Stem Cell

Mcam inhibits macrophage-mediated development of mammary gland through non-canonical Wnt signaling – Nature Communications

Time Stamp: January 1, 2024 7:00 PM
Source Node: 440161

Mcam as a potential regulator of MECs by GeCKO screening

To identify potential regulatory factors of MECs, we performed clone formation assay²⁴ and transplantation assay^25,26 for screening using a pooled mouse GeCKO library containing 130 209 single guide RNAs (sgRNAs), targeting 20 611 protein-coding genes, and 1000 control sgRNAs (non-targeting sgRNAs) in the mouse genome²⁷. The mammary progenitor-enriched population was isolated using standard protocols²⁴ and transduced with a lentivirus prepared by the GeCKO library. Multiplicity of infection (MOI) was tested and applied to ensure that each cell contained a single sgRNA. The non-transfected cells were screened out by puromycin. After 10-14 days, we selected large clones (diameter ≫ 100 μm, enriched with mammary progenitor cells) for deep sequencing. The stably transfected cells were used for a 5-week transplantation assay, and the regenerated mammary tissues were collected for deep sequencing (Fig. 1a). Thus, based on the above methods, we performed gene enrichment of representative sgRNAs from plasmids, stably transfected cells, large clones, and regenerated mammary tissues.

**Fig. 1: Mcam is a negative regulator of MECs and mammary gland development.**

For sequencing, we first compared the overall distributions of sgRNAs from all samples to demonstrate the dynamics of the sgRNA library (Fig. S1a) and validate the sequencing quality. The detected number of sgRNAs in the plasmids was close to that in the original library, showing that most sgRNAs in the library were represented (Figs. S1b and S1c). The sgRNAs in the large clones and regenerated mammary tissues, retained less sgRNAs, reflecting that sgRNAs were enriched in these samples. The global patterns of sgRNA distribution in different samples were distinct, as evident by the strong shifts in respective cumulative distribution functions (Fig. S1d). In addition, we compiled a list of genes (Supplementary Dataset 1) showing enrichment of 1, 2, or 3 or more sgRNAs, along with the proportion of samples where these genes demonstrated enrichment (Fig. S1e). Based on MAGeCK²⁸ (http://bitbucket.org/liulab/mageck-vispr), we obtained a candidate gene list properly accounted for mammary gland morphogenesis and/or MEC clonogenicity and regenerative capacity, and found Mcam was one of the common genes in all conditions (Fig. S2a). By further analysis, we found multiple sgRNAs targeting Mcam are significantly enriched in more than half of the regenerated mammary tissue (Fig. S2b) with two or more sgRNAs targeting Mcam enriched in each sample (Fig. S2c).

The candidates were selected according to the expression profile based on single-cell sequencing data of MECs²⁹, and sgRNAs ranked in each sample. The single-cell sequencing data showed that Mcam was highly expressed in basal cells (Fig. 1b). To validate this result, we collected basal (Lin⁻CD24⁺CD29⁺) and luminal (Lin⁻CD24⁺CD29⁻) cell populations using fluorescence-activated cell sorting (FACS) (Fig. S3a, b), with quantitative real-time polymerase chain reaction (qRT-PCR) confirming the high expression of Mcam in the basal cell population (Fig. 1c). Immunofluorescence (IF) staining showed that Mcam was co-localized with K14 (basal cell marker) but not with K18 (luminal cell marker) (Fig. 1d). Taken together, Mcam was identified as a potential key regulator of MECs.

Loss of Mcam enhances mammary gland development

To further investigate the role of Mcam in regulating MEC proliferation and mammary gland development in vivo, we firstly examined Mcam expression pattern in four stages of mammary gland development, and found Mcam highly expressed in puberty (4-5 weeks), pregnancy, and lactation (Fig. 1e). Then we bred Mcam flox mice³⁰ with MMTV-Cre mice to generated Mcam^fl/fl/MMTV-Cre mice (cKO mice)³¹ and K14-Cre mice to generate Mcam^fl/wt/K14-Cre mice (heterozygous cKO mice) with specific loss of Mcam in MECs or basal cells. Mcam was specifically depleted in MECs or basal cells based on qRT-PCR (Figs. 1f and S3c), IF staining (Figs. 1g and S3d), and western blot analysis (Figs. 1h and S3e) of whole mammary tissues. We next explored the mammary duct phenotype. Loss of Mcam significantly lengthened mammary ducts at the puberty stage (5-weeks old), and therefore increased mammary duct occupation in the fat pad (Figs. 1i, j and S3f, g). In addition to ductal length, mammary duct and terminal end bud (TEB) thickness increased significantly (Figs. 1k, l and S3h–j). The above effects were not observed in adult mice (8-weeks old) (Fig. S3k), which could be due to rare expression level of Mcam at this period (Fig. 1e).

These results showed that Mcam loss enhances mammary gland development, and this phenotype is likely due to increased proliferation and differentiation of the mammary epithelium. Based on IF staining with Ki67 (cell proliferation marker), both basal and luminal cells demonstrated a much higher percentage of Ki67-positive cells upon loss of Mcam (Figs. 2a, b and S4a–e), which is also supported by the enrichment of DNA replication pathway in cKO mice based on RNA-seq data (Fig. S4f). All cell differentiation markers, including Stat5, Elf5, Foxa1, Esr1, and Gata3³², were increased at the mRNA level (Fig. S4g). Immunohistochemistry (IHC) staining showed PR^– and PR⁺ cells occurred in cKO mice by MMTV-Cre, implying Mcam loss promoting cell differentiation (Fig. S4h). The induction of both cell proliferation and differentiation suggests that loss of Mcam enhances mammary duct morphogenesis.

**Fig. 2: Loss of Mcam promotes MEC regenerative capacity.**

Loss of Mcam promotes the regeneration of mammary gland

To investigate whether the above phenotype is driven by MaSCs, we explored changes in the MaSC-enriched cell population (Lin⁻CD24⁺CD29⁺) in cKO mice by MMTV-Cre. Compared to wild-type (WT) mice, the cKO mice exhibited a three-fold higher basal population (Figs. 2c, d and S5a, b), indicating that Mcam loss induces the basal cell compartment containing MaSCs. In addition, we examined the expression of stemness-related genes including Procr, K14, Lgr5, and sSHIP in basal cells derived from the mammary tissues of WT and cKO mice, and found these genes highly expressed in Mcam cKO mice (Fig. S5c), supporting the potential regulatory roles of Mcam for MaSCs.

To determine whether Mcam negatively regulates MaSC function in vivo, we carried out limiting-dilution transplantation assays in non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice using isolated Lin^– epithelial cells and basal population, primary MECs, respectively. All isolated Lin^– epithelial cells and basal cells (regenerated mammary tissues were harvested at 6 weeks after transplantation), primary MECs (regenerated mammary tissues were harvested at 5 weeks after transplantation) from cKO mice displayed a significantly higher rate of successful engraftment and more extensive mammary outgrowth than the WT cells (Figs. 2e, f and S5d). To illustrate whether Mcam directly regulates the MEC clonogenicity, we performed an in vitro clone assay using primary MECs (Figs. 2g–i and S5e–g). Mcam depletion significantly increased the numbers and diameters of colonies, representing MEC clonogenicity and proliferation ability. Moreover, Mcam was knocked down using a lentivirus (Fig. S5h) to explore its function in primary cultured MECs in vivo and in vitro. In the transplantation assay, knockdown of Mcam significantly improved the regenerative capacity of MECs to reconstruct whole mammary tissue (Fig. S5i). In the clone assay, both clone numbers and diameters increased (Fig. S5j–m), suggesting loss of Mcam promotes MEC proliferation. These results provide strong evidences for the Mcam roles in inhibiting MEC proliferation and mammary reconstruction.

Knockout of Mcam promotes macrophage recruitment

To identify the regulatory mechanism underlying the promotion function of MEC proliferation and mammary gland development in cKO mice by MMTV-Cre, RNA-sequencing (RNA-seq) was performed on mammary tissues from 5-week-old WT and cKO mice. First, we analyzed the differentially expressed genes (DEGs) identified between WT and cKO mammary tissues (total breast cells) (Fig. S6a) and performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of upregulated DEGs in cKO mammary tissues. Results indicated strong activation of pathways associated with immune regulation (Fig. 3a). Notably, the DEGs were significantly enriched in the GO categories “immune system process” and “immune response” (Supplementary Dataset 2) and in the KEGG pathways “activation of innate immune response”, “regulation of immune effector process”, and “adaptive immune response” (Fig. 3a). Immune cells are key components of the mammary microenvironment³³, therefore, we analyzed the abundance of various immune cells in WT and cKO mammary tissues using FACS.

**Fig. 3: Loss of Mcam promotes macrophage recruitment.**

Data showed a significant increase in the abundances of macrophages and neutrophils in cKO mice, but not of dendritic cells (DC), B cells, or T cells (Figs. 3b and S6b). Furthermore, using macrophage markers (F4/80 and CD11b), macrophages showed consistent and strong elevation in the Mcam cKO mice (Figs. 3c and S6c), implying that macrophages play important roles in Mcam cKO mice. In addition to stromal macrophages, tissue-resident ductal macrophages are also found in the mammary gland epithelium³⁴. Thus, we explored changes in both types of macrophages in the mammary gland using CD206 (stromal macrophage marker)³⁵ and Cx3cr1 (tissue-resident macrophage marker)³⁴. Data showed that both cell populations were significantly elevated (Figs. 3d–g and S7a–d). To further validate macrophages indeed were recruited by MECs, isolated primary macrophages and macrophage cell line RAW264.7 were co-cultured with conditional medium (CM) from either WT or cKO MECs by MMTV-Cre. Compared with WT-CM treatment, cKO-CM induced more macrophages migration (Fig. 3h, i and S7e, f). Thus, depletion of Mcam in the mammary tissue leads to an increase in macrophages.

Depletion of macrophages blocks the elevating activity of MECs upon Mcam loss

As macrophages can maintain MaSC stemness³⁶, we mixed macrophages with basal cells from WT or cKO mice by MMTV-Cre to implement a co-culture mammosphere assay for confirming Mcam roles on the clonogenicity promotion by macrophages. After co-culturing with isolated primary macrophages for 10-14 days, clone numbers and clone diameters increased significantly in the basal cells of cKO mice (Fig. 4a–c). To confirm the mediating roles of macrophages in the above function, we performed rescue assays using clodronate liposome (CL) to specifically deplete macrophages in both WT and cKO mice³⁷. After injection of CL, the percentages of macrophages in WT and cKO mice were effectively reduced, as shown by FACS analysis (Figs. 4d and S8a), as was the percentages in the spleen (Fig. S8b, c), a macrophage-enriched organ.

**Fig. 4: Depletion of macrophages blocks the elevating clonogenicity and regenerative capacity of MECs upon Mcam loss.**

We next determined the abundance of CD206⁺ macrophages in the mammary tissues after CL treatment and found that the elevating CD206⁺ signals upon Mcam loss were diminished completely (Fig. 4e, f). Based on whole-mount staining, while Mcam cKO promoted mammary duct elongation within the fat pad (Fig. 1i), CL treatment eliminated these promotion effects (Figs. 4g, h and S8d). To further investigate the inhibitory effects of mammary morphogenesis after CL treatment, we determined the proliferation of MECs, from the IF (Fig. 4i, j) and IHC (Fig. S8e) staining results, both the elevating proliferation in basal and luminal cells by Mcam loss were flattened significantly after CL treatment, further indicating that amplification of both MEC lineages after Mcam loss was dependent on macrophages.

We next isolated and transplanted MECs from WT and cKO mice by MMTV-Cre into NOD/SCID mice to examine the repopulating ability of MECs. The MECs were mixed with CL to remove migrated macrophages, as performed in previous research³⁶. While loss of Mcam increased the frequencies of MEC reconstitution (Fig. 2), we observed nearly complete inhibition of reconstitution in the fat pad when MECs were injected together with CL (Fig. 4k), suggesting that the promotion of MEC regenerative capacity upon Mcam cKO depends on the presence of macrophages. To further investigate the inhibitory effects of MaSCs after CL treatment, we determined the basal cell population percentages using FACS and found that systemic ablation of macrophages reversed the increase in Lin^–CD24⁺CD29⁺ cell abundance in cKO mice (Figs. 4l–m and S8f), suggesting that CL treatment rescued MaSC proliferation by Mcam loss. Taken together, these findings suggest that Mcam roles on clonogenicity, regeneration, and ductal morphogenesis of mammary glands are dependent on macrophages.

Mcam recruits and activates macrophages through Il4-Stat6 axis

We next examined the molecular mechanism underpinning communications between macrophages and MECs. As macrophages were enriched by Mcam cKO (Fig. 3), we explored how the increased macrophages was recruited. Based on the RNA-seq DEGs (Fig. S6a), cytokines and chemokines increased in the cKO mammary tissues (Fig. 5a). The qRT-PCR results demonstrated that Il4 was the most significantly elevated factors upon Mcam KO by both MMTV-Cre and K14-Cre (Figs. 5b and S9a). Il4 and Il13, which are secreted by MECs³⁸, and downstream Stat6 are essential for promoting macrophage polarization and activation^39,40. Here, we found that Stat6 phosphorylation levels increased significantly in cKO mice (Figs. 5c and S9b), suggesting that Il4-Stat6 may mediate the role of Mcam in macrophages. To further validate this result, we blocked Il4 by its antibody in vitro and in vivo to examine macrophage migration. Blocking Il4 in vitro reversed the elevated macrophage migration upon Mcam loss (Fig. 5d, e). After injecting with Il4 blocking antibody orthotopically, we found the same rescue effects for CD206⁺ and Cx3cr1⁺ macrophages abundance (Fig. 5f–i), which leads to the inhibition of ductal elongation in Mcam cKO mice by MMTV-Cre ultimately (Fig. 5j, k). These data demonstrate Mcam roles for recruiting macrophages through Il4-Stat6 axis.

**Fig. 5: Mcam loss recruits macrophages through Il4-Stat6 axis.**

Macrophages regulate MECs through non-canonical Wnt signaling upon Mcam loss

We also investigated the mechanism underlying the role of macrophages within mammary tissues. Previous reports have shown that macrophages play critical roles in MaSCs³⁶. Based on gene set enrichment analysis (GSEA) of RNA-seq data, the gene cluster, canonical Wnt signaling pathway, showed no significant difference between WT and Mcam cKO mice by MMTV-Cre (Fig. S9c). Upon Mcam KO, all canonical Wnt ligands, including Wnt1, Wnt3, Wnt3a, Wnt10a, and Wnt16, were undetectable by qRT-PCR. For the downstream factors of canonical Wnt signaling, representative effectors, including Axin2, GSK-3β, and β-catenin, also showed no significant differences between WT and Mcam cKO mice (Fig. 6a). These data indicate that the promotion of MECs induced by Mcam deletion was not mediated by the canonical Wnt signaling pathway.

**Fig. 6: Macrophages regulate MECs through Wnt5a upon Mcam loss.**

For non-canonical Wnt ligands, including Wnt5a, Wnt5b, and Wnt11, the level of Wnt5a was remarkable higher in the Mcam cKO group than in the WT group (Figs. 6a). This elevation of Wnt5a was verified at the protein level by IHC staining (Figs. 6b and S9d). Published data²¹ have shown that Wnt5a is expressed in both basal and luminal epithelial cells. To confirm the Wnt5a surge in Mcam cKO mice is derived from macrophages, we isolated various cell populations in mammary tissues for measuring Wnt5a levels and taking Wnt5a expression in adipocyte of WT as a control, qRT-PCR showed macrophages (CD45⁺F4/80⁺) induced nearly 400-fold Wnt5a levels in cKO mice, and stromal cells containing macrophages also showed induced Wnt5a levels in cKO mice (Figs. 6c and S9e). In addition, we depleted Wnt5a in macrophages (Fig. S9f) and transplanted these cells with basal cells to examine whether the absence of Wnt5a in macrophages could still enhance mammary development. While Mcam cKO promotes ductal elongation, loss of Wnt5a by shRNA in macrophages did not have the capacity to reconstruct whole mammary tissue (Fig. 6d), confirming Wnt5a is responsible for the phenotype observed.

We next tested whether Wnt5a mediates the maintenance of MEC clonogenicity and proliferation by adding Wnt5a blocking antibody to the culture medium in the coculture assay between cKO MECs and isolated macrophages. Results showed that in the absence of Wnt5a, Mcam loss failed to increase both clone numbers and clone diameters (Fig. 6e–g), indicating that Wnt5a mediates clonogenicity and proliferation promotion upon Mcam loss. To test the linkage between Il4 and Wnt5a, we blocked Il4 to prove that elevated Wnt5a expression levels are induced by Il4 stimulation. The results found the elevated Wnt5a levels were rescued by Il4 antibody injections (Fig. 6h, i), suggesting increased Il4 upon Mcam loss indeed activates Wnt5a expression.

Loss of Mcam promotes MEC proliferation via Ryk

Next, we investigated how Wnt5a interacts with MECs. Wnt5a binds to two different receptors, i.e., Ror2 to inhibit branching morphogenesis and Ryk to promote MaSC/progenitor proliferation²³. Thus, Ryk may be the potential receptor for our phenotype. Here, we explored the expression pattern of Ryk at the single-cell level using MEC single-cell RNA-seq from our and other laboratories. Based on the online data²⁹, Ryk was highly expressed in both basal and luminal cells, especially basal cells (Fig. 7a). Using a single-cell RNA-seq dataset generated by our laboratory⁴¹, we found the same Ryk expression pattern (Fig. S10a). Interestingly, we analyzed the correlation between Mcam and Ryk, and found that Mcam and Ryk were co-expressed in many basal cells (Fig. 7b). This result was validated at the protein level using IF staining (Fig. 7c).

We next explored the expression levels of the two receptors using qRT-PCR. Results showed that upon Mcam loss Ror2 was decreased and Ryk was elevated (Fig. 7d), especially in basal cells (Fig. 7e), consistent with the promoting role of Ryk in MaSCs found in previous study²³. The increase in Ryk expression upon Mcam cKO was confirmed by WB or IHC (Figs. 7f and S10b). Moreover, Wnt5a blocking resulted in a decrease of Ryk expression, whereas exogenous recombinant Wnt5a (rWnt5a) increased Ryk expression (Fig. S10c, d), which is consistent with the conclusion that Ryk is correlated with the levels of its ligand – Wnt5a^42,43,44. These results imply that the phenotypes observed in the current study is mediated by the Ryk receptor. To further test this hypothesis, we blocked Ryk using its antibody and performed a rescue assay. Results showed that both the increased clone numbers and clone diameters caused by Mcam loss were suppressed to a level similar to that of WT after blocking Ryk (Fig. 7g–i), suggesting that the Mcam functions on MECs is dependent on Ryk. Taken together, loss of Mcam leads to an increase of Wnt5a, which binds to the Ryk receptor, thereby contributing to the clonogenicity, regenerative capacity, and proliferation of MECs.