Scribble modulates the MAPK/Fra1 pathway to disrupt luminal and ductal integrity and suppress tumour formation in the mammary gland

Abstract

Polarity coordinates cell movement, differentiation, proliferation and apoptosis to build and maintain complex epithelial tissues such as the mammary gland. Loss of polarity and the deregulation of these processes are critical events in malignant progression but precisely how and at which stage polarity loss impacts on mammary development and tumourigenesis is unclear. Scrib is a core polarity regulator and tumour suppressor gene however to date our understanding of Scrib function in the mammary gland has been limited to cell culture and transplantation studies of cell lines. Utilizing a conditional mouse model of Scrib loss we report for the first time that Scrib is essential for mammary duct morphogenesis, mammary progenitor cell fate and maintenance, and we demonstrate a critical and specific role for Scribble in the control of the early steps of breast cancer progression. In particular, Scrib-deficiency significantly induced Fra1 expression and basal progenitor clonogenicity, which resulted in fully penetrant ductal hyperplasia characterized by high cell turnover, MAPK hyperactivity, frank polarity loss with mixing of apical and basolateral membrane constituents and expansion of atypical luminal cells. We also show for the first time a role for Scribble in mammalian spindle orientation with the onset of mammary hyperplasia being associated with aberrant luminal cell spindle orientation and a failure to apoptose during the final stage of duct tubulogenesis. Restoring MAPK/Fra1 to baseline levels prevented Scrib-hyperplasia, whereas persistent Scrib deficiency induced alveolar hyperplasia and increased the incidence, onset and grade of mammary tumours. These findings, based on a definitive genetic mouse model provide fundamental insights into mammary duct maturation and homeostasis and reveal that Scrib loss activates a MAPK/Fra1 pathway that alters mammary progenitor activity to drive premalignancy and accelerate tumour progression.

Figure 1. Scrib loss in the mammary gland results in duct hyperplasia.

A. Confirmation of Scrib deletion in mammary glands of MMTV-Cre;Scrib^flox/− 6 week old virgin mice by immunoblotting. B. Mammary whole mount analysis of 6 and 12 week old MMTV-Cre, MMTV-Cre;Scrib^flox/+ and MMTV-Cre;Scrib^flox/− virgin mice. Scale bar = 2 mm. C. Quantitation of increased terminal end bud number during duct morphogenesis of peripubescent MMTV-Cre;Scrib^flox/− 6 wk virgin mice compared to control mice. ± SEM. Mann Whitney t-test, (n = 5–9). D. Quantitation of increased branching in mammary glands of post-pubescent MMTV-Cre;Scrib^flox/− 12 wk virgin mice compared to control mice. ± SEM. Mann Whitney t-test, (n = 3–6). E. Histological analysis by H&E staining show normal duct formation in peri-pubescent 6 wk virgin MMTV-Cre (a) and MMTV-Cre;Scrib^flox/− (b) mice. Longitudinal sections of mature mammary ducts show organized epithelial bi-layer in 12 wk virgin MMTV-Cre control mice (e), whereas mammary ducts from 12 wk MMTV-Cre;Scrib^flox/− mice (f) have a highly disorganized epithelium with luminal occlusion resulting from an abundance of poorly differentiated and multilayered mammary epithelial cells. IHC shows basolateral localization of Scribble in luminal epithelial cells of 6 (c) and 12 week old (g) MMTV-Cre control mice and validates Scrib loss in mammary epithelium of 6 (d) and 12 week old (h) MMTV-Cre;Scrib^flox/− mice. Scale bar = 100 µm. F. Quantitation of multilayering phenotype performed by counting layers of luminal epithelial cells along length of longitudinal duct sections. ± SEM. Mann Whitney t-test, (n = 3–5). See also Figure S1.

Figure 2. Scrib-deficient lesions are characterized by polarity loss and the expansion of rapidly cycling poorly differentiated luminal mammary epithelial cells.

A. Histopathology by H&E staining of well differentiated luminal epithelial cells from MMTV-Cre control mice which are cuboidal with small tightly compacted nuclei and consistent cell-cell adhesion boundaries. Mammary ductal cells of MMTV-Cre;Scrib^flox/− virgin mice are no longer cuboidal, have loosely packed atypical nuclei, increased cytoplasm and inconsistent cell-cell adhesions (Lu: Lumen) B. Immunostaining in mammary ducts of 12 week mice show normal distribution of luminal (Cytokeratin 8, green and Cytokeratin 18 by IHC), and basal (Cytokeratin 5, red) cell populations in ducts from MMTV-Cre mice, whereas an expansion of Cytokeratin 8/18 luminal cells is observed in ducts from MMTV-Cre;Scrib^flox/− mice. Scale bar = 50 µm. C. Immunofluorescence of adhesion junctional proteins E-Cadherin (a,b) and β-catenin (c,d) in 12 wk virgin mice show normal lateral staining in luminal epithelial cells of MMTV-Cre mice (arrow) and ectopic randomized membrane staining in mammary epithelial cells of MMTV-Cre;Scrib^flox/− mice (arrowhead). Z-projection included for E-cadherin staining. Loss of membrane segregation in mammary ducts of 12 week MMTV-Cre;Scrib^flox/− virgin mice by confocal immunofluorescence microscopy of apical membrane marker pERM (green) and E-cadherin (red) (e, f). Red squares indicate regions in which 3D reconstructions were made from confocal z-series (g, h,)(Movie S1, S2) Scale bar = 50 µm. D. IHC and quantitation of proliferation (BrdU) and E. apoptosis (Cleaved Caspase 3) in ducts of 12 week virgin mice. (Additional quantitation of intraluminal cells not in contact with myoepithelial/basement membrane region was performed in MMTV-Cre;Scrib^flox/− mice and expressed as percentage of total luminal epithelial cell population. ± SEM. Mann Whitney t-test, (n = 3–5). Scale bar 100 µm. See also Figure S1.

Figure 3. Hyperplasia is preceded by defects in cell polarity, apoptosis and spindle orientation during the remodelling and maturation of Scrib-deficient mammary ducts.

A. Confocal immunofluorescence microscopy of apical membrane marker pERM (green) and lateral membrane marker E-cadherin (red) show polarised membrane segregation in mammary ducts of 6 week MMTV-Cre control mice (a), and partial to complete disruption of polarity in ducts of MMTV-Cre;Scrib^flox/− mice (b, c, d). Scale bar = 50 µm. B. IHC and quantitation of proliferation (BrDU) and C. apoptosis (Cleaved Caspase 3) in terminal end buds and immature ducts of 6 week old MMTV-Cre, MMTV-Cre;Scrib^flox/+ and MMTV-Cre;Scrib^flox/− virgin mice. ± SEM. Mann Whitney t-test, (n = 3–5, 5–9 TEB/mouse). Scale bar 100 µm. D. Spindle orientation of dividing luminal cells in 6 wk virgin mice. Mitotic cells were identified by PHH3 (light blue) immunofluorescence and spindle angle determined by centrosome positioning (pericentrin, red) compared to the lower layer of myoepithelial cells (CK5, green) (a). Representative images of a division within a Z-stack (b). Rosette plots showing percentage of spindle orientations in ducts of MMTV-Cre and MMTV-Cre;Scrib^flox/− mice (c).

Figure 4. Scrib loss impacts on basal progenitors activity but not mammary stem cell turnover in vivo.

A. Limiting dilution analysis of the repopulating frequency of CD29^hiCD24⁺ cells from MMTV-Cre;Scrib^flox/− and Control mice. Number of positive outgrowths is shown as number of outgrowths per number of mammary fat pads injected. B. Colony formation assay measuring increased clonogenic potential of FACS purified lin⁻/CD24⁺/CD29^hi basal cell populations from MMTV-Cre;Scrib^flox/− mice grown on irradiated 3T3s. n = 3.

Figure 5. Scribble loss disrupts Notch signalling and drives MAPK/Fra1 activation to allow inappropriate luminal expansion.

A. Representative FACS scatter plot of lin⁻/CD24⁺/CD29^hi basal and lin⁻/CD24⁺/CD29^lo luminal cell populations in 8–10 week old Scrib mutant mice. ± SEM. (n = 4–5 per group) B. Q-RT-PCR of Notch target genes Hey1 and Hey2, and ductal Gata3 in FACS purified lin⁻/CD24⁺/CD29^hi basal and lin⁻/CD24⁺/CD29^lo luminal cell populations and IHC confirmation of increased GATA3, and decreased Hey1 staining in hyperplastic ducts of 12 wk MMTV-Cre;Scrib^flox/− virgin mice. Expression levels of luminal maker CK8 and basal marker αSMA confirm purity of cell populations. ± SEM. students t-test, (n = 3, 8–10 week old mice). C. Q-RT-PCR of MAPK effector Fra1 in FACS purified lin⁻/CD24⁺/CD29^hi basal and lin⁻/CD24⁺/CD29^lo luminal cell populations (a), and confirmation of MAPK/ERK pathway activation by IHC (b–d), show significantly increased pERK positive nuclei (unpaired t test with Welch's correction, n = 3–4). Scale bar = 100 µm. D. Treatment of 6 week old MMTV-Cre, MMTV-Cre;Scrib^flox/+ and MMTV-Cre;Scrib^flox/− virgin mice with 20 mg/kg/day PD0325901 5 days on, 2 days off for two weeks blocked the onset of multilayering in MMTV-Cre;Scrib^flox/− mice. Detected by H&E staining (a–d) E. Multilayering determined and quantified by counting layers of luminal epithelial cells along length of longitudinal duct H&E sections (i). ± SEM. Mann Whitney t-test, (n = 7–9, 20–39 ducts/mouse). F. MAPK inhibition was assessed by IHC of pERK (a–d). Scale bar = 100 µm. G. Polarity status by IF of pERM (green) and E-cadherin (red) (a–d). Scale bar = 50 µm. see also Figure S2 and S4, S5.

Figure 6. Scribble loss enhances mammary tumourigenesis.

A. Histological analysis of ductal architecture in 75 week mice by H&E staining show hyperplastic alveolar nodules (HAN) in MMTV-Cre;Scrib^flox/− mice (c, d) compared to control (a, b) as indicated by appearance of circular lobules and lipid droplets (d, black arrow). B. Quantitation of rare (<20%, +), frequent (20–80%, ++), or extensive (>80%, +++) observable alveolar or lipid droplets within mammary epithelium. C. IHC of Scribble (a, b), pSTAT5 (c, d) and Ki67 (e, f) in mammary ducts (du) or HAN of 75 week MMTV-Cre control or MMTV-Cre;Scrib^flox/− mice. Scale bar = 100 µm. D. Tumour-free plot for MMTV-Cre (n = 19), MMTV-Cre;Scrib^flox/+ (n = 20) and MMTV-Cre;Scrib^flox/− (n = 19) virgin mice palpated weekly for mammary tumours. Scribble loss detected by IHC in tumours from MMTV-Cre;Scrib^flox/− aged mice. Scale bar = 100 µm. E. Incidence of mammary tumours amongst cohorts of MMTV-Cre, MMTV-Cre;Scribflox/+ and MMTV-Cre;Scribflox/− 525 day old virgin mice. F. Tumours from MMTV-Cre;Scrib^flox/− aged mice show increased cell turnover rates as quantified by PCNA and Cleaved caspase 3 IHC. Scale bar = 100 µm. G. Comparative tumour pathologies show tumours from MMTV-Cre;Scrib^flox/− aged mice (n = 11) are more progressed and lack structural acinar/glandular characteristics compared to tumours from MMTV-Cre mice (n = 11). Representative H&E staining of tumour classifications. Scale bar = 100 µm. see also Figure S6.

Figure 7. Model summarizing impacts of Scribble in mammary development and homeostasis.

In normal mammary development, Scribble moderates basal progenitor clonogenic potential and controls spindle orientation to maintain a balanced epithelial hierarchy. Scribble is also required for remodelling and to establish apical-basal polarity during late stage duct morphogenesis. Scribble loss leads to hyperactivation of the Ras/MAPK pathway and the expansion of ectopic luminal epithelial cells to promote tumour development.