/streptomycin. Cell lines were authenticated using short tandem repeat profiling (Promega).

To generate primary bone marrow–derived macrophages (BMDM), bone marrow cells were first isolated from the femurs of C57BL/6 mice. These cells were then differentiated in RPMI supplemented with 10% FBS, 1% HEPES, 40 ng/mL MCSF (Peprotech), and 50 μmol/L β-mercaptoethanol for 7 days to produce BMDM. Primary human monocyte-derived macrophages (MDMΦ) were generated from monocytes isolated from whole blood (Research Blood Component) using a Ficoll–Paque gradient and the EasySep Monocyte Enrichment Kit (STEMCELL Technologies). These cells were cultured with IMDM supplemented with 2% l-glutamine and AB serum for 7 days to generate MDMΦ. All cells were cultured in a humidified incubator at 5% CO2 and 37°C.

Microfluidic 3D cell migration assay

To quantify macrophage-assisted cancer cell migration in 3D ECM, a microfluidic cell migration assay was used (Fig. 1A; Supplementary Fig. S1A in Supplementary Information SI). This assay consists of a polydimethylsiloxane microfluidic device (20) with a collagen gel flanked by two microchannels containing media. Cancer cells (2.3 × 106 cells/mL) and/or macrophages (0.92 × 106 cells/mL) treated with Cell Tracker Red CMTPX were suspended in 2.5 mg/mL rat-tail collagen type I ECM (BD Biosciences) introduced to the central chamber of the device. For a detailed description of seeding protocols, see Supplementary Information.

Figure 1.

Macrophages enhance cancer cell migration total speed and directedness in 3D ECM. A, Schematics of the microfluidic device. Cancer cells and macrophages were suspended in 3D collagen I ECM (orange) encased in the device. B, Representative MDA-MB-231 cancer cell (MDA231) migration trajectories for cancer cell monoculture (left) and cancer cell-Raw 264.7 macrophages (MΦ) coculture (right). C and D, Coculture of Raw macrophages (MΦ) with cancer cells significantly enhanced cancer cell migration total speed (C) and directedness (D) for MDA231 cells, PC3 prostate cancer cells, and MDA-MB-435S melanoma cells (MDA435). E and F, Coculture of primary human MDMΦ, as well as murine BMDMs, with MDA231 cells enhanced migration total speed (E) and directedness (F) of MDA231 cells. **, P < 0.01; ***, P < 0.001. μm/h, μm per hour.

Figure 1.

Macrophages enhance cancer cell migration total speed and directedness in 3D ECM. A, Schematics of the microfluidic device. Cancer cells and macrophages were suspended in 3D collagen I ECM (orange) encased in the device. B, Representative MDA-MB-231 cancer cell (MDA231) migration trajectories for cancer cell monoculture (left) and cancer cell-Raw 264.7 macrophages (MΦ) coculture (right). C and D, Coculture of Raw macrophages (MΦ) with cancer cells significantly enhanced cancer cell migration total speed (C) and directedness (D) for MDA231 cells, PC3 prostate cancer cells, and MDA-MB-435S melanoma cells (MDA435). E and F, Coculture of primary human MDMΦ, as well as murine BMDMs, with MDA231 cells enhanced migration total speed (E) and directedness (F) of MDA231 cells. **, P < 0.01; ***, P < 0.001. μm/h, μm per hour.

Close modal

After overnight incubation, the microfluidic device was transferred to a fluorescent microscope (Zeiss) fitted with an environmental chamber operating at 37°C and 5% CO2. Time-lapse microscopy was employed to record cancer cell movement in the 3D collagen I ECM. Images were taken every 15 minutes for 18 hours.

ImageJ (NIH, Bethesda, MD) was used to track cancer cell movement to produce cell migration trajectories. The migration trajectories were analyzed with Chemotaxis and Migration software (Ibidi) to quantify the dynamics of cell migration, such as total speed and directedness. Total speed was calculated as total distance the cell traveled divided by migration time, whereas directedness was calculated as the displacement of the cell divided by total distance (Supplementary Fig. S1B).

As appropriate, various concentrations of neutralizing antibodies, inhibitors, or recombinant growth factors (listed in Supplementary Information) were added to the cell culture media in the device.

Real-time PCR and Western blot analysis

Total RNA was extracted and purified using the RNeasy Mini Kit and RNase-free DNase Set (QIAGEN). Expression levels of mRNA were analyzed by real-time RT-PCR using High-Capacity RNA-to-cDNA Kit and SYBR Green Master Mixture (Applied Biosystems). The sequences of the primers used can be found in Supplementary Information. Data were analyzed according to the comparative Ct method and were normalized to GAPDH expression in each sample.

Cell lysate was extracted with RIPA buffer containing protease inhibitor and PMSF. Equal amount of total protein (30 μg) was resolved on 4% to 12% NuPAGE electrophoresis gels (Invitrogen) and transferred onto nitrocellulose membranes. The membranes were probed with various primary antibodies listed in Supplementary Information, followed by secondary antibodies conjugated to horseradish peroxidase. The immunoreactive bands were detected with ECL Chemiluminescent substrates. Densitometry analysis was performed using Alpha Innotech software to quantify Western blot images. The densitometry quantification for each protein was normalized to the appropriate loading control (β-actin, GAPDH, Lamin B1) before further normalization to the control group of each experiment.

Statistical analysis

All statistical analyses were performed using GraphPad Prism with a P value of <0.05 considered statistically significant. The difference between groups was evaluated by two-tailed Student t test or one-way ANOVA. In all figures, ns represents not significant; *, P < 0.05; **, P < 0.01; and ***, P < 0.001. For cell migration quantification, bars represent mean ± SEM of data from 40 to 100 cells from three independent experiments. For Western blot and qRT-PCR quantification, bars represent mean ± SEM of data (fold increase relative to no-treatment or monoculture control) from three independent experiments.

Macrophages enhance cancer cell migration total speed and directedness in 3D ECM

To determine the effects of macrophages on the dynamics of cancer cell migration in 3D matrix, we tracked the movement of cancer cells inside the collagen I ECM in the microfluidic devices. We chose to use collagen I ECM to mimic the tumor matrix as collagen I has been shown to be the major component of tumor-associated stromal tissue (3, 21), and it has also been implicated in metastasis (22). From the cell tracking, we quantified cancer cell migration total speed and directedness (defined in Supplementary Fig. S1B), and we compared the migration dynamics of cancer cells cultured alone with that of cancer cells cocultured with Raw macrophages (Fig. 1B; Supplementary Fig. S1C). We found that Raw macrophages significantly enhanced cancer cell migration total speed and directedness (Fig. 1C and D) in 3D ECM for MDA-MB-231, PC3, and MDA-MB-435S cells. Similar to Raw macrophages, primary macrophages, such as human MDMΦ and murine BMDM, were also observed to increase cancer cell migration total speed and directedness (Fig. 1E and F) in 3D ECM. These results indicate that macrophages allow cancer cells to move faster and more persistently, contributing to increases in cancer cell invasion rate (ratio between the displacement of cell and migration time), which is an endpoint measurement of cell invasiveness (Supplementary Fig. S2). These results are in stark contrast to results obtained from a 2D migration assay, where we found that macrophages only slightly enhanced cancer cell migration total speed but markedly reduced cancer cell migration directedness (Supplementary Fig. S2F). Moreover, the abilities of macrophages to enhance cancer cell migration dynamics in 3D ECM were not affected by the seeding ratio of the cells or the addition of Matrigel into the collagen I ECM (Supplementary Fig. S2G and S2H). Hence, these results suggest that there are fundamental differences in how macrophages affect cancer cell migration on 2D substrates versus in 3D ECM.

Macrophage-induced cancer cell migration in 3D ECM is mediated via cancer cell MMP expression

Next, we investigated the molecular mechanisms that control how fast and how persistently the cancer cell migrates in 3D ECM. We hypothesized that MMPs produced by cancer cells are involved, as the migration of cells in the dense 3D matrix critically depends on their ability to degrade ECM (4, 19). To test this hypothesis, we treated MDA231 cancer cells with a pan-MMP inhibitor GM6001. We found that inhibiting MMP activities in cancer cells significantly reduced cancer cell migration total speed and directedness (Fig. 2B and C). Further evidence for the role of MMPs was obtained using confocal reflectance microscopy, which revealed that the migration of MDA231 cells in ECM produced microtracks of empty space (Fig. 2A). However, when these cells were treated with GM6001, the formation of cell protrusions, as well as the ability of cells to degrade ECM, was reduced compared with control samples (Supplementary Fig. S3A and S3B). These results illustrate that cancer cells migrate in our experimental system in an MMP-dependent fashion, and the production of MMPs is a critical determinant of cancer cell migration dynamics (total speed and directedness) in 3D ECM.

Figure 2.

Cancer cell migration speed and directedness are MMP dependent, and macrophages enhance cancer cell MMP expression. A, Representative confocal image showing MDA231 cells (green) migrating through dense collagen I ECM (magenta) by degrading the matrix, leaving behind a microtrack (arrow). B and C, Compared with the untreated and DMSO controls, inhibition of MMP activity by GM6001 significantly reduced MDA231 migration total speed (B) and directedness (C). D and E, Representative Western blot images (left) and quantification (right) showing that coculture of Raw macrophages with MDA231 cells (MDA231 Cocul) in 3D collagen I gels significantly enhanced the expression of MMP1 (D) and MT1-MMP protein (E) in MDA231 relative to monoculture control (MDA231; Ctrl, control). *, P < 0.05; ***, P < 0.001; ns, not significant.

Figure 2.

Cancer cell migration speed and directedness are MMP dependent, and macrophages enhance cancer cell MMP expression. A, Representative confocal image showing MDA231 cells (green) migrating through dense collagen I ECM (magenta) by degrading the matrix, leaving behind a microtrack (arrow). B and C, Compared with the untreated and DMSO controls, inhibition of MMP activity by GM6001 significantly reduced MDA231 migration total speed (B) and directedness (C). D and E, Representative Western blot images (left) and quantification (right) showing that coculture of Raw macrophages with MDA231 cells (MDA231 Cocul) in 3D collagen I gels significantly enhanced the expression of MMP1 (D) and MT1-MMP protein (E) in MDA231 relative to monoculture control (MDA231; Ctrl, control). *, P < 0.05; ***, P < 0.001; ns, not significant.

Close modal

On the basis of these findings, we examined the role of macrophages in regulating MMP1 and MT1-MMP expression by cancer cells. We chose to study these two MMPs as these proteases are responsible for the breakdown of collagen I matrix. Moreover, MT1-MMP and MMP1 have been shown to be present in the tumor microenvironment, and they have been implicated in tumor metastasis (23–27). To study how macrophages influence MMP expressions in cancer cells, we cocultured cancer cells with macrophages in a transwell system and assessed the cancer cell expression of MMP1 and MT1-MMP via Western blotting. We found that coculture of MDA231 cancer cells with Raw macrophages, as well as BMDMs, significantly enhanced cancer cell expression of MMP1 and MT1-MMP (Fig. 2D and E; Supplementary Fig. S3C). This result was reproduced in PC3 prostate cancer cells cocultured with macrophages (Supplementary Fig. S3D).

Next, we tested whether it is necessary for macrophages to be in direct physical contact with cancer cells to promote migration. Instead of culturing Raw macrophages together with MDA231 cancer cells in the ECM, we cultured macrophages in the microchannels flanking the ECM (Supplementary Fig. S4A). The macrophages seeded in the microchannels were not in physical contact with the ECM or the cancer cells, but they were able to communicate with the cancer cells via the secretion of paracrine factors. Interestingly, we found that macrophages cultured in the microchannels increased cancer cell migration total speed and directedness to the same degree as macrophages cultured in the collagen ECM (Supplementary Fig. S4B), suggesting that direct contact between macrophages and cancer cells is not necessary to enhance cancer cell migration. We also found that the conditioned media from Raw macrophages and BMDM significantly upregulated the expression of MMP1 and MT1-MMP (Supplementary Fig. S4C–S4G) in cancer cells. These results indicate that the effects of macrophages on cancer cell migration dynamics and MMP expressions are mediated primarily through paracrine factors secreted by macrophages.

Macrophage-released TNFα and TGFβ1 are responsible for the increases in cancer cell migration total speed and directedness

We next performed experiments to identify the paracrine factors released by macrophages that were responsible for the increases in cancer cell migration dynamics. We hypothesized that TNFα and TGFβ1 secreted by macrophages are involved in promoting cancer cell migration, as these two factors are major secretory products of macrophages in the tumor microenvironment (28–31), and they have been implicated in tumor metastasis (32, 33). Indeed, primary macrophages, such as MDMΦ and BMDM, have been shown to secrete TNFα and TGFβ1 (34–37). We first verified, using ELISA, that Raw macrophages used in our study also secreted TNFα and TGFβ1 (Supplementary Fig. S4H). To test our hypothesis further, we treated cancer cell–macrophage coculture with neutralizing antibodies against TNFα and/or TGFβ1 and measured cancer cell migration total speed and directedness as before. The antibodies used in this study were designed to act against mouse TNFα and TGFβ1. This allowed us to specifically inhibit TNFα and TGFβ1 secreted by Raw 264.7 mouse macrophages. Antibody blocking results showed that neutralizing TNFα in coculture slightly decreased macrophage-enhanced cancer cell migration total speed, while blocking TGFβ1 almost completely abrogated the effects of macrophages on total speed (Fig. 3A and B). Coblocking both TNFα and TGFβ1 did not further reduce cancer cell migration total speed when compared with the blocking of only TGFβ1 (Fig. 3C). These results suggest that TGFβ1 is primarily responsible for the ability of macrophages to enhance cancer cell migration total speed.

Figure 3.

Macrophage-released TNFα and TGFβ1 are responsible for the increase in cancer cell migration total speed and directedness. MDA231 cancer cells (CC) cocultured with Raw cells (MΦ) were treated with neutralizing antibodies against TNFα (a-TNFα) and/or TGFβ1 (a-TGFβ1). A–C, Neutralizing TNFα released by macrophages (CCMΦ a-TNFα) led to a decrease in MDA231 migration total speed compared with no-treatment control (CCMΦ; A). However, inhibiting macrophage-released TGFβ1 (CCMΦ a-TGFβ1) led to an almost complete inhibition of macrophage's effect on MDA231 migration total speed (B), similar to the simultaneous inhibition of both TNFα and TGFβ1 (C). D–F, Neutralizing macrophage-released TNFα (CCMΦ a-TNFα) or TGFβ1 (CCMΦ a-TGFβ1) alone did not significantly reduce MDA231 migration directedness (D and E). However, simultaneous inhibition of both TNFα and TGFβ1 led to an almost complete abolishment of macrophage-enhanced MDA231 migration directedness (F). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 3.

Macrophage-released TNFα and TGFβ1 are responsible for the increase in cancer cell migration total speed and directedness. MDA231 cancer cells (CC) cocultured with Raw cells (MΦ) were treated with neutralizing antibodies against TNFα (a-TNFα) and/or TGFβ1 (a-TGFβ1). A–C, Neutralizing TNFα released by macrophages (CCMΦ a-TNFα) led to a decrease in MDA231 migration total speed compared with no-treatment control (CCMΦ; A). However, inhibiting macrophage-released TGFβ1 (CCMΦ a-TGFβ1) led to an almost complete inhibition of macrophage's effect on MDA231 migration total speed (B), similar to the simultaneous inhibition of both TNFα and TGFβ1 (C). D–F, Neutralizing macrophage-released TNFα (CCMΦ a-TNFα) or TGFβ1 (CCMΦ a-TGFβ1) alone did not significantly reduce MDA231 migration directedness (D and E). However, simultaneous inhibition of both TNFα and TGFβ1 led to an almost complete abolishment of macrophage-enhanced MDA231 migration directedness (F). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Close modal

Surprisingly, when we assessed the effects of antibody blocking on cancer cell migration directedness, we found that inhibiting either TNFα or TGFβ1 in coculture did not lead to significant decreases in cancer cell migration directedness (Fig. 3D and E). In contrast, blocking TNFα and TGFβ1 simultaneously almost completely abolished the ability of macrophages to promote migration directedness (Fig. 3F), suggesting that both TNFα and TGFβ1 are important to macrophage-enhanced migration directedness. Interestingly, this result is in contrast to the antibody blocking results for total speed, which seems to suggest that macrophage-enhanced cancer cell migration total speed and directedness are controlled by two different pathways. Finally, to verify whether blocking antibody treatments were specific to macrophage-secreted TNFα and TGFβ1, we treated cancer cell monocultures with anti-mouse neutralizing antibodies that we used in coculture experiments. We found this to have no significant effect on MDA231 cell migration total speed and directedness (Supplementary Fig. S5), indicating that the antibody inhibition was macrophage specific.

We next demonstrated that coblocking of both TNFα and TGFβ1 in coculture resulted in almost complete inhibition of macrophage-induced MMP1 and MT1-MMP protein expression in cancer cells (Supplementary Fig. S6A and S6B). These results further support our previous conclusion that macrophage-enhanced cancer cell migration total speed and directedness in 3D ECM are controlled by MMPs. Finally, as expected, because both migration total speed and directedness contribute to cancer cell invasion rate (Supplementary Fig. S2), blocking of TNFα or TGFβ1 cannot completely abrogate macrophage-enhanced cancer cell invasion rate. In contrast, when both macrophage TNFα and TGFβ1 were inhibited, cancer cell invasion rate in coculture was reduced to the level of the cancer cell monoculture control (Supplementary Fig. S6C).

Macrophage-released TGFβ1 enhances cancer cell migration total speed via MT1-MMP, whereas macrophage-released TNFα and TGFβ1 synergistically increase cancer cell migration directedness via MMP1

We then proceeded to examine the detailed mechanisms by which macrophage-released TNFα and TGFβ1 affect cancer cell migration dynamics. We also sought to elucidate the seemingly distinct pathways that are involved in promoting migration total speed and directedness. As it is difficult to perform a detailed and well-controlled study on molecular mechanism with blocking antibodies alone, we elected to treat cancer cell monocultures with TNFα and/or TGFβ1 and assess the resulting cell migration dynamics and MMP expressions. We found that the treatment of MDA231 cancer cell with TNFα slightly increased cancer cell migration total speed, whereas TGFβ1 treatment significantly enhanced total speed. No additional increase in migration total speed was observed for TNFα and TGFβ1 cotreatment over the TGFβ1 monotreatment condition (Fig. 4A). These results parallel the blocking antibody experiments and further support our prior conclusion that macrophage-released TGFβ1 is the main contributor to the increase in cancer cell migration total speed.

Figure 4.

TGFβ1 increases cancer cell migration total speed via the induction of MT1-MMP expression, whereas TNFα and TGFβ1 synergistically increase cancer cell migration directedness via the induction of MMP1 expression. MDA231 monoculture (CC) was treated with TNFα and/or TGFβ1, and the resulting cell migration dynamics and MMP expressions were analyzed. A–C, Treatment of MDA231 with TGFβ1 (CC TGFβ1) led to larger increases in MDA231 migration total speed (A), MT1-MMP mRNA (B), and protein (C) expressions than TNFα monotreatment (CC TNFα). However, cotreatment of both TNFα and TGFβ1 led to no further increase in migration total speed, MT1-MMP mRNA, and protein expressions compared with TGFβ1 monotreatment. Data in A–C follow a similar trend. D–F, TNFα and TGFβ1 synergistically increased MDA231 migration directedness (D), MMP1 mRNA expression (E), and MMP1 protein production (F) when compared with monotreatment conditions. Data in D–F follow a similar trend. Data in C and F were obtained from cells cultured in 3D collagen I ECM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Figure 4.

TGFβ1 increases cancer cell migration total speed via the induction of MT1-MMP expression, whereas TNFα and TGFβ1 synergistically increase cancer cell migration directedness via the induction of MMP1 expression. MDA231 monoculture (CC) was treated with TNFα and/or TGFβ1, and the resulting cell migration dynamics and MMP expressions were analyzed. A–C, Treatment of MDA231 with TGFβ1 (CC TGFβ1) led to larger increases in MDA231 migration total speed (A), MT1-MMP mRNA (B), and protein (C) expressions than TNFα monotreatment (CC TNFα). However, cotreatment of both TNFα and TGFβ1 led to no further increase in migration total speed, MT1-MMP mRNA, and protein expressions compared with TGFβ1 monotreatment. Data in A–C follow a similar trend. D–F, TNFα and TGFβ1 synergistically increased MDA231 migration directedness (D), MMP1 mRNA expression (E), and MMP1 protein production (F) when compared with monotreatment conditions. Data in D–F follow a similar trend. Data in C and F were obtained from cells cultured in 3D collagen I ECM. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

Close modal

In contrast to its effects on total speed, TNFα or TGFβ1 monotreatment did not significantly enhance cancer cell migration directedness. When the cancer cells were treated with both TNFα and TGFβ1, however, there was a synergistic increase in cancer cell migration directedness that cannot be explained by the additive effects of TNFα and TGFβ1 monotreatment (Fig. 4D). Combined, these results provide further evidence that macrophage-induced cancer cell 3D migration total speed and directedness are controlled by two distinct mechanisms. Specifically, cancer cell migration total speed is controlled primarily by macrophage-released TGFβ1, whereas the directedness is controlled by the combined effects of macrophage-released TNFα and TGFβ1. Finally, we found that cotreatment of cancer cell monoculture with TNFα and TGFβ1 led to levels of migration total speed and directedness (Supplementary Fig. S7) comparable with those in coculture, indicating that TNFα and TGFβ1 from macrophages are, indeed, the main factors responsible for the enhancement in cancer cell migration.

For further verification that cancer cell migration total speed and directedness are controlled through two distinct pathways, we varied the concentration of TNFα and TGFβ1 in the cotreatment condition. Specifically, we treated MDA231 cancer cells with 5 ng/mL TNFα + 0.5 ng/mL TGFβ1, or 0.5 ng/mL TNFα + 5 ng/mL TGFβ1, or 5 ng/mL TNFα + 5 ng/mL TGFβ1. Interestingly, treating cancer cells with a low concentration of TGFβ1 (0.5 ng/mL), even in the cotreatment conditions, resulted in slight or no increase in the migration total speed (Supplementary Fig. S8A). These results further illustrate that cancer cell migration total speed is mainly controlled by TGFβ1. In comparison, treating cancer cells with various concentrations of TNFα or TGFβ1 in the cotreatment regimen resulted in similar levels of increase in cancer cell migration directedness over the no-treatment control. Moreover, addition of even a minute amount (0.5 ng/mL) of TGFβ1 to TNFα monotreatment resulted in sharp increases in cancer cell migration directedness. A similar response was observed when a minute amount of TNFα (0.5 ng/mL) was added to TGFβ1 monotreatment (Supplementary Fig. S8B). These results further verify that TNFα and TGFβ1 synergistically enhance cancer cell migration directedness.

As cancer cell migration in 3D ECM depends on the cell's ability to express MMPs, it seemed that the effects of TNFα and TGFβ1 on cancer cell migration dynamics might also be mediated through MMPs. To test for this hypothesis, we treated MDA231 monoculture with TNFα and/or TGFβ1 and evaluated the resulting MMP1 and MT1-MMP mRNA and protein expression. We found that the treatment of cancer cells with TNFα resulted in a slight increase in MT1-MMP mRNA and protein expression. In comparison, the treatment of cells with TGFβ1 alone markedly enhanced MT1-MMP mRNA and protein expression, whereas cotreatment of both TNFα and TGFβ1 led to no further increase in MT1-MMP expressions (Fig. 4B and C). We noted that these trends in the increases in MT1-MMP mRNA and protein expressions match the trend in the increases in cancer cell migration total speed (Fig. 4A). This observation points to the possibility that TGFβ1-induced increase in cell migration total speed is mediated mainly via MT1-MMP. Furthermore, we observed that TNFα and TGFβ1 synergistically enhanced cancer cell expression of MMP1 mRNA and protein (Fig. 4E and F). These findings are similar to the observation that TNFα and TGFβ1 synergistically promote cancer cell migration directedness (Fig. 4D), suggesting that TNFα/TGFβ1–induced cancer cell migration directedness is mediated mainly by MMP1 expression. Indeed, Pearson correlation analysis revealed that MT1-MMP expression levels resulting from TNFα and/or TGFβ1 treatments strongly correlate with cancer cell migration total speed, but not directedness. Similarly, MMP1 expression levels in cancer cells strongly correlate with migration directedness, but not total speed (Supplementary Fig. S9). These results led us to hypothesize that macrophage-induced cancer cell migration total speed is controlled by MT1-MMP expression in cancer cells, whereas the directedness is controlled by MMP1 expression.

To test whether or not macrophage-induced cancer cell migration total speed and directedness are controlled by two different MMPs, we treated cancer cell–macrophage coculture with blocking antibodies against MT1-MMP and MMP1. We found that treating the coculture with anti–MT1-MMP antibody resulted in a significant decrease in cancer cell migration total speed with little effect on directedness (Fig. 5A and B). In contrast, we observed that blocking MMP1 in coculture with anti-MMP1 antibody had almost no effect on macrophage-induced increase in cancer cell migration total speed, while the increase in cancer cell migration directedness was significantly reduced (Fig. 5C and D). Furthermore, we treated cancer cell monoculture with exogenously supplied recombinant MMP1 and observed an enhancement in cancer cell migration directedness but no significant change in migration total speed (Fig. 5E and F). These findings, coupled with previous observations that macrophage-released TNFα and TGFβ1 upregulated cancer cell expression of MMPs (Supplementary Fig S6), strongly support the conclusion that macrophage-induced MMP1 expression is responsible for the increase in cancer cell migration directedness, whereas the induction of MT1-MMP expression is responsible for the increase in total speed. Taken together, these results (Figs. 3–5) demonstrate that macrophages influence cancer cell migration in 3D ECM via two different mechanisms: (i) macrophage-released TGFβ1 increase cancer cell migration total speed (speed) via the upregulation of MT1-MMP expression; and (ii) macrophage-released TNFα and TGFβ1 synergistically enhance cancer cell migration directedness (persistence) through the induction of MMP1 expression. Hence, these results strongly suggest that both of these two pathways need to be inhibited to effectively reduce metastasis. Indeed, using a 4T1 orthotopic breast tumor model in BALB/c mice, we found that inhibiting both TNFα and TGFβ1 in these mice resulted in a more significant reduction in lung metastasis formation compared with inhibiting TNFα or TGFβ1 alone (Supplementary Fig. S10).

Figure 5.

Macrophage-induced cancer cell migration total speed is mediated via MT1-MMP, whereas directedness is mediated by MMP1. MDA231 cancer cells (CC)–Raw macrophages (MΦ) coculture was treated with blocking antibodies against MT1-MMP or MMP1. A and B, Treatment of coculture with anti-MT1-MMP antibody (CCMΦ a-MT1-MMP) decreased MDA231 migration total speed (A) but not directedness (B). C and D, Treatment of coculture with anti-MMP1 antibody (CCMΦ a-MMP1) reduced MDA231 migration directedness (D) while having a minimal effect on total speed (C). E and F, Treatment of MDA231 cancer cell monoculture with recombinant MMP1 (CC MMP1) enhanced MDA231 migration directedness (F), while having a minimal effect on total speed (E). *, P < 0.05; ***, P < 0.001; ns, not significant.

Figure 5.

Macrophage-induced cancer cell migration total speed is mediated via MT1-MMP, whereas directedness is mediated by MMP1. MDA231 cancer cells (CC)–Raw macrophages (MΦ) coculture was treated with blocking antibodies against MT1-MMP or MMP1. A and B, Treatment of coculture with anti-MT1-MMP antibody (CCMΦ a-MT1-MMP) decreased MDA231 migration total speed (A) but not directedness (B). C and D, Treatment of coculture with anti-MMP1 antibody (CCMΦ a-MMP1) reduced MDA231 migration directedness (D) while having a minimal effect on total speed (C). E and F, Treatment of MDA231 cancer cell monoculture with recombinant MMP1 (CC MMP1) enhanced MDA231 migration directedness (F), while having a minimal effect on total speed (E). *, P < 0.05; ***, P < 0.001; ns, not significant.

Close modal

Finally, we found a similar synergistic response in MMP1 secretion due to TNFα and TGFβ1 cotreatment (Supplementary Fig. S11A), which mirrors the results of cancer cell migration directedness (Fig. 4D). The synergistic induction in MMP1 protein production was also observed in MDA435 and PC3 cells (Supplementary Fig. S11B and S11C).

TNFα and TGFβ1 synergistically increase nuclear localization of NF-κB

To further understand the synergistic effects of TNFα and TGFβ1 on the expression of MMP1 in cancer cells, we tested whether TNFα and/or TGFβ1 could alter the expression or nuclear localization of NF-κB, a transcription factor for MMP1 (38). We first treated MDA231 cancer cells with TNFα and/or TGFβ1 for 48 hours and found that these two factors did not change the protein production of NF-κB by cancer cells (Fig. 6A). We then tested whether the treatment of these two factors could alter the nuclear localization of NF-κB. Indeed, cotreatment of TNFα and TGFβ1 synergistically enhanced NF-κB expression inside the nucleus of the cancer cells (Fig. 6B and C). These results support the conclusion that TNFα and TGFβ1 act together to enhance the expression of MMP1 via the synergistic induction of NF-κB nuclear translocation. Similar results were also observed in MDA435 cells (Supplementary Fig. S12).

Figure 6.

TNFα and TGFβ1 synergistically increase NF-κB nuclear localization. A, Western blot quantification (top) and representative images (bottom) showing 48-hour TNFα and/or TGFβ1 treatments of MDA231 cells did not alter the production of NF-κB (NF-κB total). B and C, Western blot quantifications (top) and representative images (bottom) showing 2-hour TNFα and TGFβ1 cotreatment of MDA231 synergistically increased the level of NF-κB in the nuclear fraction of MDA231 (NF-κB Nuc; C), but not in the cytoplasmic fraction of the cells (NF-κB Cyto; B). *, P < 0.05; ns, not significant.

Figure 6.

TNFα and TGFβ1 synergistically increase NF-κB nuclear localization. A, Western blot quantification (top) and representative images (bottom) showing 48-hour TNFα and/or TGFβ1 treatments of MDA231 cells did not alter the production of NF-κB (NF-κB total). B and C, Western blot quantifications (top) and representative images (bottom) showing 2-hour TNFα and TGFβ1 cotreatment of MDA231 synergistically increased the level of NF-κB in the nuclear fraction of MDA231 (NF-κB Nuc; C), but not in the cytoplasmic fraction of the cells (NF-κB Cyto; B). *, P < 0.05; ns, not significant.

Close modal

Macrophages in the tumor microenvironment are key promoters of cancer cell metastasis (8), suggesting that the control of these cells and their released factors can be a viable strategy in treating metastasis. Yet, it is still unclear how macrophages affect different aspects of cancer cell migration, such as speed and persistence, especially in 3D ECM that closely mimics the in vivo tumor microenvironment. To address this gap in knowledge, we utilized a microfluidic 3D cell migration assay that allows us to study, in high resolution, the effects of macrophages on cancer cell migration speed (total speed) and persistence (directedness) in 3D ECM.

From our study, we found that macrophages increase cancer cell migration speed and persistence in 3D collagen I ECM, suggesting that macrophages may help cancer cells invade and gain access to intravasation sites more efficiently. In contrast to the 3D results, we discovered that on 2D tissue culture plastic, macrophages tend to increase cancer cell migration speed but decrease persistence, so that cancer cells move faster, but more randomly. This disparity, similar to results obtained from previous works, illustrates a fundamental difference in how cancer cells migrate in 2D compared with 3D (19). We also note that the cancer cells in our 3D microfluidic system migrated at a total speed of 5 to 11 μm/hour, a value that closely matches the speed values obtained from in vivo intravital imaging experiments (39). In sum, these results demonstrate the advantages of our microfluidic assay, which allows us to perform physiologically relevant studies with precise control of experimental conditions.

In 3D, cell migration critically depends on the ability of cancer cells to degrade ECM. Indeed, MMP expression is dispensable for 2D cell migration, but not for 3D (19). In the current study, we showed that macrophages enhanced cancer cell migration in 3D ECM via the upregulation of MMP expression in cancer cells. We further identified that macrophage-released TGFβ1 increased cancer cell migration speed, whereas macrophage-released TNFα and TGFβ1 synergistically enhanced cancer cell migration persistence. Previous studies have shown that EGF released by macrophages can promote cancer cell migration (14). Here, we report that macrophage-released TNFα and TGFβ1 can also promote cancer cell migration. The clinical relevance of this finding is demonstrated by the fact that the expression levels of TNFα and TGFβ1 in tumor-associated macrophages correlate with metastasis for human tumors (28, 40). Moreover, this study, to our knowledge, is the first to report that macrophage-released TNFα and TGFβ1 control different aspects of cancer cell migration (speed vs. persistence) differently. Thus, although prior studies have implicated TNFα and TGFβ1 in cancer cell invasion and metastasis (32, 33), our results now demonstrate subtle but important differences in their effects on cancer cell migration.

We also found that TGFβ1 released by macrophages promotes cancer cell migration speed through upregulation of MT1-MMP. We suspect this is due to the fact that MT1-MMP can influence cell-intrinsic migration behaviors as well as cell-extrinsic matrix properties, both of which are known determinants of cell migration speed in 3D matrix (19, 41). Two examples of cell-intrinsic behaviors that control cell migration speed are the activities of kinases and the expression of integrin. It has been shown that intermediate levels of integrin α2β1 contribute to an optimum cell migration speed (42, 43), and inhibiting integrin could lead to a decrease in cell migration speed, but not persistence (17). Similarly, inhibiting Akt/PI3K activities in cells has been reported to result in a decrease in migration speed (44). In addition to cell-intrinsic properties, cell-extrinsic matrix properties, such as the pore size of the matrix that can be modified by MMPs, also affect cell migration speed in 3D (41). Unlike MMP1, which primarily degrades collagen I matrix to alter the cell-extrinsic properties, MT1-MMP modifies both cell-intrinsic and cell-extrinsic properties. Besides degrading collagen I matrix, MT1-MMP can process integrin (45), mediate Akt phosphorylation (46), and promote syndecan shedding (47), all of which are parts of cell-intrinsic pathways of migration. Hence, as both intrinsic and extrinsic properties control cell migration speed, it stands to reason that MT1-MMP should be the major determinant of migration speed over MMP1.

We further demonstrated that macrophage-enhanced cancer cell migration persistence, in contrast to total speed, was mediated primarily by the expression of MMP1, but not MT1-MMP. This result may be explained by the fact that MMP1 is more efficient in degrading collagen I matrix and altering extrinsic matrix properties (pore size) than MT1-MMP (48). Although cell-intrinsic properties (such as Rac activities; ref. 49) control migration persistence in 2D, it has been reported that cell-extrinsic matrix properties seem to dominate over intrinsic property as the primary determinant of 3D migration persistence (19, 50). As MMP1 is more efficient in degrading collagen I ECM and altering extrinsic matrix properties than MT1-MMP, MMP1 should therefore be a major contributor to migration persistence.

On the basis of our findings, we propose a novel mechanism whereby macrophages promote cancer cell migration speed (total speed) and persistence (directedness) via two distinct mechanisms (Fig. 7). First, macrophage-released TNFα and TGFβ1 synergistically induce nuclear translocation of NF-κB in cancer cells, leading to synergistic increases in the expressions of MMP1, which result in a synergistic enhancement in cancer cell migration persistence. In contrast to the mechanism for the persistence, macrophages increase cancer cell migration speed, mainly through TGFβ1, by the upregulation of cancer cell MT1-MMP expression. These findings establish that TNFα and TGFβ1 released by macrophages influence speed and persistence of cancer cell migration differently, and both of these factors need to be targeted to effectively inhibit macrophage-assisted cancer cell 3D migration and metastasis. Moreover, these findings also broaden our current view on the molecular determinants of 3D migration, suggesting that MT1-MMP primarily controls cell migration speed, whereas MMP1 mainly controls migration persistence. In conclusion, our findings provide new insights into macrophage-assisted cancer cell migration in 3D tumor microenvironment, and these could ultimately lead to novel therapeutic strategies to effectively inhibit tumor invasion and metastasis.

Figure 7.

Proposed mechanism explaining the effects of macrophages (MΦ) on cancer cell (CC) migration speed and persistence. Macrophage-released TNFα and TGFβ1 synergistically enhance NF-κB nuclear localization in cancer cells, leading to synergistic increases in cancer cell MMP1 mRNA expression, protein production, and protein secretion. This increase in MMP1 secretion by cancer cells leads to an increase in cancer cell migration persistence (directedness). Meanwhile, macrophages increase cancer cell migration speed (total speed), mainly through TGFβ1-induced cancer cell expression of MT1-MMP.

Figure 7.

Proposed mechanism explaining the effects of macrophages (MΦ) on cancer cell (CC) migration speed and persistence. Macrophage-released TNFα and TGFβ1 synergistically enhance NF-κB nuclear localization in cancer cells, leading to synergistic increases in cancer cell MMP1 mRNA expression, protein production, and protein secretion. This increase in MMP1 secretion by cancer cells leads to an increase in cancer cell migration persistence (directedness). Meanwhile, macrophages increase cancer cell migration speed (total speed), mainly through TGFβ1-induced cancer cell expression of MT1-MMP.

Close modal

R.D. Kamm is a co-founder (chair of scientific advisory board) of, has ownership interest (including patents) in, and is a consultant/advisory board member for AIM Biotech. No potential conflicts of interest were disclosed by the other authors.

Conception and design: R. Li, J.D. Hebert, R.O. Hynes, R.D. Kamm

Development of methodology: R. Li, H. Xing

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Li, J.D. Hebert, T.A. Lee, H. Xing, A. Boussommier-Calleja

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Li, H. Xing

Writing, review, and/or revision of the manuscript: R. Li, R.O. Hynes, D.A. Lauffenburger, R.D. Kamm

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Li

Study supervision: R.D. Kamm

The authors thank Drs. Ioannis Zervantonakis and Michael Hemann for their advice. We also thank the NIH, Koch Institute, and NSF for their support.

This study was supported by the NIH grant U01 CA202177-01 (R. Li and R.D. Kamm), NIH Pre-Doctoral Training grant T32GM007287 (J.D. Hebert), Koch Institute Graduate Student Fellowship (R. Li), and NSF Graduate Fellowship (R. Li).

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.
Balkwill
FR
,
Capasso
M
,
Hagemann
T
. 
The tumor microenvironment at a glance
.
J Cell Sci
2012
;
125
:
5591
6
.
2.
Joyce
JA
,
Pollard
JW
. 
Microenvironmental regulation of metastasis
.
Nat Rev Cancer
2009
;
9
:
239
52
.
3.
Lu
P
,
Weaver
VM
,
Werb
Z
. 
The extracellular matrix: a dynamic niche in cancer progression
.
J Cell Biol
2012
;
196
:
395
406
.
4.
Friedl
P
,
Wolf
K
. 
Tumour-cell invasion and migration: diversity and escape mechanisms
.
Nat Rev Cancer
2003
;
3
:
362
74
.
5.
Egeblad
M
,
Werb
Z
. 
New functions for the matrix metalloproteinases in cancer progression
.
Nat Rev Cancer
2002
;
2
:
161
74
.
6.
Kerkela
E
,
Saarialho-Kere
U
. 
Matrix metalloproteinases in tumor progression: focus on basal and squamous cell skin cancer
.
Exp Dermatol
2003
;
12
:
109
25
.
7.
Yamaguchi
H
,
Condeelis
J
. 
Regulation of the actin cytoskeleton in cancer cell migration and invasion
.
Biochim Biophys Acta
2007
;
1773
:
642
52
.
8.
Pollard
JW
. 
Tumour-educated macrophages promote tumour progression and metastasis
.
Nat Rev Cancer
2004
;
4
:
71
8
.
9.
Lewis
CE
,
Pollard
JW
. 
Distinct role of macrophages in different tumor microenvironments
.
Cancer Res
2006
;
66
:
605
12
.
10.
Chen
P
,
Huang
Y
,
Bong
R
,
Ding
Y
,
Song
N
,
Wang
X
, et al
Tumor-associated macrophages promote angiogenesis and melanoma growth via adrenomedullin in a paracrine and autocrine manner
.
Clin Cancer Res
2011
;
17
:
7230
9
.
11.
Wyckoff
JB
,
Wang
Y
,
Lin
EY
,
Li
J
,
Goswami
S
,
Stanley
ER
, et al
Direct visualization of macrophage-assisted tumor cell intravasation in mammary tumors
.
Cancer Res
2007
;
67
:
2649
56
.
12.
Zervantonakis
IK
,
Hughes-Alford
SK
,
Charest
JL
,
Condeelis
JS
,
Gertler
FB
,
Kamm
RD
. 
Three-dimensional microfluidic model for tumor cell intravasation and endothelial barrier function
.
Proc Natl Acad Sci U S A
2012
;
109
:
13515
20
.
13.
Patsialou
A
,
Wyckoff
J
,
Wang
Y
,
Goswami
S
,
Stanley
ER
,
Condeelis
JS
. 
Invasion of human breast cancer cells in vivo requires both paracrine and autocrine loops involving the colony-stimulating factor-1 receptor
.
Cancer Res
2009
;
69
:
9498
506
.
14.
Goswami
S
,
Sahai
E
,
Wyckoff
JB
,
Cammer
M
,
Cox
D
,
Pixley
FJ
, et al
Macrophages promote the invasion of breast carcinoma cells via a colony-stimulating factor-1/epidermal growth factor paracrine loop
.
Cancer Res
2005
;
65
:
5278
83
.
15.
Polacheck
WJ
,
Zervantonakis
IK
,
Kamm
RD
. 
Tumor cell migration in complex microenvironments
.
Cell Mol Life Sci
2013
;
70
:
1335
56
.
16.
Haessler
U
,
Teo
JCM
,
Foretay
D
,
Renaud
P
,
Swartz
MA
. 
Migration dynamics of breast cancer cells in a tunable 3D interstitial flow chamber
.
Integr Biol
2012
;
4
:
401
9
.
17.
Maheshwari
G
,
Lauffenburger
D a
. 
Deconstructing (and reconstructing) cell migration
.
Microsc Res Tech
1998
;
43
:
358
68
.
18.
Polacheck
WJ
,
Charest
JL
,
Kamm
RD
. 
Interstitial flow influences direction of tumor cell migration through competing mechanisms
.
Proc Natl Acad Sci U S A
2011
;
108
:
11115
20
.
19.
Kim
H-D
,
Guo
TW
,
Wu
AP
,
Wells
A
,
Gertler
FB
,
Lauffenburger
DA
. 
Epidermal growth factor-induced enhancement of glioblastoma cell migration in 3D arises from an intrinsic increase in speed but an extrinsic matrix- and proteolysis-dependent increase in persistence
.
Mol Biol Cell
2008
;
19
:
4249
59
.
20.
Farahat
WA
,
Wood
LB
,
Zervantonakis
IK
,
Schor
A
,
Ong
S
,
Neal
D
, et al
Ensemble analysis of angiogenic growth in three-dimensional microfluidic cell cultures
.
PLoS One
2012
;
7
:
e37333
.
21.
Kauppila
S
,
Stenbäck
F
,
Risteli
J
,
Jukkola
A
,
Risteli
L
. 
Aberrant type I and type III collagen gene expression in human breast cancer in vivo
.
J Pathol
1998
;
186
:
262
8
.
22.
Provenzano
PP
,
Inman
DR
,
Eliceiri
KW
,
Knittel
JG
,
Yan
L
,
Rueden
CT
, et al
Collagen density promotes mammary tumor initiation and progression
.
BMC Med
2008
;
6
:
11
.
23.
Liu
H
,
Kato
Y
,
Erzinger
SA
,
Kiriakova
GM
,
Qian
Y
,
Palmieri
D
, et al
The role of MMP-1 in breast cancer growth and metastasis to the brain in a xenograft model
.
BMC Cancer
2012
;
12
:
583
.
24.
Poola
I
,
DeWitty
RL
,
Marshalleck
JJ
,
Bhatnagar
R
,
Abraham
J
,
Leffall
LD
. 
Identification of MMP-1 as a putative breast cancer predictive marker by global gene expression analysis
.
Nat Med
2005
;
11
:
481
3
.
25.
Lu
X
,
Wang
Q
,
Hu
G
,
Van Poznak
C
,
Fleisher
M
,
Reiss
M
, et al
ADAMTS1 and MMP1 proteolytically engage EGF-like ligands in an osteolytic signaling cascade for bone metastasis
.
Genes Dev
2009
;
23
:
1882
94
.
26.
Jacob
A
,
Prekeris
R
. 
The regulation of MMP targeting to invadopodia during cancer metastasis
.
Front Cell Dev Biol
2015
;
3
:
4
.
27.
Ha
HY
,
Moon
HB
,
Nam
MS
,
Lee
JW
,
Ryoo
ZY
,
Lee
TH
, et al
Overexpression of membrane-type matrix metalloproteinase-1 gene induces mammary gland abnormalities and adenocarcinoma in transgenic mice
.
Cancer Res
2001
;
61
:
984
90
.
28.
Miles
DW
,
Happerfield
LC
,
Naylor
MS
,
Bobrow
LG
,
Rubens
RD
,
Balkwill
FR
. 
Expression of tumour necrosis factor (TNF alpha) and its receptors in benign and malignant breast tissue
.
Int J Cancer
1994
;
56
:
777
82
.
29.
Chong
H
,
Vodovotz
Y
,
Cox
GW
,
Barcellos-Hoff
MH
. 
Immunocytochemical localization of latent transforming growth factor-beta1 activation by stimulated macrophages
.
J Cell Physiol
1999
;
178
:
275
83
.
30.
Mantovani
A
,
Sozzani
S
,
Locati
M
,
Allavena
P
,
Sica
A
. 
Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes
.
Trends Immunol
2002
;
23
:
549
55
.
31.
Loercher
AE
,
Nash
MA
,
Kavanagh
JJ
,
Platsoucas
CD
,
Freedman
RS
. 
Identification of an IL-10-producing HLA-DR-negative monocyte subset in the malignant ascites of patients with ovarian carcinoma that inhibits cytokine protein expression and proliferation of autologous T cells
.
J Immunol
1999
;
163
:
6251
60
.
32.
Massagué
J
. 
TGFbeta in cancer
.
Cell
2008
;
134
:
215
30
.
33.
Balkwill
F
. 
TNF-alpha in promotion and progression of cancer
.
Cancer Metastasis Rev
2006
;
25
:
409
16
.
34.
Cudejko
C
,
Wouters
K
,
Fuentes
L
,
Hannou
SA
,
Paquet
C
,
Bantubungi
K
, et al
p16INK4a deficiency promotes IL-4-induced polarization and inhibits proinflammatory signaling in macrophages
.
Blood
2011
;
118
:
2556
66
.
35.
Li
D
,
Guabiraba
R
,
Besnard
A-G
,
Komai-Koma
M
,
Jabir
MS
,
Zhang
L
, et al
IL-33 promotes ST2-dependent lung fibrosis by the induction of alternatively activated macrophages and innate lymphoid cells in mice
.
J Allergy Clin Immunol
2014
;
134
:
1422
32.e11
.
36.
Daigneault
M
,
Preston
JA
,
Marriott
HM
,
Whyte
MKB
,
Dockrell
DH
. 
The identification of markers of macrophage differentiation in PMA-Stimulated THP-1 cells and monocyte-derived macrophages
.
PLoS One
2010
;
5
:
e8668
.
37.
Nacu
N
,
Luzina
IG
,
Highsmith
K
,
Lockatell
V
,
Pochetuhen
K
,
Cooper
ZA
, et al
Macrophages produce TGFβ1 (BIGH3) following ingestion of apoptotic cells and regulate MMP14 levels and collagen turnover in fibroblasts
.
J Immunol
2008
;
180
:
5036
44
.
38.
Vincenti
MP
,
Coon
CI
,
Brinckerhoff
CE
. 
Nuclear factor kappaB/p50 activates an element in the distal matrix metalloproteinase 1 promoter in interleukin-1beta-stimulated synovial fibroblasts
.
Arthritis Rheum
1998
;
41
:
1987
94
.
39.
Gligorijevic
B
,
Bergman
A
,
Condeelis
J
. 
Multiparametric classification links tumor microenvironments with tumor cell phenotype
.
PLoS Biol
2014
;
12
:
e1001995
.
40.
Ye
X
,
Xu
S
,
Xin
Y
,
Yu
S
,
Ping
Y
,
Chen
L
, et al
Tumor-associated microglia/macrophages enhance the invasion of glioma stem-like cells via TGF-β1 signaling pathway
.
J Immunol
2012
;
189
:
444
53
.
41.
Wolf
K
,
Te Lindert
M
,
Krause
M
,
Alexander
S
,
Te Riet
J
,
Willis
AL
, et al
Physical limits of cell migration: control by ECM space and nuclear deformation and tuning by proteolysis and traction force
.
J Cell Biol
2013
;
201
:
1069
84
.
42.
Palecek
SP
,
Loftus
JC
,
Ginsberg
MH
,
Lauffenburger
DA
,
Horwitz
AF
. 
Integrin-ligand binding properties govern cell migration speed through cell-substratum adhesiveness
.
Nature
1997
;
385
:
537
40
.
43.
Huttenlocher
A
,
Horwitz
AR
. 
Integrins in cell migration
.
Cold Spring Harb Perspect Biol
2011
;
3
:
a005074
.
44.
Montero
J-A
,
Kilian
B
,
Chan
J
,
Bayliss
PE
,
Heisenberg
C-P
. 
Phosphoinositide 3-Kinase is required for process outgrowth and cell polarization of gastrulating mesendodermal cells
.
Curr Biol
2003
;
13
:
1279
89
.
45.
Deryugina
EI
,
Ratnikov
BI
,
Postnova
TI
,
Rozanov
D V
,
Strongin
AY
. 
Processing of integrin alpha(v) subunit by membrane type 1 matrix metalloproteinase stimulates migration of breast carcinoma cells on vitronectin and enhances tyrosine phosphorylation of focal adhesion kinase
.
J Biol Chem
2002
;
277
:
9749
56
.
46.
Eisenach
PA
,
Roghi
C
,
Fogarasi
M
,
Murphy
G
,
English
WR
. 
MT1-MMP regulates VEGF-A expression through a complex with VEGFR-2 and Src
.
J Cell Sci
2010
;
123
:
4182
93
.
47.
Endo
K
,
Takino
T
,
Miyamori
H
,
Kinsen
H
,
Yoshizaki
T
,
Furukawa
M
, et al
Cleavage of syndecan-1 by membrane type matrix metalloproteinase-1 stimulates cell migration
.
J Biol Chem
2003
;
278
:
40764
70
.
48.
Imai
K
. 
Membrane type 1matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules
.
J Biol Chem
1997
;
272
:
2446
51
.
49.
Pankov
R
,
Endo
Y
,
Even-Ram
S
,
Araki
M
,
Clark
K
,
Cukierman
E
, et al
A Rac switch regulates random versus directionally persistent cell migration
.
J Cell Biol
2005
;
170
:
793
802
.
50.
Wu
P-H
,
Giri
A
,
Sun
SX
,
Wirtz
D
. 
Three-dimensional cell migration does not follow a random walk
.
Proc Natl Acad Sci U S A
2014
;
111
:
3949
54
.

Supplementary data