Catenin and APC Reveal Roles for Canonical Wnt Signaling in Lens Differentiation

Home , Inke Nathke, WNT10A, WNT11, WNT4, WNT6, WNT7A

Conditional Mutations of ␤-Catenin and APC Reveal Roles for Canonical Wnt Signaling in Lens Differentiation

Gemma Martinez,1,2 Mary Wijesinghe,1,2 Kirsty Turner,1 Helen E. Abud,3 Makoto M. Taketo,4 Tetsuo Noda,5 Michael L. Robinson,6 and Robb U. de Iongh1

PURPOSE. Previous studies indicate that the Wnt/␤-catenin–signal- orphogenesis of the vertebrate lens is initiated when the ing pathway is active and functional during murine lens develop- Moutgrowing optic vesicle induces overlying ectoderm to ment. In this study, the consequences of constitutively activating form a lens placode, which in turn invaginates to form a hollow the pathway in lens during development were investigated. ball of epithelial cells, the lens vesicle. Posterior lens vesicle cells elongate and differentiate into primary lens fiber cells, METHODS. To activate Wnt/␤-catenin signaling, ␤-catenin (Catnb) and adenomatous polyposis coli (Apc) genes were whereas anterior vesicle cells adopt a cuboid epithelial phenotype. During embryonic development, the anterior epithelium conditionally mutated in two Cre lines that are active in whole is highly proliferative. However, during fetal and postnatal lens (MLR10) or only in differentiated fibers (MLR39), from development, proliferation becomes restricted to a small pop- E13.5. Lens phenotype in mutant lenses was investigated by ulation of epithelial cells in the germinative zone, located histology, immunohistochemistry, BrdU labeling, quantitative anterior to the lens equator. The progeny from these cell RT-PCR arrays, and TUNEL. divisions may contribute to the anterior epithelium or move RESULTS. Only intercrosses with MLR10 resulted in ocular phe- posteriorly into the transitional zone where they differentiate notypes, indicating Wnt/␤-catenin signaling functions in lens into secondary lens fiber cells, a process that continues epithelium and during early fiber differentiation. Mutant lenses throughout life.1 How cell fate is assigned in the differentiating were characterized by increased progression of epithelial cells lens has been a question that has been the focus of studies for through the cell cycle, as shown by BrdU labeling, and phos- the past 50 years, since Coulombre and Coulombre2 showed phohistone 3 and cyclin D1 labeling, and maintenance of that signals emanating from the ocular media regulate lens epithelial phenotype (E-cadherin and Pax6 expression) in the polarity. fiber compartment. Fiber cell differentiation was delayed as Several growth factor families have been shown to be in- shown by reduced expression of c-maf and ␤-crystallin and volved in regulating lens induction and differentiation.3,4 delay in expression of the CDKI, p57kip2. From E13.5, there These include the fibroblast growth factor (FGF), transforming were numerous cells undergoing apoptosis, and by E15.5, growth factor-␤ (TGF␤), and bone morphogenetic protein 4–11 there was evidence of epithelial–mesenchymal transition with (BMP) families. TGF␤ has also been shown to initiate the numerous cells expressing ␣-smooth muscle actin. Quantita- epithelial-mesenchymal transition (EMT) of lens epithelial cells, tive PCR analyses revealed large changes in expression of Wnt a process that underlies anterior subcapsular cataract and pos- 4,12 target genes (Lef1, Tcf7, T (Brachyury), and Ccnd1), Wnt terior capsule opacification. More recently, the canonical Wnt signaling pathway has also been implicated in various inhibitors (Wif1, Dkk1, Nkd1, and Frzb) and also several Wnts 4,13–15 (Wnt6, Wnt10a, Wnt8b, and Wnt11). stages of lens development. ␤ ␤ In the absence of Wnt ligand binding, cytoplasmic -catenin CONCLUSIONS. These data indicate that the Wnt/ -catenin pathway is recruited to a degradation complex (axin/GSK3␤/CK1 and plays key roles in regulating proliferation of lens stem/progenitor APC) where it is phosphorylated, resulting in its ubiquitylation cells during early stages of fiber cell differentiation. (Invest Oph- and subsequent proteasomal degradation. The canonical Wnt thalmol Vis Sci. 2009;50:4794–4806) DOI:10.1167/iovs.09-3567 signaling pathway is activated upon binding of Wnt ligands to Frizzled (Fzd)/LDL-related protein (LRP) coreceptor com- plexes, leading to activation of Fzd and the phosphorylation of 1 disheveled (Dsh) proteins. This process in turn leads to the From the Ocular Development Laboratory, Department of Anatomy ␤ and Cell Biology, University of Melbourne, Victoria, Australia; and 3Anat- recruitment of the -catenin destruction complex to the mem- omy and Cell Biology, Monash University, Clayton, Victoria, Australia; the brane. Inhibition of the destruction complex components axin 4Department of Pharmacology Graduate School of Medicine, Kyoto Uni- and GSK3␤ is mediated by active Dsh and leads to increased versity, Kyoto, Japan; the 5Department of Cell Biology, The Cancer Insti- cytoplasmic levels of hypophosphorylated ␤-catenin, which tute, Japanese Foundation for Cancer Research, Tokyo, Japan; and the translocates to the nucleus where it interacts with the LEF/TCF 6Department of Zoology, Miami University, Oxford, Ohio. 2 transcription factors to activate the transcription of target Contributed equally to the work and therefore should be consid- genes including c-myc, cyclin D1, and the ephrins. In this ered equivalent authors. context, ␤-catenin acts as the central transcriptional regulator Supported by Project Grant 400174 from the National Health and ␤ Medical Research Council (NHMRC) (RUdI, HEA) and the Rebecca L. of the canonical Wnt signaling pathway. -Catenin also func- Cooper Research Foundation for equipment. tions as a structural protein in adherens junctions where, Submitted for publication February 13, 2009; revised April 9, together with ␣-catenin, it links membrane cadherins to the 2009; accepted July 2, 2009. actin cytoskeleton. Disclosure: G. Martinez, None; M. Wijesinghe, None; K. APC has been shown to be an essential component of the Turner, None; H.E. Abud, None; M.M. Taketo, None; T. Noda, destruction complex that negatively regulates Wnt signaling None; M.L. Robinson, None; R.U. de Iongh, None via different functions of some of its domains (reviewed in The publication costs of this article were defrayed in part by page Refs. 16–18). The function of the destruction complex relies charge payment. This article must therefore be marked “advertise- ␤ ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. on the ability of APC to promote phosphorylation of -cate- Corresponding author: Robb U. de Iongh, Anatomy and Cell Biol- nin’s centrally located amino acid repeat motifs. Failure to ogy, The University of Melbourne, Parkville VIC 3010, Australia; phosphorylate ␤-catenin results in its accumulation in the cy- [email protected]. toplasm as seen in colorectal cancer exhibiting APC trunca-

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4795

tions, which do not allow the binding of APC to axin. APC has ␤-catenin (Catnblox(Ex3)) have also been described previously.34 Cre- also been shown to have additional roles unrelated to regula- mediated recombination at this locus removes the phosphorylation tion of ␤-catenin. It has been shown to bind directly and and ubiquitination sites in exon 3 and results in a stable form of indirectly to the plus end of microtubules via its C-terminal- ␤-catenin that constitutively activates Wnt/␤-catenin signaling. located basic domain, thus playing a major role in regulating MLR10 or MLR39 mice were mated with Apclox(580S) and the cytoskeleton as well as affecting cell migration and forma- Catnblox(Ex3) mice and the resultant progeny were screened by PCR tion of mitotic spindles.19–23 It has also been associated with for the desired genotype. Heterozygous progeny (Apclox(580S)/wt/ the actin cytoskeleton.21,24–26 Creϩ and Catnblox(Ex3/wt/Creϩ) were crossed until mice homo- There is considerable evidence that inhibition of canonical zygous for the Cre transgene and the floxed alleles were gener- Wnt signaling is essential for proper eye field specification in ated. Resultant mutant mice were maintained by inbreeding to various species (reviewed in Ref. 14). Consistent with this, obtain mice with desired genotypes: APC10 (Apclox(580S)/lox(580S)/ conditional loss of ␤-catenin, in murine periocular ectoderm, MLR10Creϩ), APC39 (Apclox(580S)/lox(580S)/MLR39Creϩ), ␤CATEx310 before eye formation, results in the formation of ectopic lenses (Catnblox(Ex3)/lox(Ex3)/MLR10Creϩ or CatnbEx3/lox(Ex3)/MLR10Creϩ), by the loss of Wnt signals and in abnormal ocular morphogen- or ␤CATEx339 (CatnbEx3/lox(Ex3)/MLR39Creϩ). esis by the disruption of adherens junctions.15,27 By contrast, conditional activation of ␤-catenin by deletion of exon3, which Histology and Immunofluorescence harbors the phosphorylation/ubiquitination sites, results in suppression of lens cell fate in periocular ectoderm.15,28 Other Eyes (postnatal day [P]1–P21) and embryos (embryonic day [E]13.5– E17.5) were obtained from superovulated female mutant and Wt mice. studies have also indicated that Wnt signaling plays roles dur- 6 ing later stages of lens differentiation. Components of the Wnt Superovulations were performed as previously described. Eyes or pathway, including genes for Wnts, Fzd receptors, and LRPs, as embryo heads were fixed in either 10% neutral buffered formalin (NBF) well as modulators of Wnt activity, such as the secreted friz- and embedded in paraffin or in fresh 4% paraformaldehyde (PFA) in zled-related proteins (Sfrps) and Dikkopf (Dkks), have been PBS before cryoprotection with 30% sucrose and embedding in OCT compound (Tissue-Tek; Sakura Fintek, Tokyo, Japan). found to be expressed primarily in the lens epithelium from ␮ the early lens vesicle stages through to the postnatal stages Paraffin sections (4–7 m) were dewaxed and rehydrated through (reviewed in Ref. 14). Moreover, analysis of the Lef/TCF re- decreasing graded ethanol series to water before staining with hema- porter lines has shown activity of Wnt signaling in embryonic toxylin and eosin, periodic acid Schiff or immunofluorescence. Paraf- lens epithelium.29,30 The requirement for Wnt signaling during fin-embedded sections used for immunostaining were subjected to heat-mediated antigen retrieval with 0.01 M citrate buffer (pH 6.0) lens differentiation was first demonstrated in Lrp6 null mice, 13 which have lenses with a deficient anterior epithelium.31 More containing 0.05% Tween-20, as described previously. Frozen and recent studies of mice with a conditional null mutation of paraffin-embedded sections were blocked with 3% to 5% normal goat ␤-catenin13 show that Wnt signaling affects lens epithelial serum (NGS) in 0.1% BSA/PBS for 20 minutes before they were incu- stem/progenitor cells; these mice also have a deficient epithe- bated with primary antibodies diluted in blocking solution. The pri- lial layer with loss of E-cadherin expression, epithelial cell mary antibodies used were mouse anti-E-cadherin (1:200; cat. no. 610182; BD Transduction, Lexington, KY), mouse anti-␤-crystallin cycle arrest at the G1-S transition and premature cell cycle exit. ␤ (Clone 3H9; hybridoma supernatant, the kind gift of Robert C. Au- Fiber cell differentiation in these -catenin-null lenses was also ␤ compromised, with disrupted cell elongation, decreased gusteyn, Vision CRC, Sydney, Australia), mouse anti-active -catenin ␤ (1:100; 05-665; Millipore, Billerica MA), rabbit anti-APC (kind gift of -crystallin expression, and abnormal extracellular matrix 35 (ECM) expression. Inke Nathke, University of Dundee), rabbit anti-phospho-histone H3 (1:200, 07-424; Millipore), mouse anti-␣-smooth muscle actin (1:200, In this study, we report that activation of the Wnt pathway kip2 in the lens by truncation of APC or by an activating mutation of A2547; Sigma-Aldrich, Castle Hill, NSW, Australia), rabbit anti-p57 ␤-catenin, results in the expansion of lens progenitor cells, (1:200; 4058-500; AbCam, Cambridge, UK) and rabbit monoclonal inhibition of fiber cell differentiation and partial commitment anti-cyclin D1 (cat no. 21699 (SP4); Abcam) at 4°C overnight. APC to a mesenchymal fate. These effects are only observed if the immunohistochemistry required sections to be permeabilized with pathway is activated in lens epithelium and fiber cells but not CSK buffer (50 mM NaCl, 300 mM sucrose, 10 mM PIPES [pH 6.3], 3 mM MgCl with 0.5% Triton X-100) at room temperature for 10 min- in differentiated fiber cells alone. The results indicate that the 2 Wnt pathway plays an essential role in the regulation of lens utes and blocked with 20% NGS and 1% BSA in PBS before incubation stem/progenitor cells and early phases of fiber cell differentiation. with primary antibody. Secondary antibodies used were HRP-conjugated goat-anti mouse (1:500, 65-6120; Invitrogen, Mt. Waverley, VIC, Australia), HRP-conjugated goat-anti rabbit (1:200, 65-6520; Invitrogen) MATERIALS AND METHODS or AlexaFluor488-conjugated anti-mouse or anti-rabbit IgG (1:500; A-11008, A-11001; Invitrogen) diluted in blocking solution for 2 hours All experimental procedures on animals conformed to the ARVO State- at room temperature. Sections labeled with HRP-conjugated secondary ment for the Use of Animals in Ophthalmic and Vision Research and antibodies were developed with diaminobenzidine (DAB) and were were approved by the Animal Ethics Committees of the University of counterstained with hematoxylin, dehydrated, and mounted in DPX. Melbourne. For fluorescent labeling of nuclei, the sections were stained with Hoechst dye (Sigma-Aldrich) for 10 minutes before they were washed Generation of Conditional Mutants with PBS and mounted with fluorescent mounting medium (Dako Cytomation, Inc., Carpinteria, CA). Transgenic mice expressing Cre recombinase under the control of the ␣ A-crystallin promoter (fiber cells only; MLR39) or with the inclusion Bromodeoxyuridine (BrdU) Labeling of a Pax6 enhancer element (epithelial and fiber cells; MLR10) have been described previously.32 The floxed APC mouse (Apclox(580S))in BrdU labeling was performed to quantify cell proliferation in lenses. which LoxP sites have been inserted into the introns flanking exon 14 Pregnant mice were injected with 0.1 mg 5-bromo-2Ј-deoxyuridine of the Apc gene have also been described previously.33 Cre-mediated (BrdU) in PBS/g body weight and 0.1 mg 5-fluoro-2Јdeoxyuridine recombination at this locus results in deletion of exon 14 and a (FldU) in PBS per 10g body weight 1 to 1.5 hours before collection and downstream frameshift mutation at codon 580. The resultant mutant fixation of embryos. Rehydrated paraffin-embedded sections were APC protein is unable to participate in ␤-catenin phosphorylation, treated with 2 M HCl at room temperature for 30 minutes to denature ubiquitylation and degradation. Mice harboring the floxed exon 3 of DNA followed by neutralization with 0.1 M sodium borate at room

Downloaded from iovs.arvojournals.org on 09/25/2021 4796 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

temperature for 10 minutes. The sections were washed with PBS and incubated with 0.12% trypsin in PBS at 37°C for 25 minutes and then processed for anti-BrdU immunofluorescence. They were blocked with 3% NGS in 0.1% BSA in PBS at room temperature for 20 minutes and incubated with monoclonal anti-BrdU antibody (1:50; Bioclone, Syd- ney, NSW, Australia) at 4°C overnight. The secondary antibody was HRP-conjugated goat anti-mouse (1:300) incubated at room temperature for 2 hours. Color development was performed with DAB, and the sections were counterstained with hematoxylin, dehydrated, and mounted in DPX. BrdUϩ and total cell counts were quantified in the epithelial and fiber compartments in sections from three to six E13.5 embryos. Data are expressed as the mean (Ϯ SEM) percentage of BrdUϩ cells in each compartment, analyzed by ANOVA and Student’s t-test.

TUNEL Assay To detect cells undergoing apoptosis, a TUNEL system (DeadEnd Fluoro- metric; Promega, Sydney, NSW, Australia) was used on rehydrated paraf- ﬁn-embedded sections, according to the manufacturer’s instructions.

PCR Array Profiling Mouse Wnt PCR arrays (Signaling Pathway RT2 Profiler PCR Arrays; SuperArray Bioscience Corp., Frederick, MD) were used to interrogate the expression of 84 genes involved in Wnt-mediated signal transduction. The arrays include primers coding for members of the Frizzled and Wnt families, Wnt signaling pathway regulators and competitive antagonists, intracellular signaling mediators, pathway target genes, and genes involved in protein modification downstream of Wnt signaling, including genes involved in kinase and phosphatase activity and ubiquitination. The arrays also contain primers coding for a panel of five housekeeping genes to normalize PCR array data, genomic DNA control primers, reverse transcription controls and positive PCR controls. Embryonic lenses were dissected from at least two litters of wild- type, APC10 or ␤CATEx310 embryos at E14.5, taking care to remove surrounding optic cup tissue. As the embryonic lens is invested with the tunica vasculosa lentis, inclusion of endothelial cells and blood cells could not be excluded. Total RNA was isolated from pooled lens vesicles (RNeasy Mini Kit; Qiagen Pty Ltd., Doncaster, VIC, Australia) with on-column DNase-digestion. RNA integrity and concentration were quantified with a spectrophotometer (NanoDrop; ThermoFisher Scientific, Wilmington, DE). For each array, 1 ␮g RNA from each sample was reverse transcribed (RT2 First Strand kit; SuperArray Bio- science) and the resultant cDNA was amplified by real-time PCR (RT2 FIGURE 1. Ocular phenotypes at P21 and histology at P2 of ␤CA- Ex3 Real-Time SYBR Green PCR master mix; SuperArray Bioscience) as T 10 (B, F), APC39 (C, G), and APC10 (D, H) mutant eyes compared follows: melting for 10 minutes at 95°C, 40 cycles of two-step PCR with wild-type (Wt) eyes (A, E). Only mutants derived from the MLR10 Cre line (B, D) showed an abnormal ocular phenotype (microphthal- including melting for 15 seconds at 95°C, and annealing for 1 minute mia), whereas mutants derived from the MLR39 Cre line (C) were at 60°C. The cycling program was immediately followed by a melting similar to Wt (A). Wt and APC39 lenses showed the characteristic curve analysis to confirm generation of a single amplicon. All reverse epithelial (e) and fiber (f) cell morphology and arrangement (E, G). transcriptions and real-time PCR analyses were performed in triplicate ␤CATEx310 lenses (F) retained an epithelial layer (arrowhead) and had ⌬⌬ and average Ct values for each gene were analyzed by the Ct method large accumulations of nuclei in the fiber mass (arrows). APC10 lenses using commercial analysis template (Excel; SuperArray Bioscience). (H) lacked a well-defined epithelium and had extensive vacuolization Relative expression values were determined between wild-type and of the fiber mass that often ruptured through the posterior lens capsule ␤ Ex3 ⌬ (arrowheads) into the vitreous chamber. Scale bars: (A–D) 3 mm; APC10 or CAT 10 samples by normalizing differences in Ct values ⌬⌬ (E–F) 200 ␮m. ( Ct) to the average of five housekeeping genes. The change (x-fold) of gene expression in samples was calculated as 2Ϫ⌬⌬Ct. lenses were characterized by an irregular multilayered anterior epithelium and an abnormal fiber compartment that contained RESULTS numerous epithelioid cells and poorly elongated fiber cells (Fig. 1F). The hyaloid vasculature appeared hypertrophic. The Ocular Phenotype of Mutant Mice APC10 lenses at P2 were characterized by the lack of a well Mutant mice derived from MLR10 crosses (␤CATEx310, defined epithelium and highly vacuolated fiber mass that often APC10,) had microphthalmia that was evident from eye open- ruptured through the posterior lens capsule into the vitreous ing (Figs. 1B, 1D). By contrast, mice derived from MLR39 chamber (Fig. 1F). ␤ Ex3 crosses (APC39, Fig. 1C; CAT 39, not shown) had appar- Epithelial Hypertrophy and Disrupted Fiber ently normal ocular phenotype similar to that of wild-type mice (Fig 1A). Histologic examination of neonatal (P2) eyes revealed Differentiation in Embryonic Lenses gross disturbances of lens differentiation in APC10 and Previous studies have shown that Cre-mediated recombination ␤CATEx310, but not APC39 or ␤CATEx339 lenses. ␤CATEx310 in MLR10- and MLR39-derived lines is complete at E13.5.13

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4797

Ex3 Ex3 FIGURE 2. Lens phenotypes at E13.5 and E15.5. H&E-stained sections of Wt (A, E), ␤CAT 39 (B, F), ␤CAT 10 (C, G), and APC10 (D, H) at E13.5 (A–D) and E15.5 (E–H). Lenses from ␤CATEx339 mice (B, F) showed normal histologic structure similar to that in Wt mice (A, E) at both ages. ␤CATEx310 (C) and APC10 (D) lenses at E13.5 had a defined epithelial layer (arrowheads) but the fiber cell compartment was populated by numerous epithelial-like cells (arrows), which appeared to have migrated posteriorly along the lens capsule. By E15.5, the phenotypes of the ␤CATEx310 (G) and APC10 (H) had diverged considerably. ␤CATEx310 lenses (G) had a single monolayer of anterior epithelial cells, but the posterior fiber compartment was populated by epithelial cells and a large vesicle, containing amorphous material, occupied the anterior fiber cell compartment. APC10 lenses at E15.5 (H) had a multilayered anterior epithelium that was indistinguishable from the overlying corneal endothelium. The anterior multilayered epithelium (arrowheads) comprised small cuboid-to-squamous epithelial cells and also enlarged cells (double arrowheads, inset). The fiber compartment was highly vacuolated, with clumps of epithelial-like cells along the posterior capsule (arrows), and often contained numerous red blood cells (✽). Scale bars: (A–H) 100 ␮m; (inset)50␮m.

Histologic analysis of wild-type and mutant lenses at E13.5 and protein that recognizes part of the protein downstream of the at E15.5 showed that MLR39-derived mice had the character- frameshift mutation.35,36 APC immunostaining of E13.5 wild- istic epithelial and fiber cell arrangement similar to wild-type type lenses showed abundant expression of wild-type APC in lenses (Figs. 2A, 2B, 2E, 2F). In these lenses, the epithelium epithelial and fiber cells (Fig. 3A). By contrast, APC10 lenses covered the anterior surface of the lens and cells at the equator had a marked loss of APC reactivity in both epithelial and fiber began to elongate into lens fiber cells. However, in E13.5 cells but not in the overlying cornea (Fig. 3B), indicating ␤CATEx310 and APC10 lenses, the cells at the equator failed to efficient truncation of APC in the lens. Localization patterns of elongate, maintained an epithelial phenotype, and continued APC in E15.5 wild-type and APC10 lenses were similar to that to migrate along the posterior capsule (Figs. 2C, 2D). Fiber cell observed at E13.5 (data not shown). To confirm that there was elongation appeared to be inhibited in both lines and there was activation of Wnt signaling we used an antibody that recog- no evidence of the characteristic bow zone. The structure of nizes active ␤-catenin. In wild-type lenses, this antibody recog- the optic cup neuroepithelium in both lines appeared normal nizes membrane-associated ␤-catenin (Fig. 3E), with very little and similar to wild-type. nuclear reactivity, similar to previous studies31; however, in The phenotypes of the ␤CATEx310 and APC10 lenses began APC10 lenses, there was increased cytoplasmic staining, par- to diverge at E15.5. In E15.5 ␤CATEx310 lenses, there was a ticularly in the epithelioid cells populating the fiber compart- distinct anterior epithelium but the fiber compartment was ment (Fig. 3F). Distinct nuclear ␤-catenin reactivity was evi- populated by numerous epithelioid cells surrounding a large dent in many lens cells but not in cells outside the lens (Figs. space containing degenerate cells (Fig 2G). By contrast, the 3G, 3H). E15.5 APC10 lenses comprised a multilayered anterior epithe- In Catnblox(Ex3) mice, exon3 of the ␤-catenin gene was lium that was indistinguishable from the overlying corneal flanked by LoxP sites, and Cre-mediated recombination re- endothelium and overlaid a mass of enlarged swollen cells. The sulted in a form of ␤-catenin that cannot be phosphorylated fiber mass was largely degenerated and vacuolated with numer- and targeted for proteasomal destruction. In wild-type lenses, ous clumps of cells populating the posterior capsule (Fig. 2H). ␤-catenin was detectable in all cells but was particularly strong In many cases the lens fiber mass also contained numerous red along epithelial cell membranes (Fig. 3C). However, in ␤CA- blood cells indicating penetration of the hyaloid vessels into TEx310 lenses, there was dramatically increased localization of the lens. ␤-catenin throughout the lens, particularly in the epithelioid cells in the fiber compartment (Fig. 3D). The reactivity was Activation of the Wnt Pathway in Mutant Lenses localized along membranes, in the cytoplasm and also in nu- In Apclox(580S) mice, the Apc gene has LoxP sites flanking exon clei. The localization patterns strongly indicate that in 14. Cre-mediated excision of exon 14 in APC10 conditional these lenses there is activation of the Wnt pathway, consistent mutants results in a frameshift truncation mutation of the APC with previous studies using Apclox(580S) and Catnblox(Ex3) protein that can no longer bind ␤-catenin in lens epithelial and mice.33,34,37–40 fiber cells. As a result ␤-catenin is not ubiquitinated or de- To further confirm activation of the Wnt pathway, RNA was graded and accumulates in the cytoplasm and nuclei. To con- isolated from E14.5 APC10 and ␤CATEx310 lens vesicles and firm effective recombination of the Apc gene in APC10 lenses, assayed using quantitative PCR arrays. When this approach was immunohistochemistry was performed with an APC antibody used, 38 of 84 Wnt pathway-associated genes in the array raised to amino acid residues 1034 to 2130 of the human APC showed at least a twofold significant (P Ͻ 0.05) difference in

Downloaded from iovs.arvojournals.org on 09/25/2021 4798 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

and 11-fold), Tcf7 (7- and 5-fold) and Ccnd1 (2.8, 1.5-fold). There was also dramatic upregulation of Wnt inhibitors such as Wif1 (392, 287-fold), Dkk1 (70, 21-fold), Frzb (14-fold) and the naked cuticle homologue, Nkd1 (25- and 13-fold). Of interest, some inhibitors such as Sfrp1, Sfrp2, the Groucho-related gene (Aes), and the ubiquitin ligase complex gene (Btrc) were sig- niﬁcantly downregulated. Many of the Wnt ligands were also highly upregulated, including Wnts associated with canonical (Wnt10a, Wnt8b) and noncanonical (Wnt4, Wnt5a, and Wnt11) signaling and Wnt6, which has been implicated in EMT.41,42 Three Wnt genes, Wnt7a, Wnt7b, and Wnt8a were signiﬁcantly downregulated; both Wnt7b and Wnt8a are known to have canonical effects, whereas Wnt7a appears to initiate both canonical and noncanonical signals, depending on the Frizzled receptor (Fzd5 or Fzd10) it binds.43 Of the Wnt receptors, Fzd8 and Lrp5 were upregulated but Fzd5 and Fzd3 were moderately downregulated.

Maintenance of Epithelial Phenotype by Constitutive Activation of Wnt Signaling A predominant feature of the lens phenotypes in APC10 and ␤CATEx310 lenses was the persistence of epithelial-like cells in the posterior fiber compartment. To determine whether these cells retained epithelial cell markers, we examined the expression of E-cadherin and Pax6 by immunofluorescence at E13.5. In wild-type and ␤CATEx339 lenses, E-cadherin expression was detected uniquely in the epithelium, with little or no staining below the equator in the transitional zone or fiber cell compartment (Figs. 4A, 4B). ␤CATEx310 and APC10 lenses also expressed E-cadherin in the epithelium, but in both mutants, persistent E-cadherin reactivity was detected in transitional zone cells below the equator and in the fiber compartment (Figs. 4C, 4D). Intriguingly, E-cadherin was downregulated in many cells at the posterior pole, suggesting that other factors may downregulate E-cadherin. In older ␤CATEx310 lenses at E15.5, similar aggregations of E-cadherin-positive, epithelial- like cells were observed in the fiber compartment (data not

IGURE ␤ shown). F 3. Expression of APC and -catenin in wild-type (A, C, E; Wt) Ex3 and mutant ␤CATEx310 (D) and APC10 (B, F, G, H) lenses at E13.5. (A) In wild-type and ␤CAT 39 lenses, Pax6 was localized in the In Wt lenses, distinct reactivity for APC protein (brown stain) was nuclei and cytoplasm of epithelial and transitional zone cells, with detected in lens epithelial (e) and fiber (f) cells by an antibody that a sharp reduction in expression in differentiating fiber cells (Figs. detects the C-terminal region. Strong staining was also detected in the 4E, 4F). In ␤CATEx310 and APC10 mutant lenses, nuclear Pax6 corneal epithelial cells (c). (B) In APC10 lenses, there was marked loss reactivity was detected in all cells, including the epithelioid cells of APC staining with only occasional fiber cells still showing staining in the fiber cell compartment (Figs. 4G, 4H). (✽). (C) In Wt lenses, ␤-catenin was distinctly detected along cell membranes of epithelial (e) and fiber (f) cells and also in the optic cup Ex3 (oc). (D)In␤CATEx310 lenses, ␤-catenin expression was extremely Delayed Fiber Cell Differentiation in ␤CAT 10 strong with increased reactivity throughout the lens. (F) In APC10 and APC10 Lenses lenses, increased ␤-catenin staining compared with Wt (E) was detected, particularly in the cytoplasm of epithelial-like cells (arrows). The epithelial marker expression indicated that cells in the (G, H) In some lenses at E13.5, distinct nuclear ␤-catenin was also fiber compartment retained aspects of an epithelial phenotype. detectable, particularly in the epithelioid cells in the fiber compart- To examine whether these cells had undergone fiber cell ment (arrows), but not in cells outside the lens (arrowheads). Scale differentiation, we examined the expression of ␤-crystallin, ␮ ␮ ␮ bars: (A, B, E, F)50 m; (C, D) 100 m; (G, H)30 m. c-Maf, and Prox1. ␤-Crystallin is a characteristic marker for initiation of fiber differentiation. At E13.5, ␤-crystallin is nor- expression in either APC10 or ␤CATEx310 compared with mally expressed in the transitional zone cells undergoing ex- wild-type (Table 1 and Supplementary Data, http://www.iovs. tensive elongation and in terminally differentiated fibers in the org/cgi/content/full/50/10/4794/DC1). The data showed re- center of the lens (Fig. 5A). A similar expression was detected markably concordant results for both APC10 and ␤CATEx310 in ␤CATEx339 lenses (Fig. 5B). However, in ␤CATEx310 and mice, with 20 genes upregulated and 14 genes downregulated APC10 mutant lenses, ␤-crystallin expression was significantly significantly (P Ͻ 0.05) in both samples. Of genes showing a delayed, with most epithelioid cells not showing any reactivity greater than 1.5-fold significant change, only five genes (Wnt1, (Figs. 5C, 5D). Only cells in the center of the lens showed Wnt4, Fshb, Fzd2, and Fzd5) showed discordant patterns, distinct ␤-crystallin expression. being significantly upregulated in one and significantly down- The transcription factors, c-Maf (Figs. 5E, 5F) and Prox1 regulated in the other (Supplementary Data). Consistent with (not shown) are predominantly expressed in fiber cell nuclei of the activation of canonical Wnt signaling in both lines, there wild-type lenses.44,45 In mutant lenses, expression of c-Maf was a similar significant upregulation of many Wnt target genes appeared to be markedly reduced and delayed (Figs. 5G, 5H). such as Lef1 (59- and 31-fold, respectively), Brachyury (T; 16- By contrast, Prox1 expression did not seem to be greatly

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4799

Ex3 TABLE 1. Changes in Expression of Wnt Pathway Genes in APC10 and ␤CAT 10 Mouse Lenses at E14.5

Average ⌬Ct ؊ (Ct[GOI] Ct[HKG]) Change (x-Fold) t-Test Change (x-Fold) t-Test Gene APC10/Wt P ␤CATEx310/Wt P Wt APC10 ␤CATEx310

Wnt target genes Lef1 58.8 8.5E-07 30.8 2.7E-06 9.69 3.81 4.74 T 16.5 4.4E-06 11.1 1.0E-06 10.78 6.73 7.31 Tcf7 7.5 4.2E-07 5.0 1.1E-05 4.72 1.82 2.39 Ccnd1 2.8 9.1E-05 1.5 2.3E-03 3.33 1.83 2.71 Jun 2.4 5.8E-05 2.0 1.4E-04 4.49 3.21 3.50 Pitx2 1.2 1.5E-01 Ϫ2.6 4.2E-03 8.40 8.11 9.80 Foxn1 Ϫ3.2 7.4E-04 Ϫ2.6 4.3E-04 7.52 9.17 8.88 Wnt pathway components Rhou 2.2 2.2E-03 1.6 2.0E-02 7.91 6.80 7.26 Kremen1 2.1 2.7E-04 1.7 7.6E-03 4.91 3.82 4.18 Ep300 Ϫ1.0 7.2E-01 Ϫ2.0 2.5E-03 4.84 4.90 5.84 Ctnnb1 1.2 3.5E-01 Ϫ2.4 9.7E-04 1.36 1.12 2.63 Wnt Inhibitors Wif1 392.5 1.2E-06 287.4 1.E-06 12.04 3.43 3.88 Dkk1 69.7 1.9E-06 21.3 7.1E-06 11.82 5.70 7.41 Nkd1 25.3 2.8E-06 13.3 4.8E-06 5.29 0.63 1.55 Frzb 13.8 8.6E-05 13.6 1.1E-04 11.11 7.32 7.34 Tle1 1.9 3.0E-04 2.1 5.7E-05 6.26 5.33 5.16 Sfrp4 7.8 1.0E-04 1.8 5.6E-03 13.28 10.32 12.40 Sfrp2 Ϫ2.1 9.8E-04 Ϫ1.1 4.2E-01 1.23 2.27 1.34 Aes Ϫ2.4 1.0E-03 Ϫ1.1 4.9E-01 2.72 4.04 2.87 Btrc Ϫ1.9 5.5E-06 Ϫ2.6 1.2E-06 5.75 6.68 7.12 Sfrp1 Ϫ3.0 3.1E-05 Ϫ3.1 8.4E-06 0.01 1.61 1.64 Wnts Wnt6 508.2 8.7E-08 353.9 6.0E-07 12.49 3.50 4.02 Wnt10a 100.3 1.3E-06 62.9 2.4E-05 15.41 8.77 9.44 Wnt8b 18.6 2.2E-05 2.7 2.0E-02 15.41 11.19 13.98 Wnt11 4.2 3.1E-02 3.4 9.6E-02 13.60 11.55 11.82 Wnt4 4.1 7.8E-04 Ϫ2.1 2.4E-02 11.72 9.69 12.75 Wnt5a 2.5 7.4E-03 1.7 3.4E-02 9.94 8.62 9.16 Wnt7a Ϫ2.3 1.1E-03 Ϫ6.5 1.3E-04 4.30 5.51 7.00 Wnt7b Ϫ2.4 1.1E-03 Ϫ4.1 2.3E-04 2.99 4.27 5.01 Wnt8a Ϫ13.5 2.7E-01 Ϫ2.8 6.5E-01 12.52 16.28 13.99 Wnt Receptors Lrp5 2.4 2.E-04 2.4 9.8E-04 7.84 6.56 6.57 Fzd8 2.1 7.4E-02 2.0 6.7E-03 13.18 12.10 12.17 Fzd5 Ϫ1.5 3.7E-03 5.7 2.2E-05 6.76 7.37 4.26 Fzd3 Ϫ1.6 7.1E-03 Ϫ2.0 1.3E-03 3.26 3.92 4.23

⌬ GOI, gene of interest; HKG, house-keeping genes; Ct, change in threshold cycle number.

affected, being clearly detectable in nuclei of fiber compart- lens epithelial progenitor cells, we examined cell cycle mark- ment cells of ␤CATEx310 mutant lenses (not shown). ers. BrdU incorporation assays showed that wild-type and ␤CATEx339 lenses had normal patterns of epithelial cell prolif- ␤ Ex3 Abnormal ECM Protein Expression in CAT 10 eration, similar to that documented previously (Figs. 7A, 7B). and APC10 Lenses Only an occasional BrdUϩ cell was detected, just below the As the phenotype at postnatal ages suggested insufficiency of lens equator. However, in mutant ␤CATEx310 and APC10 the lens capsule we examined the expression of two key lens lenses, BrdU incorporation was detected, not only in the ante- capsule proteins: laminin and collagen IV. In wild-type lenses, rior epithelium but also in cells of the posterior fiber compart- both laminin and collagen IV were predominantly detected in ment (Figs. 7C, 7D). Quantification of BrdU incorporation in epithelial cells and in the posterior capsule (Figs. 6A, 6D). wild-type lenses showed that, compared with wild-type lenses, However, in both ␤CATEx310 and APC10 mutant lenses (Figs. there was a significant (P Ͻ 0.05) increase in cell proliferation 6B, 6C, 6E, 6F), cellular expression of these proteins was in the anterior epithelial layer of ␤CATEx310 (39.1% Ϯ 3.4% vs. increased and extended into the fiber compartment. Paradox- 24.2% Ϯ 1.8%), but not APC10 (18.6% Ϯ 1.6% vs. 14.8% Ϯ ically, the distinct expression in the basement membrane of 1.7%) lenses. The most dramatic changes occurred in the fiber the posterior lens capsule was decreased and difficult to de- compartment (Figs. 7E, 7F), where both mutants showed sig- tect. This result suggests that, although there is increased nificant increases in BrdUϩ cells compared to wild-type cellular production of ECM proteins, there appears to be al- (␤CATEx310, 20.2% Ϯ 2.3% vs. 0.2% Ϯ 0.2%; APC10, 14.0% Ϯ tered deposition or organization of the basement membrane. 1.2% vs. 1.9% Ϯ 0.5%). To assess changes in the cell cycle, we also examined Dysregulated Cell Cycle in ␤CATEx310 and markers for M-phase (phosphohistone 3; PH3), G1-S phase APC10 Lenses (cyclin D1), and cell cycle exit (p57kip2). PH3 expression is

Wnt signals are known to regulate the cell cycle and, as the associated with condensed chromosomes at the G2-M transi- phenotype of mutant lenses suggested altered regulation of tion. In wild-type (Fig. 8A) and ␤CATEx339 (Fig. 8B) lenses,

Downloaded from iovs.arvojournals.org on 09/25/2021 4800 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

Ex3 Ex3 FIGURE 4. Expression of the epithelial markers E-cadherin and Pax6 in wild-type (A, E; Wt), ␤CAT 39 (B, F), ␤CAT 10 (C, G), and APC10 (D, H) lenses at E13.5. (A–D) E-cadherin in Wt and ␤CATEx339 lenses was restricted to epithelial cells, with little or no expression in differentiating fiber cells below the equator (dashed line). However, in ␤CATEx310 and APC10 lenses, E-cadherin was more intense in epithelial cells and was distinctly detected in cells (arrows) below the equator, indicating the persistence of epithelial phenotype in the fiber compartment. (E–H) Pax6 expression in Wt and ␤CATEx339 lenses, was predominantly epithelial and decreased rapidly in differentiating fiber cells below the equator (dashed line). In ␤CATEx310 and APC10 lenses, distinct nuclear Pax6 expression was maintained in the fiber compartment (arrows). Scale bar, 100 ␮m.

PH3 expression was detected in a few anterior epithelial cells. cyclin-dependent kinase inhibitor p57kip2 showed that there was By contrast, ␤CATEx310 (Fig. 8C) and APC10 (Fig. 8D) lenses delayed cell cycle exit in these cells. In wild-type and ␤CATEx339 contained PH3-positive cells in both the epithelial layer and in lenses, p57kip2 expression was detected in cells just below the the fiber compartment, suggesting persistence of cell cycle lens equator (Figs. 8I, 8J). However, in ␤CATEx310 and APC10 activity in this normally postmitotic lens compartment. mutants, many of the epithelioid cells below the equator were In E13.5 wild-type and ␤CATEx339 lenses, cyclin D1 was negative for p57kip2 expression (Figs. 8K, 8L) reflecting a delay strongly expressed in cells of the germinative and early transi- in cell cycle exit and concomitant fiber differentiation. tional zones (Figs. 8E, 8F) with expression decreasing as cells Ex3 undergo terminal differentiation. Lenses of ␤CAT 10 and Increased Apoptosis in ␤CATEx310 and APC10 mice, however, displayed abnormal and persistent cy- APC10 Lenses clin D1 expression in cells of the fiber compartment (Figs. 8G, 8H), suggesting persistence of the cell cycle. Consistent with Disruptions of cell cycle often lead to apoptosis. TUNEL assays these observations of BrdU, PH3, and cyclin D1, analysis of the showed that in wild-type lenses at E13.5 and at E15.5, apopto-

Ex3 Ex3 FIGURE 5. Expression of the fiber cell markers ␤-crystallin and c-Maf in wild-type (A, E; Wt), ␤CAT 39 (B, F), ␤CAT 10 (C, G), and APC10 (D, H) lenses at E13.5. ␤-Crystallin expression in Wt (A) and ␤CATEx339 (B) lenses was restricted to fiber cells below the equator (dashed line), with no expression in anterior epithelial cells. In ␤CATEx310 (C) and APC10 (D) lenses, ␤-crystallin expression in the fiber compartment was delayed (arrowheads). c-Maf expression in Wt (E) and ␤CATEx339 (F) lenses was evident as intense nuclear reactivity as cells leave the germinative zone (arrow) and cross the equator (dashed line). In ␤CATEx310 (G) and APC10 (H) lenses, c-Maf reactivity was decreased and delayed (arrowheads). Scale bar, 100 ␮m.

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4801

FIGURE 6. The ECM markers, laminin and collagen in Wt (A, D), ␤CATEx310 (B, E), and APC10 (C, F) lenses at E13.5. (A, D) In wild-type (Wt) lenses, laminin and collagen IV were predominantly detected in epithelial cells and in the posterior lens capsule (arrowheads). Distinct reactivity was also detected in the vitreous chamber (✽) and RPE (r). In ␤CATEx310 (B, C) and APC10 (E, F) lenses, reactivity for both laminin and collagen IV was increased in the ﬁber cell compartment (arrows) below the equator (dashed line), but was less reliably detected in the posterior lens capsule. Scale bar, 100 ␮m.

sis was virtually absent (Figs. 9A, 9D). However, in both mutant However, immunostaining for ␣-sma of ␤CATEx310 and APC10 lenses at E13.5 and E15.5 there were abundant TUNELϩ cells in lenses showed discrete expression in numerous cells (Figs. 10B, both the epithelial and fiber cell compartments (Figs. 9B, 9C, 10D), which had a highly elongated morphology and resembled 9E, 9F). myofibroblasts. In ␤CATEx310 mutants, ␣-sma-positive cells were detected in both the anterior epithelial and posterior fiber com- EMT in Mutant Lenses partments. By contrast in APC10 lenses, ␣-sma-positive cells were The morphology of cells in the mutant lenses, particularly the detected only in the anterior part of the lens. APC10 lenses at E15.5, was clearly abnormal (Fig 2H). As activa- ␤ tion of the Wnt/ -catenin pathway has been associated with EMT, DISCUSSION we examined the expression of the mesenchymal marker, ␣-smooth muscle actin (␣-sma). Normally, distinct fibrillar reactiv- In several studies, ␤-catenin and the canonical Wnt signal- ity for ␣-sma was not detectable in wild-type lenses (Fig. 10A). ing have been found to participate in lens develop-

Ex3 Ex3 Ex3 FIGURE 7. BrdU incorporation in wild-type (A; Wt), ␤CAT 39 (B), ␤CAT 10 (C), and APC10 (D) lenses at E13.5. In Wt (A) and ␤CAT 39 (B) lenses, BrdU incorporation occurred only in the anterior epithelial cells (arrows). However, in ␤CATEx310 (C) and APC10 (D) lenses, pronounced BrdU incorporation was also detected in the fiber compartment (arrowheads). (E, F) Quantification of BrdU-positive cells as the percentage of total cells shows increased cell proliferation in both the epithelial and fiber compartments of ␤CATEx310 (E) and APC10 (F) lenses compared with Wt. Scale bar, 100 ␮m.

Downloaded from iovs.arvojournals.org on 09/25/2021 4802 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

Ex3 FIGURE 8. Expression of the cell cycle markers phosphohistone3 (PH3), cyclin D1, and p57kip2 in wild-type (A, E, I; Wt), ␤CAT 39 (B, F, J), ␤CATEx310 (C, G, K), and APC10 (D, H, L) lenses at E13.5. PH3 (an M-phase marker) was detected only in the epithelial cells of Wt (A) and ␤CATEx339 (B) lenses. In ␤CATEx310 (C) and APC10 (D) lenses, PH3-positive cells were detected in the fiber and epithelial compartments. Cyclin ␤ Ex3 ␤ Ex3 D1 (G1-S phase transition) in Wt (E) and CAT 39 (F) lenses was detected in epithelial and equatorial cells; however, in CAT 10 (G) and APC10 (H) lenses, pronounced cyclin D1 reactivity was detected throughout the fiber compartment. p57kip2 (cell cycle exit) in Wt (I) and ␤CATEx339 (J) lenses, p57kip2 was detected only in cells that had exited the cell cycle below the lens equator (dashed line) and decreased as fibers differentiated (✽). However, in ␤CATEx310 (K) and APC10 (L) lenses, p57kip2 was delayed (arrows) in cells below the equator. Scale bar, 100 ␮m.

ment.13,15,27,31,46 During early stages of lens induction, Wnt ing the Wnt/␤-catenin signaling pathway during early lens signals in the presumptive lens ectoderm must be suppressed differentiation. for lens formation to occur.15,27 However, during lens differ- 31 entiation, inhibition of Wnt signals by loss of Lrp6 or by loss Effect of Wnt Pathway Activation on Epithelial ␤ 13 of -catenin results in abnormal lens differentiation, charac- and Early Differentiating Fiber Cells terized by deﬁcits in the epithelium due to diminished cell proliferation and defective ﬁber cell differentiation with re- Cre-mediated recombination at the APC580S allele33 results in a duced levels of ␤-crystallin. These studies suggest that there mutant APC protein that inactivates the ␤-catenin degradation are distinct stages of Wnt activity during lens induction and complex and thus constitutively activates the Wnt/␤-catenin differentiation. In this study, we used conditional mutations pathway. Similarly, deletion of exon-3 of ␤-catenin results in a of Catnb and Apc alleles to examine the effects of overactivat- protein that is resistant to ubiquitination and degradation and

FIGURE 9. Cell death (TUNEL) in wild-type (A, D; Wt), ␤CATEx310 (B, E), and APC10 (C, F) lenses at E13.5 and E15.5. TUNELϩ nuclei were rarely detected in Wt lenses at E13.5 (A), but were present in both ␤CATEx310 (B) and APC10 (C) lenses (arrowheads) and were more prevalent in APC10 lenses. At E15.5, no TUNELϩ nuclei were detected in Wt lenses (D) but were abundant in ␤CATEx310 lenses (E), particularly in the ﬁber compartment, where there were accumulations of degenerating cells (✽) that were TUNELϩ. In APC10 lenses at E15.5 (F) occasional TUNELϩ nuclei were detected. Scale bar, 100 ␮m.

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4803

whereas the ␤-catenin exon-3 mutation resulted in lenses with recognizable structure and numerous epithelial-like cells in the fiber compartment. However, examination of the phenotypes at E13.5, just after gene recombination occurs, revealed that initially the phenotypes were highly similar and appeared to be due to Wnt-pathway–mediated effects. Subsequently, from E15.5 onward, the phenotypes in these mice started to diverge. Cells in E15.5 APC10 lenses ceased to proliferate, whereas cells in the E15.5 ␤CATEx310 lens continued to proliferate. The patterns of cell differentiation and apoptosis also varied. The APC mutation resulted in the loss of a large part of the carboxyl domains of the APC protein, with functions not only in Wnt signaling but also in microtubule association (via Kap3/KIF and EB1 domains), actin binding (via ASEF and PDZ domains) and nuclear export and localization (via nuclear export sequences and nuclear localization sequences).18,48,49 Therefore, APC10 lenses may have Wnt-pathway–independent defects in pro- cesses regulated by these domains, such as cell migration, proliferation, and death, thus accounting for the divergence in phenotype between APC10 and ␤CATEx310 lenses. Consistent with these putative roles of APC in cytoskeletal dynamics and migration, there are subtle differences in phenotype already apparent at E15.5 (Figs. 2, 10), suggestive of migration defects. In APC10 lenses, the epithelial cells accumulate at the anterior margin as opposed to the posterior compartment in Ex3 Ex3 FIGURE 10. Expression of ␣-sma in Wt (A), ␤CAT 10 (B), and ␤CAT 10 lenses. Similarly, the ␣-sma-positive cells in the APC10 (C, D) lenses at E15.5. (A) In wild-type (Wt) eyes, ␣-sma was not APC10 lenses are restricted to the anterior compartment but detectable in the lens, but was present in the muscle of the eyelids and are distributed in both compartments of ␤CATEx310 lenses. Ex3 in the iris (arrowheads). In ␤CAT 10 and APC10 lenses there were Studies are in progress to investigate in more detail the phe- ␣ ␤ Ex3 numerous -sma-reactive cells (arrowheads). In CAT 10 lenses (B) notypes of these mice at stages after E15.5 to determine the they were found in both anterior and posterior compartments; however, in APC10 lenses (D), they were localized in the abnormal multi- contribution of these other functions of APC to the mutant layered anterior cells. Many of the cells that were immunoreactive phenotypes. were elongated with a distinct spindle shape. (C) In APC10 lenses, Consistent with previous studies, Cre-mediated mutation of many of the central degenerating cells in the fiber compartment Apc and CatnbEx3 alleles resulted in overactivation of the Wnt stained nonspecifically (✽) with nonimmune serum (NIS). Scale bars: signaling pathway in ␤CATEx310 and APC10 lenses. In both (A, B) 100 ␮m; (C, D)75␮m. models, there was increased expression and nuclear localization of ␤-catenin at E13.5, which is when the recombination is complete. Further analyses by real-time PCR at E14.5 were also also constitutively activates this pathway. In this study, we consistent with activation of Wnt signaling, showing dramatic took advantage of the specificity of two lens-specific transgene promoters13,32 that drive Cre expression in the whole lens upregulation of Wnt target genes (Lef1, Brachyury, Tcf7, and (MLR10) or in the lens fibers only (MLR39) to determine cyclin D1), canonical Wnt inhibitor genes (Wif1, Dkk1, Frzb, whether there were any lens compartment-specific effects of kremen1, and Nkd1) and putative noncanonical Wnt pathway activating the Wnt pathway. The lack of phenotype in mice genes (Rhou, Wnt4, Wnt5a, Wnt11, and Fz8). The upregula- derived from the MLR39 line indicates that activation of the tion of noncanonical Wnt signaling components is intriguing, as these pathways have been reported to be antagonistic Wnt pathway in differentiating fiber cells of the lens cortex 50 does not alter their fate or the differentiation process. Similarly, to canonical Wnt signaling. It has been suggested that the our previous study has shown that loss of ␤-catenin in these Wnt/planar cell polarity pathway functions in regulating cytoskeletal polarity and cell adhesion during fiber cell dif- cells, which are already committed to differentiate, also does 51,52 not affect lens fiber differentiation.13 Lyu and Joo47 showed ferentiation, thus increased expression of components in that in vitro, Wnt signals can act synergistically with FGFs to this pathway could compound the phenotypes observed in ␤ Ex3 mediate expression of ␤B2-crystallin in lens epithelial cells CAT 10 and APC10 mice. It is unclear whether the ob- undergoing fiber differentiation. In the present study we did served upregulation of several Wnt inhibitors (Wif1, Dkk1, not investigate whether there were any alterations in the ratio Nkd1, Frzb, and Sfrp4) in these lenses is due to direct activa- of crystallins expressed in these lenses. However, the normal tion by ␤-catenin at Lef/Tcf sites or whether it reflects a more total ␤-crystallin reactivity and the lack of any morphologic general feedback response to the overactivation of ␤-catenin effects on fiber differentiation in APC39 and ␤CATEx339 mice signaling. An intriguing finding is that, Sfrp2, a Wnt inhibitor suggest that any synergistic effects of Wnt signals with FGFs that was not upregulated in ␤CATEx310 and APC10 lenses, has are likely to occur before the stage of fiber differentiation at been shown in transgenic overexpression experiments to bind which Cre becomes active in these mice. Wnt7a and to affect cytoskeletal organization of differentiating By contrast, both lines generated with the MLR10 line, lens fiber cells,52 suggesting that distinct Wnt-Frz and antago- APC10, and ␤CATEx310, showed severe microphthalmia, indi- nist combinations regulate canonical and noncanonical signals cating that overactivation of the Wnt pathway has profound in the lens. effects on lens epithelial cells and early differentiating fibers in In this study we limited our analyses predominantly to the transitional zone. The end-stage phenotypes observed in E13.5, the stage at which gene recombination in the lens cells weanling and newborn mice suggest that truncation of APC is complete and thus only investigated the immediate effects of and loss of ␤-catenin exon-3 result in different pathologies. The activating the Wnt/␤-catenin pathway. This method also al- APC mutation resulted in more severely degenerate lenses, lowed comparison with our previous analyses of mice with

Downloaded from iovs.arvojournals.org on 09/25/2021 4804 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

conditional deletion of ␤-catenin in epithelium and fibers to to promote expression of ␤B2-crystallin. Taken together, these contrast with loss of Wnt/␤-catenin signals.13 data suggest that Wnt signals are essential for regulating cell cycle progression and the exit from the cell cycle (via p57kip2) ␤ Effect of Activation of the Wnt Pathway on necessary for differentiation. However, initiation of -crystallin Epithelial Phenotype and Fiber Differentiation expression during fiber differentiation is independent of Wnt/ ␤-catenin signals and is probably dependent on FGF4 and BMP5 In both APC10 and ␤CATEx310 mice at E13.5, the lens pheno- signals. Once differentiation is initiated, low levels of endoge- type was characterized by overproliferation of lens epithelial nous Wnt/␤-catenin signals appear to be necessary to enhance cells and their delayed differentiation. The cells maintained ␤-crystallin expression.13,47 However, overactivation of Wnt/ expression of the epithelial markers E-cadherin and Pax6, but ␤-catenin signals inhibits ␤-crystallin expression. Thus, FGF had delayed expression of the fiber cell markers ␤-crystallin and Wnt signals must be tightly regulated and coordinated and c-maf. Moreover analysis of cell cycle markers indicated during lens development, particularly at the lens equator, to that these epithelioid cells in the fiber compartment failed to ensure progression/arrest of cell cycle and optimal differenti- exit the cell cycle (lack of p57Kip2 expression) and continued ation of lens fiber cells.

to progress through the G1, S, and M phases of the cell cycle. 13 ␤ Similarly, our previous studies showed that loss of -cateinin Effect of Wnt Pathway on EMT and Apoptosis in the lens results in cell cycle arrest in G1. These observations are consistent with other studies showing that Wnt/␤-catenin The abnormal levels of cell proliferation in APC10 and signals regulate cell proliferation in the brain53 and other ␤CATEx310 lenses were accompanied by elevated levels of tissues.37,38,54 apoptosis. Similar induction of apoptosis has been reported in The maintenance of lens epithelial phenotype by activation various other tissues with similar conditional mutations of APC of Wnt signaling in this study is consistent with in vitro stud- and ␤-catenin.37,59 The mechanism by which active Wnt sig- ies,55 which assessed the effects of LiCl on lens epithelial nals induces apoptosis in lens cells is unclear. It may be due to explants. LiCl is known to inhibit GSK3␤ and thus potentially dysregulation of cell cycle checkpoints and DNA damage, as activates Wnt signaling. In an intriguing finding, although ad- active ␤-catenin and APC truncations have been show to cause dition of LiCl induced migratory lens cells to adopt a more spindle and chromosomal segregation abnormalities.23,60–63 epithelial phenotype, with stabilization of tight junctions and Alternatively, alterations in the composition of the basement E-cadherin-based adhesion junctions, there was decreased cell membrane may have initiated an anoikis response. Consistent proliferation and inhibition of ␣-sma expression in response to with this, most epithelioid cells undergoing apoptosis are not TGF␤. This result contrasts with the increased proliferation in contact with the lens capsule (Fig. 9) and there are alter- and induction of ␣-sma in this study in response to activation of ations in the expression of laminin and collagen IV (Fig. 6), Wnt signaling. As indicated by these workers, it is likely that which may result in altered cell-matrix interactions. Truncated LiCl has effects other than just activating the Wnt pathway and APC also targets to mitochondria where it binds Bcl2 and thus may also affect signaling via Src kinase and expression of Slug may directly affect intrinsic apoptotic pathways.64 In colon and Smad4.55 Similarly, the in vivo context of the current cancer cell lines, truncated APC appears to have antiapoptosis studies cannot account for other signaling pathways that influ- effects64,65 and binding of truncated APC with Bcl2 is hypoth- ence lens cell responses. In addition, active Wnt signaling in esized to provide a survival advantage. In APC10 lenses, the vivo, detected using Lef/TCF reporters, only occurs in lens at levels of apoptosis were less extensive than in ␤CATEx310 E13–14. The lack of activity of these reporters in the lens at lenses. However, it seems unlikely that truncated APC confers later postnatal stages, from which explants are usually ob- improved cell survival in these lenses, as the phenotype at tained, suggests that the differences may also be stage depen- postnatal stages is more severe. dent. Activation of the Wnt/␤-catenin signaling pathway has The results presented herein indicate that inappropriate been associated previously with EMT during embryonic de- Wnt/␤-catenin signals also impact on early stages of fiber cell velopment and cancer metastasis.66,67 In both APC10 and differentiation. Exit from the cell cycle is essential for fiber cell ␤CATEx310 lenses, evidence for EMT was present at E15.5, differentiation and is temporally linked with ␤-crystallin ex- with numerous cells showing ␣-sma expression. However, as pression during normal fiber differentiation.56 Key regulators this occurred 2 days after genetic recombination and initiation of crystallin expression are the Maf transcription factors.57 of Wnt signaling at E13.5, it is not clear whether the EMT in Overactivation of Wnt/␤-catenin signals in APC10 and these lenses is a direct effect of Wnt/␤-catenin signaling or ␤CATEx310 lenses results in delayed cell cycle exit (p57kip2 secondary to potential activation of TGF␤ signaling in these expression), reduced c-maf and delayed ␤-crystallin expres- abnormal lenses. Although TGF␤ signaling is the major inducer sion. By contrast, loss of ␤-catenin results in precocious ex- of EMT in the lens,4,12 there is increasing evidence that EMT is pression of p57kip2 and c-maf in epithelial cells, yet normal mediated by a network of various signaling pathways.67,68 initiation of ␤-crystallin expression in fiber cells, albeit at de- TGF␤ signaling can result in the activation of Wnt-responsive creased levels.13 These results indicate that Wnt signals regu- genes by translocation of ␤-catenin to the nucleus after late cell cycle exit and c-maf expression. Although c-maf ex- E-cadherin downregulation. Similarly, TGF␤ has also been pression and cell cycle exit are associated with ␤-crystallin shown to upregulate Wnts and Frizzleds in lens and cause expression in the fiber compartment, precocious expression of nuclear translocation of ␤-catenin in lens epithelial cells.69 The c-maf and p57kip2 is insufficient to induce ␤-crystallin expres- results presented herein suggest that Wnt/␤-catenin signaling is sion in the epithelial compartment, indicating that other sig- an initiator of EMT in the lens, which is a characteristic of nals (FGF and BMPs) and transcription factors are necessary to subcapsular cataract and posterior capsule opacification. initiate ␤-crystallin expression. Indeed, previous studies of Maf In conclusion, this study showed that Wnt/␤-catenin signal- proteins indicate that they can act as repressors or activators of ing plays important roles in regulating lens stem/progenitor crystallin expression, depending on context and expression of cell populations and in maintaining lens epithelial cell pheno- other transcription factors such as Pax6 and Sox1 to -3.57,58 type. Wnt/␤-catenin signaling also plays important roles during Consistent with the notion that Wnt signals regulate crystallin early stages of fiber differentiation, regulating c-maf and ␤-crys- expression, Lyu and Joo47 showed, in epithelial explants, that tallin expression. In this context, it remains to be determined Wnt3a-conditioned medium synergizes with low levels of FGF2 how Wnt and FGF signals combine to regulate the precise

Downloaded from iovs.arvojournals.org on 09/25/2021 IOVS, October 2009, Vol. 50, No. 10 Role of Wnt Signaling in Lens Development 4805

temporal and spatial sequence of events during lens fiber cell 24. Moseley JB, Bartolini F, Okada K, Wen Y, Gundersen GG, Goode differentiation. Dysregulation of Wnt signaling also affects the BL. Regulated binding of adenomatous polyposis coli protein to EMT of lens cells and thus may afford targets for pharmaceu- actin. J Biol Chem. 2007;282:12661–12668. tical intervention for cataract and posterior capsule opacification. 25. Tirnauer JS. A new cytoskeletal connection for APC: linked to actin through IQGAP. Dev Cell. 2004;7:778–780. 26. Akiyama T, Kawasaki Y. Wnt signalling and the actin cytoskeleton. References Oncogene. 2006;25:7538–7544. 27. Kreslova J, Machon O, Ruzickova J, et al. Abnormal lens morpho- 1. McAvoy JW, Chamberlain CG, de Iongh RU, Hales AM, Lovicu FJ. genesis and ectopic lens formation in the absence of beta-catenin Lens development. Eye. 1999;13:425–437. function. Genesis. 2007;45:157–168. 2. Coulombre JL, Coulombre AJ. Lens development: fiber elongation 28. Miller LA, Smith AN, Taketo MM, Lang RA. Optic cup and facial and lens orientation. Science. 1963;142:1489–1490. patterning defects in ocular ectoderm beta-catenin gain-of-function 3. Lang RA, McAvoy JW. Lens induction and determination. In: mice. BMC Dev Biol. 2006;6:14. Lovicu FJ, Robinson ML, eds. Development of the Ocular Lens. 29. Liu H, Mohamed O, Dufort D, Wallace VA. Characterization of Wnt Cambridge, UK: Cambridge University Press; 2004:261–289. signaling components and activation of the Wnt canonical path- 4. Lovicu FJ, McAvoy JW. Growth factor regulation of lens develop- way in the murine retina. Dev Dyn. 2003;227:323–334. ment. Dev Biol. 2005;280:1–14. 30. Liu H, Thurig S, Mohamed O, Dufort D, Wallace VA. Mapping 5. Boswell B, Overbeek PA, Musil LS. Essential role of BMPs in canonical Wnt signaling in the developing and adult retina. Invest FGF-induced secondary lens fiber differentiation. Dev Biol. 2008; Ophthalmol Vis Sci. 2006;47:5088–5097. 324:202–212. 31. Stump RJ, Ang S, Chen Y, et al. A role for Wnt/beta-catenin 6. de Iongh RU, Lovicu FJ, Overbeek PA, et al. Requirement for signaling in lens epithelial differentiation. Dev Biol. 2003;259:48– TGFbeta receptor signaling during terminal lens fiber differentia- 61. tion. Development. 2001;128:3995–4010. 32. Zhao H, Yang Y, Rizo CM, Overbeek PA, Robinson ML. Insertion of 7. Faber SC, Dimanlig P, Makarenkova HP, Shirke S, Ko K, Lang RA. a Pax6 consensus binding site into the alphaA-crystallin promoter Fgf receptor signaling plays a role in lens induction. Development. acts as a lens epithelial cell enhancer in transgenic mice. Invest 2001;128:4425–4438. Ophthalmol Vis Sci. 2004;45:1930–1939. 8. Faber SC, Robinson ML, Makarenkova HP, Lang RA. Bmp signaling 33. Shibata H, Toyama K, Shioya H, et al. Rapid colorectal adenoma is required for development of primary lens fiber cells. Develop- formation initiated by conditional targeting of the Apc gene. Sci- ment. 2002;129:3727–3737. ence. 1997;278:120–123. 9. Furuta Y, Hogan BL. BMP4 is essential for lens induction in the 34. Harada N, Tamai Y, Ishikawa T, et al. Intestinal polyposis in mice mouse embryo. Genes Dev. 1998;12:3764–3775. with a dominant stable mutation of the beta-catenin gene. EMBO 10. Robinson ML. An essential role for FGF receptor signaling in lens J. 1999;18:5931–5942. development. Semin Cell Dev Biol. 2006;17:726–740. 35. Nathke IS, Adams CL, Polakis P, Sellin JH, Nelson WJ. The adeno- 11. Zhao H, Yang T, Madakashira BP, et al. Fibroblast growth factor matous polyposis coli tumor suppressor protein localizes to receptor signaling is essential for lens fiber cell differentiation. Dev plasma membrane sites involved in active cell migration. J Cell Biol. 2008;318:276–288. Biol. 1996;134:165–179. 12. de Iongh RU, Wederell E, Lovicu FJ, McAvoy JW. Transforming 36. Brocardo M, Nathke IS, Henderson BR. Redefining the subcellular growth factor-beta-induced epithelial-mesenchymal transition in location and transport of APC: new insights using a panel of the lens: a model for cataract formation. Cells Tissues Organs. antibodies. EMBO Rep. 2005;6:184–190. 2005;179:43–55. 37. Sansom OJ, Reed KR, Hayes AJ, et al. Loss of Apc in vivo immedi- 13. Cain S, Martinez G, Turner K, et al. Differential requirement for ately perturbs Wnt signaling, differentiation, and migration. Genes beta-catenin in epithelial and fiber cells during lens development. Dev. 2004;18:1385–1390. Dev Biol. 2008;321:420–433. 38. Colnot S, Decaens T, Niwa-Kawakita M, et al. Liver-targeted dis- 14. de Iongh RU, Abud HE, Hime GR. WNT/Frizzled signaling in eye ruption of Apc in mice activates beta-catenin signaling and leads to development and disease. Front Biosci. 2006;11:2442–2464. hepatocellular carcinomas. Proc Natl Acad SciUSA.2004;101: 17216–17221. 15. Smith AN, Miller LA, Song N, Taketo MM, Lang RA. The duality of 39. Liu F, Chu EY, Watt B, et al. Wnt/beta-catenin signaling directs beta-catenin function: a requirement in lens morphogenesis and multiple stages of tooth morphogenesis. Dev Biol. 2008;313:210– signaling suppression of lens fate in periocular ectoderm. Dev 224. Biol. 2005;285:477–489. 40. Zhang Y, Andl T, Yang SH, et al. Activation of beta-catenin signal- 16. Bienz M, Hamada F. Adenomatous polyposis coli proteins and cell ing programs embryonic epidermis to hair follicle fate. Develop- adhesion. Curr Opin Cell Biol. 2004;16:528–535. ment. 2008;135:2161–2172. 17. Nathke IS. The adenomatous polyposis coli protein: the Achilles 41. Krawetz R, Kelly GM. Wnt6 induces the specification and epithe- heel of the gut epithelium. Annu Rev Cell Dev Biol. 2004;20:337– lialization of F9 embryonal carcinoma cells to primitive endoderm. 366. Cell Signal. 2008;20:506–517. 18. Aoki K, Taketo MM. Adenomatous polyposis coli (APC): a multi- 42. Schubert FR, Mootoosamy RC, Walters EH, et al. Wnt6 marks sites functional tumor suppressor gene. J Cell Sci. 2007;120:3327–3335. of epithelial transformations in the chick embryo. Mech Dev. 19. Mogensen MM, Tucker JB, Mackie JB, Prescott AR, Nathke IS. The 2002;114:143–148. adenomatous polyposis coli protein unambiguously localizes to 43. Carmon KS, Loose DS. Secreted frizzled-related protein 4 regulates microtubule plus ends and is involved in establishing parallel two Wnt7a signaling pathways and inhibits proliferation in endo- arrays of microtubule bundles in highly polarized epithelial cells. metrial cancer cells. Mol Cancer Res. 2008;6:1017–1028. J Cell Biol. 2002;157:1041–1048. 44. Kawauchi S, Takahashi S, Nakajima O, et al. Regulation of lens fiber 20. Barth AI, Caro-Gonzalez HY, Nelson WJ. Role of adenomatous cell differentiation by transcription factor c-Maf. J Biol Chem. polyposis coli (APC) and microtubules in directional cell migration 1999;274:19254–19260. and neuronal polarization. Semin Cell Dev Biol. 2008;19:245–251. 45. Wigle JT, Chowdhury K, Gruss P, Oliver G. Prox1 function is 21. Langford KJ, Askham JM, Lee T, Adams M, Morrison EE. Examina- crucial for mouse lens-fibre elongation. Nat Genet. 1999;21:318– tion of actin and microtubule dependent APC localisations in living 322. mammalian cells. BMC Cell Biol. 2006;7:3. 46. Nadauld LD, Chidester S, Shelton DN, et al. Dual roles for adeno- 22. Collin L, Schlessinger K, Hall A. APC nuclear membrane associa- matous polyposis coli in regulating retinoic acid biosynthesis and tion and microtubule polarity. Biol Cell. 2008;100:243–252. Wnt during ocular development. Proc Natl Acad SciUSA.2006; 23. Green RA, Wollman R, Kaplan KB. APC and EB1 function together 103:13409–13414. in mitosis to regulate spindle dynamics and chromosome align- 47. Lyu J, Joo CK. Wnt signaling enhances FGF2-triggered lens fiber ment. Mol Biol Cell. 2005;16:4609–4622. cell differentiation. Development. 2004;131:1813–1824.

Downloaded from iovs.arvojournals.org on 09/25/2021 4806 Martinez et al. IOVS, October 2009, Vol. 50, No. 10

48. Senda T, Shimomura A, Iizuka-Kogo A. Adenomatous polyposis 59. Hasegawa S, Sato T, Akazawa H, et al. Apoptosis in neural crest coli (Apc) tumor suppressor gene as a multifunctional gene. Anat cells by functional loss of APC tumor suppressor gene. Proc Natl Sci Int. 2005;80:121–131. Acad SciUSA.2002;99:297–302. 49. McCartney BM, Nathke IS. Cell regulation by the Apc protein Apc 60. Dikovskaya D, Schiffmann D, Newton IP, et al. Loss of APC induces as master regulator of epithelia. Curr Opin Cell Biol. 2008;20:186– polyploidy as a result of a combination of defects in mitosis and 193. apoptosis. J Cell Biol. 2007;176:183–195. 50. Veeman MT, Axelrod JD, Moon RT. A second canon: functions and 61. Caldwell CM, Green RA, Kaplan KB. APC mutations lead to cyto- mechanisms of beta-catenin-independent Wnt signaling. Dev Cell. kinetic failures in vitro and tetraploid genotypes in Min mice. J Cell 2003;5:367–377. Biol. 2007;178:1109–1120. 51. Chen Y, Stump RJ, Lovicu FJ, McAvoy JW. A role for Wnt/planar 62. Draviam VM, Shapiro I, Aldridge B, Sorger PK. Misorientation and cell polarity signaling during lens fiber cell differentiation? Semin reduced stretching of aligned sister kinetochores promote chro- Cell Dev Biol. 2006;17:712–725. mosome missegregation in EB1- or APC-depleted cells. EMBO J. 52. Chen Y, Stump RJ, Lovicu FJ, Shimono A, McAvoy JW. Wnt signal- 2006;25:2814–2827. ing is required for organization of the lens fiber cell cytoskeleton 63. Dikovskaya D, Newton IP, Nathke IS. The adenomatous polyposis and development of lens three-dimensional architecture. Dev Biol. coli protein is required for the formation of robust spindles formed 2008;324:161–176. in CSF Xenopus extracts. Mol Biol Cell. 2004;15:2978–2991. 53. Zechner D, Fujita Y, Hulsken J, et al. beta-Catenin signals regulate 64. Brocardo M, Lei Y, Tighe A, Taylor SS, Mok MT, Henderson BR. cell growth and the balance between progenitor cell expansion Mitochondrial targeting of adenomatous polyposis coli protein is and differentiation in the nervous system. Dev Biol. 2003;258: stimulated by truncating cancer mutations: regulation of Bcl-2 and 406–418. implications for cell survival. J Biol Chem. 2008;283:5950–5959. 54. Rulifson IC, Karnik SK, Heiser PW, et al. Wnt signaling regulates pancreatic beta cell proliferation. Proc Natl Acad SciUSA. 65. Tighe A, Johnson VL, Taylor SS. Truncating APC mutations have 2007;104:6247–6252. dominant effects on proliferation, spindle checkpoint control, survival and chromosome stability. J Cell Sci. 2004;117:6339– 55. Stump RJ, Lovicu FJ, Ang SL, Pandey SK, McAvoy JW. Lithium stabilizes the polarized lens epithelial phenotype and inhibits 6353. proliferation, migration, and epithelial mesenchymal transition. 66. Vincan E, Barker N. The upstream components of the Wnt signal- J Pathol. 2006;210:249–257. ling pathway in the dynamic EMT and MET associated with colo- 56. Lovicu FJ, McAvoy JW. Spatial and temporal expression of rectal cancer progression. Clin Exp Metastasis. 2008;25:657–663. p57(KIP2) during murine lens development. Mech Dev. 1999;86: 67. Moustakas A, Heldin CH. Signaling networks guiding epithelial- 165–169. mesenchymal transitions during embryogenesis and cancer pro- 57. Yang Y, Cvekl A. Large Maf transcription factors: cousins of AP-1 gression. Cancer Sci. 2007;98:1512–1520. proteins and important regulators of cellular differentiation. Ein- 68. Thiery JP, Sleeman JP. Complex networks orchestrate epithelial- stein J Biol Med. 2007;23:2–11. mesenchymal transitions. Nat Rev Mol Cell Biol. 2006;7:131–142. 58. Kataoka K. Multiple mechanisms and functions of maf transcrip- 69. Chong CC, Stump RJ, Lovicu FJ, McAvoy JW. TGFbeta promotes tion factors in the regulation of tissue-specific genes. J Biochem. Wnt expression during cataract development. Exp Eye Res. 2009; 2007;141:775–781. 88(2):307–313.

Downloaded from iovs.arvojournals.org on 09/25/2021