Ectopic Activation of Wnt/β-Catenin Signaling in Lens Fiber Cells Results in Cataract Formation and Aberrant Fiber Cell Differentiation

The Wnt/β-catenin signaling pathway controls many processes during development, including cell proliferation, cell differentiation and tissue homeostasis, and its aberrant regulation has been linked to various pathologies. In this study we investigated the effect of ectopic activation of Wnt/β-catenin signaling during lens fiber cell differentiation. To activate Wnt/β-catenin signaling in lens fiber cells, the transgenic mouse referred to as αA-CLEF was generated, in which the transactivation domain of β-catenin was fused to the DNA-binding protein LEF1, and expression of the transgene was controlled by αA-crystallin promoter. Constitutive activation of Wnt/β-catenin signaling in lens fiber cells of αA-CLEF mice resulted in abnormal and delayed fiber cell differentiation. Moreover, adult αA-CLEF mice developed cataract, microphthalmia and manifested downregulated levels of γ-crystallins in lenses. We provide evidence of aberrant expression of cell cycle regulators in embryonic lenses of αA-CLEF transgenic mice resulting in the delay in cell cycle exit and in the shift of fiber cell differentiation to the central fiber cell compartment. Our results indicate that precise regulation of the Wnt/β-catenin signaling activity during later stages of lens development is essential for proper lens fiber cell differentiation and lens transparency.


Introduction
Mouse lens morphogenesis begins with the formation of a lens placode in the head surface ectoderm in response to inductive signals from several tissues including the underlying optic vesicle. The lens placode subsequently invaginates and gives rise to the lens pit and then to the lens vesicle composed of epithelial cells, forming a single layer of cuboidal epithelium on the anterior lens pole, and fiber cells which form the remainder of the lens. Primary fiber cells differentiate from cells comprising the posterior part of the lens vesicle. Secondary fiber cells are continually generated in the equatorial region of the lens. There, the transitional zone is formed, and multiple signaling molecules converge on lens precursors, cells exit the cell cycle, and begin to elongate and differentiate into secondary fiber cells [1]. Differentiation of lens fiber cells is characterized by a change of the cell shape and by accumulation of fiber cell-specific proteins, band c-crystallins, which are the major structural, water soluble lens proteins responsible for the optical properties of the lens [2]. Transcription factors Pax6, Prox1, c-Maf, and Sox1 are essential regulators of fiber cell differentiation, since they regulate expression of crystallins and their loss results in arrest of fiber cell differentiation [3][4][5][6][7][8][9][10]. Cell cycle inhibitors p27 Kip1 and p57 Kip2 [4] are required in lens cells for the cell cycle exit at the transitional zone and for terminal differentiation and elongation of lens fiber cells, as they inhibit cyclin-dependent kinases involved in G1/S transition [11]. Fiber cell differentiation is completed by the loss of intracellular organelles and nuclei, which ensures the lens transparency [12].
Beyond transcription factors, several growth factors have been implicated as regulators of lens fiber cell differentiation. Fibroblast growth factors (FGFs) [13][14][15], bone morphogenetic protein family (BMP) [16,17], and members of the transforming growth factor-b family (TGFb) [17][18][19] have been considered as key regulators of this process. The Wnt/b-catenin (canonical) signaling pathway represents one of the key mechanisms controlling cell-fate decisions both during embryonic development and in adult tissues (reviewed in [20]). Therefore, it is not surprising that the Wnt/bcatenin signaling pathway has also been implicated in the regulation of various stages of lens development [21][22][23][24][25][26], including lens fiber and epithelial cell differentiation [21,26,27]. In the absence of Wnt ligand, b-catenin is bound to the destruction complex, phosphorylated on serine-threonine residues encoded by exon 3 (N-terminal part of b-catenin), and thus targeting the protein for proteasomal degradation. Upon binding of Wnt ligand to the Frizzled/LRP receptor complex, the destruction complex is shifted to the cell membrane and disintegrated. As a result, bcatenin accumulates in the cytoplasm and enters the nucleus, where it acts as a transcriptional co-activator and in cooperation with transcription factors of the TCF/LEF family initiates transcription of the Wnt/b-catenin target genes. b-catenin has a dual role in the cell: besides its critical role as a transcriptional coactivator of the Wnt/b-catenin signaling pathway, it also functions as a structural protein on the cell membranes in cadherinmediated cell adhesion at adherens junctions [28]. Both of these bcatenin functions, transcriptional and cytoskeletal, have been shown to play a critical role in proper lens development and differentiation [21][22][23][24][25][26]. In early stages of the lens development, Wnt/b-catenin signaling has an important role in patterning the surface ectoderm. It has to be inhibited in the region of periocular ectoderm, where the lens develops, and b-catenin stabilization at E9.5 in this region results in suppression of the lens fate [23,24]. On the contrary, active Wnt/b-catenin signaling is required in the surrounding head ectoderm. Inhibition of Wnt/b-catenin signaling at E9.5 results in formation of ectopic lenses and in defective lens cell adhesion, but has no impact on lens fate marker acquisition [22,25]. In later stages of lens development and especially during differentiation of lens cells, the requirement for Wnt/b-catenin signaling was demonstrated in mice with null mutation of Lrp6 that exhibit incompletely formed lens epithelium [26]. Conditional deletion of b-catenin in the whole lens, or only in differentiating fiber cells at E12.5, showed that b-catenin was required in the lens epithelium and during early fiber differentiation [21]. Loss of b-catenin in the whole lens resulted in the loss of lens epithelial marker expression (E-cadherin, Pax6) and abnormal expression of differentiating fiber cell markers (c-Maf, p57 Kip2 ) in the lens epithelium, epithelial cell cycle arrest at G1-S, and premature cell cycle exit [21]. Fiber cell differentiation was also affected, with poor cell elongation, decreased b-crystallin expression, and improper formation of adhesion junctions in differentiating fiber cells [21]. b-catenin stabilization, using the Cre-loxP system in the whole lens from E12.5 onward, resulted in the ocular phenotype comprising increased progression of epithelial cells through the cell cycle and maintenance of epithelial phenotype in the fiber compartment [27]. In contrast, mice in which b-catenin stabilization was induced exclusively in lens fiber cells by embryonic day 12.5 lacked the ocular phenotype [27]. Here we show, however, that ectopic activation of Wnt/b-catenin signaling in lens fiber cells using a transgenic mouse model results in aberrant lens fiber cell differentiation and cataract formation.

Ethics Statement
Housing of animals and in vivo experiments were performed after approval by the Animal Care Committee of the Institute of Molecular Genetics (study ID#174/2010) and in compliance with national and institutional guidelines (ID#12135/2010-17210).

Generation and Screening of Transgenic Mice
To generate aA-CLEF construct the fragment containing the coding sequence for the human C-terminal part of b-catenin (aa 696-781), followed by the mouse Lef1 coding sequence and HAtag, was liberated from the pCMV-catCLEF1 vector [29]. This fragment was inserted into the pACP3 vector containing the aAcrystallin promoter, the SV40 small T antigen intron and polyadenylation sequence. To generate the delb-CAT construct the fragment containing the coding sequence for HA-tagged Nterminally truncated human b-catenin (aa 91-781) was inserted into pACP3. The linearized transgenic inserts were injected into the male pronuclei of fertilized FVB/N oocytes by the Genetic Engineering Facility of National Eye Institute. aACLEF transgenic mice were further maintained on the C57Bl/6 background and identified by PCR analysis of DNA obtained by tail biopsy with the following primers: 59 GAGGGCTGGAACGCTAGCTCA 39, and 59 GTCCATACCCAAGGCCTCCTG 39. Transgenic delb-CAT mice were identified by PCR analysis of DNA obtained by tail biopsy using primers derived from the SV40 small T intron.

Tissue Collections and Histology
Mouse embryos were staged by designation the noon of the day when the vaginal plug was observed as embryonic day 0.5 (E0.5). Embryos of the required developmental stage were harvested in cold PBS, fixed in 4% PFA/PBS (w/v) for 2-4 h on ice. Tissue was cryoprotected by overnight incubation in 30% sucrose (w/w) at 4uC, embedded and frozen in OCT (Tissue-Tek, Sakura Finetek). Horizontal cryosections of 12 mm were prepared, stored at 220uC and used up to two weeks.
For paraffin sections, E16.5 embryos were fixed in 8% PFA/ PBS (w/v) overnight on ice, processed and embedded in paraffin and sectioned 7 mm, stored at 4uC and used up to one month.

Immunofluorescence
The cryosections were air-dried for at least 30 min, washed with PBS, and permeabilized with PBS/0,1% Triton X-100 (PBT) for 15 min prior to blocking. For some antibodies the antigen retrieval was performed: slides were immersed in 10 mM citrate buffer (pH 6.0) in coplin jars and boiled for 5-15 min, then let cool down to room temperature. The paraffin sections were air-dried for at least 30 min, deparaffinized and rehydrated. Epitope retrieval was performed in 10 mM citrate buffer (pH 6.0) for 15 min in pressure cooker. All sections were blocked for 30 min in 10% BSA/PBT Total RNA Isolation, cDNA Synthesis and RT-PCR Total RNA was isolated from dissected eyes or lenses with TrizolH Reagent (Invitrogen), contaminating DNA was eliminated by DNAse I digestion and RNA was repurified with RNeasy Micro kit (Qiagen). Random-primed cDNA was generated from 500 ng of total RNA using SuperScript VILO cDNA Synthesis kit (Invitrogen).
Primers 59 GTGAAGGAACCT-TACTTCTGTGGTG 39, and 59 GTCCTTGGGGTCTTCTACCTTTCTC 39, derived from the SV40 small T intron, were used for RT-PCR detection of transgenic CLEF mRNA. Total RNA was isolated from two lenses of one E16.5 embryo, or newborn with TrizolH Reagent. At least three different E16.5 embryos were used for total RNA isolation. Random-primed cDNA was generated from 200 ng of total RNA using SuperScript VILO cDNA Synthesis kit (Invitrogen) and two independent synthesis of cDNA were performed from one total RNA sample.

Quantitative RT-PCR (qRT-PCR)
qRT-PCR were run in the LightCyclerH 480 Instrument (Roche) using LightCyclerH 480 DNA SYBR Green I Master (Roche) according to the standard manufacturer's protocol; typically, 5 ml reaction mixture was used. PCR reactions were performed in triplicate for each primer set of primers, with three different cDNA (biological triplicate). Control reactions (containing corresponding aliquots from cDNA synthesis reactions that were performed without reverse transcriptase; minus RT controls) were run in parallel. Crossing point (Cp) values were calculated by LightCyclerH 480 Software (Roche) using the second-derivate maximum algorithm. The average Cp values of all biological and technical replicates were normalized by Cp values of housekeeping genes. Statistical significance of the change in mRNA expression was calculated by a two-tailed Student t-test in Microsoft Excel. Finally, the change in mRNA expression of aA-CLEF lenses is presented as the ratio aA-CLEF/wild-type in log with base 2 scale. Primer sequences are listed in Table S1.

Western Blot Analysis
Lenses were dissected from 8-10-month-old wild-type and aA-CLEF mice, mechanically homogenized in total lens protein lysis buffer [30] or in soluble lens protein lysis buffer [31], sonicated (4610 sec, 30%), centrifuged at 14 0006 g/4uC/8 min, the supernatant was collected, and the protein concentration was determined using BCA TM Protein Assay Kit and NanoDrop ND1000. One mg (for b-crystallin detection) and 15 mg (for aand c-crystallin detection) of protein were denatured in 26 SDS, boiled and subsequently separated by SDS-PAGE, and transferred onto nitrocellulose membranes. Membranes were blocked with 5% non-fat milk/PBT for 2 h at RT, incubated with primary antibody (anti-a-crystallin and b-crystallin, Samuel Ziegler; c- crystallin, Hisato Kondoh; HRP-conjugated anti-b-actin, Sigma; anti-b-catenin, Sigma) diluted in 5% milk/PBT at 4uC overnight, washed 3 times in PBT and incubated with HRP-conjugated species-specific secondary antibody for 2 h (HRP-conjugated antirabbit swine secondary antibody, Dakocytomation), and washed 3 times in PBT. The signal was detected by enhanced chemiluminescence (ECL) kit (Amersham Bioscience). To compare protein levels, quantification of the blots was performed using Aida Image Analyser Software. Around 15 lenses dissected from E16.5 wild-type or aA-CLEF embryos were pooled and homogenized in 10 ml of total lens protein lysis buffer; 50 mg of protein was loaded on SDS-PAGE. CLEF transgenic protein was detected with primary anti-HA-High Affinity rat primary antibody and subse-quently with HRP-conjugated anti-rat goat secondary antibody (Pierce).

b-catenin Stabilization in Lens Fiber Cells Results in Cataract Development in Adulthood
Activation by Wnt/b-catenin signaling leads to inhibition of the GSK3 kinase that causes phosphorylation of b-catenin's N terminus and targets its cytoplasmic pool for ubiquitin-mediated degradation. N-terminal truncation of b-catenin leading to increased levels of b-catenin was therefore previously used to activate Wnt/b-catenin signaling in the mouse skin [32]. To  investigate the effect of ectopic activation of Wnt/b-catenin signaling in eye lens we have initially generated a transgenic mouse, referred to as delb-CAT mouse, expressing N-terminally truncated b-catenin (aa 91-781) in lens fiber cells under the control of aA-crystallin promoter (2342/+49), (Fig. 1A). Presence of transgenic delb-CAT protein in adult mutant lenses was detected by anti-HA tag antibody on western blot (Fig. 1B). N-terminally truncated b-catenin was detected with anti-b-catenin antibody confirming that the delb-CAT protein level was significantly higher than the expression of endogenous b-catenin in adult delb-CAT lenses (Fig. 1C). Upregulated expression of Axin2, known Wnt/b-catenin signaling target gene, in newborn lenses indicated the Wnt/b-catenin signaling activation in delb-CAT lenses. Several independent founders of delb-CAT mice developed cataract starting from 3 weeks of age (Fig. 1E-L'). These results suggested that expression of stabilized b-catenin is not compatible with the normal development and homeostasis of the lens. However, as b-catenin stabilization can result either in transcriptional activation of Wnt/b-catenin target genes or in accumulation of b-catenin on the cell membranes [28], it was not obvious which of the b-catenin functions was responsible for the cataract development in adult delb-CAT mice. In order to eliminate the impact of a cytoskeletal function of b-catenin on cataract formation, we generated another transgenic mouse where only the transcriptional function of b-catenin was manipulated to

Wnt/b-catenin Signaling Activation in Lens Fiber Cells Results in Cataract Development and Disrupted Lens Morphology
To activate Wnt/b-catenin signaling in lens fiber cells from the stage E12.0 onward, the transgenic construct designated aA-CLEF was generated. Constitutive activation of Wnt/b-catenin signaling was achieved by expression of transgenic protein CLEF containing the C-terminal activation domain of b-catenin fused to the amino terminus of the full-length protein Lef1. An HA-tag was fused to the carboxy terminus of Lef1 to simplify detection of the transgene expression. An identical fusion protein has previously been used to mimic activation of Wnt/b-catenin signaling [29,33]. Since we have used the CLEF fusion protein to activate Wnt/bcatenin signaling, we can exclude the possibility that the phenotypes observed here are due to the cytoskeletal function of b-catenin.
Fiber cell-specific expression of the CLEF fusion protein was achieved by aA-crystallin promoter (2342/+49) [34], (Fig. 2A). Three independent founders harboring the transgene were generated. All three founders developed cataracts in adulthood (data not shown) and their lenses appeared to have disrupted morphology (Fig. S1). Two independent transgenic lines (MB06-05 and MB06-12) were established and further examined. As animals of both lines manifested the same ocular phenotype, the MB06-5 line was used for all of the subsequent analyses and was further referred to as aA-CLEF. aA-CLEF mice were maintained and analyzed as heterozygotes on the C57Bl/6 background. In order to establish the onset of CLEF mRNA expression in the lens, a sensitive RT-PCR was initially performed. Since aA-crystallin promoter is active from embryonic day 12.5 [34], we examined the transgene expression in eyes starting from developmental stage E11.5. Only weak expression of the transgene was detected at E12.5, whereas from E13.5 onward the CLEF mRNA expression was easily detected in eyes of aA-CLEF transgenics (Fig. 2B). The presence of the CLEF fusion protein in lenses of aA-CLEF transgenic mice was confirmed at E16.5 using anti-HA-tag antibody (Fig. 2C). In order to detect the CLEF protein within the embryonic lens, anti-Lef1 antibody was used for immunofluorescent detection. Under the conditions used, Lef1 expression was not detected in wild-type lenses (Fig. 2D). As expected, given the known selectivity of aA-crystallin promoter, the CLEF protein was clearly present in the nuclei of fiber cells already in transitional zone and in the central fiber cell compartment of aA-CLEF transgenic mice (Fig. 2E). Since endogenous mouse Lef1 mRNA was not upregulated in aA-CLEF transgenics (unpublished data), the signal obtained by immunofluorescence represents the exogenous CLEF protein. Wnt/b-catenin signaling activation was confirmed by upregulated mRNA expression of known Wnt/ b-catenin target genes -Axin2, Nkd1, cyclin D1 and cyclin D2 in E16.5 aA-CLEF lenses (Fig. 2F). Adult transgenic aA-CLEF mice developed cataract and microphthalmia (Fig. 2G, H). Cataract was present already in lenses of two-week-old (P14) transgenic aA-CLEF mice and lenses were smaller compared to lenses from wildtype littermates at that stage (Fig. 2 I, J). Histological examination of eyes from the F0 adult aA-CLEF animals confirmed smaller size of the whole eye and exhibited a number of morphological alterations. Detailed view of the transitional region of mutant lens revealed random distribution of fiber cell nuclei throughout the fiber cell region, and some regions of the lens contained vacuoles (Fig. 2 K-N). Additional histological analysis of mutant lenses of subsequent generations confirmed that the ocular phenotype was transmitted (data not shown).

aA-CLEF Lenses are Characterized by Decreased Levels of c-crystallins
It is well established that band c-crystallins are present in very high concentrations in the lens, and in altered conditions can precipitate within the lens cells to form the insoluble fraction, thus contributing to cataract formation (reviewed in [35]). We therefore examined whether the crystallin protein level was not changed in adult cataractous aA-CLEF lenses. Total and soluble protein extracts were prepared from lenses of 8-10-month-old mice, when the cataract was obvious in aA-CLEF lenses. At least two independent samples were prepared from wild-type lenses and from aA-CLEF lenses, and a-, band c-crystallin proteins were detected by specific antibodies on western blot. Less c-crystallin was detected in lenses of adult aA-CLEF mice in the total lens protein extract (Fig. 3A, B), as well as in the soluble lens protein extract (Fig. 3C, D). In order to distinguish whether lower levels of c-crystallin protein in adult aA-CLEF lenses were caused by aggregation (and therefore impossible to be detected either in total or in soluble protein fraction), or by a lower level of transcription of c-crystallin coding genes, qRT-PCR was performed. Quantitative RT-PCR analysis revealed that the expression of cA-, cC-, cD-, and cE/F-crystallin mRNA was significantly decreased in adult aA-CLEF lenses in comparison to wild-type lenses of equal age (Fig. 3E). As the presence of CLEF mRNA and CLEF fusion protein was confirmed in E13.5 and E16.5 aA-CLEF lenses, respectively, the expression of crystallin genes was examined at these stages of lens development. First, the expression of band ccrystallin in developing lens at E13.5 and at E16.5 in wild-type and aA-CLEF lenses was analyzed by imunofluorescence. No apparent difference was observed in the expression levels of band ccrystallin in E13.5 or in E16.5 aA-CLEF lenses compared to wildtype lenses (Fig. S2). However, qRT-PCR analysis showed that the expression of c-crystallin mRNA was significantly decreased in aA-CLEF lenses in comparison to wild-type lenses already at E16.5 (Fig. 3F). In summary, we provide evidence of decreased levels of c-crystallins in aA-CLEF lenses beginning at E16.5 and persisting until adulthood.

Fiber Cell Junctions are not Affected in Embryonic aA-CLEF Lenses
In order to exclude the impact of aA-CLEF transgene on normal fiber cell junctions, localization of tight junction protein zona occludens-1 (ZO-1) and localization of adherens junction molecule N-cadherin was examined at E13.5 and E16.5 in wildtype and aA-CLEF lenses. ZO-1 has been previously shown to localize to the apical membrane of epithelial cells (epithelium-fiber cell interface) within the developing lens [36]. Strong expression of ZO-1 was observed on the apical aspect of the lens epithelial cell layer in wild-type as well as in aA-CLEF lenses at E13.5 and E16.5 (Fig. S3A-D). N-cadherin has been shown to be expressed in both lens epithelial cells and fiber cells during lens differentiation [37], the same expression pattern was observed in aA-CLEF lenses at E13.5 and E16.5 (Fig. S3E-H). Because no significant difference was observed in the localization of ZO-1 and N-cadherin in aA-CLEF lenses, we assumed that fiber cell junctions were not affected in aA-CLEF lenses.

Epithelial-mesenchymal Transition and Apoptosis are not Induced in Embryonic aA-CLEF Lenses
As active Wnt/b-catenin signaling has been asociated with epithelial-mesenchymal transition (EMT), presence of a-SMA (asmooth muscle actin), marker of EMT, was examined in E13.5 and E16.5 aA-CLEF lenses. a-SMA is normaly expressed within the overlying iris tissue and muscles of eyelids. We did not observe any positive staining for a-SMA in E13.5 or in E16.5 aA-CLEF lenses (Fig. S4A-D) indicating the absence of EMT during the process of abnormal lens differentiation in aA-CLEF transgenic mice.
To determine whether Wnt/b-catenin signaling activation leads to apoptosis in aA-CLEF lens fiber cells, we examined the levels of cleaved caspase 3 (marker of apoptotic cells)-positive cells by immunofluorescence. No cleaved caspase 3-positive nuclei of fiber cells were detected in E13.5 and E16.5 aA-CLEF lenses, similary to E13.5 and E16.5 wild-type lenses (Fig. S4E-H).

Expression of Transcriptional Regulators in Embryonic Lenses of aA-CLEF Transgenic Mice
To address the sequence of events leading to cataract formation in adulthood, we examined the localization of fiber cell nuclei and expression of key lens regulatory proteins in developing embryonic lenses. We have chosen E13.5 as the starting point since at this developmental stage CLEF mRNA begins to be highly expressed. Despite abnormal fiber cell nuclei localization in adult aA-CLEF lenses, DAPI staining showed that the distribution of fiber cell nuclei was unaltered at E13.5 (Fig. 4C, D). Transcription factors c-Maf, Sox1 and Prox1 are involved in proper fiber cell differentiation and crystallin expression [4,5,38] and at E13.5 are predominantly expressed in fiber cell nuclei of wild-type lenses [4,5,38]. Similar expression was observed in aA-CLEF lenses (Fig. 4A, B, E, F, I, J). In contrast, Pax6 expression remains high in the anterior lens epithelium and lens equator, and weakens as the cells undergo differentiation at the transitional zone [39]. Expression of Foxe3, a known target gene of Pax6, specifies lens epithelial cells [40,41]. The expression of Foxe3 in aA-CLEF lenses appeared to be unaltered (Fig. 4K, L), whereas Pax6 expression was stronger in the nuclei of fiber cells in the fiber cell compartment compared to wild-type (Fig. 4G, H).
To further explore changes in the lens fiber cell differentiation, we investigated the expression of fiber cell differentiation regulators and the position of fiber cell nuclei in a later stage of lens development, at E16.5. In lenses of E16.5-old controls, the nuclei of the equatorial transitional zone, where lens epithelial cells undergo differentiation, were organized in a characteristic bow pattern (Fig. 5A, C). By contrast, in aA-CLEF lenses, nuclei of differentiating fiber cells were disorganized in the entire lens fiber region (Fig. 5B, D). The same disorganization of fiber cell nuclei showed staining of Prox1-positive nuclei; moreover, Prox1 reactivity appeared more intensive in the nuclei in the central part of the fiber compartment of aA-CLEF lenses compared to wild-type lenses (Fig. 5O, P). Based on immunostaining of CLEF transgenic protein (Fig. 5E, F) it is apparent that the fiber cells from the transitional zone towards the center of the lens were affected by the expression of transgene at E16.5 aA-CLEF lenses. At E16.5 the expression of transcription factors regulating the differentiation of fiber cells was almost lost or weaker in the central part of the fiber cell region of wild-type lenses (Fig. 5I, G, K). However, in aA-CLEF lenses, Pax6, Sox1 and c-Maf strongly positive nuclei were detected in the fiber cell region of the lens (Fig. 5J, H, L). In contrast, the expression of Foxe3 in E16.5 aA-CLEF lenses appeared to be unaltered (Fig. 5M, N). Combined, our data indicate abnormal expression of some key transcription factors during lens fiber cell differentiation in aA-CLEF transgenic mice.

Activated Wnt/b-catenin Signaling in Lens Fiber Cells Affects the Cell Cycle Exit of Differentiating Fiber Cells in aA-CLEF Lenses
Since active Wnt/b-catenin signaling is known to regulate the cell cycle [42], and as the persistence of fiber cell differentiation markers in the fiber cell compartment of transgenic aA-CLEF lenses suggested delayed differentiation of fiber cells, we examined the expression of cell cycle promoting factors cyclin D1, cyclin D2 and the expression of negative regulators of cell cycle p27 Kip1 and p57 Kip2 [43]. In wild-type E13.5 lenses, cyclin D1 and cyclin D2 were strongly expressed in transitional and germinative zones, and cyclin D1 was also weakly expressed in some of the lens epithelial cells (Fig. 6A, E). Virtually the same expression pattern was observed later at E16.5 (Fig. 6C, G). Cyclin-dependent kinase inhibitors p27 Kip1 and p57 Kip2 were strongly expressed in transitional zone and further in fiber cells in the fiber cell compartment of E13.5 wild-type lenses (Fig. 6I, M), whereas later at E16.5 the expression was limited to the transitional zone (Fig. 6K, O). Immunofluorescent analysis of E13.5 aA-CLEF lenses has shown that both cyclin D1 and cyclin D2 were abnormally expressed already at this developmental stage, as they persisted in the nuclei of fiber cells in the fiber cell compartment (Fig. 6B, F). The expression of cyclin-dependent kinase inhibitors p27 Kip1 and p57 Kip2 , which regulate cell cycle exit, appeared unaltered at E13.5 (Fig. 6J, N). In E16.5 aA-CLEF lenses expression of cyclin D1, cyclin D2 as well as p27 Kip1 and p57 Kip2 persisted in the fiber compartment (Fig. 6D, H, L, P) in contrast to wild-type lenses, where the expression of these cell cycle regulators was reduced to the lens epithelium and to the transitional zone (Fig. 6C, G, K, O). Our data provide evidence of aberrant expression of cell cycle regulators in embryonic lenses of aA-CLEF transgenic mice. Combined, our observations suggest that active Wnt/b-catenin signaling in fiber cells results in a delay in the cell cycle exit and a shift of the fiber cell differentiation to the central fiber cell compartment.

Discussion
The Wnt/b-catenin signaling pathway controls many processes during development, including cell proliferation, cell differentiation and tissue homeostasis, and its aberrant regulation has been linked to various diseases in man (reviewed in [20]). In this study we have investigated the effect of ectopic activation of Wnt/bcatenin signaling during lens fiber cell differentiation. A previous study has shown that activation of Wnt/b-catenin signaling using stabilization of b-catenin in lens fiber cells by means of the Cre/ loxP system did not manifest an obvious lens phenotype [27]. In contrast, we show here that Wnt/b-catenin activation using a transgenic approach results in a conspicuous lens phenotype, which includes aberrant fiber cell differentiation detectable as early as at E13.5 during embryogenesis, and cataract development and microphthalmia accompanied by low levels of c-crystallin proteins and mRNA. The main results of both studies are summarized in Figure 7. The reasons for an apparent difference in phenotypes presented here and in [27] are currently unclear. It is likely, however, that they are due to the different experimental approaches used to activate the Wnt/b-catenin signaling pathway. We have observed cataract formation in transgenic mice designated delb-CAT, which express a stabilized (non-destructible) form of b-catenin, as well as in mice expressing a fusion protein of the Lef1 transcription factor and transactivation domain of bcatenin. In both cases, the mouse aA-crystallin promoter known to be selectively expressed in lens fiber cells [34] was used to drive expression of the transgenic protein. N-terminal truncation of bcatenin and direct fusion of the transactivation domain of bcatenin to its nuclear partner Lef1 have previously been used to mimic activation of Wnt/b-catenin signaling [29,32,44]. In particular, the strain of mice designated Catnb lox(Ex3) , in which exon 3 of the b-catenin gene has been flanked by loxP sites [44], has been extensively used to study the effect of activated Wnt/bcatenin signaling. Exon 3 encodes serine/threonine residues of bcatenin, which are targets of phosphorylation by GSK3 b kinase, the event that leads to b-catenin degradation. The product of Cre/ loxP recombination in Catnb lox(Ex3) mice is the stabilized form of bcatenin, functionally equivalent to the delb-CAT protein used here and elsewhere [32]. As the study presented here and that of [27] both employed stabilized forms of b-catenin, the difference in phenotypes must lie in the regulatory regions used. In case of bcatenin stabilization using the Cre/loxP system [27], b-catenin accumulation in the nucleus is the result of a two-step process. First, Cre recombinase is produced from the aA-crystallin promoter in lens fiber cells beginning at E13.5 (MLR39 Cre line). After recombination, transcription of the stabilized b-catenin is dependent on the activity of endogenous regulatory elements of the b-catenin gene in differentiating lens fiber cells. The level of bcatenin expression in lens fiber cells after this particular genetic manipulation is unclear, as no proof of stabilized b-catenin expression has been provided [27]. In contrast, the mouse aAcrystallin promoter is a reliable and frequently used regulatory element which allows medium-level expression of heterologous proteins in the lens fiber cell compartment [18,34,39,[45][46][47][48]. Alternative explanation of the different findings in our and Martinez et al. study [27], could be the phase of differentiation of manipulated fiber cells. Based on CLEF detection in aA-CLEF lenses, in the present study we have targeted early fiber cells (cells in transitional zone). Thus cells manipulated here were initiating the differentiation, whereas in previous study [27] only the population of fiber cells that have already initiated differentiation (outer cortical fiber cells) were affected.
Lens cataract is very often characterized by conversion of soluble crystallins into insoluble form [49], and this can be initiated by mutation in crystallins [50] or by changed physiological conditions inside the lens cells [49]. c-Crystallins are specifically expressed in differentiating and differentiated lens fiber cells; moreover, their expression is associated with decreased mitotic activity of the fiber cell [2]. We have observed downregulation of c-crystallins in aA-CLEF lenses. It is at present unclear if this is a direct effect of CLEF transcription factor or a consequence of an incomplete fiber cell differentiation in aA-CLEF lenses, although the second scenario is more likely. Most of the functional studies of c-crystallin promoters have been performed with cDand cF-crystallin promoters, however the direct regulation of c-crystallin genes by Lef/Tcf family of transcription factors has not been described (reviewed in [51]). Moreover, we were unable to identify Lef/Tcf binding sites (sequence logo CTTTGAT) in the known regulatory regions of cDand cF-crystallins (unpublished data). It was shown previously that transcription factor Pax6 can act as a transcriptional regulator of cF-crystallin promoter [52,53]. Pax6 was detected in the more mature lens fiber cells (in the central part of the lens) of aA-CLEF mice. It is very well possible that Pax6 participates in the repression of c-crystallin gene expression, thus contributing to the lower levels of c-crystallin protein in adult aA-CLEF lenses.
Disorganized distribution of fiber cell nuclei in aA-CLEF embryonic lenses was the first indication of impaired differentiation of fiber cells. The persistence of fiber cell differentiation markers in the fiber cell compartment of transgenic aA-CLEF lenses suggested delayed differentiation of fiber cells, which was shifted from the transitional zone to the central part of the lens. This is reminiscent of the effect of b-catenin stabilization within the entire lens [27] that resulted, among other effects, in delayed differentiation of lens epithelial cells. Since Wnt/b-catenin signaling is known to regulate the cell cycle [42], we examined the expression of cell cycle promoting factors cyclin D1 and cyclin D2, and the expression of negative regulators of the cell cycle, cyclin-dependent kinase inhibitors p27 Kip1 and p57 Kip2 . Cyclin D1 and cyclin D2 are closely related G1 cyclins [54]. Importantly, cyclin D1 is also a known target of the Wnt/b-catenin signaling pathway [55,56]. We thus expected its altered expression pattern due to the Wnt/b-catenin signaling activation in lens fiber cells. Cyclin D1 and cylin D2 are normally expressed in proliferating lens epithelial cells and in equatorial lens fiber cells because they are required for G1 to S phase transitional, whereas their expression is absent from fiber cells that have already completed the differentiation [43,46,57]. As cyclin-dependent kinase inhibitor (Cdk) p27 Kip1 is required together with p57 Kip2 for inhibition of Cdk's responsible for G1/S transition [11], p27 Kip1 expression coincides with the cell cycle exit during the fiber cell differentiation [46]. Since expression of p27 Kip1 and p57 Kip2 in the transitional zone appears to require the activity of transcription factor Prox1 [4], aberrant expression of p27 Kip1 at E16.5 in the central part of the lens may very well be due to the stronger Prox1 expression in this compartment in aA-CLEF lenses.
It was shown previously that the loss of b-catenin in the entire lens (epithelium and fiber cells) results in downregulation of cyclin D1 in lens epithelial cells, and at the same time in apparent upregulation of cyclin D1 in the more matured differentiating fiber cells [21], a somewhat surprisingly contradictory effect. It is well established that active Wnt/b-catenin signaling induces cyclin D1 and thereby triggers G1-progression [55,56], preventing cell cycle exit to G 0 . It is therefore not surprising that activation of Wnt/bcatenin in fiber cells in aA-CLEF mice induces an increase of cyclin D1 and cyclin D2 in the more matured differentiating fiber cells (located in the central part of the lens). Analysis of R26p-Fucci2 mouse embryos at E13.5 provides convincing visual confirmation that lens epithelial cells are in S/G2/M phase, whereas differentiating fiber cells in the lens equatorial region are in G1 phase [58]. It is likely that the differentiating fiber cells can be kept in G1 by the aberrantly present cyclin D1 and cyclin D2, activated via Wnt/b-catenin signaling at E13.5 in aA-CLEF mice.
In summary, we have shown here that ectopic activation of Wnt/b-catenin signaling in lens fiber cells during embryonic development results in delayed fiber cell differentiation and cataract formation. A more direct link between the activated Wnt/b-catenin signaling in lens fiber cells during embryonic development and cataract formation in adulthood is still missing and needs deeper investigation of the changes in lens physiology in consequence of the changed molecular regulation of the lens development.