Numb Expression and Asymmetric versus Symmetric Cell Division in Distal Embryonic Lung Epithelium

Abstract

Proper balance between self-renewal and differentiation of lung-specific progenitors is absolutely required for normal lung morphogenesis/regeneration. Therefore, understanding the behavior of lung epithelial stem/progenitor cells could identify innovative solutions for restoring normal lung morphogenesis and/or regeneration. The Notch inhibitor Numb is a key determinant of asymmetric or symmetric cell division and hence cell fate. Yet Numb proximal-distal expression pattern and symmetric versus asymmetric division are uncharacterized during lung epithelial development. Herein, the authors find that the cell fate determinant Numb is highly expressed and asymmetrically distributed at the apical side of distal epithelial progenitors and segregated to one daughter cell in most mitotic cells. Knocking down Numb in MLE15 epithelial cells significantly increased the number of cells expressing the progenitor cell markers Sox9/Id2. Furthermore, cadherin hole analysis revealed that most distal epithelial stem/progenitor cells in embryonic lungs divide asymmetrically; with their cleavage, planes are predicted to bypass the cadherin hole, resulting in asymmetric distribution of the cadherin hole to the daughter cells. These novel findings provide evidence for asymmetric cell division in distal epithelial stem/progenitor cells of embryonic lungs and a framework for future translationally oriented studies in this area.

Control of epithelial progenitor cells is critical for correct development of the lung (Warburton et al. 2000; Warburton 2008). Lethal defects of gas diffusion capacity such as the common congenital forms of lung hypoplasia and bronchopulmonary dysplasia as well as the limited capacity of the lung to recover from them could be explained by a significant deficiency of stem/progenitor cells (Warburton et al. 2008, 2010; Shi and Warburton 2009). Therefore, understanding how to achieve a proper balance between self-renewal and differentiation of lung-specific progenitors could provide innovative solutions to restoring normal lung morphogenesis and possibly regeneration of the gas diffusion surface. Asymmetric cell division (ACD) during development is indeed crucial for balancing self-renewal and differentiation as well as correct spatial and temporal specification of cell lineages in epithelia (Knoblich 2001). Yet, ACD has not been analyzed before in embryonic lung epithelial progenitor cells.

The cell fate determinant Numb is a key determinant of asymmetric versus symmetric cell division (reviewed in Knoblich 2001, 2010). Both ACD and Numb expression as well as function during development have been comparatively well studied in Drosophila and the mammalian nervous system (Betschinger and Knoblich 2004), but they are entirely uncharacterized in the lung.

Herein we report on the proximal-distal Numb expression pattern and ACD in stem/progenitor cells at the distal epithelial tips of murine lungs at E14.5, which was chosen as the most informative developmental stage to analyze ACD behavior of distal embryonic lung epithelium because of intense Numb expression and relatively high cell proliferation at that time in gestation.

Materials and Methods

Animals

Breeding of wild-type mice and isolation of lungs were carried out as described before by us (El-Hashash et al. 2011). All animals were handled and maintained in accordance with institutional guidelines and the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Antibody Staining and Western Blot Analysis

Both antibody staining on paraffin sections or fixed MLE-15 cells and Western blot analysis were performed using standard protocols as described (El-Hashash et al. 2005, 2011). Antibodies used in this study are described in Table 1. The specificity of Numb antibody was confirmed by Western blot on E14.5 lung proteins, which shows a proper protein band of 79 kilodaltons (kDa) that corresponds to Numb protein size (El-Hashash and Warburton 2011). Antibody staining was performed in triplicate and analyzed using a Leica confocal or fluorescence microscope (Leica Mikrosysteme Vertrieb GmbH; Wetzlar, Germany). DAPI (4,′6-diamidino-2-phenylindole) was used for nuclear staining. An irrelevant antibody of the same isotype (Invitrogen; Carlsbad, CA) acted as a negative control, which showed no staining for the different experimental controls reported in this study (data not shown).

Table 1.

Primary Antibodies Used in This Study

Antibody	Description	Dilution	Antigen	Source
Sox9	Rabbit polyclonal	1:100 (IHC, ICC)	Sox9	Santa Cruz Biotechnology, Santa Cruz, California
Id2	Rabbit polyclonal	1:50	Id2	CalBioreagents, Foster City, California
Numb	Rabbit polyclonal	1:50 (IHC, ICC)	Numb	Abcam, San Francisco, California
α-Adaptin	Mouse monoclonal	1:50 (IHC)	α-Adaptin	BD Biosciences, San Jose, California
E-cadherin	Mouse monoclonal	1:100 (IHC)	E-cadherin	BD Biosciences
Activated Notch1	Rabbit polyclonal	1:100 (ICC)	Activated Notch1	Abcam, San Francisco, California
Hes-1	Rabbit polyclonal	1:100 (ICC)	Hes-1	Santa Cruz Biotechnology
Collagen type IV	Rabbit polyclonal	1:100 (IHC)	Collagen type IV	Abcam, Cambridge, Massachusetts

ICC, immunocytochemistry; IHC, immunohistochemistry/immunoperoxidase.

Cell Culture and Transfection

Lung epithelial cells MLE15, a gift from Jeffrey Whitsett (Cincinnati University), were grown in culture and treated as well as processed for transfection as described (Tefft et al. 2002; El-Hashash et al. 2011). Briefly for transfection, MLE15 cells were seeded in eight chamber slides (BD Biosciences; San Jose, CA) at 50% confluency. The next day, they were transfected with small interfering RNA (siRNA) sequences targeted to Numb or control siRNA (Operon Biotechnologies; Huntsville, AL) at a final concentration of 80 nM using Lipofectamine LTX (Invitrogen; Carlsbad, CA). Fresh medium was added 6 hr later. Cells were fixed with 4% paraformaldehyde 3 days after transfection. The knockdown efficiency was analyzed by immunostaining of targeted protein. Also, we used an expression vector encoding a VP16 fusion protein, and the transfection efficiency was further monitored by fluorescence staining using anti-VP16 antibody (Santa Cruz Biotechnology; Santa Cruz, CA).

Statistical Analysis

Data are presented as means ± SD. Statistical significance between cells treated with control siRNA or Numb siRNA was calculated using Student’s t-test. Analyses were done using GraphPad Prism software (GraphPad Software, San Diego, CA; http://www.graphpad.com), as described before (El-Hashash et al. 2011).

Results

In Drosophila and mammalian epithelial cells, ACD is mediated by preferential segregation of intrinsic cell fate determinants (CFDs) (e.g., Numb) into one of two sibling daughter cells. CFDs are asymmetrically localized in dividing cells and define the axis of polarity that will determine the orientation of the apical-basal cell division plane. This allows a rapid switch from proliferation, wherein two similar daughter cells are born, to diversification, wherein different-shaped daughter cells are generated (Betschinger and Knoblich 2004). Numb protein, a Notch signaling inhibitor, is expressed uniformly in the cytoplasm in interphase but is localized asymmetrically in dividing cells. Hence, Numb is segregated to only one daughter cell, enabling this cell to adopt a different fate from that of its sibling. The cell receiving high levels of Numb suppresses extrinsic Notch signaling and differentiates, whereas the cell with low Numb levels maintains high Notch activity and thus has a stem cell fate (Frise et al. 1996; Guo et al. 1996).

Herein, we detected very weak, if any, Numb expression in the distal lung epithelial tips, which contain the epithelial progenitor pool that expresses Sox9, Id2, and N-myc (Liu and Hogan 2002; Okubo et al. 2005; Rawlins 2008; Rawlins et al. 2009) at E11.5–12.5 (Fig. 1A), whereas at E14.5, strong Numb signals were observed in the distal rather than proximal lung epithelium (Fig. 1B, C). Closer inspection revealed that Numb was asymmetrically distributed and highly concentrated at the apical side of distal epithelial progenitor cells, which also expressed epithelial progenitor markers Id2/Sox9 (Fig. 1E, and data not shown for Id2), with little or no staining at the basal pole (Fig. 1D). A similar expression pattern was found at the apical cell side for the Numb-associated signal α-adaptin (Fig. 1F, G), which is an endocytic protein that is required for Numb-mediated ACD (Berdnik et al. 2002). No staining was evident with irrelevant primary antibodies for above tested proteins (data not shown).

Figure 1.

(A–D) Immunofluorescence with specific antibodies shows very weak Numb staining in the distal epithelial tips of E12.5 lungs (A) but strong polarized Numb signals at the apical cell side in the distal tips of E14.5 lungs (C, D; arrowheads). Numb polarized staining is particularly more intense in distal (B, C; double arrowheads) than proximal (C; arrows) lung epithelium. C represents high magnification of the boxed area in B. Thin dashed lines in A and D represent the collagen IV–stained basement membrane that is shown in H. (E) Sox9 antibody staining shows strong Sox9 expression in the distal epithelial tips. (F, G) Immunofluorescence with α-adaptin antibody shows strong polarized α-adaptin signals at the apical cell side of distal epithelial tips of E14.5 lungs (arrowheads). G represents high magnification of boxed area in F. Scale bars = 50 µm.

Next, to test the function of Numb in lung epithelium, we performed siRNA-mediated Numb knockdown in MLE15 epithelial cells, which showed intense expression of asymmetrically localized Numb in addition to the expression of both epithelial progenitor cell markers Id2/Sox9 and differentiation marker SP-C similar to lung distal epithelium at E14.5 (Figs. 1E, 2H and El-Hashash el al. 2011). As shown in Figure 2 (B–D, F–G), the number of MLE15 cells that express the epithelial progenitor cell markers Sox9/Id2 increased fivefold upon Numb knockdown in vitro (n=878; p<0.05; Student’s t-test). Similarly, a significant increase in the number of MLE15 cells with strong Notch1 activity was observed in Numb siRNA-transfected cells (Fig. 2D). No change was observed in MLE15 negative controls (i.e., Lipofectamine without any siRNA) for these siRNA transfection experiments (Fig. 2A, E). Numb knockdown was verified by immunocytochemistry with a specific anti-Numb antibody (Fig. 2H, I). Together, these data suggest a conserved function for Numb in controlling Notch1 signaling and cell proliferation in epithelial lung cells.

Figure 2.

(A–C, E–G) Immunocytochemistry shows increased Sox9- (C) or Id2-positive cells (G) in MLE-15 cells transfected with Numb siRNA for 3 days in culture. Note no changes in the expression of Sox9 (A) or Id2 (E) in MLE15 cells of Lipofectamine controls for Numb siRNA transfection experiments. (D) Quantitation of the experiments shown in B–C and F–G for Sox9/Id2 as well as for cells stained positively for active Notch1 signaling (n = 878). Bars carrying the same letter (a, b) are significantly different from one another (*p<0.05; Student’s t-test). (H, I) Verification of Numb knockdown in MLE15 cells by immunocytochemistry with a specific anti-Numb antibody. Note strong polarized Numb signal in H (arrows). Scale bars = 50 µm.

In many different organs, epithelial cells characteristically show apical-basal polarity. They also have a distinct shape, such that only a subtle deviation in cleavage plane from the normal orientation suffices to result in an asymmetric rather than a symmetric distribution of their apical plasma membrane and adjacent adherent junctions to the daughter cells (Nelson 2003; Kosodo et al. 2004). Upon immunostaining for E-cadherin, which is a constituent of the apico-lateral junctional complex and lateral epithelial cell plasma membrane (Woods et al. 1997), the “cadherin hole” in the plasma membrane of mitotic epithelial cells appears as a relatively small, unstained segment of the cell surface (Kosodo et al. 2004 and Fig. 3A, B, D, E). Symmetric versus asymmetric distribution of the plasma membrane to daughter cells can thus be predicted from the orientation of the cleavage plane relative to the cadherin hole (Kosodo et al. 2004).

Figure 3.

(A, B, D, E) Analysis of cleavage plane orientation of E14.5 distal epithelial progenitors relative to the cadherin hole (yellow bar marks location, with yellow arrows marking the two edges of the cadherin hole). The cleavage plane (thick dashed white lines) is deduced from the orientation of the sister chromatids at anaphase/telophase, and either bisects (A, B) or bypasses (D, E) the cadherin hole. (C) Quantitation of the cleavage plan orientation relative to the cadherin hole for experiments shown in A, B, D, E, which is expressed as a percentage of all divisions (n = 48). (F) E-cadherin immunostaining of E14.5 distal lung epithelial tip progenitors. Thick dashed lines represent the cleavage plane of mitotic cells. Thin dashed lines in F and K represent the collagen IV–stained basement membrane. (G) Quantitation of the cleavage plane orientation relative to the basement membrane (thin dashed line in F), as shown in F, which is expressed as a percentage of all divisions (n = 48). (H, I, K, L) Low (H) and high magnification (I, K, L) of E14.5 lung epithelial cells stained with Numb or Sox9 antibody shows asymmetric nuclear localization of Numb in one daughter cell of a mitotic distal lung epithelial cell at late M phase of the cell cycle (K with inset shows low magnification of the same cell). (I, L) In some late mitotic cells, Numb is not detectable in both daughters/chromatids (I), which express Sox9 (L). (J) Quantification of data shown in I and K for Numb expression by one daughter or no Numb expression in both daughter cells. This is expressed as a percentage of all divisions in the distal epithelium (n=42). Scale bars: 50 µm. (M) Model of Numb inheritance in mitotic distal lung epithelial cells. The difference in Numb (green) expression levels between two daughter cells mediates asymmetric cell division (ACD), whereas lack of Numb inheritance by both daughters will allow them to execute the stem cell self-renewal program by maintaining Notch1 activity and thus allowing symmetric cell division.

To determine the distribution of the cadherin hole upon division of progenitor cells of distal lung epithelial tips at E14.5, we deduced the orientation of the cleavage plane in mitotic cells in anaphase/telophase, at which time the final orientation of the mitotic spindle and the cleavage plane is set in eukaryotic cells (Gonczy and Hyman 1996, Fig. 3A, B, D, E). Cadherin hole analysis then revealed that the alteration in cleavage plane orientation in distal epithelium progenitors correlated well with the distribution of the cell membrane/cadherin hole during cell division. As shown in Figure 3A–E, in distal epithelial progenitors, 94% of the cleavage planes were predicted to bypass the cadherin hole, resulting in asymmetric distribution of the cadherin hole to the daughter cells, whereas 6% of the cleavage plane orientations were predicted to bisect the cadherin hole, resulting in a symmetric distribution to both daughter cells (n=48 mitotic cells). These data suggest that most progenitors in distal embryonic lung epithelial tips divide asymmetrically.

We recently reported that most cell divisions occur perpendicular to the basement membrane in the distal lung epithelial progenitors at E14–14.5, as judged by pericentrin-stained centrosome orientation relative to the basement membrane (El-Hashash and Warburton 2011). Perpen-dicular cell division is strictly correlated with ACD in different mammalian epithelial cells because they undergo asymmetric division by shifting the spindle orientation from parallel to perpendicular (Lechler and Fuchs 2005). To further confirm this in distal lung epithelial progenitors, we evaluated the orientation of the cleavage plane to better assess cell division in distal epithelial tips of the lung. To do this, we used LGN (kindly provided by J. Blumer, Medical University of South Carolina, Charleston, SC) and mouse Inscuteable (mInsc; kindly provided by J. Knoblich, Institute of Molecular Biotechnology, Vienna, Austria) proteins, which localize to the apical cell side only during apical-basal division (Lechler and Fuchs 2005), as markers of perpendicularly dividing cells. We found that most cleavage planes were horizontal or nearly horizontal (i.e., oblique) to the basement membrane in mitotic epithelial progenitors (91%; Fig. 3F, G; n=48, and El-Hashash and Warburton 2011 for LGN/mInsc expression pattern). This further suggests that most progenitor cells at distal epithelial tips divide perpendicularly and asymmetrically. No staining was evident with the irrelevant primary antibody (data not shown).

During ACD, the Notch signaling inhibitor Numb is localized asymmetrically in the nucleus or cytoplasm (depending on the mitotic phase) in dividing cells. Hence, Numb is segregated to only one daughter cell, enabling this cell to adopt a different fate from that of its sibling. The cell receiving high Numb levels suppresses extrinsic Notch signaling and differentiates, whereas the cell with low Numb levels maintains high Notch1 activity and thus has a stem cell fate (Frise et al. 1996; Guo et al. 1996; Juven-Gershon et al. 1998; Yan et al. 2008). Therefore, to further confirm that most cell divisions occur asymmetrically in the lung epithelium and that Numb is preferentially expressed by one daughter cell in mitotic lung epithelial progenitors, we tracked Numb versus Sox9 expression as markers of cells acquiring a differentiated cell fate versus stem/progenitor cell fate, respectively, in daughter cells of mitotic distal lung epithelial cells at the mitotic (M) phase of the cell cycle. As shown in Figure 3K, Numb was asymmetrically expressed in the nucleus of one daughter of dividing cells at M phase of the cell cycle, in agreement with reports in other cell types that show nuclear Numb expression in mitotic cells (Juven-Gershon et al. 1998; Yan et al. 2008). In some mitotic cells, Numb was not detected in the nuclei of both daughter cells, which showed intensive expression of the progenitor cell marker Sox9 (Fig. 3I, L). Quantification of these data strongly suggest ACD occurs in the lung epithelium, as evident by asymmetric expression of Numb in one daughter of most dividing cells at M phase of the cell cycle (92%), whereas about 8% of mitotic cells (n=42) had no detectable expression of Numb in both daughter cells that will together acquire a progenitor cell fate, as indicated by Sox9 expression (Fig. 3J, L, M).

Taken together, we conclude that the more perpendicular/asymmetric cell division is, the more likely it is to segregate Numb preferentially to one daughter cell in mitotic lung epithelial progenitors. These results, to our knowledge, provide the first evidence for ACD in distal epithelial stem/progenitor cells of embryonic lungs.

Discussion

Most recently, we have reported that distal lung epithelial progenitors are polarized, using Numb as a polarity marker (El-Hashash and Warburton 2011). However, no specific or detailed report on the temporospatial and proximal-distal expression pattern of Numb during lung development has appeared yet to our knowledge. Identifying the temporospatial and proximal-distal expression pattern of Numb is significant for the understanding of asymmetric cell division in both proximal and distal lung epithelial progenitor cells. We have also reported that lung distal epithelial progenitors are polarized with perpendicular rather than parallel divisions (El-Hashash and Warburton 2011). However, perpendicular cell division is not a direct evidence for asymmetric division because some parallel-dividing cells also divide asymmetrically. Thus, without the cleavage plane and cadherin hole analyses described herein, one could not conclude that distal epithelial progenitors cells divide asymmetrically in the embryonic lung. In this work, we describe a proximal-distal expression pattern of Numb and analyze, for the first time, cleavage plan orientation and asymmetric cell division in embryonic lung distal epithelial progenitor cells.

Several recent reports suggest that undifferentiated epithelial progenitors undergo multiple division-linked cell fate decisions (symmetric and asymmetric) in the lung, which lead to an apparently homogeneous expansion of the progenitor cell population (Lu et al. 2008; Rawlins 2008). No reports about ACD in the embryonic lung have appeared as yet to our knowledge. Our current and recently published (El-Hashash and Warburton 2011) studies provide several lines of evidence that suggest that asymmetric cell divisions are common in embryonic distal lung epithelial cell populations, which contain progenitor cells (Okubo et al. 2005; Rawlins et al. 2009). For instance, the cleavage plane orientations are predicted to bypass the cadherin hole, resulting in asymmetric distribution of the cadherin hole to the daughter cells in most distal epithelial progenitor cells (this study; Fig. 3). Moreover, our report that most of the distal epithelial cells have apically localized Par, LGN, NuMA, and mInsc polarity proteins, with mitotic spindles aligned perpendicular to the basement membrane and a characteristic asymmetric segregation/inheritance of Numb (El-Hashash and Warburton 2011), provides further evidence that they are dividing asymmetrically. Indeed, a strict correlation exists between ACD and the apical localization of polarity proteins Par/LGN/NuMA/mInsc, which control spindle orientation in mammalian epithelial mitotic cells (Lechler and Fuchs 2005), perpendicular alignment of mitotic spindles, and asymmetric Numb segregation in different Drosophila and mammalian epithelial cell types (Cayouette and Raff 2002, 2003; Haydar et al. 2003; Noctor et al. 2004; Lechler and Fuchs 2005). Further studies are needed to determine asymmetric versus symmetric division in embryonic lung distal epithelial progenitors.

During embryonic development, multipotent epithelial stem/progenitor cells localize within the distal lung epithelial buds/airways (Rawlins and Hogan 2006; Rawlins 2008; Rawlins et al. 2009). ACD likely mediates the balance between lung epithelial stem/progenitor cell maintenance and differentiating cell populations at distal epithelial tips. This ACD-mediated balance is critical for the long-term maintenance of tissue self-renewal during development and in diseases. For instance, congenital lung hypoplasia and bronchopulmonary dysplasia, wherein a significant deficiency of stem/progenitor cells probably occurs, are common features of human prematurity and/or lung injury and are thus major public health problems in human infancy. Therefore, proper balance between self-renewal and differentiation of lung-specific progenitors, which is mediated by ACD, is absolutely required for normal lung morphogenesis and regeneration. Thus, controlled outgrowth and branching of the epithelial tubes generate a sufficiently large gas diffusion surface to sustain life. Developmental defects in this smooth progression lead to defective differentiation and postnatal respiratory distress (Warburton et al. 2008, 2010).

Therefore, understanding the behavior of lung epithelial progenitors, as described herein, could identify innovative solutions to restoring normal lung morphogenesis. Moreover, characterizing ACD (as in this study) and identifying novel factors/mechanisms regulating ACD and behavior of lung epithelial progenitors, as key mechanisms that regulate the balance between lung progenitor cell self-renewal and differentiation, can help to identify novel targets for the prevention and rescue therapy of fatal lung disease in infancy and childhood and for lung regeneration after injury. In addition, identification of the molecular programs regulating the balance between the proliferation and differentiation of endogenous lung-specific progenitor cells is critical for developing techniques that harness the ability of these cells to regenerate diseased and damaged lungs. Our data provide a conceptual framework for future translationally orientated and mechanistic studies in this area.