SEARCH

SEARCH BY CITATION

Keywords:

  • Canonical Wnt pathway;
  • Retina;
  • Ciliary marginal zone;
  • Stem cell;
  • Proliferation;
  • Xenopus laevis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Vertebrate retinal stem cells, which reside quiescently within the ciliary margin, may offer a possibility for treatment of degenerative retinopathies. The highly proliferative retinal precursor cells in Xenopus eyes are confined to the most peripheral region, called the ciliary marginal zone (CMZ). Although the canonical Wnt pathway has been implicated in the developing retina of different species, little is known about its involvement in postembryonic retinas. Using a green fluorescent protein-based Wnt-responsive reporter, we show that in transgenic Xenopus tadpoles, the canonical Wnt signaling is activated in the postembryonic CMZ. To further investigate the functional implications of this, we generated transgenic, hormone-inducible canonical Wnt pathway activating and repressing systems, which are directed to specifically intersect at the nuclear endpoint of transcriptional Wnt target gene activation. We found that postembryonic induction of the canonical Wnt pathway in transgenic retinas resulted in increased proliferation in the CMZ compartment. This is most likely due to delayed cell cycle exit, as inferred from a pulse-chase experiment on 5-bromo-2′-deoxyuridine-labeled retinal precursors. Conversely, repression of the canonical Wnt pathway inhibited proliferation of CMZ cells. Neither activation nor repression of the Wnt pathway affected the differentiated cells in the central retina. We conclude that even at postembryonic stages, the canonical Wnt signaling pathway continues to have a major function in promoting proliferation and maintaining retinal stem cells. These findings may contribute to the eventual design of vertebrate, stem cell-based retinal therapies.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Author contributions: T. Denayer: conception and design, manuscript writing, collection and/or assembly of data, data analysis and interpretation; M.L.: conception and design, collection and/or assembly of data, data analysis and interpretation; T. Denayer and M.L. contributed equally to this work; C.B.: collection and/or assembly of data, data analysis and interpretation; T. Deroo and A.H.: generation of specific plasmid constructs; S.J.: collection and/or assembly of data; F.V.R.: financial support; M.P.: conception and design, manuscript writing, collection and/or assembly of data, data analysis and interpretation, final approval of manuscript, financial support; K.V.: conception and design, manuscript writing, data analysis and interpretation, final approval of manuscript, financial support.

Quiescent retinal stem cells are present within the ciliary epithelium (a structure located between the retina and the iris) of adult birds and mammals, including humans [1, [2], [3], [4], [5]6]. These cells represent a potential tool for developing stem cell transplantation therapies to treat retinopathies [7].

Fish and amphibians are excellent models for molecular investigation of retinal stem cells. Their retinas grow continuously throughout their lifetime, keeping pace with the increasing size of the larval and adult animals. New types of retinal cells are generated by stem and progenitor cells located in the ciliary marginal zone (CMZ) [6, 8, 9]. These stem cells are multipotent, giving rise to all types of retinal neurons and glia [10]. Within the CMZ, self-renewing stem cells are the closest to the periphery and highly proliferating progenitors lie in the middle, whereas cells at the central edge have stopped dividing and entered the differentiation process [11, 12].

Several pathways, including Notch [13], fibroblast growth factor [2], and Hedgehog (Hh) [14, [15], [16], [17], [18], [19], [20], [21]22], were reported to sustain retinal stem cell properties. Another important cell-extrinsic factor involved in regulating development of the eye and retina is the canonical Wnt pathway [23, 24]. This evolutionarily conserved pathway has crucial roles in numerous developmental processes, where it orchestrates alterations in gene expression that can influence cell proliferation, survival, and fate. In addition, canonical Wnt signaling is involved, during adult life, in the regulation of several postnatal stem/progenitor populations [25, [26]27]. Constitutive activation of the Wnt pathway in humans is implicated in the initiation of various oncogenic malignancies [27, [28]29] and in ocular diseases [24]. Canonical Wnt signaling is initiated by binding of a Wnt ligand to its cognate receptors, Fz and LRP-5/-6, resulting in stabilization of β-catenin and its accumulation in the cytoplasm. Subsequently, β-catenin translocates to the nucleus and, upon functional association with transcription factors of the LEF-1/TCF family, induces the transcriptional activation of specific Wnt target genes, such as Cyclin D1 [30, 31]. In the absence of Wnts, cytoplasmic β-catenin is phosphorylated and thereby targeted for ubiquitination and proteasomal degradation, whereas in the nucleus the transcription of prospective Wnt targets is repressed through binding of LEF-1/TCF factors [32, 33]. Notably, besides its signaling function, β-catenin also plays a role at the adherens junctions, where it directly binds to the classic cadherins [34].

In the developing central nervous system and in the adult subventricular zone, the canonical Wnt pathway regulates the proliferation of neural precursors [35, [36], [37], [38]39, 40]. This mitogenic effect may be in part regulated through direct transcriptional control of Cyclin D1 [35]. However, canonical Wnt signaling was also reported to promote neuronal differentiation of neural progenitor cells [41], indicating that precursor cell modulation by canonical Wnt signaling might be stage- or context-dependent.

From the dynamic expression patterns of various canonical Wnt pathway components, including Wnts, Fzs, and Wnt antagonists, an essential role in vertebrates during early eye development became evident [23, 42]. Mice and zebrafish transgenic for a LEF-1/TCF-responsive reporter exhibit active canonical Wnt signaling in the peripheral part of the developing retina [43, 44]. In the developing chick retina, LEF-1 expression is strong at the periphery and gradually decreases toward the central retina [45], a pattern consistent with Wnt reporter activity [46]. Chick (C)Wnt-2b is expressed in the peripheral CMZ region and likely signals via the canonical Wnt pathway [45]. Overexpression of CWnt-2b promotes prolonged proliferation of retinal progenitors and suppresses neuronal differentiation [45, 47, 48]. This presumably occurs by downregulation of proneural and neurogenic gene expression, independently of cell cycle progression [48]. Conversely, expression of dominant-negative LEF-1 blocks cell proliferation and causes premature cell differentiation [45]. In contrast, when expressed ectopically in the early chick retina, CWnt2b or constitutively active β-catenin triggers the conversion of retinal cells into peripheral cells, whereas loss-of-function of the canonical Wnt pathway has the opposite effect [46]. Wnt-2b was further proposed to be essential for laminar organization of the chick retina [47]. Regarding the embryonic mouse retina, Liu et al. [49] demonstrated that ectopic expression of stabilized β-catenin did not induce proliferation but led to transdifferentiation into peripheral cells. However, in vitro studies in mammals support the implication of the canonical Wnt pathway, via Wnt3a or other pathway components, in promoting the proliferation of explanted adult retinal stem/progenitor cells [50, 51] and in contributing to regeneration following injury [52].

In embryonic Xenopus retinas, XFz5 is specifically confined to the CMZ region [53] and acts through canonical Wnt signaling [54]. When XFz5 is blocked in the developing retina, Sox2 expression is reduced, resulting in decreased proliferation of retinal progenitors, loss of proneural gene expression, and bias of the retinal progenitors toward a glial fate [54]. Early eye formation in zebrafish was also shown to depend on the coordinated activity of Fz5 and the canonical Wnt pathway [55]. Ectopic expression of stabilized β-catenin in developing zebrafish retinas results in inhibition of cell cycle exit and increased proliferation [44].

These data point to canonical Wnt signaling as a mitogenic regulator of the embryonic retinal stem/progenitor cells in the developing eye. Nonetheless, the signaling events or underlying mechanisms that are involved in vivo during postembryonic stages remain largely unknown and warrant further investigation. In addition, phenotypes obtained by knockout or overexpression of (stabilized) β-catenin may be complicated by the role that β-catenin plays in cadherin-mediated cell adhesion [56].

In this study, we demonstrated in postembryonic Xenopus retinas that canonical Wnt pathway activity is restricted to the CMZ. To further explore the functional role of the Wnt pathway in the retina, we developed and optimized a hormone-inducible transgenic system that can activate the canonical Wnt pathway at its transcriptional level, making it possible to uncouple the analysis from the earlier developmental events in which the canonical Wnt pathway is involved. Because of its inducibility, this elegant system is particularly relevant for studying postembryonic retinal stages. Also, in contrast to overexpression of stabilized β-catenin, this system cannot interfere with cadherin-mediated cell adhesion. In addition to our gain-of-function strategy, we deployed an inducible Wnt repressing transgenic system. Our results indicate that canonical Wnt signaling regulates the postembryonic proliferation potential of retinal stem/progenitor cells. As inferred from a pulse-chase experiment on 5-bromo-2′-deoxyuridine (BrdU)-labeled retinal precursors, we propose that cell cycle exit regulation is the underlying mechanism. Finally, our data suggest that canonical Wnt signaling-dependent control of postembryonic retinoblast proliferation is, at least in part, mediated by XFz5, similar to the embryonic situation.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Construction of Plasmids

The construction of the pCMV-LEFΔNVP16GR, pVP16, pC-VP16, pEnR, pC-EnR, and pDnFz5 plasmids is presented in supplemental online Materials and Methods.

Cell Culture, Transient Transfection, and TKTOP Reporter Assay

Human embryonic kidney (HEK293T) and Xenopus XTC cells were cultured, transfected, and analyzed in a TKTOP reporter assay as described in supplemental online Materials and Methods.

In Vitro Transcription and mRNA Injection

Constructs were linearized with NotI, and capped RNA was synthesized using SP6 RNA polymerase as described [57]. Eggs were collected, fertilized, and dejellied as described [58]. Embryos at the four-cell stage were injected with 25 pg of mRNA (5 nl) in a single blastomere. Injection was in the ventral marginal region for scoring axis duplication and in the animal pole region for reverse transcription (RT)-polymerase chain reaction (PCR) analysis. During injection, embryos were incubated as described [58].

Transgenesis Procedure

Generation of transgenic Xenopus laevis embryos was carried out essentially as described [58]. pVP16, pC-VP16, and pDnFz5 were linearized with ApaI; pEnR and pC-EnR were linearized with SpeI.

Dexamethasone Treatment

Dexamethasone (Dex) was used at 10 μM, and embryos or cells were treated as indicated.

Western Blot Expression Analysis

Total protein lysates were prepared, and proteins were separated, transferred, and detected as described previously [57].

RNA Extraction and Semiquantitative RT-PCR Analysis

RNA extraction and semiquantitative RT-PCR analysis are presented in supplemental online Materials and Methods.

Histology

TOPTKiGFP transgenic tadpoles were fixed in 4% paraformaldehyde (PFA)/phosphate-buffered saline (PBS), cryosectioned, and viewed directly as previously described [58].

BrdU Labeling

Stage 46 embryos were injected in the abdomen with 100 nl of BrdU (5-Bromo-2′-Deoxyuridine Labeling and Detection kit I; Roche Molecular Biochemicals, Indianapolis, http://www.roche-applied-science.com), fixed in 4% PFA/PBS, after 6 hours or 7 days, and cryostat sectioned at 12 μm.

Immunohistochemistry and Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling Analysis

Immunohistochemistry was performed as previously described [14]. Several primary antibodies were used: monoclonal anti-BrdU (Roche Molecular Biochemicals), polyclonal anti-phosphorylated histone 3 (anti-P-H3) (Euromedex), monoclonal anti-syntaxin (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), polyclonal anti-glycine (ImmunoSolution), polyclonal anti-CRALBP, polyclonal anti-rhodopsin (R2-12N), and monoclonal anti-calbindin (Swant, Bellinzona, Switzerland, http://www.swant.com). Specific binding sites were visualized using anti-mouse or anti-rabbit fluorescent secondary antibodies (Alexa 488 or 594; Molecular Probes, Eugene, OR, http://probes.invitrogen.com). Sections were then stained in Hoechst solution (Sigma-Aldrich) and mounted in Fluorsave mounting medium (Calbiochem, San Diego, http://www.emdbiosciences.com) or in 4,6-diamidino-2-phenylindole-containing Vectashield mounting medium (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Cell apoptosis was assayed using the In Situ Cell Death Detection kit, TM red (Roche Molecular Biochemicals).

In Situ Hybridization

Digoxigenin-labeled antisense RNA probes were generated according to the manufacturer's instructions (Roche Molecular Biochemicals). Whole-mount in situ hybridization was carried out as previously described [59], except for proteinase K treatment, which was lengthened to 1 hour. Embryos were then cryostat-sectioned (14 μm).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Canonical Wnt Signaling Activity in Postembryonic Retinal Stem/Progenitor Cells of X. laevis

Published experiments on mice, zebrafish, and chick have shown that canonical Wnt pathway activity during embryonic stages is confined to the peripheral region of the retina [43, 44, 46]. However, activity of the canonical Wnt pathway during postembryonic retinal stages has so far not been determined. To answer this question, we used X. laevis as a model, in which we transgenically introduced the TOPTKiGFP Wnt reporter tool, a fluorescent tracer that can be induced at any stage of development by adding Dex to the water [58]. At stage 40, all structural layers in the Xenopus retina are established, and thereafter it is considered postembryonic. Dex treatment of stage 46 TOPTKiGFP transgenic tadpoles demonstrated strong activity of the Wnt reporter in several tissues and organs, including the eyes. Histological sectioning revealed that the canonical Wnt signaling activity in the retina is confined to the CMZ, where it can be detected both centrally and in the most peripheral region, known to harbor retinal stem cells (Fig. 1). This indicates that the canonical Wnt pathway continues to be activated in the retinal mitotic cells throughout the postembryonic stages. Because it is not possible to sharply delimitate the border between the stem cell area and the progenitor region in the CMZ, both populations will be referred to together as precursor cells.

thumbnail image

Figure Figure 1.. Canonical Wnt signaling is specifically active in the ciliary marginal zone (CMZ) of postembryonic tadpoles. Cryosections of a TOPTKiGFP transgenic Xenopus tadpole (stage 46) treated with dexamethasone for 24 hours reveal enhanced green fluorescent protein (eGFP) expression in the retinal CMZ. Pictures show 4,6-diamidino-2-phenylindole staining (A), eGFP expression (B), and overlay of both (C). Insets are higher magnifications of the regions indicated by the white rectangles. Scale bar = 100 μm.

Download figure to PowerPoint

Conditional Interference with the Canonical Wnt Pathway at Postembryonic Stages

To understand the underlying role of canonical Wnt signaling in the postembryonic CMZ, we investigated the effect of ectopic activation versus repression of the pathway, specifically at these postembryonic retinal stages. We designed a binary transgenic system that transcriptionally activates the canonical Wnt pathway at its nuclear endpoint of target gene activation. The system remains quiescent until it is externally triggered at a desired stage of development. The active component of the system is a chimeric fusion of the DNA-binding domain of LEF-1 (LEFΔN; consisting mainly of the high mobility group [HMG] box) and the transactivation domain of herpes simplex protein VP16. This construct, which has been described by Vleminckx et al. [60] as an efficient activator of canonical Wnt target genes at the transcriptional level, was C-terminally linked to the Dex-responsive hormone-binding domain of the human glucocorticoid receptor (GR). This rendered this fusion, LEFΔNVP16GR, hormone-inducible by Dex (Fig. 2A). As assayed in vitro in HEK293T cells cotransfected with the TKTOP-Luc Wnt reporter, pCMV-LEFΔNVP16GR induces an ∼10-fold increase in Wnt pathway activity in the presence of Dex (Fig. 2B). Its hormone inducibility, specificity, and effectiveness were further validated by an axis duplication assay in RNA-injected X. laevis embryos [61] and an RT-PCR analysis for the known Wnt target gene siamois (Fig. 2C, 2D).

thumbnail image

Figure Figure 2.. Overview of the optimized, hormone-inducible, transgenic systems designed to either activate or repress the canonical Wnt pathway and validation of the Wnt pathway activating system. (A): Schematic structure of the Wnt pathway-activating fusion, pCMV-LEFΔNVP16GR, and its control counterpart, pCMV-NLSVP16GR. ▾, unique sites for linearizing the vectors for RNA in vitro transcription. (B): TKTOP-Luc reporter assay in HEK293T cells transiently cotransfected with the indicated fusion construct or multicomponent vector, in the presence or absence of Dex. The bars represent average values ± SD. (C): Axis duplication assay following ventral marginal microinjection of LEFΔNVP16GR or NLSVP16GR RNA in a single blastomere of four-cell-St. X. laevis embryos. Pictures were taken at the early tadpole and neurula St. (insets). Axis duplication was observed only in LEFΔNVP16GR-injected embryos treated with Dex, whereas normal development occurred in other conditions. (D): Reverse transcription-polymerase chain reaction (RT-PCR) analysis for the siamois Wnt target gene. Siamois expression was clearly induced in Dex-treated animal caps isolated from blastula St. embryos that had been injected at the four-cell St. with LEFΔNVP16GR RNA. EF-1α was amplified as a control. (E): Schematic overview of pVP16, pC-VP16, pEnR, and pC-EnR. ▾, unique sites used to linearize the vectors for transgenesis. (F): eGFP expression at St. 41, revealing the bona fide transgenic embryos. Shown are bright-field (left) and fluorescent (right) images. (G): Western blot analysis of LEFΔNVP16GR and NLSVP16GR expression in HEK293T cells transfected with the respective multicomponent vectors and treated with Dex for 24 hours. EGFP (27 kDa) was expressed upon transfection of the multicomponent vectors. Detection of actin (43 kDa) served as a loading control. (H): Western blot expression analysis of LEFΔNVP16GR in transgenic pVP16 embryos that were selected on the basis of their eGFP expression (bottom pictures) and incubated without or with Dex for 24 hours before analysis. WT embryos were included as negative control. LEFΔNVP16GR expression was clearly induced in the Dex-treated transgenic tadpoles. Minor levels of LEFΔNVP16GR protein were observed in the noninduced transgenic embryo at St. 25, which, however, had very high EGFP expression, likely reflecting high copy-number transgene integration. Detection of actin served as a loading control. (I): RT-PCR analysis of the En-2 Wnt target gene expression in St. 12.5 transgenic pVP16 and WT embryos treated with Dex for 14 hours. The expression of En-2 is strongly induced in the transgenic embryo compared with the WT. EF-1α was amplified as control. Abbreviations: 14xUAS, 14 repeats of upstream activating sequences; CMV, cytomegalovirus promoter; Dex, dexamethasone; E1b, carp basal E1b promoter; EF-1α, Xenopus EF-1α promoter; eGFP, enhanced green fluorescent protein; EnR, transrepression domain of Drosophila Engrailed; GAL4, DNA-binding domain of GAL4 yeast transactivator; GR, hormone-binding domain of human glucocorticoid receptor; HA, hemagglutinin epitope; hECpr, minimal human E-cadherin promoter; LEFΔN, mouse LEF-1 DNA-binding domain; NI, noninjected; NT, nontransfected cells; RLU, relative light units; RT-, negative control reaction in which reverse transcriptase was omitted; St., stage; VP16, transactivation domain of the herpes simplex virus VP16 protein; WT, wild-type.

Download figure to PowerPoint

To further ensure effective fusion dormancy in the absence of hormone and thereby avoid any leakage in transgenic embryos, the expression of LEFΔNVP16GR was also made Dex-dependent by using a GAL4/upstream activating sequence (UAS) binary induction loop as an extra level of temporal control. The system comprises a Dex-inducible GAL4VP16GR transactivator under transcriptional control of the ubiquitously active Xenopus EF-1α promoter (Fig. 2E). Downstream in this cassette, LEFΔNVP16GR is transcriptionally regulated by the binding of active GAL4VP16GR to a 14xUAS in combination with a minimal E1b promoter [62]. In addition, a cassette containing a minimal human E-cadherin promoter driving enhanced green fluorescent protein (eGFP) expression was incorporated as a selection marker to recognize the genuine nonmosaic transgenic embryos (Fig. 2F). In summary, both the expression and the functionality of the LEFΔNVP16GR fusion are triggered by adding Dex to the tadpole swimming water, hence ensuring tightly controlled activity. This Wnt pathway activating system is hereafter referred to as pVP16. Since the HMG box of LEF-1/TCF transcription factors represents the main DNA-binding domain [63, 64], its deletion theoretically disrupts the capacity to recognize and bind DNA. Therefore, in parallel, we constructed a control multicomponent vector, pC-VP16, which lacks this DNA-binding domain but, importantly, still possess a nuclear localization signal to allow nuclear translocation of the fusion (Fig. 2E). Including this tight control system likely bans putative aspecific side effects that can be caused by modular elements present in the multicomponent vector.

The hormone inducibility of pVP16 was confirmed in a Wnt reporter assay. As expected, its negative controls (pCMV-NLSVP16GR and pC-VP16) could not induce TKTOP luciferase reporter activity regardless of the presence or absence of Dex (Fig. 2B). In the absence of Dex, pVP16 showed less basal activity than pCMV-LEFΔNVP16GR, and it reached a lower activity in the presence of Dex, possibly because of a delayed effect of the extra GAL4/UAS loop. However, relative induction (with vs. without Dex) resembled that for pCMV-LEFΔNVP16GR. A Western blot expression analysis, in both transfected cells and transgenic embryos, confirmed that the translation of LEFΔNVP16GR and NLSVP16GR is Dex-inducible (Fig. 2G, 2H). Furthermore, RT-PCR analysis on pVP16 transgenic embryos revealed the Dex-induced transcriptional activation of En-2 (Fig. 2I), a known direct target of canonical Wnt signaling during postgastrula Xenopus development [65]. Notably, in the absence of Dex, pVP16 transgenic embryos develop normally (supplemental online Fig. 1), confirming that our Wnt pathway-activating system is tightly regulated.

Similarly, we used a Dex-inducible Wnt interfering system in which LEFΔN was fused to the transrepression domain of Drosophila Engrailed (EnR) and coupled to the GR domain. Expression of this construct was also made inducible via a GAL4VP16GR/UAS loop as described [57], and it will henceforth be referred to as pEnR. A control vector, lacking the DNA-binding domain and called pC-EnR, was also designed (Fig. 2E). Hence, this optimized transgenic platform, including an inducible gain- and loss-of-function system along with its stringent controls, is ideally positioned for further exploration of the canonical Wnt pathway during the postembryonic stages in various organs and tissues, including the retina.

Postembryonic Activation of the Canonical Wnt Pathway in the Xenopus Retina

The canonical Wnt pathway has emerged as a critical player in regulating the proliferation of stem/progenitor cells in the developing retinas of different species, including Xenopus [44, 48, 54]. We therefore investigated whether this function is maintained beyond the embryonic stages. To that end, we generated pVP16 and pC-VP16 F0 transgenic X. laevis embryos using nuclear transfer. The bona fide transgenic embryos were selected on the basis of their eGFP expression. To functionally induce the system, Dex was added from stage 40 onward, when all structural layers in the Xenopus retina are established and the retina is considered postembryonic. Dex-treated wild-type embryos were also included as negative controls. At stage 46, transgenic and control tadpoles were pulse-labeled for 6 hours with BrdU, and positive cells were then quantified on retinal sections. Transcriptional activation of the Wnt pathway in the postembryonic retina significantly increased the number of S-phase cells in the dorsal and ventral CMZ, in comparison with transgenic pC-VP16 and wild-type tadpoles (Fig. 3A, 3B). A significant increase in the number of cells positive for the mitotic marker P-H3 was also discernable (Fig. 3A, 3C). This was not due to enhanced cell survival, as terminal deoxynucleotidyl transferase dUTP nick-end labeling analysis revealed no difference in cell apoptosis in the retinas of pVP16 transgenics compared with those of the controls (supplemental online Fig. 2A, 2B).

thumbnail image

Figure Figure 3.. Canonical Wnt signaling control precursor cell proliferation in the postembryonic CMZ. (A–C): Transcriptional activation of canonical Wnt signaling. (D–F): Repression of canonical Wnt signaling. (A, D): BrdU and P-H3 staining of stage 46 retinas from WT and transgenic tadpoles as indicated, treated with dexamethasone from stage 40 onward. Sections were counterstained with Hoechst or DAPI. In each series, the lower panels are higher-magnification images of the regions indicated by the white squares in the upper images. (B,C, E,F): Quantification of BrdU-positive (B,E) and P-H3-positive (C,F) cells in the CMZ per section. Values are given as mean ± SEM, and significant differences between means were analyzed by Student's t-test (***, p < .0001; **, p < .002; *, p < .02). Scale bars = 100 μm. Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; CMZ, ciliary marginal zone; DAPI, 4,6-diamidino-2-phenylindole; P-H3, phosphorylated histone 3; WT, wild-type.

Download figure to PowerPoint

To test whether increased BrdU incorporation and P-H3 labeling in the CMZ could result from delayed cell cycle exit, we set up a pulse-chase experiment in which wild-type tadpoles were injected with BrdU at stage 46 and fixed 3, 5, 7, or 10 days later. Three days after the pulse, BrdU-labeled cells were predominantly confined to the CMZ and then progressively moved toward the central retina, incorporating into the different layers as they exited the cell cycle and differentiated (Fig. 4A, 4B; data not shown). Quantitative analysis in transgenics at day 7 revealed more BrdU-positive cells in pVP16 tadpoles (Fig. 4C, 4F), as observed above after a 6-hour BrdU pulse. Strikingly, we found that the proportion of BrdU-labeled cells that remained in the CMZ was also significantly higher than in the controls (Fig. 4G). In addition, CMZ size was greatly increased (Fig. 4E), whereas no changes in cell apoptosis levels could be detected (supplemental online Fig. 2C, 2D). These data collectively constitute a strong indication that activation of the canonical Wnt pathway maintains precursor cells longer in a proliferative state.

thumbnail image

Figure Figure 4.. Retinal precursors show delayed cell cycle exit but are correctly directed toward retinal cell layers upon canonical Wnt pathway activation. Pulse-chase experiment of BrdU-labeled cells in WT, pC-VP16, or pVP16 embryos, treated with dexamethasone from stage 40 onward. Tadpoles were injected with BrdU at stage 46 and fixed 7 days later. (A–C): Sections of retinas showing integration of BrdU-labeled cells in the ganglion and in the inner and outer cell layers. Note an extreme case of CMZ enlargement in the pVP16 retina (indicated by a dotted line). Lower panels are higher magnifications of the dorsal part of the retina. The central limit of the CMZ was determined with respect to the initiation of the plexiform layers (white dotted line). (D): Quantification of BrdU-positive cells in the different cell layers, showing equivalent distribution in all conditions. (E–F): Quantification of total CMZ cell number in the dorsal and ventral retina, as estimated by counting Hoechst-labeled nuclei (E) and total BrdU cell number (F). (G): Percentage of BrdU-labeled cells remaining in the CMZ following the pulse. Values are given as mean ± SEM, and significant differences between means were analyzed by Student's t test (**, p < .01; *, p < .05). Scale bar = 100 μm. Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; CMZ, ciliary marginal zone; WT, wild-type.

Download figure to PowerPoint

We next wondered whether precursors in which the Wnt pathway is activated correctly generate the different retinal cell types. Analysis of the proportion of BrdU-labeled cells (among those that have exited the CMZ) within the ganglion cell layer, the inner nuclear layer, or the photoreceptor layer revealed equivalent distribution in all conditions (Fig. 4D). Although we cannot exclude differences in cell type proportions within the inner nuclear layer (amacrine, bipolar, horizontal, and Müller cells), these data suggest that activation of the canonical Wnt signaling at postembryonic stages does not influence the proliferation of particular lineage-committed progenitor cells, nor does it affect retinoblast fate determination.

Importantly, next to the increase in CMZ cell proliferation, differentiated cells in the central retina and pigmented epithelium of pVP16 embryos did not seem to be affected. In line with this, immunostaining for markers of photoreceptors, amacrine cells, Müller glia cells, inner plexiform layer, and optic nerve was similar to that of control embryos (Figs. 5A–5R′). Quantification of the number of rods, cones, and glycinergic amacrine cells also showed no significant difference (Fig. 5S–5U). These results indicate that activation of the canonical Wnt pathway in the postembryonic retina specifically maintains retinal stem/progenitor cells in a proliferative state but does not affect differentiated cells.

thumbnail image

Figure Figure 5.. Transcriptional activation of the canonical Wnt pathway in the postembryonic retina does not affect differentiated cells. (A–R): Immunofluorescence analysis of cell type-specific markers in stage 46 retinas of WT (A–F), pC-VP16 (G–L), and pVP16 (M–R) embryos treated with dexamethasone from stage 40 onward. (A′–R′) show only the immunostaining, whereas (A–R) merge it with Hoechst staining. Expression of calbindin (cone marker), rhodopsin (rod marker), CRALBP (Müller glial cell marker), glycine (glycinergic amacrine cell marker), and syntaxin (inner plexiform layer and optic nerve marker) were not significantly affected by transcriptional activation of the canonical Wnt pathway. (C″, I″, O″) show higher magnifications of the dotted square regions delineated in (C′, I′, O′). Scale bars = 50 μm (F, L,R) and 100 μm (all other panels). (S–U): Graphs indicating the average numbers of calbindin-, R2-12N-, and glycine-positive cells per retinal section in each condition. Values are given as mean ± SEM. No significant differences were found between WT and pC-VP16 or between pC-VP16 and pVP16 (Student's t-test). Abbreviation: WT, wild-type.

Download figure to PowerPoint

Postembryonic Repression of the Canonical Wnt Pathway in the Xenopus Retina

In a complementary approach exploring the implications of transcriptional repression of Wnt target genes during postembryonic retinal stages, we generated pEnR and pC-EnR transgenic Xenopus tadpoles. They were also concomitantly treated with Dex from stage 40 onward and were BrdU-pulse-labeled for 6 hours at stage 46. Transcriptional repression of canonical Wnt signaling significantly reduced the number of BrdU-positive cells in the dorsal and ventral CMZ of pEnR tadpoles compared with the controls (Fig. 3D, 3E). This significant decrease in proliferation was also demonstrated by P-H3 staining of M-phase cells (Fig. 3D, 3F). Noticeably, no increase in cell death could be detected in the retina of Dex-induced pEnR tadpoles (supplemental online Fig. 2E, 2F). In addition, as in the gain-of-function experiment, pEnR retinas did not exhibit any obvious defects in laminar organization or retinal cell marker expression (supplemental online Fig. S3). Together, and in accordance with the data obtained by the gain-of-function approach, these results suggest that the canonical Wnt pathway has a specific and prominent function in promoting the proliferation and maintenance of retinal precursor cells at postembryonic stages but does not seem to alter differentiated retinal cells.

Comparison of Embryonic and Postembryonic Stem Cells

To address the question of whether different or similar mechanisms regulate embryonic versus postembryonic retinal stem cells, we looked at components of the canonical Wnt pathway that have been implicated in embryonic retinogenesis. We first analyzed the status of Cyclin D1, a known direct target gene of canonical Wnt signaling, upon activation or repression of the Wnt pathway. In situ hybridization revealed a strong decrease of Cyclin D1 expression in the retina of stage 46 pEnR transgenic tadpoles (Fig. 6A–6C). In contrast, we could not detect any obvious changes in the intensity of Cyclin D1 staining in pVP16 CMZ (Fig. 6D–6F). However, we observed that the size of the expression domain was broader in pVP16 retinas than in controls, consistent with the observed increase in the CMZ size (Fig. 6G).

thumbnail image

Figure Figure 6.. Regulation of Cyclin D1 expression upon repression or activation of the canonical Wnt pathway. (A–F): Expression of Cyclin D1 on cryostat retinal sections after whole-mount in situ hybridization performed on stage 46 WT or transgenic tadpoles as indicated, treated with dexamethasone from stage 40 onward. (A–C): Repressing canonical Wnt signaling leads to a strong decrease of Cyclin D1 expression in the CMZ (C). (D–F): Transcriptional activation of the canonical Wnt pathway does not obviously modify relative levels of Cyclin D1 expression but expands the size of its expression domain compared with control retinas (F). Scale bar = 25 μm. (G): Graph indicating the average area per section (arbitrary units, measured using ImageJ software) of Cyclin D1 expression domain in each condition. Values are given as mean ± SEM, and a significant difference could be observed between pC-VP16 and pVP16 (Student's t-test; ***, p < .001). Abbreviations: CMZ, ciliary marginal zone; WT, wild-type.

Download figure to PowerPoint

Second, we asked whether postembryonic proliferation in the CMZ could be mediated by Fz5, as this Wnt receptor has been shown to be involved in embryonic retinogenesis [54]. We cloned a dominant-negative Xenopus Fz5 receptor (DnFz5) into a Dex-inducible GAL4VP16GR/UAS loop, resulting in the pDnFz5 construct (Fig. 7A). Its efficiency and inducibility were verified in a luciferase in vitro Wnt reporter assay. As expected, upon cotransfection with XWnt5a and XFz5, pDnFz5 efficiently repressed Wnt pathway activity only in the presence of Dex (Fig. 7B). Next, we generated pDnFz5 transgenics and analyzed BrdU incorporation in the CMZ. We found a significant decrease in BrdU incorporation similar to that observed in pEnR transgenics (Fig. 7C, 7D). Again, no obvious defects in laminar organization or retinal cell marker expression could be detected in the central retina of pDnFz5 transgenic tadpoles (Fig. 7E–7G). Together, these data suggest that the canonical Wnt pathway controls postembryonic retinal precursor proliferation at least in part through XFz5.

thumbnail image

Figure Figure 7.. Inhibition of Fz5 function results in decreased cell proliferation in the ciliary marginal zone (CMZ) at postembryonic stages but does not affect differentiated cells. (A): Schematic overview of the inducible pDnFz5 construct. ▾, unique sites used to linearize the vectors for transgenesis. (B): Validation of pDnFz5 by a TKTOP-Luc reporter assay in XTC cells transiently cotransfected with the indicated constructs, in the presence or absence of Dex. Upon Dex addition, pDnFz5 repressed the Wnt pathway activity induced by XWnt5a and XFz5, similar to pEnR. The bars represent average values ± SD. (C): Hoechst and BrdU staining of stage 46 retinas from WT and pDnFz5 transgenics, treated with Dex from stage 40 onward. Scale bar = 50 μm. (D): Quantification of the BrdU-labeled cells in the dorsal and ventral CMZ, per section. Values are given as mean ± SEM, and significant differences could be observed between pDnFz5 transgenic and control tadpoles (Student's t test; *, p < .05). (E): Immunofluorescence analysis of cell type-specific markers (glycine, syntaxin, calbindin) in stage 46 retinas of WT and pDnFz5 transgenic tadpoles treated with Dex from stage 40 onward. The lower images present a merge with a Hoechst staining. Scale bar = 100 μm. (F,G): Graphs indicating the average numbers of glycine- and calbindin-positive cells per retinal section in each condition. Values are given as mean ± SEM. No significant difference could be observed between WT and pDnFz5 (Student's t-test). Abbreviations: 14xUAS, 14 repeats of upstream activating sequences; BrdU, 5-bromo-2′-deoxyuridine; Dex, dexamethasone; DnFz5, dominant-negative Xenopus Fz5; E1b, carp basal E1b promoter; EF-1α, Xenopus EF-1α promoter; eGFP, enhanced green fluorescent protein; GAL4, DNA-binding domain of GAL4 yeast transactivator; GR, hormone-binding domain of human glucocorticoid receptor; hECpr, minimal human E-cadherin promoter; RLU, relative light units; VP16, transactivation domain of the herpes simplex virus VP16 protein; WT, wild-type.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

In this study, we used several transgenic strategies to show that the canonical Wnt pathway regulates the proliferation potential of stem/progenitor cells of the postembryonic Xenopus retina. First, using a transgenic Wnt-responsive reporter, we showed that the canonical Wnt pathway is active in the CMZ of the Xenopus retina at a postembryonic tadpole stage. Second, using an inducible gain-of-function strategy, we demonstrated that ectopic activation of the canonical Wnt pathway at the level of target gene induction increases the number of proliferating cells in the CMZ by maintaining retinal precursor cells longer in a proliferative state. Third, by using an inducible loss-of-function approach, we showed that the transcriptional outcome of the canonical Wnt pathway is required to maintain cell proliferation within the CMZ. Importantly, activation or repression had no effect on the differentiated cells in the central retina or the laminar organization or on retinal cell death. Finally, we illustrate that this role of the canonical Wnt pathway in the proliferation of postembryonic retinal progenitor/stem cells is mediated at least in part by XFz5.

Several Wnt ligands and Fz receptors, along with Wnt antagonists, are dynamically expressed in the peripheral and more central retina, as has been demonstrated for successive stages of mouse and chick retinal development [23, 42]. Most Wnts and Fzs are not restricted to signaling exclusively via the canonical β-catenin-dependent pathway or through the noncanonical β-catenin-independent pathway [66, 67]. Hence, proposals for functional implications of canonical or noncanonical downstream Wnt signaling should not be based only on expression studies of those molecules. By contrast, the use of a reporter containing specific consensus LEF-1/TCF sites is a suitable and faithful tool for studying the dynamics of canonical Wnt signaling. It has been reported that integration of a Wnt-responsive reporter in the mouse, zebrafish, and chick results in expression of the reporter in the peripheral retinal region of the developing retina [43, 44, 46]. However, the functional implications of canonical Wnt signaling in the postembryonic retina have so far not been fully addressed.

To further examine the implications of this postembryonic Wnt activity in the Xenopus CMZ, we used here functional interference by inducibly activating or inhibiting the canonical Wnt pathway. Temporal control of transgene expression and function allows the separation of later postembryonic phenotypes from events that occurred earlier in development. Our constructs act at the endpoint of the canonical Wnt pathway. This is important since many components of the Wnt pathway are also involved in other cellular processes and not all Wnt signals are activating the canonical β-catenin-mediated signaling branch. This is complicating the phenotypic analysis when these components are ectopically expressed, suppressed, or mutated. In line with this, it was recently proposed that the lamination defects observed after elimination of β-catenin in the retina can be attributed to its role in cell adhesion, rather than to its function as a nuclear effector in canonical Wnt signaling [56]. Hence, our optimized and highly complementary activating and repressing systems, in parallel with their tight controls, constitute a powerful experimental system for manipulating Wnt signaling in late developmental stages. Within the scope of the present study, this transgenic platform was effectively suited for postembryonic retinal analysis of the canonical Wnt pathway.

The canonical Wnt pathway is critically involved in regulating the proliferation of postnatal stem cell in several organs and tissues [25, [26]27], including the brain [35, [36], [37]38]. In line with this, we showed that transcriptional activation of the canonical Wnt pathway in transgenic pVP16 tadpoles at postembryonic stages leads to significant increases in the number of S- and M-phase cells in the CMZ, as determined by both BrdU and P-H3 labeling. Moreover, a BrdU pulse-chase experiment indicated that this most likely relies on delayed cell cycle exit of retinal precursor cells. Conversely, transcriptional repression of the pathway resulted in significantly decreased cell proliferation in this compartment. In support of our results, Van Raay et al. [54] reported that the canonical Wnt pathway stimulates proliferation in the early embryonic Xenopus retina. In the developing chick retina, a similar role in promoting prolonged proliferation was described [45, 48]. Also, injection of stabilized β-catenin in wild-type zebrafish embryos led to increased BrdU incorporation in the retina [44]. In contrast, overexpression of stabilized β-catenin in the developing mouse retina did not stimulate cell proliferation, although it expanded the ciliary margin. Transdifferentiation from neural retina to ciliary margin was proposed to be the underlying mechanism [56]. Our data indicate that in the postembryonic Xenopus retina, canonical Wnt signaling is dedicated to the control of retinal precursor proliferation. Interestingly, although in our setting the Wnt pathway activating construct was ubiquitously expressed, the increase in cell proliferation was restricted to the CMZ, showing that the differentiated cells in the central retina do not re-enter the cell cycle. Therefore, it is not likely that cells in the central Xenopus retina dedifferentiate or transdifferentiate upon activation of canonical Wnt signaling.

As inferred from the BrdU pulse-chase experiment, the most likely mechanism used by the canonical Wnt pathway to regulate proliferation in postembryonic CMZ is control of cell cycle exit. However, we cannot exclude an additional effect on cell cycle kinetics. Recently for Hh signaling, Locker et al. [21] illustrated a dynamic mechanism controlling both cell cycle length and cell cycle exit, during retinogenesis. Such a dual control was proposed to sustain the conversion from slow-cycling stem cells to faster-cycling transit-amplifying progenitor cells. Whether a comparable or different mechanism holds true for the canonical Wnt pathway remains to be verified. It is, however, tempting to speculate that Wnt and Hedgehog pathways act in different manners on retinal precursor cells, as Hedgehog tends to bring them closer to their final division [21], whereas, as shown here, Wnt activation seems to delay cell cycle exit. Our transgenic systems can be ideal tools to address the potential interdependence of Wnt and Hh signaling in retinal cell proliferation.

Notably, neither activation nor repression of the canonical Wnt pathway affected the differentiated cells in central retina or the laminar organization or retinal cell apoptosis. Also, postembryonic activation of the canonical Wnt pathway did not bias the proliferating progenitors toward a particular lineage, as shown by the pulse-chase experiment. As such, the canonical Wnt signaling in the postembryonic retina essentially and specifically promotes proliferation and maintenance of CMZ precursor cells, without affecting already differentiated cells or influencing the fate of newly generated cells.

Mitogenic effects of the canonical Wnt pathway on neural stem/progenitor cells [35, [36], [37]38] might rely in part on Cyclin D1 regulation [38]. Cyclin D1 is a key component of G1/S transition and, importantly, a direct Wnt target gene [30, 31]. In this study, using our repressing system, we demonstrated that the canonical Wnt pathway is indeed necessary for proper expression of Cyclin D1 in the postembryonic CMZ. Using the activating system, however, no increased Cyclin D1 levels were found, but rather an expanded zone of expression, consistent with a delayed cell cycle exit, was found. In line with this, Van Raay et al. previously showed that in the developing Xenopus retina, interference with Wnt signaling did not affect Cyclin D1 expression levels but reduced its expression domain [54].

Further, consideration of other downstream Wnt pathway components that could be involved in the postembryonic Xenopus CMZ led us to XFz5 as a good candidate. This receptor is exclusively expressed in the CMZ at stage 40 [53] and signals via the canonical Wnt pathway, at least in the embryonic retina [54]. Blocking XFz5 function in the postembryonic retina revealed decreased proliferation, comparable to that observed in pEnR transgenic animals. This receptor thus likely mediates Wnt signaling in the postembryonic retina, as it does during embryonic retinogenesis. Since XFz5 is confined to the central CMZ and excluded from the stem cell area [53], this may suggest that Wnt-dependent proliferation effects essentially concern progenitor cells. However, as we showed active Wnt signaling in retinal stem cells, their proliferative properties might also be controlled by this pathway, although through additional frizzled receptors. Other pathway components remain to be unveiled.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

During the postembryonic stages, canonical Wnt pathway activity is specifically confined to the CMZ of the Xenopus retina, where stem cells reside peripherally. By using optimized transgenic systems, ectopic activation of the canonical Wnt pathway significantly increased the number of proliferative cells in this specific region, most likely because of delayed cell cycle exit. The opposite effect of blocking cell proliferation was evident upon inhibition of the pathway. Canonical Wnt signaling in the postembryonic retina was at least partly mediated by Fz5 and presumably involved regulation of Cyclin D1 expression. These results indicate that the canonical Wnt pathway plays a major role in regulating proliferation and maintenance of retinal stem/progenitor cells at postembryonic stages, implying that these findings might be of help in designing new retinal stem cell-based therapies. It was recently reported that activation of canonical Wnt signaling in the mouse can indeed promote retinal regeneration after damage, as well as in a mouse model for retinal degeneration [52]. Of potential therapeutic relevance is our observation that ectopic activation of the Wnt pathway in postembryonic retinas stimulates proliferation without significantly affecting the differentiated cells or the structural organization of the retina. If this holds true for higher vertebrates, it could offer therapeutic opportunities for temporal in situ activation of the canonical Wnt pathway to boost the residential stem cells in the retina.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

We thank Daniel Kessler, Reinhard Köster, Marc Van de Wetering, Monica Vetter, Shinichi Ohnuma, and Geert Berx for sharing plasmids, as well as N. Colley and J. Saari for providing antibodies. We are also indebted to J. Hamdache, K. Parain, and G. Van Imschoot for technical assistance and to A. Bredan for editing the manuscript. We thank the anonymous reviewers for helpful suggestions. T. Denayer and C.B. are postdoctoral fellows of the Research Foundation–Flanders (FWO) and of the Agence National de la Recherche (ANR), respectively. This work was supported by grants from the Interuniversitaire Attractiepolen, the Foundation against Cancer, and the FWO (to K.V.) and by an ANR jeunes chercheurs, Association pour la Recherche sur le Cancer, and Retina France Grants (to M.P.). T. Deroo is currently affiliated with Division of Developmental Neurobiology, National Institute for Medical Research, London, U.K.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of Potential Conflicts of Interest
  9. Acknowledgements
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
SC-07-0900_Supplemental_Data.pdf94KSupplemental Data
SC-07-0900_Supplemental_Figure_1.jpg247KSupplemental Figure 1
SC-07-0900_Supplemental_Figure_2.jpg676KSupplemental Figure 2
SC-07-0900_Supplemental_Figure_3.jpg740KSupplemental Figure 3

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.