|Home | About | Journals | Submit | Contact Us | Français|
Morphogenesis and shape of the ocular lens depend on epithelial cell elongation and differentiation into fiber cells, followed by the symmetric and compact organization of fiber cells within an enclosed extracellular matrix-enriched elastic capsule. The cellular mechanisms orchestrating these different events however, remain obscure. We investigated the role of the Rac1 GTPase in these processes by targeted deletion of expression using the conditional gene knockout (cKO) approach. Rac1 cKO mice were derived from two different Cre (Le-Cre and MLR-10) transgenic mice in which lens-specific Cre expression starts at embryonic day 8.75 and 10.5, respectively, in both the lens epithelium and fiber cells. The Le-Cre/Rac1 cKO mice exhibited an early-onset (E12.5) and severe lens phenotype compared to the MLR-10/Rac1 cKO (E15.5) mice. While the Le-Cre/Rac1 cKO lenses displayed delayed primary fiber cell elongation, lenses from both Rac1 cKO strains were characterized by abnormal shape, impaired secondary fiber cell migration, sutural defects and thinning of the posterior capsule which often led to rupture. Lens fiber cell N-cadherin/β-catenin/Rap1/Nectin-based cell-cell junction formation and WAVE-2/Abi-2/Nap1-regulated actin polymerization were impaired in the Rac1 deficient mice. Additionally, the Rac1 cKO lenses were characterized by a shortened epithelial sheet, reduced levels of extracellular matrix (ECM) proteins and increased apoptosis. Taken together, these data uncover the essential role of Rac1 GTPase activity in establishment and maintenance of lens shape, suture formation and capsule integrity, and in fiber cell migration, adhesion and survival, via regulation of actin cytoskeletal dynamics, cell adhesive interactions and ECM turnover.
Vertebrate lens morphogenesis, which is initiated from a single cell type, provides a unique model system to explore the interrelationship between epithelial cell elongation and differentiation into fiber cells, and the intercellular and cell: extracellular matrix adhesive interactions which drive the migration and symmetric packing arrangement of fiber cells (Chow and Lang, 2001; McAvoy, 1980). These key determinants of lens morphogenesis, architecture and ultimately its optical properties are thought to be regulated in part, by various intrinsic and extrinsic factors that control dynamic reorganization of the actin cytoskeleton and cell adhesive interactions (Bassnett et al., 1999; Beebe et al., 2001; Danysh and Duncan, 2009; Menko S., 2004; Quinlan, 2004; Rafferty and Scholz, 1991; Ramaekers et al., 1981; Rao and Maddala, 2006; Straub et al., 2003). Importantly, disruption of actin cytoskeletal organization and cell adhesive interactions have been shown to impair lens epithelial cell elongation and differentiation, underlining the importance of actin cytoskeletal organization and cell-cell interactions in these cellular processes (Beebe and Cerrelli, 1989; Cain et al., 2008; Maddala et al., 2004; Menko S., 2004; Mousa and Trevithick, 1977; Pontoriero et al., 2009; Weber and Menko, 2006a). Although important insights have emerged regarding the external cues which control lens epithelial cell proliferation and differentiation (Chow and Lang, 2001; Lovicu and McAvoy, 2005), little is known about the cellular signaling pathways regulating lens fiber cell actin cytoskeletal dynamics, migration, adhesion and shape (Chen et al., 2008; Cooper et al., 2008; Grove et al., 2004; Maddala et al., 2004; Maddala et al., 2008; Rivera et al., 2009; Weber and Menko, 2006a; Weber and Menko, 2006b; Zelenka, 2004). Therefore, to understand how a functional lens is formed from a single cell type and how this organ attains its distinct architecture, it is important to identify the molecular mechanisms that regulate the cytoarchitecture and adhesive interactions of lens epithelial and fiber cells.
Rho GTPases are GTP-binding proteins that play a pivotal role in regulating actin cytoskeletal organization, cell adhesive interactions, and influence cell polarity, morphogenesis, migration, vesicle trafficking, cell cycle progression and transcriptional activity (Burridge and Wennerberg, 2004; Etienne-Manneville and Hall, 2002; Ridley, 2001a; Ridley, 2011). Specific functions are attributed to the Rho, Rac and Cdc42 GTPases, with Rac GTPase regulating membrane ruffling, protrusive activity and lamellipodial formation via its effects on actin polymerization and cadherin-mediated cell–cell adhesion in various cell types (Burridge and Wennerberg, 2004; Etienne-Manneville and Hall, 2002; Fukata and Kaibuchi, 2001; Ridley, 2001a; Ridley, 2011). Mammals express three highly homologous Rac isoforms, Rac1, Rac2 and Rac3, with the ubiquitously expressed Rac1 being the best studied member of this family (Heasman and Ridley, 2008). Rac1 regulates actin dynamics through two main downstream effector pathways, one of which is comprised of the p21-activated kinase (PAK) which signals to the actin cytoskeleton via LIM kinase and cofilin (Bokoch, 2003; Yang et al., 1998). The second effector pathway involves the verprolin-homologous protein (WAVE), which is a member of the Wiskott-Aldrich syndrome protein (WASP) family (Eden et al., 2002; Ridley, 2011; Stradal et al., 2004; Takenawa and Suetsugu, 2007). Additionally, Rac and its exchange factor Tiam1 regulate cadherin/catenin-based cell-cell adhesion by facilitating translocation of cadherins to the adhesion site and stabilizing the adherens junctions (Fukata and Kaibuchi, 2001; Hordijk et al., 1997; Malliri and Collard, 2003; Takai et al., 2008).
Lens epithelial and fiber cells express Rac1 and Rac 2 GTPases (Maddala et al., 2001; Rao et al., 2004), and Rac GTPase activity is known to be stimulated by various growth factors, PI3 kinase and lipid oxides in lens epithelial cells (Girao et al., 2003; Maddala et al., 2003; Weber and Menko, 2006b). Further, the lens fiber cell plasma membrane has been shown to contain abundant levels of Rac GTPase relative to all other Rho family GTPases (Bassnett et al., 2009). Importantly, transgenic mouse lenses overexpressing Rho GDIα, a negative regulator of Rho GTPases, exhibit defects in development and tissue architecture (Maddala et al., 2008). Additional indirect evidence for the participation of Rac GTPase in lens fiber cell migration, adhesion and differentiation has also been documented in WNT/PCP (Chen et al., 2008) and Abi-2 (Grove et al., 2004) targeted and PI3-kinase (Weber and Menko, 2006b) inhibitor treated lenses.
While there is circumstantial evidence for the participation of Rac GTPase in lens fiber cell differentiation, migration and adhesion, its direct role in these events and lens morphogenesis remains to be defined. To explore this objective, we generated lens specific Rac1 cKO mouse models by suppressing Rac1 expression in distinct regions and at different stages of the developing lens. Collectively, these Rac1 cKO mouse models reveal that Rac1 plays a crucial role in establishment of lens shape, suture formation, fiber cell migration and survival via its effects on actin cytoskeletal dynamics, cell adhesive interactions and ECM turnover.
The Rac1flox, Le-Cre, MLR10-Cre and MLR39-Cre transgenic mice used in this study have been described previously (Ashery-Padan et al., 2000; Chen et al., 2009; Glogauer et al., 2003; Zhao et al., 2004). Briefly, the conditional Rac1flox/flox mice obtained from the Jackson Laboratory contain loxP sites flanking (floxed) exon 1 on a C57BL/6J background. Mice homozygous for this allele are viable, fertile, normal in size and do not display any physical or behavioral abnormalities. The Le-Cre transgenic mice used in this study express Cre recombinase at embryonic day 8.75 under the control of a Pax6 P0 enhancer/promoter, with Cre being expressed in both lens epithelium and fiber cells as well as other surface ectoderm-derived eye structures (Ashery-Padan et al., 2000). The MLR10-Cre mice used in this study express Cre at embryonic day10.5 in both the lens epithelium and fibers under the transcriptional control of the minimal αA-crystallin promoter, which has been modified by the insertion of a Pax6 consensus binding element (Zhao et al., 2004). MLR39-Cre mice start expressing Cre at embryonic day 12.5 mainly in the lens fibers via the minimal αA-crystallin promoter, with little to no activity being present in the lens epithelium (Zhao et al., 2004). MLR10-Cre and MLR39-Cre mice, which were on FVB background, were backcrossed to wild type C57BL/6J mice to generate Cre transgenic mice on the C57BL/6J background. Obtained Le-Cre mice were in the C57BL/6J background. Mice that were homozygous for a Rac1 allele (Rac1flox/flox) were crossed with mice heterozygous for a floxed Rac1 allele (Rac1flox/wild type) and expressing Cre recombinase in the lens to generate Rac1 GTPase conditional knockout (cKO) offspring. Mice carrying the Le-Cre transgene (Fig. 1A), MLR10-Cre transgene (Fig.1B) and the Rac1 floxed alleles were genotyped by PCR using tail DNA. Animals were housed in a pathogen-free vivarium and used in accordance with institutional policies approved by the Duke University Institutional Animal Care & Use Committee (IACUC).
Embryonic heads (E12.5, 14.5, 15.5 and 17.5) and whole eyes (day1) of Rac1 cKO and wild type (WT) littermate mice were fixed in 50 mM cacodylate buffer (pH 7.2) containing 2.5% glutaraldehyde, 4% sucrose and 2 mM CaCl2 for 2 h, and transferred to 10% buffered formalin as described earlier by us (Maddala et al., 2004). The specimens were subsequently embedded in glycol methacrylate, sectioned (2-µm sections) using a JB-4 microtome and stained with hematoxylin and eosin (H&E). Micrographs were captured using a Zeiss Axio Imager equipped with Hamamatsu Orca ER monochrome CCD camera.
Tissue sections derived from both cryosectioning and paraffin embedding were used in immunofluorescence analyses as we described earlier (Maddala, 2011). The following primary antibodies were used with paraffin sections at a 1:200 dilution for 24 h at 4°C: Rac1 (monoclonal, 610651, BD Transduction Laboratories, San Jose, CA), N-cadherin (monoclonal, clone: 3b9, Invitrogen, Camarillo CA), β-catenin (monoclonal, c7082, Sigma-Aldrich, St Louis, MO), Rap1A/Rap1B (polyclonal, 2399, Cell Signaling Technologies, Danvers, MA) and Nectin-1 (polyclonal, SC-28639, Santa Cruz Biotechnology, CA). Air dried tissue cryosections were immunostained with the respective polyclonal antibodies to fibronectin, laminin, collagen IV (gift from Harold Erickson, Duke University), Abi-2 (gift from Ann Marie Pendergast, Duke University), WAVE2 (Millipore, AB4226, Temecula CA) and E-cadherin (Cell Signaling Technologies, 4065, Danvers, MA), all used at a dilution of 1:1000. After incubation with primary antibodies (both paraffin-embedded and cryo-preserved tissue sections), the slides were washed and incubated in the dark for 2 h at room temperature (RT), with either Alexa fluor 488 or 594 conjugated secondary antibodies (Invitrogen). Slides were mounted using Vector mount and nail polish and photographed using a Nikon Eclipse 90i confocal laser scanning microscope. For F-actin staining, the pre-blocked sections were labeled with phalloidin conjugated with tetra rhodamine isothiocyanate (TRITC) (500 ng/ml; Sigma-Aldrich) as described above. All representative immunofluorescence data reported in this study were based on a minimum of 3 to 6 tissue sections derived from three independent specimens per group.
To evaluate changes in levels of Rac1, WAVE-2, integrins (αvβ1 (ABCAM ab75472), αvβ3 (Santa Cruz Biotec, SC-7312), αvβ5 (Santa Cruz Biotec, SC-13588) and β1 (Millipore, 04-1109), phospho-cofilin (C8992 Sigma-Adrich), Abi-2 (ABCAM ab70144), Nap1 (Sigma-Aldrich, N3788), fibronectin, laminin, collagen IV, β-catenin, N-cadherin, Rap1A/ Rap1B, Nectin-1, and β and γ crystallins (gifts from Sam Zigler, Johns Hopkins University) and β-tubulin (Sigma-Aldrich T5293) in Rac1 cKO lenses, 6–8 intact postnatal day 1 (P1) lenses were pooled from both the WT and MLR-10/Rac1cKO mice. Pooled lens samples were homogenized in 10 mM Tris buffer pH 7.4 containing 0.2 mM MgCl2, 5 mM N-ethylmaleimide, 2.0 mM Na3 Vo4, 10 mM NaF, 60 µm phenyl methyl sulfonyl fluoride (PMSF), 0.4 mM iodoacetamide, Protease inhibitor cocktail tablets (complete, Mini, EDTA-free REF 11836170001) and PhosSTOP Phosphatase Inhibitor Cocktail Tablets (REF 04906837001) (1each/10ml buffer, Roche (Basel, Switzerland)). Homogenates were centrifuged at 800×g for 10 min at 4 °C. Protein concentration was estimated in supernatants using the Bio-Rad reagent (Cat. 500-0006). This fraction was centrifuged further at 100,000xg for 1 h at 4 °C and the insoluble pellet derived was resuspended in the tissue homogenization buffer. This centrifugation step was repeated twice and the resultant pellets were pooled and suspended in homogenization buffer containing 5 M urea, 2 M thiourea and 2% CHAPs and used as the membrane enriched fraction. Equal amounts of protein derived from the lens homogenate (800×g supernatant) or membrane enriched insoluble fraction were resolved on SDS-PAGE gels, followed by electrophoretic transfer to nitrocellulose membrane as described earlier (Maddala, 2011). Immunoblots were developed by enhanced chemiluminescence (ECL), and scanned densitometrically using a FOTO DYNE Gel Doc scanner equipped with TL100 software, Densitometry analyses were carried out using ImageJ software (Maddala, 2011).
In situ terminal transferase dUTP nick end labeling (TUNEL) staining was performed using an ApopTag Plus Fluorescein kit (Chemicon, S7111, Temecula, CA) to evaluate and compare apoptotic cell death in lens sections from Rac1 cKO and WT littermate mice as we described earlier (Maddala et al., 2008). Apoptotic cells were scored using a fluorescence microscope (Zeiss Axioplan-II).
For transmission electron microscope-based histological analysis, freshly enucleated eyes from the Le-Cre/Rac1 cKO (E15.5), MLR-10/Rac1 cKO (E17.5), and WT mice were fixed in 10% buffered formalin and processed as we described earlier (Maddala, 2011). Electron microscopic images were captured with a Jeol JEM-1400 transmission electron microscope equipped with an Orius CCD digital camera (JEOL, Tokyo, Japan).
Actin filament fluorescence was quantified using either ImageJ or Metamorph. Signal intensity profiles from phalloidin-labelled lens pit cryosections (one central cryosection for n=3 eyes) were generated using ImageJ. A line was drawn around the lens pit for each example, the signal intensity was then tabulated. The intensity profile data was normalized to the average Hoechst 33258 signal for each lens pit, and then normalized for lens pit size. An average intensity profile was then generated. Similar measurements were done for each of the developmental stages studied. In the case of Metamorph-based measurements, the average fluorescence was quantified as the mean pixel intensity per unit area within each region. The camera measures the intensity of each pixel and gives a grey-scale value between 0 and 4096. Each measurement was corrected for the background fluorescence. Imaging conditions were set constant for both WT and Rac1 mutant sections, and 12-bit images were captured with an image exposure time of 100 milliseconds. The background corrected pixel intensity values were exported into excel sheet and saved for further analysis. A minimum of four measurements from each of six sections derived from three independent specimens were used to calculate average pixel intensities for each developmental stages tested.
Capsule width was measured using the measurement tool in Adobe Photoshop CS3 Extended software. Briefly, the image with scale bar was opened in Photoshop and the scale bar was used as reference to measure the number of pixels. The measurement tool was used to draw a line across the width from one end to the capsule to other end, and the lengths were tabulated. A similar measurement strategy was used to determine the lens epithelial sheet length. Briefly, using scale bar as reference, a line was drawn on 10× sagittal sections, with the start point being an epithelial cell adjacent to one showing signs of elongation and extending to the other end, to an epithelial cell that is adjacent to a cell starting to elongate. For all the measurements six alternate sections were taken from the center of the lens, from three independent mouse lenses.
Where required, the Student’s t-test was performed to determine significance of differences noted between the Rac1 mutant and WT specimens using Sigma plot. Values are represented as Mean ± Standard Error of the Mean (SEM).
To investigate the role of Rac1 GTPase in lens morphogenesis and architecture, we conditionally deleted Rac1 by expressing Cre recombinase under the control of lens regulatory sequences using three different transgenic mouse lines. As described earlier, the Le-Cre line initiates Cre expression by embryonic day 8.75 in the head surface ectoderm that includes the presumptive lens ectoderm (Ashery-Padan et al., 2000). Within the eye, the MLR10-Cre expresses Cre lens-specifically initiating at E10.5 and being active in both the lens epithelium and fibers, while, MLR39-Cre is active only in fiber cells, initiating lens specific expression at E 12.5 (Zhao et al., 2004). Recombination of the conditional Rac1 alleles led to an expected and dramatic reduction (~90%) in Rac1 protein levels in lens total homogenates (800×g supernatant; immunoblot analysis; Fig. 1C; densitometric analysis: Fig. 1D; n=3) and the intact lens (immunofluorescence analysis, Fig 1E) of P1 Le-Cre/Rac1 and MLR10/Rac1 null mice, respectively, relative to Rac1 levels in littermate WT mice. In Fig. 1C, lanes 1 and 2 represent two independent specimens from each group. As shown in Fig.1E, Rac1 expression is suppressed in both lens epithelium (Epi) and fibers (LF) in Le-Cre/Rac1 (lens is indicated with line drawing) and MLR10/Rac1 cKO mice. Compared to the Le-Cre/Rac1 and MRL10/Rac1 cKO lenses, the MLR39/Rac1 cKO mouse lenses exhibited only a 60% reduction in Rac1 protein levels (Fig. 1C & D). As expected, this reduction in Rac1 protein level was specific to fiber cells but not noted in the epithelium (Fig. 1E). Correspondingly, while both Le-Cre/Rac1 and MLR10/Rac1 cKO (weaning and older) mice exhibited severe, bilateral microphthalmia, the MLR39/Rac1 cKO mice had eyes of normal size (data not shown). For immunoblot analyses of Rac1 protein levels in the Le-Cre/Rac1 cKO mice, intact lenses from day 1 old mice, which were found only in a small number of specimens (as shown in Fig. 1E), were used.
Histological analysis (sagittal sections) of developing lenses by hematoxylin and eosin (H&E) staining revealed morphological abnormalities in both Le-Cre/Rac1 and MLR10/Rac1 cKO mice. At E12.5, the developing lenses of Le-Cre/Rac1 cKOs were smaller in size with the primary lens fiber cells displaying delayed elongation compared to littermate WT lenses (Fig.2A, arrows). By E14.5, the lens epithelial sheet length in Le-Cre/Rac1 cKO mice becomes markedly shorter than that of WT lenses. Further, the Le-Cre/Rac1 cKO mice exhibit defective lens fiber cell organization and reduction in lens size (Fig.2A). The lens then becomes progressively and severely deformed through E15.5, E17.5 and day 1, with the lens tissue attaching to the cornea, followed ultimately by complete rupture of the lens posterior capsule and leakage of lens content into the vitreous body (Fig.2B). Considerable variability was noted with respect to severity of lens phenotype within the same line. For example, a small number of eyes derived from the Le-Cre/Rac1 cKO mice at P1 contained intact lens although with much reduced size as shown in Fig. 1E. In contrast to the Le-Cre/Rac1 cKO lenses, the MLR10/Rac1 cKO lenses showed no notable defects prior to E15.5 (not shown). At E15.5, although the lens epithelium and fiber cell elongation are normal, lens size is much smaller and primary fiber cells display abnormal orientation with respect to anterior to posterior polarity, compared to the WT (Fig. 2B). By E17.5, all the fibers were visibly disoriented compared to the WT, and by day1 (P1), the lens appears more cup-shaped as opposed to the spherical shape typical of normal lenses, with fibers exhibiting abnormal migration pattern and ultimately leading to a ruptured posterior capsule in some lenses. In both Le-Cre/Rac1 (E14.5) and MLR10/Rac1 (E17.5 and P1) cKOs however, expression of fiber cell differentiation markers including β/γ-crystallins and aquaporin-0 was found to be comparable to WT lenses by immunohistochemical and immunoblot analyses (not shown). On a relative basis, lens developmental defects were more severe, and displayed an earlier onset (by almost 2.5 to 3 days) in the Le-Cre/Rac1 cKO mice compared to the MLR10/Rac1 cKO mice, consistent with the expected earlier onset of Cre expression in the Le-Cre/Rac1 cKO mice. In contrast, the MLR39/Rac1 cKO lenses (both neonatal and adult) did not show any observable defects in either lens development or architecture (data not shown). Based on these histological observations, we only used the embryonic (Le-Cre/Rac1 and MLR10/Rac1) and P1 (MLR10/Rac1) intact lens specimens from cKO mice along with their respective littermate WT mice for all further comparative analyses.
The lenses derived from both the Le-Cre/Rac1 and MLR10/Rac1 cKO mice exhibit consistent abnormalities in lens shape, suture formation, fiber cell migration pattern and orientation with severe and earlier onset in the Le-Cre/Rac1 cKO lenses compared to the MLR10/Rac1 cKO (Figs. 2 and and3).3). In both Le-Cre/Rac1 and MLR10/Rac1 cKO lenses, differentiating fiber cells starting from embryonic stages (E12.5 and E15.5, respectively) fail to transition to a typical concave orientation from their initial convex orientation. These changes were easily detected in histological sections of MLR10/Rac1 cKO lenses since the initial phenotype is milder in these mice relative to the Le-Cre/Rac1 cKO lenses (Fig. 3). As shown in the schematic drawings (Figs.3A and C), early differentiating WT fibers elongate at the germinative zone, migrating in a convex orientation initially, gaining a concave orientation and migrating toward the lens interior by reorganizing their adhesive interactions posteriorly with basement membrane of the capsule and anteriorly with epithelial cells. This transition in fiber cell morphology occurs after they develop contacts with the lens epithelium through the apical terminals, and with the posterior capsule via their posterior terminals. Subsequently, these fiber cell terminals (both apical and basal) detach from the epithelium and posterior capsule, and the fiber cell terminals at the anterior and posterior poles interact with each other deriving from the opposite poles and form sutures (Kuszak et al., 2004). In the Rac1 cKO lenses (in both Le-Cre and MLR10 Rac1 cKO), the fiber cells fail to migrate, orient normally and gain the expected concave orientation (Fig. 3A–C), and to form the regular lens sutures (Fig. 3D, arrows). The extent of these defects, especially the pattern of fiber cell migration, orientation and sutures appears to vary from lens to lens as depicted in the schematics (Fig.3). These schematics represent only a small number of fiber cells of the lens. Further, in both Le-Cre/Rac1 (E14.5) and MLR10/Rac1 (P1) cKO lenses, epithelial sheet length is significantly smaller compared to the respective WT lenses based on quantitative measurements (Fig. 3E). Unlike in WT lenses, the lens epithelium in Rac1 cKOs does not extend to the bow regions or the germinative zone (Fig.3A–E). It is likely that there is a shift in the germinative zone, with the bow region shifting upwards due to the flattening of the anterior hemisphere in these Rac1 cKO lenses. These different phenotypes collectively lead to the abnormal shape of both Le-Cre/Rac1 and MLR10/Rac1 cKO lenses as depicted in the schematic drawings (Fig.3).
As described above, one of the predominant phenotypes in the Rac1 deficient mice include microphthalmic eyes with reduced lens size and shortened lens epithelial sheet suggesting impairment in cell survival mechanisms (Figs. 1–3). To determine whether the absence of Rac1 GTPase impacted cell survival of both epithelium and fiber cells of the lens and led to size and shape changes in Rac1 deficient lenses as noted in Figs 1–3, we evaluated the status of apoptosis using TUNEL analysis of E14.5 Le-Cre/Rac1, E17.5 MLR10/Rac1 cKO, and corresponding WT eyes. Quantitative analysis of TUNEL positive cells (based on direct counting, n=4 independent samples counted) revealed a significant increase in apoptosis in both epithelium and fiber cells of the Le-Cre/Rac1 (red/orange staining, arrows) and MLR10/Rac1 (green/yellow, arrows) cKO lenses compared to WT lenses (Fig. 4A and B). In Fig 4A, while the Le-Cre/Rac1 specimens were counterstained for cell nuclei (blue fluorescence) with Hoechst, the MLR10/Rac1 specimens were stained with propidium iodide (red fluorescence). The green fluorescence in the Le-Cre/Rac1 null specimens is from the GFP expression in the Le-Cre mice. The diffused green stain in the WT specimen represents nonspecific background.
To determine the effects of the lack of Rac1 expression on lens epithelial and fiber cell actin cytoskeletal organization, we examined F-actin staining of lens cryosections with rhodamine-phalloidin (red fluorescence). Lenses derived from Le-Cre/Rac1 (E12.5, E14.5 and E15.5) and MLR10/Rac1 cKO mice (E15.5, E16.5, E17.5 and P1) from different age groups were used in these analyses. Both Le-Cre/Rac1 and MLR10/Rac1 cKO lenses showed a gradual and progressive reduction in F-actin staining in both epithelium and fiber cells starting from E12.5 and E15.5 in the Le-Cre/Rac1 and MLR10/Rac1 cKO lenses, respectively, as compared to WT lenses (Fig. 5). Further, F-actin fluorescence intensity quantitated (n=4) by pixel counting using ImageJ (A) or Metamorph (B) confirmed a significant reduction in the E14.5 Le-Cre/Rac1 and day 1 MLR10/Rac1 cKO lenses as compared to littermate WT lenses (presented as % change in F-actin signal intensity from WT; Fig. 5C), indicating the disruption of actin cytoskeletal organization and actin polymerization in the Rac1 deficient lenses.
To obtain further insight into the molecular basis for the changes in F-actin distribution in Rac1 deficient lenses, we assessed the expression and distribution profiles of downstream regulatory components of Rac1 signaling pathways that are known to regulate actin nucleation and branching. Expression of WAVE-2, Abi-2, and Nap1 was assessed by immunofluorescence and immunoblot analysis of either E15.5 Le-Cre/Rac1, E17.5 MLR10-Cre/Rac1 (immuno-fluorescence), or P1 MLR10/Rac1 cKO (immunoblot) lenses. Immunofluorescence staining of both WAVE-2 and Abi-2 were found to be markedly decreased in both the epithelium (Epi) and fiber cells (LF) in Rac1 cKO lenses compared to WT lenses (Fig.6A). Additionally, immunoblot analysis of P1 MLR10/Rac1 cKO intact lens homogenates (800×g supernatants) confirmed a significant decrease (n=3, pooled specimens; protein levels normalized to β-tubulin) in levels of WAVE-2, Abi-2 and Nap1 proteins compared to littermate WT lenses (Fig. 6B & C). In addition to these Rac1 effector molecules, we also evaluated for changes in phosphorylation status of cofilin in the P1 MLR10-Cre/Rac1 cKO lens homogenates (800×g supernatants). Cofilin, a well characterized downstream target of Rac GTPase signaling with actin filament severing activity and role in cell migration, is negatively regulated by phosphorylation at serine 3 (Arber et al., 1998; Huang et al., 2006; Yang et al., 1998). Therefore, changes in cofilin phosphorylation were quantified by immunoblot analysis using a phospho-specific cofilin antibody. Intriguingly, the levels of phospho-cofilin were found to be significantly elevated (by ~ 48%, n=3, pooled specimens) in the Rac1 deficient lenses compared to WT specimens (Fig. 6B & C).
To explore the effects of absence of Rac1 GTPase on lens epithelial and fiber cell-cell interactions, we examined the distribution pattern of E-cadherin in the lens epithelium, and β-catenin, Rap1A/B, Nectin-1 and N-cadherin in the lens fiber mass by immunofluorescence analyses using Le-Cre/Rac1 (E15.5) and MLR10/Rac1 cKO (E17.5 and P1) lenses. Lens sagittal and equatorial sections were analyzed for lens epithelial and fiber cell-cell junctions, respectively. The lens epithelium of both Le-Cre/Rac1 and MLR10/Rac1 cKO specimens showed a notable reduction in E-cadherin staining at cell-cell junctions as compared to WT specimens (Fig. 7A & B). Insets in Fig. 7B depict magnified (2.5 X) areas of the central epithelium (indicated with a box). While N-cadherin, β-catenin, Nectin-1 and Rap1A/B were distributed discretely to the cell-cell junctions of the hexagonal fiber cells (LF) in the WT specimens, these proteins did not localize to the cell-cell junctions in Rac1 deficient lenses (in E17.5 and P1 MLR10/Rac1, Fig. 7C, data shown for the P1 specimens). Additionally, the changes in cell adhesive interactions in the Rac1 deficient lenses were associated with a significant decrease in N-cadherin, β-catenin, and Nectin-1 protein levels based on immunoblot analysis of the membrane enriched insoluble fraction derived from the P1 Rac1 cKOs as compared to WT lenses (Fig. 7D & E; n=3, pooled specimens, normalized to β-tubulin protein levels of 800×g lens supernatant), indicating impaired formation and stability of adherens junctions in the Rac1 deficient lenses. Due to the severe phenotype associated with the Le-Cre/Rac1 cKOs, we could not obtain suitable equatorial plane paraffin sections for these analyses.
To explore whether the changes in the adherens junction complexes described above in the Rac1 deficient lenses influence the hexagonal geometry and symmetric organization of fiber cells, we performed transmission electron microscope-based histological analysis using equatorial sections of lenses from E15.5 Le-Cre/Rac1, E17.5 MLR10/Rac1 ckO lenses, and their respective WT lenses. As shown in Fig. 7F, the WT lens fibers showed an expected hexagonal shape with long (arrows) and short arms (arrow heads) with close packing among the adjacent fiber cells. In contrast, the Rac1-deficient lens specimens exhibited disorganized fiber cells with disrupted hexagonal cell shape (Fig.7F). These observations are based on a minimum of 3 independent analyses.
Since a ruptured posterior capsule was one of the consistent phenotypes noted in the Rac1 deficient lenses, we examined the possible influence of Rac1 deficiency on the organization and distribution of extracellular matrix proteins (ECM) of the lens by immunofluorescence analysis. Antibodies against collagen IV, laminin and fibronectin were used to analyze E15.5 Le-Cre/Rac1 and P1 MLR10/Rac1 ckO lenses, alongside littermate WT specimens. As shown in Fig. 8A, all three antibodies revealed specific and intense staining distributing only to the lens capsule. Additionally, based on these staining patterns, in both Le-Cre/Rac1 and MLR10/Rac1 cKO lenses, we observed a consistent thinning of the posterior capsule more so than the anterior capsule, relative to WT controls. Furthermore, measurements of posterior central capsule thickness revealed a significant (by 60%) decrease in Rac1-deficient lenses compared to WT controls (Fig.8B, n=6). Additionally, Rac-deficient lenses exhibited marked reduction in immunofluorescence staining for all three ECM molecule distribution assessed (fibronectin, laminin and collagen IV), relative to WT specimens (Fig. 8A). Based on these observations, we also analyzed protein levels of collagen IV, laminin and fibronectin, in lysates (800×g supernatants) of the P1 MLR10/Rac1 cKO lenses by immunoblot analysis (Fig. 8C). Consistent with the immunofluorescence data, there were significant decreases in the levels of collagen IV, laminin and fibronectin in Rac1 deficient lenses compared to WT specimens based on densitometric analysis (Fig. 8D, n=3, pooled specimens, normalized to β-tubulin). In Fig. 8C, lanes 1 and 2 represent two independent specimens from each group.
Since the ECM organization of lens capsule, fiber cell adhesion and migration were severely impaired in the Rac1-deficient lenses (Figs.3 and and7),7), we then evaluated the levels of integrin proteins, which function as ECM receptors. Immunoblot analysis of lens lysates (800×g supernatants) derived from the P1 MLR10/Rac1 cKO mice showed a significant decrease in the levels of αvβ1, αvβ3, αvβ5 and β1 integrins compared to WT lenses (Fig.8E and F, n=3, pooled specimens), indicative of down regulation or degradation in association with the decreases in ECM noted in the absence of Rac1 in cKO lenses (Fig. 8).
Based on these different observations, we also evaluated the organization and interaction of fiber cell posterior terminals with the lens capsule by TEM using sagittal sections of both Le-Cre/Rac1 (E15.5) and MLR10/Rac1 (E17.5) cKO lenses along with respective WT specimens. Interestingly, both the Le-Cre/Rac1 and MRL10/Rac1 cKO lens specimens exhibited distinct morphological abnormalities at the interface of fiber cell terminals and the lens capsule basement membrane (Fig.8G). While the WT specimens showed well organized and spread out, web-like membrane/cytoplasmic extensions (Fig.8G, arrows) attaching fiber cell terminals to the lens capsule basement membrane, these web-like protrusive structures were smaller, broken and detached from the lens capsule in the Rac1 deficient specimens (Fig. 8G, arrow heads). It is not clear whether these structures are membrane protrusions of the fiber cell terminals enriched in basal membrane complexes (BMCs) known to be composed of cell adhesive proteins, β1-integrin, myosin and actin (Bassnett et al., 1999). However, there seem to be definite abnormalities in the attachment of the membrane extensions to the capsule in the Rac1-deficient lens, providing morphological evidence for the disruption in the membrane extensions and adhesive interactions between the basement membrane of capsule and the fiber cell posterior terminals. These conclusions were based on data derived from four independent specimens from each group.
Using a tissue-specific conditional gene deletion approach, we demonstrate an essential role for Rac1 GTPase during lens morphogenesis and maturation. We show that Rac1 GTPase regulates crucial aspects of lens morphogenesis and cytoarchitecture including epithelial cell survival, fiber cell elongation, migration and adhesion. This study also reveals the critical role for Rac1 GTPase in establishment of lens shape, suture formation, capsule basement membrane integrity, and fiber cell hexagonal geometry, which are important determinants of lens function.
Rac1 is expressed uniformly throughout the mouse lens including in the epithelium and fiber cells. Lack of Rac1 expression in the developing lens resulted in abnormal lens shape and reduction in size (Figs.2 & 3), with the prominent phenotype being manifested only when Rac1 expression is deleted from both lens epithelium and fibers, as evidenced by the lack of a lens phenotype in MLR39/Rac1 cKO mice in which Rac1 expression is deleted only in the fiber cells (Figs. 1). This observation demonstrates the importance of the epithelium and its interaction with fiber cells in establishment of lens shape and growth. Further, this phenotype appears to be a direct result of shortening of the epithelial sheet, increased apoptosis, and the abnormal fiber cell migration and orientation noted in the Rac1 deficient lenses (Fig.3 & 4). While the molecular basis for the shortened epithelial sheet in Rac1 deleted lenses however, is not entirely clear at present, we speculate that the increase in epithelial cell apoptosis (Fig. 4) which is consistent with the known anti-apoptotic activity of Rac (Murga et al., 2002; Nishida et al., 1999) could be partly responsible for this phenotype. Additionally, the lens epithelium in Rac1 null mice exhibited abnormal cell cycle progression based on in vivo BrdU incorporation profile (Fig. S1) in association with defective E-cadherin-based cell-cell interactions, suggesting compromised cell survival and proliferation in the absence of Rac1 GTPase in lens. However, it is important to point out that the MLR39 Cre transgenic mice start expressing the Cre protein from E12.5, unlike the Le-Cre (at E8.75) and the MLR10 (at E10.5) mice. Further, the promoters driving Cre expression are different in the three Cre transgenic mice used in this study (Ashery-Padan et al., 2000; Zhao et al., 2004). Therefore, the possibility that lack of phenotype in the MLR39/Rac1 mutant mice is partly attributable to the differences in amount and onset of Cre expression in fibers cells cannot be completely ruled out. It is possible that manifestation of the abnormal lens shape and fiber cell orientation phenotype result from the absence of Rac1 GTPase expression during early embryonic lens development, perhaps before E12.5.
Rac1 GTPase has been demonstrated to control directional cell migration in various cell types via regulating cell adhesive interactions and the formation of cell membrane protrusions and lamellipodia at the leading edges, by acting downstream of growth factor receptors, integrins, PI3 kinase and its lipid products (Burridge and Wennerberg, 2004; Raftopoulou and Hall, 2004; Ridley, 2001b; Ridley et al., 2003). However, the role of Rac in cell migration in the context of organogenesis is obscure (Heasman and Ridley, 2008; Wang and Zheng, 2007), and this study provides direct evidence for the involvement of Rac1 in fiber cell migration and organization in the intact lens, as shown in Fig. 3. The mouse lens expresses both Rac1 and Rac2 (Rao et al., 2004) and it is evident from the lens phenotype and other lens changes reported in this study that Rac1 plays a non-redundant role in lens. Rac1 GTPase regulates actin polymerization during lamellipodial extension and membrane ruffle formation by activating the actin–nucleating ARP2/3 complex through WASP-family verprolin-homologous proteins (WAVE) (Eden et al., 2002; Ridley, 2011; Stradal et al., 2004; Takenawa and Suetsugu, 2007; Yamazaki et al., 2003). Importantly, WAVE-2 has been demonstrated to regulate ARP2/3 mediated actin assembly and branching in a wide range of systems, acting in concert with Sra1, Nap1, Abi and HSPC300, which exist as a complex with WAVE-2 (Eden et al., 2002; Ridley, 2011; Stradal et al., 2004; Takenawa and Suetsugu, 2007). Rac1 GTPase interacts with and activates the WAVE-2 complex through the Sra1 subunit (Kunda et al., 2003; Pollitt and Insall, 2009; Stradal et al., 2004; Takenawa and Suetsugu, 2007). WAVE-2 and Abi-2 were found to be abundantly expressed in lens fibers relative to the epithelium and other ocular tissues based on distribution analysis (Fig.6A). Moreover, distribution analysis of actin filaments, WAVE-2 and its associated proteins Abi-2 and Nap1 demonstrated reduced actin filament staining and downregulation of WAVE-2, Abi-2 and Nap1, in the lens epithelium and fiber cells of Rac1 cKO mice, implying impairment in actin nucleation possibly via the ARP2/3 complex. These observations, taken together with the earlier report on the lens phenotype of defective fiber cell migration and cell-cell interactions in the Abi-2 null mice (Grove et al., 2004), confirm the importance of Rac1 GTPase regulated WAVE-2 complex activity in controlling fiber cell migration. Abi-1 and Abi-2 both have been found to exist as part of the WAVE protein complex and to regulate actin dynamics (Eden et al., 2002; Soderling et al., 2002).
Stimulation of PAK kinase and LIM kinase by Rac1 GTPase results in the phosphorylation-dependent inactivation of cofilin (Arber et al., 1998; Huang et al., 2006; Yang et al., 1998), and leads to decreased actin filament severing and cell movement under normal circumstances. Intriguingly, deletion of Rac1 expression did not result in the expected decrease in PAK and LIM kinase-mediated cofilin phosphorylation, with Rac1 KO lenses exhibiting an increase in cofilin phosphorylation (Fig. 6B and C). This observation suggests the involvement of a Rac1-independent pathway in the inhibition of cofilin activity in the Rac1 KO lenses. Such a pathway is likely upregulated as an adaptive response to the Rac1 deficiency-induced alterations in actin dynamics and cell mobility in the Rac1 KO lenses. We have not analyzed the status of Slingshot phosphatase activity, which is known to dephosphorylate and control cofilin activity (Huang et al., 2006), in Rac1 cKO lenses.
It is important to note that, in addition to involvement of Rac GTPase downstream signaling pathways in fiber cell actin dynamics and migration discussed above, the fact that Rac1 deficient lenses and Sfrp2 (secreted WNT antagonist) overexpressing lenses (Chen et al., 2008) appear to share a striking number of abnormalities including changes in shape, shortened epithelial sheet length, and abnormal fiber cell migration and cytoskeletal organization, further suggests that Rac1 is also an important component of the WNT/PCP signaling pathway in the lens.
Another prominent phenotype in the Rac1 deficient lenses was disorganization of fiber cells, defective suture formation and disruption of the prismoid morphology of fiber cells. Lens suture formation depends on fiber cell migration, spreading and turnover of cell-cell adhesive interactions at the interface of apical fiber tips and epithelium, the basal tips of fiber cells and basement membrane of capsule (Kuszak, 2004), and between the lateral membranes of fiber cells. As discussed above, in addition to defective fiber cell migration, N-cadherin/β-catenin and Rap1/Nectin mediated cell-cell interactions were disrupted in the Rac1 deficient lenses. E-cadherin-based cell-cell junctions regulate Rac activity (Fukuyama et al., 2006; Kawakatsu et al., 2002; Sander et al., 1998; Yap and Kovacs, 2003) and activated Rac in turn regulates actin cytoskeletal interactions with the adherens junctions (Baum and Georgiou, 2011; Fukata and Kaibuchi, 2001), which are important for the stabilization of the adherens junctions (Baum and Georgiou, 2011; Fukata and Kaibuchi, 2001; Halbleib and Nelson, 2006; Hordijk et al., 1997; Takai et al., 2008). Similarly, Rap1/Nectin, which act upstream of Rac GTPase, collectively regulate adherens junctions in various other cell types (Kooistra et al., 2007; Takai et al., 2008). The disruption of fiber cell organization and hexagonal shape in Rac1 deficient lenses as assessed by TEM (Fig.7F) therefore implies that the deficits in N-cadherin/β-catenin and Rap1/Nectin mediated cell-cell interactions noted in these lenses likely compromise lens suture formation and cytoarchitecture. It is not clear however, why and how the levels of Rap1 and Nectin-1 are decreased in Rac1 KO lenses. It is unlikely that these changes are secondary to the primary insult since we not only used intact lenses for the analyses but also noted that the Rac1 deficient lenses derived from the MLR10-Cre mice at P1 did not exhibit the typical secondary changes such as extensive vacuole accumulation and degeneration of fiber cells, as noted for many types of cataractous lenses (Chen et al., 2008; Cooper et al., 2008; Simirskii et al., 2007).
The lens capsule is a basement membrane that surrounds and serves to control the shape, elastic and mechanical properties of the ocular lens. Additionally, the ECM components of the capsule play a crucial role in lens epithelial cell proliferation, differentiation, migration and cell adhesion via engagement of integrin receptors (Danysh and Duncan, 2009; Menko S., 2004). Interestingly, the posterior capsule was frequently ruptured very early on (embryonic and neonatal) in the Le-Cre/Rac1 and MLR10/Rac1 cKOs, a phenotype which was found to be associated with decreased content of specific ECM proteins (Fig. 8). ECM components are produced and secreted by the lens epithelium and fiber cells (Danysh and Duncan, 2009). It is possible that the absence of Rac1 GTPase activity results in impaired transcriptional activity and MAP kinase activation, (Coso et al., 1995; Etienne-Manneville and Hall, 2002; Hill et al., 1995; Minden et al., 1995) which in turn may be associated with, or partly responsible for, the altered basement membrane composition in the Rac1 deficient lenses. Indeed Rac GTPase has been shown to regulate the activity of several different matrix metalloproteases (Engers et al., 2001; Kheradmand et al., 1998). Therefore, the absence of Rac1 in lens appears to have a direct influence on ECM turnover. Importantly, expression of various integrins is downregulated at the level of protein in these Rac1 deficient lenses. Integrins regulate lens epithelial proliferation, survival and differentiation (Walker and Menko, 2009) and mediate cell adhesive interactions through fiber cell basal membrane complex and through interactions with epithelial cells (Bassnett et al., 1999; Simirskii et al., 2007; Walker and Menko, 2009). ECM and integrins regulate Rac GTPase activity and cell adhesive interactions (del Pozo et al., 2000; Hotchin and Hall, 1995; Price et al., 1998) and conversely, Rac1 and Rap1 control integrin activation and cell adhesion in certain cell types (D'Souza-Schorey et al., 1998; Vielkind et al., 2005). Therefore, the changes in the basement membrane protein organization and composition in the Rac1 deficient mice very likely play a key role in the compromised cell survival and defective fiber cell migration and suture formation noted in the Rac1 cKO lenses. Interestingly, although the absolute requirement of Rac1 in cell migration has remained somewhat questionable (Wells et al., 2004), Rac1 deficient fiber cells in intact lenses exhibit defective cell membrane/cytoplasmic extensions required for cell adhesion and migration, similar to the effects noted in Schwann cells in the peripheral nerves (Benninger et al., 2007).
In conclusion, this study uncovers a crucial role for Rac1 GTPase activity in lens morphogenesis and architecture via regulation of epithelial cell survival, cell cycle progression, fiber cell migration, cell adhesion, and ECM synthesis. Importantly, Rac1 deficiency significantly impacts lens growth, shape, capsule thickness, cell survival and cytoarchitecture.
We thank Ying Hao for TEM histology and Corey Morris for genotyping and schematic illustrations. We thank Sam Zigler, Jr for providing the aquaporin-0 and crystallin antibodies, Herald Erickson for antibodies to fibronectin, laminin and collagen IV and Ann Marie Pendergast for Abi-2 antibody. This work was supported by the National Institutes of Health grants to P.V. Rao (EY12201 and EY018590), R.A. Lang (EY17848), and M. L. Robinson (EY12995) and by a National Eye Institute Core Grant for Vision Research (P30-EY5722).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.