Arthritis augments breast cancer metastasis: role of mast cells and SCF/c-Kit signaling

SKG mice were established from a closed breeding colony of Balb/C mice [20]. At about 3 months old, mice were injected with 1 × 10⁵ 4T1 cells (in 100 μl of PBS) in the mammary fat pad. Balb/C mice were used as the nonarthritic controls. After 35 days of tumor challenge, all mice were killed, and analysis was conducted.

PyV MT oncogenic mice were originally a gift from Dr. W. J. Muller (McGill University, Toronto, Ontario, Canada) [17]. The PyV MT mice were bred to be congenic on the C57Bl/6 background and were used in several of our prior publications [9, 21‐24]. Genotyping of these mice was conducted as described previously [21].

All mice were bred and maintained in specific pathogen-free conditions in the UNCC Animal Facility. All experimental procedures were conducted according to Institutional Animal Care and Use Committee guidelines. All protocols were approved by the UNCC Internal Animal Care Review Committee (IACUC ID: 08-036.0 and 11-015.0).

The 4T1 cells were purchased from The American Type Cell Culture Collection (Manassas, VA, USA). This highly metastatic BC cell line is derived from a spontaneously arising Balb/C mammary tumor. Cells were maintained in complete RPMI [8]. 4T1 tumors resemble human late-stage metastatic BC [25‐27].

The PyV MT cell lines were generated from PyV MT tumors and cultured as previously described in complete DMEM [9].

The PyV MT mice were injected with 50 μl of 2-mg/ml CII (MD Biosciences, St. Paul, MN, USA) in CFA (Difco Laboratories, Detroit, MI, USA) intradermally about 1.5 cm distal from the base of the tail at 12 weeks of age. Fifty-sixty percent of mice developed arthritis within 15 to 30 days after collagen injection [28]. All mice were killed at 22 weeks, and analysis was conducted.

BMMCs were stained with FITC rat anti-mouse CD117 antibody (BD Pharmingen, San Diego, CA, USA). Acquisition was performed on a FACS Fortessa Cytometer, and analysis was performed by using the Flow Jo program.

We performed BCA assay to load equal quantities of tumor lysates onto sodium dodecylsulfate-polymerase chain reaction (SDS-PAGE) gels. SCF and β-actin antibodies were used at 1:200 dilution (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and were used according to manufacturer's recommendations.

The densitometric analyses of immunoblots as shown in Additional file 1 were performed by using NIH Imaging program. Results are presented as mean values of arbitrary densitometric units corrected for background intensity and normalized to the expression of β-actin.

Lungs, bones, and tumor sections were processed as previously described [8, 9]. Paraffin-embedded blocks were prepared, and 4-μm-thick sections were cut for hematoxylin/eosin (H&E) and pancytokeratin staining. The pancytokeratin antibody was purchased from Santa Cruz and used at 1:50 dilution by following the same protocol as described in our previous publications [8, 9]. To establish the expression of MCs on the tumors, lungs, and bones, anti-mast cell tryptase antibody (ab134932) was purchased from Abcam, Cambridge, MA, USA.

To establish further the expression of MCs, toluidine blue purchased from Fluka Analytical (Exeter, Devon, UK) distributed by Sigma-Aldrich (Sigma Aldrich, St. Louis, MO, USA) was used as directed by the manufacturer. All images were taken by using Olympus DP71 light microscopy.

Cells were plated on chamber slides and grown to 75% confluency followed by immunofluorescence (IF) staining with SCF rabbit antibody at 1:200 dilution and Alexafluor 594 donkey anti-rabbit (Invitrogen) secondary at 1:1,000. Staining was performed as previously described [29]. Slides were examined under Olympus FV1000 fluorescence microscopy, and pictures were taken at 400× magnification.

BM cells were harvested from femurs of mice, and 2 × 10⁶ cells were cultured, as previously described [10]. The cells were cultured in the presence of 10 ng/ml IL-3 and 50 ng of recombinant murine SCF (PeproTech, Rocky Hill, NJ, USA), and the nonadherent cells were passaged every 3 days. Four weeks later, the cells were used as MCs for experiments [10, 30]. The MCs were identified and counted in three different ways: (a) light microscopy: we counted the MCs with a hemocytometer and used trypan blue to detect any dead cells [30]; (b) flow cytometry: BMMCs were stained with FITC rat anti-mouse CD117 antibody, and the percentage of MCs was analyzed with flow cytometry. CD117 is the cKit receptor present on mast cells; and (c) toluidine staining [30]. We counted five fields of toluidine-positive BMMCs generated from BM derived from each mouse.

Blood samples (0.5 to 1.0 ml) were obtained by cardiac puncture from individual tumor-bearing animals by using heparin as the anticoagulant and processed as described previously [29, 31].

Cells were cultured in complete DMEM for 21 days.

SCF expressing 4T1 and PyV MT cells were plated over transwell inserts in the upper chamber (BD Biosciences, San Diego, CA, USA) and permitted to migrate toward MCs in the lower chamber for 15 hours. Percentage migration was determined as described earlier [29]. We added neutralizing SCF antibody (R&D Systems, Minneapolis, MN, USA) to tumor cells or anti-cKit antibody (eBioscience, San Diego, CA, USA) to the MCs as independent groups for 12 hours and analyzed the migration of BC cells toward MCs.

The Pix array 100 x-ray machine (Bi Optic Inc, Santa Clara, CA, USA) was used for bone imaging, as described [32].

When the tumors were about 5 mm in length (after about 6 days of 4T1 injections), mice were injected (IP) with neutralizing rat anti-mouse cKit-receptor antibody (50 μg) (eBioscience) or 20 μg goat anti-mouse-SCF antibody (R&D Systems, Minneapolis, MN, USA) once a week for 4 weeks. The mice were euthanized 24 hours after the last injections at about 35 days after tumor challenge. For treatment experiments, 4T1 cells transfected with green fluorescent protein (4T1-GFP) were injected.

Data were analyzed by using GraphPad software. Results are expressed as mean ± SEM and are representative of three or more separate experiments. Comparison of groups was done by using one-way or two-way ANOVA followed by the Bonferroni posttest for multiple comparisons (*P < 0.05, **P < 0.01, ***P < 0.001). The Student t test was used for comparing the level of significance between two experimental groups.

We first demonstrated that the 4T1 tumor burden is significantly higher in the arthritic SKG versus the nonarthritic Balb/C mice (Figure 1A). Second, we substantiated this finding in the PyV MT mice with higher tumor burden in the arthritic PyV MT versus the nonarthritic PyV MT mice (Figure 1B). Third, metastases to the lungs and bones were compared. We report that nine of 10 SKG mice showed lung metastasis, whereas only four of 10 Balb/C mice showed metastasis (Figure 1C). Similarly, seven of 10 arthritic PyV MT mice showed lung metastasis, whereas only three of 10 nonarthritic Py VMT mice showed the same (Figure 1C). Similar trends were observed with bone metastasis, with eight of 10 SKG mice developing bone metastasis, whereas only three of 10 Balb/C developing the same (Figure 1D). Likewise, five of 10 arthritic PyV MT mice developed bone metastasis, whereas none of the nonarthritic PyV MT mice developed bone metastasis (Figure 1D).

Circulating tumor cells (CTCs) are being aggressively explored as a prognostic tool and a measure of metastasis and response to therapy [29, 31]. One way to evaluate CTCs in mice is to culture the blood and detect colonies of tumor cells. We detected higher levels of CTCs (that formed colonies within 3 weeks of culture) in the blood of the tumor-bearing arthritic mice compared with the blood from the tumor-bearing nonarthritic mice (Figure 1E). The data corroborate our previous studies [8, 9] and further establish that the proinflammatory microenvironment created in bones and lungs by AA is highly conducive for BC cells to form metastases.

We first assessed the expression of MCs in the tumor and metastatic niches (bone and lung). Tryptase is the most abundant secretory granule-derived serine contained in MCs that has been used as a marker for MC activation. In addition, tryptase has been shown to be a sensitive and specific marker for the localization of MCs in tissues [33‐35]. By doing MC tryptase staining by IHC, we detected significantly higher numbers of MCs in the tumors of arthritic versus nonarthritic mice with BC (Figure 2A). We detected infiltration of MCs in the tumors of SKG versus Balb/C mice with BC (Figure 2B) and in the PyV MT mice induced with arthritis versus PyV MT mice with no arthritis (Figure 2C). We further identified significantly higher levels of MCs in the sites of metastasis: lungs (Figure 3) and bones (Figure 4) of arthritic versus nonarthritic mice in both (SKG and PyV MT) models. Thus, data suggest that an increase in MC migration and activation occurs within the tumor as well as in the metastatic niches, as indicated by the increased MC expression.

However, to confirm further the in vivo infiltration of MCs, toluidine staining was also conducted on tissue sections. MCs contain granules (metachromatic) composed of heparin and histamine. Toluidine blue stains MC blue-purple (metachromatic staining) and the background blue (orthochromatic staining). Clearly, in both the arthritic PyV MT and SKG mice, we observe a significantly increased infiltration of MCs within the tumor (Figure 5A, B), lung (Figure 5C, D), and bone (Figure 5E, F). Shown in Figures 2 through 5 are representative MC staining in the tissues from one mouse per experimental group. Also shown in Figure 5B is a magnified image of MC infiltration in the 4T1 tumor section of an arthritic SKG mouse. To quantitate the results from Figures 2 through 5, MCs were counted under the microscope (six mice per experimental group and 10 fields per section per organ), and the mean numbers reflected in the graph.

Most important, we noted that in the non-tumor-bearing arthritic mice (SKG and CII-injected C57BL/6 mice), the number of MCs was increased compared with nonarthritic tumor-bearing Balb/C or PyV MT mice (Figures 3 through 5), indicating that the arthritic milieu by itself can increase the MC population in the bones and lungs, creating a conducive niche that attracts the SCF-expressing tumor cells. The MC expression in the lungs and bones of nonarthritic Balb/C and C57BL/6 mice with no tumor was found to be negligible (data not shown). We therefore believe that once tumor cells populate the bone and lungs, the tumor-derived SCF binds to the c-Kit receptor on MCs and helps in the differentiation, maturation, and survival of MCs, remodeling the microenvironment and further increasing the population of MCs [10] and Figures 2 through 5. Thus, the higher tumor burden and metastases seen in the arthritic mice (Figure 1) may be due to the increase in MC infiltration in the tumors and metastatic niches.

To confirm that MC activation and proliferation in the tumors and sites of metastasis may be mediated by tumor-derived SCF, we assessed the expression of SCF on 4T1 and PyV MT tumors in vivo by Western blotting. We observed that SCF was highly expressed by the 4T1 and PyV MT tumors in vivo (Figure 6A). No difference was found in SCF protein expression levels between tumors derived from arthritic versus nonarthritic mice, indicating that the arthritic milieu does not influence SCF levels. SCF expression was also confirmed on 4T1 and PyV MT tumor cell lines in vitro by IF (Figure 6B).

Because we observed increased MCs in the arthritic mice even in the absence of a tumor, we determined whether the differentiation of MCs from the bone marrow-derived precursors was affected by the arthritic milieu. We cultured 2 × 10⁶ BM cells in the presence of IL-3 and SCF for 30 days to induce the precursor cells to differentiate into MCs. BM cells differentiate into MCs under selection with recombinant IL-3 and SCF [30]. We observed significant increase in the bone marrow-derived mast cells (BMMCs) in non-tumor-bearing arthritic SKG and C57BL/6+CII mice versus Balb/C with tumor or PyV MT mice (Figure 7A and 7C; Tables 1 and 2). Balb/C or C57BL/6 had no detectable levels of mast cells (Table 1 and 2). Numbers of MCs are represented in Figure 7A and 7C, respectively. Light-microscopic images of differentiated MCs from representative groups are shown in Figure 7B and 7D, respectively. Significant increase in MC differentiation was observed in tumor-bearing SKG and arthritic PyV MT mice as compared with tumor-bearing nonarthritic Balb/C mice and nonarthritic PyV MT mice (Figure 7A through D). Because differentiated mast cells should express the c-Kit receptor, we stained the cells with a c-Kit antibody (CD117) and analyzed with flow cytometry. A representative histogram is shown in Figure 7E and 7F. To confirm that the differentiated cells were indeed MCs, toluidine staining was conducted, and representative staining is shown in Figure 7G. Data suggest that the BM-derived precursor cells that are predestined to differentiate into MCs are significantly higher in the arthritic BM milieu before tumors are formed and that the tumors further enhance the differentiation (Tables 1 and 2). It can therefore be speculated that AA probably affects hematopoiesis and induces the production of MCs, thus creating a microenvironment appropriate for tumor cells to home and form metastases.

To determine that the chemotactic migration of tumor cells into metastatic sites is indeed mediated by MCs, and SCF on tumors and c-Kit on mast cells are necessary for this, we conducted an in vitro migration assay. SCF expressing 4T1 and PyV MT cells (as shown in Figure 6B) were placed in the upper chamber, and BMMCs that express c-Kit (Figure 7E, F) of arthritic and nonarthritic mice in the lower chamber. We found significant increase in the migration of 4T1 and PyV MT cells toward the MCs that were derived from 4T1-bearing arthritic SKG or PyV MT mice (Figure 7H and 7I, respectively). We further showed that pretreatment of 4T1 and PyV MT cells with anti-SCF antibody or adding anti-cKit antibody to the BMMCs in the lower chamber significantly decreased the migration of the tumor cells toward the MCs (Figure 7H and 7I). The data suggest that the SCF/c-Kit signaling may be one of the drivers of BC-associated bone and lung metastasis.

A key finding from this study indicates that the SCF/c-Kit signaling was necessary for the tumor cells to migrate toward the MCs in an in vitro migration assay. This set the stage for examining the effects of blocking this signaling in vivo. When mice were treated with therapy to target the c-Kit receptor or SCF, a significant decrease in tumor burden was noted in the mice treated with anti-cKit antibody but not with anti-SCF antibody (Figure 8A). However, metastasis to the lungs (Figure 8B through E) and bone (Figure 8F through I) was significantly reduced in both treatment groups. We observed about threefold and about twofold decrease in lung metastasis in mice treated with anti-cKit and anti-SCF, respectively (Figure 8B). Figure 8C shows representative images of GFP-positive 4T1 cells in the lungs of SKG mice injected with 4T1 cells with no treatment versus the treated groups. Figure 8D shows the metastatic lesions in the lungs and the pancytokeratin staining, confirming that the metastatic patches seen are epithelial tumor cells (Figure 8E). Similarly, we observed about threefold and about twofold decrease in metastatic bone lesions in mice treated with anti-cKit and anti-SCF, respectively (Figure 8F). Pancytokeratin staining in the BM and bone tissues (Figure 8G, H) and representative radiographic x-ray bone images (Figure 8I) confirmed the presence of epithelial tumor cells in the bones of these mice.

Finally, to determine the levels of mast cells in the treated mice, we assessed the number of mast cells in the tumor, bone, and lung of treated mice. We observed a significant decrease in MC infiltration in the tumor, lungs, and bones of mice treated with anti-cKit receptor antibody (Figures 10A through C and 11A through C). Representative tryptase and toluidine staining of tumor, lung, and bone tissue sections is shown in Figures 10D through F and 11D through F, respectively. Interestingly, mice treated with anti-SCF also showed significant decrease in MC accumulation in the tumor, lung, and bone (Figure 10D(ii), E(ii), F(ii), and 11D(ii), E(ii), and 11F(ii), respectively), confirming the role of SCF in MC proliferation [10]. Taken together, these findings confirm that the decrease in metastasis in the treated groups (Figure 8) was driven by low MC accumulation.