Rebalancing of actomyosin contractility enables mammary tumor formation upon loss of E-cadherin

Abstract

E-cadherin (CDH1) is a master regulator of epithelial cell adherence junctions and a well-established tumor suppressor in Invasive Lobular Carcinoma (ILC). Intriguingly, somatic inactivation of E-cadherin alone in mouse mammary epithelial cells (MMECs) is insufficient to induce tumor formation. Here we show that E-cadherin loss induces extrusion of luminal MMECs to the basal lamina. Remarkably, E-cadherin-deficient MMECs can breach the basal lamina but do not disseminate into the surrounding fat pad. Basal lamina components laminin and collagen IV supported adhesion and survival of E-cadherin-deficient MMECs while collagen I, the principle component of the mammary stromal micro-environment did not. We uncovered that relaxation of actomyosin contractility mediates adhesion and survival of E-cadherin-deficient MMECs on collagen I, thereby allowing ILC development. Together, these findings unmask the direct consequences of E-cadherin inactivation in the mammary gland and identify aberrant actomyosin contractility as a critical barrier to ILC formation.

Fig. 1

E-cadherin loss drives cell extrusion towards the basal lamina. a Schematic overview of engineered alleles in Wcre;Cdh1^F/F;mTmG mice. b, c Examination of GFP-positive Wcre activity in mammary glands of 6-week-old Wcre;mTmG female mice by immunofluorescence (IF) analysis (n = 6). b IF staining of GFP was examined in cytokeratin-8 (CK8)-positive luminal mouse mammary epithelial cells (MMECs) and c cytokeratin-14 (CK14)-positive myoepithelial cells. Scale bar, 10 µm. d Identification of E-cadherin inactivated MMECs by IF staining of E-cadherin, GFP, CK14, and CK8 in 6-week-old Wcre;Cdh1^F/F;mTmG female mice (n = 6). Arrows indicate events of cell extrusion upon inactivation of E-cadherin. Scale bare, 10 µm. e Immunohistochemical detection of GFP-positive E-cadherin inactivated MMECs in mammary gland sections of 3-, 4-, and 6-month-old Wcre;Cdh1^F/F;mTmG female mice and age-matched WCre;mTmG control mice (n = 3). Scale bar, 20 µm. f Examination of E-cadherin expression in mammary gland sections of 3-month-old Wcre;Cdh1^F/F;mTmG female mice and age-matched Wcre;mTmG control mice by IF analysis of GFP, E-cadherin, CK14, and Hoechst. Asterisk indicates area of zoom. Scale bar, 50 µm. g Quantification of the amount of extruded GFP-positive cells in 3-month-old Wcre;Cdh1^F/F;mTmG (n = 3) female mice and age-matched Wcre;mTmG (n = 3) control mice. Data are of nine images per group. h Quantification of the average size of extruded GFP-positive cell clusters in the mammary glands of Wcre;Cdh1^F/F;mTmG mice at the ages of 3, 5, and 12 months. Data are of three mice per time point and 10 images per mouse. All data are depicted as mean ± standard deviation. All p values were calculated using an unpaired two tailed t-test. Source data are provided as a Source Data file

Fig. 2

E-cadherin loss increases actomyosin contractility. a Still images derived from in vivo intravital imaging of the mammary gland of 8-week-old Wcre;mTmG and Wcre;Cdh1^F/F;mTmG mice displaying GFP-positive Cre-switched MMECs and mTomato non-switched MMECs and stromal cells. Zooms reveal motile GFP-positive E-cadherin inactivated MMECs in Wcre;Cdh1^F/F;mTmG mice. Scale bars, 20 µm. b Quantification of the percentage of GFP-positive motile cells among Wcre;mTmG (n = 2) and Wcre;Cdh1^F/F;mTmG (n = 3) mice. Data are of four and three images per group, respectively. c Still image of in vivo intravital imaging of Wcre;Cdh1^F/F;mTmG mice demonstrates extensive cell blebbing of GFP-positive E-cadherin inactivated MMECs. Scale bar, 10 µm. d Quantification of the percentage of GFP-positive blebbing cells among Wcre;mTmG (n = 2) and Wcre;Cdh1^F/F;mTmG (n = 3) mice. Data are of four and three images per group, respectively. e Examination of myosin light chain (MLC) activity by IF analysis of pSer¹⁹ MLC in GFP-positive E-cadherin-deficient MMECs in 5-month-old Wcre;mTmG and Wcre;Cdh1^F/F;mTmG mice. Scale bar, 20 µm. f Quantification of the amount of GFP⁺ pSer¹⁹ MLC^high cells Wcre;mTmG (n = 3) and Wcre;Cdh1^F/F;mTmG (n = 3) mice. Data are of 51 and 60 images, respectively. All data are depicted as mean ± standard deviation. All p values were calculated using an unpaired two tailed t-test. Source data are provided as a Source Data file

Fig. 3

Actomyosin relaxation enables survival upon E-cadherin loss. a Merged brightfield and GFP images of MMECs isolated from Wcre;mTmG and Wcre;Cdh1^F/F;mTmG female mice in the absence or presence of 10 µM Y-27632. Asterisks indicate areas of zoom, red dotted lines display outgrowth of E-cadherin-deficient MMECs. Zooms display altered cell adhesion in the absence or presence of Y-27632 of GFP-positive E-cadherin-deficient cells derived from Wcre;Cdh1^F/F;mTmG female mice. Scale bar, 50 µm. b Western blot analysis of Wcre;mTmG and Wcre;Cdh1^F/F;mTmG MMECs stained for E-cadherin and tubulin (loading control). c Western blot analysis of subclones derived from the Wcre;mTmG and Wcre;Cdh1^F/F;mTmG MMECs stained for E-cadherin and tubulin (loading control). d Brightfield images of MMECs 3 h post seeding in the absence or presence of 10 µM Y-27632 or 10 µM blebbistatin. Scale bar, 20 µm. e Western blot analysis of pSer⁴⁷³ Akt, Akt, cleaved caspase-3, and actin (loading control) in Wcre;Cdh1^F/F;mTmG subclones grown in the absence or presence of 10 µM Y-27632. f, g Representative images (f) and quantification (g) of clonogenic assays with Wcre;mTmG (circles) and Wcre;Cdh1^F/F;mTmG (squares) subclones grown in the presence or absence of 10 µM Y-27632 or 10 µM blebbistatin (7 days after seeding the cells). Data are of three independent experiments with three clones per experiment. h Western blot analysis of WT FVB and Cdh1^∆/∆ MMECs stained for E-cadherin and actin (loading control). i, j Representative images (i) and quantification (j) of clonogenic assays with Cdh1^∆/∆ and WT control MMEC lines with or without 10 µM Y-27632. Data are of three independent experiments. k Representative images of immunoblots of active RhoA pull-down assays and ser¹⁹ myosin light chain phosphorylation (pMLC) from Cdh1^∆/∆ and WT control MMECs harvested 3 h post seeding. l Quantification of active RhoA pull-down assays by densitometry normalized to the actin loading control. Data are of three independent experiments. All data are depicted as mean ± standard deviation. All p values were calculated using an unpaired two tailed t test. Source data are provided as a Source Data file

Fig. 4

Balanced actomyosin contractility is essential for survival. a, b Representative images (a) and quantification (b) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMECs cultured with different concentrations of blebbistatin. Data are of three independent experiments with three clones per experiment. c, d Representative images (c) and quantification (d) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMECs cultured with different concentrations of H1152. Data are of three independent experiments with three clones per experiment. All data are depicted as mean ± standard deviation. Source data are provided as a Source Data file

Fig. 5

MYPT1 drives actomyosin relaxation and cell adhesion. a Schematic overview of the doxycycline (dox) inducible t-MYPT1 (MYPT1^1–413) construct. b Western blot analysis of Wcre;Cdh1^F/F;mTmG MMECs transduced with dox-inducible Flag-tagged t-MYPT1 in the absence, presence, or following washout of 2 mg/mL dox as stained for FLAG and actin (loading control). c, d IF staining (c) and quantification (d) of dox-inducible Flag-tagged t-MYPT1 Wcre;Cdh1^F/F;mTmG MMECs cultured in the presence of 2 µg/mL dox or 50 h post washout stained for phospho myosin light chain (pMLC), actin, and Hoechst. Data are of 15 images per condition. e, f Representative images (e) and quantification (f) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMECs expressing Dox-inducible Flag-tagged t-MYPT1 cultured with or without 2 mg/mL dox and/or 10 µM Y-27632. Data are of three independent experiments with three clones per experiment. g Schematic overview of full-length MYPT1, t-MYPT1, and t-MYPT1ΔPP1 (MYPT1^92–413) displaying the PP1 recognition motif and the ankyrin repeats (ANK) domain, the leucizine zipper domain and inhibitory phosphorylation sites (T696 and T853). h, i IF staining (h) and quantification (i) of Wcre;Cdh1^F/F;mTmG MMECs transduced with MYPT1-wt, t-MYPT1, t-MYPT1ΔPP1, or GFP 48 h post washout of 10 µM Y-27632 and stained for pMLC, actin, and Hoechst. Data are of 15 images per condition. j, k Representative images (j) and quantification (k) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMECs transduced with t-MYPT1, t-MYPT1ΔPP1 or GFP cultured in the presence or absence of 10 µM Y-27632 (7 days after seeding the cells). Data are of three independent experiments with three clones per experiment. All data are depicted as mean ± standard deviation. All p values were calculated using an unpaired two tailed t-test. Source data are provided as a Source Data file

Fig. 6

Collagen I does not support adhesion upon loss of E-cadherin. a IF analysis of GFP (green), Laminin (red), and Keratin 8 (cyan) expression in mammary gland sections from 12-week-old Wcre;mTmG (top panels) and Wcre;Cdh1^F/F;mTmG mice (middle and bottom panels). Scale bar is 20 µM b, c Representative images (b) and quantification (c) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMEC subclones seeded on laminin-coated wells or uncoated plastic wells. Data are of three independent experiments with three clones per experiment. d GFP immunohistochemistry staining combined with a methylene blue staining (collagen). Scale bar is 20 µM. e, f Representative images (e) and quantification (f) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMEC subclones seeded on laminin-coated or collagen I-coated wells in the presence or absence of Y-27632 (10 µM) or blebbistatin (10 µM). Data are of three independent experiments with three clones per experiment. g–i Immunoblots (g) and quantifications (h, i) of active RhoA pull-down assays from Cdh1^∆/∆ MMECs harvested 3 h post plating on laminin- or collagen I-coated plates. Data are of three independent experiments. j, k Representative images (j) and quantification (k) of clonogenic assays with Wcre;Cdh1^F/F;mTmG MMECs transduced with t-MYPT1, t-MYPT1ΔPP1, or GFP cultured seeded on laminin-coated or collagen I-coated wells. Data are of three independent experiments with three clones per experiment. All data are depicted as mean ± standard deviation. All p values were calculated using an unpaired two tailed t-test. Source data are provided as a Source Data file

Fig. 7

Relaxation of actomyosin contractility drives ILC formation. a Schematic representation of intraductal injections performed in WapCre;Cdh1^F/F mice with high-titer lentiviruses produced from vectors encoding MYPT1-wt, t-MYPT1, t-MYPT1ΔPP1, or empty vector (EV). b Quantified tumor burdens of Wcre;Cdh1^F/F females 20 weeks after intraductal injection of Lenti-EV (n = 16 glands), Lenti-MYPT1-wt, (n = 12 glands), Lenti-t-MYPT1 (n = 24 glands), or Lenti-t-MYPT1ΔPP1 (n = 14 glands). Data are depicted as mean ± standard deviation. p values were calculated using an unpaired two-tailed t-test. c Representative images of tumor sections stained with H&E or stained for E-cadherin and CK8 and Collagen (Sirius red). Scale bar is 20 µM. Source data are provided as a Source Data file