Eya1 controls cell polarity, spindle orientation, cell fate and Notch signaling in distal embryonic lung epithelium

Abstract

Cell polarity, mitotic spindle orientation and asymmetric division play a crucial role in the self-renewal/differentiation of epithelial cells, yet little is known about these processes and the molecular programs that control them in embryonic lung distal epithelium. Herein, we provide the first evidence that embryonic lung distal epithelium is polarized with characteristic perpendicular cell divisions. Consistent with these findings, spindle orientation-regulatory proteins Insc, LGN (Gpsm2) and NuMA, and the cell fate determinant Numb are asymmetrically localized in embryonic lung distal epithelium. Interfering with the function of these proteins in vitro randomizes spindle orientation and changes cell fate. We further show that Eya1 protein regulates cell polarity, spindle orientation and the localization of Numb, which inhibits Notch signaling. Hence, Eya1 promotes both perpendicular division as well as Numb asymmetric segregation to one daughter in mitotic distal lung epithelium, probably by controlling aPKCζ phosphorylation. Thus, epithelial cell polarity and mitotic spindle orientation are defective after interfering with Eya1 function in vivo or in vitro. In addition, in Eya1(-/-) lungs, perpendicular division is not maintained and Numb is segregated to both daughter cells in mitotic epithelial cells, leading to inactivation of Notch signaling. As Notch signaling promotes progenitor cell identity at the expense of differentiated cell phenotypes, we test whether genetic activation of Notch could rescue the Eya1(-/-) lung phenotype, which is characterized by loss of epithelial progenitors, increased epithelial differentiation but reduced branching. Indeed, genetic activation of Notch partially rescues Eya1(-/-) lung epithelial defects. These findings uncover novel functions for Eya1 as a crucial regulator of the complex behavior of distal embryonic lung epithelium.

Fig. 1.

Eya1 and polarity proteins are expressed in the lung. (A-C) Antibody staining shows widespread expression of Eya1 in both lung epithelium and mesenchyme at E11.5-E12.0 (arrowheads), and strong polarized Eya1 signals in the distal epithelium at E12.5-E14.0 (B,C; arrowheads). (D-F) Immunofluorescence shows very week Numb expression at E11.5-E12.0 distal epithelium (D), and strong polarized Numb signals in the distal (inset a″ in E and F; double arrowheads) rather than proximal epithelium (inset b″ in E; arrowhead) from E13-E13.5. (G) Western blot shows no apparent changes in the expression of polarity proteins in E14 Eya1^−/− lungs. Scale bars: 50 μm.

Fig. 2.

Eya1 deletion causes mislocalization of spindle-regulatory proteins, and increases parallel spindle alignments in mitotic distal epithelium. (A,A′,B,B′,D,D′,E,E′,G,G′,H,H′) Immunofluorescence with specific antibodies shows that LGN, NuMA and Insc specifically localize to apical cell sides of wild-type distal epithelial cells (A,A′,D,D′,G,G′; arrowheads) and have a diffuse, basolateral or basal localization in Eya1^−/− distal epithelial cells (B,B′,E,E′,H,H′; arrowheads). Broken line represents the collagen IV-stained basement membrane. A′,B′,D′,E′,G′ are electronic magnifications from areas marked with asterisks in A,B,D,E,G, respectively. (C,F,I) Quantification of mitotic distal epithelial cells with apical localization of LGN, NuMA or Insc for the experiments shown in A-H′. This is expressed as a percentage of all mitotic distal epithelial cells. *Significantly different from control (P<0.05; Student's t-test). Error bars indicate s.e.m. (J,K) Schematic representation of LGN localization in wild-type (J) or Eya1^−/− (K) distal epithelial cells. Each dot represents the centre of an LGN crescent in a single mitotic cell. (L,M) Schematic representation of spindle orientation in E14 wild-type (L) or Eya1^−/− (M) distal epithelium. Each line represents the spindle axis of a single late mitotic cell. (N,O) Examples of distal epithelial mitotic cells that divide perpendicularly, as represented by the perpendicular orientation of pericentrin-stained centrosomes (arrowheads/arrows) relative to the basement membrane (broken line; N), and others that have their centrosomes aligned parallel to the basement membrane (O; arrowheads). (P) Quantitation of the spindle orientations, which is expressed as a percentage of all divisions in the distal epithelium, of the experiments shown in N,O for E14 wild-type/Eya1^−/− lungs. Mitotic cells are quantified based on centrosome orientation relative to the basement membrane in order to distinguish parallel from perpendicular spindle alignments.Bars carrying the same letter (a,b) are significantly different from one another (*P<0.05; Student's t-test). Data are mean±s.e.m. Scale bars: 50 μm.

Fig. 3.

Functions of polarity proteins in lung epithelium in vitro. (A,B) Immunocytochemistry with α-tubulin antibody shows well-organized and oriented spindle fibers (arrowheads) in MLE15 cells during early (A) and late (B) mitosis. (C,D) Spindle fibers are disorganized/disoriented (arrowheads) in mitotic MLE15 cells after Insc or Gpsm2 knockdown. (E-M) Immunocytochemistry shows that MLE15-positive cells (arrowheads) for Sox9 (F), Hes-1 (K) or Hes-5 (M) increase with strong nuclear staining, while SP-B-positive cells (H) decrease after Numb knockdown. (I) Quantitation of Sox9- or SP-B-positive cells, which is expressed as a percentage of all counted MLE15 cells, of the experiments shown in E-H. Bars carrying the same letter (a,b) in I, O or P are significantly different from one another (*P<0.05; Student's t-test). Data are mean±s.e.m. (N) Western blot of the experiments shown in J-M. (O) Means fluorescence intensity of Hes-1 or Hes-5 staining for experiments showing in J-M. (P) Quantitation of Hes-1-positive cells, which is expressed as a percentage of all counted MLE15 cells, of the experiments shown in J-K. In O,P, Error bars indicate s.e.m. Scale bars: 50 μm.

Fig. 4.

Eya1 is required for polarized apical localization and phosphorylation of Numb in distal epithelial cells. (A,B) Immunofluorescence for Numb shows preferential Numb localization to the apical side of distal epithelial cells in wild-type lungs (A; arrowheads). (B) Increased Numb expression but loss of its asymmetric localization in Eya1^−/− distal epithelium (arrowheads). (C-E) High magnification of wild-type/mutant mitotic distal epithelial cells at anaphase/telophase shows that Numb localizes asymmetrically, and is inherited only by one daughter cell in wild-type lungs (a in C). a″ is a daughter cell that does not inherit Numb. (D,E) Numb (arrowheads) fails to localize asymmetrically and is inherited by both daughter cells (b,b″,c,c″ in D and d,d″,e,e″ in E) in Eya1^−/− mitotic distal epithelium cells with planer/parallel (D) or perpendicular (E) divisions (E, which represents the area marked with an asterisk in B), relative to the collagen IV-stained basement membrane (thick white broken lines in A-E). (F) Quantification of late mitotic distal epithelial progenitors, with Numb inherited by one (1daug.) or both (2daug.) daughter cells in E14 wild-type or Eya1^−/− lungs. Bars carrying the same letter (a,b) in F, L are significantly different from one another (*P<0.05; Student's t-test). Data are mean±s.e.m. (G) Morphometric analysis of Numb signal intensity/distribution of selected areas (thin yellow broken lines in A,B) shows loss of polarized/asymmetric localization of Numb, which is distributed at both apical and basal sides of Eya1^−/− distal epithelium. (H-I′) Immunostaining with specific Ser295 phospho-Numb antibody shows increased Numb phosphorylation in E14 Eya1^−/− distal epithelium (I,I′; arrowheads: staining in the cytoplasmic side of cell membrane and in the nuclei) compared with control lungs (H,H′; arrowheads). (H′,I′) High magnification of boxed areas in panels H,I, respectively. (J,K) Western blots of E14 lungs with anti-Numb (J), anti-Ser-295 phospho-Numb or anti-tyrosine phosphorylated aPKCζ antibody (K) show increased Numb/aPKCζ phosphorylation in Eya1^−/− lungs. Bars in J represent quantified western blot signals (mean±s.e.m., **P<0.001). Blue numbers in K represent relative band intensity. (L) Mean fluorescence intensity of total Numb or phospho-Numb staining compared between wild-type and Eya1^−/− distal epithelium for experiments showing in A,B and H-I′. Error bars indicate s.e.m. (M,N) In control mitotic MLE15 cells, Numb (arrowheads) segregated asymmetrically and was inherited by one daughter cell in anaphase/telophase. Upon Eya1 knockdown, Numb segregated to both daughters (N; arrowheads). Scale bars: 50 μm.

Fig. 5.

Eya1 regulates aPKCζ and Numb phosphorylation. (A,B,G,H,M,N) Antibody staining of MLE15 cells shows changes of Numb distribution/Ser295 phosphorylation (B,H; arrowheads) and aPKCζ tyrosine phosphorylation (N; arrowheads) after Eya1 knockdown compared with control cells. Arrowheads in A indicate polarized Numb staining. (C,D,I,J,O,P) Rescue of endogenous Eya1 function by co-transfection of murine siRNA and murine wild-type or enzymatically inactive mutant Eya1 constructs (48 hours) in MLE15 cells reveals that Numb and aPKCζ distribution/phosphorylation (arrowheads) are dependent on Eya1 phosphatase activity. (E,K) Inhibition of aPKCζ in Eya1 siRNA-transfected MLE15 cells rescues Numb distribution/Ser295 phosphorylation (arrowhead). (F,L) Increased aPKCζ tyrosine phosphorylation (arrowheads) in Eya1^−/− distal lung epithelium. (Q) Quantitation of Sox9- or SP-B-positive cells, which is expressed as a percentage of all counted MLE15 cells, after interfering with the function of Numb and/or Eya1. Bars carrying the same letter (a,b) are significantly different from the control of the same protein (*P<0.05; ANOVA-Dunnett test). Data are mean±s.e.m. (R) Western blot of Numb, phospho-Numb or phospho-aPKCζ for experiments showing in A-E,G-K,M-P. Bar graphs represent quantified western blot signals (mean±s.e.m.). Bars carrying the same letter (a,b,c) are significantly different from the control of the same protein (**P<0.001; ANOVA-Dunnett test). (S) Endogenous Eya1 was immunoprecipitated from AEC2 cells with a specific Eya1 antibody and western blotting was performed with antibodies specific to different polarity proteins. Anti-Eya1IP of Eya1 siRNA-transfected cells was used as a control. (T) siRNA knockdown of endogenous aPKCζ or different polarity proteins in epithelial cells (48 hours) and subsequent IP for Eya1 and western blot for different polarity proteins. (U) In vitro phosphatase assay using immunopurified wild-type Eya1 or enzymatically inactivated mutant protein (Eya1 D323A) and aPKCζ protein. Graph represents quantified western blot signals normalized to input (n=3; mean±s.e.m., **P<0.001). Scale bars: 50 μm.

Fig. 6.

Inhibition of Notch signaling in Eya1^−/− lung distal epithelium. (A-F) Immunohistochemistry with specific antibodies shows reduced staining of activated-Notch1 (B), Hes1 (D) and Hes5 (F) in E14 Eya1^−/− distal epithelium (arrowheads) compared with control lungs (A,C,E; arrowheads). (G) Western blots show reduction of activated-Notch1 and Hes-5 in E14 Eya1^−/− lungs. (H) Western blot of activated-Notch1 for experiments shown in I-M. Blue numbers in G,H,S represent relative band intensity. (I,J) Immunocytochemistry shows reduced activated-Notch1 expression in MLE-15 after Eya1 knockdown. (K,L) Rescue of endogenous Eya1 function by co-transfection of murine siRNA and murine wild-type or enzymatically inactive mutant Eya1 constructs for 48 hours in MLE15 cells reveals that Notch1 signaling/activity is dependent on Eya1 phosphatase activity. (M) Inhibition of aPKCζ in Eya1 siRNA-transfected MLE15 cells rescues activated-Notch1 expression. (N) Mean fluorescence intensity of Hes1 staining for experiments showing in O-R. *Significantly different from control (P<0.05; ANOVA-Dunnett test). Error bars indicate s.e.m. (O-Q) Immunocytochemistry of MLE-15 cells shows decreased Hes1 expression after Eya1 knockdown (O,P), but increased Hes-1 expression upon Eya1 overexpression (Q; arrowheads). (R) Hes1-positive cells with strong nuclear staining further increase after co-transfection of Numb siRNA and wild-type Eya1 expression vector in MLE-15 cells (arrowheads). (S) Western blot of Hes1 for experiments showing in O-R. (T) Quantitation of Hes1-positive cells, which is expressed as a percentage of all counted MLE15 cells, of the experiments shown in O-R. *Significantly different from control (P<0.05; ANOVA-Dunnett test). Data are mean±s.e.m. Scale bars: 50 μm.

Fig. 7.

Genetic activation of Notch signaling in Eya1^−/− lungs partially rescues epithelial progenitor defects and branching phenotype. (A,B,D-F) External appearance (A,B) and histological analysis (D-F) of control versus Eya1^−/− lungs show reduced epithelial branching and size of Eya1^−/− lungs, which are restored in NICD; Spc-rtTA^+/−-tet(O) Cre^+/−Eya1^−/− compound mutant lungs (A,B,F). The last panel in B is high magnification of the yellow boxed area. (C) Quantitation of Sox9- or SP-B-positive cells, which is expressed as a percentage of all counted MLE15 cells, after interfering with the function of Notch1 and/or Eya1. *Bars carrying the same letter (a,b) are significantly different from the control of the same protein (P<0.05; ANOVA-Dunnett test). Data are mean±s.e.m. (G,H,L,M,Q,R,U,V) Specific antibody staining shows similar polarized localization of polarity proteins (arrowheads) between compound mutant and wild-type lung distal epithelium. Broken lines represent the collagen IV-stained basement membranes. (I-K,N-P) Immunohistochemistry on serial sections shows reduced expression of progenitor markers Sox9 and Id2 in E14.5 Eya1^−/− distal epithelium compared with control lungs (arrowheads). (K,P) Sox9/Id2 expression is substantially rescued in NICD; Spc-rtTA^+/−-tet(O) Cre^+/−Eya1^−/− lungs (arrowheads). (S,T) Western blots show changes of the expression of Sox9, Id2 and SP-C between wild-type, Eya1^−/− and compound mutant lungs. Blue numbers represent relative band intensity. Scale bars: 50 μm.