Pygo2 expands mammary progenitor cells by facilitating histone H3 K4 methylation

Abstract

Recent studies have unequivocally identified multipotent stem/progenitor cells in mammary glands, offering a tractable model system to unravel genetic and epigenetic regulation of epithelial stem/progenitor cell development and homeostasis. In this study, we show that Pygo2, a member of an evolutionarily conserved family of plant homeo domain-containing proteins, is expressed in embryonic and postnatal mammary progenitor cells. Pygo2 deficiency, which is achieved by complete or epithelia-specific gene ablation in mice, results in defective mammary morphogenesis and regeneration accompanied by severely compromised expansive self-renewal of epithelial progenitor cells. Pygo2 converges with Wnt/beta-catenin signaling on progenitor cell regulation and cell cycle gene expression, and loss of epithelial Pygo2 completely rescues beta-catenin-induced mammary outgrowth. We further describe a novel molecular function of Pygo2 that is required for mammary progenitor cell expansion, which is to facilitate K4 trimethylation of histone H3, both globally and at Wnt/beta-catenin target loci, via direct binding to K4-methyl histone H3 and recruiting histone H3 K4 methyltransferase complexes.

Figure 1.

Pygo2 expression in mammary epithelial progenitor cells. (A, C, and F) Sketch diagrams showing anatomical features of embryonic (A), pubertal (C), and pregnant (F) mammary glands. (B, left) Strong Pygo2 expression in E15.5 mammary bud, particularly in the bulb region (asterisk). The inset is an enlarged image showing nuclear localization of Pygo2. The white dashed lines indicate basement membrane separating epithelial and mesenchymal compartments. The absence of signal in Pygo2^−/− mammary bud (right) confirms antibody specificity. The arrowhead points to Pygo2-expressing mesenchymal cells. (D) Pygo2 expression in cap (outlined by keratin 14⁺ or K14⁺; arrow) and body cells (indicated by an asterisk) of TEBs from 3-wk-old females. (E) Pygo2 expression in 6-wk-old mammary epithelium. Note the reduced Pygo2 expression as cells leave the TEB (arrow). (G) Ubiquitous Pygo2 expression in lobuloalveolar cells of 18.5-d pregnant mammary glands. The planes of sections for D/E and G are indicated by the dashed lines in C and F, respectively. WT, wild type. Bars: (B) 30 µm; (D and E) 25 µm; (G) 50 µm.

Figure 2.

Embryonic and postnatal mammary phenotypes of Pygo2-deficient mice. (A) Whole-mount preparations of carmine-stained skins of E18.5 wild-type (WT; top) and Pygo2^−/− (bottom) embryos showing select individual mammary glands (MG). (B) Summary of results obtained on all ten glands from multiple mutant embryos (n = 4). The mean branching point is calculated to be 3.25 ± 0.63 and 6.5 ± 1.04 for mutant and wild type, respectively (asterisk; P = 0.04). (C) Whole-mount LacZ-stained skin preparation of an E11.5 K14-Cre/Rosa26R embryo. A transversal section of placode 3 (arrow) is shown at the bottom. (D) LacZ-stained cross sections (counterstained with hematoxylin and eosin) through mammary duct (top) and TEB (bottom) of a 3-wk-old K14-Cre/Rosa26R female. (E) Genomic PCR of Pygo2 alleles showing recombination in epidermis but not dermis or other tissues. Co, colon; De, dermis; Ep, epidermis; He, heart; In, intestine; Li, liver; Lu, lung; St, stomach. The plus and minus signs denote the controls for PCR. (F) Absence of Pygo2 in ductal epithelia and TEBs (insets) of virgin SSKO mammary glands. (G and H) Carmine red–stained whole-mount preparations of mammary gland 4 from 3-wk-old (n = 4; G) and 6-wk-old (n = 3; H) females. Note the absence of TEB structures (short arrows in G) and the defective ductal elongation (marked by long arrows in H) in mutants. The control genotypes shown are K14-Cre/Pygo2^flox/+ (F and G) and Pygo2^flox/+ (G). Bars: (A) 112 µm; (C) 30 µm; (D and F) 25 µm; (G) 107 µm; (H) 937 µm.

Figure 3.

Pygo2 SSKO mammary epithelium contains fewer stem/progenitor cells. (A) Results of FACS analysis revealing a decreased number of Lin⁻CD24⁺CD29^High cells in 10-wk-old Pygo2-deficient mammary glands. Representative FACS profiles from a single pair are shown on the left, and mean values from three different pairs are shown on the right. (B) Reduced presence of K6⁺ progenitor cells in 8-wk-old Pygo2-deficient mammary duct and ductal termini (insets). (C) ER expression is not affected by Pygo2 loss. (D) Representative results from limiting dilution transplantations of MECs derived from 9–12-wk-old control or Pygo2 SSKO (n = 6) mice. (E) Reduced colony formation and K6 expression in Matrigel culture of MECs isolated from 8-wk-old SSKO mice. A quantitative analysis (right) reveals a statistically significant reduction in the number of K6⁺ colonies and K6⁺ cells per positive colony. The control genotypes shown are two Pygo2^flox/+ and one wild type (A), Pygo2^flox/+ (B), K14-Cre/Pygo2^flox/+ (C), four Pygo2^flox/+ and two K14-Cre/Pygo2^flox/+ (D), and two K14-Cre/Pygo2^flox/+ and one Pygo2^flox/+ (E). (A and E) Error bars represent standard deviation. Bars: (B and C) 25 µm; (E) 37.5 µm.

Figure 4.

Loss of Pygo2 leads to reduced mammary epithelial proliferation and compromised G1–S transition. (A) Immunohistochemical detection of BrdU incorporation in mammary buds of E15.5 control and Pygo2^−/− embryos. The BrdU-labeling index, calculated as the percent of BrdU-positive cells per total number of cells in mammary buds, is shown on the right. P < 0.001. The dashed lines indicate the basement membranes. WT, wild type. (B) Reduced growth of Pygo2-depleted MCF10A cells in high-density culture. (C) Reduced colony formation by Pygo2-depleted cells. (D) Cell cycle analysis of control and Pygo2-depleted cells. (E) Altered expression of cell cycle genes in Pygo2-depleted cells. mRNAs were collected 3 d after siRNA transfection and quantified by quantitative RT-PCR. (A–C and E) Error bars represent standard deviation. Bars, 20 µm.

Figure 5.

Pygo2 is required for Wnt/β-catenin target gene expression in mammary epithelium. (A and B) Whole-mount LacZ staining of Pygo2 wild-type (WT; top) and deficient (bottom) BAT-gal embryos at E11.5 (A) and E15.5 (B). Numbers (2 and 3) next to the arrows indicate mammary buds 2 and 3. Corresponding sections of the boxed areas are shown on the right. Closed and open arrows indicate externally visible and invisible mammary glands, respectively. (C) Whole-mount in situ hybridization for Lef1 on E13.5 wild-type and mutant embryos. (D) Results of quantitative RT-PCR showing reduced Lef1 expression in Pygo2 SSKO mammary glands (n = 3). The error bar represents standard deviation. Bars: (A) 195 µm; (B) 315 µm; (C) 900 µm; (insets) 20 µm.

Figure 6.

Loss of Pygo2 rescues the mammary outgrowth phenotype of K14-ΔN–β-catenin mice. (A) Whole-mount carmine staining of mammary glands from 12-wk-old wild-type (WT) and K14-ΔN–β-catenin (n = 5) females. (B and C) Persistent K6 expression in adult glands from K14-ΔN–β-catenin (12 wk old; B, right) and Cby1^−/− (39 wk old; C, right) mice. The basal compartment is outlined by K14 staining. The insets show images of TEBs. (D) Whole-mount carmine-stained skin of 12-wk-old K14-ΔN–β-catenin and K14-ΔN–β-catenin/Pygo2 SSKO (n = 4) littermates. (E) Quantitative analysis of branch points. Error bars represent standard deviation. β-cat, β-catenin. Bars: (A and D) 500 µm; (B and C) 25 µm.

Figure 7.

The H3K4me2/3-binding activity of Pygo2 is required for colony formation by mammary progenitor cells. (A) Sequence alignment of PHD fingers in Pygo and ING2/bromodomain PHD finger transcription factor (BPTF) proteins. Note that the residues critical for H3K4me3 binding in ING2 and bromodomain PHD finger transcription factor (highlighted in red) are conserved among Pygo PHDs. Residues required for BCL9 binding are highlighted in blue. The asterisks represent conserved cysteine and histidine amino acids of the C₄HC₃ PHD. (B) Coomassie blue staining of calf thymus histone pull-downs. (C) Western blot analysis of pull-down samples in B with antibodies against mono-, di-, and trimethylated H3 K4. (D) Coimmunoprecipitation/Western blot analysis of Pygo2–H3K4me3 interaction in 293T cells. (E) Pull-down of histone H3 N-terminal peptides, with eluates analyzed by Western blotting with anti-GST antibody. (F) Mapping residues in Pygo2 PHD required for H3K4me3 (b) or BCL9 HD1 (c) interaction. (a) Coomassie blue staining of GST, GST–Pygo2 PHD derivatives, and 6× His–BCL9 HD1 domain proteins purified from bacteria. GST pull-downs of histones were analyzed by Western blotting with anti-H3K4me3 (b), whereas nickel (6× His) pull-downs of GST-PHD derivatives were analyzed with anti-GST antibodies (c). (G) Synergistic interactions among H3K4me3–Pygo2–BCL9. Purified proteins were incubated as indicated and pulled down by nickel beads, and eluates were analyzed by Western blotting with the indicated antibodies. (H) Methyl histone H3 binding of Pygo2 is required for colony formation by MCF10A cells. (a) Representative images revealing the effects of exogenous Pygo2 derivatives are shown. (b) Quantification of three independent experiments. Error bars represent standard deviation. *, P = 1.0; **, P = 0.045; ‡, P = 0.010; and †, P = 0.030. P-values were calculated using two-tailed t tests assuming equal variance. Note that wild-type and mutant 2× Flag–flagged Pygo2 proteins were expressed at similar levels in infected cells (c). IB, immunoblot; IP, immunoprecipitation.

Figure 8.

Pygo2 regulates H3K4me3 levels by facilitating chromatin association of WDR5. (A) Requirement of Pygo2 for global H3K4me3 in MCF10A cells. The H3K4me3/H3 ratio in cells treated with 30 nM of control siRNA was arbitrarily set to be 1. (B) Western blot analysis showing that Pygo2 knockdown specifically affects di- and trimethylation of H3 K4. (C) Reduced H3K4me3 levels in Pygo2 SSKO mammary glands. The H3K4me3/H3 ratio in wild-type glands was arbitrarily set to be 1. (D) Results of ChIP experiments revealing reduced H3K4me3 at the c-Myc enhancer and Lef1 promoter. Note that the total H3 levels at these loci are not significantly affected. (E) Association of Pygo2 with WDR5 in 293T cells with and without BIO treatment. (F) Pygo2–WDR5 interaction in MCF10A cells. Asterisks label cross-reacting IgGs. (G) Reduction of chromatin association of WDR5 upon knockdown of Pygo2 in MCF10A cells. The asterisk indicates a cross-reacting band. WCE, whole-cell extract; S2, soluble cytosolic fraction; S3, soluble nuclear fraction; P3, chromatin-enriched fraction. (H) ChIP analysis showing decreased WDR5 occupancy at the c-Myc enhancer and Lef1 promoter in Pygo2-depleted MCF10A cells. (D and H) Error bars represent standard deviation. IB, immunoblot; IP, immunoprecipitation.