SEARCH

SEARCH BY CITATION

Keywords:

  • Wnt signaling;
  • cell cycle;
  • CyclinD1;
  • medaka;
  • proliferation;
  • retinal development

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Differentiation of neural retinal precursor (NRP) cells in vertebrates follows an established order of cell-fate determination associated with exit from the cell cycle. Wnt signaling regulates cell cycle in colon carcinoma cells and has been implicated in different aspects of retinal development in various species. To better understand the biological roles of Wnt in the developing retina, we have used a transgenic and pharmacological approach to manipulate the Wnt signaling pathway during retinal development in medaka embryos. With the use of both approaches, we observed that during the early phase of retinal development Wnt signaling regulated cell cycle progression, proliferation, apoptosis, and differentiation of NRP cells. However, during later phases of retinal development, proliferation and apoptosis were not affected by manipulation of Wnt signaling. Instead, Wnt regulated Vsx1 expression, but not the expression of other retinal cell markers tested. Thus, the response of NRP cells to Wnt signaling is stage-dependent. Developmental Dynamics 239:297–310, 2010. © 2009 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The retina has been used extensively as a model system to understand the mechanisms of cellular proliferation and differentiation during early development of the central nervous system (CNS) in vertebrates. The layered structure of the retina is evolutionarily conserved and consists of six types of neurons plus one type of glia (Ramón y Cajal, 1892), which differentiate from NRP cells in a sequential order during embryonic development (reviewed by Livesey and Cepko, 2001). During the development of the vertebrate retina, NRP cells respond to different cues that orchestrate their proliferation and subsequent differentiation to neuronal and glial cell types. In the Xenopus retina, early exit from the cell cycle promotes differentiation of precursor cells into different cell fates pointing at the existence of an intrinsic differentiation program (Ohnuma et al., 2002). On the other hand, in chicken and other vertebrates, NRP differentiation is controlled by extrinsic cues such as FGF and sonic hedghog (Shh). In these models, the expression of the proneural gene Ath5 is initiated by FGF (Martínez-Morales et al., 2005) at the optic stalk and is propagated by Shh to spread across the retina like a wave (Masai et al., 2000; Neumann and Nuesslein-Volhard, 2000).

The Wnt signaling pathway regulates many aspects of embryonic development, particularly in the CNS where it provides instructions for cell-fate specification of neural crest stem cells (Lee et al., 2004) and dorsal pallium in the forebrain (Backman et al., 2005). In other regions of the CNS, it controls the cell-cycle machinery (Megason and McMahon, 2002); for example, β-Catenin (β-Cat) induces proliferation in the cortex (Chenn and Walsh, 2002). Several components of the Wnt signaling pathway are expressed in the retina during early eye development and Wnt activity has been detected in the retina using different reporters, suggesting that Wnt signaling participates in the correct patterning of this structure (Jasoni et al., 1999; Esteve et al., 2000; Jin et al., 2002; Fuhrmann et al., 2003; Kubo et al., 2003; Liu et al., 2003, 2006; Yokoi et al., 2003; Van Raay and Vetter, 2004; Van Raay et al., 2005; Veien et al., 2005). In medaka, at least Wnt8b is expressed in the retina in the central optic vesicle at st. 22 and the dorsal ciliary margin at st. 24 (Yokoi et al., 2003). Moreover, in zebrafish Wnt2b and Wnt8 are expressed in the dorsal retinal pigment epithelium (RPE) and there are also several members of the Tcf family of transcription factors that mediate Wnt signaling, which are expressed dynamically in the developing eye (Veien et al., 2008). Additionally, Dickkopf 3 (Dkk3), which normally functions as a secreted inhibitor of Wnt, is expressed in Müller glia and retinal ganglion cells (RGCs) and, surprisingly, acts as a positive regulator of Wnt signaling in the mouse retina (Nakamura et al., 2007). Fz5 is another member of the Wnt signaling pathway, which regulates neural differentiation in the Xenopus retina. Inhibition of Fz5 at early stages of embryo development reduces cell proliferation in the optic cup and increases the differentiation of progenitors to Müller glia (Van Raay et al., 2005). This suggests that Fz5, and thus Wnt signaling, has a specific role in cell differentiation during eye formation; however, the early inhibition of Fz5 may alter early processes of eye development that subsequently affect retinal development. In fact, in zebrafish Fz5 and Wnt11 act to suppress Wnt signaling and facilitate the formation of the eye field (Cavodeassi et al., 2005).

Although the many different functions of Wnt signaling in the retina could be attributed to the various components of the pathway, which exert specific roles during development, there are also examples where multiple roles are assigned to a single component such as Wnt2b. In the ciliary marginal zone (CMZ), Wnt2b regulates proliferation of cells in the lens (Jasoni et al., 1999) and of adult retinal stem cells (RSC) (Kubo et al., 2003). It also induces proliferation of embryonic NRP cells, thereby inhibiting their differentiation (Kubo et al., 2005). However, in other experiments, Wnt2b induces differentiation towards peripheral fates of the eye, the ciliary body, and iris (Cho and Cepko, 2006). In addition, Wnt signaling promotes proliferation of NRP cells and regeneration of explanted damaged retinas in mammals (Osakada et al., 2007). Recently, it has also been shown that Wnt signaling regulates proliferation of RSC in Xenopus tadpole eyes (Denayer et al., 2008). Thus, a number of reports have assigned different roles to Wnt signaling during CNS development and, in particular, eye development. However, there are discrepancies in the results such as those regarding the function of Wnt2b. Some of these inconsistencies may be the result of comparing the role of Wnt at different developmental stages and/or the different experimental techniques used to analyze Wnt function, particularly, the use of in vitro techniques. Therefore, it is important to elucidate the primary role of Wnt signaling during retinal development in vivo and to determine whether the function of this pathway changes at different stages.

In this study, we have performed conditional overexpression and inhibition of Wnt signaling using a genetic and a pharmacological approach. Since the overexpression of Wnt genes activates the Wnt pathway (Gordon and Nusse, 2006), we created two transgenic medaka lines for conditional gene expression using the heat shock (HS) promoter: one that activated Wnt8 and another that activated DNWnt8 expression, a truncated form of Wnt8 that functions as a specific dominant negative of the Wnt signaling pathway (Hoppler et al, 1996; Itoh and Sokol, 1999; Tada and Smith, 2000). To further validate the results obtained from the transgenic approach we also activated the Wnt signaling pathway using specific GSK3β inhibitors. Using these strategies, we have identified specific temporal roles for Wnt signaling during distinct stages of retinal development. Our results demonstrate that early in development Wnt signaling plays a role in regulating cell cycle progression, which in turn affects proliferation, apoptosis, and early differentiation of NRP cells. Subsequently, Wnt signaling regulates Vsx1 expression, but has no effect on any of the early Wnt regulated processes. Thus, Wnt signaling has different roles during retinal development and its function depends on the stage of development.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Conditional Expression of Wnt8 or DNWnt8 Modifies Wnt Signaling

To understand the role of Wnt signaling during vertebrate retinal development, we used the HS promoter to induce conditional expression of genes and manipulate Wnt signaling. We generated two constructs using the HS promoter to regulate expression of the genes Wnt8 or the C-terminal truncated form DNWnt8, which has previously been used as a specific dominant-negative form for the canonical Wnt pathway (Hoppler et al, 1996; Itoh and Sokol, 1999; Tada and Smith, 2000; Mullor et al., 2001). Both constructs were microinjected into CAB medaka embryos at the one-cell stage to produce germline transgenic animal lines. To achieve an increased transgenesis rate, the constructs contained flanking I-SceI meganuclease sites (Thermes et al., 2002), and to facilitate identification of transgenic embryos, the constructs also included Green Fluorescent Protein (GFP) under a constitutively active CSKA promoter (Fig. 1A). These constructs allowed us to conditionally manipulate the Wnt signaling pathway at precise stages of retinal development, avoiding interference with early gene functions like anterior-posterior (A-P) patterning (Erter et al., 2001; Lekven et al., 2001; Houart et al., 2002; Kudoh et al., 2002) or eye field establishment (Esteve et al., 2004; Cavodeassi et al., 2005), which could affect retinal development indirectly. We have focused on early stages of medaka eye development and retinal progenitor cells differentiation. The time points at which eyes were treated or isolated represent important differentiation processes that occur around those stages: st. 17 (optic lobe formation), st. 22 (eye vesicle), st. 26 (retinal ganglion cell differentiation and beginning of pigmentation), st. 30 (formation of the plexiform layer), and st. 35 (photoreceptor differentiation) (Kitambi and Malicki, 2008).

thumbnail image

Figure 1. Conditional expression of Wnt8 and DNWnt8 modifies Wnt signaling. A: Schematic representation of the transgene constructs with the flanking I-SceI meganuclease sites, the HS promoter, the Wnt8 or DNWnt8 genes and the CSKA promoter regulating GFP expression for transgenic embryo identification. B: RT-PCR detection of transgene expression of cDNA synthesized from st. 18 embryos after HS treatment at st. 17 and detected using transgene specific primers. Transgene expression was not detected in CAB control embryos or transgenic embryos without the HS treatment. Actin expression was used as a control and three independent RT-PCR reactions were performed using cDNA from two different experiments. C–F: WMISH detection of transgene activation in st. 18 embryos after HS at st. 17 (100% of the embryos) detected using WMISH for transgene specific probes (in purple), transgenic embryos without HS treatment (C, E), and after HS treatment (D, F). Anterior is to the right, eye vesicles are inside marked areas. For each treatment, a total of 30 embryos were used in three independent experiments. Scale bars = 100 μm. G: Axin2 (ch.1) and Lef1 expression detected using RT-PCR at st. 22 or 30 after HS treatment at st. 17 or st. 26, respectively. Actin expression was used as a control and four independent RT-PCR reactions were performed on cDNA from two independent experiments obtaining similar results. H: Axin1 and Axin2 (in chr. 1 and 8) expression detected using qRT-PCR at st. 22 after HS treatment at st. 17 or at st. 26 after HS treatment at st.30. *P < 0.01, **P < 0.001. The data represent the mean ± SD (error bars; see Experimental Procedures for details). I: Lef1 expression detected using qRT-PCR at st. 22 after HS treatment at st. 17 or at st. 26 after HS treatment at st. 30. *P < 0.01, **P < 0.001. The data represent the mean ± SD (error bars; see Experimental Procedures for details). J: CAB embryo at st. 30 after HS treatment at st. 10. Normal anterior and posterior embryonic structures are present in all embryos (100% present the normal phenotype, n = 84). K: HS-Wnt8 transgenic embryo at st. 30 after HS treatment at st. 10. Visible anterior embryonic structures were missing (64.2% present the phenotype, 35.8% seemed normal, n = 42 embryos). L: HS-DNWnt8 transgenic embryo at st. 30 after HS treatment at st. 10. Posterior embryonic structures were missing and only the head was visible (76.2% present the phenotype, 23.8% seemed normal, n = 38 embryos). Three independent experiments were performed and n is the total number of embryos. Scale bars = 100 μm. M: Axin2chr. 1 expression detected using RT-PCR at st. 22 or 26 after incubation with LiCl and Indirubin from st. 17 to st. 22 or from st. 17 to st. 26 using two different concentrations of LiCl or Indirubin. Three independent RT-PCR reactions were performed on cDNA from two independent experiments obtaining similar results. N: Axin2 (chr. 1) and Lef1 expression detected using qRT-PCR at st. 22 or 26 after incubation with 0.3 M LiCl or 10 mM Indirubin from st. 17. *P < 0.01, **P < 0.001. The data represent the mean ± SD (error bars; see Experimental Procedures for details). O: Axin2 (chr. 1) and Lef1 expression detected using qRT-PCR at st. 22 after incubation with 0.15 M LiCl or 5 mM Indirubin from st. 17. The data represent the mean ± SD (error bars; see Experimental Procedures for details).

Download figure to PowerPoint

To functionally validate our expression system, we HS-treated transgenic animals at st. 17 and analyzed transgene expression. After administering the HS treatment to the transgenic embryos, we detected activation of the transgenes using reverse transcriptase PCR (RT-PCR; Fig. 1B) and whole-mount in situ hybridization (WMISH) using specific probes for the Wnt8 and DNWnt8 transgenes (Fig. 1C–F). This demonstrated the correct function of our HS conditional expression constructs in the transgenic medaka and that exogenous Wnt8 or DNWnt8 were expressed in the embryo, including the developing eye vesicles, only after HS treatment. To verify that the exogenous genes modulated the function of endogenous Wnt signaling, we HS treated transgenic embryos at st. 17 or st. 26 and isolated their eyes at st. 22 or st. 30, respectively. We used RT-PCR and quantitative RT-PCR (qRT-PCR; Supp. Table S1, which is available online) to detect any changes in the expression of endogenous Wnt signaling target genes. One of these targets was the Wnt signaling nuclear mediator/target Lymphoid enhancer factor 1 (Lef1, Filali et al., 2002) and the other was Axin2, which associates directly with β-catenin, GSK-3β, and APC and is implicated in down-regulating Wnt signaling (Yan et al., 2001; Lustig et al., 2002; Jho et al., 2002). Both genes were detected using RT-PCR on isolated eye vesicles and eyes, although gene expression was not detected using WMISH, probably due to less sensitive to low or transient levels of gene expression (Supp. Fig. S1). Activation of the HS-Wnt8 transgene either at st. 17 or st. 26 and analysis of gene expression at st. 22 or st. 30, respectively, revealed an increase in Lef1 and Axin2 chromosome (ch.) 1 expression after HS treatment. Conversely, activation of the HS-DNWnt8 transgene at st. 17 or st. 26 and analysis of gene expression at st. 22 or st. 30, respectively, inhibited Lef1 and Axin2 expression (Fig. 1G–I). However, in medaka, Axin1 and Axin2 are duplicated with one copy each in ch. 1 and 8. In order to quantitatively determine which Axin genes were activated by Wnt signaling, we analyzed the expression of all four Axin genes using qRT-PCR (Supp. Table S1). After HS treatment at st. 17 or st. 26, eyes were collected from st. 22 or st. 30 embryos, respectively, and analyzed. HS-Wnt8 and HS-DNWnt8 increased and decreased Axin2 (ch. 1) expression, respectively, with no changes induced in Axin1 (ch. 1 and 8) or Axin2 (ch. 8; Fig. 1H). Thus, transgene activation affects the expression of the endogenous Wnt target gene Lef1 and Axin2 (ch. 1) at st. 17 and st. 26.

Finally, to test whether the transgenes could induce whole embryo phenotypes, we focused on previously described Wnt phenotypes of A-P axis patterning. The activation of HS-Wnt8 in embryos at st. 10 caused a head-less phenotype lacking anterior structures of the embryo (Fig. 1K) similar to the phenotypes caused by Wnt gain-of-function experiments in other vertebrates (Erter et al., 2001; Lekven et al., 2001; Houart et al., 2002; Kudoh et al., 2002). On the other hand, activation of HS-DNWnt8 in embryos at st. 10 yielded a head-only phenotype with a total lack of posterior structures in the embryo (Fig. 1L), similar to the phenotype produced by inhibition of the Wnt pathway (Erter et al., 2001; Lekven et al., 2001; Houart et al., 2002). These results confirm that Wnt signaling can be manipulated using our transgenic constructs, which, upon HS treatment, express functional Wnt8 and DNWnt8 proteins.

Wnt Signaling Regulates NRP Cell Differentiation During Early Stages of Eye Development

In the developing medaka retina, β-Catenin (β-Cat) was detected in and near the CMZ at st. 24 (Supp. Fig. S2D), where Wnt signaling is active (Kubo et al., 2003, 2005). At earlier stages, high levels of β-Cat were detected in the retina and nuclear β-Cat could be detected in 17.8 ± 1.2% of the cells following a scattered pattern, whereas in the lens no nuclear β-Cat was detected (Supp. Fig. S2A, B). This result indicates that Wnt signaling is active in the retina and may be acting in a transient manner as described in mice (Fuhrmann et al., 2009). To investigate the role of Wnt signaling in the differentiation of NRP cells in the medaka retina, we analyzed the expression of neuronal differentiation markers in the retina after manipulation of Wnt signaling. As described in other vertebrates (Prada et al., 1991; Cayouette et al., 2006; Wallace, 2008), medaka NRP cells follow a temporal pattern of differentiation: Ath5 expression appears first in NRP cells at st. 25 (Supp. Fig. S3) and Ath5-expressing cells will give rise to RGCs (Brown et al., 1998; Kanekar et al., 1998) and horizontal, amacrine, and photoreceptor cells (Poggi et al., 2005). Then, Prox1 begins to be expressed at st. 27 (Supp. Fig. S3) and will label amacrine, horizontal, and bipolar cells (Dyer et al., 2003). Next to be expressed is Vsx1 at st. 30 (Supp. Fig. S3) and Rhodopsin at st. 34 (Supp. Fig. S3). Vsx1 will label bipolar cells in zebrafish (Passini et al., 1998) and Rhodopsin will label rod photoreceptors (Vihtelic et al., 1999; Kitambi and Malicki, 2008). Due to the multiple retinal cell lineages that can arise from NRP cells expressing these markers, we used these markers as temporal indicators of NRP cell differentiation, which will be further modulated towards specific retinal cell types. The markers used begin expression at different stages of development (Supp. Fig. S3). Thus, Ath5 and Prox1 each mark the beginning of an early NRP differentiation process and Vsx1 and Rhod indicate late NRP differentiation processes.

We initially analyzed the role of Wnt signaling in regulating Ath5 expression using two pharmacological inhibitors of GSK3β, LiCl and indirubin. By inhibiting GSK3β, a crucial kinase in the Wnt signal transduction cascade, these inhibitors provoke an activation of the Wnt pathway (Osakada et al., 2007). To determine the concentrations required for these compounds to be efficient altering the Wnt signaling pathway in the medaka eye, we incubated st. 17 medaka embryos with two different concentrations of LiCl (0.15 or 0.3 M) or indirubin (5 or 10 μM) until st. 22 or st. 26 and performed RT-PCR and qRT-PCR analysis on cDNA to identify any changes in Axin2 (ch.1) expression (Fig. 1M–O). Eyes of embryos treated with a concentration of 0.3M for LiCl or 10 μM for indirubin from st. 17 onwards showed an increase in Axin2 (ch.1) expression at st. 22 and st. 26 when compared to the control (Fig. 1M–O). At these concentrations, both LiCl and indirubin are able to efficiently activate Wnt signaling. Thus, we used these conditions for the pharmacological treatment of embryos in subsequent experiments.

To analyze the role of Wnt signaling in regulating Ath5, embryos from an Ath5-GFP transgenic line were used (del Bene et al. 2007). After incubating Ath5-GFP embryos with either LiCl or indirubin from st. 17 to st. 26 (Fig. 2A), we observed an increase in the GFP expression domain at st. 30, when only RGC are differentiating (Fig. 2B–D). This result was confirmed by analyzing embryos with WMISH to detect endogenous expression of Ath5 mRNA (Supp. Fig. S3A–C). However, at st. 35, the distribution of GFP cells in Ath5-GFP transgenic-treated embryos was similar to that of control embryos (Fig. 2E–G).

thumbnail image

Figure 2. Wnt signaling regulates Ath5 and Prox1 genes expression. A, H, Q, X: Drawing showing the stages when LiCl, indirubin, or HS treatments were applied and the stages when embryos were fixed and gene expression analyzed (arrows). B–G: Ath5-GFP fluorescent expression in embryonic eyes after pharmacological inhibition of GSK3β. Control CAB (B, E) embryos and LiCl- (C, F) or indirubin- (D, G) treated embryos showing an increase in the Ath5-GFP domain at st. 30 (B–D; LiCl increased retinal GFP expression in 83% of embryos tested, n = 36 and indirubin in 68%, n = 21, three independent experiments, n is the total number of embryos). The Ath5-GFP domain showed no observable difference from control embryos in st. 35-treated embryos (E–G). I–N: Analysis of Ath5 expression (purple) detected by WMISH in control and transgenic embryos after HS treatments. Control CAB (I, L) embryos were HS treated at the same stages as the transgenic embryos. At st. 30, the activation of HS-Wnt8 (J) increased the expression of Ath5 (72.5%, n = 62) and the activation of HS-DNWnt8 (K) decreased the expression of Ath5 (85.1%, n = 68), but these differences were not detected at stage 35 (M, N; n = 40). In all cases, three independent experiments were performed and n is the total number of embryos. Scale bars = 50 μm. O: RT-PCR analysis of Ath5 expression. Stages shown indicate when the HS treatment and RNA extraction are performed for each RT-PCR. Three independent RT-PCR reactions were performed on cDNA from two independent experiments. P: qRT-PCR analysis of Ath5 expression relative to Actin control expression in embryonic eyes after HS treatment at st. 17 and analysis at st. 30 or st. 35 and HS treatment at st. 26 and analysis at st. 30. *P < 0.001, **P < 0.0001. The data represent the mean ± SD (error bars). R–W, Y–A′: Analysis of Prox1 expression (purple) detected by WMISH in control and transgenic embryos after HS treatments. Control CAB (R, U, Y) embryos were HS treated at the same stages as the transgenic embryos. When the HS treatment was administered at st. 17, activation of HS-Wnt8 (S) increased the expression of Prox1 at st. 32 (73.1%, n = 41) and activation of HS-DNWnt8 (T) decreased the expression of Prox1 (84.8%, n = 46). However, these differences were not detected at stage 35 (V, W; n = 37). When the HS was administered at st. 26, there was no change in Prox1 at st. 32 (Y–A′; n = 45). In all cases, staging was performed counting somites and according to morphological development of the embryo. No apparent embryonic morphological defects were detected that could influence eye development. Three independent experiments were performed and n is the total number of embryos. Scale bars = 50 μm. B′: qRT-PCR analysis of Prox1 expression relative to Actin control expression in embryonic eyes after HS treatment at st. 17 and analysis at st. 32 or 35 and HS treatment at st. 26 and analysis at st. 32. *P < 0.001, **P < 0.0001.The data represent the mean ± SD (error bars).

Download figure to PowerPoint

To dissect more precisely the effect of Wnt signaling on Ath5 gene expression, we used HS-Wnt8 and HS-DNWnt8 transgenic embryos to manipulate Wnt signaling. These transgenic embryos were exposed to HS treatments at st. 17, 22, and 26 to activate the HS-Wnt8 or HS-DNWnt8 transgene expression during retinal development. Subsequently, Ath5 gene expression was detected by WMISH at st. 30 (Fig. 2H). Activation of HS-Wnt8 increased the expression domain of Ath5 (Fig. 2J) whereas activation of HS-DNWnt8 decreased the expression domain of Ath5 (Fig. 2K). Moreover, an early single HS treatment had a similar effect on Ath5 expression, whereby the activation of HS-Wnt8 at st. 17 increased the expression domain of Ath5 (74.6%, n=32, images not shown) and activation of HS-DNWnt8 decreased the expression domain of Ath5 (78.6%, n=47, images not shown). However, in both of the above experiments, these alterations of Ath5 expression were not detected at st. 35 (Fig. 2L–N and data not shown) or when the activation of the transgenes was induced at st. 26 (data not shown), a stage at which Lef1 and Axin2 (chr. 8) expression was altered by a single HS treatment (Fig. 1G–I). The Ath5 expression results after a single HS treatment were confirmed by RT-PCR and qRT-PCR to quantify Ath5 expression (Fig. 2O, P). These results suggest that Wnt signaling regulates Ath5 expression only during early developmental stages.

To determine whether these observations reflected a specific effect on Ath5 expression or a more general effect on NRP cell differentiation, we explored the effects of Wnt signaling on other retinal cell types. Prox1 is a homeobox protein that is expressed in all amacrine cells during differentiation (Dyer et al., 2003). After activation of HS-Wnt8 at st. 17 (Fig. 2Q), we detected an increase in the expression of Prox1 at st. 32 (Fig. 2S), but at st. 35 Prox1 mRNA levels were equivalent to control embryos (Fig. 2V). When HS-DNWnt8 was activated at st. 17, we detected a decrease of Prox1 expression at st. 32 (Fig. 2T), but at st. 35 the levels of Prox1 were similar to those of control eyes (Fig. 2W). On the other hand, after activation of the transgenes at st. 26, we did not detect any changes in Prox1 expression at st. 32 (Fig. 2X–A′). The Prox1 expression results after a single HS treatment were confirmed using qRT-PCR to quantify Prox1 expression (Fig. 2B′). These results were similar to those obtained for Ath5 expression and suggest that there is an early phase of retinal development in which Wnt signaling regulates Ath5 and Prox1 expression.

Wnt Signaling Regulates the Late Differentiation of Retinal Neural Precursors

NRP cells follow a temporal differentiation sequence whereby bipolar (Vsx1) and rod (rhodopsin) photoreceptor cells are two of the latest cell types to appear. To analyze whether the Wnt pathway also modulates the differentiation of these late-developing cell types, we induced Wnt8 or DNWnt8 transgene expression using three consecutive HS treatments at st. 17, 22, and 26 (Fig. 3A) or we treated embryos with indirubin or LiCl from st. 17 to 35. Activation of HS-Wnt8 increased expression of Vsx1 at st. 35 (Fig. 3C) and similar results were obtained after pharmacological inhibition of GSK3β (Supp. Fig. S4D–G). As expected, activation of HS-DNWnt8 resulted in a decrease of Vsx1 expression at st. 35 (Fig. 3D). However, in contrast to the effects on Ath5 and Prox1, activation of either of the transgenes at st. 17 did not alter Vsx1 expression at st. 35 (Fig. 3E–H). Surprisingly, a single HS treatment for activation of HS-Wnt8 at st. 26 increased expression of Vsx1 and activation of HS-DNWnt8 inhibited Vsx1 expression (Fig. 3I–L). These results were confirmed using RT-PCR (Fig. 3M) and qRT-PCR (Fig. 3Y). Manipulation of Wnt signaling did not change Rhodopsin expression, neither by activating the transgenes at st. 17 or st. 26 (Fig. 3M, Y), nor by activating the transgenes at st. 17, st. 22 and st. 26 (Fig. 3N–T), nor by pharmacologically inhibiting GSK3β with indirubin from st. 17 to st. 35 (Supp. Fig. S4D, H, I). With the use of the α-Zpr1 antibody that labels the retinal photoreceptor layer (Kitambi and Malicki, 2008), we did not detect any changes in the photoreceptor layer of HS-treated transgenic embryos (Fig. 3U–X). Thus, after st. 26, Wnt signaling regulates Vsx1 expression, but does not affect Rhodopsin, Ath5, or Prox1 expression.

thumbnail image

Figure 3. Wnt signaling regulates Vsx1 gene expression, but not Rhodopsin gene expression. A, E, I, N, U: Scheme showing the stages when HS treatments were applied and the stages when embryos were fixed and gene expression analyzed (arrows). In all cases, three independent experiments were performed and n is the total number of embryos. B–D, F–H, J–L: Analysis of Vsx1 expression (purple) detected by WMISH in control and transgenic embryos after HS treatments. Control CAB (B, F, J) embryos were HS treated at the same stages as the transgenic embryos. A–D: After 3 HS treatments (st. 17, 22, and 26), the activation of HS-Wnt8 (C) increased the expression of Vsx1 at st. 35 (80.4%, n = 41) and the activation of HS-DNWnt8 (D) decreased the expression of Vsx1 (85%, n = 47). E–H: When HS treatment was administered at st. 17, there was no change of Vsx1 expression at st. 35 (n = 45). I–L: When the HS was administered at st. 26, the activation of HS-Wnt8 increased the expression of Vsx1 at st. 35 (74.3%, n = 39) and the activation of HS-DNWnt8 decreased the expression of Vsx1 (74.3%, n = 43). Scale bars = 50 μm. M: RT-PCR analysis for Vsx1 and Rhodopsin expression at st. 35 after HS treatment at st. 17 or after HS treatment at st. 26. Three independent RT-PCR reactions were performed on cDNA from two independent experiments. O–T: Analysis of Rhodopsin expression (purple) detected by WMISH in control and transgenic embryos after HS treatments. Control CAB (O, R) embryos were HS treated at the same stages as the transgenic embryos. After 3 HS treatments (st. 17, 22, and 26), there is no change in Rhodopsin expression at st. 35 (P, Q; n = 40) or 36 (S, T; n = 43). V–X: α-Zpr1 immunostaining labeling the retinal photoreceptor layer of control and transgenic embryos at st. 35 after HS treatment at st. 17, 22, and 26. There was no observable change in α-Zpr1 distribution at st. 35 in transgenic embryos (W, X) compared to control (V; Control n = 10; HS-Wnt8 n = 7; HS-DNWnt8 n = 9). Scale bars = 50 μm. Y: qRT-PCR analysis of Vsx1 and Rhodopsin expression relative to Actin control expression in embryonic eyes after HS treatment at st. 17 or st. 26 and analysis at st. 35. *P < 0.001, **P < 0.0001. The data represent the mean ± SD (error bars).

Download figure to PowerPoint

Wnt Signaling Regulates Apoptosis in Early NRP Cells

During early neural development in medaka, many apoptotic cells are detected in the neural tube from st. 20 to st. 27, but the number of apoptotic cells decreases after st. 29 (Ishikawa et al., 2007). Apoptosis is another important patterning process during embryonic development. Thus, we investigated the effects of Wnt signaling on apoptosis of NRP cells using TUNEL staining. After HS treatments at st. 17, 22, and 26 of HS-Wnt8 transgenic embryos, the relative number of apoptotic NRP cells increased at st. 22 (Fig. 4A, C); however, at st. 26 and st. 30, the relative number of apoptotic cells was reduced (Fig. 4B, C and Supp. Table S2). In contrast, after HS treatments at st. 17, 22, and 26 of the HS-DNWnt8 transgenic embryos, we observed that inhibition of Wnt signaling decreased the relative number of apoptotic cells at st. 22 (Fig. 4A, C). However, the relative number of apoptotic cells increased at st. 26 and st. 30 (Fig. 4B, C and Supp. Table S2). To discern whether Wnt signaling could have opposing effects on apoptosis at different stages, we activated transgene expression at st. 17 or st. 22 and counted apoptotic cells at st. 22 and st. 30 (Fig. 4D and Supp. Table S2) or st. 26 and st. 30 (Fig. 4E and Supp. Table S2), respectively. Apoptosis was affected in a similar manner after HS treatment at one stage as with consecutive HS treatments at three stages. However, when Wnt signaling was manipulated at st. 26, apoptosis was not altered (Fig. 4F and Supp. Table S2). These results indicate that Wnt signaling regulates apoptosis during early but not late eye development.

thumbnail image

Figure 4. Apoptosis of NRP cells is regulated by Wnt signaling. A, B: TUNEL-positive cells in a control, HS-Wnt8 and a HS-DNWnt8 embryo after HS treatment at st. 17, and TUNEL staining at st. 22 (A) or eyes at st. 30 (B). In A, anterior is to the left and eye cups are inside marked areas. Scale bars = 50 μm. C–F: TUNEL-positive cells in treated retinas were counted and compared to control retinas (see Experimental Procedures for details). Numbers relative to the control (baseline) are represented. C: Embryos collected at st. 22 were HS treated at st. 17, embryos collected at st. 26 were HS treated at st. 17 and st. 22, and embryos collected at st. 30 were HS treated at st. 17, st. 22 and st. 26. D: Embryos were HS treated at st. 17 and collected at st. 22 and 30. E: Embryos were HS treated at st. 22 and collected at st. 26 and 30. F: Embryos were HS treated at st. 26 and collected at st. 30. In every group, n = 35. *P < 0.05; **P < 0.01; ***P < 0.001. Bar graphs represent the mean ± SD (error bars) from three independent experiments. See Supp. Table S2 and Supp. Fig. S6 for absolute data numbers and representation.

Download figure to PowerPoint

Wnt Signaling Regulates Proliferation of Retinal Precursor Cells During Early Development

To test whether another developmental patterning process such as proliferation could also be affected in the eye by manipulation of the Wnt pathway, we analyzed the effects of Wnt signaling on NRP proliferation using phospho-histone H3 (pH3) as a marker of proliferating cells in M phase. When LiCl or indirubin were used to activate the Wnt signaling pathway from st. 17 to st. 22 or 26, we observed an increase in the relative number of pH3-positive cells when compared to control retinas at st. 22, but this increase was not maintained at st. 26, when the relative number of pH3-positive cells decreased (Fig. 5A and Supp. Table S3). Similarly, HS treatments at st. 17, 22, and 26 in HS-Wnt8 transgenic embryos were associated with an increase in the relative number of pH3-positive cells at st. 22 (HS at st. 17); this increase was undetectable at st. 26 (HS at st. 17 and 22). Moreover, at st. 30 (HS at st. 17, 22, and 26) there was a decrease in the relative number of pH3-positive cells (Fig. 5B and Supp. Table S3). On the other hand, activation of the HS-DNWnt8 construct in transgenic embryos after HS treatments at st. 17, st. 22, and st. 26 resulted in a decrease in the relative number of pH3-positive cells at st. 22 (HS at st. 17). This effect was undetectable by st. 26 (HS at st. 17 and 22) and at st. 30 (HS at st. 17, 22, and 26) there was an increase in the relative number of pH3-positive cells (Fig. 5B and Supp. Table S3).

thumbnail image

Figure 5. Proliferation of NRP cells is regulated by Wnt signaling. A–C, E, F: pH3-positive cells in the treated retinas were counted and compared to control retinas (see Experimental Procedures for details). Numbers relative to the control (baseline) are represented. A: CAB embryos were incubated with LiCl or indirubin between st. 17 until fixation at st. 22 or 26. B: Embryos were HS treated at st. 17 and collected at st. 22, HS treated at st.17 and st.22 and collected at st.26 or HS treated at st.17, st, 22 and st. 26 and collected at st.30. C: Embryos were HS treated at st. 17 and collected at st. 22 and 30. D: pH3-positive cells in a control, HS-Wnt8, and a HS-DNWnt8 embryo eye after HS treatment at st. 22 and pH3 immunostaining at st. 26. The retinas are inside the marked areas. Scale bars = 50 μm. E: Embryos were HS treated at st. 22 and collected at st. 26, 30, and 32. F: Embryos were HS treated at st. 26 and collected at st. 30. Proliferation was not affected at this stage by Wnt signaling. In every group, n = 60. *P < 0.05; **P < 0.01; ***P < 0.001. Bar graphs represent the mean ± SD (error bars) from three independent experiments. See Supp. Table S3 and Supp. Fig. S7 for absolute data numbers and representation.

Download figure to PowerPoint

To distinguish whether the above results were the consequence of Wnt signaling regulating proliferation with opposing actions at different stages of development, we administered single HS treatments at different developmental stages and registered proliferation values at later stages. Activation of HS-Wnt8 at st. 17 resulted in an increase in the relative number of pH3-positive cells at st. 22 (Fig. 5C and Supp. Table S3). However, at st. 30, we detected a decrease. On the other hand, activation of HS-DNWnt8 at st. 17 resulted in a decrease in proliferation at st. 22 (Fig. 5C and Supp. Table S3) and an increase at st. 30. Moreover, when the HS treatment was performed at st. 22 similar results were obtained for each transgenic construct with the compensation effect detected at st. 32 (Fig. 5D, E and Supp. Table S3). Finally, when the transgenes were induced at st. 26, we did not observe any changes in the relative number of pH3-positive cells (Fig. 5F). Thus, we conclude that NRP proliferation can be regulated by Wnt signaling at early stages of retinal development (<st. 26).

We further analyzed the effect on proliferation by studying the cell cycle of HS-treated transgenic embryos by using propidium iodide (PI) staining and flow cytometry analysis. Embryos were HS-treated at st. 17 and eyes collected at st. 22. Eyes were processed and used for flow cytometry studies. HS treatment to activate HS-Wnt8 expression at st. 17 resulted in a decrease of the percentage of cells in G0/G1 phases and an increase in the percentage of cells at S and G2/M phases in st. 22 eyes (Fig. 6A, B). This supports our previous pH3 staining results and suggests that overexpression of Wnt signaling causes G1-S phase transition to proceed faster resulting in cell accumulation in S and G2-M phases. Activation of HS-DNWnt8 expression at st. 17 resulted in an increase in the percentage of cells at G0/G1 phase and a decrease in the percentage of cells at S and G2/M phase in st. 22 eyes when compared to the control (Fig. 6A, B). However, HS treatments at st. 26 and eyes collected at st. 30 resulted in no significant changes in the cell cycle (Fig. 6C, D). These results confirm our previous pH3 staining results and suggest that Wnt signaling is necessary for cell cycle progression at the G1-S phase transition. Thus, Wnt signaling regulates cell cycle progression of NRP cells during early eye development.

thumbnail image

Figure 6. Wnt signaling regulates cell cycle kinetics, G1-S transition, and CnnD1 expression. A: IP flow cytometry histograms of cells isolated from the eyes of HS treated embryos eyes at st. 17 and collected at st. 22. B: Bar graphs representing the percentage of cells in G0/G1, S, and G2/M from embryonic eyes at st. 22 after HS at st. 17. *P < 0.05, **P < 0.01. The data represent the mean ± SD (error bars) from three independent experiments. C: IP flow cytometry histograms of cells isolated from the eyes of HS treated embryos eyes at st. 26 and collected at st. 30. D: Bar graphs representing the percentage of cells in G0/G1, S, and G2/M from embryonic eyes at st. 30 after HS at st. 26. *P < 0.05, **P < 0.01. The data represent the mean ± SD (error bars) from three independent experiments. E, F: RT-PCR analysis of gene expression for several regulatory proteins of different cell cycle phases at st. 22 after HS at st. 17 (E) or analysis at st. 30 after HS treatment at st. 26 (F). Three independent RT-PCR reactions were performed on cDNA from two independent experiments with similar results. G: qRT-PCR analysis of CcnD1 and CcnA expression relative to Actin control expression in embryonic eyes at st. 22 after HS treatment at st. 17 or analysis at st. 30 after HS treatment at st. 26. *P < 0.0001. The data represent the mean ± SD (error bars).

Download figure to PowerPoint

Subsequently, we tested the effects of Wnt signaling manipulation on the expression of several regulatory proteins of different cell cycle phases using RT-PCR. We observed that only CyclinD1 (CcnD1) expression was increased after activation of HS-Wnt8 and decreased after activation of HS-DNWnt8 after HS treatment at st. 17 but not after HS treatment at st. 26. No other detectable change of expression was observed for CyclinA (CcnA), Cdk2, Cdk4, Cdk6, Cdc25a or p53 after HS treatment at st. 17 or st. 26 (Fig. 6E, F). The change in CcnD1 expression, which regulates G1-S phase transition, and the expression of CcnA, which regulates G2-M phase transition, was further analyzed by performing qRT-PCR (Supp. Table S1). As observed previously with semi-quantitative RT-PCR, when using qRT-PCR, we detected that activation of HS-Wnt8 at st. 17 resulted in an increase in CcnD1 expression and no changes in CcnA expression in st. 22 eyes. Furthermore, activation of HS-DNWnt8 at st. 17 resulted in a decrease in CcnD1 expression and no change in CcnA expression in eyes at st. 22 (Fig. 6G). However, no changes in CcnD1 and CcnA expression were observed when the HS treatment was administered at st. 26 and expression analysis was performed at st. 30 (Fig. 6G). Therefore, our results suggest that Wnt signaling can regulate NRP proliferation at early stages of retinal development (<st. 26) by regulating CcnD1 expression and G1-S phase transition.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Using conditional gene expression techniques in transgenic embryos combined with pharmacological activation of the Wnt pathway, we have performed a stage-specific analysis of Wnt signaling function during eye development and have identified two different temporal roles for Wnt signaling during retinal development. Our results show for the first time in the eye that Wnt signaling has a role in regulating cell cycle progression and CcnD1 expression in NRP cells during the early stages of eye development (between st. 17 and 26). It has been proposed that cell cycle exit regulates retinal differentiation (Cayoutte et al., 2006). Wnt regulation of the cell cycle could explain the changes observed after Wnt signaling perturbation in proliferation, apoptosis, and expression of Ath5 and Prox1. These results would suggest that at early stages retina differentiation is tightly linked to cell cycle progression. The second role of Wnt signaling, which is exerted during the later stages of retinal development, is to regulate the expression of Vsx1 and, thus, participate in the differentiation of NRP. During these later stages (after st. 26), Wnt signaling has no effect on the expression of genes that regulate cell cycle progression and has no effect on proliferation, apoptosis, or expression of Ath5, Prox1, and Rhodopsin. Thus, in this model, Wnt signaling would participate in retinal development through an initial cell cycle–dependent cell differentiation phase and, later, through a signal-dependent cell differentiation phase.

In the present study, we have used a canonical Wnt member, Wnt8, and a specific canonical Wnt signaling inhibitor, DN-Wnt8 (Hoppler et al., 1996, Itoh and Sokol, 1999; Tada and Smith, 2000), to produce transgenic medaka embryos. Moreover, both LiCl and indirubin act by inhibiting GSK3β, a critical kinase in the canonical Wnt signal transduction cascade (Osakada et al., 2007). Activation of either Wnt8 or DNWnt8 upregulates or inhibits Lef1, a coactivator of βCat, and Axin2 expression, both of which are target genes of Wnt signaling. In addition, the A-P axis phenotypes induced by our transgenes are similar to those obtained after manipulation of the canonical Wnt pathway (Erter et al., 2001; Lekven et al., 2001; Houart et al., 2002; Kudoh et al., 2002). These results suggest that the different temporal roles of Wnt signaling during retinal development may be mediated by the canonical pathway. On the other hand, the lack of TOP-GFP activity in the early stages of retinal development in zebrafish suggests that canonical Wnt signaling is not active in the retina (Dorsky et al., 2002; Veien et al., 2008). However, at least one member of the Tcf family, Tcf7, is differentially expressed in the zebrafish retina (Veien et al., 2005) and canonical Wnt signaling from the RPE is necessary for the correct patterning of the retina (Veien et al., 2008). In fact, Tcf activation in the mouse retina does not necessarily correlate with TOP-gal activation suggesting that this reporter construct may not reflect all canonical Wnt signaling (Fuhrmann et al., 2009).

The effects of Wnt on multiple developmental processes of the retina may seem contradictory, especially for proliferation and apoptosis, which both appear to be regulated by Wnt signaling. However, it has been demonstrated in Xenopus that, during retinal development, Shh also regulates multiple developmental processes, including cell cycle kinetics where its effect is to shorten the cell cycle and increase the differentiation of NRP cells (Locker et al., 2006). Therefore, the simultaneous effects of Wnt that we have observed could be explained by the participation of Wnt signaling in the regulation of cell cycle kinetics. Our results suggest that between st. 17 to 26 of medaka embryonic development, Wnt signaling regulates cell cycle transition through G1-S by regulating CcnD1 expression. This is in accordance with previous data showing that Wnt signaling regulates CcnD1 in colon carcinoma cells to control proliferation (Tetsu and McCormick, 1999; Shtutman et al., 1999).

The fact that two processes like proliferation and apoptosis are regulated by Wnt signaling in the developing retina raises the question of a compensatory mechanism. For example, an increase in proliferation would cause a subsequent increase in apoptosis to maintain a constant number of cells and vice versa. However, we have observed that both processes respond to Wnt signaling simultaneously with no observable delay. In addition, although we could disrupt proliferation and apoptosis at early stages by the conditional activation of the Wnt transgenes, retinal development was restored to normal by st. 40, suggesting the existence of an intrinsic rescue mechanism. The nature of this rescue may be associated with cell cycle regulation by other cues where an increase in cell cycle kinetics will later be compensated by a decrease in cell cycle kinetics.

The role of Wnt signaling in proliferation and differentiation in medaka is not in agreement with previous reports regarding the manipulation of βCat expression in the mouse model. In mice, βCat has been linked to lamination of the retina but not to neurogenesis or proliferation (Fu et al., 2006). Deletion of βCat in cells of the Six3 expression domain caused a severely disorganized central retina. However, similar experiments showed that TOP-gal activity was maintained even in the absence of βCat (Fuhrmann et al., 2009), suggesting that βCat may not be necessary for activation of other Wnt pathway components. Also, the defects induced by manipulation of βCat, a late component of the Wnt signaling transduction cascade, may be different from those induced by manipulation of Wnt gene expression, which is subject to modulation by other transcriptional elements.

Our results regarding cell cycle regulation are consistent with previous in vitro studies with cultured retinas showing that Wnt3a promotes proliferation of NRP cells (Das et al., 2006; Inoue et al., 2006; Osakada et al., 2007). Furthermore, microinjection of stabilized βCat in zebrafish and Xenopus embryos promotes proliferation in the retina (Van Raay et al., 2005; Yamaguchi et al., 2005). In explants of chicken retinas, initial reports suggested that Wnt2b only controls proliferation (Kubo et al., 2003, 2005). However, these reports have been challenged by more recent in vivo studies demonstrating that Wnt2b functions only in determining the peripheral fates of retinal cells (Cho and Cepko, 2006). Our results suggest that the main reason for the discrepancies among these reports may be the stages at which Wnt signaling was activated, which could alter significantly the response of NRP cells. Additionally, our hypothesis that Wnt signaling controls cell-cycle kinetics could explain the proliferation results observed in retina explants (Kubo et al., 2003, 2005); Wnt signaling would decrease cell-cycle length in these explanted retinas, thereby increasing proliferation.

It is interesting to note the Wnt signaling also plays a role in proliferation in the postnatal retina. In Xenopus tadpoles, Wnt signaling is active in the CMZ and controls proliferation of RSC (Denayer et al., 2008). The role of Wnt signaling regulating the proliferation of postnatal RSC is similar to the early role we describe in teleost fish during development. However, since stem cells have an indefinite proliferation capacity, acceleration of the cell cycle kinetics by Wnt signaling will not encounter any additional developmental programs or cues to stop it and proliferation will be maintained. It would be interesting to analyze the presence or the lack of any possible RSC depletion effects caused by Wnt signaling in this post-embryonic context.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Fish Embryos and Treatments

Adult medaka (Oryzias latipes) CAB strain animals (kindly provided by Paola Bovolenta, CSIC, Madrid) were kept in recirculating water aquaria at 28°C on a 14-hr light/10-hr dark daily cycle. Embryos were collected by natural spawning in 1× Yamamoto (Yamamoto, 1975) and staged as described (Iwamatsu, 2004). Embryos were raised at 25°C.

For pharmacological treatments, CAB or Ath5-GFP transgenic embryos (kindly provided by Dr. Joachim Wittbrodt; Del Bene et al., 2007) at the desired stages were treated with LiCl at 0.15M or 0.3M (Hedgepeth et al., 1997) or with indirubin at 5 μM or 10 μM (Castelo-Branco et al., 2004). Indirubin was dissolved in DMSO according to the manufacturer's instructions. For each HS treatment, HS-Wnt8 or HS-DNWnt8 transgenic embryos and CAB control embryos at the desired stages were treated three consecutive times for 30 min at 37°C followed by 1 hr at 25°C (Grabher and Wittbrodt, 2004). When HS treatments were performed at three different stages, the described procedure was repeated at each stage. When embryos older than st. 25 were used, retinal pigmentation was inhibited by incubating the embryos in phenylthiourea (PTU; 0.003%) from st. 17 until fixation. Treated embryos were staged according to the number of somites and described morphological characteristics (Iwamatsu, 2004).

Transgenesis

The two medaka transgenic lines HS-Wnt8 and HS-DNWnt8 were generated using two distinct plasmids, pHS-Wnt8-CSKA-GFP and pHS-DNWnt8-CSKA-GFP both of which contain two I-SceI meganuclease sites (Thermes et al., 2002). Between these sites, there is GFP under the control of the CSKA constitutive promoter (gift from J. Wittbrodt) and an HS inducible promoter that controls either Xenopus Wnt8 from the plasmid pHSXwnt-8HSG (gift from T. Czerny) or a C-terminal truncated form of Wnt8 (DNWnt8; gift from S. Sokol) that acts as a dominant-negative protein. Details of the constructs are available upon request. The resulting plasmids were verified by sequencing. For DNA microinjection, plasmids were digested with the enzyme I-SceI (Roche), according to the manufacturer's instructions. Linearized DNA at a concentration of 10 ng/μl was injected through the chorion into the cytoplasm of one-cell-stage CAB embryos using a pressure Narishige IM300 microinjector and borosilicate glass capillaries (WPI). GFP-positive embryos were selected, raised to sexual maturity, and crossed to establish the transgenic lines (Supp. Fig. S8).

RT-PCR and qRT-PCR Analysis

Total RNA was extracted from eyes dissected from groups of 40 embryos using Trizol reagent (Sigma, St. Louis, MO). For HS treatment experiments, groups were formed from HS-treated or control embryos at the appropriate stages. For optimization of the pharmacological compounds, groups were formed from embryos treated with LiCl, indirubin, or control embryos. cDNAs were synthesized from 1 μg of total eye RNA using Random primers hexamers (Roche, Nutley, NJ) and Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Equal amounts of first-strand cDNAs were then amplified using primer pairs for the studied genes (Supp. Table S4). For Wnt8 and DNWnt8 gene amplification, the same primer pair was used to amplify a fragment in the 5′region of these genes. The PCR reactions were performed using the appropriate number of cycles (see Supp. Table S4) with a denaturation step at 94°C for 35 sec, annealing at the appropriate temperature (see Supp. Table S4) for 50 sec and extension at 72°C for 30 sec. After amplification, the reaction product was analyzed in a 2% agarose gel containing ethidium bromide (1 μg/mL) and images of the bands were captured on a Versadoc Image analyzer (Bio-Rad, Hercules, CA). The number of cycles was optimized by determining the exponential range of amplification for each primer pair. A negative control to confirm the absence of gDNA was performed using Actin primers on RNA without Reverse Transcriptase. RT-PCR experiments for each gene were performed three times.

For qRT-PCR, we used 1 μL of cDNA in 1 × SYBR Green Master Mix (Bio-Rad) and 10 mM of each forward and reverse primer. The PCR amplification and fluorescence detection were performed in a Real-Time PCR thermal cycler (MiniOpticon, Bio-Rad). The PCR conditions were: 35 cycles of denaturation at 95°C for 15 sec, annealing for 45 sec at the appropriate temperature (see Supp. Table S1), and extension for 30 sec at 72°C. The melting curve was constructed by plotting fluorescence data against temperature (55–95°C with an interval of 1°C). Samples with good melting curve were considered for analysis. Control samples without cDNA were included in all assays to confirm the absence of non-specific amplification products. For each sample, the threshold cycle (Ct) for the internal control (Actin) amplification was subtracted from the threshold cycle of the corresponding transcription factor amplification (Ct, transcription factor) to yield ΔCt. The Pfaffl Method was used to calculate the ratio of relative gene expression related to Actin (internal control), and represented in the bar graphs. Primer pair efficiencies were calculated for each primer pair by performing dilution curves (for primer efficiencies, see Supp. Table S1). qRT-PCR reactions were performed in triplicate for each experiment. Bar graphs represent the results from six independent experiments performed on cDNA samples from two different treated groups of embryos (three per sample).

Whole Mount In Situ Hybridization, Immunohistochemistry and TUNEL Aassay

pProx1, pVsx1, pAxin2(Ch.1) and pLef1 plasmids were made using the following primer pairs to amplify around 400-bp fragments of the Prox1, Vsx1, Axin2(Ch.1) and Lef1 genes. Primer sequences are as follows: Prox1F, GATAACATTGCTCGAGCCA AAG; Prox1R, TTATTCTTGAGAAG CTGG; Vsx1F, AGTCCATCCT CAACTCAG; Vsx1R, GTCCTAA-ATTTTCTCAGC; Axin2(Ch.1)F, CAA AGTCTCT GCATTTCC; Axin2(Ch.1)R, GCTTGA AGATAACTTTGGC; Lef1F, ATGAAA AACACAGAGACC; Lef1R, TTTGTT GGACGTTCTGG. The PCR fragments amplified were cloned into the pBS vector within the EcoRI and XhoI sites. The Prox1,Vsx1, Axin2 (ch.1), and Lef1 antisense DIG-labeled RNA probes were synthesized by digesting pProx1,pVsx1, pAxin2(Ch.1), and pLef1 with NotI and transcribing with T7 polymerase. Ath5 and Rhodopsin probes (gift from Dr. Paola Bovolenta, CSIC, Cajal Institute) were produced by digesting plasmids with NotI or EcoRI and transcribing with SP6 or T3 polymerase respectively. Wnt8 and DNWnt8 probes were produced by digesting pHS-DNWnt8-CSKA-GFP with EcoRI and transcribing with T7 polymerase. WMISH was performed as described (Loosli et al., 1998) on embryos fixed with 4% paraformaldehyde (PFA)/2×PBSTw previously exposed to pharmacological or HS treatment at the desired stages. Stained embryos were subsequently mounted in 87% glycerol for imaging. Sense probes were used as controls (Supp. Fig. S5).

Immunohistochemistry to detect endogenous canonical Wnt signalling activity was performed using the antibody β-Catenin (ref. 9562, Cell Signaling, Danvers, MA; this antibody was used to detect nuclear β-Catenin in Lee et al., 2009, among others) at 1:200. CAB embryos at st. 24 were processed for cryosections (Vihtelic and Hyde, 2000) and 5–10-μm sections were stained for β-Catenin following the manufacturer's immunostaining instructions. Embryos were incubated with goat secondary antibody Alexa488 α-rabbit (Invitrogen) at 1:500 and DAPI (Sigma) and mounted using fluoromount G medium (Southern Biotech, Birmingham, AL) for imaging. Serial confocal images (35 images per section) and z-axis reconstruction (Leica LCS Lite software) were used to detect nuclear β-Cat staining. Only strong β-Cat expression surrounded by DAPI staining in all three axis was considered as positive nuclear staining (Supp. Fig. S2A, B). The numbers of total nuclei (DAPI) and nuclear β-Cat-positive cells were counted in cryosections from three different eyes.

Immunohistochemistry to assay cell proliferation was performed using the antibody α-pH3 at 1:200 (Cell Signaling). HS-Wnt8 or HS-DNWnt8 transgenic embryos and CAB control embryos, treated with HS, LiCl, or indirubin at the appropriate stages, were fixed overnight at 4°C in 4% PFA/2×PBSTw, dechorionated, and processed according to the manufacturer's immunostaining protocol for pH3. Embryos were incubated with secondary antibody Alexa 488 α-mouse (Invitrogen) at 1:500 and mounted using fluoromount G medium. Zpr1 (ZIRC) immunohistochemistry was performed at 1:200. HS-Wnt8 or HS-DNWnt8 transgenic embryos and CAB control embryos, treated with HS at the appropriate stages, were processed for cryosections and 5–10-μm sections were stained for Zpr1 following the manufacturer's immunostaining instructions. Embryos were incubated with goat secondary antibody Alexa633 α-mouse (Invitrogen) at 1:500 and DAPI (Sigma) and mounted using fluoromount G medium (Southern Biotech) for imaging.

For the apoptosis assay, HS-Wnt8, HS-DNWnt8, or CAB control embryos treated with HS at the desired stages were fixed in 4% PFA/2×PBSTw overnight at 4°C and dechorionated. TUNEL-positive cells were detected by whole-mount TUNEL (Yamamoto and Henderson, 1999) using an ApopTag Peroxidase In situ Apoptosis Detection Kit following the manufacturer's instructions (Chemicon International Inc., Temecula, CA). Embryos were mounted with 87% glycerol for imaging.

Images were obtained with a Leica MZ16FA stereomicroscope, Leica DM 5000B fluorescent microscope, and confocal images with a Leica DM IRE2 microscope.

Cell Cycle Analysis

Eyes were extracted from groups of 20 embryos and trypsinized (Tripsin/EDTA no. 25200-072, Invitrogen) at 37°C for 5 min. Cell suspension was centrifuged at 5,000 rpm for 6 min at 4°C, washed with PBS, and fixed with 70% ethanol overnight at −20°C. For flow cytometric determination of DNA content in single cells, cell pellets were resuspended in 300 μl propidium iodide (PI) solution containing PI (50 μg/ml) and RNAse A (10 μg/ml) and stained overnight at 4°C. Cell suspensions were analyzed in a Cytomics FC500 flow cytometer (Beckman Coulter, Fullerton, CA). For fitting cell cycle phases from the primary cytometric data, the public domain software WinMDI 2.8 (http://facs.scripps.edu/software.html) and Cylchred (http://www.facslab.toxikologie.uni-mainz.de/engl.%20Websites/Downloads-engl.jsp) were used. Three independent experiments were performed and analyzed using GraphPad Prism (GraphPad Software Inc., San Diego, CA) and t-test analysis (P < 0.05).

Imaging and Statistical Analysis

All images were collected using an Olympus DP70 camera connected to a Leica M716FA dissecting microscope or a Leica DM5000B vertical microscope. For cell proliferation and TUNEL assays, total numbers of pH3-positive or apoptotic cells of retinas were counted and numbers relative to the control are represented in Figures 4 and 5. Statistical analysis to determine significant changes were performed on GraphPad Prism (Version 4.00, 1992-2003 GraphPad Software Inc.) using one-way ANOVA plus Tukey postHoc test. For all data. a level of 5% or less (P < 0.05) was taken as statistically significant.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank J. Wittbrodt for providing the Ath5-GFP transgenic line and cDNA, T. Czerny and S. Sokol for providing cDNA, M. Schartl and Y. Wakamatsu for help with medaka, and D. Burks, A. Ruiz i Altaba, and P. Sánchez for discussions and critical reading of the manuscript. We specially thank P. Bovolenta for discussions, initial guidance, materials, and critical reading of the manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_22168_sm_SuppFig1.tif4932KSupp. Fig. S1. Axin2 (Ch.1) and Lef1 expression pattern. A, C: Analysis of Axin2(Ch.1) expression (purple) detected by WMISH at st. 19 (A) and st. 24 (C). B, D: Analysis of Lef1 expression (purple) detected by WMISH at st. 19 (B) and st. 24 (D). Arrowheads, regions where both genes are expressed. Arrows, regions where only Axin2(Ch.1) is expressed. Scale bars = 100 μm.
DVDY_22168_sm_SuppFig2.tif16326KSupp. Fig. S2. Endogenous canonical Wnt signaling in the medaka retina. A: β-Cat immunohistochemistry and DAPI staining on eye cryosections. Z-sections cross at a dot of strong β-Cat staining (a). Z-axis projections are shown to the right (y axis) and below (x axis). B: Magnification of the area around (a) and corresponding z-sections, which show that β-Cat staining is surrounded by DAPI staining in all three axis. C: Control immunohistochemistry experiment using only secondary antibody showing no signal. D: β-Cat immunohistochemistry detection in the CMZ (arrows) at st. 24. Scale bars = 10 μm (A); 5 μm (B); 20 μm (C,D).
DVDY_22168_sm_SuppFig3.tif9757KSupp. Fig. S3. Marker gene expression pattern in the medaka developing eye. Expression patterns of (A) Ath5, (B) Prox1, (C) Vsx1, and (D) Rhod genes during different stages of eye development (sts. 23, 25, 27, 30, 34, 36).
DVDY_22168_sm_SuppFig4.tif8175KSupp. Fig. S4. Pharmacological inhibition of GSK3β regulates Ath5 and Vsx1 expression, but not Rhodopsin expression. A, D: Scheme showing the stages when LiCl or indirubin were administered and the stages when embryos were fixed and gene expression analyzed (arrows). CAB embryos were incubated with LiCl or indirubin from st. 17 until fixation at st. 30 or 35 and Ath5, Vsx1, or Rhodopsin expression was analyzed. B, C: Pharmacological inhibition of GSK3β activated Ath5 expression at st. 30. Indirubin increased Ath5 mRNA expression in 71% of embryos tested, n = 20. E–G: Pharmacological inhibition of GSK3β activated Vsx1 expression at st. 35. LiCl increased the expression of Vsx1 in 90% of embryos, n = 20, and indirubin in 75% of the embryos, n = 20. H, I: Pharmacological inhibition of GSK3β did not affect Rhodopsin expression at st. 35, n = 25. Scale bars = 50 μm.
DVDY_22168_sm_SuppFig5.tif5231KSupp. Fig. S5. WMISH using sense probes. Control WMISH analysis using sense probes for Ath5 at st. 30, Prox1 at st. 32, and Vsx1 or Rhodopsin at st. 35. Scale bars = 50 μm.
DVDY_22168_sm_SuppFig6.tif1833KSupp. Fig. S6. Apoptosis of NRP cells is regulated by Wnt signaling. Bar graphs of the TUNEL analysis representing the absolute numbers of apoptotic cells scored in HS-Wnt8 and HS-DNWnt8 transgenic embryos after the described HS treatments. A: Embryos collected at st. 22 were HS treated at st. 17, embryos collected at st. 26 were HS treated at st. 17 and st. 22, and embryos collected at st. 30 were HS treated at st. 17, st. 22, and st. 26. B: Embryos were HS treated at st. 17 and collected at st. 22 and 30. C: Embryos were HS treated at st. 22 and collected at st. 26 and 30. D: Embryos were HS treated at st. 26 and collected at st. 30. In every group, n = 35. *P < 0.05; **P < 0.01; ***P < 0.001. Bar graphs represent the mean ± SD (error bars) from three independent experiments.
DVDY_22168_sm_SuppFig7.tif2288KSupp. Fig. S7. Proliferation of NRP cells is regulated by Wnt signaling. Bar graphs of the pH3 immunostaining analysis representing the absolute numbers of proliferating cells scored in HS-Wnt8 and HS-DNWnt8 transgenic embryos after the described HS treatments. A: CAB embryos were incubated with LiCl or indirubin between st. 17 until fixation at st. 22 or 26. B: Embryos were HS treated at st. 17 and collected at st. 22, HS treated at st. 17 and st. 22 and collected at st. 26, or HS treated at st. 17, st. 22, and st. 26 and collected at st. 30. C: Embryos were HS treated at st. 17 and collected at st. 22 and 30. D: Embryos were HS treated at st. 22 and collected at st. 26, 30, and 32. E: Embryos were HS treated at st. 26 and collected at st. 30. Proliferation was not affected at this stage by Wnt signaling. In every group n = 60. *P < 0.05; **P < 0.01; ***P < 0.001. Bar graphs represent the mean ± SD (error bars) from three independent experiments.
DVDY_22168_sm_SuppFig8.tif1877KSupp. Fig. S8. GFP-positive transgenic embryos. GFP expression of HS-Wnt8 and HS-DNWnt8 transgenic embryos at st.19. Scale bars = 100 μm.
DVDY_22168_sm_SuppTables.doc157KSupplementary table 1. Primers and PCR conditions used for qRT-PCR. Supplementary table 2. Data results from apoptosis analysis (TUNEL) in HS-treated transgenic embryos. The data represent the number of cells stained with TUNEL in the developing retina. n is the number of retinas counted for each condition. This data was used to calculate relative values using the control condition as 1. Supplementary table 3. Data results for proliferation using pH3 staining in treated embryos. The data represent the number of cells stained with pH3 in the developing retina. n is the number of retinas counted for each condition. This data was used to calculate relative values using the control condition as 1. Supplementary table 4. List of primers used in RT-PCR and qRT-PCR, and RT-PCR conditions.

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.