The chemokine receptor CCR6 facilitates the onset of mammary neoplasia in the MMTV-PyMT mouse model via recruitment of tumor-promoting macrophages

To first determine whether CCR6 may play a role in the regulation of mammary neoplasia, we investigated expression of the receptor in CD45-negative normal mouse mammary cells, and cells from various tumor stages (representative H&E pictures in Fig. 1a). CCR6 was expressed on a low proportion of normal mammary cells, but this proportion was greatly amplified in accordance with increasingly higher grades of MMTV-PyMT cancer including initial hyperplasia, early carcinoma and late carcinoma as indicated (Fig. 1a). This is consistent with human breast cancer [23] and other mouse models of cancer [17]. Additionally, in both non-PyMT Ccr6 ^WT and MMTV-PyMT Ccr6 ^WT mammary tissues the ligand for CCR6, CCL20, was highly expressed at concentrations over 50 ng/mg tissue (Fig. 1b). These data raise the possibility of a role for CCR6 in breast cancer development.

To next establish the role of CCR6 deletion on mammary tumorigenesis, we compared the rate and total extent of PyMT-driven neoplasia between MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mice. Tumor onset was significantly delayed in MMTV-PyMT Ccr6 ^−/− mice (Fig. 1c), with some mice not developing palpable tumors until 150 days old (21 weeks) compared to a maximum onset age of 130 days old (18 weeks) for MMTV-PyMT Ccr6 ^WT counterparts (Fig. 1d).

In order to assess the impact of CCR6 on the later stages of cancerogenesis, MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mice were sacrificed at 22–24 weeks of age and the total number of mammary tumors per mouse was determined. We found that MMTV-PyMT Ccr6 ^−/− mice had significantly decreased tumor incidence compared to MMTV-PyMT Ccr6 ^WT animals (Fig. 1e). Together, these results implicate CCR6 as being an important player in breast oncogenesis.

We then sought to examine whether CCR6 influenced early hyperplasia of mammary glands during tumor initiation as well as late stage tumorigenesis. Glands from 8-week-old MMTV-PyMT Ccr6 ^−/− and Ccr6 ^WT mice were extracted and whole mounted for quantitation of hyperplastic/early-neoplastic lesions (representative images from both genotypes shown in Fig. 2a). We found that the deletion of CCR6 significantly reduced the initial hyperplastic outgrowth within the gland (Fig. 2b), a common indicator of future breast cancer development. As the total area of PyMT-driven hyperplastic outgrowth per gland was reduced by threefold in CCR6-null animals, we concluded that the effect of CCR6 on mammary tumorigenesis is manifested very early on in cancer development.

The difference seen in early tumor initiation between MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mice can potentially result from a difference in normal mammary development, which may then have translated into decreased hyperplasia. We therefore extracted pubertal mammary glands from non-PyMT 6-week-old Ccr6 ^WT and Ccr6 ^−/− mice. Representative glands are shown in Additional file 1: Figure S1a. When ductal epithelial growth was quantitated, we observed no statistically significant difference in ductal length, number of terminal end structures or branching between Ccr6 ^WT and Ccr6 ^−/− mice (Additional file 1: Figure S1b-d), and hence the overall effect of CCR6 deletion on normal mammary gland biology appears to be minimal and is unlikely to account for differences in PyMT-driven tumor development between the two genotypes. Furthermore, the levels of CCL20 were not statistically different between non-PyMT Ccr6 ^WT and MMTV-PyMT Ccr6 ^WT mammary tissues (Fig. 1b), demonstrating that the expression of CCL20 is not affected by the process of tumorigenesis. Taken together, these data show that early stage tumorigenesis is mediated by a CCR6-dependent mechanism, without affecting normal mammary morphogenesis.

To investigate the mechanism underlying CCR6-driven mammary tumorigenesis, we studied the epithelial cell population to determine if CCR6 was having a direct effect on cell proliferation. Cells at the stage of early neoplasia from MMTV-PyMT Ccr6 ^WT mammary glands were assayed for proliferation upon stimulation with CCL20. No differences in cell proliferation were observed (Fig. 3a). Furthermore, Ki67 staining of sectioned hyperplastic mammary glands from MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mice showed that epithelial cells in knock-out mice are still able to adequately proliferate and staining of Ki67 is equal to that in the Ccr6 ^WT (Fig. 3b). This suggests that the role of CCR6 in breast cancer is independent of epithelial cells.

We next determined whether CCR6 may exert its effect by skewing distinct cell populations within the bulk epithelium, as we have reported previously for the chemokine receptor CCR7 [29]. The current prevailing paradigm has mammary epithelial and breast cancer cells hierarchically organized with a self-renewing, quiescent, multipotent progenitor (or stem-like cell) population giving rise to basal and luminal progenitors which in turn differentiate into specific lineages making up the mammary gland and heterogenous breast tumors [30]. Recently, a number of immune mediators including chemokine receptors have been implicated in maintenance of the cancer stem-like cells within mammary tumors (reviewed in [31]). We therefore tested the potential link between the tumor-promoting function of CCR6 and breast cancer stem-like cell pools.

Freshly isolated MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mammary cells from pre-neoplastic mice at 8–9 weeks-old were assayed by flow cytometry for expression of cell surface markers CD24 and CD29 [32] (representative plots shown in Fig. 4a), which were previously used to define stem cells in the MMTV-PyMT [33, 34] and other breast cancer mouse models [35, 36]. We found that the deletion of CCR6 did not alter the proportions of the stem cell-enriched basal population (CD24⁺CD29^hi) nor the luminal population (CD24⁺CD29^lo) (Fig. 4b) in hyperplastic mammary glands.

We also investigated the effect of CCR6 ablation on functional stem-like mammary cancer cell pools using the mammosphere assay, which is used to select for colonies of early stem-like progenitors [37]. Pre-neoplastic MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mammary cells were seeded into non-adherent mammosphere culture and allowed to grow for 7 days. The deletion of CCR6 did not alter mammosphere-forming efficiency (Fig. 4c), and when sphere cultures were stimulated with varying concentrations of CCL20, no change in their ability to form mammospheres (Fig. 4d) was observed, supporting results obtained using flow cytometric analysis.

Having found no difference in the proportion or function of stem cell-like pools within early tumorigenic lesions, we next tested a bona fide property of cancer stem cells – their ability to propagate tumors upon transplantation. MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− donor mammary tissue was transferred into recipient fat pads of syngeneic non-PyMT Ccr6 ^WT mice. Using this approach, we found no significant difference between MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− tissue in the ability to form outgrowths when transplanted into Ccr6 ^WT recipients (Fig. 4e), indicating that the deletion of CCR6 did not reduce the tumor-propagating capability of mammary epithelium.

Altogether, these results demonstrate that the role of CCR6 in breast cancer is independent of breast epithelial and progenitor cells, raising the possibility that its mechanism of action involves the tumor microenvironment.

To test whether the reduced mammary tumorigenesis caused by the deletion of CCR6 was due to an effect of the CCR6-CCL20 axis on the tumor microenvironment we next investigated by flow cytometry the levels and identity of tumor-infiltrating leukocytes in mammary tumors at the stage of early carcinoma (see Fig. 1a). Tumor-associated macrophages have been extensively implicated in tumor promotion both in the mammary gland and elsewhere, due to their role in angiogenesis, cell proliferation and tissue remodeling [5]. To initially examine the polarization of TAMs in the MMTV-PyMT mice, macrophages were assessed for expression of prototypic markers interleukin-4-receptor (IL4-R) and mannose receptor (CD206), which have been used previously in flow cytometric analysis to distinguish alternatively-activated M2 macrophages from classically-activated M1 [38‐42]. We estimated using these markers that a high proportion of TAMs were of an M2-like phenotype (Additional file 1: Figure S2a), as has been suggested previously for MMTV-PyMT mammary tumors [43]. Interestingly, CCR6 was found to be highly expressed within the TAM population as it was detected on greater than 60 % of total macrophages (Additional file 1: Figure S2b). Most importantly, CCR6 was expressed at higher levels and on a significantly higher proportion of putative M2 macrophages (up to 90 %) than M1 (Fig. 5a and Additional file 1: Figure S2b), using both IL4-R and CD206 to delineate the populations. This strong correlation potentially implicates CCR6 in the regulation of pro-tumorigenic macrophages within the mammary gland microenvironment.

We then assessed the levels of macrophages in mammary tumors from MMTV-PyMT Ccr6 ^WT and Ccr6 ^−/− mice, and found that the proportion and overall numbers of TAMs were significantly reduced in MMTV-PyMT Ccr6 ^−/− mammary tumors relative to Ccr6 ^WT (Fig. 5b). Furthermore, the deletion of CCR6 resulted in a shift towards an M1 macrophage phenotype, defined by both IL4-R and CD206 prototypic markers (Fig. 5c).

The deletion of CCR6 also resulted in reduced trafficking of dendritic cells to the tumor site, consistent with previous studies, which demonstrated a reduced overall migratory ability of dendritic cells in CCR6-null mice [44, 45]. Although, overall numbers of tumor-infiltrating dendritic cells were much lower than the corresponding macrophage population. Notably, the majority of tumor-infiltrating dendritic cells were CCR6-positive, consistent with previous findings [12] (Additional file 1: Figure S3). We have also assessed CCR6 expression on B cells and specific T cell subsets (helper T cells (Th), cytotoxic T cells (Tc), and regulatory T cells (Treg)) within mammary tumors. In agreement with earlier reports for various biological settings [12, 46], all tested infiltrating leukocyte subsets expressed CCR6 at varying levels. When CCR6 was ablated, only the Tc cell subset showed a slight increase in the MMTV-PyMT Ccr6 ^−/− as a proportion of CD45⁺ tumor-infiltrating cells, however no significant differences were found in total cell numbers between the two genotypes (Additional file 1: Figures S4 and S5).

Our findings thus suggest that CCR6 promotes mammary tumorigenesis through an epithelium-independent mechanism involving tumor-infiltrating macrophages.

We then sought to provide definitive evidence for the macrophage-mediating function of CCR6 in mammary tumor promotion using an in vivo macrophage reconstitution assay. Reconstitution assays, sometimes referred to as “add-back” assays, are frequently used to underscore a role for various cellular subsets in multiple pathological settings, and macrophage reconstitution has been used previously in mammary gland studies [47]. A schematic of the experimental setup is shown in Fig. 6a. MMTV-PyMT mammary tumor cells from Ccr6 ^WT donor mice were purified and transplanted into the inguinal mammary fat pads of non-PyMT Ccr6 ^WT and Ccr6 ^−/− recipients at 5 weeks of age. Two days post-transplantation, TAMs (CD45⁺F4/80⁺) were sorted from excised and dissociated MMTV-PyMT Ccr6 ^WT tumors (Fig. 6b) and orthotopically injected into a group of Ccr6 ^−/− recipients as specified in Fig. 6a. All other recipients received sham injections with vehicle only, followed by assessment of mammary tumor growth 6 weeks later.

In agreement with the results from spontaneous tumorigenesis studies, we have found that tumors grew significantly slower in the Ccr6 ^−/− hosts compared to Ccr6 ^WT , indicating that CCR6 is required in the mammary stroma for robust tumor development. However, when the reduced macrophage phenotype was restored in Ccr6 ^−/− mice through orthotopical injections, the tumor latency was significantly shortened, approaching that of the Ccr6 ^WT mice (Fig. 6c).

It was further found that the supplementation of Ccr6 ^−/− mice with TAMs restored the efficiency of tumor growth (measured by weight of tumor-bearing mammary glands) within these mice to that seen in the Ccr6 ^WT , whilst Ccr6 ^−/− mice that received sham injections displayed reduced tumorigenesis (Fig. 6d) as seen in the spontaneous model (see Fig. 1).

Enumeration of macrophages within grafted tumors in Ccr6 ^WT and Ccr6 ^−/− mice (Fig. 6e) paralleled the results seen in spontaneous mammary tumors (see Fig. 5b). Whilst there was an upward trend towards increased TAMs in the Ccr6 ^−/− mice that received TAM injections, the difference was not statistically significant (Fig. 6e). This, combined with the decreased tumor growth in the CCR6-null mice, indicated that the support of macrophages is essential at the early stages of tumor growth.

We have thus established an essential role for CCR6 in the tumor microenvironment, providing a causative link between this receptor, infiltrating macrophages and mammary tumor development. Hence, therapeutic opportunities may be explored to control breast cancer progression, via manipulation of the CCR6-CCL20 axis to control tumor-promoting macrophages.