Volume 3, Issue 1, Pages 70-78 (January 2013) Lgr5-Expressing Cells Are Sufficient and Necessary for Postnatal Mammary Gland Organogenesis  Vicki Plaks, Audrey Brenot, Devon A. Lawson, Jelena R. Linnemann, Eline C. Van Kappel, Karren C. Wong, Frederic de Sauvage, Ophir D. Klein, Zena Werb  Cell Reports  Volume 3, Issue 1, Pages 70-78 (January 2013) DOI: 10.1016/j.celrep.2012.12.017 Copyright © 2013 The Authors Terms and Conditions

Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure 1 Lgr5 Expression Is Restricted to a Rare Subpopulation of Cytokeratin 14+, Lin−CD24+CD49fhigh Mammary Basal Cells (A) The expression of Lgr5 was examined in cryosections from 7-week-old Lgr5-EGFP MGs with an anti-GFP antibody (green). Carmine stain of a representative MG whole mount demonstrates that Lgr5+ ducts are located to the nipple area, but not to the invading front. Around the lymph node (LN), there are some positive and negative ducts. (B) Cryosections costained with anti-GFP and anti-K14. Lgr5+ cells (green) are located to the suprabasal layer of the ducts and are a subpopulation of the myoepithelial K14+ cells (red). (C) MGs were isolated from Lgr5-EGFP mice and analyzed by flow cytometry for the expression of the cell surface markers Ter119, CD45, CD31 (Lin), CD24, and CD49f. Lgr5+ cells (GFP+) were part of the Lin−CD24+CD49fhigh cells (stem cell-enriched population). Lgr5+ cells are 0.3% of total mammary cells and 2.5% of Lin−CD24+CD49fhighbasal cells. GFP+ cells within the luminal population are 0.009% of total. (D) Summary of flow cytometry data in Figure 1C, Lgr5+ cells in 7.5-week-old pubertal female mice, percentage (%) of Lgr5+ cells of total (n = 14), and of Lin−CD24+CD49fhighbasal cells (n = 7). See also Figure S1. (E) Real-time, quantitative PCR analysis of the Lgr5+ cell population (relative to Lgr5− mammary cells) revealed that they are high for basal but not luminal markers. PR, progesterone receptor; ERα, estrogen receptor α. See also Table S1. Bars represent SE. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure 2 Within the Lin−CD24+CD49fhigh Basal Population, Lgr5+ Cells Are Highly Potent in Generating Functional Mammary Outgrowths (A) Lgr5+ (GFP+) and nonexpressing (GFP−) cells from Lgr5-EGFP were isolated by flow cytometry from the Lin−CD24+CD49fhigh basal population and injected (10, 50, or 100 cells) into cleared mammary fat pads. Outgrowths were analyzed 6 weeks posttransplantation. (B) Transplanted basal Lgr5+ cells have higher numbers of outgrowths compared to the basal Lgr5− cells. Data are pooled from three different experiments. (C) Whole-mount carmine-stained representative outgrowths show that ten basal Lgr5+ cells are able to reconstitute a full MG versus no outgrowth for basal Lgr5-transplanted cells. (D) Mice transplanted with ten Lgr5+ cells were mated with males, and their MGs were analyzed on day 18.5 (E18.5) of pregnancy. (E) Whole-mount carmine-stained mammary epithelial outgrowths from E18.5 pregnant female mice transplanted with ten basal Lgr5+ cells that underwent full lobuloalveolar differentiation (basal Lgr5+), comparable to the endogenous epithelium in MG #3 of the recipient mouse (upper panels). MG sections from the same mice stained positive for the milk protein, β-casein (lower panels; brown). See also Figure S2 and Table S2. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure 3 Lgr5+ Cells Can Regenerate a MG from a Single Cell and Maintain Regenerative Potential through Serial Transplantations (A) Single mammary Lgr5+ (GFP+) cells from Lgr5-EGFP crossed into the LifeAct-RFP mice were isolated by flow cytometry into 96-well plates and transplanted into cleared mammary fat pads. Outgrowths were analyzed at 8 weeks posttransplantation. (B) From transplants of single adult mammary Lgr5+ cells in 54 MGs, 13 mammary outgrowths were observed. (C) A representative RFP+ mammary outgrowth from a single Lgr5+ cell, exhibiting a full epithelial tree (left) with ductal structures at higher magnification of boxed area (right). (D) Outgrowths from single Lgr5+ cells differentiate into the myoepithelial (K14+ in red) and luminal (K8+ in green) lineages (left). Boxed area magnified (right). See also Figure S3. (E) Mammary outgrowths from two mice transplanted with 100 Lgr5+ cells (isolated from Lgr5-EGFP crossed into the LifeAct-RFP mice) were collected and retransplanted into ten mice each for secondary and the same for tertiary outgrowths. (F and G) Lgr5+ outgrowths retain their regenerative potential through secondary (F) and tertiary (G) transplants. RFP images are representative of the mammary outgrowths. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure 4 Depletion Experiments Demonstrate that Lgr5+ Cells Are Necessary for MG Epithelial Reconstitution (A) Depletion of Lgr5+ cells was achieved utilizing Lgr5-DTR:GFP crossed into actin-RFP mice, injected with 50 ng/g BW DTx, analyzed 24 hr post-DTx i.p. (Lgr5+ cells are 0.1% of total dissociated mammary cells versus 0% in DTx-injected mice). (B) Isolated primary MECs of Lgr5-DTR:GFP mice or WT littermates transplanted into contralateral precleared mammary fat pads with or without DTx administration. MGs collected 3 weeks posttransplantation had significantly impaired outgrowths in the Lgr5-DTR:GFP transplants versus the WT controls. (C) To assess the growth potential of the Lgr5-DTR:GFP and control littermate, mice transplanted with the same cells as in (B) but not treated with DTx reveal no difference between the two contralateral sides. (D) Outgrowth area for Lgr5-DTR:GFP epithelial transplants (including impaired ducts) relative to the contralateral WT transplants is significantly reduced in DTx-treated mice (∗p = 0.006). Bars represent SE. See also Figure S4. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure 5 Depletion of Lgr5+ Cells during Pubertal Development Results in Impaired Ductal Invasion and TEB Formation (A) Carmine-stained MG of 4.5-week-old Lgr5-DTR:GFP mice (n = 6) or WT littermates (n = 4) that were i.p. injected with DTx demonstrates significantly reduced ductal invasion in the Lgr5-DTR:GFP mice. (B) Quantification of data presented in (A). ∗p = 0.004. (C) Depletion of Lgr5+ cells from Lgr5-DTR:GFP mice resulted in significant reduction in the number of TEBs per MG versus WT littermates. Arrows indicate the TEBs. (D) Quantification of data presented in (C). ∗p = 0.0015. (E) Whole mounts of 5-week-old Lgr5-EGFP-IRES-creERT2/Rosa-Tomato mice 1 week past start of Tamoxifen (TAM) induction indicated that Lgr5+ cell progeny is close to the nipple area (left) and, according to its localization and shape, mark myoepithelial cells (middle, enlargement of red boxed area in left) and not TEBs in the invading front (carmine-stained tissue on right). Bars represent SE. See also Figure S5. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure S1 Localization of Lgr5+ Cells in Older Mammary Glands, Related to Figure 1 Flow cytometry of mammary glands (MGs) from 27-week-old mice, shows that the Lgr5+ cells were still mostly located within the Lin-CD24+CD49fhigh basal cell population. Within the basal cells, Lgr5+ cells are restricted to a CD24highCD49fhigh population (Lgr5+ cells are 1.6% of total mammary cells and 16.3% of Lin-CD24+CD49fhigh basal cells; GFP+ cells are 0.003% of total within the luminal population). Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure S2 Single-Cell PCR Analysis on Lgr5+ Cells, Related to Figure 2 Single basal Lgr5+ cells versus Lin-CD24+CD49fhigh basal Lgr5- cells were characterized to elucidate whether the different functional mammary reconstitution abilities of the two subsets are based on differences in gene expression, similar to the rationale of analyzing intestinal Lgr5+ cell (Muñoz et al., 2012). Therefore, the single cell expression analysis explored only a few genes and focused on markers related to stem/progenitor capacity, rather than on signaling molecules. Single cell PCR analysis was performed on cells sorted using the flow cytometry methodology described in Figures 2A and 3A. + or - indicate GFP expression within the basal population, as sorted by flow cytometry. (A) Heatmap of unsupervised clustering of the expression data of the 29 genes tested [markers of mammary lineage differentiation, stem cell and signaling pathways (Wnt, Notch), pluripotency and epithelial-to-mesenchymal transition] reveals that within the basal population, sorted basal Lgr5+ cells (n = 8/14), cluster together versus other basal Lgr5- cells (n = 10/11). Cells were considered Lgr5+ based on higher GFP expression within the basal population. Orange box demarcates the clustering of Lgr5+cells. The color scale represents the centered and scaled log transformed expression values, and gray represents failed reactions. The heatmap is a representative of 3 experiments; each tested 25-60 single cells. (B) All single basal Lgr5- cells versus basal Lgr5+ cells represented in a bar graph according to their relative expression levels reveal that all cells sorted as low for GFP within the basal population display relatively low Lgr5 expression levels. Cells that were sorted as high for GFP show higher levels of Lgr5 expression in most cells (10/14) but also contain cells with expression levels similar to the Lgr5- cells (4/14). This could also explain why not all Lgr5+ sorted cells were able to generate mammary outgrowths, as described in Figures 2B and 3B. The bar graph is based on the data presented in (A). (C) Analysis of the Delta Ct average fold change values relative to the basal Lgr5-cells, summarizing data from the 3 experiments, reveals that the Lgr5+ cells are overall higher for basal markers (red bars- K5, p63) and lower for luminal markers (blue bars- K8, K18, ERα, Muc1, β-casein) within the basal population. These results summarize data from 3 experiments. The genes that exhibited the biggest and significant changes across 3 experiments were p63 and β-casein (p = 0.01). Interestingly, the basal marker p63, which is also necessary for the normal development of human mammary glands, was found to be a positive prognostic factor in endocrine-treated ER positive breast cancer (Hanker et al., 2010), thereby indicating that Lgr5+ cells can serve as cancer initiating cells. The observation that β-casein is downregulated supports the notion that the Lgr5+ cells are more stem-like and therefore, less differentiated than the other basal cells. Bars represent standard error. See also Table S2 for primer sequences. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure S3 Further Examples of Outgrowths from Single Lgr5+ Cell Transplants, Related to Figure 3 Additional examples of mammary ductal outgrowths from mammary transplants of single Lgr5+ cells isolated from LifeAct-RFP+ mammary glands and transplanted into cleared fat pads from unlabeled female nude mice. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure S4 Effects of Diphtheria Toxin-Mediated Depletion of Lgr5+ Cells on Mammary Outgrowths and In Vitro Sphere Formation, Related to Figure 4 (A) Outgrowths from additional recipient mice to those presented in Figure 4B, transplanted with epithelial cells from Lgr5-DTR:GFP x actin-RFP or actin-RFP littermate into contralateral pre-cleared mammary fat pads with DTx administration. Lower panels are higher magnifications of the boxed areas in the corresponding upper panels. These data demonstrate runted and fewer ducts in outgrowths of mice administered with DTx. The 60% decrease in ductal area shown in Figure 4D for the Lgr5-DTR:GFP transplants treated with DTx includes these impaired structures, although they do not form proper ductal structures. (B) Carmine stains of the PBS-or DTx-injected recipient non-transplanted MGs #3 of 8-week-old nude mice exhibit normal ductal structures, indicating that DTx is not toxic to endogenous WT mature mammary epithelial tissue. (C) In vitro mammary sphere-forming assay using 104 primary mammary epithelial cells from Lgr5-DTR:GFP mice (the same number of cells used for the in vivo depletion experiments) show that cells cultured in the presence of DTx are unable to form spheres, as analyzed on day 8 of culture. Data are summarized from 2 independent experiments each with cells from 4 mice and 2-4 wells per condition. Bars represent SE (∗p = 0.003). Representative images of spheres are included. Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions

Figure S5 Detection of Lgr5+ Cell Progeny in Tamoxifen-Induced Lgr5-EGFP-IRES-creERT2/Rosa-Tomato MGs, Related to Figure 5 (A) By whole mount fluorescence analysis, Lgr5+ cell progeny are localized to myoepithelial cells at 2 and 3 weeks post the start of Tamoxifen (TAM) induction (red; images on right are enlargement of boxed are on left), according to their localization and shape. TAM induction was started at 4 weeks of age so that 2 and 3 weeks induction corresponded to 6 and 7 weeks of age, respectively. Note that over-exposed images on left confirm that TEBs are negative for red fluorescence. (B) MG #3 also demonstrates how Lgr5+cell progeny (marked area on left) is localized to a very restricted area of the entire ductal tree (marked area on right) and does not localize to TEBs. (C) Flow cytometry analysis of MGs from 5-week-old Lgr5-EGFP-IRES-creERT2/Rosa-Tomato mice, 1 week after the start of TAM induction. Note that most of the Tomato+ progeny still maintain their Lgr5 expression (lower left). When Lgr5+ progeny is backgated on mammary subsets (upper right), it is found to localize exclusively to the basal population, which includes myoepithelial cells (lower right plot). Cell Reports 2013 3, 70-78DOI: (10.1016/j.celrep.2012.12.017) Copyright © 2013 The Authors Terms and Conditions