All of the neurons and glia in the adult central nervous system (CNS) are derived from immature neuroepithelial cells that line the embryonic neural tube. During CNS development, these progenitor cells proliferate, diversify, migrate, and establish synaptic connections that are required for normal brain function. These events are controlled by a well-orchestrated series of cell–cell interactions; however, in many instances, these interactions as well as the molecules that mediate them have not been well-characterized.
The mammalian retina is an excellent model system in which to study the role of signaling molecules in patterning and neuronal diversification, as it is very accessible and the different retinal cell types are well characterized and can be identified with cell-type–specific markers and by their position within the retinal laminae. The lineage of the cells in the retina is also known—all of the cells, except for astrocytes, are derived from multipotential precursor cells (Turner and Cepko, 1987; Holt et al., 1988; Wetts and Fraser, 1988). The generation of the different cell types is under temporal control with retinal ganglion cells (RGCs), cone photoreceptors, horizontal cells, and half of the amacrine cells generated first, followed by rod photoreceptors, bipolar cells, Müller glia, and the remaining amacrine cells generated in a second wave of histogenesis, which extends into the postnatal period (Young, 1985). This temporal control of retinal cell type development reflects the operation of both cell intrinsic and extrinsic patterning mechanisms. Retinal precursor cells change over time in terms of their proliferative capacity, gene expression, and competence to generate specific cell types (Watanabe and Raff, 1990; Alexiades and Cepko, 1996, 1997; Brown et al., 1998; Belliveau and Cepko, 1999). Intercellular signaling mediated by diffusible signals, as well as the Notch pathway, also plays a role in patterning the eye cup and precursor cell proliferation and differentiation (Altshuler and Cepko, 1992; Watanabe and Raff, 1992; Altshuler et al., 1993; Austin et al., 1995; Dorsky et al., 1995, 1997; Lillien, 1995; Henrique et al., 1997; Jensen and Wallace, 1997; Levine et al., 1997; Belliveau and Cepko, 1999; Furukawa et al., 2000; Wang et al., 2002).
Members of the Wnt signaling family are also good candidates for signaling molecules that pattern the retina. Wnts are secreted glycoproteins that play important roles in patterning and cell fate in many regions of the vertebrate embryo (Nusse and Varmus, 1992; Parr and McMahon, 1994; Cadigan and Nusse, 1997; Moon et al., 1997; Wodarz and Nusse, 1998). Wnt receptors, members of the Frizzled gene family, are 7-transmembrane-domain cell surface receptors that bind Wnts by means of their extracellular cysteine-rich domain (CRD; Vinson et al., 1989; Bhanot et al., 1996; Wang et al., 1996). Although there is evidence for other signaling pathways activated by Wnts (Kuhl et al., 2000; Thorpe et al., 2000; Bienz, 2001), signaling through the canonical Wnt pathway results in the stabilization of cytoplasmic β-catenin, which then associates with HMG-box transcription factors belonging to the TCF/Lef family, and results in the up-regulation of TCF-dependent gene transcription (van de Wetering et al., 1997). Wnt signaling is also modulated in the extracellular environment by the antagonistic activity of members of the secreted-frizzled-related protein (Sfrp) family, which contain CRDs that bind Wnts, preventing them from interacting with Frizzleds (Lin et al., 1997; Rattner et al., 1997). Wingless, the Drosophila homologue of the vertebrate Wnt-1, is an important regulator of eye development in flies (Treisman and Rubin, 1995). Given the striking conservation in the requirement for similar transcription factors and signaling molecules in patterning the eyes of flies and vertebrates, it is likely that Wnts will also play a role in mammalian eye patterning. Although the expression of members of the Wnt, Frizzled, and Sfrp gene families has been documented in the eyes of several vertebrate species (Wang et al., 1996; Leimeister et al., 1998; Zakin et al., 1998; Borello et al., 1999; Jasoni et al., 1999), a comprehensive analysis of the expression patterns of these genes in the developing and adult mammalian retina has been lacking. As a first step toward understanding the function of Wnt signaling in eye development, we performed reverse transcriptase-polymerase chain reaction (RT-PCR) and in situ hybridization analyses to determine when and where components of the Wnt signaling pathway are expressed in the developing and adult murine eye. We also examined reporter gene activation in the eyes of TCF/Lef-LacZ transgenic mice, a canonical Wnt reporter strain. We show that components of the Wnt signaling pathway are expressed in a dynamic pattern during eye development, and our results implicate a role for activation of the canonical Wnt signaling pathway in the development of non-neural derivatives of the optic cup.
Expression of Wnt Pathway Genes in the Embryonic Retina
To examine the Wnt expression pattern in the neural retina, we performed RT-PCR analysis by using Wnt-specific primer pairs on RNA isolated from embryonic day 14.5 (E14.5) and E18.5 neural retinas. Care was taken to remove the retinal pigment epithelium (RPE) during the dissection process, so as to reduce the likelihood of amplifying RPE-specific products. Wnt-3, -5a, -5b, -7b, and -13–specific gene products were amplified from the E14.5 and E18.5 neural retina (Fig. 1). No specific signal was detected in the neural retina at any age for Wnt-2, -3a, -4, -6, -7a, -8b, -10b, and -11 (Fig. 1).
To localize Wnt gene expression in ocular structures, we performed in situ hybridization on eye tissues obtained from developing mice with digoxigenin (DIG) -labeled antisense riboprobes corresponding these genes (Fig. 2; Table 1). As a control for probe specificity, we confirmed that each of the riboprobes demonstrated specific hybridization in tissues where their expression had been reported previously (Fig. 2 and data not shown).
Table 1. Wnt Transcript Localization in the Eye From E12.5 to E14.5a
NR, neural retina; RPE, retinal pigment epithelium; ISH, in situ hybridization; RT-PCR, reverse transcriptase-polymerase chain reaction; E, embryonic day.
Expression also detected by ISH in the postnatal day 7 and adult retina.
Expression also detected by RT-PCR at E18.5 and adult (Wnt-13).
The developing eye begins as an outgrowth, or optic vesicle, originating from the ventral diencephalon and growing toward the surface ectoderm where it invaginates to form a bilayered optic cup. The outer layer differentiates as the RPE and the inner layer differentiates as the neural retina. At E12.5, the neural retina consists of a single layer of proliferating retinal precursor cells, but by E18.5, the neural retina consists of two layers—the outer neuroblast layer, which contains retinal precursor cells, and the inner retinal ganglion cell layer, which contains the RGCs. Our in situ hybridization analysis revealed that Wnt-3, -5a, -5b, and -7b were expressed diffusely in the neural retina at E12.5 and that Wnt-3, -5b, and -7b expression was down-regulated in the central retina by E14.5 (Fig. 2). Although we detected a Wnt-1 product by RT-PCR analysis in the neural retina, it is likely that this is not a specific signal, as we were unable to localize Wnt-1 transcripts in the developing retina by in situ hybridization (Fig. 2 and data not shown). Compared with the other Wnt transcripts, we detected high levels of Wnt-13 transcript in the RPE located in the most peripheral part of the optic cup, which corresponds to the ciliary margin, a region of the optic cup that is fated to give rise to the iris and ciliary body (Fig. 2). The expression of Wnt-13 in the RPE in this region was confirmed by in situ hybridization analysis on sections from albino (CD1) mice (Fig. 2).
To complement our analysis of Wnt gene expression in the retina, we also examined the expression of mouse frizzled (Mfz) and Sfrp genes in the embryonic retina. To date, nine Mfz genes have been described and we examined the expression of Mfz-3, -4, -6, and -7 by in situ hybridization and confirmed the expression in the neural retina of a fifth Mfz gene, Mfz-5, by RT-PCR. Our in situ hybridization analysis revealed that these four Mfz genes were expressed in the developing eye, but that they had unique expression patterns, especially in the ciliary margin (Table 2). At E12.5, Mfz-3, -4, -6, and -7 were expressed throughout the neural retina and ciliary margin, and Mfz-4 was also expressed in the RPE (Fig. 3 and data not shown). Differences in the expression pattern of Mfz genes in the ciliary margin of the optic cup were apparent by E14.5. Although low levels of Mfz-6 expression were maintained in the ciliary margin, Mfz-3 expression was down-regulated and Mfz-4 expression was up-regulated at the distal part of the ciliary margin, in a region that gives rise to the iris. Mfz-7 expression became restricted to the proximal half of the ciliary margin, in a region that gives rise to the ciliary body (Fig. 3). The expression of Mfz-4 in the ciliary body and developing iris was maintained until at least postnatal day 7 (P7; data not shown). At E14.5, Mfz-3 and Mfz-7 were also expressed in astrocyte precursor cells in the optic disc, as indicated by the intense triangular pattern of the in situ hybridization signal in this region of the retina and in the optic nerve (Fig. 3).
Table 2. Mfz Expression in the Eye From E12.5 to E18.5a
We examined the expression of the four murine Sfrp genes (Sfrp-1, -2, -3, and -4) in the developing eye by in situ hybridization. Sfrp-4 expression was not detected in the retina at E14.5 and E18.5 and was not studied further. Sfrp-1, -2, and -3 were expressed in the retina in distinct patterns (Table 3). Sfrp-1 was expressed throughout the neural retina at all stages, with higher levels of expression observed in the pigmented and nonpigmented layer of the ciliary margin as early as E12.5 until at least P7 (Fig. 4 and data not shown). Although Sfrp-2 was highly expressed in the neural retina throughout the developmental stages, it was down-regulated in the ciliary margin as early as E12.5 and, from E14.5 onward, it was down-regulated in the RGC layer (Fig. 4). A high level of Sfrp-1 and -2 expression was also observed in astrocyte precursor cells at the optic disc at E14.5. Sfrp-3 expression in the retina was not initiated until after E18.5 (Fig. 4); see text below.
Table 3. Sfrp Expression in the Developing Eye and Adult Neural Retinaa
Wnt Reporter Transgene Activity in the Developing Retina
To identify those regions of the developing eye that respond to Wnt signaling by means of the canonical pathway, we examined reporter gene activation in TCF/Lef-LacZ transgenic mice. In this reporter line, LacZ expression is under the control of a minimal HSP promoter and six copies of TCF/Lef-responsive element; thus, detection of β-gal activity provides an indication of where TCF-dependent gene expression is occurring within a tissue. From E11.5 to E14.5, high levels of β-gal activity were detected in the ciliary margin in both the presumptive RPE and nonpigmented inner layer but not in the proximal RPE at both of E13.5 and E14.5 (Fig. 5A–F and data not shown). At E17.5, β-gal activity was detected exclusively in the future ciliary body and iris (Fig. 5G). The pattern of β-gal activity in the ciliary margin overlapped with and was adjacent to the Wnt-13 expression domain (Fig. 2), the only identified Wnt gene expressed in this region. At E13.5, β-gal activity was also detected, albeit at lower levels compared with the ciliary margin, in cells located in the outer neuroblast layer of the neural retina, and in the anterior lens epithelium (Fig. 5A). By E14.5, however, β-gal activity in both of these regions was substantially reduced.
Expression of Wnt Pathway Genes in Perinatal and Adult Retina
Wnt signaling is likely to be important at perinatal and adult stages of the retina, as our RT-PCR analysis revealed that Wnt-5a, -5b, -10a, and -13 were expressed in the adult neural retina (Fig. 1). We, therefore, determined localization of Wnt-5a, -5b, and -13 in the P7 and adult retina by in situ hybridization. By P7, cells in the retina are organized into three layers; however, retinal histogenesis is incomplete, as several important processes, such as photoreceptor differentiation and synaptogenesis, are occurring at this age. Wnt-5a, -5b, and -13 were expressed in the inner nuclear layer (INL) at P7, in what are likely to be amacrine cells (Fig. 6 and data not shown), as determined by staining of serial sections with anti-Pax6 antibodies, which identify amacrine cells in this layer (Davis and Reed, 1996; Nishina et al., 1999). Wnt-13 expression was also detected in the RGC layer at P7. Wnt-5b and -13 transcripts were localized to the INL in the adult retina (Fig. 6). Although we detected the expression of Wnt-5a in the adult retina by RT-PCR analysis, we were unable to localize Wnt-5a transcripts in the adult retina by in situ hybridization.
Mfz-3, -6, and -7 were expressed in the RGC and INL layer at P7, but their expression was restricted to the INL in the adult retina. Mfz-4 was expressed in all three cellular layers at P7 and restricted to INL and in photoreceptors in the outer nuclear layer (ONL) in the adult (Fig. 6). The expression of the Mfz genes in the INL was diffuse, making it difficult to correctly identify the Mfz-expressing cells in this layer.
The Sfrp genes were also expressed at late stages of retinal development (Fig. 7; Table 3). Sfrp-1 expression was detected in all three cellular layers at P7 and in the INL and ONL in the adult retina (Fig. 7). At P7, Sfrp-2 and -3 were expressed in a complementary pattern: Sfrp-2 was expressed in the INL in the peripheral retina but not in central retina, and Sfrp-3 was expressed in the INL in the central but not in the peripheral retina (Fig. 7). We also noted that Sfrp-2 was expressed at higher levels in the nasal compared with the temporal retina at this stage. In the adult retina, Sfrp-2 expression was down-regulated and Sfrp-3 expression was restricted to the INL, possibly in bipolar cells, as indicated by the expression of Chx-10 (Fig. 7 and data not shown), a bipolar cell marker (Burmeister et al., 1996; Hatakeyama et al., 2001). The complementary distribution of Sfrp-2 and -3 transcripts could reflect the central to peripheral maturation gradient in the retina, with Sfrp-2 expression marking immature and Sfrp-3 expression marking more differentiated regions of the retina.
Expression of Wnt Pathway Genes in the Eyelid, Cornea, and Lens
Several of the Wnts, Mfzs, and Sfrps were expressed in other regions of the eye, such as the eyelid, cornea, and lens (Tables 1–3), implicating a role for Wnt signaling in the development of these structures. In the developing eyelid, for example, Wnt-1, -3, -4, -5b, -6, and Mfz-3, -4, -6, -7, and Sfrp-1 and -2 were expressed in the epithelial layer, whereas Wnt-5a, Mfz-7, and Sfrp-2 were expressed in the underlying mesenchyme. Wnts-2, -3, -4, -6, and Mfz-3, -4, -6, -7, and Sfrp-1 were expressed in the cornea epithelium. In the lens, Wnt-3, Mfz-3, -4, -7, and Sfrp-1 and -2 were expressed in the epithelium; Wnt-5b, Mfz-3, -6, and Sfrp-1 were expressed at the lens equator; and Wnt-7a and Sfrp-1 was expressed in the lens fibers (Figs. 2–4). In TCF/Lef-LacZ transgenic mice, β-gal activity was detected in the lens epithelium before E13.5 but was not detected after E14.5, suggesting that Wnt signaling is down-regulated in this region during development (Fig. 5 and data not shown). Mfz-7 expression was also detected in the extraocular mesenchyme at E12.5 and E14.5 (Fig. 3).
Although the expression of members of the Wnt signaling cascade has been reported in the vertebrate retina, this study represents the first systematic characterization of Wnt, Mfz, and Sfrp gene expression in the developing and adult murine retina. We show the localization of transcripts for several Wnt signaling components and Wnt reporter transgene activation in the developing eye. We also show that several members of the Wnt signaling cascade are expressed in the perinatal and adult retina, implicating a posthistogenesis role for this pathway in the eye.
Wnt Signaling in the Developing Neural Retina
Based on our in situ hybridization analysis, we have localized Wnt-3, -5a, -5b, and -7b transcripts to the embryonic mouse neural retina. Wnt expression has been reported previously in the chick RPE (Wnt-3, -5a, and -11) and neural retina (Wnt-11; Jin et al., 2002). We did not detect Wnt-3 and -5a expression in the RPE by in situ hybridization in pigmented (C57Bl/6) or albino (CD1) mice (Fig. 2 and data not shown), and we did not detect Wnt-11 expression in the mouse retina by RT-PCR (Fig. 1). The discrepancy between our results in the mouse and those reported in the chick could be related to species-specific differences in Wnt signaling in eye development. Wnt-7b expression has been reported in the mouse optic stalk and dorsal optic vesicle at E9.5 (Parr et al., 1993) but not at later stages in mouse development.
Wnt signaling by means of the canonical pathway involving β-catenin stabilization and TCF-dependent gene transcription appears to be activated in neural retina, ciliary margin, and other ocular structures, such as the lens. We assessed TCF-dependent transcription in the eye by examining β-gal activity in TCF/Lef-LacZ transgenic mice. Reporter gene activity was localized to the outer neuroblast layer of the neural retina at E13.5, which could reflect the activity of Wnt-3, -5a, -5b, and -7b that are expressed at low levels in the neural retina at this stage. The level of β-gal activity in the neural retina decreased from E13.5 to E14.5, indicating that Wnt signaling in this region is down-regulated during embryogenesis, which correlates with the reduction in the in situ hybridization signal of these Wnt genes in the neural retina from E12.5 to E14.5.
We find Wnt gene expression and TCF-dependent reporter gene activation in the neuroblast layer at a time when retinal precursor cells are proliferating and making cell fate choices; therefore, it is possible that Wnt signaling may be involved in these processes. Notably, at E13.5, β-gal activity was localized to cells in the neuroblast layer close to the RPE, which correspond to the region where presumptive photoreceptors are localized, raising the possibility that Wnt signaling plays a role in photoreceptor development. Wnt-3, -5a, and -7b have been shown to play a role in precursor cell proliferation in several tissues including the limb, hair follicle, lung, and caudal regions of the embryo (McMahon et al., 1992; Gofflot et al., 1998; Millar et al., 1999; Yamaguchi et al., 1999; Reddy et al., 2001; Shu et al., 2002). Wnt-5b is expressed in the developing gut (Lickert et al., 2001); however, the function of Wnt-5b in embryonic development is not known. Wnts have also been shown to be downstream targets of other signaling pathways in developing tissues. For example, Wnt-5a appears to be a target of Shh signaling in the hair follicle (Reddy et al., 2001). Because Shh signaling has been implicated previously in growth control and organization in the developing retina (Jensen and Wallace, 1997; Levine et al., 1997), it will be important to determine whether there is a link between Hedgehog and Wnt signaling pathways in retinal development.
The reception of a Wnt signal by a cell will depend in large part on the binding of Wnts to Mfz receptors and the presence of secreted Wnt antagonists, such as Sfrps, in the extracellular milieu that compete for Wnt binding. The expression of several Mfz genes in the eye has been characterized by in situ hybridization at early embryonic stages and by RNAse protection on RNA derived from total adult eye (Wang et al., 1996; Borello et al., 1999). We have extended those findings by characterizing the localization of four Mfz genes at later stages in embryogenesis and in the adult retina. The expression pattern of Sfrp-2 that we observed in the embryonic eye is in good agreement with previous studies reported by Leimeister et. al. (Leimeister et al., 1998); however, our observation that Sfrp-1 transcripts localize to the neural retina, in addition to the RPE, ciliary margin, and lens, is different from previous reports (Rattner et al., 1997; Leimeister et al., 1998). It is unlikely that this difference in the Sfrp-1 expression pattern between our study and the previous studies reflects a problem with probe specificity, as we find no evidence that the Sfrp-1 probe cross-hybridizes to Sfrp-2 (data not shown).
Our analyses and those of other groups have revealed a considerable overlap between Wnt, Mfz, and Sfrp expression in the neural retina, as well as in several ocular structures. This overlap in Mfz expression suggests the possibility that there are differences in the binding specificities of Wnts for the different Mfz and Sfrp and/or that signaling through different Mfz receptors results in different cellular responses. Based on previous reports showing an interaction between Wnt-5a and Mfz-5 in the Xenopus axis duplication assay (Ishikawa et al., 2001), it is possible that Wnt-5a preferentially signals by means of Mfz-5, which is also expressed in the neural retina (Wang et al., 1996; Borello et al., 1999). It is not clear, however, which Mfz the other Wnts present in the neural retina bind to.
Wnt-13 Signaling at the Ciliary Margin
Wnt signaling may be playing an important role in patterning and development of the distal eyecup or ciliary margin. We find that Wnt-13 is expressed in the cells at the tip of and in the RPE overlying the ciliary margin and that β-gal activity in TCF/Lef-LacZ reporter mice is up-regulated in the ciliary margin in cells within and adjacent to the Wnt-13 expressing cells. Our findings are in good agreement with previous reports showing Wnt-13 expression in this region of the chick eye (Jasoni et al., 1999; Fuhrmann et al., 2000; Kubo et al., 2003) and with the finding of Dorsky et. al., who reported activation of a Lef1/β-catenin–dependent reporter gene in the ciliary margin in developing zebrafish eye (Dorsky et al., 2002). Although previous studies on chick and on E9.5 mouse embryo have shown that Wnt-13 expression is restricted to the dorsal half of optic cup (Zakin et al., 1998; Jasoni et al., 1999), we find no evidence for asymmetric expression of Wnt-13 in the dorsal eyecup at E12.5 and later stages (data not shown).
The ciliary margin can be subdivided into a proximal and distal region that differentiates as ciliary body and iris, respectively. The pigmented cells in iris are also the source of retinal stem cells found in the adult neural retina (Tropepe et al., 2000). Our observation that Mfz genes are expressed in nonoverlapping subregions within the ciliary margin is consistent with the possibility that patterning within this region could be mediated by differences in Wnt signal reception through the different Mfz receptors. Wnt-13 signaling might also be kept localized in the ciliary margin by the antagonistic action of Sfrp-2, which is expressed in the adjacent neural retina but not in the ciliary margin.
Wnt Signaling in the Perinatal and Adult Retina
Wnt signaling may also be playing a role at late stages in retinal development and in the homeostasis and/or function of the adult retina. The expression of Mfz and Sfrp transcripts in the adult eye has been reported previously (Wang et al., 1996; Rattner et al., 1997; Chang et al., 1999), and we extend those findings by showing the localization of Wnts, Mfz-3, -4, -6, and -7, and Sfrp-1 and -3 in the mouse neural retina at perinatal and adult stages. However, Wnt genes are expressed at very low levels compared with Mfz or Sfrp genes. In contrast to Chang et al., who showed Sfrp-2 expression in the INL of the bovine retina (Chang et al., 1999), we find that Sfrp-2 expression is down-regulated in the murine retina after P7. Similarly, Rattner et al. (1997) reported that Sfrp-1 transcripts are restricted to the ciliary margin and lens in the adult mouse eye; however, we find that Sfrp-1 transcripts are also present in the neural retina at this stage.
Wnt signaling has been shown to play a role in synapse and axon remodeling and terminal arborization in other regions of the CNS, which raises the possibility that Wnt signaling performs a similar function in the perinatal retina (Salinas et al., 1994). Retinal histogenesis proceeds from the center to the periphery, such that by P7, the cells in the central retina have ceased proliferating and are undergoing differentiation and synaptogenesis (Cepko, 1993). The regression of Sfrp-2 expression toward the periphery of the retina, together with the onset of Sfrp-3 expression in the central retina, is consistent with the possibility that Sfrp-2 is involved in early events in histogenesis, whereas Sfrp-3 plays a role in later events, such as differentiation and synaptogenesis. Although Wnt expression in the adult CNS has been documented (Gavin et al., 1990; Roelink and Nusse, 1991; Salinas et al., 1994), the functional significance of Wnt signaling in this context is not known.
Wnt Signaling in the Developing Lens and Optic Disc
Based on our in situ hybridization analysis, it is apparent that Wnt signaling is playing a role in the development of several ocular structures, including the lens, cornea, and eyelid. We show that Wnt-3, -5b, and -7b, at least 4 Mfz genes, and Sfrp-1 and -2 are expressed in the developing lens, which is in good agreement with previous reports that localized Mfz and Sfrp gene expression to the lens (Rattner et al., 1997; Leimeister et al., 1998; Borello et al., 1999; Jin et al., 2002). Jasoni et al. (1999) reported the expression of Wnt-13 in the lens epithelium in the chick; however, we did not detect Wnt-13 expression in the mouse lens at any stages from E12.5 to E18.5. Some aspects of Wnt signaling in the lens epithelium likely go through the canonical Wnt signaling pathway, as we observed β-gal reporter gene activation in the lens epithelium in E13.5 TCF/Lef-LacZ embryos, which is in good agreement with a previous study showing activation of a Lef1/β-catenin-dependent reporter gene in the zebrafish lens (Dorsky et al., 2002). TCF/Lef-LacZ reporter gene activation was down-regulated in the lens at later stages, suggesting that Wnt-mediated up-regulation of TCF-dependent gene expression in the lens might be under strict temporal control.
Wnt pathway is likely also involved in the development of cells at the optic disc/nerve. The optic disc consists of a cuff of optic stalk-derived glioblasts that initially serve as a conduit for retinal ganglion cell axons into the optic nerve, and they eventually give rise to the astrocytes that are located in the nerve fiber layer on the vitreal side of the neural retina, which are intimately associated with retinal blood vessels. Hedgehog signaling likely also plays a role in the development of this population (Wallace and Raff, 1999); however, few other signaling molecules that regulate this population have been identified. We show that Mfz-3, -7, Sfrp-1, and -2 are highly expressed in the astrocyte population at the optic disc at E14.5, implicating a role for Wnt signaling in the development of the astrocyte lineage of the optic disc.
Based on the data we have shown here, it is reasonable to predict that Wnt-3, -5a, -5b, and -7b signaling may be involved in progenitor cell fate determination and proliferation in the developing neural retina and in homeostasis in the adult retina and that Wnt-13 signaling functions in the patterning and development of distal eyecup structures. Wnt signaling is likely to contribute to the development of the lens, optic nerve, cornea epithelium, and eyelid. It is also possible that additional Wnts could play a role in retinal development, as we did not always observe a complete overlap in the expression patterns of the Wnt, Wnt receptor, and Wnt antagonist genes.
Animals, In Situ Hybridization, and β-Galactosidase Activity
To prepare E12.5–18.5 embryos for in situ hybridization, time-mated C57BL/6 and CD1 mice (the day of vaginal plug was considered as E0.5) were anesthetized by CO2 and killed by cervical dislocation. Whole embryos were rinsed in cold Dulbecco's phosphate buffered saline (PBS) and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, overnight at 4°C. Tissues were then cryoprotected in 30% sucrose/PBS for 16 hr, equilibrated in 50:50 mixture of OCT and 30% sucrose for 1 hr, embedded, frozen, and stored at −80°C. To obtain postnatal day 7 (P7) and adult retinal tissues for in situ hybridization, mice were anesthetized with euthanyl and perfused with 4% paraformaldehyde/0.1 M phosphate buffer, pH 7.4. After removal of the lens, the eyes were prepared, as described in above. Tissues were sectioned at 14 μm and processed for in situ hybridization, as described previously (Jensen and Wallace, 1997) with the following DIG-labeled antisense riboprobes: Wnt-1, -2, -3, -4, -5a, -5b, -6, -7a, and -7b (a kind gift of A. McMahon); Wnt-13 (a kind gift of L. Zakin); Mfz-3, -4, -6, and -7 (a kind gift of J. Chin Hsieh); and Sfrp-1, -2, -3, and -4 (a kind gift from A. Rattner). The generation and characterization of TCF/Lef-LacZ reporter mice is described elsewhere (Mohamed et al., manuscript submitted for publication). Heterozygous TCF/Lef-LacZ mice were time-mated with wild-type CD1 females, and the embryos were dissected and embedded directly in OCT, frozen, cryosectioned at 14 μm, and stained for β-galactosidase activity, as previously described (Dufort et al., 1998). Sections were viewed under a Zeiss Axioplan microscope, and digital images were captured by using an Axio Vision 2.05 (Zeiss) camera and processed with Adobe Photoshop.
Total RNA was extracted by using TRI REAGENT (Sigma) from freshly dissected embryonic and adult neural retina. Care was taken during the dissection to remove the RPE and all other ocular structures. Five micrograms of total RNA was reverse transcribed using Invitrogen oligo (dT) primer and Thermoscript RT, following the manufacturer's instructions. The resultant first-strand cDNA was diluted threefold and 3 μl was used for PCR analysis with the following intron-spanning primer pairs: Wnt-1: F 5′-aaatcgcccaacttctgca-3′, R 5′-aatacccaaagaggtcacagc-3′; Wnt-4: F 5′-tgtacctggccaagctgtcat-3′, R 5′-tccggtcacagccacactt-3′; Wnt-7b: F 5′-accgtcttcgggcaagaact-3′, R 5′-cctggcgttctttttgatct-3′; Wnt-13: F 5′-tgtactctgcgcacctgct-3′, R 5′-tgcactcacactgggtgac-3′. Primer sequences for Wnt-2, -3, -3a, -5a, -5b, -6, -7a, -8b, -10a, -10b, and -11 were the same as those used by Lako et al. (2001). Amplification of untranscribed RNA was used as a negative control and amplification of cDNA with a primer pairs for the mHPRT gene (F 5′-ctgctttccggagcggtagc-3′, R 5′-caacttgcgctcatcttagg-3′) served as a positive control. The cycling parameters were as follows: 2 min at 94°C, followed by 30 cycles consisting of 1 min at 94°C, 1 min at the annealing temperature (determined for each Wnt primer pair), 45 sec at 72°C, followed by a final 7 min extension at 72°C.
We thank Drs. McMahon, Zakin, Hsieh, and Rattner for probes and Dr. Dakubo and members of the Wallace laboratory for helpful discussion. V.A.W. was funded by grants from the National Cancer Institute of Canada with funds from the Terry Fox Run and the Canadian Institutes of Health Research (CIHR) and D.D. was funded by the CIHR. V.A.W. is the recipient of a CIHR New Investigator Award, and D.D. is a Chercheur Boursier du Fonds de la Recherche en Sante du Quebec (FRSQ).