- Email: [email protected]
glypican-3 Controls Cellular Responses to Bmp4 in Limb Patterning and Skeletal Development
Developmental Biology 225, 179 –187 (2000) doi:10.1006/dbio.2000.9831, available online at http://www.idealibrary.com on
glypican-3 Controls Cellular Responses to Bmp4 in Limb Patterning and Skeletal Development Stephenie Paine-Saunders,* Beth L. Viviano,* Joel Zupicich,† William C. Skarnes,† and Scott Saunders* ,‡ ,1 *Division of Newborn Medicine, Department of Pediatrics, Washington University School of Medicine and St. Louis Children’s Hospital, St. Louis, Missouri 63110; †Department of Molecular and Cell Biology, University of California at Berkeley, Berkeley, California 94720; and ‡Department of Molecular Biology and Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110
Glypicans represent a family of six cell surface heparan sulfate proteoglycans in vertebrates. Although no specific in vivo functions have thus far been described for these proteoglycans, spontaneous mutations in the human and induced deletions in the mouse glypican-3 (Gpc3) gene result in severe malformations and both pre- and postnatal overgrowth, known clinically as the Simpson–Golabi–Behmel syndrome (SGBS). Mice carrying mutant alleles of Gpc3 created by either targeted gene disruption or gene trapping display a wide range of phenotypes associated with SGBS including renal cystic dysplasia, ventral wall defects, and skeletal abnormalities that are consistent with the pattern of Gpc3 expression in the mouse embryo. Previous studies in Drosophila have implicated glypicans in the signaling of decapentaplegic, a BMP homolog. Our experiments with mice show a significant relationship between vertebrate BMP signaling and glypican function; GPC3-deficient animals were mated with mice haploinsufficient for bone morphogenetic protein-4 (Bmp4) and their offspring displayed a high penetrance of postaxial polydactyly and rib malformations not observed in either parent strain. This previously unknown link between glypican-3 and BMP4 function provides evidence of a role for glypicans in vertebrate limb patterning and skeletal development and suggests a mechanism for the skeletal defects seen in SGBS. © 2000 Academic Press
Key Words: heparan sulfate; glypican; proteoglycans; bone morphogenetic protein; limb; skeleton; polydactyly; Simpson– Golabi–Behmel syndrome.
INTRODUCTION The essential requirement of heparan sulfate glycosaminoglycans for normal development has been emphasized only recently by the discovery that mutations in heparan sulfate biosynthesis cause significant developmental defects in Drosophila (Bellaiche et al., 1998; Binari et al., 1997; Hacker et al., 1997; Haerry et al., 1997), in mice (Bullock et al., 1998), and in humans (Lind et al., 1998; McCormick et al., 1998). This has stimulated interest in understanding the potential functions of specific protein gene products that bear these posttranslational modifica1
To whom correspondence should be addressed at the Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8208, St. Louis, MO 63110. Fax: (314) 286-2893. E-mail: [email protected]. 0012-1606/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
tions. Although heparan sulfate is found ubiquitously on cell surfaces, these carbohydrates are carried predominantly by the protein products of only two gene families, syndecans (Saunders et al., 1989) and glypicans (David et al., 1990). All members of the glypican family share a common mechanism of attachment to the cell membrane through glycosylphosphatidylinositol linkages, and the characterization of amino acid sequence homologies has suggested the existence of subfamilies of structurally related vertebrate glypicans (Paine-Saunders et al., 1999). Whether these individual glypicans have distinct or a related function is unknown; however, human mutations in a single glypican gene, glypican-3 (Gpc3), result in an overgrowth phenotype known as Simpson–Golabi–Behmel syndrome (SGBS). This X-linked condition is characterized by both pre- and postnatal overgrowth, a distinct facial appearance, macroglos-
179
180
Paine-Saunders et al.
sia, and multiple anomalies, including cleft palate, congenital heart and renal defects, imperforate vagina, umbilical hernias, supernumerary nipples, vertebral and rib defects, and postaxial polydactyly (Behmel et al., 1984; Garganta and Bodurtha, 1992; Hughes-Benzie et al., 1992; Pilia et al., 1996). Nothing is known about the mechanism by which these in vivo mutations of Gpc3 lead to the described phenotypic features in humans. Nonetheless, as described in this study and elsewhere (Cano-Gauci et al., 1999), GPC3-deficient mice recapitulate the SGBS phenotype and therefore provide an excellent in vivo model system for evaluating the functions of vertebrate glypicans. Although the mechanism of action of vertebrate glypicans is unknown, the function of the Drosophila glypican dally has been evaluated genetically. These studies have revealed a role for dally in both wingless (wg; the Drosophila ortholog of vertebrate Wnt) signaling during early embryonic patterning (Lin and Perrimon, 1999; Tsuda et al., 1999) and decapentaplegic (dpp; the Drosophila ortholog of vertebrate bone morphogenetic proteins) signaling during postembryonic morphogenesis (Jackson et al., 1997). The mechanism of Dally function has not been experimentally established; however, these findings are consistent with the proposed function for heparan sulfate proteoglycans as coreceptors for a number of growth factors in vitro. Moreover, it is clear from the above studies that glypican functions are highly context and tissue dependent in Drosophila and are likely to be similarly context dependent in vertebrates. Therefore it is not obvious which signaling pathways might be impacted by the loss of function of a specific vertebrate glypican in any particular tissue during development. We have therefore chosen to focus on understanding the function of GPC3 in the pathogenesis of a single phenotypic trait in vertebrates, polydactyly. Although commonly seen in human patients with SGBS, this phenotype is never seen in GPC3-deficient mice. We therefore speculated that these animals may represent a sensitized strain for glypican function in limb development and that genetic interaction studies might point toward the signaling pathway(s) mediated by GPC3 in this tissue. Indeed, as has been suggested by others, the identification of genetic modifiers provides an important pathway to understanding the cellular and genetic interactions underlying embryonic development and the pathogenesis of human malformations (Dunn et al., 1997; Erickson, 1996; Winter, 1996). We have found that the distribution of Gpc3 expression overlaps with the patterns of a number of gene products implicated in limb patterning and whose biological activities are known to be dependent in part on heparan sulfate proteoglycans. These include fibroblast growth factors (FGFs) (Rapraeger et al., 1991; Yayon et al., 1991), Wnts (Lin and Perrimon, 1999; Reichsman et al., 1996; Tsuda et al., 1999), bone morphogenetic proteins (BMPs) (Jackson et al., 1997; Ruppert et al., 1996), and members of the vertebrate hedgehog family (The et al., 1999). In particular, Bmp4 expression in the developing limb is highly coincident with
Gpc3 expression in the progress zone, in the anterior and posterior necrotic zones, and in the interdigital web where it plays an active role in mediating critical apoptotic events in the development of these structures. Furthermore, Bmp4 haploinsufficient mice show a phenotype implicating a role for BMPs in anterior–posterior patterning (Dunn et al., 1997). Therefore we decided to test genetically whether GPC3 might be important in BMP4 signaling in the developing limb and have found a strong genetic interaction between mutants of these two genes that result in a high penetrance of both postaxial polydactyly and rib malformations. This study provides the first in vivo evidence of a role for glypicans in vertebrate development and establishes a previously unknown link between GPC3 and BMP4 function, which suggests a mechanism for the skeletal defects seen in SGBS.
MATERIALS AND METHODS Generation of mutant mice. A 7.5-kb genomic fragment containing the first exon of Gpc3, which includes the signal peptide, was isolated from a strain 129SV genomic library in the Lambda FIX II vector (Stratagene) using a 5⬘-specific probe from human Gpc3 cDNA. The DNA targeting vector was constructed by replacing the first exon of Gpc3 (from approximately ⫺50 bp to just downstream of the first splice site) with the AscI fragment containing the PGK-neo gene from pKO SelectNeo V800 (Lexicon). This resulted in a vector flanked with 5.5 kb of homologous sequence upstream and 2.0 kb of homologous sequence downstream. Additionally, an RsrII fragment containing the RNA Pol II diphtheria toxin gene from pKO SelectDT V840 (Lexicon) was inserted at the 3⬘ end to facilitate negative selection. The targeting vector was linearized with NotI before electroporation into RW4 ES cells. G418-resistant ES cell clones were screened for targeting by Southern blot analysis. One targeted ES cell line, No. 108, was microinjected into C57Bl/6J blastocysts, resulting in multiple high-percentage chimeric mice. Chimeric males were mated to C57Bl/6J females (The Jackson Laboratory). Agouti offspring that inherited the targeted allele (Gpc3 tm108Sau) were backcrossed to either C57Bl/6J or C57Bl/ 6-Bmp4 tm1blh (The Jackson Laboratory). The Gpc3 TgN(geo)ex136Ska allele was recovered in a screen, as described previously (Skarnes et al., 1995), whereby the pGT1.8TM vector inserted within the second intron of Gpc3 in 129 Ola ES cells. The Gpc3 tm108Sau allele was identified by PCR using a Gpc3-specific forward primer, Dx22099, 5⬘-ATCTCCCAGCTAAGCTCCTT-3⬘ (⫺187 to ⫺168), and reverse primer derived from the neo gene, 5⬘-ATGCTCCAGACTGCCTTG-3⬘, while the wild-type allele was identified using a reverse primer designed from the deleted exon, Dx22141, 5⬘-CCCAAGCCTAGCAGCATC-3⬘ (184 –201). Details are available on request. The Gpc3TgN(geo)ex136Ska allele was followed by detection of the lacZ reporter gene contained in the gene trapping vector (Skarnes et al., 1995), while Bmp4 knockout genotyping (Winnier et al., 1995) and sex determination (Hogan et al., 1994) were both done by PCR as described elsewhere. Timed pregnancies were determined by visualization of a vaginal plug which was defined as day 0.5. Southern analysis. ES cell clones were harvested and genomic DNA was prepared, restriction digested with HindIII and XbaI, and fractionated by agarose gel electrophoresis (George and Hynes,
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
181
GPC3 in Limb Patterning and Skeletal Development
1994). DNA was transferred to Zeta Probe (Bio-Rad) using alkaline transfer and hybridized with Rapid-Hyb (Amersham) at 65°C for 1 h using a random-labeled probe generated from a 0.5-kb EcoRV–XbaI genomic fragment immediately 3⬘ of the targeted sequence. The hybridized targeted allele gives a restriction fragment of 2.1 kb compared with the endogenous allele of 5.2 kb. Western analysis. Heparan sulfate proteoglycans were extracted from embryonic day 16.5 mouse liver in 50 mM Tris–HCl (pH 8.0), 0.15 M NaCl, 6 M urea, 1% Triton X-100, 1 mM EDTA, 1 g/ml pepstatin, 0.25 mg/ml N-ethylmaleimide, 0.5 mM phenylmethylsulfonyl fluoride and purified by DEAE chromatography as described elsewhere (Herndon and Lander, 1990). Heparitinase (Sigma) digests (0.002 U/ml) were carried out in 20 mM Tris–HCl, pH 7.5, 5 mM CaCl 2 at 37°C. Following electrophoresis in 4 –15% SDS–PAGE gel (Bio-Rad) and electroblotting to Immobilon-P (Millipore), GPC3 was detected using a rabbit polyclonal antisera generated against a synthetic peptide corresponding to murine GPC3 amino acids 524 –541. X-gal staining of tissues. Whole embryos at day 11.5, or dissected limbs also at day 11.5 and later, were fixed 30 min in 1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl 2, 5 mM EGTA, 0.02% NP-40 in PBS. Tissues were washed 3⫻ 30 min in 0.2% NP-40 in PBS and stained overnight at 37°C in 1 mg/ml X-gal, 5 mM K 3Fe(CN) 6, 5 mM K 4(CN) 6, 2 mM MgCl 2, 0.01% deoxycholate, 0.02% NP-40 in PBS. After staining, tissues were postfixed in 1% formaldehyde, 0.2% glutaraldehyde in PBS for 1 h before dehydration to 30% saline/70% EtOH, sectioned in paraffin, and lightly counterstained with eosin. Skeleton preparations. Embryonic day 16.5 animals were fixed and their skeletons were stained with alcian blue and alizarin red and cleared with KOH/glycerine as described by McLeod (1980).
RESULTS AND DISCUSSION Null Alleles of Glypican-3 Model Human SGBS To evaluate the functions of GPC3 in vivo, we created and characterized two mutant Gpc3 alleles generated by independent methods. The first allele, designated tm108, was created by homologous recombination, resulting in targeted deletion of both the transcriptional and the translational initiation sites, as well as nucleotides encoding the signal peptide and first 33 amino acids of the putative mature protein (Fig. 1A). The second allele, designated Ex136, was identified in a gene trap experiment designed to identify mutations in secreted or cell surface proteins (Skarnes et al., 1995). Insertion of the gene trap vector into the second intron of Gpc3 was confirmed by sequence analysis of 5⬘ RACE products derived from a fusion transcript (Townley et al., 1997) (Fig. 1B). Both ES cell clones were successfully transmitted through the germ line of chimeric mice following blastocyst injection. To determine the functional consequence of these mutations we evaluated the presence of GPC3 protein in DEAE-purified proteoglycan extracts from hemizygous mutant and wild-type animals by Western blotting using an antipeptide antibody recognizing the core protein of murine GPC3. While this antiserum readily identified the GPC3 heparan sulfate proteoglycan in wild-type extracts, as well as the core protein following enzymatic removal of the heparan sulfate
FIG. 1. Gene targeting, trapping, and molecular characterization. (A) Structure of the murine Gpc3 locus. In the targeting vector the neo gene was flanked on the 5⬘ side with 5.5 kb of genomic DNA and on the 3⬘ side with 2 kb of genomic DNA, introducing a deletion of the first exon of Gpc3. The identified probe hybridizes to a 2.2-kb HindIII/XbaI restriction fragment at a correctly targeted locus (tm108) compared with a 5-kb HindIII restriction fragment from the endogenous locus. (B) The trapped allele of Gpc3 (Ex136) was the result of a random integration event of the gene trapping vector pGT1.8TM within the second intron of Gpc3 as indicated. Identification of the site of integration within the Gpc3 locus was determined by 5⬘ RACE of the resulting fusion transcript. (C) A rabbit polyclonal antiserum, specific to a GPC3-derived peptide, detects the GPC3 proteoglycan and its core protein (as well as a previously described lower molecular weight fragment of GPC3) following heparitinase digestion in extracts from wild type, but not in extracts from either the targeted (tm108) or the trapped (Ex136) allele of glypicans, thus indicating that both alleles are null mutations for GPC3.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
182
Paine-Saunders et al.
FIG. 2. Mice bearing mutant alleles of Gpc3 recapitulate many of the phenotypic features of human patients with Simpson–Golabi– Behmel syndrome. (A) Male embryonic day 16.5 embryos, bearing either hemizygous Ex136 (column 4) or tm108 (column 5) mutant alleles weigh approximately 30% larger (P ⬍ 0.0001) than wildtype littermate controls (column 1), while heterozygous females for either the Ex136 (column 2) or the tm108 (column 3) allele show intermediate growth properties (P ⬍ 0.0001). (B) The kidneys of Gpc3 mutants show disregulated development of the collecting system resulting in cystic (arrowhead) medullar changes of the renal parenchyma. (C and D) Subtle skeletal abnormalities are apparent in Gpc3-deficient animals, including bifurcations of the xyphoid process as seen in D compared with a wild-type littermate control in C.
chains, no detectable GPC3 was found in extracts from animals hemizygous for either mutant allele (Fig. 1C). In contrast, immunostaining of the same blot with a monoclonal antibody recognizing the core protein of syndecan-1 detected similar levels of that heparan sulfate proteoglycan in all extracts (data not shown). Therefore we have concluded that both the tm108 and Ex136 alleles of Gpc3 are functionally null alleles. Consistent with the findings of Gpc3 mutations in humans, we found that mice bearing either the tm108 or the Ex136 mutation displayed a range of phenotypic features similar to human patients with SGBS, and no obvious differences between the phenotypes of mice bearing these two different alleles were observed (Fig. 2 and data not shown). In particular, hemizygous male mouse embryos, of either allele, were 30% larger than wild-type littermate controls (Fig. 2A), while heterozygous female animals (presumably mosaic due to X inactivation) showed intermediate growth properties. Animals deficient for GPC3 also displayed medullary cystic dysplasia of their kidneys (Fig. 2B) and a high frequency of perinatal lethality, phenotypes reported in human patients with SGBS. On embryonic day 16.5, embryos derived from heterozygous female (129SVJ/C57B16) and wild-type male (C57B16) matings were found at the expected Mendelian ratios including 51% wild type (⫹/⫹,
27/53), 23% heterozygote (Gpc3/⫹, 12/53), and 26% hemizygote (Gpc3/Y, 14/53). By the time of weaning, however, there was a significant deviation from the expected frequency of affected animals with 58% wild type (⫹/⫹, 39/67), 25% heterozygote (Gpc3/⫹, 17/67), and only 16% hemizygote (Gpc3/Y, 11/67). The majority of these deaths appear to occur within the first 2 days of life, but the reasons for this lethality have not been addressed. Some of these phenotypic features have recently been reported by another group who has independently targeted the Gpc3 allele (Cano-Gauci et al., 1999). However, a number of additional phenotypic traits seen in human SGBS were also apparent in our mutant alleles of Gpc3. Specifically, we found that ventral wall closure defects occur with incomplete penetrance (33%, n ⫽ 12). In most cases these consisted of small- to moderate-sized umbilical herniations but in one stillborn animal this was manifested as an omphalocele. In addition, stained skeletal prepara-
FIG. 3. Expression of the Gpc3 promoter is dynamically regulated in the developing limb as reflected by the distribution of -galactosidase expression from the trapped Ex136 allele. Wholemount -galactosidase staining of embryonic day 11.5 (A–E), day 12.5 (F), and day 13.5 (G) limbs, followed by paraffin infiltration and sectioning (C–G). Gpc3 is not expressed in the apical epidermal ridge (AER) as seen in whole mount (A and B) or longitudinal section (arrowheads). At embryonic day 11.5, Gpc3 is intensely expressed in the distal mesenchyme and in the proximal anterior and posterior limb bud as seen in longitudinal section (D) and is asymmetrically expressed preferentially in the dorsal third of the limb bud as seen in cross section at a level just below the AER (E). At later stages (F and G), Gpc3 expression becomes restricted to regions of cartilage differentiation in developing bone and within the interdigital webs. a, anterior; p, posterior; d, dorsal; v, ventral.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
183
GPC3 in Limb Patterning and Skeletal Development
tions revealed subtle malformations in GPC3-deficient pups, including bifurcations of the xyphoid process (Figs. 2C and 2D). The identification of these additional traits further supports the validity of murine GPC3 deficiency as a model for human SGBS and the potential value of these animals in understanding the mechanism of GPC3 function. Given the complex nature of the SGBS phenotype, and the evidence that the Drosophila glypican dally may have multiple mechanisms of action which are both tissue and developmental stage dependent (Jackson et al., 1997; Lin and Perrimon, 1999; Tsuda et al., 1999), we chose to focus on understanding the function of GPC3 in the development of a single phenotypic trait, polydactyly.
Gpc3 Expression Is Dynamically Regulated during Murine Limb Patterning As a first step in evaluating the possible roles of GPC3 in vertebrate limb patterning, we examined the normal distribution of expression of Gpc3 during limb morphogenesis using the -galactosidase reporter allele Ex136. Preliminary studies (Paine-Saunders and Saunders, unpublished) have established that expression of -galactosidase from this trapped allele faithfully recapitulates the normal expression pattern of Gpc3 mRNA as detected by in situ hybridization. Gpc3 expression in the developing limb was seen to be dynamically regulated in association with limb patterning events. At early embryonic stages (embryonic day 11.5), Gpc3 expression was clearly absent from the specialized signaling structure the apical epidermal ridge (AER) (Figs. 3A, 3B, and 3C), yet was intensely expressed in the underlying mesenchyme which responds to inductive cues from the AER. However, expression in the mesenchyme at this stage was asymmetric. As seen in longitudinal sections, Gpc3 expression was most intense in the regions of the distal mesenchyme within the progress zone and within the proximal anterior and posterior limb bud, regions which ultimately define the anterior and posterior necrotic zones (ANZ and PNZ, respectively) (Fig. 3D). On cross section of the limb bud at a level just proximal to the AER, it was apparent that Gpc3 was asymmetrically expressed more abundantly in the dorsum of the developing limb bud as well. Whether this pattern reflects positive regulation of Gpc3 gene expression in these peripheral regions of the limb mesenchyme or alternatively a suppression of Gpc3 expression in the central condensing mesenchyme region cannot be currently distinguished. At later developmental ages such as embryonic days 12.5 (Fig. 3F) and 13.5 (Fig. 3G), glypican-3 expression was restricted to the interdigital webs and the regions of chondrocytic differentiation of the developing bones.
GPC3 and BMP4 Mutations Interact to Result in Postaxial Polydactyly and Other Skeletal Defects A number of factors that regulate cell proliferation, differentiation, or death have been implicated in vertebrate
limb patterning. One of these is insulin-like growth factor-II (IGF-II). Because mutant mice with elevated serum and tissue levels of IGF-II share several phenotypic features with SGBS, including postaxial polydactyly (Eggenschwiler et al., 1997), there has been considerable speculation about the potential role of GPC3 as a negative regulator of IGF-II function (Pellegrini et al., 1998; Pilia et al., 1996). However, there has been no experimental evidence in support of this hypothesis. Indeed, on the contrary it has been impossible to demonstrate a direct physical interaction between the rat GPC3 core protein and IGF-II (Song et al., 1997), as was initially reported for human GPC3 (Pilia et al., 1996). Furthermore, recent experiments have found no alteration in tissue or serum levels of IGF-II in embryonic GPC3deficient animals (Cano-Gauci et al., 1999), a result that we have independently confirmed by quantitative radioimmunoassay (data not shown). In contrast, there is considerable in vitro evidence for a general role of heparan sulfate proteoglycans in the signaling of a number of developmentally important growth factors with spatial patterns of action that overlap with the distribution of Gpc3 expression. These include FGFs in the progress zone, Wnts in the dorsal limb bud, and BMPs within the ANZ, PNZ, and interdigital web. Furthermore, as discussed above, there is evidence clearly supporting a role for the invertebrate glypican dally in either wg or dpp signaling in the context of distinct developmental events in Drosophila (Jackson et al., 1997; Lin and Perrimon, 1999; Tsuda et al., 1999). Because of the overlapping expression pattern of GPC3 and BMP4, and the evidence of a role for BMP4 in anterior–posterior patterning (Dunn et al., 1997), we tested genetically whether GPC3 might be important in BMP4 signaling in the developing limb. Bmp4 homozygous deficient animals die during early embryonic development before limb morphogenesis (Winnier et al., 1995). Animals heterozygous for a targeted disruption of the Bmp4 gene (Bmp4 tm1blh), however, are viable and display a haploinsufficient phenotype that includes defects in anterior–posterior limb patterning (Dunn et al., 1997). In particular, it has been previously reported that on the C57Bl/6 genetic background Bmp4 heterozygotes display preaxial polydactyly of the right hindlimb with 12% penetrance (Dunn et al., 1997). We have observed similar results with these animals on a mixed 129SV/ C57Bl/6 background (Fig. 4F). But on closer inspection of embryonic day 16.5 Bmp4 heterozygotes, we also observed a small vestigial nubbin on the postaxial side of the right forelimb with approximately 50% penetrance (Fig. 4B). This nubbin typically showed little or no ossification (Fig. 4I, arrowhead) and frequently was no longer visible on newborn animals. When present at birth it typically formed an ectopic nail at this location. When animals were mated to produce animals heterozygous for a disrupted Bmp4 allele and hemizygous for either mutant Ex136 or tm108 Gpc3 allele, the nature and severity of the limb phenotype changed dramatically. Now, 100% of the animals showed some degree of polydactyly with vari-
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
184
Paine-Saunders et al.
FIG. 4. Combined genetic deficiency for Gpc3 and haploinsufficiency for Bmp4 results in postaxial polydactyly. Forelimbs (A–D, I) and hindlimbs (E–H, J, K) of embryonic day 16.5 mice bearing heterozygous mutations for Bmp4 either alone or in combination with hemizygous deficiencies for Gpc3. All animals bearing mutations in Gpc3 alone show normal limb patterning (not shown) as do the majority of animals haploinsufficient for Bmp4 mutations alone (A and E), while 10% of Bmp4 haploinsufficient animals display preaxial polydactyly of the right hindlimb and 50% display a small vestigial postaxial nubbin (B, arrow) of the right forelimb. This nubbin displays little to no ossification as shown by skeletal preparation (I). In contrast, 100% of animals bearing either hemizygous targeted (tm108) or trapped (Ex136) alleles, in combination with a heterozygous mutation for Bmp4, develop postaxial polydactyly. Forelimbs frequently reveal bilateral abnormalities involving more severe duplications of the fifth digit, which include well-ossified triphalangeal duplications (J, asterisk). The hindlimbs of these animals display postaxial polydactyly with 66% penetrance, and all display well-patterned and ossified duplications of the terminal fifth digit (K, asterisk). There were no differences in severity or penetrance of these phenotypes between either allele of Gpc3.
able severity (n ⫽ 36). In the forelimbs, the phenotype demonstrated a clear increase in the severity of the phenotype observed in Bmp4 heterozygous animals alone. Frequently both forelimbs were associated with well-formed bifurcations of the fifth digit (Figs. 4C and 4D). The redundant structures in these animals contained ossified bone in the form of an ectopic triphalangeal duplication branching from the fifth metatarsal bone (Fig. 4J), a defect never seen in Bmp4 heterozygous mutants alone.
In the hindlimb, the change in the polydactylous phenotype was even more dramatic. Here 66% of the animals had postaxial polydactyly, primarily of the right hindlimb, of a form similar to that which is seen in human patients with SGBS. The expressivity of this phenotype was complete, with all animals displaying well-formed branched bifurcations of the fifth digit identical to those shown in Figs. 4G and 4H and identical to those seen in human patients with SGBS. Similar to the forelimb defects, these structures contained ossified bone as well, containing an ectopic triphalangeal digit bifurcating from a common fifth metatarsal bone (Fig. 4K). We found no differences in severity or penetrance of polydactyly whether the targeted tm108 or the trapped Ex136 mutant allele of Gpc3 was combined with heterozygotes for Bmp4. We therefore conclude that these two alleles of Gpc3 are equivalent with respect to limb patterning and, because they represent independently generated mutant alleles of Gpc3, that the phenotypes seen are specific to a loss of GPC3 function in combination with a reduction in BMP4 function. On close examination of the remainder of the skeletons of the animals null for Gpc3 and heterozygous for Bmp4, we determined that there were other significant skeletal abnormalities with apparently complete penetrance. These animals, compared with littermate controls (Figs. 5A and 5B), had markedly abnormal chests with deformed and abnormally fused ribs (Figs. 5C and 5D). In addition, these animals were found to have dual ossification centers throughout the length of their sternum (Fig. 5F). These skeletal abnormalities were not seen in either Gpc3 null or Bmp4 heterozygous animals alone and were apparent only as a result of genetic interaction between these two mutant alleles. Interestingly, these phenotypic features are quite reminiscent of those reported in association with a homozygous deficiency of another Bmp gene, Bmp7 (Luo et al., 1995). However, in contrast to the high penetrance of this phenotype in animals deficient for Gpc3 and haploinsufficient for Bmp4, this feature was found inconsistently in Bmp7-deficient animals. This may be consistent with both the hypothesis that GPC3 is involved in a pathway that impacts multiple BMP family members and the speculation that some structurally related BMPs, such as BMP4 and BMP7, may display certain redundancies in function during development (Dudley and Robertson, 1997; Katagiri et al., 1998). However, we found that a genetic interaction was not evident in all tissues in which GPC3 and BMP4 (or other BMP genes) are coexpressed, suggesting that, like Dally, GPC3 may function in a tissue-dependent manner.
GPC3 Modulation of BMP4 Signaling This study represents the first genetic evidence of a role for glypicans in the control of cellular responses to BMPs during vertebrate development. Here we report that the loss of GPC3 results in both enhanced and synthetic phenotypes in association with BMP4 heterozygous deficiencies. As outlined below, we believe that this genetic interaction
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
185
GPC3 in Limb Patterning and Skeletal Development
FIG. 5. Combined genetic deficiency for Gpc3 and haploinsufficiency for Bmp4 results in other skeletal abnormalities. Animals bearing heterozygous null mutations for Bmp4 alone (A and B), as well as animals bearing hemizygous null mutations for Gpc3 alone (Fig. 2D and data not shown), have normally formed rib cages. In contrast, all animals with combined deficiency for Gpc3 and haploinsufficiency for Bmp4 display markedly abnormal chest walls with multiple rib fusions (C and D) as well as dual ossification centers of the sternum (F).
argues for a direct role for GPC3 in the control of cellular responses to BMP4. Although controversy remains as to whether BMPs function through short-range interactions or over longer ranges through the establishment of morphogenetic gradients, it is clear that the quantity of BMP protein available in the target tissue is critical for achieving the necessary developmental response. This is evident from several studies which have now established, in both vertebrate and invertebrate species, patterning abnormalities associated with heterozygous null mutants for BMPs, which presumably express half the normal levels of these gene products (Dunn et al., 1997; Green, 1968; Irish and Gelbart, 1987; Katagiri et al., 1998; McPherron et al., 1999). Therefore, understanding the molecular mechanisms that modify the strength of BMP signals or the cellular response to these signals in vivo is essential to deciphering the role of BMPs in the regulation of morphogenesis. One strategy that has been particularly effective in this
regard has been the identification of genetic modifiers of BMP function. For example, several mutations that modify the phenotype of weak dpp alleles in Drosophila have been identified (Ferguson and Anderson, 1992a,b; Raftery et al., 1995), and the characterization of these gene products, and their vertebrate homologs, has led rapidly to new information about the downstream signaling pathways activated by BMP and about the secreted proteins that modify BMP activities. Significantly, the Drosophila glypican Dally has been shown to modify Dpp function in at least some postembryonic developmental events in imaginal tissue (Jackson et al., 1997). In most of these tissues, mutations of dally enhance the phenotype associated with reduced dpp expression while suppressing phenotypes associated with ectopic dpp expression, and dally mutants display reduced expression of dpp target genes, suggesting that Dally directly affects cellular responses to Dpp. Similarly, we have discovered that loss of a vertebrate glypican, GPC3, modifies BMP4 function in the mouse limb. Interestingly, not only is this genetic interaction between Gpc3 and Bmp4 mutants in mice particularly strong, but the polydactylous phenotype seen in these animals is identical to the phenotype seen in human patients with Gpc3 deficiency alone. Taken together, the results of our studies and previous work in Drosophila demonstrate that the regulation of cellular responses to BMP is an evolutionarily conserved function of glypicans. There are multiple mechanisms by which Dally and GPC3 could function. For example, they could affect a parallel signaling pathway that alters cellular responses to BMPs. Alternatively, they could more directly modulate the BMP signaling pathway. By analogy with the role of another heparan sulfate proteoglycan, betaglycan, which directly potentiates TGF- binding to its signaling receptor, it has been suggested that Dally could function as a coreceptor for Dpp (Jackson et al., 1997) and GPC3 could have a similar function with respect to BMP4 in the vertebrate limb. Consistent with this model, Drosophila Dpp (Groppe et al., 1998) and vertebrate BMP2, -3, -4, and -7 (Ruppert et al., 1996; Sampath et al., 1987, 1992; Wang et al., 1988) all bind to heparin with high affinity in vitro and therefore likely bind to heparan sulfate in vivo. However, the consequence of this interaction to BMP function remains unclear. For example, although heparin potentiates the activity of BMP2 in a chick limb bud differentiation model in vitro, as would be expected if it were serving a coreceptor function, so does destruction of the N-terminal heparin binding domain of BMP2 or addition of a peptide corresponding to the N-terminal heparin binding domain of BMP2 in these same assays (Ruppert et al., 1996). These latter results can be interpreted as evidence that at least some heparan sulfate-containing proteins actually sequester BMPs in vivo and serve to suppress rather than enhance BMP functions. Another likely mechanism by which GPC3 could directly affect BMP4 activity in the developing limb could be through regulation of the distribution or level of accumu-
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
186
Paine-Saunders et al.
lation of a BMP antagonist, several of which are known to bind to heparin with high affinity. In particular, the heparin binding BMP antagonist Gremlin has recently been shown to mediate the SHH/FGF4 feedback loop in vertebrate limb buds through direct BMP antagonism (Zuniga et al., 1999). Loss of the heparan sulfate proteoglycan GPC3 could result in the increased accumulation and/or a broader distribution of Gremlin protein in the developing limb fields, thus resulting in an increase in signaling through this feedback loop and therefore development of redundant postaxial structures. Further experimentation will be necessary to carefully evaluate each of these hypotheses about the role of GPC3 in BMP signaling and vertebrate limb patterning. This analysis will not be trivial because heparan sulfate proteoglycans have been ascribed with various other functions, both in vitro and in vivo, and GPC3 has not yet been specifically tested for these functions. The potential promiscuity with which GPC3 may carry out the regulation of cellular processes introduces unique challenges to the understanding of its in vivo function.
ACKNOWLEDGMENTS We thank Jonathan Gitlin, Dave Wilson, and Ralph Sanderson for critical review of the manuscript. These studies were supported by National Institutes of Health Grant DK56063 (S.S.) and March of Dimes Birth Defects Foundation Grant 6-FY99-441 (S.S.). S.S. was also supported as both a Spoehrer Scholar in Pediatrics and a Scholar of the Child Health Research Center of Excellence in Developmental Biology at Washington University School of Medicine (HD33688). W.C.S. is a 1998 Searle Scholar.
REFERENCES Behmel, A., Plochl, E., and Rosenkranz, W. (1984). A new X-linked dysplasia gigantism syndrome: Identical with the Simpson dysplasia syndrome? Hum. Genet. 67, 409 – 413. Bellaiche, Y., The, I., and Perrimon, N. (1998). Tout-velu is a Drosophila homologue of the putative tumor suppressor EXT-1 and is needed for Hh diffusion. Nature 394, 85– 88. Binari, R. C., Staveley, B. E., Johnson, W. A., Godavarti, R., Sasisekharan, R., and Manoukian, A. S. (1997). Genetic evidence that heparin-like glycosaminoglycans are involved in wingless signaling. Development 124, 2623–2632. Bullock, S. L., Fletcher, J. M., Beddington, R. S., and Wilson, V. A. (1998). Renal agenesis in mice homozygous for a gene trap mutation in the gene encoding heparan sulfate 2-sulfotransferase. Genes Dev. 12, 1894 –1906. Cano-Gauci, D. F., Song, H. H., Yang, H., McKerlie, C., Choo, B., Shi, W., Pullano, R., Piscione, T. D., Grisaru, S., Soon, S., Sedlackova, L., Tanswell, A. K., Mak, T. W., Yeger, H., Lockwood, G. A., Rosenblum, N. D., and Filmus, J. (1999). Glypican3-deficient mice exhibit developmental overgrowth and some of the abnormalities typical of Simpson–Golabi–Behmel syndrome. J. Cell Biol. 146, 255–264. David, G., Lories, V., Decock, B., Marynen, P., Cassiman, J.-J., and Berghe, H. V. d. (1990). Molecular cloning of a phosphati-
dylinositol-anchored membrane heparan sulfate proteoglycan from human lung fibroblasts. J. Cell Biol. 111, 3165–3176. Dudley, A. T., and Robertson, E. J. (1997). Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 deficient embryos. Dev. Dyn. 208, 349 –362. Dunn, N. R., Winnier, G. E., Hargett, L. K., Schrick, J. J., Fogo, A. B., and Hogan, B. L. M. (1997). Haploinsufficient phenotypes in BMP4 heterozygous null mice and modification by mutations in Gli3 and Alx4. Dev. Biol. 188, 235–247. Eggenschwiler, J., Ludwig, T., Fisher, P., Leighton, P. A., Tilghman, S. M., and Efstratiadis, A. (1997). Mouse mutant embryos overexpressing IGF-II exhibit phenotypic features of the Beckwith– Wiedemann and Simpson–Golabi–Behmel syndromes. Genes Dev. 11, 3128 –3142. Erickson, R. P. (1996). Mouse models of human genetic disease: Which mouse is more like a man? BioEssays 18, 993–998. Ferguson, E. L., and Anderson, K. V. (1992a). decapentaplegic acts as a morphogen to organize dorsal–ventral pattern in the Drosophila embryo. Cell 71, 451– 461. Ferguson, E. L., and Anderson, K. V. (1992b). Localized enhancement and repression of the activity of the TGF- family member, decapentaplegic, is necessary for dorsal–ventral pattern formation in the Drosophila embryo. Development 114, 583–597. Garganta, C. L., and Bodurtha, J. N. (1992). Report of another family with Simpson–Golabi–Behmel syndrome and a review of the literature. Am. J. Med. Genet. 44, 129 –135. George, E. L., and Hynes, R. O. (1994). Gene targeting and generation of mutant mice for studies of cell– extracellular matrix interactions. Methods Enzymol. 245, 386 – 420. Green, M. C. (1968). Mechanism of the pleiotropic effects of the short ear mutant gene in the mouse. J. Exp. Zool. 167, 129 –150. Groppe, J., Rumpel, K., Economides, A. N., Stahl, N., Sebald, W., and Affolter, M. (1998). Biochemical and biophysical characterization of refolded Drosophila DPP, a homolog of bone morphogeneic proteins 2 and 4. J. Biol. Chem. 273, 29052–29065. Hacker, U., Lin, X., and Perrimon, N. (1997). The Drosophila sugarless gene modulates Wingless signaling and encodes an enzyme involved in polysaccharide biosynthesis. Development 124, 3565–3573. Haerry, T. E., Heslip, T. R., Marsh, J. L., and O’Connor, M. B. (1997). Defects in glucuronate biosynthesis disrupt wingless signaling in Drosophila. Development 124, 3055–3064. Herndon, M. E., and Lander, A. D. (1990). A diverse set of developmentally regulated proteoglycans is expressed in the rat central nervous system. Neuron 4, 949 –961. Hogan, B., Beddington, R., Costantini, F., and Lacey, E. (1994). “Manipulating the Mouse Embryo: A Laboratory Manual,” 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Hughes-Benzie, R. M., Hunter, A. G., Allanson, J. E., and MacKenzie, A. E. (1992). Simpson–Golabi–Behmel syndrome associated with renal dysplasia and embryonal tumors: Localization of the gene to Xqcen– q21. Am. J. Med. Genet. 43, 428 – 435. Irish, V. F., and Gelbart, W. M. (1987). The decapentaplegic gene is required for dorsal–ventral patterning of the Drosophila embryo. Genes Dev. 1, 868 – 879. Jackson, S. M., Nakato, H., Sugiura, M., Jannuzi, A., Oakes, R., Kaluza, V., Golden, C., and Selleck, S. B. (1997). Division abnormally delayed, a Drosophila glypican, controls cellular responses to the TGF- related morphogen, Dpp. Development 124, 4113– 4120.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
187
GPC3 in Limb Patterning and Skeletal Development
Katagiri, T., Boorla, S., Frendo, J. L., Hogan, B. L., and Karsenty, G. (1998). Skeletal abnormalities in doubly heterozygous Bmp4 and Bmp7 mice. Dev. Genet. 22, 340 –348. Lin, X., and Perrimon, N. (1999). Dally cooperates with Drosophila Frizzled 2 to transduce Wingless signalling. Nature 400, 281–284. Lind, T., Tufaro, F., McCormick, C., Lindahl, U., and Lidholt, K. (1998). The putative tumor suppressors EXT1 and EXT2 are glycosyltransferases required for the biosynthesis of heparan sulfate. J. Biol. Chem. 273, 26265–26268. Luo, G., Hofmann, C., Bronckers, A. L. J. J., Sohocki, M., Bradley, A., and Karsenty, G. (1995). BMP-7 is an inducer of nephrogenesis, and is also required for eye development and skeletal patterning. Genes Dev. 9, 2808 –2820. McCormick, C., Leduc, Y., Martindale, D., Mattison, K., Esford, L. E., Dyer, A. P., and Tufaro, F. (1998). The putative tumor suppressor EXT1 alters the expression of cell-surface heparan sulfate. Nat. Genet. 19, 158 –161. McLeod, M. J. (1980). Differential staining of cartilage and bone in whole mouse fetuses by alcian blue and alizarin red S. Teratology 22, 299 –301. McPherron, A. C., Lawler, A. M., and Lee, S. J. (1999). Regulation of anterior/posterior patterning of the axial skeleton by growth/ differentiation factor 11. Nat. Genet. 22, 260 –264. Paine-Saunders, S., Viviano, B. L., and Saunders, S. (1999). GPC6, a novel member of the glypican gene family, encodes a product structurally related to GPC4 and is colocalized with GPC5 on human chromosome 13. Genomics 57, 455– 458. Pellegrini, M., Pilia, G., Pantano, S., Lucchini, F., Uda, M., Fumi, M., Cao, A., Schessinger, D., and Forabosco, A. (1998). Gpc3 expression correlates with the phenotype of the Simpson– Golabi–Behmel syndrome. Dev. Dyn. 213, 431– 439. Pilia, G., Hughes-Benzie, R. M., MacKenzie, A., Baybayan, P., Chen, E. Y., Huber, R., Neri, G., Cao, A., Foarabosco, A., and Schlessinger, D. (1996). Mutations in GPC3, a glypican gene, cause the Simpson–Golabi–Behmel overgrowth syndrome. Nature Genet. 12, 241–247. Raftery, L. A., Twombly, V., Wharton, K., and Gelbart, W. M. (1995). Genetic screens to identify elements of the decapentaplegic signaling pathway in Drosophila. Genetics 139, 214 –254. Rapraeger, A., Krufka, A., and Olwin, B. B. (1991). Requirement of heparan sulfate for bFGF-mediated fibroblast growth and myoblast differentiation. Science 252, 1705–1708. Reichsman, F., Smith, L., and Cumberledge, S. (1996). Glycosaminoglycans can modulate extracellular localization of the wingless protein and promote signal transduction. J. Cell Biol. 135, 819 – 827. Ruppert, R., Hoffmann, E., and Sebald, W. (1996). Human bone morphogenetic protein 2 contains a heparin-binding site which modifies its biological activity. Eur. J. Biochem. 237, 295–302. Sampath, T. K., Maliakal, J. C., Hauschka, P. V., Jones, W. K., Sasak, H., Tucker, R. F., White, K. H., Coughlin, J. E., Tucker, M. M., Pang, R. H. L., Corbett, C., Ozkaynak, E., Oppermann, H.,
and Rueger, D. C. (1992). Recombinant human osteogenic protein-1 (hOP-1) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates osteoblast proliferation and differentiation in vitro. J. Biol. Chem. 267, 20352–20362. Sampath, T. K., Muthukumaran, N., and Reddi, A. H. (1987). Isolation of osteogenin, an extracellular matrix-associated, boneinductive protein, by heparin affinity chromatography. Proc. Natl. Acad. Sci. USA 84, 7109 –7113. Saunders, S., Jalkanen, M., O’Farrell, S., and Bernfield, M. (1989). Molecular cloning of syndecan, an integral membrane proteoglycan. J. Cell Biol. 108, 1547–1556. Skarnes, W. C., Moss, J. E., Hurtley, S. M., and Beddington, R. S. (1995). Capturing genes encoding membrane and secreted proteins important for mouse development. Proc. Natl. Acad. Sci. USA 92, 6592– 6596. Song, H. H., Shi, W., and Filmus, J. (1997). OCI-5/rat glypican-3 binds to fibroblast growth factor-2 but not to insulin-like growth factor-2. J. Biol. Chem. 272, 7574 –7577. The, I., Bellaiche, Y., and Perrimon, N. (1999). Hedgehog movement is regulated through tout velu-dependent synthesis of a heparan sulfate proteoglycan. Mol. Cell 4, 633– 639. Townley, D. J., Avery, B. J., Rosen, B., and Skarnes, W. C. (1997). Rapid sequence analysis of gene trap integrations to generate a resource of insertional mutations in mice. Genome Res. 7, 293–298. Tsuda, M., Kamimura, K., Nakato, H., Archer, M., Staatz, W., Fox, B., Humphrey, M., Olson, S., Futch, T., Kaluza, V., Siegfried, E., Stam, L., and Selleck, S. B. (1999). The cell-surface proteoglycan Dally regulates Wingless signalling in Drosophila. Nature 400, 276 –280. Wang, E. A., Rosen, V., Cordes, P., Hewick, R. M., Kriz, M. J., Luxenberg, D. P., Sibley, B. S., and Wozney, J. M. (1988). Purification and characterization of other distinct bone-inducing factors. Proc. Natl. Acad. Sci. USA 85, 9484 –9488. Winnier, G., Blessing, M., Labosky, P. A., and Hogan, B. L. M. (1995). Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev. 9, 2105– 2116. Winter, R. M. (1996). Analysing human developmental abnormalities. BioEssays 18, 965–971. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64, 841– 848. Zuniga, A., Haramis, A. G., McMahon, A. P., and Zeller, R. (1999). Signal relay by BMP antagonism controls the SHH/FGF4 feedback loop in vertebrate limb buds. Nature 401, 598 – 602. Received for publication April 12, 2000 Revised May 26, 2000 Accepted June 16, 2000
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
glypican-3 Controls Cellular Responses to Bmp4 in Limb Patterning and Skeletal Development Stephenie Paine-Saunders,* Beth L. Viviano,* Joel Zupicich,† William C. Skarnes,† and Scott Saunders* ,‡ ,1 *Division of Newborn Medicine, Department of Pediatrics, Washington University School of Medicine and St. Louis Children’s Hospital, St. Louis, Missouri 63110; †Department of Molecular and Cell Biology, University of California at Berkeley, Berkeley, California 94720; and ‡Department of Molecular Biology and Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110
Glypicans represent a family of six cell surface heparan sulfate proteoglycans in vertebrates. Although no specific in vivo functions have thus far been described for these proteoglycans, spontaneous mutations in the human and induced deletions in the mouse glypican-3 (Gpc3) gene result in severe malformations and both pre- and postnatal overgrowth, known clinically as the Simpson–Golabi–Behmel syndrome (SGBS). Mice carrying mutant alleles of Gpc3 created by either targeted gene disruption or gene trapping display a wide range of phenotypes associated with SGBS including renal cystic dysplasia, ventral wall defects, and skeletal abnormalities that are consistent with the pattern of Gpc3 expression in the mouse embryo. Previous studies in Drosophila have implicated glypicans in the signaling of decapentaplegic, a BMP homolog. Our experiments with mice show a significant relationship between vertebrate BMP signaling and glypican function; GPC3-deficient animals were mated with mice haploinsufficient for bone morphogenetic protein-4 (Bmp4) and their offspring displayed a high penetrance of postaxial polydactyly and rib malformations not observed in either parent strain. This previously unknown link between glypican-3 and BMP4 function provides evidence of a role for glypicans in vertebrate limb patterning and skeletal development and suggests a mechanism for the skeletal defects seen in SGBS. © 2000 Academic Press
Key Words: heparan sulfate; glypican; proteoglycans; bone morphogenetic protein; limb; skeleton; polydactyly; Simpson– Golabi–Behmel syndrome.
INTRODUCTION The essential requirement of heparan sulfate glycosaminoglycans for normal development has been emphasized only recently by the discovery that mutations in heparan sulfate biosynthesis cause significant developmental defects in Drosophila (Bellaiche et al., 1998; Binari et al., 1997; Hacker et al., 1997; Haerry et al., 1997), in mice (Bullock et al., 1998), and in humans (Lind et al., 1998; McCormick et al., 1998). This has stimulated interest in understanding the potential functions of specific protein gene products that bear these posttranslational modifica1
To whom correspondence should be addressed at the Washington University School of Medicine, 660 South Euclid Avenue, Campus Box 8208, St. Louis, MO 63110. Fax: (314) 286-2893. E-mail: [email protected]. 0012-1606/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
tions. Although heparan sulfate is found ubiquitously on cell surfaces, these carbohydrates are carried predominantly by the protein products of only two gene families, syndecans (Saunders et al., 1989) and glypicans (David et al., 1990). All members of the glypican family share a common mechanism of attachment to the cell membrane through glycosylphosphatidylinositol linkages, and the characterization of amino acid sequence homologies has suggested the existence of subfamilies of structurally related vertebrate glypicans (Paine-Saunders et al., 1999). Whether these individual glypicans have distinct or a related function is unknown; however, human mutations in a single glypican gene, glypican-3 (Gpc3), result in an overgrowth phenotype known as Simpson–Golabi–Behmel syndrome (SGBS). This X-linked condition is characterized by both pre- and postnatal overgrowth, a distinct facial appearance, macroglos-
179
180
Paine-Saunders et al.
sia, and multiple anomalies, including cleft palate, congenital heart and renal defects, imperforate vagina, umbilical hernias, supernumerary nipples, vertebral and rib defects, and postaxial polydactyly (Behmel et al., 1984; Garganta and Bodurtha, 1992; Hughes-Benzie et al., 1992; Pilia et al., 1996). Nothing is known about the mechanism by which these in vivo mutations of Gpc3 lead to the described phenotypic features in humans. Nonetheless, as described in this study and elsewhere (Cano-Gauci et al., 1999), GPC3-deficient mice recapitulate the SGBS phenotype and therefore provide an excellent in vivo model system for evaluating the functions of vertebrate glypicans. Although the mechanism of action of vertebrate glypicans is unknown, the function of the Drosophila glypican dally has been evaluated genetically. These studies have revealed a role for dally in both wingless (wg; the Drosophila ortholog of vertebrate Wnt) signaling during early embryonic patterning (Lin and Perrimon, 1999; Tsuda et al., 1999) and decapentaplegic (dpp; the Drosophila ortholog of vertebrate bone morphogenetic proteins) signaling during postembryonic morphogenesis (Jackson et al., 1997). The mechanism of Dally function has not been experimentally established; however, these findings are consistent with the proposed function for heparan sulfate proteoglycans as coreceptors for a number of growth factors in vitro. Moreover, it is clear from the above studies that glypican functions are highly context and tissue dependent in Drosophila and are likely to be similarly context dependent in vertebrates. Therefore it is not obvious which signaling pathways might be impacted by the loss of function of a specific vertebrate glypican in any particular tissue during development. We have therefore chosen to focus on understanding the function of GPC3 in the pathogenesis of a single phenotypic trait in vertebrates, polydactyly. Although commonly seen in human patients with SGBS, this phenotype is never seen in GPC3-deficient mice. We therefore speculated that these animals may represent a sensitized strain for glypican function in limb development and that genetic interaction studies might point toward the signaling pathway(s) mediated by GPC3 in this tissue. Indeed, as has been suggested by others, the identification of genetic modifiers provides an important pathway to understanding the cellular and genetic interactions underlying embryonic development and the pathogenesis of human malformations (Dunn et al., 1997; Erickson, 1996; Winter, 1996). We have found that the distribution of Gpc3 expression overlaps with the patterns of a number of gene products implicated in limb patterning and whose biological activities are known to be dependent in part on heparan sulfate proteoglycans. These include fibroblast growth factors (FGFs) (Rapraeger et al., 1991; Yayon et al., 1991), Wnts (Lin and Perrimon, 1999; Reichsman et al., 1996; Tsuda et al., 1999), bone morphogenetic proteins (BMPs) (Jackson et al., 1997; Ruppert et al., 1996), and members of the vertebrate hedgehog family (The et al., 1999). In particular, Bmp4 expression in the developing limb is highly coincident with
Gpc3 expression in the progress zone, in the anterior and posterior necrotic zones, and in the interdigital web where it plays an active role in mediating critical apoptotic events in the development of these structures. Furthermore, Bmp4 haploinsufficient mice show a phenotype implicating a role for BMPs in anterior–posterior patterning (Dunn et al., 1997). Therefore we decided to test genetically whether GPC3 might be important in BMP4 signaling in the developing limb and have found a strong genetic interaction between mutants of these two genes that result in a high penetrance of both postaxial polydactyly and rib malformations. This study provides the first in vivo evidence of a role for glypicans in vertebrate development and establishes a previously unknown link between GPC3 and BMP4 function, which suggests a mechanism for the skeletal defects seen in SGBS.
MATERIALS AND METHODS Generation of mutant mice. A 7.5-kb genomic fragment containing the first exon of Gpc3, which includes the signal peptide, was isolated from a strain 129SV genomic library in the Lambda FIX II vector (Stratagene) using a 5⬘-specific probe from human Gpc3 cDNA. The DNA targeting vector was constructed by replacing the first exon of Gpc3 (from approximately ⫺50 bp to just downstream of the first splice site) with the AscI fragment containing the PGK-neo gene from pKO SelectNeo V800 (Lexicon). This resulted in a vector flanked with 5.5 kb of homologous sequence upstream and 2.0 kb of homologous sequence downstream. Additionally, an RsrII fragment containing the RNA Pol II diphtheria toxin gene from pKO SelectDT V840 (Lexicon) was inserted at the 3⬘ end to facilitate negative selection. The targeting vector was linearized with NotI before electroporation into RW4 ES cells. G418-resistant ES cell clones were screened for targeting by Southern blot analysis. One targeted ES cell line, No. 108, was microinjected into C57Bl/6J blastocysts, resulting in multiple high-percentage chimeric mice. Chimeric males were mated to C57Bl/6J females (The Jackson Laboratory). Agouti offspring that inherited the targeted allele (Gpc3 tm108Sau) were backcrossed to either C57Bl/6J or C57Bl/ 6-Bmp4 tm1blh (The Jackson Laboratory). The Gpc3 TgN(geo)ex136Ska allele was recovered in a screen, as described previously (Skarnes et al., 1995), whereby the pGT1.8TM vector inserted within the second intron of Gpc3 in 129 Ola ES cells. The Gpc3 tm108Sau allele was identified by PCR using a Gpc3-specific forward primer, Dx22099, 5⬘-ATCTCCCAGCTAAGCTCCTT-3⬘ (⫺187 to ⫺168), and reverse primer derived from the neo gene, 5⬘-ATGCTCCAGACTGCCTTG-3⬘, while the wild-type allele was identified using a reverse primer designed from the deleted exon, Dx22141, 5⬘-CCCAAGCCTAGCAGCATC-3⬘ (184 –201). Details are available on request. The Gpc3TgN(geo)ex136Ska allele was followed by detection of the lacZ reporter gene contained in the gene trapping vector (Skarnes et al., 1995), while Bmp4 knockout genotyping (Winnier et al., 1995) and sex determination (Hogan et al., 1994) were both done by PCR as described elsewhere. Timed pregnancies were determined by visualization of a vaginal plug which was defined as day 0.5. Southern analysis. ES cell clones were harvested and genomic DNA was prepared, restriction digested with HindIII and XbaI, and fractionated by agarose gel electrophoresis (George and Hynes,
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
181
GPC3 in Limb Patterning and Skeletal Development
1994). DNA was transferred to Zeta Probe (Bio-Rad) using alkaline transfer and hybridized with Rapid-Hyb (Amersham) at 65°C for 1 h using a random-labeled probe generated from a 0.5-kb EcoRV–XbaI genomic fragment immediately 3⬘ of the targeted sequence. The hybridized targeted allele gives a restriction fragment of 2.1 kb compared with the endogenous allele of 5.2 kb. Western analysis. Heparan sulfate proteoglycans were extracted from embryonic day 16.5 mouse liver in 50 mM Tris–HCl (pH 8.0), 0.15 M NaCl, 6 M urea, 1% Triton X-100, 1 mM EDTA, 1 g/ml pepstatin, 0.25 mg/ml N-ethylmaleimide, 0.5 mM phenylmethylsulfonyl fluoride and purified by DEAE chromatography as described elsewhere (Herndon and Lander, 1990). Heparitinase (Sigma) digests (0.002 U/ml) were carried out in 20 mM Tris–HCl, pH 7.5, 5 mM CaCl 2 at 37°C. Following electrophoresis in 4 –15% SDS–PAGE gel (Bio-Rad) and electroblotting to Immobilon-P (Millipore), GPC3 was detected using a rabbit polyclonal antisera generated against a synthetic peptide corresponding to murine GPC3 amino acids 524 –541. X-gal staining of tissues. Whole embryos at day 11.5, or dissected limbs also at day 11.5 and later, were fixed 30 min in 1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl 2, 5 mM EGTA, 0.02% NP-40 in PBS. Tissues were washed 3⫻ 30 min in 0.2% NP-40 in PBS and stained overnight at 37°C in 1 mg/ml X-gal, 5 mM K 3Fe(CN) 6, 5 mM K 4(CN) 6, 2 mM MgCl 2, 0.01% deoxycholate, 0.02% NP-40 in PBS. After staining, tissues were postfixed in 1% formaldehyde, 0.2% glutaraldehyde in PBS for 1 h before dehydration to 30% saline/70% EtOH, sectioned in paraffin, and lightly counterstained with eosin. Skeleton preparations. Embryonic day 16.5 animals were fixed and their skeletons were stained with alcian blue and alizarin red and cleared with KOH/glycerine as described by McLeod (1980).
RESULTS AND DISCUSSION Null Alleles of Glypican-3 Model Human SGBS To evaluate the functions of GPC3 in vivo, we created and characterized two mutant Gpc3 alleles generated by independent methods. The first allele, designated tm108, was created by homologous recombination, resulting in targeted deletion of both the transcriptional and the translational initiation sites, as well as nucleotides encoding the signal peptide and first 33 amino acids of the putative mature protein (Fig. 1A). The second allele, designated Ex136, was identified in a gene trap experiment designed to identify mutations in secreted or cell surface proteins (Skarnes et al., 1995). Insertion of the gene trap vector into the second intron of Gpc3 was confirmed by sequence analysis of 5⬘ RACE products derived from a fusion transcript (Townley et al., 1997) (Fig. 1B). Both ES cell clones were successfully transmitted through the germ line of chimeric mice following blastocyst injection. To determine the functional consequence of these mutations we evaluated the presence of GPC3 protein in DEAE-purified proteoglycan extracts from hemizygous mutant and wild-type animals by Western blotting using an antipeptide antibody recognizing the core protein of murine GPC3. While this antiserum readily identified the GPC3 heparan sulfate proteoglycan in wild-type extracts, as well as the core protein following enzymatic removal of the heparan sulfate
FIG. 1. Gene targeting, trapping, and molecular characterization. (A) Structure of the murine Gpc3 locus. In the targeting vector the neo gene was flanked on the 5⬘ side with 5.5 kb of genomic DNA and on the 3⬘ side with 2 kb of genomic DNA, introducing a deletion of the first exon of Gpc3. The identified probe hybridizes to a 2.2-kb HindIII/XbaI restriction fragment at a correctly targeted locus (tm108) compared with a 5-kb HindIII restriction fragment from the endogenous locus. (B) The trapped allele of Gpc3 (Ex136) was the result of a random integration event of the gene trapping vector pGT1.8TM within the second intron of Gpc3 as indicated. Identification of the site of integration within the Gpc3 locus was determined by 5⬘ RACE of the resulting fusion transcript. (C) A rabbit polyclonal antiserum, specific to a GPC3-derived peptide, detects the GPC3 proteoglycan and its core protein (as well as a previously described lower molecular weight fragment of GPC3) following heparitinase digestion in extracts from wild type, but not in extracts from either the targeted (tm108) or the trapped (Ex136) allele of glypicans, thus indicating that both alleles are null mutations for GPC3.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
182
Paine-Saunders et al.
FIG. 2. Mice bearing mutant alleles of Gpc3 recapitulate many of the phenotypic features of human patients with Simpson–Golabi– Behmel syndrome. (A) Male embryonic day 16.5 embryos, bearing either hemizygous Ex136 (column 4) or tm108 (column 5) mutant alleles weigh approximately 30% larger (P ⬍ 0.0001) than wildtype littermate controls (column 1), while heterozygous females for either the Ex136 (column 2) or the tm108 (column 3) allele show intermediate growth properties (P ⬍ 0.0001). (B) The kidneys of Gpc3 mutants show disregulated development of the collecting system resulting in cystic (arrowhead) medullar changes of the renal parenchyma. (C and D) Subtle skeletal abnormalities are apparent in Gpc3-deficient animals, including bifurcations of the xyphoid process as seen in D compared with a wild-type littermate control in C.
chains, no detectable GPC3 was found in extracts from animals hemizygous for either mutant allele (Fig. 1C). In contrast, immunostaining of the same blot with a monoclonal antibody recognizing the core protein of syndecan-1 detected similar levels of that heparan sulfate proteoglycan in all extracts (data not shown). Therefore we have concluded that both the tm108 and Ex136 alleles of Gpc3 are functionally null alleles. Consistent with the findings of Gpc3 mutations in humans, we found that mice bearing either the tm108 or the Ex136 mutation displayed a range of phenotypic features similar to human patients with SGBS, and no obvious differences between the phenotypes of mice bearing these two different alleles were observed (Fig. 2 and data not shown). In particular, hemizygous male mouse embryos, of either allele, were 30% larger than wild-type littermate controls (Fig. 2A), while heterozygous female animals (presumably mosaic due to X inactivation) showed intermediate growth properties. Animals deficient for GPC3 also displayed medullary cystic dysplasia of their kidneys (Fig. 2B) and a high frequency of perinatal lethality, phenotypes reported in human patients with SGBS. On embryonic day 16.5, embryos derived from heterozygous female (129SVJ/C57B16) and wild-type male (C57B16) matings were found at the expected Mendelian ratios including 51% wild type (⫹/⫹,
27/53), 23% heterozygote (Gpc3/⫹, 12/53), and 26% hemizygote (Gpc3/Y, 14/53). By the time of weaning, however, there was a significant deviation from the expected frequency of affected animals with 58% wild type (⫹/⫹, 39/67), 25% heterozygote (Gpc3/⫹, 17/67), and only 16% hemizygote (Gpc3/Y, 11/67). The majority of these deaths appear to occur within the first 2 days of life, but the reasons for this lethality have not been addressed. Some of these phenotypic features have recently been reported by another group who has independently targeted the Gpc3 allele (Cano-Gauci et al., 1999). However, a number of additional phenotypic traits seen in human SGBS were also apparent in our mutant alleles of Gpc3. Specifically, we found that ventral wall closure defects occur with incomplete penetrance (33%, n ⫽ 12). In most cases these consisted of small- to moderate-sized umbilical herniations but in one stillborn animal this was manifested as an omphalocele. In addition, stained skeletal prepara-
FIG. 3. Expression of the Gpc3 promoter is dynamically regulated in the developing limb as reflected by the distribution of -galactosidase expression from the trapped Ex136 allele. Wholemount -galactosidase staining of embryonic day 11.5 (A–E), day 12.5 (F), and day 13.5 (G) limbs, followed by paraffin infiltration and sectioning (C–G). Gpc3 is not expressed in the apical epidermal ridge (AER) as seen in whole mount (A and B) or longitudinal section (arrowheads). At embryonic day 11.5, Gpc3 is intensely expressed in the distal mesenchyme and in the proximal anterior and posterior limb bud as seen in longitudinal section (D) and is asymmetrically expressed preferentially in the dorsal third of the limb bud as seen in cross section at a level just below the AER (E). At later stages (F and G), Gpc3 expression becomes restricted to regions of cartilage differentiation in developing bone and within the interdigital webs. a, anterior; p, posterior; d, dorsal; v, ventral.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
183
GPC3 in Limb Patterning and Skeletal Development
tions revealed subtle malformations in GPC3-deficient pups, including bifurcations of the xyphoid process (Figs. 2C and 2D). The identification of these additional traits further supports the validity of murine GPC3 deficiency as a model for human SGBS and the potential value of these animals in understanding the mechanism of GPC3 function. Given the complex nature of the SGBS phenotype, and the evidence that the Drosophila glypican dally may have multiple mechanisms of action which are both tissue and developmental stage dependent (Jackson et al., 1997; Lin and Perrimon, 1999; Tsuda et al., 1999), we chose to focus on understanding the function of GPC3 in the development of a single phenotypic trait, polydactyly.
Gpc3 Expression Is Dynamically Regulated during Murine Limb Patterning As a first step in evaluating the possible roles of GPC3 in vertebrate limb patterning, we examined the normal distribution of expression of Gpc3 during limb morphogenesis using the -galactosidase reporter allele Ex136. Preliminary studies (Paine-Saunders and Saunders, unpublished) have established that expression of -galactosidase from this trapped allele faithfully recapitulates the normal expression pattern of Gpc3 mRNA as detected by in situ hybridization. Gpc3 expression in the developing limb was seen to be dynamically regulated in association with limb patterning events. At early embryonic stages (embryonic day 11.5), Gpc3 expression was clearly absent from the specialized signaling structure the apical epidermal ridge (AER) (Figs. 3A, 3B, and 3C), yet was intensely expressed in the underlying mesenchyme which responds to inductive cues from the AER. However, expression in the mesenchyme at this stage was asymmetric. As seen in longitudinal sections, Gpc3 expression was most intense in the regions of the distal mesenchyme within the progress zone and within the proximal anterior and posterior limb bud, regions which ultimately define the anterior and posterior necrotic zones (ANZ and PNZ, respectively) (Fig. 3D). On cross section of the limb bud at a level just proximal to the AER, it was apparent that Gpc3 was asymmetrically expressed more abundantly in the dorsum of the developing limb bud as well. Whether this pattern reflects positive regulation of Gpc3 gene expression in these peripheral regions of the limb mesenchyme or alternatively a suppression of Gpc3 expression in the central condensing mesenchyme region cannot be currently distinguished. At later developmental ages such as embryonic days 12.5 (Fig. 3F) and 13.5 (Fig. 3G), glypican-3 expression was restricted to the interdigital webs and the regions of chondrocytic differentiation of the developing bones.
GPC3 and BMP4 Mutations Interact to Result in Postaxial Polydactyly and Other Skeletal Defects A number of factors that regulate cell proliferation, differentiation, or death have been implicated in vertebrate
limb patterning. One of these is insulin-like growth factor-II (IGF-II). Because mutant mice with elevated serum and tissue levels of IGF-II share several phenotypic features with SGBS, including postaxial polydactyly (Eggenschwiler et al., 1997), there has been considerable speculation about the potential role of GPC3 as a negative regulator of IGF-II function (Pellegrini et al., 1998; Pilia et al., 1996). However, there has been no experimental evidence in support of this hypothesis. Indeed, on the contrary it has been impossible to demonstrate a direct physical interaction between the rat GPC3 core protein and IGF-II (Song et al., 1997), as was initially reported for human GPC3 (Pilia et al., 1996). Furthermore, recent experiments have found no alteration in tissue or serum levels of IGF-II in embryonic GPC3deficient animals (Cano-Gauci et al., 1999), a result that we have independently confirmed by quantitative radioimmunoassay (data not shown). In contrast, there is considerable in vitro evidence for a general role of heparan sulfate proteoglycans in the signaling of a number of developmentally important growth factors with spatial patterns of action that overlap with the distribution of Gpc3 expression. These include FGFs in the progress zone, Wnts in the dorsal limb bud, and BMPs within the ANZ, PNZ, and interdigital web. Furthermore, as discussed above, there is evidence clearly supporting a role for the invertebrate glypican dally in either wg or dpp signaling in the context of distinct developmental events in Drosophila (Jackson et al., 1997; Lin and Perrimon, 1999; Tsuda et al., 1999). Because of the overlapping expression pattern of GPC3 and BMP4, and the evidence of a role for BMP4 in anterior–posterior patterning (Dunn et al., 1997), we tested genetically whether GPC3 might be important in BMP4 signaling in the developing limb. Bmp4 homozygous deficient animals die during early embryonic development before limb morphogenesis (Winnier et al., 1995). Animals heterozygous for a targeted disruption of the Bmp4 gene (Bmp4 tm1blh), however, are viable and display a haploinsufficient phenotype that includes defects in anterior–posterior limb patterning (Dunn et al., 1997). In particular, it has been previously reported that on the C57Bl/6 genetic background Bmp4 heterozygotes display preaxial polydactyly of the right hindlimb with 12% penetrance (Dunn et al., 1997). We have observed similar results with these animals on a mixed 129SV/ C57Bl/6 background (Fig. 4F). But on closer inspection of embryonic day 16.5 Bmp4 heterozygotes, we also observed a small vestigial nubbin on the postaxial side of the right forelimb with approximately 50% penetrance (Fig. 4B). This nubbin typically showed little or no ossification (Fig. 4I, arrowhead) and frequently was no longer visible on newborn animals. When present at birth it typically formed an ectopic nail at this location. When animals were mated to produce animals heterozygous for a disrupted Bmp4 allele and hemizygous for either mutant Ex136 or tm108 Gpc3 allele, the nature and severity of the limb phenotype changed dramatically. Now, 100% of the animals showed some degree of polydactyly with vari-
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
184
Paine-Saunders et al.
FIG. 4. Combined genetic deficiency for Gpc3 and haploinsufficiency for Bmp4 results in postaxial polydactyly. Forelimbs (A–D, I) and hindlimbs (E–H, J, K) of embryonic day 16.5 mice bearing heterozygous mutations for Bmp4 either alone or in combination with hemizygous deficiencies for Gpc3. All animals bearing mutations in Gpc3 alone show normal limb patterning (not shown) as do the majority of animals haploinsufficient for Bmp4 mutations alone (A and E), while 10% of Bmp4 haploinsufficient animals display preaxial polydactyly of the right hindlimb and 50% display a small vestigial postaxial nubbin (B, arrow) of the right forelimb. This nubbin displays little to no ossification as shown by skeletal preparation (I). In contrast, 100% of animals bearing either hemizygous targeted (tm108) or trapped (Ex136) alleles, in combination with a heterozygous mutation for Bmp4, develop postaxial polydactyly. Forelimbs frequently reveal bilateral abnormalities involving more severe duplications of the fifth digit, which include well-ossified triphalangeal duplications (J, asterisk). The hindlimbs of these animals display postaxial polydactyly with 66% penetrance, and all display well-patterned and ossified duplications of the terminal fifth digit (K, asterisk). There were no differences in severity or penetrance of these phenotypes between either allele of Gpc3.
able severity (n ⫽ 36). In the forelimbs, the phenotype demonstrated a clear increase in the severity of the phenotype observed in Bmp4 heterozygous animals alone. Frequently both forelimbs were associated with well-formed bifurcations of the fifth digit (Figs. 4C and 4D). The redundant structures in these animals contained ossified bone in the form of an ectopic triphalangeal duplication branching from the fifth metatarsal bone (Fig. 4J), a defect never seen in Bmp4 heterozygous mutants alone.
In the hindlimb, the change in the polydactylous phenotype was even more dramatic. Here 66% of the animals had postaxial polydactyly, primarily of the right hindlimb, of a form similar to that which is seen in human patients with SGBS. The expressivity of this phenotype was complete, with all animals displaying well-formed branched bifurcations of the fifth digit identical to those shown in Figs. 4G and 4H and identical to those seen in human patients with SGBS. Similar to the forelimb defects, these structures contained ossified bone as well, containing an ectopic triphalangeal digit bifurcating from a common fifth metatarsal bone (Fig. 4K). We found no differences in severity or penetrance of polydactyly whether the targeted tm108 or the trapped Ex136 mutant allele of Gpc3 was combined with heterozygotes for Bmp4. We therefore conclude that these two alleles of Gpc3 are equivalent with respect to limb patterning and, because they represent independently generated mutant alleles of Gpc3, that the phenotypes seen are specific to a loss of GPC3 function in combination with a reduction in BMP4 function. On close examination of the remainder of the skeletons of the animals null for Gpc3 and heterozygous for Bmp4, we determined that there were other significant skeletal abnormalities with apparently complete penetrance. These animals, compared with littermate controls (Figs. 5A and 5B), had markedly abnormal chests with deformed and abnormally fused ribs (Figs. 5C and 5D). In addition, these animals were found to have dual ossification centers throughout the length of their sternum (Fig. 5F). These skeletal abnormalities were not seen in either Gpc3 null or Bmp4 heterozygous animals alone and were apparent only as a result of genetic interaction between these two mutant alleles. Interestingly, these phenotypic features are quite reminiscent of those reported in association with a homozygous deficiency of another Bmp gene, Bmp7 (Luo et al., 1995). However, in contrast to the high penetrance of this phenotype in animals deficient for Gpc3 and haploinsufficient for Bmp4, this feature was found inconsistently in Bmp7-deficient animals. This may be consistent with both the hypothesis that GPC3 is involved in a pathway that impacts multiple BMP family members and the speculation that some structurally related BMPs, such as BMP4 and BMP7, may display certain redundancies in function during development (Dudley and Robertson, 1997; Katagiri et al., 1998). However, we found that a genetic interaction was not evident in all tissues in which GPC3 and BMP4 (or other BMP genes) are coexpressed, suggesting that, like Dally, GPC3 may function in a tissue-dependent manner.
GPC3 Modulation of BMP4 Signaling This study represents the first genetic evidence of a role for glypicans in the control of cellular responses to BMPs during vertebrate development. Here we report that the loss of GPC3 results in both enhanced and synthetic phenotypes in association with BMP4 heterozygous deficiencies. As outlined below, we believe that this genetic interaction
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
185
GPC3 in Limb Patterning and Skeletal Development
FIG. 5. Combined genetic deficiency for Gpc3 and haploinsufficiency for Bmp4 results in other skeletal abnormalities. Animals bearing heterozygous null mutations for Bmp4 alone (A and B), as well as animals bearing hemizygous null mutations for Gpc3 alone (Fig. 2D and data not shown), have normally formed rib cages. In contrast, all animals with combined deficiency for Gpc3 and haploinsufficiency for Bmp4 display markedly abnormal chest walls with multiple rib fusions (C and D) as well as dual ossification centers of the sternum (F).
argues for a direct role for GPC3 in the control of cellular responses to BMP4. Although controversy remains as to whether BMPs function through short-range interactions or over longer ranges through the establishment of morphogenetic gradients, it is clear that the quantity of BMP protein available in the target tissue is critical for achieving the necessary developmental response. This is evident from several studies which have now established, in both vertebrate and invertebrate species, patterning abnormalities associated with heterozygous null mutants for BMPs, which presumably express half the normal levels of these gene products (Dunn et al., 1997; Green, 1968; Irish and Gelbart, 1987; Katagiri et al., 1998; McPherron et al., 1999). Therefore, understanding the molecular mechanisms that modify the strength of BMP signals or the cellular response to these signals in vivo is essential to deciphering the role of BMPs in the regulation of morphogenesis. One strategy that has been particularly effective in this
regard has been the identification of genetic modifiers of BMP function. For example, several mutations that modify the phenotype of weak dpp alleles in Drosophila have been identified (Ferguson and Anderson, 1992a,b; Raftery et al., 1995), and the characterization of these gene products, and their vertebrate homologs, has led rapidly to new information about the downstream signaling pathways activated by BMP and about the secreted proteins that modify BMP activities. Significantly, the Drosophila glypican Dally has been shown to modify Dpp function in at least some postembryonic developmental events in imaginal tissue (Jackson et al., 1997). In most of these tissues, mutations of dally enhance the phenotype associated with reduced dpp expression while suppressing phenotypes associated with ectopic dpp expression, and dally mutants display reduced expression of dpp target genes, suggesting that Dally directly affects cellular responses to Dpp. Similarly, we have discovered that loss of a vertebrate glypican, GPC3, modifies BMP4 function in the mouse limb. Interestingly, not only is this genetic interaction between Gpc3 and Bmp4 mutants in mice particularly strong, but the polydactylous phenotype seen in these animals is identical to the phenotype seen in human patients with Gpc3 deficiency alone. Taken together, the results of our studies and previous work in Drosophila demonstrate that the regulation of cellular responses to BMP is an evolutionarily conserved function of glypicans. There are multiple mechanisms by which Dally and GPC3 could function. For example, they could affect a parallel signaling pathway that alters cellular responses to BMPs. Alternatively, they could more directly modulate the BMP signaling pathway. By analogy with the role of another heparan sulfate proteoglycan, betaglycan, which directly potentiates TGF- binding to its signaling receptor, it has been suggested that Dally could function as a coreceptor for Dpp (Jackson et al., 1997) and GPC3 could have a similar function with respect to BMP4 in the vertebrate limb. Consistent with this model, Drosophila Dpp (Groppe et al., 1998) and vertebrate BMP2, -3, -4, and -7 (Ruppert et al., 1996; Sampath et al., 1987, 1992; Wang et al., 1988) all bind to heparin with high affinity in vitro and therefore likely bind to heparan sulfate in vivo. However, the consequence of this interaction to BMP function remains unclear. For example, although heparin potentiates the activity of BMP2 in a chick limb bud differentiation model in vitro, as would be expected if it were serving a coreceptor function, so does destruction of the N-terminal heparin binding domain of BMP2 or addition of a peptide corresponding to the N-terminal heparin binding domain of BMP2 in these same assays (Ruppert et al., 1996). These latter results can be interpreted as evidence that at least some heparan sulfate-containing proteins actually sequester BMPs in vivo and serve to suppress rather than enhance BMP functions. Another likely mechanism by which GPC3 could directly affect BMP4 activity in the developing limb could be through regulation of the distribution or level of accumu-
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
186
Paine-Saunders et al.
lation of a BMP antagonist, several of which are known to bind to heparin with high affinity. In particular, the heparin binding BMP antagonist Gremlin has recently been shown to mediate the SHH/FGF4 feedback loop in vertebrate limb buds through direct BMP antagonism (Zuniga et al., 1999). Loss of the heparan sulfate proteoglycan GPC3 could result in the increased accumulation and/or a broader distribution of Gremlin protein in the developing limb fields, thus resulting in an increase in signaling through this feedback loop and therefore development of redundant postaxial structures. Further experimentation will be necessary to carefully evaluate each of these hypotheses about the role of GPC3 in BMP signaling and vertebrate limb patterning. This analysis will not be trivial because heparan sulfate proteoglycans have been ascribed with various other functions, both in vitro and in vivo, and GPC3 has not yet been specifically tested for these functions. The potential promiscuity with which GPC3 may carry out the regulation of cellular processes introduces unique challenges to the understanding of its in vivo function.
ACKNOWLEDGMENTS We thank Jonathan Gitlin, Dave Wilson, and Ralph Sanderson for critical review of the manuscript. These studies were supported by National Institutes of Health Grant DK56063 (S.S.) and March of Dimes Birth Defects Foundation Grant 6-FY99-441 (S.S.). S.S. was also supported as both a Spoehrer Scholar in Pediatrics and a Scholar of the Child Health Research Center of Excellence in Developmental Biology at Washington University School of Medicine (HD33688). W.C.S. is a 1998 Searle Scholar.
REFERENCES Behmel, A., Plochl, E., and Rosenkranz, W. (1984). A new X-linked dysplasia gigantism syndrome: Identical with the Simpson dysplasia syndrome? Hum. Genet. 67, 409 – 413. Bellaiche, Y., The, I., and Perrimon, N. (1998). Tout-velu is a Drosophila homologue of the putative tumor suppressor EXT-1 and is needed for Hh diffusion. Nature 394, 85– 88. Binari, R. C., Staveley, B. E., Johnson, W. A., Godavarti, R., Sasisekharan, R., and Manoukian, A. S. (1997). Genetic evidence that heparin-like glycosaminoglycans are involved in wingless signaling. Development 124, 2623–2632. Bullock, S. L., Fletcher, J. M., Beddington, R. S., and Wilson, V. A. (1998). Renal agenesis in mice homozygous for a gene trap mutation in the gene encoding heparan sulfate 2-sulfotransferase. Genes Dev. 12, 1894 –1906. Cano-Gauci, D. F., Song, H. H., Yang, H., McKerlie, C., Choo, B., Shi, W., Pullano, R., Piscione, T. D., Grisaru, S., Soon, S., Sedlackova, L., Tanswell, A. K., Mak, T. W., Yeger, H., Lockwood, G. A., Rosenblum, N. D., and Filmus, J. (1999). Glypican3-deficient mice exhibit developmental overgrowth and some of the abnormalities typical of Simpson–Golabi–Behmel syndrome. J. Cell Biol. 146, 255–264. David, G., Lories, V., Decock, B., Marynen, P., Cassiman, J.-J., and Berghe, H. V. d. (1990). Molecular cloning of a phosphati-
dylinositol-anchored membrane heparan sulfate proteoglycan from human lung fibroblasts. J. Cell Biol. 111, 3165–3176. Dudley, A. T., and Robertson, E. J. (1997). Overlapping expression domains of bone morphogenetic protein family members potentially account for limited tissue defects in BMP7 deficient embryos. Dev. Dyn. 208, 349 –362. Dunn, N. R., Winnier, G. E., Hargett, L. K., Schrick, J. J., Fogo, A. B., and Hogan, B. L. M. (1997). Haploinsufficient phenotypes in BMP4 heterozygous null mice and modification by mutations in Gli3 and Alx4. Dev. Biol. 188, 235–247. Eggenschwiler, J., Ludwig, T., Fisher, P., Leighton, P. A., Tilghman, S. M., and Efstratiadis, A. (1997). Mouse mutant embryos overexpressing IGF-II exhibit phenotypic features of the Beckwith– Wiedemann and Simpson–Golabi–Behmel syndromes. Genes Dev. 11, 3128 –3142. Erickson, R. P. (1996). Mouse models of human genetic disease: Which mouse is more like a man? BioEssays 18, 993–998. Ferguson, E. L., and Anderson, K. V. (1992a). decapentaplegic acts as a morphogen to organize dorsal–ventral pattern in the Drosophila embryo. Cell 71, 451– 461. Ferguson, E. L., and Anderson, K. V. (1992b). Localized enhancement and repression of the activity of the TGF- family member, decapentaplegic, is necessary for dorsal–ventral pattern formation in the Drosophila embryo. Development 114, 583–597. Garganta, C. L., and Bodurtha, J. N. (1992). Report of another family with Simpson–Golabi–Behmel syndrome and a review of the literature. Am. J. Med. Genet. 44, 129 –135. George, E. L., and Hynes, R. O. (1994). Gene targeting and generation of mutant mice for studies of cell– extracellular matrix interactions. Methods Enzymol. 245, 386 – 420. Green, M. C. (1968). Mechanism of the pleiotropic effects of the short ear mutant gene in the mouse. J. Exp. Zool. 167, 129 –150. Groppe, J., Rumpel, K., Economides, A. N., Stahl, N., Sebald, W., and Affolter, M. (1998). Biochemical and biophysical characterization of refolded Drosophila DPP, a homolog of bone morphogeneic proteins 2 and 4. J. Biol. Chem. 273, 29052–29065. Hacker, U., Lin, X., and Perrimon, N. (1997). The Drosophila sugarless gene modulates Wingless signaling and encodes an enzyme involved in polysaccharide biosynthesis. Development 124, 3565–3573. Haerry, T. E., Heslip, T. R., Marsh, J. L., and O’Connor, M. B. (1997). Defects in glucuronate biosynthesis disrupt wingless signaling in Drosophila. Development 124, 3055–3064. Herndon, M. E., and Lander, A. D. (1990). A diverse set of developmentally regulated proteoglycans is expressed in the rat central nervous system. Neuron 4, 949 –961. Hogan, B., Beddington, R., Costantini, F., and Lacey, E. (1994). “Manipulating the Mouse Embryo: A Laboratory Manual,” 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Hughes-Benzie, R. M., Hunter, A. G., Allanson, J. E., and MacKenzie, A. E. (1992). Simpson–Golabi–Behmel syndrome associated with renal dysplasia and embryonal tumors: Localization of the gene to Xqcen– q21. Am. J. Med. Genet. 43, 428 – 435. Irish, V. F., and Gelbart, W. M. (1987). The decapentaplegic gene is required for dorsal–ventral patterning of the Drosophila embryo. Genes Dev. 1, 868 – 879. Jackson, S. M., Nakato, H., Sugiura, M., Jannuzi, A., Oakes, R., Kaluza, V., Golden, C., and Selleck, S. B. (1997). Division abnormally delayed, a Drosophila glypican, controls cellular responses to the TGF- related morphogen, Dpp. Development 124, 4113– 4120.
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.
187
GPC3 in Limb Patterning and Skeletal Development
Katagiri, T., Boorla, S., Frendo, J. L., Hogan, B. L., and Karsenty, G. (1998). Skeletal abnormalities in doubly heterozygous Bmp4 and Bmp7 mice. Dev. Genet. 22, 340 –348. Lin, X., and Perrimon, N. (1999). Dally cooperates with Drosophila Frizzled 2 to transduce Wingless signalling. Nature 400, 281–284. Lind, T., Tufaro, F., McCormick, C., Lindahl, U., and Lidholt, K. (1998). The putative tumor suppressors EXT1 and EXT2 are glycosyltransferases required for the biosynthesis of heparan sulfate. J. Biol. Chem. 273, 26265–26268. Luo, G., Hofmann, C., Bronckers, A. L. J. J., Sohocki, M., Bradley, A., and Karsenty, G. (1995). BMP-7 is an inducer of nephrogenesis, and is also required for eye development and skeletal patterning. Genes Dev. 9, 2808 –2820. McCormick, C., Leduc, Y., Martindale, D., Mattison, K., Esford, L. E., Dyer, A. P., and Tufaro, F. (1998). The putative tumor suppressor EXT1 alters the expression of cell-surface heparan sulfate. Nat. Genet. 19, 158 –161. McLeod, M. J. (1980). Differential staining of cartilage and bone in whole mouse fetuses by alcian blue and alizarin red S. Teratology 22, 299 –301. McPherron, A. C., Lawler, A. M., and Lee, S. J. (1999). Regulation of anterior/posterior patterning of the axial skeleton by growth/ differentiation factor 11. Nat. Genet. 22, 260 –264. Paine-Saunders, S., Viviano, B. L., and Saunders, S. (1999). GPC6, a novel member of the glypican gene family, encodes a product structurally related to GPC4 and is colocalized with GPC5 on human chromosome 13. Genomics 57, 455– 458. Pellegrini, M., Pilia, G., Pantano, S., Lucchini, F., Uda, M., Fumi, M., Cao, A., Schessinger, D., and Forabosco, A. (1998). Gpc3 expression correlates with the phenotype of the Simpson– Golabi–Behmel syndrome. Dev. Dyn. 213, 431– 439. Pilia, G., Hughes-Benzie, R. M., MacKenzie, A., Baybayan, P., Chen, E. Y., Huber, R., Neri, G., Cao, A., Foarabosco, A., and Schlessinger, D. (1996). Mutations in GPC3, a glypican gene, cause the Simpson–Golabi–Behmel overgrowth syndrome. Nature Genet. 12, 241–247. Raftery, L. A., Twombly, V., Wharton, K., and Gelbart, W. M. (1995). Genetic screens to identify elements of the decapentaplegic signaling pathway in Drosophila. Genetics 139, 214 –254. Rapraeger, A., Krufka, A., and Olwin, B. B. (1991). Requirement of heparan sulfate for bFGF-mediated fibroblast growth and myoblast differentiation. Science 252, 1705–1708. Reichsman, F., Smith, L., and Cumberledge, S. (1996). Glycosaminoglycans can modulate extracellular localization of the wingless protein and promote signal transduction. J. Cell Biol. 135, 819 – 827. Ruppert, R., Hoffmann, E., and Sebald, W. (1996). Human bone morphogenetic protein 2 contains a heparin-binding site which modifies its biological activity. Eur. J. Biochem. 237, 295–302. Sampath, T. K., Maliakal, J. C., Hauschka, P. V., Jones, W. K., Sasak, H., Tucker, R. F., White, K. H., Coughlin, J. E., Tucker, M. M., Pang, R. H. L., Corbett, C., Ozkaynak, E., Oppermann, H.,
and Rueger, D. C. (1992). Recombinant human osteogenic protein-1 (hOP-1) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates osteoblast proliferation and differentiation in vitro. J. Biol. Chem. 267, 20352–20362. Sampath, T. K., Muthukumaran, N., and Reddi, A. H. (1987). Isolation of osteogenin, an extracellular matrix-associated, boneinductive protein, by heparin affinity chromatography. Proc. Natl. Acad. Sci. USA 84, 7109 –7113. Saunders, S., Jalkanen, M., O’Farrell, S., and Bernfield, M. (1989). Molecular cloning of syndecan, an integral membrane proteoglycan. J. Cell Biol. 108, 1547–1556. Skarnes, W. C., Moss, J. E., Hurtley, S. M., and Beddington, R. S. (1995). Capturing genes encoding membrane and secreted proteins important for mouse development. Proc. Natl. Acad. Sci. USA 92, 6592– 6596. Song, H. H., Shi, W., and Filmus, J. (1997). OCI-5/rat glypican-3 binds to fibroblast growth factor-2 but not to insulin-like growth factor-2. J. Biol. Chem. 272, 7574 –7577. The, I., Bellaiche, Y., and Perrimon, N. (1999). Hedgehog movement is regulated through tout velu-dependent synthesis of a heparan sulfate proteoglycan. Mol. Cell 4, 633– 639. Townley, D. J., Avery, B. J., Rosen, B., and Skarnes, W. C. (1997). Rapid sequence analysis of gene trap integrations to generate a resource of insertional mutations in mice. Genome Res. 7, 293–298. Tsuda, M., Kamimura, K., Nakato, H., Archer, M., Staatz, W., Fox, B., Humphrey, M., Olson, S., Futch, T., Kaluza, V., Siegfried, E., Stam, L., and Selleck, S. B. (1999). The cell-surface proteoglycan Dally regulates Wingless signalling in Drosophila. Nature 400, 276 –280. Wang, E. A., Rosen, V., Cordes, P., Hewick, R. M., Kriz, M. J., Luxenberg, D. P., Sibley, B. S., and Wozney, J. M. (1988). Purification and characterization of other distinct bone-inducing factors. Proc. Natl. Acad. Sci. USA 85, 9484 –9488. Winnier, G., Blessing, M., Labosky, P. A., and Hogan, B. L. M. (1995). Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev. 9, 2105– 2116. Winter, R. M. (1996). Analysing human developmental abnormalities. BioEssays 18, 965–971. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64, 841– 848. Zuniga, A., Haramis, A. G., McMahon, A. P., and Zeller, R. (1999). Signal relay by BMP antagonism controls the SHH/FGF4 feedback loop in vertebrate limb buds. Nature 401, 598 – 602. Received for publication April 12, 2000 Revised May 26, 2000 Accepted June 16, 2000
Copyright © 2000 by Academic Press. All rights of reproduction in any form reserved.