Macrophages are critical drivers of tumor growth, invasion, and metastasis. Movement of macrophages into tumors requires the activity of cell surface proteases such as plasmin. In this study, we offer genetic evidence that plasminogen receptor S100A10 is essential for recruitment of macrophages to the tumor site. Growth of murine Lewis lung carcinomas or T241 fibrosarcomas was dramatically reduced in S100A10-deficient mice compared with wild-type mice. The tumor growth deficit corresponded with a decrease in macrophage density that could be rescued by intraperitoneal injection of wild-type but not S100A10-deficient macrophages. Notably, macrophages of either genotype could rescue tumor growth if they were injected into the tumor itself, establishing that S100A10 was required specifically for the migratory capability needed for tumor homing. Conversely, selective depletion of macrophages from wild-type mice phenocopied the tumor growth deficit seen in S100A10-deficient mice. Together, our findings show that S100A10 is essential and sufficient for macrophage migration to tumor sites, and they define a novel rate-limiting step in tumor progression. Cancer Res; 71(21); 6676–83. ©2011 AACR.
There is an increasingly large body of evidence correlating tumor-associated macrophage (TAM) density with poor prognosis in a varied number of solid tumors (1, 2). TAM density is directly dependent upon the recruitment of monocytic precursor cells from circulation in response to varied chemotactic signals from the tumor (3). The tumor cytokine milieu, including macrophage colony-stimulating factor, interleukin (IL) 4, IL-13, and IL-10, causes the recruited monocytes/macrophages to differentiate into the alternatively activated M2 macrophage phenotype rather than the tumoricidal M1 phenotype (4–6). The M2 TAMs promote tumor growth by mediating inflammation, stimulating angiogenesis, suppressing antitumor immunity, and by matrix remodeling (6–8). Little is known, however, about the mechanism by which monocytes/macrophages move through tissue barriers to arrive at the tumor site. Plow and colleagues hypothesized that upon activation, circulating monocytes mobilize cell surface plasminogen receptors to generate plasmin, facilitating their movement out of the vessel, through the basement membrane and interstitial space and into the tumor site (9–11). Several activators have been proposed to mediate cellular plasmin generation, and most recently, the plasminogen receptor S100A10 has been shown to mediate the recruitment of macrophages to an inflammatory stimulus (12).
S100A10, also known as p11, is a member of the S100 family of small, EF hand containing dimeric proteins. It is present on the cell surface in a heterotetrameric form in which 2 annexin A2 monomers secure the S100A10 dimer to the cell surface. S100A10 binding to plasminogen mediates its activation by plasminogen activators, either tissue-type plasminogen activator (tPA) or urokinase-type plasminogen activator (uPA), facilitating the conversion of plasminogen to plasmin (13). Active plasmin both degrades fibrin directly and activates members of the matrix metalloproteases family, creating a localized proteolytic hub (14, 15). Recent work by our laboratory has shown that S100A10 is responsible for the conversion of plasminogen to plasmin by many transformed and normal cells including macrophages (12, 16, 17). Here, we show that compared with wild-type mice, tumor cells injected into S100A10-null mice fail to thrive because of a loss of macrophage recruitment to the tumor site. Thus, we show for the first time that macrophages use S100A10 to migrate to the tumor site and that this recruitment is the rate-limiting step in tumor growth.
Materials and Methods
Cell lines and reagents
The LLC cell line was obtained from and authenticated by American Type Culture Collection in 2006. The T241 cell line was obtained from Dr. Y. Cao (Karolinska Institute, Stockholm, Sweden) in 2007. Both LLC and T241 have been found negative for mycoplasma (PCR-based test; VenorGem, within 6 months) as well as negative for a panel of murine pathogens (Charles River comprehensive mouse panel 1) conducted after the cells were received and passaged. The cell lines were not verified after arrival in our laboratory. Cell lines were maintained in complete Dulbecco's Modified Eagle's Media (Gibco) with 10% fetal calf serum (Hyclone) and penicillin/streptomycin (Hyclone).
In vivo tumor growth and macrophage addback
Mouse experiments were carried out under the approval of the Dalhousie University Carlton Animal Care Facility, and mice were maintained in pathogen-free facilities. Tumors were established by subcutaneous injection of 2 × 106 cells in the right flank of female 6- to 8-week-old mice. Volume calculated as (length × width2)/2. Intraperitoneal transfer of 9 × 106 thioglycolate-elicited macrophages was conducted 1 day prior to tumor cell inoculation, whereas intratumoral transfer of 0.25 × 106 thioglycolate-elicited macrophages was conducted 4 days after tumor cell inoculation.
Macrophages were collected by peritoneal lavage with 5 mL RPMI (Gibco) 4 days after intraperitoneal injection of 2.5 mL of 4% Brewers thioglycolate (Sigma).
Immunofluorescence and immunohistochemistry
Harvested tumors were snap frozen and sectioned to 10 μm for immunostaining with F4/80 (Abcam) with an alkaline phosphatase secondary (Biocare Medical); CD31 (Abcam) with an Alexa Fluor 488 secondary as well as anti-Fibrin (ADI) with a horseradish peroxidase (HRP) secondary. For proliferating cell nuclear antigen (PCNA) staining, tumors were fixed in 10% formalin, paraffin embedded, and sectioned to 8 μm before PCNA (Abcam) immunostaining with HRP secondary.
Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling assay
Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) was conducted on snap-frozen, cryosectioned tumors as per the manufacturer's instructions with an alkaline phosphatase converter (In situ cell death detection; Roche).
Clodronate (Sigma) was encapsulated in liposomes as in Zeisberger and colleagues (18).
Array conducted by RayBiotech Inc.
A total of 1 × 105 wild-type and S100A10-null peritoneal macrophages were added to the upper reservoir of Matrigel-coated invasion chambers (BD; 8-μm pore) in the presence or absence of 0.5 μmol/L plasminogen (American Diagnostica). Macrophages were allowed to invade toward 1 × 105 LLC cells in the lower reservoir for 48 hours with or without aprotinin (2.2 μmol/L; Sigma-Aldrich). Macrophages invading to the underside of the Transwell insert were stained with hematoxylin and eosin (Sigma-Aldrich) and counted.
Plasminogen activation assay
A total of 2.5 × 105 wild-type or S100A10-null thioglycolate-elicited macrophages were plated in 96-well plates for 24 hours. Cells were washed and incubated with or without uPA (50 nmol/L) for 10 minutes prior to the addition of plasminogen (0.5 μmol/L; American Diagnostica) with or without aprotinin (2.2 μmol/L; Sigma-Aldrich) and S2251, a chromogenic plasmin substrate (500 μmol/L; Chromogenix, Diapharma Group). The rate of appearance of the S2251 cleavage product was measured spectrophotometrically at 405 nm with readings taken every minute for 2 hours.
A total of 1 × 105 macrophages and 0.2 × 105 LLC cells plated for 24 hours prior to a 4-hour incubation with varying levels of clodrolip. Following a 24-hour recovery period, cells were subjected to an MTT assay (CellTiter 96; Promega) to assess viability.
MTT was conducted on macrophages and LLC cells as per the manufacturer's instructions (CellTiter 96; Promega).
All statistical calculations were conducted in Prism (GraphPad Prism version 5.0 Software). All errors were expressed as SEM and all the Student t tests conducted were 2-tailed. ANOVA with Tukey post analysis was used to calculate the statistical significance between more than 2 groups.
Results and Discussion
Our previous studies had shown that macrophage migration in response to thioglycolate-induced peritonitis was dramatically diminished in S100A10-null mice (12). The inability of these macrophages to respond to this inflammatory stimulus was due to the reduced capability of the S100A10-null macrophages to convert plasminogen to plasmin. Because the site of tumor growth is also a site of inflammation, we tested the possibility that S100A10 might be required for the recruitment of macrophages to the tumor site. Therefore, we inoculated wild-type and S100A10-null mice with LLC cells and measured the kinetics of tumor growth. Surprisingly, we found that S100A10-null tumor growth ceased at approximately 7 days, whereas tumors in the wild-type controls displayed nearly exponential growth (Fig. 1A). Eighteen days after the initiation of tumor growth, we observed that LLC tumors in wild-type mice were greater than 10-fold the weight and larger in size than those in the S100A10-null mice (Fig. 1B and C). To rule out the possibility that the tumor growth deficit was limited to Lewis lung carcinoma, we examined another tumor model system and observed that the tumor growth rate from T241 murine fibrosarcoma cells was also dramatically reduced in the S100A10-null mice (Fig. 1D). Immunohistochemical assessment of the PCNA status indicated that there was no difference in the proliferative rates of the tumor cells in either mouse type (Fig. 2A). There was, however, a moderate increase in apoptotic cells in the tumors on S100A10-null mice, identified by TUNEL (Fig. 2B). Taken together, these observations suggest that apoptosis contributes to the tumor growth inhibition in the S100A10-null environment, although the apoptotic mechanism is unclear.
Immunohistochemical analysis indicated that the tumors grown in the S100A10-null mice contained much fewer macrophages than the wild-type tumors, suggesting that the S100A10-null macrophages failed to colonize the LLC (Fig. 3A and B) and T241 tumors (Fig. 3C). S100A10-null macrophages could only be observed along the absolute tumor tissue border, whereas the wild-type macrophages were abundant throughout the tumor (Fig. 3B). Macrophages are known to infiltrate and support tumor growth (13). Therefore, these results suggested that S100A10 plays a key role in the recruitment of macrophages to the tumor site and that the inability of S100A10-null macrophages to reach the tumor site might result in a tumor growth deficit, although we could not rule out the possibility that S100A10 played a role in the function of other tumor-associated cells.
We had also considered the possibility that the difference in macrophage density could also result from lower levels of the cytokines emanating from the tumor cells in the S100A10-null mice. However, the MCP-1, CCL5, granulocyte macrophage colony-stimulating factor, and macrophage colony-stimulating factor levels were either the same or elevated in the S100A10-null environment, indicating that the reduction in recruitment of S100A10-null macrophages was not due to insufficient stimuli for migration of the macrophages (Supplementary Fig. S1).
To test the possibility that the tumor growth deficit observed in the S100A10-null mice was due to an inability of the macrophages to activate plasminogen to plasmin, facilitating tumor infiltration, we conducted a plasminogen activation assay. The activation of plasminogen to plasmin was quantified spectrophotometrically by the appearance of a cleavage product from a plasmin-specific substrate, S2251. We observed that S100A10-null macrophages had a significantly impaired rate of plasminogen activation compared with macrophages from wild-type mice and that plasmin activity could be abrogated by the addition aprotinin, a plasmin specific protease inhibitor (Fig. 4A).
To assess whether this impairment in plasminogen dependent proteolysis affected the invasive capability of the S100A10-null macrophages, an invasion assay was conducted in which macrophages invaded through a Transwell insert coated with a synthetic basement membrane (Matrigel) toward an LLC monolayer in the presence of plasminogen. A drastic decrease in invasive capability was observed with S100A10-null macrophages, of which approximately 80% fewer cells were capable of crossing the barrier compared with macrophages from the wild-type mice. The invasion of the macrophages was dependent on the presence of plasminogen in the upper chamber and was blocked by the addition of aprotinin, a plasmin inhibitor (Fig. 4B). When the invasion assays were repeated in the absence of a Matrigel barrier, the migration of macrophages from wild-type and S100A10-null mice through the inserts was indistinguishable, suggesting that the protease-independent macrophage migration was unaffected by loss of S100A10 (Fig. 4C). Collectively, these data show that the reduction in recruitment of S100A10-null macrophages to tumors was not due to an insufficient stimulus for migration because the levels of macrophage chemoattractants were similar in all tumors (Supplementary Fig. S1). Furthermore, while the wild-type and S100A10-null macrophages are capable of migrating toward a chemoattractant (Fig. 4C), the capacity of the S100A10-null macrophages to generate plasmin (Fig. 4A) and to invade was lacking (Fig. 4B). These data suggest that S100A10-dependent proteolysis plays a key role in macrophage invasion into the tumor site.
Angiogenesis, the formation of new blood vessels from preexisting vessels, is commonly associated with TAM function (4, 19). We observed that compared with tumors grown in the wild-type mice, tumors from the S100A10-null mice had a 58% decrease in vessel density (Fig. 5A). In addition, multiple fibrin-occluded vessels were identified in the vasculature of the tumors grown in the S100A10-null mice (Fig. 5B and C). The absence of S100A10 from the tumor vasculature of the S100A10-null mouse was confirmed by anti-S100A10 immunohistochemistry (Supplementary Fig. S2). Tumors have an increased ability to activate fibrinogen to fibrin, forming clots which accumulate in blood vessels (20). Vessels in the wild-type mice express S100A10 and can produce plasmin to clear the fibrin accumulation. Studies with the S100A10-null mouse have shown that he S100A10-null endothelial cells have a reduced fibrinolytic activity, thereby causing the accumulation of fibrin in the tissue (21). Thus, the loss of tumor growth in the S100A10-null mice could be explained, in part, by the loss in TAM recruitment, the loss in tumor angiogenesis, and an inability of the existing vasculature to maintain vascular fluidity.
To directly access the role that macrophages played in our tumor model system, we tested the effect of injecting macrophages, collected from wild-type mice, into the peritoneum of S100A10-null mice prior to subcutaneous tumor implantation. We observed that the wild-type macrophages stimulated tumor growth (Fig. 6A and C) and produced an increase in tumor vascular density to that observed in tumors grown in S100A10-null mice (Fig. 5A). However, wild-type macrophages injected into the peritoneum of wild-type mice had no effect on tumor vascular density (Fig. 5A) or tumor growth in these wild-type mice (Fig. 6A and C). Importantly, an intraperitoneal injection of macrophages isolated from S100A10-null mice had no effect on tumor growth in S100A10-null mice (Fig. 6A and C). These results indicate that the wild-type macrophages, but not S100A10-null macrophages, were capable of stimulating both angiogenesis and subsequent tumor growth in the S100A10-null mice. In contrast, we observed that the direct injection of either wild-type or S100A10-null macrophages into the tumors growing in S100A10-null mice stimulated tumor growth to similar levels to that of wild-type mice (Fig. 6B and D). These results indicate that the tumor growth deficit could be overcome by S100A10 expressing wild-type macrophages infiltrating into the tumor from the peritoneum or by the direct intratumoral delivery of S100A10-null macrophages. Thus, macrophages lacking S100A10 are capable of stimulating tumor growth but lack the ability to transit into the tumor. In addition, the transfer of wild-type macrophages by either route had no effect on tumor growth in the wild-type mice, suggesting that endogenous macrophage recruitment was sufficient to support tumor growth in wild-type mice.
Tumor growth is promoted by macrophages as well as by other cells including neutrophils and fibroblasts, all of which lack S100A10 in our mouse model (6). We therefore sought to address whether selective depletion of macrophages would also inhibit tumor growth. It has been shown that macrophages may be selectively depleted in mice using liposome-encapsulated clodronate (clodrolip; ref. 18). We treated LLC tumor–bearing wild-type mice with intraperitoneal inoculations of clodrolip or with an empty liposome control at various points throughout tumor progression. We conducted fluorescence-assisted cell-sorting analysis of the 16-day-old tumors and observed that clodrolip treatment resulted in an approximate 75% depletion of the macrophage content of the tumors compared with the control-treated tumors (Fig. 7A, left and middle). We observed that clodrolip-mediated reduction of TAMs also caused a dramatic reduction in tumor growth (Fig. 7B) and mass (Fig. 7C) and produced a strikingly similar tumor growth profile to that in the S100A10-null mouse (Fig. 7B and C). To rule out the possibility that the tumor growth deficit was due to a cytotoxic effect of clodrolip on the tumor cells, we incubated LLC cells and macrophages with varying concentrations of clodrolip. Macrophages showed a significant decrease in survival relative to clodrolip concentration, whereas LLC cell survivability remained unaffected (Supplementary Fig. S3).
Tumor growth and metastasis require an intimate relationship between macrophages and cancer cells, the current study shows that the recruitment of macrophages to the tumor site is mediated by S100A10. Although other mechanisms cannot be excluded, we propose that S100A10-dependent plasmin generation plays a critical role in the movement of macrophages to the tumor site. These findings highlight a new therapeutic modality in which tumor progression may be controlled by targeting tumor-associated macrophages.
Disclosure Potential of Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors were supported by a grant from the Canadian Cancer Society Research Institute and Canadian Institutes of Health Research.
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.
S100A10-null mice were donated by P. Svenningsson (Karolinska Institute), and P. Greengard (Rockefeller University). The authors thank B. Hildebrand and the Dalhousie University Carleton Animal Care Facility for animal support, R. De Antueno (Dalhousie University) for liposome technical support, Pat Colp (Histology Research Services, Dalhousie University) for immunohistochemical technical support, and M. Taboski (Dalhousie University) for helpful comments.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received May 24, 2011.
- Revision received August 29, 2011.
- Accepted August 30, 2011.
- ©2011 American Association for Cancer Research.