Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Biol. Author manuscript; available in PMC 2010 August 15.
Published in final edited form as:
PMCID: PMC2917894

BMP7 and SHH Regulate Pax2 in Mouse Retinal Astrocytes by Relieving TLX Repression


Pax2 is essential for development of the neural tube, urogenital system, optic vesicle, optic cup and optic tract. In the eye, Pax2 deficiency is associated with coloboma, a loss of astrocytes in the optic nerve and retina, and abnormal axonal pathfinding of the ganglion cells axons at the optic chiasm. Thus, appropriate expression of Pax2 is essential for astrocytes determination and differentiation. Although BMP7 and SHH have been shown to regulate Pax2 expression, the molecular mechanism by which this regulation occurs is not well understood. In this study, we determined that BMP7 and SHH activate Pax2 expression in mouse retinal astrocyte precursors in vitro. SHH appeared to play a dual role in Pax2 regulation; 1) SHH may regulate BMP7 expression, and 2) the SHH pathway cooperates with the BMP pathway to regulate Pax2 expression. BMP and SHH pathway members can interact separately or together with TLX, a repressor protein in the tailless transcription factor family. Here we show that the interaction of both pathways with TLX relieves the repression of Pax2 expression in mouse retinal astrocytes. Together these data reveal a new mechanism for the cooperative actions of signaling pathways in astrocyte determination and differentiation and suggest interactions of regulatory pathways that are applicable to other developmental programs.

Keywords: SHH, BMP7, TLX, PAX2, astrocytes, retina


During induction of the eye, the interaction of microenvironmental cues, signaling pathways, and transcription factors are critical to the proper patterning of the optic cup, lens, and optic stalk. There are a large variety of inductive factors, such as retinoic acid, fibroblast growth factors (FGFs), sonic hedgehog (SHH), WNTs, activins, and bone morphogenetic proteins (BMPs) that are essential for various aspects of patterning. In addition, there are a large number of transcription factors whose expression patterns rely in part on the presence of these inductive factors (Fuhrmann et al., 2000; Vogel-Hopker et al., 2000; Zhao et al., 2001). Although the regulation of transcription factor expression by these factors is partially characterized, our understanding of the mechanisms by which these factors cooperate needs investigation.

One protein that has been identified as critical to optic cup development is PAX2, a member of the Paired Box family of transcription factors. PAX2 plays a pivotal role in eye, ear, spinal cord and kidney development (Dressler et al., 1990; Eccles et al., 1992; Nornes et al., 1990). Loss of Pax2 expression has been linked to various developmental and functional abnormalities such as coloboma (lack of choroid fissure formation), a marked reduction in optic nerve and retinal astrocytes, misprojections of the optic nerve, inner ear malformations, and kidney hypoplasia (Dressler and Douglass, 1992). Interestingly, Pax2 overexpression in the optic cup can phenocopy the loss of Pax2 expression in some respects, namely the formation of colobomas, albeit the formation of the colobomas in gain- and loss-of-function studies appears to be different (Sehgal et al., 2008). In the vertebrate eye, PAX2 is co-expressed with other transcription factors like PAX6 and CHX10, throughout the optic vesicle stage. As development proceeds, inductive factors restrict the expression patterns of the various transcription factors to lens, retina, retinal pigmented epithelium (RPE) or optic stalk. At the optic cup stage, PAX2 is expressed in ventral optic cup and optic stalk and eventually becomes restricted entirely to the optic stalk, the cells of which eventually give rise to the astrocytes of the optic nerve (Chu et al., 2001). A subpopulation of the optic nerve astrocytes then migrate into the optic cup to generate retinal astrocytes (Chan-Ling and Stone, 1991; Zhang and Stone, 1997)

The BMPs are a large family of secreted proteins that are known to be critical for the formation of a large array of tissues in the developing organism (Chen et al., 2004). BMPs can signal through SMAD-dependent and SMAD-independent pathways (Derynck and Zhang, 2003; Herpin and Cunningham, 2007). In the canonical signaling pathway, BMP signaling is mediated by the receptor SMADs (R-SMADs), 1, 5, and 8. Upon activation, these SMADs form a complex with SMAD4 and the complex is translocated to the nucleus to regulate transcription (Dudley and Robertson, 1997). SMADs can regulate gene expression by binding directly to DNA sequences, or by interacting with co-activators and co-repressors (Lee et al., 2000; Miyazono, 1999; Miyazono, 2000a; Miyazono, 2000b; Miyazono et al., 2005; Zhang and Derynck, 2000).

Many of the BMPs are expressed in tissues surrounding the eyefield, optic vesicles, and optic cups, and within portions of the eye itself in the developing and mature organ (Belecky-Adams and Adler, 2001; Dale and Jones, 1999; Hocking and McFarlane, 2007; Koshiba-Takeuchi et al., 2000; Wilson et al., 2007; Wordinger et al., 2002). Organogenesis of the eye appears to be reliant on BMPs, in particular BMP4 and 7 (Adler and Belecky-Adams, 2002; Dudley et al., 1995; Koshiba-Takeuchi et al., 2000; Sakuta et al., 2001; Trousse et al., 2001). BMP7 is hypothesized to be critical for many steps in the progressive development of the eye, including the patterning of the eyefield, optic stalk, and optic nerve head, and differentiation of the retinal pigmented epithelium, lens placode, anterior segment, and retinal ganglion cells (Adler and Belecky-Adams, 2002; Dale et al., 1997; Dale and Jones, 1999; Dudley and Robertson, 1997; Furuta et al., 1997; Hung et al., 2002; Jena et al., 1997; Morcillo et al., 2006; Muller et al., 2007; Wawersik et al., 1999; Wordinger et al., 2002; Zhao et al., 2002). At the eye field stage, BMP7, is expressed along with SHH in the prechordal mesoderm underlying the diencephalon and controls the identity of the rostral diencephalon (Dale et al., 1997). BMP7 has been shown by several investigators to regulate the expression of Pax2, although it is unclear whether this regulation is direct or indirect (Adler and Belecky-Adams, 2002; Dale et al., 1997; Morcillo et al., 2006).

SHH is a member of the hedgehog family that, much like the BMPs, appears to be involved in the development of many tissues, including midline structures, limbs, gastrointestinal tract, and central nervous system (CNS; (Chari and McDonnell, 2007). In vertebrates, the hedgehog proteins signal through the GLI1, GLI2 and GLI3 transcription factors (Hooper and Scott, 2005). How the GLI proteins are regulated in vertebrates is still not well understood. In both Drosophila and vertebrates, the SHH binds to the patched homologues (PTC) and relieves repression on another transmembrane protein, smoothened (SMO). In Drosophila, active SMO controls the proteolytic processing of the GLI homologue cubitus interruptus (Ci). Ci is bifunctional; in its full length form the protein acts as a transcriptional activator, but following cleavage, the protein will act as a transcriptional repressor (Parkin and Ingham, 2008). SMO activity regulates the Ci processing through a complex that includes kinesin protein Cos-2, a serine-threonine kinase fused, and a novel protein known as suppressor of fused (Huangfu and Anderson, 2006). The signaling of the SMO protein and the conservation of pathway members is not clear in vertebrates, and only one of the GLI proteins, GLI3, has been shown to be proteolytically cleaved to give rise to the transcriptional repressor form of the protein (Huangfu and Anderson, 2006). Hence, GLI3 can function as a functional repressor and activator, while GLI1 and GLI2 are thought to act only as activators (Bai et al., 2004; Motoyama et al., 2003).

During development, SHH is released from the prechordal mesoderm to induce the formation of the floorplate in the midline neural plate (Dale et al., 1997). The floorplate itself then expresses SHH which is proposed to regulate the formation of ventral neural tube cell fates. The eyefield, which initially forms in the region overlying the prechordal mesoderm, is cut into two pieces as the floorplate of the neural plate is induced by SHH (Ekker et al., 1995; Macdonald et al., 1995; Nagase et al., 2005; Ohkubo et al., 2002). At a molecular level, SHH induces Pax2 expression in this region, which in turn reciprocally inhibits Pax6. The region that continues to express Pax2 in the ventral diencephalon becomes the optic stalk. Later in development, SHH expression is restricted to retinal ganglion cells and SHH derived from ganglion cell axons is thought to drive the proliferation of optic nerve astrocytes (Dakubo et al., 2008; Dakubo et al., 2003; Wallace and Raff, 1999). SHH has also been shown by several investigators to regulate the expression of Pax2 in the eye (Macdonald et al., 1995; Schimmenti et al., 2003; Wang et al., 2005). Bovolenta and colleagues found that Pax2 is regulated by sequential activity of SHH and BMP7 (Morcillo et al., 2006). While it is generally accepted that SHH regulates Pax2 expression, little is known about the mechanism of Pax2 regulation by these signaling molecules and whether the regulation is direct or indirect (Macdonald et al., 1995; Morcillo et al., 2006; Nakamura, 2001).

Pax2 is also regulated in the vertebrate eye by TLX, a member of the tailless class of orphan nuclear receptors, which has been shown to repress expression of Pax2 (Yu et al., 2000; Yu et al., 1994). In early stages of development, Tlx is expressed mainly in the dorsal retina with low levels of expression in the ventral retina, whereas Pax2 expression is limited primarily to the ventral retina and optic stalk (Nornes et al., 1990). In the chick optic vesicle, expression of Tlx and Pax2 is dynamic and transiently overlaps, particularly in the presumptive optic stalk. Later in development, Tlx is expressed in retinal progenitor cells and Müller glial cells in the retina. TLX represses Pax2 by binding to a conserved recognition sequence (AAGTCA) ~ 80base pairs upstream of the TATA box (Miyawaki et al., 2004; Yu et al., 2000).

The studies reported here were directed at understanding the molecular mechanism by which SHH and BMP7 regulate Pax2 expression in mouse retinal astrocytes. We show that 1) BMP7 expression appears to be regulated by SHH, and 2) SHH and BMP7 interact with TLX on the Pax2 promoter and relieve repression by TLX in retinal astrocytes.


Phospho-SMAD1 and GLI2 and 3 are present in the developing optic vesicle, optic stalk, and nerve

Previous reports from our laboratory and others have proposed that SHH and BMP7 are critical to the development of the optic stalk astrocytes and have shown that mRNA encoding intracellular signaling components of the BMP and SHH pathways are present in the developing optic stalk (Belecky-Adams and Adler, 2000; Dakubo et al., 2003; Furimsky and Wallace, 2006; Morcillo et al., 2006). Immunohistochemistry was performed to compare the localization of PAX2 with GLI2, GLI3, and pSMAD1 proteins at embryonic day 9.5 (E9.5) optic vesicle, E10.5 optic stalk, and E16 optic nerve. As has been described in previous reports (Nornes et al., 1990), PAX2 was not present in the optic stalk at E9.5 (Fig 1A–C;. However, GLI2 (Fig 1D–F), GLI3 (Fig 1G–I) and pSMAD1 (Fig 1J–L) were all present in the optic vesicle at E9.5. GLI2 and phospho-SMAD1 (pSMAD1) were both localized primarily in cells at the ventricular edge of the optic vesicle, where cells were undergoing mitosis (compare Fig 1F and L), while GLI3 immunolabel was present throughout the optic vesicle (Fig 1I). Further, GLI3 label was also present in the presumptive lens ectoderm and in a subpopulation of mesenchyme surrounding the optic vesicle (Fig 1I). At E10.5, PAX2 was localized primarily to the ventral optic cup and optic stalk (Fig 2A–D). GLI2 protein was localized to the optic stalk, optic cup and lens vesicle (Fig 2E–H). GLI3 was widely expressed throughout the optic stalk, retinal pigmented epithelium (RPE), optic cup, and lens vesicle (Fig 2I–L). Phospho-SMAD1 was also localized to the optic stalk, RPE, optic cup and lens placode (Fig 2M–P). As was noted at E9.5, GLI2 and pSMAD1 immunolabel appeared to be more intense at the ventricular edge of the dorsal and ventral optic stalk and optic cup, where dividing cells are located (arrowheads, Fig 2H and P). In the E16 eye, PAX2 is restricted to the optic nerve head and optic nerve (Fig 3A–C). GLI2 (Fig 3D–F), GLI3 (Fig 3G–I), and pSMAD1 (Fig 3J–L) were all localized to subpopulations of cells in the optic nerve head (ONH) and optic nerve (ON).

Figure 1
GLI2, GLI3, and pSMAD1 are expressed in E9.5 optic vesicle prior to PAX2
Figure 2
Expression of PAX2, GLI2, GLI3, and pSMAD1 in E10.5 optic cup
Figure 3
Expression of PAX2, GLI2, GLI3, and pSMAD1 in E16 optic nerve

Characteristics of retinal astrocyte precursors in vitro

To begin to unravel the molecular mechanism by which the BMP7 and SHH pathways might regulate the differentiation of retinal astrocytes, we turned to an in vitro model system which would allow us to uniformly treat cells with various combinations of the two growth factors and probe molecular interactions between the two pathways. We chose to use primary retinal astrocyte precursors isolated from dissociated retina obtained from 4-week-old transgenic mice that ubiquitously express a temperature sensitive SV-40 large T antigen (immorto-mice; (Jat et al., 1991)). Collagenase-treated cells were cultured following the depletion of endothelial cells using antibody-coated magnetic beads (Scheef et al., 2005). Cells were plated and grown at the permissive temperature of 33°C and surviving cells allowed to reach confluence over a 2–3 week period. Confluent cells were subsequently characterized using immunocytochemistry and fluorescence-activated cell sorting (FACs) analyses. Figure 4A and B show the morphology of the cells under phase contrast microscopy. Cells were positive for PAX2 (Fig 4C), GFAP (Fig 4D), and NG2 (Fig 4E). Consistent with previous analyses of the cells, FACs analyses indicated that nearly 100% of the cells expressed GFAP and NG2 (Fig 4F, G;(Scheef et al., 2005)). The immunocytochemistry and FACs analyses coupled with the previous analysis of the cells indicated that the cells showed characteristics consistent with multipotent retinal astrocyte precursors (Scheef et al., 2005; Stallcup and Beasley, 1987; Yokoyama et al., 2006).

Figure 4
Characterization of mouse retinal astrocyte precursors

SHH and BMP7 activate PAX2 and phospho PAX2

To determine if addition of BMP7 and/or SHH to retinal astrocytes affected morphology, cells were treated with vehicle, BMP7, SHH, or BMP7+ SHH for 2 days and immunolabeled for PAX2 or GFAP. While there were no detectable changes in appearance or number (data not shown) of the cells, there were evident increases in the amount of PAX2 immunofluorescence of treated cells in comparison to the control vehicle-treated cells (Fig 5A). To verify the increase in levels of PAX2, lysates of treated cells were subjected to western analysis. Cells were treated with vehicle, BMP7, SHH, or both BMP7 and SHH for 4 hours, cell lysates separated by SDS-PAGE, and immunoblot for PAX2 performed. As shown in Fig 5; B, there was a notable increase in both isoforms of PAX2 in lysates from treated cells in comparison to the vehicle control (Fig 5B). These assays showed the 46 and 48 kDa forms seen in previous studies (Dressler and Douglass, 1992) also were present in these cells (Fig 5B). Densitometric analysis revealed a 3-fold increase in PAX2 protein in lysates treated with BMP7 or SHH and a 4-fold increase in lysates co-treated with BMP7 and SHH compared to control cultures treated with vehicle alone (Fig 5B′).

Figure 5
SHH and BMP7 promote increased levels of PAX2 and phospho PAX2 in retinal astrocytes

PAX2 activity is regulated by phosphorylation of a carboxyl-terminal trans- activation domain that is rich in serine and threonine residues (Cai et al., 2002). To determine if there was a concomitant increase in phospho-PAX2 following treatment of retinal astrocytes with BMP7, SHH, or both, western blotting of lysates from treated cells was performed using a phospho-PAX2 specific antibody. Immunoblots showed 1.6-fold and 2-fold increases in phospho-PAX2 levels in lysates treated with BMP7 or SHH, respectively, in comparison to vehicle treated controls, while lysates co-treated with BMP7 and SHH showed an approximately 3-fold increase over controls (Fig 5C, C′).

Cells treated with SHH have increased levels of phospho SMAD1

Because the observed increases in both PAX2 and phospho-PAX2 in the cells treated with BMP7 and SHH were sub-additive, the possibility of cross talk between the two signaling pathways was investigated. Retinal astrocytes were treated with control vehicle, BMP7, SHH, or both BMP7 and SHH and lysates of cells were assayed for the downstream signaling component of the canonical BMP pathway, pSMAD1. Vehicle-treated cells showed a basal level of pSMAD1 (Fig 6A). As expected, BMP7-treated cells showed an increase in the levels of pSMAD1. Interestingly, there was also an increase in pSMAD1 following SHH treatment, and co-treatment yielded a sub-additive increase in pSMAD1 similar to what had been seen in previous experiments in which BMP7 and SHH were added to the cultures.

Figure 6
Cells incubated with SHH have increased levels of phospho SMAD1

A densitometric analysis of the western blot using β-tubulin as a loading control showed a 2-fold increase in pSMAD1 levels in BMP7- treated, a 2.7-fold increase in SHH-treated and a 3-fold increase in co-treated cells over vehicle (Fig 6A, A′). As a negative control the same blot was subjected to an antibody specific for phospho-SMAD2, the downstream signaling component of activin and other TGFβ growth factors. Quantitative analysis showed no significant changes in phospho-SMAD2 following treatment of cells in comparison to the vehicle controls (Fig 6B, B′).

Follistatin abolishes the increase in phospho SMAD1 protein in SHH-treated cells

The previous result suggested that the SHH pathway was regulating pSMAD1 levels, either through pathway cross talk or by regulating expression of one or more members of BMP pathway. If SHH regulates BMP7 expression, we hypothesized that treatment of cells with a BMP7 inhibitor should abolish the effects of both BMP7 and SHH on pSMAD1 levels. To block BMP signaling, we treated cells with each growth factor separately or in combination with follistatin and a BMP7 function blocking antibody. Cell lysates were analyzed for levels of pSMAD1. Consistent with our prior observations, treatment of the cells with BMP7, SHH or both factors yielded an increase in pSMAD1 that was greater in the lysates from control-treated cells (Fig 6D). As would be expected if SHH signaling were regulating BMP expression, co-treatment with follistatin and BMP7 function blocking antibody blocked pSMAD1 signaling following treatment with each factor alone or in combination. A densitometric analysis of the blots using α-tubulin as a loading control showed a 2-fold increase in lysates treated with BMP7, a 2.5-fold increase in SHH-treated lysates and a 3.3-fold increase in lysates co-treated with BMP7 and SHH. Lysates treated with growth factors and combination of follistatin and BMP7 function blocking serum showed >10-fold decrease in levels of pSMAD1 (Fig 6D, D′).

Cyclopamine treatment decreases BMP7 protein levels in vitro

To further test if BMP7 expression is regulated by SHH, cells were treated with vehicle or cyclopamine (an inhibitor of SHH signaling), and analyzed for levels of BMP7 by western blotting. BMP7 protein levels in cyclopamine treated cells were decreased as compared to the vehicle treated cells. Densitometric analysis showed a 2-fold decrease in BMP7 protein levels in cyclopamine treated cells compared to vehicle controls (Fig 6E, E′).

BMP7 and SHH decrease the binding of TLX protein on Pax2 promoter

Tailless (TLX) is an orphan nuclear receptor that acts as a suppressor of Pax2 expression and is essential for eye development (Yu et al., 2000). We hypothesized that TLX repression would have to be removed to allow Pax2 expression and induction of astrocyte fate in the optic stalk. To begin to examine whether the intracellular signaling pathways activated by BMP7 and SHH decreased TLX availability to interact with the proximal Pax2 promoter sequences, we performed EMSA reactions using nuclear extracts of treated astrocytes and biotin labeled oligonucleotides containing the TLX binding site in the Pax2 promoter. If the BMP and/or SHH pathways interfered with TLX repression, then one might expect to see a decrease in TLX binding to its consensus sequences in the presence of nuclear extracts from treated cells. As was hypothesized, TLX binding activity for the Pax2 promoter sequence was decreased in cells treated with both BMP7 and SHH (Fig 7A, C; n=4). Parallel EMSA reactions using in vitro translated TLX protein (Fig 7B, Lane 2) and nuclear extracts from astrocytes (Fig 7B, Lane 3) served as controls for TLX protein/DNA complex migration (Fig 7B). Further, treatment of EMSA reactions containing nuclear extracts from astrocytes incubated with the oligo representing the TLX binding site followed by incubation with an antibody specific for TLX showed a supershift of the detected band confirming the presence of TLX in the complex (Fig 7B, lane 4).

Figure 7
BMP7 and SHH decrease the binding of TLX protein on Pax2 promoter

To confirm that there was indeed less TLX bound to the TLX consensus site of the Pax2 promoter, a chromatin immunoprecipitation (ChIP) using TLX antibody was performed using lysates from retinal astrocytes treated with vehicle, BMP7, SHH, or BMP7 + SHH. Quantitative PCR (qPCR) was done using primers specific for a 60 bp region of the Pax2 promoter encompassing the TLX-binding region of the Pax2 promoter (Figure 7D). Comparable to the EMSA analysis, treatment of cells with BMP7, SHH, or BMP7 + SHH decreased the amount of Pax2 amplified, indicating that less TLX was immunoprecipitated as a complex with the Pax2 promoter in treated cells (Fig 7D).

TLX, GLI2 and pSMAD1 interact with each other

There are several ways which the SHH and BMP pathways might decrease TLX binding to its consensus site. For instance, the SMADs and/or GLIs might 1) bind to TLX to keep it from binding to DNA or release it from the consensus site, or 2) bind to DNA and sterically interfere with TLX binding to its consensus sequences. To analyze whether decreased binding of TLX following BMP7 and SHH signaling involves interaction of their downstream components pSMAD1 and GLI2 respectively, co-immunoprecipitation (Co-Ip) was performed. Retinal astrocytes were treated with vehicle and total proteins were isolated. Extracted nuclear proteins from treated cells were precipitated with control β-tubulin, GLI2 or pSMAD1 antibodies and then blotted with antibody specific to TLX (Fig 7E; a), pSMAD1 (Fig 7E; b) or GLI2 (Fig 7E; c). As expected, lysates incubated with control β-tubulin antibody do not co-precipitate TLX, GLI2, or pSMAD1 (Fig 7E; first lane). All of the proteins were detected following immunoprecipitation using the pSMAD1 or GLI2 antibodies, consistent with the hypothesis that these proteins interact with each other (n=3). To determine if TLX, GLI2 and/or pSMAD1 could also form complexes in treated cells, Co-Ips were performed using nuclear extracts from cells treated with BMP7, SHH, or BMP7 + SHH (n=3). As was seen in vehicle-treated cells, TLX, GLI2, and pSMAD1 were immunoprecipitated as a complex when cells were incubated with pSMAD1 or GLI2 antibody (data not shown).

An alternate hypothesis would be that BMP7 and SHH treatment of astrocytes downregulates the expression of TLX. To rule out BMP7 or SHH regulation of TLX expression, western blotting was performed on BMP7, SHH, and co-treated cells. No significant differences in amounts of TLX protein were apparent in the treated cells compared to controls (Fig 7F, F′).

To test whether SMAD1 or GLI2 could bind to sequences within the proximal promoter of Pax2 and thereby sterically hinder TLX binding, we performed an avidin-biotin complex DNA-binding (ABCD) assay. This assay is used to evaluate the interactions of endogenous proteins (SMAD1 and GLI2) with the biotinylated double-stranded DNA oligonucleotides containing the binding element of interest. Total cellular proteins from retinal astrocytes were isolated and incubated with biotin-labeled oligonucleotides representing the TLX binding site in the Pax2 promoter. Proteins bound to the TLX binding site oligonucleotides were purified by streptavidin agarose chromatography, eluted from the beads, separated on SDS-PAGE gels, transferred and blotted with antibody specific for TLX, pSMAD1, or GLI2. Consistent with our other observations, TLX protein was observed to bind to the TLX binding site (Fig 7G). Interestingly, neither pSMAD1 or GLI2 could be detected, although control experiments showed that pSMAD1 and GLI2 were detectable in assays using the consensus SMAD or GLI sites respectively (data not shown).

SHH and BMP7 regulate Pax2 promoter activity in the presence of TLX

To determine if SHH and BMP7 can regulate the activity of the Pax2 promoter within retinal astrocytes, luciferase assays were performed. Astrocytes were transiently co-transfected with an expression vector for TLX (CMV-TLX) and a luciferase reporter gene containing 900 bp of the proximal Pax2 upstream region (Pax2-Luc). Cells were incubated with vehicle or growth factors for 24 hours, lysed, and total proteins were analyzed for luciferase activity. As expected, cells co-transfected with CMV-TLX and Pax2-Luc vectors showed decreased levels of luciferase activity in comparison to those transfected with Pax2-Luc plus control expression vector (Fig 8A). Cells transfected with Pax2-Luc and subsequently treated with either BMP7 and/or SHH showed at least a 2-fold increase in luciferase activity. Consistent with our hypothesis, cells co-transfected with CMV-TLX and Pax2-Luc and treated with either BMP7, SHH or BMP7 and SHH were able to relieve Pax2 repression to varying degrees (Fig. 8A).

Figure 8
SMADs and GLIs regulate the Pax2 promoter in astrocytes and are present at the Pax2 promoter

P-SMAD1 and GLI2 decrease TLX-binding at the Pax2 proximal promoter

To verify that pSMAD1 and GLI2 are found at the Pax2 promoter in situ, cells were treated with vehicle or BMP7 + SHH and a ChIP analysis performed on the lysates using antibodies specific for pSMAD1, GLI2, or TLX. The results of the ChIP were then analyzed by qPCR using primers that amplified the proximal Pax2 promoter (Fig 8B). As was seen in the previous ChIP analysis, TLX appeared to be abundant at the Pax2 promoter in vehicle-treated cells. However, in lysates from treated cells there was a decrease in Pax2 amplification of approximately 2-fold. Furthermore, treated cells showed a dramatic increase in the levels of amplified Pax2 when pSMAD1 and GLI2 were used for the ChIP analysis. Together these results support the hypothesis that untreated retinal astrocyte precursors have high levels of TLX bound to the Pax2 promoter, while cells treated with BMP7 and/or SHH have relieved TLX repression of Pax2 expression by removal of TLX and a concomitant increase in pSMAD1 and GLI2 bound to the proximal Pax2 promoter.


The present study focuses on the mechanism of regulation of Pax2 by SHH and BMP7. Indirect immunohistochemistry showed that critical members of the BMP and SHH pathways were present in the optic vesicle just prior to Pax2 expression and were present in the PAX2-expressing optic stalk and optic nerve. Primary cultures of retinal astrocytes were used to probe the interactions of the two pathways in retinal astrocytes. Cells were treated with BMP7 and SHH and the results analyzed by western blotting, electrophoretic mobility shift assay, ABCD assay, co-immunoprecipitation, luciferase assays, and ChIP-qPCR. The following conclusions could be drawn from these experiments; 1) BMP7 and SHH increase PAX2 and phospho- PAX2 levels in astrocytes, 2) SHH regulates BMP7 expression in astrocytes, 3) SHH and BMP7 relieve Pax2 repression by decreasing the binding of TLX, and 4) the modulation of TLX repression of Pax2 expression is likely to include binding of both pSMAD1 and GLI2 to TLX, rather than regulation of TLX mRNA or protein levels or steric interference of TLX binding to the Pax2 promoter.

Cooperation of SHH and BMP pathways

Although the SHH and BMP pathways are typically described as having antagonistic effects in patterning of the central nervous system, our results, as well as the results of other labs, indicate that this may be a simplistic model (Arkell and Beddington, 1997; Liem et al., 1995). An excellent example of this is the interactions of BMPs and SHH in the neural tube. BMPs have been proposed as one of the “dorsalizing” factors, along with wnts and activins, whereas SHH has been proposed as the “ventralizing” factor of the neural tube (Ulloa and Briscoe, 2007). What is rarely acknowledged is the fact that some of the BMPs (as well as certain wnts) are also expressed in the prechordal mesoderm underlying the forebrain and midline portion of the neural tube (Dale et al., 1997; Furuta et al., 1997; Placzek and Briscoe, 2005). Dale et al. (1997) have shown that both BMP7 and SHH are co-expressed in the prechordal mesoderm and are required to induce the formation of ventral midline cells of the rostral diencephalon. BMP7 reportedly acts on the midline neural cells to modify their response to SHH, inducing the expression of markers found only in rostral diencephalon (Dale et al., 1997). Interestingly, studies by Morcillo et al (2006) have also suggested that the activity of BMP7 and SHH are necessary for the proper patterning of the optic nerve head and optic fissure and that the activity of BMP7 prior to SHH signaling is necessary for Pax2 expressing cells in the optic nerve head.

These findings are not restricted to the optic nerve and Pax2. Investigations in other tissues also suggest that SHH and BMPs can act cooperatively (Suzuki et al., 2004; Yuasa et al., 2002). The cooperation of SHH and BMPs has been characterized at several different levels. First, SHH and BMPs have been shown by various investigators to regulate each other’s expression (Kawai and Sugiura, 2001; Zhang et al., 2000), although the level of BMPs present appeared to determine whether the SHH expression is upregulated or inhibited (Zhang et al., 2000). That BMPs may regulate the expression of SHH is consistent with the idea that BMP7 and SHH sequentially regulate patterning and differentiation in the diencephalon (Dale et al., 1997; Morcillo et al., 2006). A second mechanism by which SHH and BMPs may act coordinately may be crosstalk of the pathways with one another. For instance, there is evidence that the BMP pathways interact with the notch, Toll, and MAP Kinase pathways (reviewed in (Herpin and Cunningham, 2007). Finally, the last way in which the BMP and SHH pathways might act coordinately is through co-regulation of genes. The SMADs have been shown to regulate gene expression either by binding to DNA directly with co-activators or co-repressors, or by binding to other transcription factors to regulate their binding to DNA (Miyazono et al., 2005). There are at least 3 consensus DNA recognition sequences for the BMP-activated SMADs (Kim et al., 1997; Kusanagi et al., 2000; Shim et al., 2002). The SMADs have also been shown to interact with many different transcription factors, although only a handful of factors have been implicated in binding to the BMP-related SMADs, including Runx, Menin, YY1, Msx1, Vent2, MyoD, Hoxc8, SIP1, FoxG1, p300, CBP, cSki, SnoN, and GCN5 (Miyazono et al., 2005). This is one of the first studies to show that BMP and SHH pathways may interact with each other to regulate gene expression.

TLX and Pax2 regulation

TLX is an orphan nuclear receptor that plays an important role in regulating cell numbers and astrocyte development in the developing retina (Miyawaki et al., 2004). TLX is also implicated in forebrain development and emotional behavior (Roy et al., 2002). TLX regulates Pax2 by binding the promoter at a recognition sequence which has been conserved in mouse and humans (Yu et al., 2000). In addition, TLX is expressed in proangiogenic astrocytes and its expression is controlled by oxygen concentration. TLX acts as proangiogenic switch which becomes active in low concentrations of oxygen. After retinal angiogenesis, TLX is downregulated and its angiogenic activity is switched off when astrocytes come in contact with the blood vessels (Uemura et al., 2006). BMPs also are expressed in endothelial cells as well as playing key roles to regulate the developmental program of hematopoietic cells in humans (Bhatia et al., 1999; Fan et al., 2002). In this study, BMP7 and SHH were shown to relieve the repression of Pax2 by TLX by interacting with TLX. Astrocyte migration into the retina precedes the formation of blood vessels (Fruttiger, 2007; Stone and Dreher, 1987). PAX2 (+) perinatal astrocytes closely associate with blood vessels (Chu et al., 2001). In addition, after completion of blood vessel formation, astrocytes lose their proangiogenic properties but continue maintain vascular homeostasis by promoting the formation and maintenance of blood-retina barrier (Abbott, 2002; Gardner et al., 1997; Janzer and Raff, 1987). We hypothesize that after hypoxia-induced blood vessel formation occurs, TLX repression of Pax2 is relieved by BMPs from the blood and then Pax2 continues its expression in astrocytes.

The BMPs have recently been implicated in the development and formation of glial scars following injury and in disease states throughout the central nervous system (Hall and Miller, 2004). The BMP-related SMADs interact with members of the ciliary neurotrophic factor(CNTF)/leukemia inhibitory factor (Stayner et al.) pathways, known as the signal transducers and transcription factors (STATs), via the FRAP-STAT pathway (Rajan and McKay, 1998; Rajan et al., 2003). The STATs and SMADs form a complex with the p300/CBP proteins to increase transcription of the type III intermediate filament protein GFAP (Yanagisawa et al., 2001). Together these two pathways potentiate astrocyte induction (Nakashima et al., 1999). BMP7 has been shown to be upregulated following spinal cord injury and to play a neuroprotective role in the central nervous system (Cox et al., 2004; Harvey et al., 2004; Setoguchi et al., 2001; Tsai et al., 2007). However, we hypothesize, that while BMPs may play a role initially in neuroprotection, they may also initiate reactive gliosis and/or maintain glial scars that prevent regeneration of central nervous system connections.

Future studies will be focused on determining the role of BMP7 and SHH in both astrocyte development as well as the in glial scar formation following injury and in disease states such as glaucoma, diabetic retinopathy, and age-related macular degeneration.

Materials and Methods

Materials used

Dulbecco’s Modified Eagle’s Medium (DMEM), cell dissociation solution (Sigma; St. Louis, MO), FBS, Penicillin/Streptomycin, Trypsin-EDTA, Phenol:Choroform, DTT, ProLong antifade mounting media with DAPI (Invitrogen; Carlsbad, CA), INFγ, mouse N-terminus SHH peptide, mouse BMP7 (R&D Systems; Minneapolis, MN), Hepes, L-Glutamine, NEAA (Cellgro; Herndon, VA), EC growth supplement, protease inhibitor cocktail, RNase free water, Heparin (Sigma; St. Louis, MO), PVDF membranes (Millipore; Bedford, MA), NaCl, Tris, NP-40, Sodium deoxycholate, EDTA, IGPEAL, Glycerol, SDS, L-Glutamine, Sodium Pyruvate, CellBind –treated dishes and flasks (Fisher Scientific; Hanover Park, IL), Roche buffer M, FuGene HD Transfection Reagent (Roche; Indianapolis, IN), Nuclear extraction kit, EMSA kit (Panomics; Fremont, CA), ECL Chemiluminescent substrate, Protein G plus beads (Pierce), Chemiluminescence films (Amersham Bio- sciences, UK), DC assay kit (Bio-Rad; Hercules, CA), Streptavidin agarose resin (Thermo Scientific; Rockford, IL), Cyclopamine (Calbiochem), SHH, BMP7, Follistatin (R&D systems, Minneapolis, MN), Lucifearse Assay System, TNT Quick Coupled Transcription/Translation system (Promega, Madison, WI), pCMV-TLX neo6 expression vector (Origene; Rockville, MD), mouse BMP7 and BMP7 function blocking serum (kind gift from Pamela J. Lein, Oregon Health and Science University, Portland, Oregon). Antibodies: Rabbit anti- PAX2 (Covance; 1μg/ml), rabbit anti-phospho PAX2 (Zymed; 2μg/ml), rabbit anti acetyl-Histone H3 (Millipore 17-615, 4μg/ml ChIP), rabbit anti-SMAD1 (Zymed; 1μg/ml), rabbit anti- phospho SMAD1 (Millipore; 1μg/ml for westerns; 10ug/ml for immunofluorescence), rabbit anti-phospho SMAD1 (Santa Cruz sc-12353; 4μg/ml for ChIP), rabbit anti-SMAD2 (Zymed; 2μg/ml), rabbit anti-TLX (Abcam ab30942; 0.4μg/ml for westerns, 4μg/ml for ChIP), mouse anti-TLX (abcam 10ug/ml for immunofluorescence), rabbit anti-GLI2 (Abcam; 1μg/ml for westerns, 10mg/ml for immunofluorescence), rabbit anti-GLI2 (Santa Cruz sc-20290; 4μg/ml ChIP); rabbit anti-GLI3 (Abcam; 10ug/ml); rabbit anti-GLI1 (Cell Signaling; 1:100), mouse anti-β-tubulin (Sigma; 1μg/ml), mouse anti-α tubulin (Sigma; 0.5μg/ml), human anti-BMP7 (R&D systems; 2μg/ml for westerns; 10ug/ml for immunofluorescence), rabbit anti-GFAP (Dako; 1:1000 for immunofluorescence; 2μg/ml for FACs); mouse anti-NG2 chondroitin sulfate proteoglycans (Fisher Scientific; 2μg/ml); ECL anti-rabbit HRP (1:5000; GE Healthcare, UK), anti-rabbit HRP for CO-IP (1:20000; e Bioscience, San Diego, CA), donkey anti-goat IgG conjugated to alexa-fluor 594 (Molecular Probes Invitrogen, 1:1000); goat anti-mouse IgG conjugated to alexa-fluor 594 (Molecular Probes Invitrogen, 1:1000), and goat anti-rabbit IgG conjugated to alexa-fluor 594 (Molecular Probes Invitrogen, 1:1000).

Experimental animals

All animal experimentation described in this study was conducted in accordance with accepted standards of humane animal care, the NIH guidelines for the care and use of experimental animals, and protocols approved by the institutional animal care and use committees for Indiana University School of Medicine and University of Wisconsin. Wild –type C3H mice and immorto-mice were housed under pathogen-free conditions in controlled light cycle and temperatures, and provided tap water and commercial mouse chow ad libidum. Mating pairs were monitored daily for vaginal plugs.

Cell Culture and Gene Regulation Assays

Retinal astrocytes were isolated from “immorto” mice back-crossed to C57BL/6J and grown as previously described (Scheef et al., 2005) and were grown on CellBind-treated dishes and flasks. For luciferase assays, a luciferase reporter gene containing 900 bp upstream of the start site of the mouse Pax2 promoter (including the TLX binding site) was constructed in the pGL4.14 vector (Promega). Astrocytes were cultured in 6×35mm dishes. 5×105 astrocytes were transiently co-transfected using Fugene HD transfection reagent with 300 ng of reporter plasmid and 200 ng of expression vector. Parallel control samples were co-transfected with equal amounts of empty vectors (n=6 for each condition). After 24 hours of transfection, cells were treated with BMP7, SHH and, BMP7 & SHH. 24 hours later, cells were lysed and luciferase activity was measured using Beckman Coulter LD400 plate reader/luminometer (Beckman Coulter, Fullerton, CA). Total protein concentration was determined using Bio-Rad DC assay and luciferase activities were normalized to protein concentration. Each experiment was repeated at least three times (n= 4+).

SDS-PAGE and Western Analysis

Astrocytes were seeded in 10 mm culture dishes at a concentration of 5×105 cells/ml. They were treated with BMP7 (200 ng/ml), SHH (200 ng/ml) either alone or in combination. After 12–16 hours of treatment, cells were lysed by adding lysis buffer (50mM Tris pH 7.4, 1% NP-40, 0.25% sodium deoxycholate, 150mM sodium chloride, 1mM EDTA, supplemented with 1 μg/ml protease inhibitors) for 30 min on ice. Cell lysates were centrifuged at 14000 rpm for 15 min at 4°C and supernatant was collected. Total protein concentration was determined using DC-Assay (Bio-Rad). Proteins were transferred to PVDF membranes overnight at 4°C and immunoblotted as described in (Nofziger et al., 2005). Densitometry of the blots was performed using ImageJ 1.34 s software. western data values for treated samples were compared with vehicle samples using one-tailed student’s t-test. Differences were considered significant when p ≤ 0.06. For total PAX2 protein densitometric analysis, the PAX2 dimers were totaled together because previous reports have shown no difference in the ability of the two alternatively spliced products to bind to DNA and activate expression (Phelps and Dressler, 1996)

Nuclear Extract Preparation

Nuclear extracts were prepared from wild type retinal astrocytes using a nuclear extraction kit according to the manufacturer’s instructions (Panomics, Fremont, CA). Cells were grown to 90% confluence and washed with PBS. Buffer A working reagent containing Buffer A (supplied with the kit), DTT, 10% IGEPAL and protease inhibitor cocktail was added to the culture flasks, kept on ice bucket and incubated for 10 minutes on rocking platform. Cells were removed using a sterile scraper and transferred to 1.5ml microfuge tube and centrifuged at 14,000 rpm for 3 min at 4°C. To the pellet, Buffer B working solution containing Buffer B (supplied with the kit), DTT and protease inhibitor cocktail was added. Tubes were vortexed and incubated horizontally on an ice bucket which was then transferred to a rocking platform for 2 hours. Samples were centrifuged and supernatants were transferred to a fresh microfuge tube. Nuclear extracts were stored at −80°C until analyzed.

Electrophoretic Mobility Shift Assay (EMSA) and In Vitro Translation

Mobility shift assays were performed using EMSA kits (Panomics) according to the manufacturer’s instructions with the following additions or changes. Probes used for EMSA analysis were as follows: Anti-sense probe (5′-3′): AGCTTTGTCTGACAAGTCATCCATCTAGCT; Sense probe (5′-3′): AGCTAGATGGATGACTTGTCAGACAAAGCT. 4 μg nuclear extract was incubated with 5x binding buffer, Poly d (I–C) for 5 min at room temperature before adding 1 μl of biotin labeled Pax2 probe (prepared by annealing sense and biotin labeled antisense oligos by boiling 100ng/μl of each oligo with 10x Roche buffer M for 5 minutes and kept at room temperature overnight). In vitro translations were performed as in Pfaeffle et al (2007).

Avidin Biotin Conjugated DNA Binding assay (ABCD)

The ABCD assay was performed as in Glass et al. (1987). Briefly, sense and biotin labeled antisense oligos were first annealed by boiling 100 pmole/μl of each oligo with 10x Roche buffer M for 5 minutes and kept at room temperature overnight. 10μl of the total sample was used for each reaction. 500μg total proteins from astrocytes were mixed with 100μl of protease inhibitor cocktail, 10μl of probe and DNA pull down buffer pH 7.5 (containing 25mM HEPES, 15mM NaCl, 0.5mM DTT, 0.5% IGEPAL, 0.1mM EDTA and 10% glycerol) was added to make total volume to 1ml in 1.5ml microfuge tubes. The solution was incubated on rocker shaker at 4°C overnight. 40 μl of 50% streptavidin-agarose resin was added and was incubated for 6hours at 4°C on rocker shaker. Tubes were centrifuged at 12000 rpm for 2 minutes and the pellet was then washed five times with DNA pull down buffer. 1x SDS-PAGE loading buffer was added to the pellet and boiled for 5 minutes. The samples were stored at −20°C until further analyzed by western blotting.


Co-immunoprecipitation was performed using the universal magnetic Co-IP kit from Active Motif with minor modifications to the manufacturer’s instructions. Briefly, nuclei were isolated form cells using enzymatic digestion. Three hundred micrograms of nuclear protein extract was incubated with 5μg of antibody specific for b-tubulin, pSMAD1, or GLI2 overnight at 4°C on a rotator. Following incubation with antibody, lysates were incubated with magnetic protein G beads for 3–4 hours at 4°C on a rotator. After washing beads 6 times in ice cold wash buffer, beads were suspended in 2X reducing loading buffer (130mm Tris pH6.8, 4% SDS, 0.02% bromophenol blue, 20% glycerol, 100mmDTT), vortexed to mix and stored at −80°C until western blotting was performed. Prior to loading, beads were boiled for 10 minutes, spun, and the elute was subjected to SDS-PAGE. Following transfer, blots were immunoblotted as in western analysis section.


Prior to immunohistochemistry, antigen retrieval was performed by placing sections in 10mM sodium citrate pH 6.0 at 65°C and heating in a microwave on high for 20 minutes. Sections were then allowed to come to room temperature prior to proceeding with the immunofluorescence protocol. Immunofluorescence on sections was performed as described in Wilson et al.(Wilson et al., 2007), with the exception that slides were mounted using ProLong Gold with DAPI. Immunocytochemistry was performed as described in Belecky-Adams et al (Belecky-Adams et al., 1996). Antibodies were used at the following dilutions: Pax2,1:200; TLX, 10μg/ml; GLI1, 1:100, Gli2, 10μg/ml, Gli3, 10μg/ml, SHH, 10μg/ml, BMP7, 10μg/ml, pSMAD1, 10μg/ml.

FACScan analysis

FACS analysis was performed as described in Scheef et al., 2005. Briefly, cells were rinsed with phosphate buffered saline (PBS) with 0.4% EDTA. Cells were incubated with 2ml of dissociation solution and rinsed from the plates with 5ml of DMEM containing 10% fetal bovine serum. Cells were spun, washed once one with tris-buffered saline (TBS), resuspended in 0.5ml of TBS with 1% goat serum, incubated for 20 minutes at 4°C, incubated with 0.5ml of rabbit anti GFAP (2μg/ml) or rabbit anti-NG2 (2μg/ml) for 30 minutes at 4°C. FACS of labeled cells was compared to samples incubated with control IgG in place of primary antibodies.

Chromatin Immunoprecipitation and Quantitative PCR

Cells were rinsed with cold PBS, removed from dish with cell scraper, and crosslinking carried out in a final concentration of 1% formaldehyde for 15 minutes on ice. Cells were subsequently incubated with glycine to stop the crosslinking and were washed twice by spinning at 4500RPM for 7 minutes and rinsing the pellet in PBS containing 1X protease inhibitor mix (Roche) and 0.1mM PMSF at 4°C. Following rinses, pellets were resuspended in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2mM EDTA, 16.7 mM Tris HCl pH8.1, 167mM NaCl with 1X protease inhibitor mix and 0.1mM PMSF) and stored at −80°C. Upon thaw, cells were spun at 6000 RPM for 5 minutes and pellets resuspended in 200mL of SDS lysis buffer (1% SDS, 10mM EDTA, 50mM Tris-HCl, pH8.1 with 1X protease inhibitor mix and 0.1mM PMSF) and incubated on ice for 10 minutes. The mixture was sonicated for 3–4 sets of 10-second pulses using a Cole Palmer, High Intensity Ultrasonic Processor/Sonicator, 50-watt model (with 2mm tip) set to 30% of maximum power. Debris was removed by spinning at 13,000 RPM for 10 minutes at 4°C, supernatant transferred to a fresh tube and 1.8ml of ChIP dilution buffer added. Two hundred microliters of sample was saved as an input control. For immunoprecipitation, antibody was added at a final concentration of 2μg/ml and rotated overnight at 4°C. After overnight incubation, samples were incubated with 60ml of protein A agarose, rotated for 1 hour at 4°C, and centrifuged at 1000RPM for 1 min at 4°C. Beads were washed at room temperature for 5 minutes on a rocker once in 1.0ml of low salt wash buffer (0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris HCl, pH8.1, 150mM NaCl), four times in 1.0ml high salt wash buffer ( 0.1% SDS, 1% Triton X-100, 2mM EDTA, 20mM Tris HCl, pH8.1, 500mM NaCl), once in LiCl wash buffer (0.25M LiCl, 1% NP-40, 1% deoxycholate, 1mM EDTA, 10mM Tris HCl, pH8.1) and twice in TE. Samples were incubated in 250ul of freshly prepared elution buffer (1% SDS,0.1M NaHCO3), incubated at room temp. for 15 minutes, beads spun out and supernatant transferred to a fresh tube. Elution was repeated with beads once more and eluates were combined. The sample and input control were incubated with 20ul of 5.0M NaCl for 4 hours at 65°C, then incubated in 20μg of proteinase K. Samples were treated with phenol chloroform and ethanol precipitated. Pellets were resuspended in 50μl of TE and stored at −20°C until qPCR performed. qPCR was performed using Taqman PCR Reagent kit and Applied Biosystems -7500 Real time PCR system according to the manufacturer’s instructions. Primer Express 3.0 by Applied Biosystems was used to design Primers and Probes. Oligonucleotide Probe had 5′ FAM and 3′TAMRA reporter and quencher dyes respectively. GAPDH (used as endogenous control) and Pax2 primers used were: GAPDH Forward Primer; 5′-GAAGGTGAAGGTCGGAGTC-3′, GAPDH Reverse Primer; 5′-GAAGATGGTGATGGGATTTC-3′, GAPDH Probe FAM-CAAGCTTCCCGTTCTCAGCC-TAMRA, PAX2 Forward Primer; 5′CGGCGCTGGCGAATC3′, PAX2 Reverse Primer; 5′CGGGAGATGGATGACTTGTCA3′ Pax2 Probe; FAM-AGAGTGGTGGAATCTA-TAMRA. Each reaction was performed in triplicates and each experiment was performed at least 3 times. The data was expressed as Log10 fold increase (mean±SD) above control.


The authors would like to thank Pam Lein (Oregon Health and Science University, Portland, OR) for mouse BMP7 and function-blocking BMP7 antibody and Greg Dressler (University of Michigan, Ann Arbor, MI) for sharing 9.1 Kbp of 5′ untranslated mouse Pax2 sequences. The authors also gratefully acknowledge Rachel Mullen for help with immunostaining. TBA is supported by the March of Dimes, American Cancer Society, and American Health Assistance Foundation, while SJR is supported by R01 HD42024 from the NIH/NICHD. NS is supported by EY016695 and Retina Research Foundation.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Abbott NJ. Astrocyte-endothelial interactions and blood-brain barrier permeability. J Anat. 2002;200:629–38. [PubMed]
  • Adler R, Belecky-Adams TL. The role of bone morphogenetic proteins in the differentiation of the ventral optic cup. Development. 2002;129:3161–71. [PubMed]
  • Arkell R, Beddington RS. BMP-7 influences pattern and growth of the developing hindbrain of mouse embryos. Development. 1997;124:1–12. [PubMed]
  • Bai CB, Stephen D, Joyner AL. All mouse ventral spinal cord patterning by hedgehog is Gli dependent and involves an activator function of Gli3. Dev Cell. 2004;6:103–15. [PubMed]
  • Belecky-Adams T, Adler R. Developmental expression patterns of bone morphogenetic proteins, receptors, and binding proteins in the chick retina. J Comp Neurol. 2001;430:562–72. [PubMed]
  • Belecky-Adams T, Cook B, Adler R. Correlations between terminal mitosis and differentiated fate of retinal precursor cells in vivo and in vitro: analysis with the “window-labeling” technique. Developmental Biology. 1996;178:304–15. [PubMed]
  • Belecky-Adams TL, Adler R. Developmental expression patterns of bone morphogenetic proteins, receptors and binding proteins in the chick retina. J Comp Neurol. 2000 in press. [PubMed]
  • Bhatia M, Bonnet D, Wu D, Murdoch B, Wrana J, Gallacher L, Dick JE. Bone morphogenetic proteins regulate the developmental program of human hematopoietic stem cells. J Exp Med. 1999;189:1139–48. [PMC free article] [PubMed]
  • Cai Y, Lechner MS, Nihalani D, Prindle MJ, Holzman LB, Dressler GR. Phosphorylation of Pax2 by the c-Jun N-terminal kinase and enhanced Pax2-dependent transcription activation. J Biol Chem. 2002;277:1217–22. [PubMed]
  • Chan-Ling T, Stone J. Factors determining the migration of astrocytes into the developing retina: migration does not depend on intact axons or patent vessels. J Comp Neurol. 1991;303:375–86. [PubMed]
  • Chari NS, McDonnell TJ. The sonic hedgehog signaling network in development and neoplasia. Adv Anat Pathol. 2007;14:344–52. [PubMed]
  • Chen D, Zhao M, Harris SE, Mi Z. Signal transduction and biological functions of bone morphogenetic proteins. Front Biosci. 2004;9:349–58. [PubMed]
  • Chu Y, Hughes S, Chan-Ling T. Differentiation and migration of astrocyte precursor cells and astrocytes in human fetal retina: relevance to optic nerve coloboma. FASEB J. 2001;15:2013–5. [PubMed]
  • Cox S, Harvey BK, Sanchez JF, Wang JY, Wang Y. Mediation of BMP7 neuroprotection by MAPK and PKC IN rat primary cortical cultures. Brain Res. 2004;1010:55–61. [PubMed]
  • Dakubo GD, Beug ST, Mazerolle CJ, Thurig S, Wang Y, Wallace VA. Control of glial precursor cell development in the mouse optic nerve by sonic hedgehog from retinal ganglion cells. Brain Res. 2008;1228:27–42. [PubMed]
  • Dakubo GD, Wang YP, Mazerolle C, Campsall K, McMahon AP, Wallace VA. Retinal ganglion cell-derived sonic hedgehog signaling is required for optic disc and stalk neuroepithelial cell development. Development. 2003;130:2967–80. [PubMed]
  • Dale JK, Vesque C, Lints TJ, Sampath TK, Furley A, Dodd J, Placzek M. Cooperation of BMP7 and SHH in the induction of forebrain ventral midline cells by prechordal mesoderm. Cell. 1997;90:257–69. [PubMed]
  • Dale L, Jones CM. BMP signalling in early Xenopus development. Bioessays. 1999;21:751–60. [PubMed]
  • Derynck R, Zhang YE. Smad-dependent and Smad-independent pathways in TGF-beta family signalling. Nature. 2003;425:577–84. [PubMed]
  • Dressler GR, Deutsch U, Chowdhury K, Nornes HO, Gruss P. Pax2, a new murine paired-box-containing gene and its expression in the developing excretory system. Development. 1990;109:787–95. [PubMed]
  • Dressler GR, Douglass EC. Pax-2 is a DNA-binding protein expressed in embryonic kidney and Wilms tumor. Proc Natl Acad Sci U S A. 1992;89:1179–83. [PubMed]
  • Dudley AT, Lyons KM, Robertson EJ. A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye. Genes Dev. 1995;9:2795–807. [PubMed]
  • Dudley AT, Robertson EJ. Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 deficient embryos. Dev Dyn. 1997;208:349–62. [PubMed]
  • Eccles MR, Wallis LJ, Fidler AE, Spurr NK, Goodfellow PJ, Reeve AE. Expression of the PAX2 gene in human fetal kidney and Wilms’ tumor. Cell Growth Differ. 1992;3:279–89. [PubMed]
  • Ekker SC, Ungar AR, Greenstein P, von Kessler DP, Porter JA, Moon RT, Beachy PA. Patterning activities of vertebrate hedgehog proteins in the developing eye and brain. Curr Biol. 1995;5:944–55. [PubMed]
  • Fan X, Valdimarsdottir G, Larsson J, Brun A, Magnusson M, Jacobsen SE, ten Dijke P, Karlsson S. Transient disruption of autocrine TGF-beta signaling leads to enhanced survival and proliferation potential in single primitive human hemopoietic progenitor cells. J Immunol. 2002;168:755–62. [PubMed]
  • Fruttiger M. Development of the retinal vasculature. Angiogenesis. 2007;10:77–88. [PubMed]
  • Fuhrmann S, Levine EM, Reh TA. Extraocular mesenchyme patterns the optic vesicle during early eye development in the embryonic chick. Development - Supplement. 2000;127:4599–609. [PubMed]
  • Furimsky M, Wallace VA. Complementary Gli activity mediates early patterning of the mouse visual system. Dev Dyn. 2006;235:594–605. [PubMed]
  • Furuta Y, Piston DW, Hogan BL. Bone morphogenetic proteins (BMPs) as regulators of dorsal forebrain development. Development. 1997;124:2203–12. [PubMed]
  • Gardner TW, Lieth E, Khin SA, Barber AJ, Bonsall DJ, Lesher T, Rice K, Brennan WA., Jr Astrocytes increase barrier properties and ZO-1 expression in retinal vascular endothelial cells. Invest Ophthalmol Vis Sci. 1997;38:2423–7. [PubMed]
  • Hall AK, Miller RH. Emerging roles for bone morphogenetic proteins in central nervous system glial biology. J Neurosci Res. 2004;76:1–8. [PubMed]
  • Harvey BK, Mark A, Chou J, Chen GJ, Hoffer BJ, Wang Y. Neurotrophic effects of bone morphogenetic protein-7 in a rat model of Parkinson’s disease. Brain Res. 2004;1022:88–95. [PubMed]
  • Herpin A, Cunningham C. Cross-talk between the bone morphogenetic protein pathway and other major signaling pathways results in tightly regulated cell-specific outcomes. FEBS J. 2007;274:2977–85. [PubMed]
  • Hocking JC, McFarlane S. Expression of Bmp ligands and receptors in the developing Xenopus retina. Int J Dev Biol. 2007;51:161–5. [PubMed]
  • Hooper JE, Scott MP. Communicating with Hedgehogs. Nat Rev Mol Cell Biol. 2005;6:306–17. [PubMed]
  • Huangfu D, Anderson KV. Signaling from Smo to Ci/Gli: conservation and divergence of Hedgehog pathways from Drosophila to vertebrates. Development. 2006;133:3–14. [PubMed]
  • Hung FC, Zhao S, Chen Q, Overbeek PA. Retinal ablation and altered lens differentiation induced by ocular overexpression of BMP7. Vision Res. 2002;42:427–38. [PubMed]
  • Janzer RC, Raff MC. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature. 1987;325:253–7. [PubMed]
  • Jat PS, Noble MD, Ataliotis P, Tanaka Y, Yannoutsos N, Larsen L, Kioussis D. Direct derivation of conditionally immortal cell lines from an H-2Kb-tsA58 transgenic mouse. Proc Natl Acad Sci U S A. 1991;88:5096–100. [PubMed]
  • Jena N, Martin-Seisdedos C, McCue P, Croce CM. BMP7 null mutation in mice: developmental defects in skeleton, kidney, and eye. Exp Cell Res. 1997;230:28–37. [PubMed]
  • Kawai S, Sugiura T. Characterization of human bone morphogenetic protein (BMP)-4 and -7 gene promoters: activation of BMP promoters by Gli, a sonic hedgehog mediator. Bone. 2001;29:54–61. [PubMed]
  • Kim J, Johnson K, Chen HJ, Carroll S, Laughon A. Drosophila Mad binds to DNA and directly mediates activation of vestigial by Decapentaplegic. Nature. 1997;388:304–8. [PubMed]
  • Koshiba-Takeuchi K, Takeuchi JK, Matsumoto K, Momose T, Uno K, Hoepker V, Ogura K, Takahashi N, Nakamura H, Yasuda K, Ogura T. Tbx5 and the retinotectum projection. Science. 2000;287:134–7. [PubMed]
  • Kusanagi K, Inoue H, Ishidou Y, Mishima HK, Kawabata M, Miyazono K. Characterization of a bone morphogenetic protein-responsive Smad-binding element. Mol Biol Cell. 2000;11:555–65. [PMC free article] [PubMed]
  • Lee KS, Kim HJ, Li QL, Chi XZ, Ueta C, Komori T, Wozney JM, Kim EG, Choi JY, Ryoo HM, Bae SC. Runx2 is a common target of transforming growth factor beta1 and bone morphogenetic protein 2, and cooperation between Runx2 and Smad5 induces osteoblast-specific gene expression in the pluripotent mesenchymal precursor cell line C2C12. Mol Cell Biol. 2000;20:8783–92. [PMC free article] [PubMed]
  • Liem KF, Jr, Tremml G, Roelink H, Jessell TM. Dorsal differentiation of neural plate cells induced by BMP-mediated signals from epidermal ectoderm. Cell. 1995;82:969–79. [PubMed]
  • Macdonald R, Barth KA, Xu Q, Holder N, Mikkola I, Wilson SW. Midline signalling is required for Pax gene regulation and patterning of the eyes. Development. 1995;121:3267–78. [PubMed]
  • Miyawaki T, Uemura A, Dezawa M, Yu RT, Ide C, Nishikawa S, Honda Y, Tanabe Y, Tanabe T. Tlx, an orphan nuclear receptor, regulates cell numbers and astrocyte development in the developing retina. J Neurosci. 2004;24:8124–34. [PubMed]
  • Miyazono K. Signal transduction by bone morphogenetic protein receptors: functional roles of Smad proteins. Bone. 1999;25:91–3. [PubMed]
  • Miyazono K. Positive and negative regulation of TGF-beta signaling. J Cell Sci. 2000a;113(Pt 7):1101–9. [PubMed]
  • Miyazono K. TGF-beta signaling by Smad proteins. Cytokine Growth Factor Rev. 2000b;11:15–22. [PubMed]
  • Miyazono K, Maeda S, Imamura T. BMP receptor signaling: transcriptional targets, regulation of signals, and signaling cross-talk. Cytokine Growth Factor Rev. 2005;16:251–63. [PubMed]
  • Morcillo J, Martinez-Morales JR, Trousse F, Fermin Y, Sowden JC, Bovolenta P. Proper patterning of the optic fissure requires the sequential activity of BMP7 and SHH. Development. 2006;133:3179–90. [PubMed]
  • Motoyama J, Milenkovic L, Iwama M, Shikata Y, Scott MP, Hui CC. Differential requirement for Gli2 and Gli3 in ventral neural cell fate specification. Dev Biol. 2003;259:150–61. [PubMed]
  • Muller F, Rohrer H, Vogel-Hopker A. Bone morphogenetic proteins specify the retinal pigment epithelium in the chick embryo. Development. 2007;134:3483–93. [PubMed]
  • Nagase T, Nagase M, Yoshimura K, Fujita T, Koshima I. Angiogenesis within the developing mouse neural tube is dependent on sonic hedgehog signaling: possible roles of motor neurons. Genes Cells. 2005;10:595–604. [PubMed]
  • Nakamura H. Regionalization of the optic tectum: combinations of gene expression that define the tectum. Trends Neurosci. 2001;24:32–9. [PubMed]
  • Nakashima K, Yanagisawa M, Arakawa H, Kimura N, Hisatsune T, Kawabata M, Miyazono K, Taga T. Synergistic signaling in fetal brain by STAT3-Smad1 complex bridged by p300. Science. 1999;284:479–82. [PubMed]
  • Nofziger C, Chen L, Shane MA, Smith CD, Brown KK, Blazer-Yost BL. PPARgamma agonists do not directly enhance basal or insulin-stimulated Na(+) transport via the epithelial Na(+) channel. Pflugers Arch. 2005;451:445–53. [PubMed]
  • Nornes HO, Dressler GR, Knapik EW, Deutsch U, Gruss P. Spatially and temporally restricted expression of Pax2 during murine neurogenesis. Development. 1990;109:797–809. [PubMed]
  • Ohkubo Y, Chiang C, Rubenstein JL. Coordinate regulation and synergistic actions of BMP4, SHH and FGF8 in the rostral prosencephalon regulate morphogenesis of the telencephalic and optic vesicles. Neuroscience. 2002;111:1–17. [PubMed]
  • Parkin CA, Ingham PW. The adventures of Sonic Hedgehog in development and repair. I. Hedgehog signaling in gastrointestinal development and disease. Am J Physiol Gastrointest Liver Physiol. 2008;294:G363–7. [PubMed]
  • Phelps DE, Dressler GR. Identification of novel Pax-2 binding sites by chromatin precipitation. J Biol Chem. 1996;271:7978–85. [PubMed]
  • Placzek M, Briscoe J. The floor plate: multiple cells, multiple signals. Nat Rev Neurosci. 2005;6:230–40. [PubMed]
  • Rajan P, McKay RD. Multiple routes to astrocytic differentiation in the CNS. J Neurosci. 1998;18:3620–9. [PubMed]
  • Rajan P, Panchision DM, Newell LF, McKay RD. BMPs signal alternately through a SMAD or FRAP-STAT pathway to regulate fate choice in CNS stem cells. J Cell Biol. 2003;161:911–21. [PMC free article] [PubMed]
  • Roy K, Thiels E, Monaghan AP. Loss of the tailless gene affects forebrain development and emotional behavior. Physiol Behav. 2002;77:595–600. [PMC free article] [PubMed]
  • Sakuta H, Suzuki R, Takahashi H, Kato A, Shintani T, Iemura S, Yamamoto TS, Ueno N, Noda M. Ventroptin: a BMP-4 antagonist expressed in a double-gradient pattern in the retina. Science. 2001;293:111–5. [PubMed]
  • Scheef E, Wang S, Sorenson CM, Sheibani N. Isolation and characterization of murine retinal astrocytes. Mol Vis. 2005;11:613–24. [PubMed]
  • Schimmenti LA, de la Cruz J, Lewis RA, Karkera JD, Manligas GS, Roessler E, Muenke M. Novel mutation in sonic hedgehog in non-syndromic colobomatous microphthalmia. Am J Med Genet A. 2003;116A:215–21. [PubMed]
  • Sehgal R, Karcavich R, Carlson S, Belecky-Adams TL. Ectopic Pax2 expression in chick ventral optic cup phenocopies loss of Pax2 expression. Dev Biol. 2008;319:23–33. [PMC free article] [PubMed]
  • Setoguchi T, Yone K, Matsuoka E, Takenouchi H, Nakashima K, Sakou T, Komiya S, Izumo S. Traumatic injury-induced BMP7 expression in the adult rat spinal cord. Brain Res. 2001;921:219–25. [PubMed]
  • Shim S, Bae N, Han JK. Bone morphogenetic protein-4-induced activation of Xretpos is mediated by Smads and Olf-1/EBF associated zinc finger (OAZ) Nucleic Acids Res. 2002;30:3107–17. [PMC free article] [PubMed]
  • Stallcup WB, Beasley L. Bipotential glial precursor cells of the optic nerve express the NG2 proteoglycan. J Neurosci. 1987;7:2737–44. [PubMed]
  • Stayner CK, Cunliffe HE, Ward TA, Eccles MR. Cloning and characterization of the human PAX2 promoter. J Biol Chem. 1998;273:25472–9. [PubMed]
  • Stone J, Dreher Z. Relationship between astrocytes, ganglion cells and vasculature of the retina. J Comp Neurol. 1987;255:35–49. [PubMed]
  • Suzuki T, Takeuchi J, Koshiba-Takeuchi K, Ogura T. Tbx Genes Specify Posterior Digit Identity through Shh and BMP Signaling. Dev Cell. 2004;6:43–53. [PubMed]
  • Trousse F, Esteve P, Bovolenta P. Bmp4 mediates apoptotic cell death in the developing chick eye. J Neurosci. 2001;21:1292–301. [PubMed]
  • Tsai MJ, Weng CF, Shyue SK, Liou DY, Chen CH, Chiu CW, Yang TH, Pan HA, Liao RI, Kuo HS, Huang MC, Huang WC, Hoffer BJ, Cheng H. Dual effect of adenovirus-mediated transfer of BMP7 in mixed neuron-glial cultures: neuroprotection and cellular differentiation. J Neurosci Res. 2007;85:2950–9. [PubMed]
  • Uemura A, Kusuhara S, Wiegand SJ, Yu RT, Nishikawa S. Tlx acts as a proangiogenic switch by regulating extracellular assembly of fibronectin matrices in retinal astrocytes. J Clin Invest. 2006;116:369–77. [PubMed]
  • Ulloa F, Briscoe J. Morphogens and the control of cell proliferation and patterning in the spinal cord. Cell Cycle. 2007;6:2640–9. [PubMed]
  • Vogel-Hopker A, Momose T, Rohrer H, Yasuda K, Ishihara L, Rapaport DH. Multiple functions of fibroblast growth factor-8 (FGF-8) in chick eye development. Mech Dev. 2000;94:25–36. [PubMed]
  • Wallace VA, Raff MC. A role for Sonic hedgehog in axon-to-astrocyte signalling in the rodent optic nerve. Development. 1999;126:2901–9. [PubMed]
  • Wang Y, Dakubo GD, Thurig S, Mazerolle CJ, Wallace VA. Retinal ganglion cell-derived sonic hedgehog locally controls proliferation and the timing of RGC development in the embryonic mouse retina. Development. 2005;132:5103–13. [PubMed]
  • Wawersik S, Purcell P, Rauchman M, Dudley AT, Robertson EJ, Maas R. BMP7 acts in murine lens placode development. Dev Biol. 1999;207:176–88. [PubMed]
  • Wilson JM, Sato K, Chernoff EA, Belecky-Adams TL. Expression patterns of chick Musashi-1 in the developing nervous system. Gene Expr Patterns. 2007;7:817–25. [PubMed]
  • Wordinger RJ, Agarwal R, Talati M, Fuller J, Lambert W, Clark AF. Expression of bone morphogenetic proteins (BMP), BMP receptors, and BMP associated proteins in human trabecular meshwork and optic nerve head cells and tissues. Mol Vis. 2002;8:241–50. [PubMed]
  • Yanagisawa M, Nakashima K, Takizawa T, Ochiai W, Arakawa H, Taga T. Signaling crosstalk underlying synergistic induction of astrocyte differentiation by BMPs and IL-6 family of cytokines. FEBS Lett. 2001;489:139–43. [PubMed]
  • Yokoyama A, Sakamoto A, Kameda K, Imai Y, Tanaka J. NG2 proteoglycan-expressing microglia as multipotent neural progenitors in normal and pathologic brains. Glia. 2006;53:754–68. [PubMed]
  • Yu RT, Chiang MY, Tanabe T, Kobayashi M, Yasuda K, Evans RM, Umesono K. The orphan nuclear receptor Tlx regulates Pax2 and is essential for vision. Proc Natl Acad Sci U S A. 2000;97:2621–5. [PubMed]
  • Yu RT, McKeown M, Evans RM, Umesono K. Relationship between Drosophila gap gene tailless and a vertebrate nuclear receptor Tlx. Nature. 1994;370:375–9. [PubMed]
  • Yuasa T, Kataoka H, Kinto N, Iwamoto M, Enomoto-Iwamoto M, Iemura S, Ueno N, Shibata Y, Kurosawa H, Yamaguchi A. Sonic hedgehog is involved in osteoblast differentiation by cooperating with BMP-2. J Cell Physiol. 2002;193:225–32. [PubMed]
  • Zhang Y, Derynck R. Transcriptional regulation of the transforming growth factor-beta -inducible mouse germ line Ig alpha constant region gene by functional cooperation of Smad, CREB, and AML family members. J Biol Chem. 2000;275:16979–85. [PubMed]
  • Zhang Y, Stone J. Role of astrocytes in the control of developing retinal vessels. Invest Ophthalmol Vis Sci. 1997;38:1653–66. [PubMed]
  • Zhang Y, Zhang Z, Zhao X, Yu X, Hu Y, Geronimo B, Fromm SH, Chen YP. A new function of BMP4: dual role for BMP4 in regulation of Sonic hedgehog expression in the mouse tooth germ. Development. 2000;127:1431–43. [PubMed]
  • Zhao S, Chen Q, Hung FC, Overbeek PA. BMP signaling is required for development of the ciliary body. Development. 2002;129:4435–42. [PubMed]
  • Zhao S, Hung FC, Colvin JS, White A, Dai W, Lovicu FJ, Ornitz DM, Overbeek PA. Patterning the optic neuroepithelium by FGF signaling and Ras activation. Development. 2001;128:5051–60. [PubMed]