Heparan sulfate dependent signaling of fibroblast growth factor (FGF) 18 by chondrocyte-derived perlecan

Abstract

Perlecan is a large multi-domain proteoglycan which is essential for normal cartilage development. In this study perlecan was localized in the pericellular matrix of hypertrophic chondrocytes in developing human cartilage rudiments. Perlecan immunopurified from medium conditioned by cultured human fetal chondrocytes was found to be substituted with heparan sulfate (HS), chondroitin sulfate (CS) and keratan sulfate (KS). Ligand and carbohydrate engagement (LACE) assays demonstrated that immunopurified chondrocyte-derived perlecan formed HS dependent ternary complexes with fibroblast growth factors (FGF) 2 and either FGFR receptors (FGFRs) 1 or 3, however these complexes were not biologically active in the BaF32 cell system. Chondrocyte-derived perlecan also formed HS dependent ternary complexes with FGF18 and FGFR3. The proliferation of BaF32 cells expressing FGFR3 was promoted by chondrocyte-derived perlecan in the presence of FGF18 and this activity was reduced by digesting the HS with either heparinase III or mammalian heparanase. These data suggest that FGF2 and 18 bind to discrete structures on the HS chains attached to chondrocyte-derived perlecan which modulate the growth factor activities. The presence and activity of mammalian heparanase may be important in the turnover of HS and subsequent signaling required for the establishment and maintenance of functional osteo-chondral junctions in long bone growth.

Cartilage provides the framework for endochondral bone formation via a process in embryonic development known as chondrogenesis. The growth of long bone begins with condensation of the mesenchymal stem cells followed by proliferation, which causes the anlage to expand in length and increase in width following the differentiation of the chondroblastic mesenchymal cells into chondrocytes. This process results in chondrocytes being arranged in morphologically distinct zones including resting, proliferating, prehypertrophic and hypertrophic (1). These zones reflect the function of the chondrocytes as they are exposed to a plethora of signaling molecules, some of which interact with their surrounding matrix (2).

Perlecan, the major heparan sulfate (HS) proteoglycan of basement membranes (3–4), has also been isolated from articular cartilage (5) and localized to chondrocytes in the hypertrophic regions of articular cartilage as well as to the growth plate of developing long bones where it was shown to surround the cells and form a “basement membrane” of the chondron (6–9). While there is no direct evidence that chondrocyte-derived perlecan HS is involved in chondrogenesis, cartilage-derived perlecan has previously been shown to bind a number of growth factors involved in bone development including those of the FGF family. Specifically, it has been shown to bind and regulate the activities of FGF1 (10), FGF2 (11–14) and FGF9 (10) by an HS-mediated mechanism and more recently FGF18 by binding to the protein core (15). FGF2 and FGF18 are involved in cartilage growth and maturation where they have been implicated in the development of functional cartilage and bone tissue (16–19). FGF2 may provide an antagonistic signal to FGF18 as FGF2 has been shown to decrease the presence of hypertrophic chondrocytes in cartilage tissues (20) and mice null for FGF2 have decreased bone mass (21). Mice overexpressing FGF2 displayed a phenotype characterized by chondrodysplasia, where FGF2 is anti-proliferative in the growth plate and increased apoptosis in the growth plate chondrocytes, leading to dwarfism (22). Mice lacking FGF18 display expanded zones of proliferating and hypertrophic chondrocytes as well as increased chondrocyte proliferation (2). Mice null for FGF18 have similar skeletal structures to those in animals null for FGFR3 characterized by elongated long bones (16, 23–24). Interestingly, humans that have an activating mutation in FGFR3 also experience a shorter stature, which supports the animal model data and suggests that it is a result of premature terminal differentiation of the chondrocytes (25–26). Together, these data suggest that perlecan, FGF18 and FGFR3 may affect the same signaling pathways that control the chondrocyte phenotype resulting in the modulation of the tissue transition zone between cartilage and bone. In this study we have investigated the HS chains of perlecan isolated from chondrocytes using structural analytical techniques as well as binding assays and the BaF32 cell system to investigate FGF2 and FGF18 complex formation and signaling.

EXPERIMENTAL PROCEDURES

Antibodies, Enzymes and Reagents

The monoclonal antibodies against perlecan domains I (CSI-076) and IV (A7L6) and the polyclonal anti-FGF2 antibody were sourced from Abcam, Cambridge, MA, USA while polyclonal anti-FGF18 was sourced from Santa Cruz Biotechnology Inc., Santa Cruz CA, USA. Antibodies against HS (10E4) and heparinase III generated HS-stubs (3G10) were purchased from Seikagaku Corp., Tokyo, Japan. Antibodies reactive for chondroitinase ABC generated unsulfated (1B5), 4-sulfated (2B6), 6-sulfated (3B3) CS-stubs and KS (5D4) were provided by Prof. Bruce Caterson, Cardiff University, Cardiff, Wales, UK. Biotinylated anti-mouse, anti-rabbit or anti-goat/sheep whole immunoglobulin (Ig) secondary antibodies, streptavidin-horse radish peroxidase (SA-HRP) and streptavidin-fluorescein (SA-FITC) were purchased from GE Healthcare, Little Chalfont Buckinghamshire, UK. Biotinylated anti-rat Ig secondary antibodies were purchased from Dako, Glostrup, Denmark. HRP conjugated anti-mouse Ig raised in sheep and HRP conjugated anti-human Fc/Fab secondary antibodies were purchased from Millipore, Billerica MA, USA. Endoglycosidase enzymes, chondroitinase ABC and heparinase III (EC 4.2.2.8) were purchased from Seikagaku Corp., Tokyo, Japan. Mammalian heparanase was obtained from human platelets as described previously (27). All other chemicals were purchased from Sigma-Aldrich, St. Louis, MO, USA unless stated otherwise.

Isolation of human fetal chondrocytes

Human fetal feet were obtained with informed consent under ethical approval of the North Sydney and Central Coast Area Health Authority Human Research and Ethics Committee, while human foetal chondrocytes sourced from the knee were harvested in accordance with institutional approval from the Human Research Ethics Committee of the University of New South Wales. The cartilage tissue used had no signs of vascularisation and was cleaned and diced into 1 mm3 pieces before it was digested with sterile 0.1% w/v Pronase (from Streptomyces griseus, Roche) in DMEM on a rocker at 37 °C for 90 min followed by centrifugation at 1000 rpm for 3 min. A second digestion with 0.5% w/v collagenase in DMEM on a rocker at 37 °C for 20 h was subsequently performed and cells were separated using a sterile 70 μm cell sieve (BD Biosciences, San Jose CA, USA) followed by centrifugation at 1000 rpm for 3 min. Chondrocytes were cultured in DMEM containing 50 μg/ml L-ascorbic acid, 10% v/v fetal bovine serum (FBS, Invitrogen, Carlsbad CA, USA) and 1% v/v penicillin/streptomycin. Conditioned medium was collected every 3 days and stored at −20 °C until required.

Immunohistochemistry

Human fetal feet (12–14 week old) were obtained fixed and paraffin embedded as described previously (28). Sections were treated twice, 5 min each, with xylene to remove paraffin and the slides were immersed in a series of ethanol solutions for 3 min each (twice in 100% v/v, once in 95% v/v, once in 70% v/v) followed by several exchanges of water. Antigen epitopes were retrieved by immersing the slides in 0.01 M sodium citrate, pH 6 followed by heat treatment in a decloaking chamber (Applied Medical CA, USA) at 121 °C for 4 min. Slides were then rinsed with deionized water followed by blocking with 3% v/v H₂O₂ for 10 min. Additionally, some slides were treated with chondroitinase ABC (0.05 U/ml) or heparinase III (0.01 U/ml) in Dulbecco’s phosphate buffered saline (DPBS) pH 7.2 for 3 h at 37 °C. The slides were washed with 50 mM Tris-HCl, 0.15 M NaCl, 0.05% w/v Tween-20, pH 7.6 (TBST) and then blocked with 1% w/v bovine serum albumin (BSA) in TBST for 1 h at room temperature (RT). The slides were incubated with primary antibodies diluted in 1% w/v BSA in TBST at 4 °C for 16 h. Primary antibodies used included A7L6 (4 μg/ml), 10E4 (4 μg/ml), 5D4 (1:1000), CS56 (1:500), 1B5 (1:500), 2B6 (1:500), 3B3 (1:500), FGF2 (4 μg/ml), FGF18 (4 μg/ml), FGFR1 (4 μg/ml), FGFR3 (4 μg/ml) and type X collagen (4 μg/ml). Slides were then washed twice with TBST before incubating with the appropriate biotinylated secondary antibodies (1:500) for 1 h at RT. The slides were washed twice with TBST then incubated for 30 min with streptavidin-HRP (1:250), rinsed 4 times with TBST before developing with NovaRED® chromogen stain (Vector Laboratories, Burlingame CA, USA). The slides were counterstained with hemotoxylin (Vector Laboratories, Burlingame CA, USA) for 3 min then rinsed with deionized water before imaging using light microscopy.

Immunocytochemistry

Chondrocytes (passages 2 – 4) were cultured to confluence on microscope slides (Ultrafrost, Lomb Scientific, Taren Point NSW, Australia), fixed with ice cold acetone for 3 min and rinsed with 50 mM Tris-HCl, 0.15 M NaCl, pH 7.6 (TBS). Selected slides were treated with chondroitinase ABC (0.05 U/ml) or heparinase III (0.01 U/ml) in PBS, pH 7.2 for 3 h at 37 °C. Slides were then blocked with 0.1% w/v casein in DPBS for 1 h at RT followed incubation with the primary antibodies at a final concentration of 2 μg/ml for 16 h at 4 °C. Primary antibodies used included CSI-076, 10E4, CS56, 2B6, 3B3, FGF2, FGF18, FGFR1, FGFR3 and type X collagen. Slides were rinsed twice with TBST and incubated with the appropriate biotinylated secondary antibodies (1:500) for 1 h at RT before rinsing twice with TBST and incubation with SA-FITC (1:250) for 30 min at RT followed by 4 washes with TBST. The slides were then counterstained with 1 μg/ml of 4′, 6-diamidino-2-phenylindole, dilactate (DAPI, Invitrogen, Carlsbad CA, USA) in DPBS for 10 min in the dark and rinsed 4 times with the deionised water before imaging using fluorescence microscopy.

Immunopurification of perlecan

Perlecan was isolated from the conditioned medium produced by cultured human fetal chondrocytes and human coronary artery endothelial cells (HCAEC) by anion exchange and monoclonal antibody affinity chromatography, as described previously (29–30).

Western blot analysis

Purified perlecan samples (10 μg/lane) were treated with heparinase III (0.01 U/ml) in DPBS pH 7.2 at 37 °C for 16 h, and electrophoresed through 3–8% Tris-Acetate NuPAGE® SDS-PAGE gels (Invitrogen, Carlsbad CA, USA) for 1 h at 200 V in tris-tricine buffer (50 mM tricine, 50 mM tris base, 0.1% w/v SDS, pH 8.3). Molecular weight markers (HiMark®, Invitrogen, Carlsbad CA, USA) were electrophoresed on each gel. Samples were then transferred to polyvinylidene difluoride (PVDF) membrane (Immobilon-P, Millipore, Billerica MA, USA) using transfer buffer (5 mM bicine, 5 mM bis tris, 0.2 mM EDTA, 0.005% SDS, 10% v/v methanol, pH 7.2) in a semi-dry blotter (Invitrogen, Carlsbad CA, USA) at 300 mA and 20 V for 1 h. The membrane was blocked with 1% w/v BSA in 20 mM tris base, 136 mM NaCl, 0.1 w/v% Tween-20, pH 7.6 (TBST-WB) for 2 h at RT followed by incubation with the perlecan domain IV antibody (A7L6, 0.1 μg/ml) diluted in 1% w/v BSA in TBST-WB for 16 h at 4 °C. Membranes were subsequently rinsed with TBST-WB, incubated with HRP conjugated antibodies (1:50,000 diluted in 1% w/v BSA in TBST-WB) for 45 min at RT, rinsed with TBST-WB and 20 mM tris base, 136 mM NaCl, pH 7.6 before being imaged using chemiluminescent reagents (Femto reagent kit, Pierce Biotechnology, Rockford, IL, USA) and X-ray film (Australian Imaging Distributors, North Ryde NSW, Australia).

ELISA analysis

Immunopurified chondrocyte-derived perlecan (5 μg/ml) was adsorbed onto wells of a 96-well ELISA plate (Greiner Bio One, GmbH, Frickenhausen, Baden-Württemberg, Germany) for 2 h at RT, rinsed twice with DPBS and blocked with 0.1% w/v casein in DPBS for 1 h at RT. Wells were rinsed twice with DPBS with 0.1% w/v Tween-20 (DPBST) and incubated with antibodies against perlecan domain IV (MAb A7L6, 2 μg/ml), HS (MAb 10E4, 2 μg/ml), CS (MAb CS56, 1:500) and KS (MAb 5D4, 1:500) diluted in 0.1% w/v casein in DPBST. Wells were then rinsed twice with DPBST followed by incubation with biotinylated secondary antibodies diluted in 0.1% w/v casein DPBST for 1 h at RT followed by two rinses with DPBST. Wells were then incubated with SA-HRP (1:500) for 30 min at RT and rinsed four times with DPBST. Binding of the antibodies to the samples was detected using the colourimetric substrate, 2,2′azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS), and absorbance measured at 405 nm.

Ligand and Carbohydrate Engagement assay (LACE)

Perlecan (5 μg/ml) was adsorbed onto 96-well high-binding ELISA plates for 16 h at 4 °C. Wells were washed with DPBS and selected wells were digested with endoglycosidase enzymes, chondroitinase ABC (0.05 U/ml), heparinase III (0.01 U/ml) or mammalian heparanase (2 ng/ml) in 0.01% w/v BSA with DPBS pH 7.2 for 3 h at 37 °C. Mammalian heparanase was found to degrade perlecan HS at pH 7.2 and inhibit the formation of ternary complexes with FGFs and their receptors. Wells were blocked with 3% w/v BSA in DPBS for 1 h at RT and washed twice with DPBST. Either recombinant human FGF2 or 18 (5 nM, Invitrogen, Carlsbad CA, USA) and either soluble recombinant human FGFR1c or 3c human-IgG chimeric proteins (5 nM, R&D Systems, Minneapolis MN, USA) were incubated in PBST for 10 min before being transferred into each well and incubated for 1 h at 37 °C. Controls were incubated with FGFR1c or 3c only. Wells were then washed twice with DPBST and incubated with HRP conjugated anti-human IgG secondary antibody (1:1000) in DPBST for 1 h at 37 °C. Wells were washed 4 times with DPBST before the addition of ABTS and the absorbance measured at 405 nm.

BaF32 cell proliferation assays

BaF32 cells are an IL-3 dependent and HS proteoglycan deficient myeloid B cell line which have been stably transfected with either FGFR1c or FGFR3c (31–32). BaF32 cells are a model system developed to identify HS and heparin structures that interact with FGFs and their receptors. The readout of this assay is cell proliferation which indicated the formation of ternary complexes on the cell surface. BaF32 cells were maintained in RPMI 1640 medium containing 10% v/v FBS, 10% v/v WEHI-3BD conditioned medium and 1% v/v penicillin/streptomycin. WEHI-3BD cells were maintained in RPMI 1640 medium supplemented with 2 g/L sodium bicarbonate, 10% v/v FBS and 1% v/v penicillin/streptomycin and the conditioned medium was collected 3 times per week and stored at −20 °C until required. For the mitogenic assays, the BaF32 cells were transferred into IL-3 depleted medium for 24 h prior to experimentation and seeded into 96-well plates at a density of 2×10⁴ cells/well in the presence of chondrocyte-derived perlecan (40 μg/ml) and either FGF2 (0.03 nM) or FGF18 (9 nM). Heparin (30 nM) with either FGF2 (0.03 nM) or FGF18 (9 nM) were used as positive controls for the assay while cells in the presence of medium only were used as a negative control. To investigate the role of the glycosaminoglycan chains attached to chondrocyte-derived perelcan, endoglycosidase digestions were also performed in situ with chondroitinase ABC (0.05 U/ml), heparinase III (0.01 U/ml) or mammalian heparanase (2 ng/ml) at pH 7.2 and 37 °C for 16 h prior to commencement of the cell assay. Cells were incubated for 96 h in 5% CO₂ at 37 °C and the number of cells present was assessed using the MTS assay. The MTS reagent (Promega, Madison WI, USA) was added to the cell cultures 6 h prior to measuring the absorbance at 490 nm.

Statistical analysis

A student’s t-test (two samples, 2-tailed distribution assuming equal variance) was used to compare statistical significance. Results of p < 0.05 were considered significant. Experiments were performed in triplicate and experiments were repeated.

RESULTS

Immunolocalization of perlecan, glycosaminoglycans, FGFs and FGFRs in the developing human anlagen

The presence of perlecan was detected in the middle zone of developing cartilage rudiments in human feet (marked by asterisks in Figure 1A) and in the developing growth plates located in the either end of the metatarsal (filled arrow – proximal end; unfilled arrow – distal end, Figure 1A). When the region identified by the filled arrow in Figure 1A was examined under higher power, it was noted that the perlecan immunoreactivity was localized to the pericellular matrix, which is in close association with the cell and lacunae membranes (Figure 1B). Chondrocytes in this same region stained positively for type X collagen (Figure 1C), indicating that these cells were hypertrophic in nature. Regions of the developing foot that had significant immunoreactivity for type X collagen also demonstrated the presence of HS stubs (Figure 1D) generated after digestion with heparinase III, native CS (Figure 1E), as well as 4-sulfated (Figure 1G) and 6-sulfated (Figure 1H) CS stubs generated after digestion with chondroitinase ABC. In contrast, the same regions showed no reactivity towards the antibody that recognized the unsulfated CS stubs (Figure 1F). Interestingly, the presence of the HS stubs (Figure 1D) and 4-sulfated CS-stubs (Figure 1G) was confined to the pericellular matrix, whereas the 6-sulfated CS-stubs (Figure 1H) were also detected in the inter-territorial matrix (Figure 1H), which was most likely attached to aggrecan. Chondrocytes in regions of type X collagen staining also stained for the presence of FGF2 (Figure 1I), FGF18 (Figure 1L), FGFR1 (Figure 1J) and FGFR3 (Figure 1K). FGF18 was also detected in the tissue between the developing cartilage joints (Figure 1M) and the perichondrium surrounding the cartilage (data not shown).

Figure 1.

Immunolocalization of perlecan, type X collagen, GAGs, FGFs and FGF receptors 1 and 3 in developing human cartilage rudiments. The sections were probed for the presence of perlecan (marked by asterisks and arrows in A, MAb A7L6). A region was selected from Figure 1A (marked by the filled arrow at 25 × magnification) and viewed at a higher (1000 ×) magnification (B, MAb A7L6). Similar regions were also investigated at the higher magnification for type X collagen (C), as well as the presence of HS-stubs generated by heparinase III enzyme digestion (D, MAb 3G10), native CS (E, MAb CS56), while CS-stubs were generated by chondroitinase ABC digestion, and probed for the presence of unsulfated CS stubs (F, MAb 1B5), 4-sulfated CS stubs (G MAb 2B6), 6-sulfated CS stubs (H, MAb 3B3). The sections were probed for the presence of FGF2 (I), FGFR1 (J), FGFR3 (K) and FGF18 (L and M). The nuclei were counterstained with hemotoxylin (scale bar is 500 μm in A, 10 μm in B – K and 20 μm in L – M).

Immunolocalization of type X collagen, perlecan, glycosaminoglycans, FGFs and FGFRs in cultured human fetal chondrocytes

When the cells were isolated from developing cartilage rudiment tissue and cultured, some of the cells were found to express type X collagen supporting the idea that some of the cells were hypertrophic (Figure 2A). Most cells expressed and secreted perlecan into the pericellular/extracellular space as microfibrillar structures (Figure 2B). HS was detected with a punctate and granular staining pattern (Figure 2C). Staining for CS showed that some cells exhibited intracellular peri-nuclear staining, such as that seen for the native CS (Figure 2D), and the 4- and 6-sulfated CS stubs (Figure 2E and F respectively). FGF2 was detected intracellularly and extracellularly where it appeared to be localized in microfibrillar structures (Figure 2G). FGF18 (Figure 2H) and FGFR1 (Figure 2I) did not show any significant staining as compared to a no primary control (data not shown), however some of the cells showed positive staining for FGFR3 (Figure 2J).

Figure 2.

The expression of perlecan, glycosaminoglycans, FGFs, FGF receptors and type X collagen in cultured human fetal chondrocytes. Human fetal chondrocytes were cultured on glass slides and probed for the presence of type X collagen (A), perlecan (B, MAb CSI-076), HS (C, MAb 10E4), native CS (D, MAb CS56), 4-sulfated CS stubs (E, MAb 2B6), 6-sulfated CS stubs (F, MAb 3B3), FGF2 (G), FGF18 (H), FGFR1 (I) or FGFR3 (J). Specific staining was shown by FITC (green) whilst the nuclei were counterstained with DAPI (blue) at 500 × magnification (scale bar is 20 μm in A–J).

Biochemical characterization of chondrocyte derived perlecan

Immunopurified chondrocyte-derived perlecan isolated from cells in culture was characterized with respect to its molecular mass and glycosaminoglycan composition. Chondrocyte-derived perlecan was found to have a molecular mass in excess of 600 kDa when probed with a perlecan protein core antibody against domain IV (Figure 3A). Immunopurified chondrocyte-derived perlecan was probed with antibodies against HS, CS and KS by ELISA which confirmed that this material contained all three glycosaminoglycan types (Figure 3B). The substructure of the CS chains was further analyzed and shown to contain both 4- and 6-sulfated CS stub structures (data not shown).

Figure 3.

Characterization of immunopurified chondrocyte-derived perlecan. Immunopurified chondrocyte-derived perlecan was electrophoresed through 3 – 8 % SDS-PAGE gels, electroblotted onto a PVDF membrane and probed with an anti-perlecan antibody (Domain IV: MAb A7L6) (A). Immunopurified chondrocyte-derived perlecan was coated onto 96 well plates (B) and probed with antibodies against perlecan (Domain IV: MAb A7L6), heparan sulfate (HS, MAb 10E4), chondroitin sulfate (CS, MAb CS56) or keratan sulfate (KS, MAb 5D4). Data is presented as mean ± standard deviation (n=3).

Ternary complexes formed between perlecan, growth factors and receptors

Chondrocyte-derived perlecan formed ternary complexes with FGF2 and either FGFR1 (Figure 4A) or FGFR3 (Figure 4B). Digestion of chondrocyte-derived perlecan with either heparinase III or mammalian heparanase reduced the ability of perlecan to form complexes. Chondrocyte-derived perlecan formed HS dependent ternary complexes with FGF18 and FGFR3 (Figure 4D), but not with FGF18 and FGFR1 (Figure 4C). In contrast, endothelial-derived perlecan supported the formation of HS-dependant complexes between FGF18 and either FGFR1 or FGFR3 (Figure 4C and D). The complexes formed between chondrocyte-derived perlecan, FGF18 and FGFR3 were also sensitive to both heparinase III and mammalian heparanase digestion however when endothelial-derived perlecan was used to form these complexes it was only sensitive to heparinase III digestion (Figure 4D). These contrasting results between chondrocyte- and endothelial-derived perlecan suggest that it is not simply a difference in the amount of HS that elicits the different responses but it is suggestive of significant differences in the structure of the HS.

Figure 4.

Ternary complexes formed between perlecan, either FGF2 or FGF18 and either FGF receptor type 1 or 3. Perlecans were adsorbed onto wells of a 96 well plate and were either undigested or digested in situ with heparinase III, mammalian heparanase or chondroitinase ABC before performing the LACE assay. Measurements were corrected for absorbance measurements detected in the absence of FGF and are presented as mean ± standard deviation (n=3). Significant differences were analyzed using a student’s t-test where ^* denotes p < 0.05 compared to undigested chondrocyte-derived perlecan and ^** denotes p < 0.05 compared to the undigested endothelial-derived perlecan.

The binding of FGFR3 to perlecan in the absence of FGF18 was assessed to ensure that both FGF18 and FGFR3 needed to be present to form a complex (Figure 5). FGFR3 itself, did not bind to either chondrocyte- or endothelial-derived perlecan in either the presence or absence of HS. This supports the finding that perlecan HS, FGF18 and FGFR3 need to be present to form a complex.

Figure 5.

Complexes formed between perlecan, FGF18 and FGFR3 or perlecan and FGFR3. Perlecans were adsorbed onto wells of a 96 well plate and were either undigested or digested in situ with heparinase III, mammalian heparanase or chondroitinase ABC before performing a LACE assay. Measurements were corrected for absorbance measurements detected in the absence of perlecan and are presented as mean ± standard deviation (n=3).

The FGF growth promoting activities of perlecan

The activity of the ternary complexes formed between perlecan, growth factors and their cognate receptors were tested using the BaF32 cell system expressing either FGFR1 (Figure 6A and C) or FGFR3 (Figure 6B and D). Commercially available heparin and either FGF2 or FGF18 were used as positive controls for the assay while cells in the presence of medium only were used as a negative control. Proliferation of the FGFR1 expressing BaF32 cells in the presence of chondrocyte-derived perlecan and FGF2 was found to support increased, but not statistically significantly (p < 0.05) greater proliferation than that of the medium only control (Figure 6A), suggesting that chondrocyte-derived perlecan elicited only weak activity in the presence of FGF2. Chondrocyte-derived perlecan did not support the proliferation of FGFR1 expressing BaF32 cells in the presence of FGF 18 (Figure 6A). Digestion of chondrocyte-derived perlecan with endoglycosidase enzymes did not significantly (p < 0.05) change the level of proliferation of FGFR1 expressing BaF32 cells in the presence of FGF2 compared to undigested chondrocyte-derived perlecan (Figure 6C). Both endothelial-derived perlecan and heparin supported the proliferation of FGFR1 expressing BaF32 cells in the presence of FGF2, but not in the presence of FGF18 (Figure 6A).

Figure 6.

The proliferation of BaF32 cells expressing either FGFR1 or FGFR3 in the presence of chondrocyte- or endothelial-derived perlecan with either FGF2 or FGF18. FGFR1 (panel A) or FGFR3 (panel B) expressing BaF32 cells were incubated with chondrocyte- or endothelial-derived perlecan in the presence of either FGF2 (black bars) or FGF18 (grey bars) and compared to heparin in the presence of either FGF2 or FGF18. Medium only was used as a negative control (white bars). Chondrocyte-derived perlecan was either left undigested or digested with heparinase III, mammalian heparanase or chondroitinase ABC in situ prior to incubation with either FGFR1 expressing BaF32 cells in the presence of FGF2 (panel C) or FGFR3 expressing BaF32 cells in the presence of FGF18 (panel D). Data are presented as mean ± standard deviation (n=3). Significant differences were analyzed using a student’s t-test where ^* denotes p < 0.05 compared to medium only and ^** denotes p < 0.05 for the endoglycosidase digested perlecan samples compared to the undigested perlecan.

Chondrocyte-derived perlecan supported the proliferation of the FGFR3 expressing BaF32 cells in the presence of FGF18 which was significantly (p < 0.05) greater than that of the negative control and equal to that of the positive heparin control (Figure 6B). Digestion of chondrocyte-derived perlecan with either heparinase III or mammalian heparanase significantly (p < 0.05) reduced the level of proliferation compared to the undigested chondrocyte-derived perlecan (Figure 6D). Interestingly, digestion with mammalian heparanase significantly reduced the amount of proliferation to a level that was statistically similar to that achieved in the medium only negative control (Figure 6D). Chondrocyte-derived perlecan did not promote the proliferation of FGFR3 expressing BaF32 cells in the presence of FGF2 above the medium only control (Figure 6B). In contrast, endothelial-derived perlecan promoted the proliferation of FGFR3 expressing BaF32 cells in the presence of either FGF2 or FGF18 (Figure 6B).

DISCUSSION

This paper shows that the HS decorating chondrocyte-derived perlecan protein core mediates the binding of FGF18 and the subsequent activation of FGFR3. It also shows that digestion of the HS decorating chondrocyte-derived perlecan can inhibit this signaling, supporting the hypothesis that FGF18 signaling in developing cartilage tissues is a result of the concerted actions and local concentrations of perlecan and the enzyme that degrades its HS.

The localization of perlecan to the pericellular matrix of chondrocytes in the developing cartilage rudiments supports the hypothesis that perlecan performs an important role in the development of these tissues (10, 28, 33–34). The staining pattern for the immunolocalization of FGF2 and perlecan was similar in some regions of the developing cartilage, supporting the idea that perlecan is a mechanotransductor by binding FGF2, which results in chondrocytes being able to receive proliferative signals in response to mechanical stimuli (14). FGF2, FGF18, FGFR1 and FGFR3 were immunolocalized to similar regions of the cartilage tissue, which contained hypertrophic chondrocytes as identified by their size and the presence of type X collagen, supporting hypotheses that these factors may be involved in the regulation of chondrocyte proliferation and terminal differentiation prior to apoptosis (18, 35–39). FGF18 was detected in the developing cartilage tissue between the developing cartilage joints and the perichondrium surrounding the cartilage rudiment as reported previously using mRNA in situ hybridization (2, 18). FGF18 could not, however, be detected in the in vitro chondrocyte cultures suggesting that FGF18 was not produced by these cells and may act as a paracrine signaling factor.

Glycosaminoglycan characterization of the chondrocyte-derived perlecan suggested that it was secreted as a full length molecule substituted with HS, KS and CS. This is in agreement with previous reports describing perlecan isolated from human and bovine cartilage tissues as a proteoglycan decorated with HS, CS or both (7, 10). However, the precise location, structure and role of the glycosaminoglycans remain to be determined.

FGF2 has been shown to bind to perlecan derived from fetal bovine growth plates in an HS-dependent manner (13) while the role of FGF18 has also been shown to be important in cartilage development (18). Thus, in this study, we were interested to determine whether the HS on the chondrocyte-derived perlecan was able to interact with FGF2 and FGF18 to form ternary complexes with either FGFR1 or FGFR3 in vitro. The HS only endothelial-derived perlecan was used as a control to compare with chondrocyte-derived perlecan in the LACE assays. Chondrocyte-derived perlecan formed HS dependent ternary complexes with FGF2 and FGFR1 or FGFR3, which were sensitive to both heparinase III and mammalian heparanase digestion. In contrast, endothelial cell-derived perlecan was only sensitive to heparinase III digestion indicating differences in the structure of HS produced by the two cell types. Chondrocyte-derived perlecan also formed HS dependent ternary complexes with FGF18 and FGFR3 whilst complexes were not formed between FGF18 and FGFR1, suggesting that the HS motif on chondrocyte-derived perlecan may be specific for certain complexes. Previous studies have shown that FGF18 binds to FGFR3 with higher specificity than to other FGFRs (18) and that FGF18 has a greater receptor selectivity than FGF2 (39–40). FGF18 has also been shown to bind to the protein core of perlecan (15), which may explain the residual binding demonstrated between heparinase III treated perlecan, FGF18 and FGFR3. However the complexes formed between FGF18, FGFR3 and perlecan are predominantly formed via interactions with the HS chains and FGFR3 alone does not bind directly to the HS chains as there was no significant binding in the absence of FGF18.

We were interested to determine whether chondrocyte-derived perlecan could promote signaling of FGF2 and FGF18 and whether digestion by mammalian heparanase affected the biological activity which was determined by the proliferation of FGFR1 and FGFR3 expressing BaF32 cells in situ. The BaF32 cell assays indicate signaling of the complexes formed between perlecan, growth factors and growth factor receptors through their proliferation. Chondrocyte-derived perlecan did not stimulate the statistically significant (p < 0.05) proliferation of FGFR1 expressing BaF32 cells in the presence of FGF2 or 18 above the medium only control and hence did not provide a biologically active signal. These data are in agreement with previous studies using perlecan isolated from bovine growth plate (13). The LACE assay demonstrated that chondrocyte-derived perlecan did not form ternary complexes with FGF18 and FGFR1, however, chondrocyte-derived perlecan was able to form ternary complexes with FGF2 and FGFR1. It has been shown previously that perlecan can sequester FGF2 away from its high affinity receptor preventing its activation (12, 13) and degradation (41). The LACE assay data together with the BaF32 cell-based assay data presented here shows that whilst chondrocyte-derived perlecan can form ternary complexes with FGF2 and FGFR1, it was not able to signal to cells suggesting that the complexes are in an inactive state. This novel phenomenon was also observed for complexes formed between chondrocyte-derived perlecan, FGF2 and FGFR3. The ability of chondrocyte-derived perlecan to form biologically inactive ternary complexes is likely due to the structure of its HS. The importance of HS structure on biological activity has been reported previously in both the endothelial and neuroepithelial systems. Different sources of endothelial-derived perlecan have been reported to provide differential regulation of FGF2 mediated cell signaling (42) while neuroepithelial cells at different stages of development have structurally distinct forms of HS (43) which selectively interact with specific FGF-FGFR complexes to regulate activation (44).

Chondrocyte-derived perlecan stimulated the proliferation of FGFR3 expressing BaF32 cells in the presence of FGF18 and this activity was modulated by either heparinase III or mammalian heparanase. This suggests a role for mammalian heparanase in vivo in the turnover and processing of the FGF18-FGFR3 signal in hypertrophic chondrocytes by cleaving the HS attached to perlecan that promotes the formation of biologically active complexes. In support of this, FGF18 has been known to signal via FGFR3 to promote chondrogenesis (36) while heparanase mRNA was recently localized at the chondro-osseus junction of murine growth plates (45–46), which further supports the hypothesis that mammalian heparanase plays an important role in the turnover of HS attached to chondrocyte-derived perlecan in cartilage development.

In conclusion, this paper demonstrated that chondrocyte-derived perlecan HS was capable of binding and supporting the biological activity of FGF18 through the formation of ternary complexes with FGFR3. Chondrocyte-derived perlecan was also capable of binding FGF2 through the formation of ternary complexes with FGFR1 or FGFR3, however they did not form biologically active complexes. The inhibition of FGF18 activity by mammalian heparanase leads to subsequent events of terminal differentiation of hypertrophic chondrocytes and apoptosis which itself is a fore-runner of the development of bone. These events are controlled by the synthesis and turnover of the growth factors, their receptors and HS, which all act in concert to produce functional components of the skeletal system with the correct architecture.

ABBREVIATIONS

ABTS

2,2′azino-bis(3-ethylbenzthiazoline-6-sulfonic acid)

BSA

bovine serum albumin

CS

chondroitin sulfate

DPBS

Dulbecco’s phosphate buffered saline

DPBST

Dulbecco’s phosphate buffered saline with 0.1% w/v Tween-20

FGF

fibroblast growth factor

FGFR

fibroblast growth factor receptor

HS

heparan sulfate

KS

keratan sulfate

LACE

ligand and carbohydrate engagement assay

PVDF

polyvinylidene difluoride

RT

room temperature

SA-FITC

streptavidin-fluorescein

SA-HRP

streptatividin-horse radish peroxidase

TBS

50 mM Tris-HCl, 0.15 M NaCl, pH 7.6

TBST

50 mM Tris-HCl, 0.15 M NaCl, 0.05% w/v Tween-20, pH 7.6

TBST-WB

20 mM tris base, 136 mM NaCl, 0.1% w/v Tween-20, pH 7.6