Slug and Sox9 cooperatively determine the mammary stem cell state

Abstract

Regulatory networks orchestrated by key transcription factors (TFs) have been proposed to play a central role in the determination of stem cell states. However, the master transcriptional regulators of adult stem cells are poorly understood. We have identified two TFs, Slug and Sox9, that act cooperatively to determine the mammary stem cell (MaSC) state. Inhibition of either Slug or Sox9 blocks MaSC activity in primary mammary epithelial cells. Conversely, transient coexpression of exogenous Slug and Sox9 suffices to convert differentiated luminal cells into MaSCs with long-term mammary gland-reconstituting ability. Slug and Sox9 induce MaSCs by activating distinct autoregulatory gene expression programs. We also show that coexpression of Slug and Sox9 promotes the tumorigenic and metastasis-seeding abilities of human breast cancer cells and is associated with poor patient survival, providing direct evidence that human breast cancer stem cells are controlled by key regulators similar to those operating in normal murine MaSCs.

Figure 1. Slug is the major EMT-TF that is expressed in MaSCs

(A) The mRNA levels of the EMT-TFs in MaSC-enriched basal (stem/basal) or luminal progenitor cells were compared to those of differentiated luminal cells (diff. luminal) by qRT-PCR in triplicate. GAPDH was used as a loading control. (B) Slug protein expression in the mammary gland was analyzed by immunofluorescence on tissue sections. An anti-keratin 8 antibody was used to label the luminal cells. (C) The expression of the Slug-YFP reporter in MaSC-enriched basal (basal/stem), luminal progenitor and differentiated luminal (diff. luminal) cells. (D) The gland-reconstituting activities of Slug-YFP⁺ and Slug-YFP⁻ MECs were measured by the limiting dilution analysis. Representative whole-mount images of carmine-stained fat pads (upper panel) and reconstitution efficiencies (lower panel) are shown. In the lower panel, each circle represents one transplanted fat pad, and the dark area of each circle represents the percentage of the fat pad that is occupied by reconstituted mammary ductal trees. P=1.6X10⁻⁵. The data are represented as mean ± standard error of the mean (SEM). See also Figure S1.

Figure 2. The ectopic expression of Slug induces MaSC activity

(A) The expression levels of EMT-associated proteins were determined by immunoblot. The primary MECs transduced with tetracycline-inducible Slug lentivirus were treated with the indicated concentration of doxycycline for 5 days. (B) The organoid-forming efficiencies of the primary MECs transduced with the indicated vectors. (C) The gland-reconstituting activity was measured using the competitive reconstitution assay. The left panel shows representative whole-mount fluorescence images of the mammary fat pads at the indicated time points post-injection. Right panel shows the ratios of GFP-expressing (either control vector or Slug) to dsRed-expressing cells, as measured by flow cytometry. (D, E) The organoid-forming efficiencies of MaSC-enriched basal cells (D) or luminal progenitor cells (E) that were transduced with the indicated vectors. The data are represented as mean ± SEM. See also Figure S2.

Figure 3. The cooperation of Sox9 with Slug in the formation of MaSCs

(A) The screening for Slug cofactor(s) that are involved in the induction of organoid-forming cells. Differentiated luminal cells transduced with the indicated vectors were treated with doxycycline for 5 days in monolayer culture and then subjected to organoid culture without further doxycycline treatment. (B) The organoid-forming efficiencies of differentiated luminal cells transduced with the indicated vectors and then treated as in (A). (C) The gland-reconstituting activity of differentiated luminal cells transduced with the indicated vectors. The fat pads were analyzed 7 weeks post-injection by whole-mount analysis (top panel) and flow cytometry (middle panel). The relative MaSC activity was quantified as the ratio of GFP- to dsRed-positive cells (lower panel). The data are representative of three independent experiments. (D) The mammary gland outgrowths generated by GFP-expressing differentiated luminal cells transduced with the indicated vectors. The cells were treated with doxycycline for 8 days in monolayer culture and then transplanted into cleared mammary fat pads at limiting dilutions. The fat pads were examined 3 months post-implantation by whole-mount imaging (left and middle panels) or immunofluorescence on tissue sections (right panel). (E) Secondary transplantation generated by Slug/Sox9-exposed differentiated luminal cells. The mice were mated 4 weeks post-transplantation. The mammary fat pads around gestation day 18 were then analyzed by whole-mount fluorescent imaging (left) or immunofluorescence on tissue sections (right). The data are represented as mean ± SEM. See also Figure S3.

Figure 4. The induction of MaSCs by Sox9 in basal cells

(A) The organoid-forming efficiencies of basal cells transduced with the indicated vectors. (B) The solid organoid- and acinus-forming efficiencies of basal cells transduced with the indicated cDNA and shRNA expression vectors. Cells were subjected to organoid culture 5 days post-infection. The shLuciferase (shLuc) shRNA was used as a control. (C) A model showing the mammary epithelial hierarchy and the actions of the forced expression of Slug and Sox9 in various mammary epithelial lineages. The dashed lines indicate that the expression of the indicated factor(s) converts differentiated cells into stem or progenitor cells. Whether Sox9 expression converts differentiated myoepithelial cells or myoepithelial progenitors into MaSCs and/or expands a pre-existing MaSC population remains to be determined. The data are represented as mean ± SEM.

Figure 5. Slug and Sox9 are required for maintaining endogenous MaSCs

(A) Confocal immunofluorescence analyses of mammary gland sections stained with rabbit anti-Slug and goat anti-Sox9 antibodies. The arrows point to Slug and Sox9 double-positive nuclei. (B) The organoid-forming efficiencies of primary MECs transduced with the indicated shRNA vectors. Cells were subjected to organoid culture 4 days post-infection. (C) The gland-reconstituting activity of primary MECs transduced with the indicated shRNA vectors. The shRNA vector-transduced and GFP-expressing primary MECs (1×10⁵) were mixed with an equal number of dsRed-expressing MECs and then transplanted into cleared mammary fat pads. The ratio of GFP- to dsRed-positive cells was normalized against that of the shLuc control to obtain the relative reconstitution efficiency. The data are represented as mean ± SEM. See also Figure S4.

Figure 6. Slug and Sox9 activate distinct auto-regulatory gene expression programs

(A) Phase-contrast and immunofluorescence images of differentiated luminal cells expressing the indicated vectors for 4 (phase-contrast) or 5 (immunofluorescence) days. (B and C) The mRNA levels of basal cell TFs (B) and luminal progenitor genes (C) in differentiated luminal cells expressing the indicated vectors for 5 days, as measured by qRT-PCR. GAPDH was used as a loading control. (D) The mRNA levels of various signature genes in tetracycline-inducible Slug and Sox9-transduced differentiated luminal cells after a 6-day doxycycline treatment (Slug/Sox9 on dox) or a 6-day doxycycline treatment plus a 6-day doxycycline withdrawal (Slug/Sox9 dox withdrawal). The mRNA levels were normalized to those of the control vector-transduced cells after a 6-day doxycycline treatment. Primers for amplifying protein-coding sequences were used to detect the total mRNA levels of Slug and Sox9 (total); and primers for amplifying the 5’UTRs were used to detect the endogenously expressed Slug and Sox9 mRNA (endo). (E) The organoid-forming efficiencies of differentiated luminal cells transduced with the indicated vectors after a 5-day doxycycline treatment (on dox) or a 5-day treatment plus a 6-day withdrawal (dox withdrawal) in monolayer culture. The data are represented as mean ± SEM. See also Figure S5.

Figure 7. Slug and Sox9 act as regulators of breast CSCs

(A) The tumor weight and incidence of MDA-MB-231 cells expressing the indicated shRNAs. Cells were injected subcutaneously at the indicated numbers. The tumor weight and incidence were determined 3 months post-injection. Each data point represents one tumor. The mean and SEM of each group are represented by horizontal and vertical bars. The table shows the tumor incidence. (B) Lung metastases formed by MDA-MB-231 cells expressing the indicated shRNA vectors upon tail vein injection. (C) Lung metastases formed by tdTomato-labeled MCF7ras cells that were transduced with the indicated vectors and injected orthotopically into the mammary fat pads. Whole-mount fluorescence lung images and the histology of the lung sections are shown. n=4 for each group. The data are representative of two independent experiments. (D) The cumulative survival rate of human breast cancer patients with primary tumors expressing high levels of both Slug and Sox9 (Slug/Sox9-high) or tumors expressing either only one factor or neither factors at high levels (Non-Slug/Sox9-high). The data are presented as mean ± SEM. See also Figure S6.