Oestrogen receptor α AF-1 and AF-2 domains have cell population-specific functions in the mammary epithelium

Abstract

Oestrogen receptor α (ERα) is a transcription factor with ligand-independent and ligand-dependent activation functions (AF)-1 and -2. Oestrogens control postnatal mammary gland development acting on a subset of mammary epithelial cells (MECs), termed sensor cells, which are ERα-positive by immunohistochemistry (IHC) and secrete paracrine factors, which stimulate ERα-negative responder cells. Here we show that deletion of AF-1 or AF-2 blocks pubertal ductal growth and subsequent development because both are required for expression of essential paracrine mediators. Thirty percent of the luminal cells are ERα-negative by IHC but express Esr1 transcripts. This low level ERα expression through AF-2 is essential for cell expansion during puberty and growth-inhibitory during pregnancy. Cell-intrinsic ERα is not required for cell proliferation nor for secretory differentiation but controls transcript levels of cell motility and cell adhesion genes and a stem cell and epithelial mesenchymal transition (EMT) signature identifying ERα as a key regulator of mammary epithelial cell plasticity.

Fig. 1

Mammary gland phenotype of AF-1⁰ and AF-2⁰ mice. a Schematic representation of the mutant genomic loci and the ERα proteins expressed in WT mice, AF-1⁰ mice with deletion of amino acids 2–148, and AF-2⁰ mice with deletion of amino acids 543–549. Only the main protein initiated by the translational initiation codon in exon 1 on amino acid 1 (ATG) is shown for each genotype. The less abundantly expressed protein is initiated in exon 2 on amino acid 178. b Whole-mount stereomicrographs of inguinal mammary glands from 7-week-old WT, AF-1⁰ and AF-2⁰ females. Arrowheads indicate TEBs. Black lines mark the borders of the ductal outgrowth. Scale bars: 5 mm (left), 2 mm (right). LN sub-iliac lymph node. c Dot plot showing extent of fat pad area filled by the engrafted WT, AF-1⁰ and AF-2⁰ epithelia at different developmental stages (n = 3–18). Shown are means ± SEM; unpaired, two-tailed, Student’s t test, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, n.s. not significant. d ERα IHC of mammary glands from 3-week-old WT, AF-1⁰, AF-2⁰ and ERα^−/− mice. Representative pictures of glands analysed from three females of each genotype are shown. Scale bar: 100 μm

Fig. 2

Steroid hormone levels in ERα AF-1⁰, ERα AF-2⁰ and ERα^−/− mice. a Box plots showing plasma levels of 17-β-estradiol, testosterone, progesterone, androstenedione, 17-OH-progesterone, corticosterone and deoxycorticosterone determined by LC/MS measured in peripubertal (3-week old), pubertal (4- to 7-week old) and adult (>8-week old) WT, AF-1⁰, AF-2⁰ and ERα^−/− females (n = 3–25). Horizontal lines outside the boxes depict minimum and maximum values, upper and lower borders of the boxes represent lower and upper quartiles, and the line inside the box identifies the median. Unpaired, Mann–Whitney test. b Mean serum levels of the different steroids (shown in a) were plotted relative to average levels in WT littermates in the context of major biosynthetic pathways. Colour code represents fold changes over levels in the WT littermates

Fig. 3

Mammary epithelial-intrinsic role of the ERα AF-1 and AF-2 domains. a–d Fluorescence stereomicrographs of contralateral inguinal mammary fat pads engrafted with mammary epithelium from AF-1⁰.GFP⁺ or WT.GFP⁺ littermates. Nulliparous (a, b) and day 16–18 pregnant (c, d) recipients are shown. Scale bars; 5 mm (top), 2 mm (bottom). e Box plot showing extent of fat pad filling by the engrafted epithelia in virgin (n = 18) and pregnant (n = 5) recipients. f–j Fluorescence stereomicroscopy of contralateral inguinal mammary fat pads engrafted with mammary epithelium from AF-2⁰.GFP⁺ or WT.GFP⁺ littermates. Nulliparous (f, g) and (P16–18) pregnant (h, i) recipients are shown. Scale bars; 5 mm (top) 2 mm (bottom). j Box plot showing extent of fat pad filling by the engrafted epithelia in virgin (n = 10) and pregnant (n = 4) recipients. For both box plots, horizontal lines outside the boxes depict minimum and maximum values, upper and lower borders of the box represent lower and upper quartiles and the line inside the box identifies the median. k Bar plot showing relative transcript levels of the ERα target genes Areg, Pgr1, Prlr and Wnt4, and a control gene, Itgb1, normalised to 36b4 and Hprt in mammary glands from peripubertal WT, AF-1⁰, AF-2⁰ and ERα^−/− females. Data are shown as means ± SEM of three independent experiments. Paired two-tailed Student’s t test, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, n.s. not significant. l PgR IHC of mammary glands from 3-week-old WT, AF-1⁰, and AF-2⁰, ERα^−/− and PR^−/− mice. Representative pictures of glands analysed from three females of each genotype are shown. Scale bar; 100 μm

Fig. 4

Contribution of AF-1⁰, AF-2⁰ and ERα^−/− cells to chimeric ducts with WT MECs. a Scheme of experimental design. After cell dissociation, 10,000 ERα mutant (ERmt) or WT GFP⁺ epithelial cells were mixed with 90,000 WT.DsRed⁺ epithelial cells and injected into the cleared mammary fat pad of peripubertal recipient mice. b Representative fluorescence stereomicrographs of chimeric epithelia from WT.GFP⁺ or ERα.mutant.GFP⁺ and WT.DsRed⁺ cells mixed in a 1:10 ratio. Hosts were analysed 10 weeks after engraftment. Scale bars; 1 mm (left) 0.2 mm (right). c Pie charts showing the proportion of engrafted mammary glands appearing exclusively DsRed⁺, exclusively GFP⁺, or mixed (red and green stripes) based on evaluation at low (7.8×) magnification of fluorescence stereomicrographs. From top to bottom, n = 51, 21, 17 and 13. d Dot plot showing the percentage of the reconstituted ductal epithelium that is GFP⁺ in virgin mice based on images at 7.8× magnification, (n = 13–51) bars indicate medians. e Dot plot showing the ratio of GFP/DsRed signal intensity of the reconstituted ductal epithelia based on images at 7.8× magnification, bars indicate medians (n = 29, 15, 7 and 7). f, g Bar graphs showing flow cytometric analysis of the percentage of GFP⁺ cells in the CD24^high CD49f^low f and CD24^low CD49f^high g cell populations of reconstituted chimeric mammary epithelia. From left to right, n = 20, 5, 8 and 6. Shown are means ± SEM; Mann–Whitney test, two-tailed, ^**p < 0.01, ^****p < 0.0001, n.s. not significant

Fig. 5

Proliferation of AF-1⁰, AF-2⁰ and ERα^−/− cells in chimeric epithelia during pregnancy. a Representative fluorescence stereomicrographs of chimeric epithelia from WT.GFP⁺ or ERα mutant.GFP⁺ and WT.DsRed⁺ cells mixed in a 1:10 ratio. Hosts were analysed at P16–18. Scale bars; 1 mm. b Pie charts showing the number of mammary glands presenting only DsRed⁺ epithelial regenerations (red) or mixed ⁺ and GFP⁺ regenerations (red and green stripes) based on evaluation on “low magnification 7.8×” fluorescence stereomicrographs; from top to bottom, n = 20, 7, 7 and 6. c Dot plot showing percentage of area filled with GFP⁺ structures over total area filled in chimeric glands during pregnancy; evaluation at 7.8× magnification; shown are means ± SEM. d Dot plot showing ratios of GFP over DsRed signal intensity of different chimeras in pregnant hosts (n = 4–16; black line: median)

Fig. 6

Intraductal engraftment of WT and ERα^−/− MECs. a Fluorescence stereomicrograph of NSG.GFP⁺ mammary gland 5 days after intraductal injection of WT.DsRed⁺ MECs, representative of nine successfully injected glands. Scale bar; 1 mm. b, c Double immunofluorescence with anti-P63 and anti-RFP antibodies b or anti-αSma and anti-RFP antibodies c counterstained with DAPI. Scale bars; 50 μm. Representative pictures of injected glands from three females. d, e Fluorescence stereomicrographs of contralateral mammary glands 5 days after intraductal injection with WT.GFP⁺ or ERα^−/−.GFP⁺ in nulliparous d or day 16–18 pregnant host (e). Scale bars; 500 μm. f Bar plot showing relative GFP signal intensity in contralateral glands injected with WT.GFP⁺ or ERα^−/−.GFP⁺ cells in nulliparous and pregnant recipients, (n = 3–6; mean ± SEM). g, h Box plots showing percentage of GFP+ cells by FACS in nulliparous g and pregnant h hosts (n = 8, 9). Whiskers depict minimum and maximum values, box borders lower and upper quartiles, line inside identifies the median. Student’s paired t test, two-tailed. i, j Bar plots overlying dot plots showing relative contribution of GFP+ cells by FACS of intraductally engrafted glands from nulliparous (i) and pregnant (j) hosts, mean ± SEM. k Box plot showing percentage of GFP and Ki67 double + over total GFP+ cells from contralateral WT.GFP⁺ or ERα mutant.GFP⁺ with WT.DsRed⁺ MECs chimeras in nulliparous recipients 4 weeks after engraftment (n = 3; mean ± SD). l Box plot showing the percentage of GFP and EdU double + over total GFP+ cells from contralateral inguinal mammary glands intraductally engrafted with WT.GFP⁺ or ERα^−/−.GFP⁺ (left panel) and WT.GFP⁺ or AF-2⁰.GFP⁺ MECs (right panel) in pregnant recipients P12.5 (n = 3; mean ± SD). m Representative FACS plots showing CD49f and CD24 expression in Lin^- cells from mammary glands of AF-1⁰, AF-2⁰, ERα^−/− females and WT littermates. For both box plots, whiskers depict minimum and maximum values, borders of the box represent lower and upper quartiles, and line inside the box identifies the median. Paired two-tailed Student’s t test, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001, n.s. not significant

Fig. 7

Global gene expression profile of ERα^−/− and WT MECs. a Principal component (PC) analysis of RNAseq data showing the importance of the Esr1 genotype for global gene expression. b Bar graph showing the number of genes whose expression increased (IE) or decreased (DE) in ERα^−/−.GFP⁺ vs. WT.GFP⁺ MECs engrafted intraductally to contralateral glands of NSG females and subsequently isolated by FACS-sorting from hosts during pregnancy. c Bar plot showing reads per kilobase million (RPKM) of ERα high, sensor cell markers expressed in ERα^−/−.GFP⁺ and WT.GFP⁺ MECs that were engrafted intraductally to contralateral glands of NSG hosts and subsequently isolated by FACS-sorting from pregnant hosts. d Bar plot showing RPKM of milk genes expressed in FACS-sorted ERα^−/−.GFP⁺ and WT.GFP⁺ MECs grafted intraductally and isolated from pregnant hosts. e, f MetaCore analysis of genes showing decreased (e) or increased (f) expression, in ERα^−/−.GFP⁺ MECs compared to WT.GFP⁺ MECs. The most significantly enriched terms are listed with p value. g, h Reactome analysis of genes with decreased g or increased h expression, in ERα^−/−.GFP⁺ MECs compared to WT.GFP⁺ MECs. The most significantly enriched terms are listed with p value. i Visual representation of GO terms with increased expression in ERα^−/−.GFP+ vs. WT.GFP+ MECs by Cytoscape. j Pie chart showing proportions of KEGG terms among the genes with increased expression in the ERα^−/−.GFP⁺ vs. WT.GFP⁺ MECs

Fig. 8

ERα-mediated control of gene expression signatures. a Heatmap of RNAseq transcriptomic analysis of WT.GFP⁺ and ERα^−/−.GFP⁺ MECs that grew intraductally in NSG females contralateral glands. Row Z scores of genes that were differentially expressed are included. b, c Gene set enrichment analysis of genes that were down-regulated, in in ERα^−/−.GFP⁺ MECs compared to WT.GFP⁺ control cells. By hallmark gene sets (b) and curated gene sets (c). The most significantly enriched terms with p values are listed

Fig. 9

ERα mRNA and protein expression ERα^−/− and WT MECs. a ERα immunofluorescence (red) and b RNAscope with anti-Esr1 probes (red) on an adjacent section counterstained with DAPI (blue) on mammary gland from 6-week-old WT female. Scale bars; 50 μm. c Bar plot showing percentage of luminal cells with detectable levels of ERα protein and Esr1 mRNA at puberty (n = 5; mean ± SD). d Bar plot showing percentage of luminal cells with different Esr1 RNAscope scores (n = 5; mean ± SD). e RNAscope with anti-Areg probes (red) counterstained with DAPI (blue) on mammary gland from 6-week-old WT female. Scale bar; 50 μm. f RNAscope with anti-Esr1 or g anti-Areg probes (red) co-stained with anti-PgR antibody (green) and counterstained with DAPI (blue). Arrowheads point to apically-located luminal cells with ERα transcript but no PgR expression (f) and Areg transcript and PgR expression (g). Empty arrowheads point to myoepithelial cells f, g. Scale bars; 100 μm. h Current model of mammary stem cells and the differentiation hierarchy after Visvader, J.E. and Stingl, J.. In grey: previously proposed ERα status of different MEC differentiation stages. In red: ERα status based on the present findings