The extracellular matrix protein WARP is a novel component of a distinct subset of basement membranes

Available online at www.sciencedirect.com

Matrix Biology 27 (2008) 295 – 305 www.elsevier.com/locate/matbio

The extracellular matrix protein WARP is a novel component of a distinct subset of basement membranes Justin M. Allen a , Bent Brachvogel a,c , Peter G. Farlie a , Jamie Fitzgerald a,d , John F. Bateman a,b,⁎ a

Murdoch Childrens Research Institute, University of Melbourne, Royal Children's Hospital, Parkville, Victoria, Australia b Department of Paediatrics, University of Melbourne, Royal Children's Hospital, Parkville, Victoria, Australia c Center for Biochemistry, Center for Molecular Medicine, Medical Faculty, University of Cologne, Cologne, Germany d Department of Orthopaedics and Rehabilitation, Oregon Health and Science University, Portland, Oregon, USA Received 11 December 2007; accepted 8 January 2008

Abstract WARP is a recently described member of the von Willebrand factor A domain superfamily of extracellular matrix proteins, and is encoded by the Vwa1 gene. We have previously shown that WARP is a multimeric component of the chondrocyte pericellular matrix in articular cartilage and intervertebral disc, where it interacts with the basement membrane heparan sulfate proteoglycan perlecan. However, the tissue-specific expression of WARP in non-cartilaginous tissues and its localization in the extracellular matrix of other perlecan-containing tissues have not been analyzed in detail. To visualize WARP-expressing cells, we generated a reporter gene knock-in mouse by targeted replacement of the Vwa1 gene with βgalactosidase. Analysis of reporter gene expression and WARP protein localization by immunostaining demonstrates that WARP is a component of a limited number of distinct basement membranes. WARP is expressed in the vasculature of neural tissues and in basement membrane structures of the peripheral nervous system. Furthermore, WARP is also expressed in the apical ectodermal ridge of developing limb buds, and in skeletal and cardiac muscle. These findings are the first evidence for WARP expression in non-cartilaginous tissues, and the identification of WARP as a component of a limited range of specialized basement membranes provides further evidence for the heterogeneous composition of basement membranes between different tissues. © 2008 Elsevier B.V./International Society of Matrix Biology. All rights reserved. Keywords: WARP; Vwa1; Extracellular matrix; Basement membrane; Reporter gene

1. Introduction Extracellular matrices (ECM) are composed of dynamic and highly ordered structural networks of collagens, proteoglycans and noncollagenous glycoproteins. The temporal and spatial regulation of ECM formation contributes to the diverse composiAbbreviations: WARP, von Willebrand A-domain related protein; Vwa1, von Willebrand A-domain-containing 1; ECM, extracellular matrix; VWA, von Willebrand A; NLS, nuclear localization sequence; PECAM, platelet/endothelial cell adhesion molecule; GuHCl, guanidine hydrochloride; BSA, bovine serum albumin; E(n), embryonic day (n); AER, apical ectodermal ridge; CNS, central nervous system; BMP, bone morphogenetic protein; FGF, fibroblast growth factor. ⁎ Corresponding author. Murdoch Childrens Research Institute, Royal Children's Hospital, Parkville, Victoria 3052, Australia. Fax: +61 3 8341 6429. E-mail address: [email protected] (J.F. Bateman).

tion and biological properties of different connective tissues and has fundamental effects on cellular phenotypes as well as tissue architecture and function (Nelson and Bissell, 2006). In addition to performing structural roles, ECM components have functions as diverse as mediating collagen fibrillar architecture, bridging between macromolecular networks, binding growth factors and influencing cell differentiation, migration or adhesion, or providing linkages between cells and the ECM (Heinegard et al., 2002). The modular structure of many ECM proteins contributes to the structural and functional diversity of connective tissues. The von Willebrand factor type A domain (VWA domain) is a component of many extracellular proteins involved in supramolecular structures and protein–protein interactions (reviewed by Whittaker and Hynes, 2002). WARP (von Willebrand A domain related protein), the product encoded by the Vwa1 (von

0945-053X/$ - see front matter © 2008 Elsevier B.V./International Society of Matrix Biology. All rights reserved. doi:10.1016/j.matbio.2008.01.005

296

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

Willebrand factor A domain-containing 1) gene, is a recently described VWA domain-containing ECM protein with a restricted expression pattern in cartilage tissues. The WARP protein comprises a single N-terminal VWA-domain-containing a putative metal ion-dependent adhesion site (MIDAS) motif, two fibronectin type III repeats, and a unique C-terminal domain (Fitzgerald et al., 2002). In embryonic skeletal tissues, WARP is expressed in the presumptive articular cartilage prior to joint cavitation, and subsequently in articular cartilage as well as in meniscus and intervertebral disc but was not detected in cartilage tissues that undergo endochondral ossification (Allen et al., 2006). Within these permanent cartilage structures, WARP is highly restricted to the pericellular matrix of the chondrocytes where it colocalizes with the basement membrane heparan sulfate proteoglycan perlecan. We have demonstrated that WARP forms disulfide-bonded multimers, and that it interacts with high affinity to discrete domains of the perlecan core protein as well as to its heparan sulfate chains (Allen et al., 2006). Perlecan is a large heparan sulfate proteoglycan and is one of the major basement membrane components along with laminins, nidogens and collagen IV (Yurchenco et al., 2004). Basement membranes are highly specialized sheet-like ECM structures that are formed early in development and are essential for tissue development and homeostasis. They impart mechanical stability, and have important roles in establishing tissue borders and barriers, as well as in compartmentalizing tissues and maintaining their function (Sasaki et al., 2004). In addition to structural roles, basement membranes can also profoundly influence cell adhesion, morphology and phenotype through direct interaction with cellular receptors and by influencing the bioavailability of growth factors and signaling molecules (Hallmann et al., 2005). Perlecan exerts a structural role in the basement membrane through its interactions with macromolecules including collagen IV, laminins, nidogens-1 and -2, fibronectin and heparin, and with receptors such as integrins and dystroglycans (reviewed by Knox and Whitelock, 2006). In addition, by

binding to signaling molecules and growth factors such as BMPs and FGFs, as well as FGF receptors via its heparan sulfate chains, perlecan can regulate their bioavailability and modulate subsequent signaling events (Knox et al., 2002; Smith et al., 2007). Analysis of perlecan-null mice has revealed that perlecan is not indispensable for the formation of basement membranes, but it has an essential influence on the structural integrity of basement membranes in response to mechanical stress (Arikawa-Hirasawa et al., 1999; Costell et al., 1999). The high affinity interaction of WARP with perlecan in the chondrocyte pericellular matrix suggests that WARP may also contribute to the structural integrity of perlecan-containing connective tissues. However the distribution of WARP in noncartilaginous tissues, in particular concerning the localization in basement membranes of different origin, has not been determined. We have generated a β-galactosidase reporter gene knock-in mouse targeting one allele of the Vwa1 locus to enable in situ visualization of WARP-expressing cells in mouse tissues. Immunolocalization of WARP protein in wild type mice was used to confirm the results of the reporter gene expression studies. We report that WARP is expressed in the vasculature of neural tissues, in basement membrane structures of the peripheral nervous system, in the apical ectodermal ridge of developing limb buds, and in skeletal and cardiac muscle. These data suggest that WARP is a novel component of a distinct subset of basement membrane tissues and provide further evidence for the biochemical heterogeneity of basement membranes between different tissues. 2. Results 2.1. Generation of Vwa1+/βgal mice To determine the expression pattern of WARP in situ, a mouse strain was generated by replacing an intact allele of the Vwa1 gene with a β-galactosidase reporter gene cassette (Fig. 1). For enhancing the sensitivity and clarity of the staining

Fig. 1. Targeted replacement of the Vwa1 coding sequence with the lacZ reporter gene. (A) Schematic diagram of the targeting construct containing the lacZ gene preceded by a nuclear localization sequence (nls-LacZ) and a neomycin selection cassette (neo) in frame with the Vwa1 coding sequence. The regions amplified by genotyping primer pairs are indicated by arrows. (B) The domain structure of the WARP protein, consisting of an N-terminal von Willebrand type A domain, two fibronectin type III repeats and a unique C-terminal sequence with no homology to any protein module. (C) Genotyping of Vwa1+/βgal mice by genomic PCR with allele-specific primers.

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

297

we integrated a nuclear localization sequence (NLS) to the 5′ end of the β-galactosidase expression gene. The targeted allele also a contained a floxed neomycin selection cassette for previous selection of positive clones in cell culture. Heterozygous Vwa1+/βgal mice were viable and fertile and did not have any apparent abnormalities. Wild type and heterozygous littermates of intercrosses from heterozygous mice were used for all X-gal staining experiments. Their genotypes were confirmed for each experiment by PCR-genotyping of genomic DNA to detect wild type (196 bp) and mutant alleles (367 bp) (Fig. 1). 2.2. WARP expression in embryonic development Whole mount X-gal staining of Vwa1+/βgal embryos was used to determine the sites of WARP expression in embryonic development. WARP-expressing cells were detected throughout the vasculature of the developing central nervous system (CNS) from E9.5 onwards, whereas no staining was found in the remaining vasculature of the embryo (Fig. 2A–C). Strong expression of WARP was also detected in the distal tip of the of the developing limb bud at E9.5 (Fig. 2A), and this expression was highly restricted to the apical ectodermal ridge (AER) at E10.5 and E11.5 (Fig. 2B–C). Strong X-gal staining was detected in the cranial ganglia at E9.5, and also in the dorsal root ganglia from E10.5 onwards (Fig. 2). From E11.5, WARP is expressed in the myotomes, transient segmented structures containing precursors of the trunk musculature. Consistent with this expression in early differentiating muscle, β-galactosidase activity was also present in the pre-muscle mass in the forelimb bud (Fig. 2C). To verify the reporter gene results, and to further investigate the role of WARP in these tissues, WARP expression was analyzed in detail during post-natal development of the tissues described above and WARP protein localization was determined by immunostaining in Vwa1+/+ wild type mice. 2.3. WARP is expressed in the vasculature of the central nervous system The vasculature of the CNS is a prominent site of WARP expression during embryogenesis. Whole mount X-gal staining of a 7-week-old Vwa1 +/βgal mouse brain clearly demonstrates that WARP-expression also persists in vasculature of adult brain meninges, as cells aligned along the blood vessels of the meninges were clearly stained (Fig. 3A–B). Furthermore, sections of a 7-week-old Vwa1 +/βgal mouse brain stained for β-galactosidase activity showed that WARPexpressing cells were detected along blood vessels throughout the brain in addition to those demonstrated in the meninges (Fig. 3C). Immunostaining using the anti-WARP antisera was used to visualize WARP protein distribution and to confirm the X-gal staining data. WARP is strongly expressed in blood vessels in sections of 7-week-old mouse brain and colocalization with perlecan by confocal microscopy indicates that WARP is a component of the vascular basement membrane (Fig. 3D–F).

Fig. 2. WARP-expressing cells visualized by whole mount X-gal staining in Vwa1+/βgal embryos. The vasculature of the developing brain and spinal cord (arrowheads) was strongly stained at E9.5 (A), E10.5 (B) and E11.5 (C). Expression in the forelimb bud (FL) at E9.5 (A) becomes restricted to the apical ectodermal ridge (AER) at E10.5 and E11.5 (B–C). Strong staining was detected in the cranial ganglia (CG) from E9.5 (A) and in the developing dorsal root ganglia (DRG) from E10.5 onwards (B–C). WARP-expressing cells in the developing musculature were detected in the myotomes (M) and the pre-muscle mass of the forelimb (asterisk) at E11.5 (C). Bar; 1 mm.

2.4. WARP is a basement membrane component in peripheral nerves Whole mount X-gal staining demonstrated that WARP is expressed in cranial and dorsal root ganglia during embryonic development. The site of WARP expression in the peripheral nervous system was therefore examined by immunohistochemistry in tissues of post-natal mice. In coronal sections adjacent to the spine of a 2-week-old mouse, WARP protein was detected throughout the fascicles of the dorsal root nerve projecting from the dorsal root ganglia (Fig. 4A, C). In the adjacent spinal facet joint, WARP immunoreactivity is also detected in the articular cartilage of the rib tubercle and the transverse process of the vertebrae at similar intensity, suggesting that WARP is expressed at similar levels in the nerve and in articular cartilage (Fig. 4A, C). The distribution of WARP protein within nerve tissue was examined by immunofluorescence staining for WARP in

298

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

Fig. 3. WARP expression in the vasculature of the central nervous system. WARP-expressing cells were identified by X-gal staining of a 7-week-old Vwa1+/βgal mouse brain in whole mount (A–B) and in tissue sections (C). Immunofluorescence of 7-week-old wild type mouse brain with WARP antibodies (D) and co-localization with antibodies to the basement membrane heparan sulfate proteoglycan perlecan (E and merge F) demonstrate that WARP is a component of the basement membrane in CNS blood vessels. Bar; 2.5 mm (A), 0.5 mm (B), 50 μm (C–F).

transverse sections of the sciatic nerve of a 6-week-old mouse. WARP was immunolocalized to fine filamentous structures surrounding the nerve fibers identified by the presence of

neuron-specific β-III tubulin (Fig. 4E). Within peripheral nerves, individual nerve fibers are surrounded by a basement membrane synthesized by Schwann cells, which are in turn

Fig. 4. WARP expression in peripheral nerves. WARP immunostaining was detected in nerve fascicles extending from the dorsal root ganglia (DRG) as well as in articular cartilage structures (arrowheads) of the ribs (Ri) and vertebral bodies (VB) in coronal sections of a 2-week-old mouse (A, C). No immunostaining was detected in sections incubated with preimmune serum as a negative control (B, D). Immunofluorescent co-staining of WARP with Tuj I antibody for axonal β-III tubulin (E) or with laminin (F–H) demonstrate that WARP is a component of the basement membrane surrounding nerve fibers. WARP is absent from the epineurium (arrowheads) where laminin is strongly expressed (F–H). X-gal staining of 4-day-old mouse sciatic nerve demonstrates WARP expression throughout the nerve (I). In the 6-week-old nerve, X-gal staining reveals expression of WARP by cells of the vasculature within the nerve (J), confirmed in transverse sections by combined X-gal staining and immunostaining for the endothelial cell marker PECAM in Vwa1+/βgal nerve (K), and by immunofluorescent co-staining for WARP and PECAM in wild type nerve (L). Bar; 400 μm (A, B), 200 μm (C, D), 10 μm (E), 20 μm (F–H), 100 μm (I, L), 50 μm (J, K).

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

encased within a loose connective tissue matrix termed the endoneurium (Parmantier et al., 1999). The entire nerve is ensheathed in a collagenous perineurium structure, and nerves containing several nerve fascicles are surrounded by a further connective tissue structure called the epineurium (Parmantier et al., 1999), although a separate epineurium and perineurium is difficult to distinguish in rodent sciatic nerve (Layton and Sastry, 2004). Co-localization of WARP with laminin demonstrates that WARP is a component of the endoneurial basement membrane surrounding Schwann cells (Fig. 4F–H). However the absence of WARP immunoreactivity from the outer epineurium layer where laminin is strongly expressed suggests an axon-specific function. To visualize WARP-expressing cells, the sciatic nerve of a 4day-old Vwa1+/βgal mouse was stained with X-gal, revealing that cells throughout the nerve express WARP (Fig. 4I). Previous analysis of sciatic nerves from newborn rodents have shown that Schwann cells comprise over 90% of cells within the nerve during early post-natal development (Wanner et al., 2006), strongly suggesting that the β-galactosidase positive cells we detected throughout the nerve are Schwann cells. However, when staining sciatic nerve tissue from 6-weekold Vwa1+/βgal mice, WARP expression could only be detected by a limited number of cells at this timepoint (Fig. 4J). The observation that WARP-expressing cells were often found in linear alignment suggested that WARP may be expressed by blood vessels within the nerve, consistent with its expression in the CNS vasculature. This was confirmed by combined X-gal staining and immunohistochemistry for CD31/PECAM (platelet/endothelial cell adhesion molecule), demonstrating that WARP-expressing cells are part of the vasa nervorum, the microvasculature within peripheral nerves (Fig. 4K). This finding was confirmed by immunofluorescence co-staining of WARP protein with PECAM in 6-week-old sciatic nerve (Fig. 4L), demonstrating that WARP is present in the basement membrane surrounding blood vessels. However, WARP protein is still present in the endoneurial basement membrane in 6-

299

week-old mice even though WARP is only being synthesized by cells associated with the vasculature (Fig. 4K), rather than by Schwann cells that are throughout the nerve. Taken together, these findings suggest that in addition to WARP expression in blood vessels within the nerve, WARP is expressed by Schwann cells during early stages of nerve development and that although the gene is not expressed in the Schwann cells of adult nerves, the protein persists in the nerve basement membrane. 2.5. Characterization of WARP in nerve tissue The expression of WARP in neural tissues was further verified by Northern blot and immunoblot analysis. A single transcript at 2.3 kb, corresponding to the predicted size of the WARP transcript, was detected by Northern blot hybridization of RNA isolated from sciatic nerve, spinal cord and brain of 2week-old mice (Fig. 5A), with strongest expression in nerve. WARP protein was detected by immunoblot analysis of 2-weekold mouse sciatic nerve homogenates solubilized in 4 M GuHCl. A single band corresponding to the WARP monomer of 45 kDa was detected when resolving the samples under reducing conditions (Fig. 5B). Our reporter gene analysis of WARP expression in nerve demonstrated that WARP is strongly expressed by Schwann cells in neonate nerves (Fig. 4I) but is undetectable in adult Schwann cells (Fig. 4K). Similarly, expression studies of laminin chains and nidogens, which are highly abundant molecules in the Schwann cell basal lamina, have demonstrated that the mRNAs encoding these molecules are expressed highly in developing nerves but are downregulated to barely detectable levels in adult nerve (Kucherer-Ehret et al., 1990; Wallquist et al., 2002; Lee et al., 2007). Furthermore, previous studies analyzing laminin protein distribution in nerve demonstrated that a high proportion of laminin in developing nerve is present in the interstitial matrix of the endoneurium and is readily extracted with physiological buffers, whereas laminin in adult nerve is highly restricted to the Schwann cell basal lamina and is only extracted from nerve tissue

Fig. 5. WARP expression and sequential extraction from nerve tissues. Northern blot analysis of WARP mRNA expression in 2-week-old sciatic nerve, spinal cord and brain, and reprobed with β-actin as a loading control (A). The migration of the 18S and 28S ribosomal RNA bands are indicated. Silver stain and immunoblot analysis of WARP protein in 2-week-old sciatic nerve GuHCl extract (B). Sequential extraction of sciatic nerve with 0.5% Triton-X100 (E1), 500 mM NaCl (E2), 4 M GuHCl (E3) and gel loading buffer containing 2% SDS and 2 M urea (E4) from 4-day-old, 4-week-old and 20-week-old mice (C). Immunoblot results are representative findings from two independent experiments.

300

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

using denaturing solutions (Kucherer-Ehret et al., 1990). Our previous studies on cartilage revealed that WARP requires denaturation with 4 M GuHCl for extraction from cartilage homogenates (Allen et al., 2006). Therefore we sought to determine whether WARP is also incorporated into the insoluble ECM of nerves during development. To investigate the solubility and the degree of integration of WARP into the nerve ECM, sciatic nerve tissues from mice of different ages were sequentially extracted with progressively disruptive buffers and the amount of WARP in the extracted material analyzed by immunoblotting (Fig. 5C). In 4-day-old mouse nerve, a small proportion of WARP was extracted with neutral buffer containing 0.5% Triton-X100, with greater amounts of WARP protein released by extraction with 0.5 M NaCl and subsequently with 4 M GuHCl. In 4-week-old nerve, virtually no WARP is extracted with 0.5% Triton-X100, a small amount with high salt buffer and most of the WARP protein is extracted with GuHCl. A further smaller pool requires extraction with urea and SDS. However in 20-week-old mouse nerve, both 0.5% Triton-X100 and high salt buffer failed to extract WARP, and the vast majority of WARP is extracted only with GuHCl or urea and SDS. In the 20-week-old mouse nerve the proportion in the most insoluble fraction, the urea and SDS extract, is greater (Fig. 5C). These data suggest that, like laminin chains, WARP becomes increasingly incorporated into the insoluble Schwann cell ECM during post-natal development of nerve tissue. 2.6. WARP has a restricted expression pattern in developing skeletal tissues The first site of WARP expression in developing skeletal tissues is in the limb bud outgrowth at E9.5 (Fig. 2A). From E10.5 onwards as the apical ectodermal ridge (AER) forms its characteristic thickened ridge structure, WARP expression is highly restricted to the cells of the AER (Fig. 6A). Immunolocalization reveals that WARP is present in the basement membrane directly underlying the cells of the AER (Fig. 6B). In addition to expression in the AER, WARP expression later in limb development has previously been described in the joint interzone prior to joint cavitation, and in articular cartilage and fibrocartilage structures (Allen et al., 2006). Consistent with these findings, β-galactosidase activity was detected in the developing articular cartilage, meniscus and intervertebral disc in embryonic and adult tissues (Fig. 6C–E). X-gal staining of the knee joint of E16.5 Vwa1+/βgal mouse embryos revealed that although WARP is highly expressed in developing articular cartilage, resting chondrocytes also weakly express WARP, with expression decreasing throughout the femoral and tibial epiphyses as the chondrocytes undergo maturation (Fig. 6C). Immunolocalization confirmed previous findings that WARP is restricted to the chondrocyte pericellular matrix in articular cartilage (Fig. 6F). 2.7. WARP is expressed in muscle tissues Whole mount X-gal staining of E11.5 Vwa1+/βgal embryos revealed reporter gene expression in the myotomes and in the

Fig. 6. WARP expression in skeletal development. Strong X-gal staining was detected in the AER of the forelimb bud (FL) and hindlimb bud (HL) in E11.5 Vwa1+/βgal embryos (A). In sections through the hindlimb bud (HL), WARP immunolabeling is restricted to the basement membrane (arrowheads) underlying the AER (arrow) (B). In skeletal tissues from E16.5 Vwa1+/βgal embryos, β-galactosidase activity was detected in articular cartilage and fibrocartilage structures. Strong staining was detected in the meniscus (Me) and in developing articular cartilage of the femur (Fe) and tibia (Ti), with weaker staining also present throughout the underlying resting zone chondrocytes (C). In the spine, WARP-expressing cells were detected in the articular cartilage surfaces (arrowheads) of the ribs (Ri) and vertebral bodies (VB), and in the annulus fibrosus (AF) of the intervertebral disc (D). In tissues from 7-week-old mice, WARP expression was detected in meniscus (Me) and in femoral (Fe) and tibial (Ti) articular cartilages by X-gal staining (E), and WARP protein was shown to have a restricted pericellular localization by immunofluorescence staining (F). Bar; 0.5 mm (A), 30 μm (B), 200 μm (C, D), 100 μm (E), 20 μm (F).

pre-muscle mass of the limb buds (Fig. 2C), suggesting that WARP is a component of developing muscle tissues. We therefore examined WARP expression in adult muscle tissue in sections of the soleus muscle and the myocardium of the heart from post-natal mice. The connective tissue structure of muscle shares a broadly similar hierarchy of tissue organization to that of peripheral nerves. Individual muscle fibers surrounded by endomysium, consisting of basement membrane and collagens, are bundled together as fascicles within a loose perimysium connective tissue matrix, and several fascicles forming a muscle group are ensheathed by a dense, collagenous epimysium layer (Jarvinen et al., 2002). X-gal staining revealed that WARP is expressed by myocytes in skeletal muscle (Fig. 7A). WARP

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

301

Fig. 7. WARP expression in muscle tissues. X-gal staining of WARP-expressing myocytes in skeletal muscle sections from a 12-week-old Vwa1+/βgal mouse (A). Immunostaining for WARP in 20-week-old skeletal muscle (B) and cardiac muscle (C) is detected in the endomysium surrounding muscle fibers and in capillaries (arrowheads) supplying the musculature. The localization of WARP to the endomysial basement membrane was confirmed by co-staining with perlecan (D–F), and capillaries (arrowheads) were identified by PECAM immunolabeling (G–I). Bar; 50 μm.

protein was detected in the endomysium closely apposed to the muscle fibers in both skeletal and cardiac muscle (Fig. 7B, C), although was not detected in the perimysium or epimysium (not shown). Co-localization with perlecan demonstrates that WARP is a component of the basement membrane within the endomysium (Fig. 7D–F). In addition, WARP immunostaining was present in capillaries throughout skeletal and cardiac muscle (Fig. 7B, C), confirmed by PECAM co-staining (Fig. 7G–I). 3. Conclusions Utilizing a reporter gene mouse model as well as immunostaining, we have demonstrated that WARP is a novel component of basement membranes in several tissues. Our previous findings described a restricted expression pattern for WARP in skeletal tissues (Allen et al., 2006), and the results of the present study are the first to demonstrate WARP expression in non-cartilaginous tissues. The expression of WARP in a subset of basement membrane-containing tissues provides further evidence for the heterogeneous composition of basement membranes in different tissues. The vasculature of the central nervous system is a prominent site of WARP expression from early stages of embryonic development, with expression detected from E9.5 onwards. Notably, expression could not be detected in the vasculature of

other tissues during early development, which may suggest that WARP has specific functions in establishing or maintaining the blood–brain barrier. Consistent with this hypothesis, WARP expression was also detected in microvessels within the endoneurium of peripheral nerves which, like the central nervous system, are characterized by a blood-neural barrier (Allt and Lawrenson, 2000). Interestingly, the multiplexin collagen XValso localizes to the basement membrane of blood vessels and other tissues, and is only expressed in the vasculature within the brain during embryonic development but is downregulated post-natally (Muona et al., 2002). This suggests that collagen XV may function in establishing rather than maintaining the blood–brain barrier, whereas WARP may have an important role in maintaining the blood–brain barrier throughout life. Analysis of WARP expression in tissues from post-natal mice revealed that vascular expression of WARP is not restricted to the nervous system. WARP was also detected in continuous capillaries within the musculature but was absent from other blood vessels such as sinusoidal capillaries in the liver (data not shown) indicating that WARP is only expressed in a subtype of blood vessels. Analysis of WARP protein, mRNA and reporter gene expression demonstrates that WARP is strongly expressed in the peripheral nervous system during embryonic development as well as in early post-natal stages during the processes of myelination and Schwann cell basement membrane assembly. However X-gal staining revealed that, like other components of

302

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

the Schwann cell ECM, WARP expression was found to be downregulated at the transcriptional level in adulthood despite the protein persisting in the tissue. Previous studies have shown that laminin α2, α4, β1, β2 and γ1 chains were all expressed at high levels in newborn rat sciatic nerve, but their levels decreased with developmental age to below the level of detection in adult nerve tissues (Wallquist et al., 2002). Similarly, expression of the basement membrane glycoproteins nidogen1 and -2 were undetectable in adult rat sciatic nerve by RT-PCR or in situ hybridization despite the presence of both proteins in the nerve ECM (Lee et al., 2007). However, expression of laminin α2, α4, β1 and γ1 chains and of both nidogens have been shown to be upregulated in animal models of nerve injury and regeneration (Wallquist et al., 2002; Lee et al., 2007). These findings suggest that the turnover of matrix components in the adult peripheral nerve is very low, so that after the initial processes of myelination and matrix assembly, the synthesis of ECM molecules is downregulated except in response to tissue remodeling and repair. The critical requirement for interactions between Schwann cells and the ECM in the myelination process has been well established by mouse models for laminin chains (Chen and Strickland, 2003) and for integrin and dystroglycan receptors (Feltri et al., 2002; Previtali et al., 2003; Saito et al., 2003). Recent data has suggested that the cell surface heparan sulfate proteoglycan glypican-1 also has a fundamental role in myelination, with siRNA knockdown of either glypican-1 or its ligand α4(V) collagen severely impeding myelination in a Schwann cell-DRG neuron culture system (Chernousov et al., 2006). The specific interaction of WARP with the heparan sulfate chains of perlecan (Allen et al., 2006) suggests that WARP could also potentially interact with cell surface heparan sulfate proteoglycans including glypican-1, although this has not yet been directly examined. Components of the ECM have complex and diverse effects in promoting and/or inhibiting growth, migration and differentiation of Schwann cells during development and tissue repair. Further characterization of the function of WARP in peripheral nerves may reveal important roles in the development and regeneration of nerve tissue. In skeletal tissues, WARP has two distinct and nonoverlapping sites of expression. Consistent with previous findings in our laboratory, WARP is expressed in developing articular cartilage during joint development, and its expression is restricted to articular cartilage and to fibrocartilage tissues (Allen et al., 2006). In the present study, weak X-gal staining was also detected in underlying epiphyseal cartilage in embryonic skeletal tissues, which may suggest that a gradient of WARP expression exists, decreasing as chondrocytes undergo differentiation and maturation, and reflects the enhanced sensitivity of β-galactosidase reporter gene activity compared to other methods of gene expression analysis. Several basement membrane components have been shown to localize to the chondrocyte pericellular matrix, suggesting that this structure may be the functional equivalent of a basement membrane in cartilage tissues (Kvist et al., 2007). The localization of WARP to the pericellular matrix as well as several basement membrane tissues provides further support for this hypothesis.

In addition to the role of WARP in permanent cartilage structures, WARP was also expressed in the limb bud during early stages of limb formation. WARP-expressing cells were highly restricted to the AER, although WARP protein was deposited as part of the basement membrane underlying the AER. This basement membrane has been shown recently to have a critical role in maintaining the structure of the AER as well as influencing gradients of growth factors important for limb development. The Japanese chick wingless mutant, which displays variable truncation of limbs, has a disrupted AER basement membrane and altered FGF8 signaling (Yamaguchi et al., 2004). Mice null for the basement membrane glycoproteins nidogen-1 and -2, and mice with compound deletions of the laminin-binding integrins α3 and α6, also displayed a disrupted AER basement membrane and resulting defects in limb development (De Arcangelis et al., 1999; Bose et al., 2006). Importantly, analysis of the nidogen double knockout mice revealed that, in addition to structural abnormalities in the formation of the AER, disruption of the AER basement membrane also resulted in an altered distribution of signaling molecules in the underlying limb bud which contributed to the syndactyly in these mice (Bose et al., 2006). The highly restricted expression of WARP by AER cells, and its localization to the AER basement membrane, suggests that WARP may also have a critical role in early limb development. Another prominent basement membrane structure we have demonstrated to contain WARP is the endomysial basement membrane surrounding muscle fibers in skeletal and cardiac muscle. The muscle basement membrane has important roles in imparting mechanical stability as well as in signaling to muscle cells. This is illustrated by the development of congenital muscular dystrophies in humans and in mouse models which can be caused by mutations in basement membrane components, their receptors and cytoskeletal linkages as well as networks that connect basement membranes with the collagenous interstitial matrix (reviewed by Sanes, 2003). The importance of VWA domain-containing molecules in the muscle ECM is well established through the analysis of mutations in the genes encoding the α chains of type VI collagen, which result in Bethlem myopathy and Ullrich congenital muscular dystrophy (Lampe and Bushby, 2005). WARP expression was detected in muscle tissues from early timepoints of development, with reporter gene expression in myotomes and pre-muscle mass structures during embryogenesis, indicating potential contributions for WARP in the differentiation and early development of muscle as well as in post-natal tissues. In summary, our expression studies have identified WARP as a novel component of a limited subset of basement membranes. Our previous studies have shown that WARP is a multimeric VWA domain-containing glycoprotein which interacts with the basement membrane molecule perlecan (Allen et al., 2006), which is important for contributing mechanical stability to basement membrane tissues (Arikawa-Hirasawa et al., 1999; Costell et al., 1999). WARP shares a similar domain organization with the matrilins, a family of multimeric VWA domaincontaining glycoproteins that putatively function as adaptor molecules in linking between ECM structures (reviewed by

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

Wagener et al., 2005), which may suggest similar functions for WARP. Matrilin-1 multimers bound to aggrecan and collagen II fibrils have been demonstrated to form interactions with collagen VI-bound small leucine-rich proteoglycans (Wiberg et al., 2003), providing an elegant demonstration of the multivalent interactions that contribute to ECM assembly. Our data describing the localization of WARP in mouse tissues suggests that WARP may also function as an adaptor molecule to cross-link or stabilize basement membrane components, contributing to their tissue-specific properties and may potentially facilitate linkage between basement membranes and other extracellular matrix networks. 4. Experimental procedures 4.1. Generation of Vwa1+/βgal reporter gene mice Heterozygous WARP lacZ knock-in mice were generated on a C57BL/6 background (Ozgene Pty Ltd., Western Australia). This strain, termed Vwa1+/βgal, has a targeted replacement of the coding region of the Vwa1 gene by the β-galactosidase reporter gene containing a mouse nuclear localization sequence (NLS), as well as a neomycin selection cassette. Males with germline transmission identified by Southern analysis were used to establish inbred lines of Vwa1+/βgal mutant mice by breeding to C57BL/6 females. All mouse experiments were approved by the Animal Ethics Committee of the Murdoch Childrens Research Institute. 4.2. Genotyping Genomic DNA was prepared by proteinase K digestion of tail biopsies, and mice were genotyped by a multiplex PCR reaction utilizing two pairs of allele-specific primers. The wild type PCR product (196 bp) was amplified by primers intron4up 5′-TGTTGTTAGAGTCCGGGTCA and intron4dw 5′-GGAGCAAGGTGTCATGCAG, and the mutant product (367 bp) was amplified by primers lacZ2up 5′-GCCAGTTTGAGGGGACGACGACAG and promo1dw 5′-TCACGGTAGGAGGGCAAGT. PCR reactions consisted of initial denaturation at 94 °C for 5 min; 35 cycles of denaturation (94 °C for 1 min), annealing (63 °C for 30 s), and elongation (72 °C for 30 s); and 7 min of elongation at 72 °C. 4.3. X-gal staining Embryos, dissected tissues or sections were fixed in lacZ fix (0.2% glutaraldehyde, 50 mM EGTA, pH 7.3, 100 mM MgCl2 in 100 mM sodium phosphate, pH 7.3). After several washes with lacZ wash buffer (2 mM MgCl2 and 0.02% Nonidet-P40 in 100 mM sodium phosphate, pH 7.3), samples were stained with 2 mg/ml X-gal in lacZ wash buffer containing 5 mM potassium ferrocyanide and 5 mM potassium ferricyanide overnight at room temperature in the dark. Samples were then washed extensively in lacZ wash buffer. X-gal stained tissue sections for co-localization studies were then post-fixed in 4% paraformaldehyde in PBS for 10 min and immunostained as described

303

below. All other sections were counterstained either with eosin or 0.1% safranin O for 30 s, dehydrated through an ethanol series and equilibrated in xylene before mounting in DPX. Whole mount images were obtained with a Leica Wild MZ8 microscope and tissue sections were analyzed with a Leica Diaplan or Nikon Eclipse 80i microscope. 4.4. Antibodies WARP immunoblotting and immunostaining were performed using a sheep antibody (Institute for Medical and Veterinary Science (Adelaide, Australia) raised against full length recombinant WARP, which had been purified from the media of transfected 293-EBNA cells (Allen et al., 2006). Results with this antibody were verified using a WARP rabbit polyclonal antibody raised against the C-terminal domain of WARP described previously (Allen et al., 2006). Co-staining was performed using rat anti-PECAM monoclonal antibody (MEC 7.46, Abcam), rat anti-perlecan monoclonal antibody (A7L6, Abcam), chicken anti-laminin polyclonal antibody (Abcam) and mouse anti-βIII tubulin (Tuj-I, Abcam). Fluorescent detection was performed using Alexa Fluor-conjugated donkey anti-sheep 488, donkey anti-rabbit 594, donkey anti-rat 594, donkey anti-rat 488, goat anti-chicken 594 and donkey anti-mouse 488 secondary antibodies (Invitrogen). 4.5. Protein extraction and immunoblotting Sciatic nerves were dissected from mice of various ages, diced into 1 mm3 cubes and snap frozen on dry ice. Nerve tissue was ground into powder and incubated in 300 μl of chilled buffer 1 (50 mM Tris HCl pH 7.5, 10 mM EDTA, 0.5% TritonX 100) overnight at 4 °C with agitation. The supernatant was clarified by centrifugation at 13,000 rpm for 20 min at 4 °C and was stored at − 20 °C. The remaining pellet was resuspended with 300 μl of buffer 2 (500 mM NaCl, 50 mM Tris HCl pH 7.5, 10 mM EDTA) and extracted overnight. The supernatant was recovered as described above and the pellet was extracted with 300 μl buffer 3 (4 M GuHCl, 50 mM sodium acetate pH 5.8, 10 mM EDTA) overnight at 4 °C with agitation. The extract was clarified by centrifugation as described. All buffers contained Complete Protease Inhibitors (Roche). The buffer 1, 2 and 3 extracts were quantitated using the BCA Protein Assay (Pierce) following the manufacturers instructions. For SDS-PAGE, 30 μl of each extract was precipitated with 90% ethanol overnight, the pellet washed with 70% ethanol and resuspended in SDS loading buffer. The insoluble pellet remaining after the final extraction step was boiled in 300 μl of SDS loading buffer containing 2 M urea and 2% SDS, and 30 μl was used for SDSPAGE analysis. After denaturation at 99 °C for 5 min and reduction with 50 mM DTT, samples were resolved on a 10% polyacrylamide gel and transferred to Immobilon™-P PVDF membrane (Millipore) for immunoblot analysis using 1:1000 dilution of the rabbit antiWARP C-terminal domain polyclonal antibody as described previously (Allen et al., 2006). Alternatively, samples were transferred to Hybond-LFP membrane (GE Healthcare) for

304

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305

immunoblot analysis using a 1:8000 dilution of sheep anti-WARP antibody, detected with 1:5000 AlexaFluor 488-conjugated donkey anti-sheep antibody (Invitrogen), and visualized with a Typhoon 9400 phosphorimager (Amersham Biosciences).

Childrens Research Institute. We gratefully acknowledge the mouse husbandry support provided by staff at the Murdoch Childrens Research Institute, as well as assistance with confocal microscopy provided by Dr. Matt Burton.

4.6. Immunostaining

References

Tissues from embryonic and adult mice were surgically removed and fixed with 4% paraformaldehyde in PBS for 10 min or overnight depending on tissue volume. Skeletal tissues were decalcified in PBS containing 7% (w/v) EDTA. Tissues were frozen in Tissue Tek OCT compound and 10 μm sections were prepared using a Leica CM3050 cryostat. Cartilage tissue sections were digested with 0.2% hyaluronidase (bovine, type IV-Sigma) and sections for immunoperoxidase staining were treated with 0.3% H2O2 (v/v) in methanol to inactivate endogenous peroxidases. Sections were blocked and permeabilised with PBS containing 1% BSA and 0.1% Triton-X 100 for 1 h, then immunolabeled overnight at 4 °C. Control sections were probed with preimmune serum or with secondary antibodies only. For immunoperoxidase staining, bound antibodies were detected using the Vectastain Elite ABC kit (Vector Laboratories) and visualized using Sigma-Fast DAB tablets. For immunofluorescence staining, antibodies were detected using the secondary antibodies described above. The slides were counterstained with DAPI (4′,6-diamidino-2-phenylindole), washed with PBS and mounted in Fluorsave (Calbiochem). Immunostaining and histology were visualized with a Leica Diaplan microscope and immunofluorescence images were visualized with a Zeiss Axio Imager M1 microscope. Laser scanning confocal microscopy was performed using a Leica TCS SP2 SE confocal microscope.

Allen, J.M., Bateman, J.F., Hansen, U., Wilson, R., Bruckner, P., Owens, R.T., Sasaki, T., Timpl, R., Fitzgerald, J., 2006. WARP is a novel multimeric component of the chondrocyte pericellular matrix that interacts with perlecan. J. Biol. Chem. 281 (11), 7341–7349. Allt, G., Lawrenson, J.G., 2000. The blood–nerve barrier: enzymes, transporters and receptors—a comparison with the blood–brain barrier. Brain Res. Bull. 52 (1), 1–12. Arikawa-Hirasawa, E., Watanabe, H., Takami, H., Hassell, J.R., Yamada, Y., 1999. Perlecan is essential for cartilage and cephalic development. Nat. Genet. 23 (3), 354–358. Bose, K., Nischt, R., Page, A., Bader, B.L., Paulsson, M., Smyth, N., 2006. Loss of nidogen-1 and -2 results in syndactyly and changes in limb development. J. Biol. Chem. 281 (51), 39620–39629. Chen, Z.L., Strickland, S., 2003. Laminin gamma1 is critical for Schwann cell differentiation, axon myelination, and regeneration in the peripheral nerve. J. Cell Biol. 163 (4), 889–899. Chernousov, M.A., Rothblum, K., Stahl, R.C., Evans, A., Prentiss, L., Carey, D.J., 2006. Glypican-1 and alpha4(V) collagen are required for Schwann cell myelination. J. Neurosci. 26 (2), 508–517. Costell, M., Gustafsson, E., Aszodi, A., Morgelin, M., Bloch, W., Hunziker, E., Addicks, K., Timpl, R., Fassler, R., 1999. Perlecan maintains the integrity of cartilage and some basement membranes. J. Cell Biol. 147 (5), 1109–1122. De Arcangelis, A., Mark, M., Kreidberg, J., Sorokin, L., Georges-Labouesse, E., 1999. Synergistic activities of alpha3 and alpha6 integrins are required during apical ectodermal ridge formation and organogenesis in the mouse. Development 126 (17), 3957–3968. Feltri, M.L., Graus Porta, D., Previtali, S.C., Nodari, A., Migliavacca, B., Cassetti, A., Littlewood-Evans, A., Reichardt, L.F., Messing, A., Quattrini, A., Mueller, U., Wrabetz, L., 2002. Conditional disruption of beta 1 integrin in Schwann cells impedes interactions with axons. J. Cell Biol. 156 (1), 199–209. Fitzgerald, J., Ting, S.T., Bateman, J.F., 2002. WARP is a new member of the von Willebrand factor A-domain superfamily of extracellular matrix proteins. FEBS Lett. 517 (1–3), 61–66. Hallmann, R., Horn, N., Selg, M., Wendler, O., Pausch, F., Sorokin, L.M., 2005. Expression and function of laminins in the embryonic and mature vasculature. Physiol. Rev. 85 (3), 979–1000. Heinegard, D., Aspberg, A., Franzen, A., Lorenzo, P., 2002. Glycosylated matrix proteins. In: Royce, P.M., Steinman, B. (Eds.), Connective Tissue and its Heritable Disorders. Wiley Liss, New York, pp. 271–291. Jarvinen, T.A., Jozsa, L., Kannus, P., Jarvinen, T.L., Jarvinen, M., 2002. Organization and distribution of intramuscular connective tissue in normal and immobilized skeletal muscles. An immunohistochemical, polarization and scanning electron microscopic study. J. Muscle Res. Cell Motil. 23 (3), 245–254. Knox, S., Merry, C., Stringer, S., Melrose, J., Whitelock, J., 2002. Not all perlecans are created equal: interactions with fibroblast growth factor (FGF) 2 and FGF receptors. J. Biol. Chem. 277 (17), 14657–14665. Knox, S.M., Whitelock, J.M., 2006. Perlecan: how does one molecule do so many things? Cell. Mol. Life Sci. 63 (21), 2435–2445. Kucherer-Ehret, A., Pottgiesser, J., Kreutzberg, G.W., Thoenen, H., Edgar, D., 1990. Developmental loss of laminin from the interstitial extracellular matrix correlates with decreased laminin gene expression. Development 110 (4), 1285–1293. Kvist, A.J., Nystrom, A., Hultenby, K., Sasaki, T., Talts, J.F., Aspberg, A., 2007. The major basement membrane components localize to the chondrocyte pericellular matrix — a cartilage basement membrane equivalent? Matrix Biol. (Electronic publication ahead of print). Lampe, A.K., Bushby, K.M., 2005. Collagen VI related muscle disorders. J. Med. Genet. 42 (9), 673–685.

4.7. Northern blot hybridization RNA was extracted from tissues of 2-week-old mice using the guanidine isothiocyanate method (Sambrook and Russell, 2001), phenol:chloroform extracted and was further purified and DNase-digested using the RNeasy kit (Qiagen) according to manufacturers instructions. Denatured RNA samples (5 μg) were fractionated by electrophoresis through 0.8% agaroseformaldehyde gel and transferred to nitrocellulose (Schleicher and Schuell) overnight by capillary transfer. RNA was immobilized to the membrane by baking at 80 °C under vacuum for one hour. Northern blots were pre-hybridized with ExpressHyb (Amresco) solution containing 100 μg/ml sheared and denatured fish sperm DNA for 1 h at 68 °C, before hybridization for 2 h at 68 °C with a [α-32P] dCTP-labeled 1.1 kb cDNA probe complementary to the 3′UTR of the Vwa1 gene (Megaprime labeling kit, Amersham). The blot was washed to a stringency of 0.1× SSC, 0.1% SDS at 50 °C and exposed to XOMAT film with intensifying screens at − 80 °C. Acknowledgements This work was supported by grants from the National Health and Medical Research Council of Australia and the Murdoch

J.M. Allen et al. / Matrix Biology 27 (2008) 295–305 Layton, B.E., Sastry, A.M., 2004. A mechanical model for collagen fibril load sharing in peripheral nerve of diabetic and nondiabetic rats. J. Biomech. Eng. 126 (6), 803–814. Lee, H.K., Seo, I.A., Park, H.K., Park, Y.M., Ahn, K.J., Yoo, Y.H., Park, H.T., 2007. Nidogen is a prosurvival and promigratory factor for adult Schwann cells. J. Neurochem. 102 (3), 686–698. Muona, A., Eklund, L., Vaisanen, T., Pihlajaniemi, T., 2002. Developmentally regulated expression of type XV collagen correlates with abnormalities in Col15a1(-/-) mice. Matrix. Biol. 21 (1), 89–102. Nelson, C.M., Bissell, M.J., 2006. Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer. Annu. Rev. Cell Dev. Biol. 22, 287–309. Parmantier, E., Lynn, B., Lawson, D., Turmaine, M., Namini, S.S., Chakrabarti, L., McMahon, A.P., Jessen, K.R., Mirsky, R., 1999. Schwann cell-derived desert hedgehog controls the development of peripheral nerve sheaths. Neuron 23 (4), 713–724. Previtali, S.C., Nodari, A., Taveggia, C., Pardini, C., Dina, G., Villa, A., Wrabetz, L., Quattrini, A., Feltri, M.L., 2003. Expression of laminin receptors in Schwann cell differentiation: evidence for distinct roles. J. Neurosci. 23 (13), 5520–5530. Saito, F., Moore, S.A., Barresi, R., Henry, M.D., Messing, A., Ross-Barta, S.E., Cohn, R.D., Williamson, R.A., Sluka, K.A., Sherman, D.L., Brophy, P.J., Schmelzer, J.D., Low, P.A., Wrabetz, L., Feltri, M.L., Campbell, K.P., 2003. Unique role of dystroglycan in peripheral nerve myelination, nodal structure, and sodium channel stabilization. Neuron 38 (5), 747–758. Sambrook, J., Russell, D.W., 2001. Molecular cloning. A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York. Sanes, J.R., 2003. The basement membrane/basal lamina of skeletal muscle. J. Biol. Chem. 278 (15), 12601–12604.

305

Sasaki, T., Fassler, R., Hohenester, E., 2004. Laminin: the crux of basement membrane assembly. J. Cell. Biol. 164 (7), 959–963. Smith, S.M., West, L.A., Govindraj, P., Zhang, X., Ornitz, D.M., Hassell, J.R., 2007. Heparan and chondroitin sulfate on growth plate perlecan mediate binding and delivery of FGF-2 to FGF receptors. Matrix Biol. 26 (3), 175–184. Wagener, R., Ehlen, H.W., Ko, Y.P., Kobbe, B., Mann, H.H., Sengle, G., Paulsson, M., 2005. The matrilins — adaptor proteins in the extracellular matrix. FEBS Lett. 579 (15), 3323–3329. Wallquist, W., Patarroyo, M., Thams, S., Carlstedt, T., Stark, B., Cullheim, S., Hammarberg, H., 2002. Laminin chains in rat and human peripheral nerve: distribution and regulation during development and after axonal injury. J. Comp. Neurol. 454 (3), 284–293. Wanner, I.B., Guerra, N.K., Mahoney, J., Kumar, A., Wood, P.M., Mirsky, R., Jessen, K.R., 2006. Role of N-cadherin in Schwann cell precursors of growing nerves. Glia 54 (5), 439–459. Whittaker, C.A., Hynes, R.O., 2002. Distribution and evolution of von Willebrand/integrin A domains: widely dispersed domains with roles in cell adhesion and elsewhere. Mol. Biol. Cell 13 (10), 3369–3387. Wiberg, C., Klatt, A.R., Wagener, R., Paulsson, M., Bateman, J.F., Heinegard, D., Morgelin, M., 2003. Complexes of matrilin-1 and biglycan or decorin connect collagen VI microfibrils to both collagen II and aggrecan. J. Biol. Chem. 278 (39), 37698–37704. Yamaguchi, K., Francis-West, P.H., Shiraishi, K., Akita, K., 2004. Premature regression of the leg apical ectodermal ridge in the Japanese chick wingless mutant. Anat. Embryol. (Berl). 208 (6), 487–497. Yurchenco, P.D., Amenta, P.S., Patton, B.L., 2004. Basement membrane assembly, stability and activities observed through a developmental lens. Matrix Biol. 22 (7), 521–538.