Increasing evidence suggests that macrophages play a role in the normal development of specific organs. In the pubertal mammary gland, macrophages are recruited to the highly mitotic terminal end buds (TEBs) from which ducts elongate and branch to give rise to a mature ductal tree. Macrophages are presumed to play a role in tissue remodeling by engulfing apoptotic epithelial cells during ductal morphogenesis 12. Colony stimulating factor-1 (Csf1; also referred to as macrophage-Csf) is a key growth factor that regulates the proliferation and survival of macrophages 3. Csf1^op/op mice, homozygous for a null mutation in Csf1, exhibit multiple defects and have reduced macrophage numbers in most tissues including the mammary gland. These mice are severely runted, toothless and osteopetrotic, owing to an osteoclast deficiency that impairs bone resorption. The mammary glands of Csf1^op/op mice display lower numbers of terminal end buds as well as reduced ductal branching and elongation 4. During pregnancy, Csf1^op/op glands develop precocious alveolar units but fail to switch to the lactational state resulting in impaired lactation 5. Mammary-specific rescue has shown that the mammary gland phenotype is not due to secondary effects induced by a systemic Csf1 deficiency 6. Targeted ablation of the receptor for Csf1 (Csf1r or c-fms), recapitulated all the phenotypes observed in Csf1^op/op mice, although the receptor-null mice exhibited slightly more severe defects 7.

The nature of the mammary stem cell niche is yet to be defined but candidate cell types that may constitute the niche include epithelial cells, macrophages and eosinophils, among other cells in the mammary stroma (fibroblasts and adipocytes). One mechanism that may underlie the requirement of macrophages for normal mammopoiesis could be their interaction with mammary stem cells (MaSCs) 89. To explore this possibility, we have utilized two in vivo approaches to determine whether macrophages are required for the function of MaSCs. We report that stem/progenitor cell activity is markedly attenuated by depletion of macrophages from the mammary gland, indicating that macrophages may influence stem cell function.

To evaluate whether the impaired ductal morphogenesis that occurs in Csf1^op/op mice (Figure 1a) reflects a perturbation in MaSC activity, freshly sorted CD29^hiCD24⁺ mammary epithelial cells were isolated from BALB/c Csf1^op/op and wild-type glands. They were then transplanted into the de-epithelialized fat pads of three-week-old BALB/c mice at limiting dilution. The CD29^hiCD24⁺ subpopulation has previously been shown to be enriched for MaSCs but also contains mature myoepithelial cells and possibly basal progenitor cells 8. Cells were purified from 12 week-old Csf1^op/op mice and eight-week-old wild-type littermates, in order to compensate for the recognised delay in filling of the fat pad that occurs in Csf1^op/op mutant mice 4. The glands were harvested for wholemount analysis eight weeks post-transplantation.

Limiting dilution analysis of four independent experiments revealed that the stem cell frequency was reduced in Csf1^op/op mice. The mammary repopulating frequency was 1 in 261 (95% confidence interval (CI) = 1/166 to 1/408) versus 1 in 97 (95% CI = 1/66 to 1/141) for Csf1^op/op and wild-type mammary glands, respectively (P < 0.001; Table 1). Estimation of the absolute number of mammary repopulating cells revealed an 18-fold decrease in mutant glands (mean of 702 versus 40 cells per gland in wild-type and Csf1^op/opmice, respectively). In addition to the decrease in repopulating frequency by CD29^hiCD24⁺ cells in the Csf1^op/op group, only 5 of the 45 transplanted fat pads (11 ± 4%; mean ± standard error of the mean (SEM)) yielded outgrowths filling more than 50% of the fat pad, whereas 22 of the 45 transplanted fat pads (49 ± 4%) were filled by outgrowths from wild-type stem cells. Thus, despite being transplanted into a normal microenvironment, the MaSC-enriched population from Csf1^op/op mice did not give rise to fully developed mammary outgrowths (Figure 1b), indicating that the function of mammary stem/progenitor cells had been attenuated in the macrophage-depleted environment.

Consistent with the transplantation data, in vitro colony forming assays on fibroblast feeder layers revealed a two-fold reduction in the clonogenic potential of CD29^hiCD24⁺ cells from Csf1^op/op mice relative to those from control mice (Figure 1c). The absolute numbers of clonogenic cells in the MaSC-enriched population per mammary gland from either wild-type or Csf1^op/opmice were estimated to be 14,666 ± 1530 and 1306 ± 348, respectively (mean ± SEM, n = 3). Notably, a 2.4-fold decrease in colony size was also apparent on comparison of colonies derived from Csf1^op/op CD29^hiCD24⁺ cells relative to those from wild-type mice. The addition of F4/80-positive mouse mammary macrophages to the clonogenic assays had no discernible effect on colony size or morphology, but may reflect inadequate culture conditions for macrophages in the epithelial cell assays (data not shown).

To investigate a potential role for macrophages in the mammary stem cell microenvironment, the wild-type MaSC-enriched population was next transplanted into the macrophage-deficient fat pads of Csf1^op/op mice. Two hundred CD29^hiCD24⁺ cells isolated from the mammary glands of young wild-type female mice (BALB/c) were transplanted into either three-week-old wild-type or six-week-old Csf1^op/op fat pads that had been cleared of endogenous epithelium. Older mutant recipient mice were used to compensate for the marked developmental delay that occurs in Csf1^op/op mice. Substantial outgrowths were observed in wild-type recipients (18 of 24), whereas only a single small outgrowth (1 of 18) comprising two ducts was evident in Csf1^op/op recipients (P < 0.00001; Figures 2a, b). These data suggest that mammary repopulating cells require the presence of resident macrophages in the mouse mammary gland.

To further investigate the effect of macrophage depletion from the mammary fat pad on stem cell activity, an alternative model of macrophage deficiency was sought. Clodronate is a bisphosphonate that can be delivered in a targeted manner when packaged into liposomes 10. The liposomes are non-toxic until ingestion by macrophages, in which they are broken down by liposomal phospholipases to reveal the drug. Sufficient clodronate uptake by macrophages causes cell death by apoptosis. In co-transplantation experiments, CD29^hiCD24⁺ cells (200 cells) from the mammary glands of Rosa26 mice were coinjected into the cleared mammary fat pads with either clodronate-containing or saline-containing (control) liposomes. Six weeks after transplantation, the glands were harvested and the outgrowths analysed by staining for LacZ activity (see Materials and methods). Three independent experiments revealed outgrowths in 22 of 28 (77 ± 5%) fat pads coinjected with saline liposomes compared with only 4 of 28 (18 ± 10%) fat pads co-injected with clodronate-containing liposomes (P < 0.000001). Of the four outgrowths evident in the clodronate group, only one branching outgrowth was observed, while the remaining three were very rudimentary structures (Figure 2c).

Our data reveal that mammary macrophages contribute to normal stem/progenitor cell function in the developing mouse mammary gland. In a macrophage-deficient environment, the MaSC-enriched subpopulation had a markedly diminished capacity to reconstitute mammary outgrowths. The role of macrophages in this organ, however, appears to be supportive rather than integral to MaSC function, because these mice harbor some MaSCs and limited development can occur in Csf1^op/op mice. Interestingly, even in a macrophage-replete system, the function of MaSCs from Csf1^op/op glands was compromised. Indeed, both the absolute number and repopulating capability of MaSCs appear to have been altered by the macrophage-deficient environment, possibly reflecting downregulation of growth factor-mediated signal transduction pathways required for stem cell function.

Here we demonstrate the efficacy of clodronate-containing liposomes for macrophage depletion in the mammary gland. Clodronate liposomes have been successfully utilized for macrophage depletion from liver, spleen, lung, lymph nodes, joints, peritoneal cavity, and testis 10. Local depletion of resident macrophages from the developing mammary gland by clodronate liposomes led to a marked decrease in repopulation by the MaSC-enriched fraction. The depletion of macrophages is likely to be temporary as the liposomes are phagocytosed within the first few days and macrophages presumably return to normal levels by the time of tissue harvest at six weeks post-transplantation. Therefore, mammary macrophages appear to be required for the early stages of repopulation by MaSCs and their progeny.

It is tempting to speculate that mammary macrophages constitute part of the normal MaSC microenvironment and facilitate the maintenance and/or proliferation of stem cells. Gene expression analysis has shown that the MaSC-enriched subpopulation expresses significantly higher levels of Csf1 (approximately five-fold by quantitiative RT-PCR) relative to the luminal epithelial subset. Only macrophages in the mammary gland appear to express the receptor (c-fms or Csf1R) for the Csf1 ligand under normal physiologic conditions, based on immunostaining for the receptor 4 as well as expression in c-fms reporter mice (data not shown). Csf1 may therefore be a chemoattractant for macrophages to the niche in which MaSCs are thought to reside in a basal location. Consistent with this notion, macrophages can be found around the cap cell region of TEBs 4, which has been proposed to be enriched for stem cells 12. The factors released by macrophages that support stem cell activity are yet to be determined but may include macrophage-secreted cytokines such as CXCL12. Evidence for the direct participation of macrophages in the MaSC niche should eventually be possible through the prospective identification of a highly specific MaSC marker. The finding that the colonic niche was located at the base of the crypts allowed the localization of macrophages to this region, while electron microscopy identified a direct physical relation between colonic epithelial progenitor cells and macrophages 13.

Our data demonstrate that depletion of mammary gland macrophages either by chemical ablation or the use of Csf-1^op/op mutant mice alters mammary stem/progenitor cell activity, reflected in a substantially reduced repopulating frequency. The diminished outgrowth potential indicates a continued requirement for normal macrophages during ductal morphogenesis and may be mediated by effects on the putative basal progenitor cell. Overall, these findings indicate that macrophages play a role in supporting normal mammary stem cell function in vivo and suggest that they may constitute part of the normal MaSC microenvironment.

CI: confidence interval; CSF1: colony stimulating factor 1; MaSC: mammary stem cell; RT-PCR: reverse transcription polymerase chain reaction; SEM: standard error of the mean; TEB: terminal endbud.

The authors declare that they have no competing interests.

DG contributed to the conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing. MA contributed to the collection and assembly of data. NvR provided the liposomal reagents. GJL contributed to the study conception and design, provision of study materials, data analysis and interpretation. JEV contributed to the conception and design, provision of study materials, data analysis and interpretation and manuscript writing. All authors read and approved the final manuscript.