Biomaterials 35 (2014) 6776e6786

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Engineering a vascular endothelial growth factor 165-binding heparan sulfate for vascular therapy Chunming Wang a,1, 3, Selina Poon a, 3, Sadasivam Murali a, Chuay-Yeng Koo a, 2, Tracey J. Bell b, Simon F. Hinkley b, Huiqing Yeong a, Kishore Bhakoo c, Victor Nurcombe a, d, Simon M. Cool a, e, * a

Glycotherapeutics Group, Institute of Medical Biology, Agency for Science, Technology and Research, Singapore The Ferrier Research Institute, Victoria University of Wellington, Lower Hutt, New Zealand Translational Molecular Imaging Group, Singapore Bioimaging Consortium, Agency for Science, Technology and Research, Singapore d Lee Kong Chian School of Medicine, Nanyang Technological UniversityeImperial College London, Singapore e Department of Orthopaedic Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 March 2014 Accepted 22 April 2014 Available online 20 May 2014

The therapeutic use of VEGF165 to stimulate blood vessel formation for the treatment of peripheral arterial disease or cardiovascular-related disease has met with limited success. Here we describe an affinity-isolated heparan sulfate glycotherapeutic (HS7þve) that binds to, and enhances the bioactivity of, VEGF165. Application of HS7þve complexed with VEGF165 results in enhanced VEGF165eVEGFR2 interaction, prolonged downstream pErk1/2 signalling, and increased cell proliferation and tube formation in HUVECs, compared with VEGF165 alone. The pro-angiogenic potential of HS7þve was further assessed in vivo using the chick embryo chorioallantoic membrane (CAM) assay. Exogenous dosing with HS7þve alone significantly enhanced the formation of new blood vessels with potencies comparable to VEGF165. These results demonstrate the potential for vascular therapy of glycotherapeutic agents targeted at augmenting the bioactivity of VEGF165. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Angiogenesis Vasculogenesis Glycosaminoglycans Blood vessel Heparin VEGF

1. Introduction Ischaemic heart and vascular disease including myocardial infarction and stroke remain the leading cause of death worldwide [1], but treatments remain problematical. The use of proangiogenic agents to stimulate the formation of new blood vessels, a therapeutic angiogenic approach, is currently being trialed to improve perfusion at ischaemic sites [2e4]. Vascular endothelial growth factor (VEGF) appears particularly promising. The most abundant isoform, VEGF165, has excited the most interest because of its powerful physiological effects [5]. However, clinical trials testing recombinant VEGF165 have so far been disappointing [5e8], in part due to its instability in physiological environments [9]. Thus,

* Corresponding author. Institute of Medical Biology, Agency for Science, Technology and Research, Singapore. Fax: þ65 6478 9477. 1 Present address: Institute of Chinese Medical Sciences, University of Macau, Avenida Padre Tomas Pereira, Taipa, Macau. 2 Present address: Division of Cancer Studies, King’s College London, UK. 3 Both authors contributed equally. http://dx.doi.org/10.1016/j.biomaterials.2014.04.084 0142-9612/Ó 2014 Elsevier Ltd. All rights reserved.

maintaining effective concentrations of VEGF165 at ischaemic sites has proven difficult, resulting in high, and often excessive dosing that leads to unwanted side effects such as aberrant angiogenesis [2,5e7]. As such, there is still a pressing need for a safe and effective therapeutic capable of restoring blood supply. Numerous studies have demonstrated the essential role of heparan sulfate (HS) in mediating VEGF165-directed angiogenesis [10,11]. HS is composed of a family of variably sulfated glycosaminoglycans (GAGs) consisting of repeating disaccharide units of glucuronic acid (GlcA) and glucosamine (GlcN) [12,13]. HS binds to the carboxyl-terminal of VEGF165, stabilises and enhances the interaction between VEGF165 and VEGF receptor 2 (VEGFR2), and so regulates endothelial proliferation, tube formation and vascular hyper-permeability [11,14,15]. As HS can be readily harvested from cultured tissues and cells, it is emerging as a new class of therapeutic compound capable of augmenting blood vessel growth. Subtle variations in disaccharide sequence, chain length and biosynthesis endow distinct HS variants in each tissue with unique growth-factor binding capacity, and thereby targeted control of downstream bioactivity [16,17]. Maximising the therapeutic potential of HS thus requires a strategy for dealing with this

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

6777

Fig. 1. Purification of VEGF165-binding HS. (A) Schematic of VEGF165 showing the VEGF receptor binding region 1e110 aa (blue) and the carboxyl-terminal heparin-binding domain 111-165aa (green), the latter of which was synthesised as a 6-aminohexanoic acid-biotin linked peptide for HS purification. The VEGF165 heparin-binding domain peptide was labelled AR55. (B) Filter-binding assay of AR55-AHX-biotin immobilised on nitrocellulose membranes and exposed to [3H]-heparin. Data is representative of duplicate experiments. (C) Schematic of VEGF165-binding HS (HS7þve) isolation from porcine mucosal HS (HSpm): HSpm was dissolved in low salt buffer (20 mM PBS, 0.15M NaCl, pH 7.2) and loaded onto a streptavidin column coupled with AR55. The column was washed with the same buffer until the baseline reached zero, and HS7þve was then eluted using high salt buffer (20 mM PBS, 1.5 M NaCl, pH 7.2). Both HS7þve and unbound HS (HS7ve) were collected. (D) Chromatogram depicting the isolation of HS7þve from HSpm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

6778

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

compositional heterogeneity. We have recently developed a platform technology that isolates particular HS variants based on their affinity for specific growth factors [18]. This technology has been recently exemplified with an HS variant tuned to the osteogenic factor BMP-2, and proved efficacious for bone healing [18]. Like BMP-2, VEGF165 has a prominent heparin-binding domain [19] that can be used to enrich for HS with complimentary binding motifs. It is known that sulfates within any particular HS are arranged in sequences that align to corresponding basic amino acids in target proteins [20]. The synthesis of such peptide domains allows for affinity-based purification methods, and thus enrichment of HS preparations with increased growth factor potency [18]. Here we exploit the affinity of HS for VEGF165 to isolate a unique binding HS fraction and characterise its composition, VEGF165binding capacity and angiogenic ability both on endothelial cellbased and organotypic CAM assays. This carbohydrate-based approach offers a new strategy for manipulating the angiogenic cascade that may prove useful for the treatment of blood vessel disease. 2. Materials and methods

was applied. Isocratic flow (1 ml/min) was monitored by UV detection at 232 nm. HSpm that did not bind to the column (HS7ve) was collected as flow-through. Bound HS (HS7þve) was then eluted with high salt buffer (HiSB) (20 mM PBS, 1.5 M NaCl, pH 7.2), lyophilised for 72 h, and then applied to HiPrep Desalting Columns (GE Healthcare). Desalted HS samples were lyophilised, weighed and stored in aliquots at 80  C. The procedure is illustrated in Fig. 1C. 2.3. Characterisation of HS7þve HS7þve, HS7ve and HSpm (1 mg) in 100 mM NaOAc with 10 mM CaOAc, pH 7 (500 ml) were digested in duplicate with 2.5 mIU each of heparinase (EC 4.2.2.7), heparitinase II (no EC number assigned) and heparitinase (EC 4.2.2.8), all from Ibex Technologies Incorporated (Montreal, Canada) at 37  C for 24 h with gentle inversion (9 rpm). A further aliquot of each enzyme was added and the solutions incubated as before, for a further 48 h. The digestion was terminated by heating (100  C, 5 min) and the resulting disaccharides analysed by capillary electrophoresis (CE). Analyses were performed using an Agilent3D CE (Agilent Technologies, Waldbronn, Germany) instrument with an uncoated fused silica capillary tube (75 mm ID, 64.5 cm total and 56 cm effective length; Agilent Technologies Inc., Santa Clara, CA, USA), in 20 mM phosphoric acid buffer, pH 3.5 at 25  C operating at a capillary voltage of 30 kV and UV detection (232 nm). Samples were introduced to the capillary tube using hydrodynamic injection (50 mbar  12 s) at the cathodic (reverse polarity) end. The disaccharides in the digests were identified by their migration times relative to an internal standard (DUA,2S-GlcNCOEt,6S) and quantified against heparin disaccharide standard calibration curves (Iduron Limited, Manchester, UK).

2.1. Materials 2.4. Gel-shift assay

Crude porcine mucosal heparan sulfate (HSpm, P/N HO-03105, Lot# HO-10697) was purchased from Celsus Laboratories (Cincinnati, OH, USA). Human umbilical vein endothelial cells (HUVECs) (Merck Millipore) were provided at passage 1 and were maintained with the EndoGROÔ-Low Serum culture media kit (Merck Millipore). HUVECs from passages 2 to 8 were used for all experiments. All other cell culture-related reagents were purchased from Invitrogen/Gibco (Singapore). Recombinant VEGF165, recombinant VEGFR2 Fc chimerae, and biotinylated antibodies against VEGF165 and VEGFR2 were purchased from R&D Systems. General chemicals and reagents were from SigmaeAldrich unless otherwise stated.

Gel-shift assays were performed following the protocol of Sato and co-workers with modifications [21,22]. Briefly, VEGF165 (100 ng) was incubated with HS7þve, HS7ve or HSpm (2 mg) at room temperature for 2 h. The samples were resolved on an SDS polyacrylamide gel and blotted onto a nitrocellulose membrane (Millipore). The membrane was blocked and incubated with specific, biotinylated anti-VEGF antibody, followed by horseradish peroxidase (HRP)-conjugated streptavidin (Invitrogen). LumiGLOÒ Chemiluminescent Substrate Kits (KPL Inc., USA) were employed to visualise immunoreactive bands.

2.2. Isolation of the VEGF-affinity HS variant

2.5. Co-immunoprecipitation (Co-IP)

Purification of the VEGF-binding HS fraction was based on its affinity for the heparin-binding domain of VEGF165. A 55-amino acid peptide (AR55) corresponding to the heparin-binding domain of VEGF165 [14] (Fig. 1A) was synthesised and linked to biotin via 6-aminohexanoic acid (AHX), and the purity of this peptide confirmed to be >95% with HPLC-MS by the manufacturer (GL Biochem, Shanghai, China). Its affinity for heparin was confirmed by a [3H]-heparin assay: briefly, the peptide was immobilised onto circular nitrocellulose membranes (Millipore) approximately 7 mm in diameter. The membranes were then incubated overnight with 0.3 mCi [3H]heparin (Perkin Elmer) in 300 ml PBS/4% BSA, then washed in PBS, immersed in 2 ml of UltimaGold (Perkin Elmer) scintillation fluid and the counts per minute (CPM) retained read the next day by liquid scintillation counting. Membranes blotted with PBS served as control (Fig. 1B). Once the heparin-binding capacity of the peptide was confirmed, 5 mg of AR55-AHX-biotin was coupled to a streptavidin column (1 ml, GE Healthcare) according to the manufacturer’s protocol. After the column was equilibrated with low salt buffer (LoSB) (20 mM PBS, 0.15 M NaCl, pH 7.2), 5 mg of HSpm

VEGF165 (100 ng/ml) was incubated with VEGFR2 Fc chimerae (600 ng/ml) in the presence or absence of 1 mg/ml or 5 mg/ml of HS7þve, HS7ve or HSpm at room temperature for 10 min. The complex formed was co-immunoprecipitated with Protein A/G Agarose (Santa Cruz Biotechnology) for 1 h at 4  C. The precipitated samples were resolved on 4e12% BiseTris gels (NuPAGE, Novex, Life Technologies) and blotted onto nitrocellulose membranes. Membranes were blocked and incubated with biotinylated anti-VEGF or anti-VEGFR2 antibodies, followed by HRPconjugated streptavidin. Immunoreactive bands were visualised using the LumiGLOÒ Chemiluminescent Substrate Kit. Densitometry analyses using NIH ImageJ software were performed on 3 separate experiments to arrive at a mean value. 2.6. Glycosaminoglycan (GAG)-binding ELISA GAG-ELISA was performed according to the manufacturer’s instructions (Iduron, UK). Briefly, each well was coated overnight with 1 mg of each nominated HS in

Fig. 2. The characterisation of HS7þve, HSpm and HS7ve. Disaccharide composition by capillary electrophoresis (CE) revealed that two disaccharide species (peaks I and III, corresponding to DUA,2S-GlcNS,6S and DUA-GlcNS,6S, respectively) were enriched in HS7þve compared to the composition of the other HS preparations.

C. Wang et al. / Biomaterials 35 (2014) 6776e6786 standard assay buffer (SAB; 100 mM NaCl, 50 mM NaOAc, 0.2% v/v Tween-20, pH 7.2). Wells incubated with SAB in the absence of HS served as controls. The next day, all wells were blocked with blocking buffer (BB; 0.4% w/v fish gelatin in SAB) for 1 h. VEGF165 or VEGFR2 were added to the wells at varying concentrations and incubated for 2 h. After incubation with biotinylated antibodies for 1 h and ExtrAvidin-AP for 30 min, wells were incubated with SigmaFAST pNPP for 40 min for colorimetric development. Absorbance was measured at 405 nm. Except for pNPP, which was prepared in water and incubated at room temperature, all other reagents were prepared in BB and incubated at 37  C.

6779

2.7. Surface-plasmon resonance (SPR)-based measurement of competitive ability of VEGF165 to bind heparin-coated SA sensor chip Biotinylated heparin was prepared as previously described [23] and immobilised onto a streptavidin (SA) chip and competitive inhibition of VEGF165 binding to the heparin support detected using a BIACORE T100 surface-plasmon resonance instrument as per the manufacturer’s protocols (GE Healthcare, Sweden). Briefly, 0.1 mg/ml of heparinebiotin conjugate prepared in 0.05% v/v Tween-20 in HBS-EP buffer (10 mM HEPES, 150 mM NaCl, pH 7.4) was injected into running buffer (HBS-

Fig. 3. Binding of HS fractions to VEGF165. (A) Gel-shift assay of either VEGF165 alone or complexed with HS7þve, HS7ve or HSpm (left to right on the gel) confirmed greater binding capacity of the retained fraction. Data is representative of two independent observations. (B) The binding ability of HS7þve coated on a GAG-binding plate for VEGF165 was assessed against its controls. The results (mean  SD) are representative of at least three independent experiments, done in triplicate. (C) Competitive inhibition of VEGF165 binding to a heparin-coated SPR chip by the various HS7 GAGs at nominated concentrations. The results are representative of duplicate independent experiments.

6780

C. Wang et al. / Biomaterials 35 (2014) 6776e6786 2.9. VEGFeVEGFR2 signal transduction HUVECs in 6-well plates were treated with the heparinase mixture then treated with VEGF165 (10 ng/ml), pre-incubated with or without various HSs (10 mg/ml). Protein samples were collected in 2 Laemmli buffer, resolved on NuPAGE (Life Technologies), and blotted onto nitrocellulose membranes. Membranes were blocked and incubated with specific primary antibodies (Cell Signaling Technology Inc.) followed by appropriate HRP-conjugated secondary antibodies (Jackson Immunoresearch). The LumiGLOÒ Chemiluminescent Substrate Kit was employed to visualise immunoreactive bands. 2.10. Cell proliferation Cell proliferation was assayed by using the BrdU Cell Proliferation Kit (Roche). HUVECs were seeded in 96-well plates in media deprived of all growth factors and supplements. VEGF165 (10 ng/ml) was pre-incubated with each of the various HSs at concentrations of 0.1, 1, 10 and 50 mg/ml for 30 min at room temperature before it was added to cells. After 24 h incubation at 37  C in the presence of various VEGF165eHS combinations, BrdU was added and allowed to incorporate into proliferating cells for 24 h. The level of BrdU was detected by specific antibodies and substrates according to the manufacturer’s instructions. Absorbance was measured at 450 nm with correction at 630 nm. The proliferation of HUVECs treated with the Erk1/2 inhibitor, U0126 (10 mM), was also assayed as described [24]. 2.11. Tube formation assay The effect of HS7þve on tube formation by HUVECs grown on Type I collagen gels was next investigated. Briefly, 300 ml of rat tail Type I collagen (BD Biosciences) was applied to 24-well plates and incubated at 37  C to allow the formation of collagen fibrils. HUVECs were seeded at a density of 8  104 cells/well for 1 h at 37  C. Media was removed and 200 ml of collagen was overlaid on the cells and incubated for 1 h at 37  C. Media was then added on top of the collagen gel layers. VEGF165 pre-incubated with or without HS was then added to the media. A total of 6 images per treatment group were taken after 20 h incubation at 37  C. The average branch point values and average tube length values were analysed using NIH ImageJ software. 2.12. In vivo chicken chorioallantoic membrane (CAM) assay

Fig. 4. Effect of HS fractions on VEGF165eVEGFR2 interaction. (A) The binding ability of HS7þve coated on a GAG-binding plate for VEGFR2 was assessed against its controls. The result (mean  SD) is representative of at least three independent experiments, done in triplicate. (B) Representative immunoblot of VEGF165 and VEGFR2 Fc chimerae co-immunoprecipitated in the presence of either 1 or 5 mg/ml of HS7þve, HS7ve or HSpm, with corresponding densitometry analysis from three separate experiments (*p < 0.05). EP buffer containing 0.1% v/v Tween-20) and applied to the sensor chip at a flow rate of 30 ml/min. Competitive inhibition was achieved by premixing samples of VEGF165 (1.25 nM) with 0, 5, 10, or 20 mg/ml of HS7þve, HS7ve or HSpm that were then injected separately over the heparin chip at a flow rate of 30 ml/min for 120 s followed by washing with running buffer for 600 s. Prior to the injection of each premixed sample, the heparin-coated sensor chip was regenerated by injecting 2 M NaCl to remove all surface bound proteins.

2.8. Flow cytometry HUVECs seeded in 6-well plates were treated with a mixture of heparinase I, II and III for 15 min at 37  C to remove endogenous HS [15], and then cultured with VEGF165 (20 ng/ml) pre-incubated separately with the various HSs (20 mg/ml). The cells were collected with TrypLE (Gibco), fixed with Flow Cytometry Fixation buffer (R&D Systems), and labelled with PE-conjugated anti-VEGF antibody (R&D Systems). Non-treated cells labelled with PE-conjugated IgG2A isotype served as a control. Flow cytometry analysis was performed on the BD FACSArray system. Data were analysed on the FlowJo workstation (Tree Star, USA).

Hatching eggs (Chew's Farm, Singapore) were incubated at 37  C. On day 4 of embryonic development (E4), eggs were cracked open and embryos cultured ex ovo after they were transferred to 100 mm culture dishes (Greiner Bio-One). At E7, cut filter paper w2 mm in size (Whatman, UK) was placed between the two main vessels of the chorioallantoic membrane as described previously by Ribatti et al. [25]. Five experimental treatment groups were investigated: (i) PBS control, (ii) VEGF165, (iii) HS7þve, (iv) HS7ve and (v) HSpm. HS and VEGF were applied at 10 mg and 50 ng per membrane, respectively. Each treatment was loaded onto the filter paper and photomicrographs taken immediately. These served as the baseline values. At E10, photos of the vessels around the filter paper were taken again. The formation of new vessels that converged on each implant was analysed and scored as described by Ribatti et al. with modifications [25]. Vascular networks within a 1 mm diameter ring superimposed on the filter paper implant were scored on a 0e5 scale, with 0 depicting no change over time: 1e2 slight increments in vascular density; and 3, 4 and 5 increasing density in vessel formation surrounding the filter paper (Supp. Fig. A). The vascular coefficient is obtained by calculating the ratio of the degree of vessel density to the highest attainable score. 2.13. Statistical analyses Quantitative data is reported as the mean  standard deviation. Statistical analyses were performed using one-way ANOVA (GraphPad v5), with significance set at a p-value < 0.05.

3. Results 3.1. Purification and characterisation of HS variants To generate an HS variant with increased affinity for the proangiogenic factor VEGF165, we utilised affinity chromatography by employing a peptide-based affinity trap consisting of the 55-amino acid heparin-binding domain of VEGF165 (AR55) [18,19] (Fig. 1A). The affinity of this peptide for heparin was confirmed by filterbinding assays with [3H]-heparin (Fig. 1B). The immobilised peptide was then used as the ligand for separation of bound (HS7þve) versus unbound (HS7ve) from the HSpm starting material (Fig. 1C & D). The bound material (HS7þve) represented 9% of the starting sugar mixture by weight. The disaccharide composition of each fraction was then determined by capillary electrophoresis (CE) (Fig. 2). Two disaccharide

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

6781

Fig. 5. HS7þve modulates the activity of VEGF165. (A) VEGF165 binding to the HUVEC cell surface was confirmed by flow cytometry. Heparinase (Hsase) pre-treatment of HUVECs reduced the amount of VEGF165 bound to the surface of HUVECs, a deficit that was recovered only by the addition of HS7þve. The data is representative of duplicate experiments. (B) Effect of HS fractions on VEGF165-induced receptor activation and signal transduction in HUVECs. Cells were pre-treated with heparinase then treated with VEGF165 pre-incubated with or without various HSs for 5 or 30 min then lysed for immunoblot analyses of signal transduction proteins. The results are representative of three independent observations.

species were found to be enriched in HS7þve compared to HS7ve or HSpm, namely the proportions of DUA,2S-GlcNS,6S and DUAGlcNS,6S (Fig. 2 peaks I and III). Furthermore, a reduced proportion of DUA-GlcNS was also apparent in HS7þve.

3.2. Affinity of HS variants for VEGF165 In order to evaluate the relative binding affinity of HS7þve for VEGF165, a combination of gel-shift, ELISA- and SPR-based assays

6782

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

Fig. 6. Positive effect of HS7þve on VEGF165-mediated HUVEC proliferation. (A) HUVECs were first seeded in media deprived of all growth factors and supplements overnight. Nominated treatment conditions were prepared and pre-incubated at room temperature for 30 min, then added to the HUVECs. After 24 h, BrdU was added and its incorporation assayed after further 24 h. (B) The effect of U0126 on HS7þveeVEGF165-mediated proliferation confirmed pathway integrity. The results are means of three independent experiments  SD (**p < 0.01; ***p < 0.001).

was performed. When exposed to HS7þve, VEGF165 was observed to migrate at a slower rate, and resulted in an HS7þveeVEGF165 complex at a higher molecular weight. Although VEGF165 formed complexes with all HS fractions assayed, there was a considerably greater density of HS7þveeVEGF165 complexes compared to those observed with either HS7ve or HSpm (Fig. 3A). Quantification by GAG-binding ELISA confirmed that HS7þve had a relatively greater affinity for VEGF165 than HSpm or HS7ve (Fig. 3B), with only minimal binding seen in control wells (data not shown). When the concentration of VEGF165 was 200 ng/ml, HS7þve bound 1.5-fold more of the growth factor than HSpm. The ability of HS7 GAGs to competitively inhibit the binding of VEGF165 to a heparin-coated SPR chip was next assessed at GAG concentrations of 0, 5, 10 or 20 mg/ml (Fig. 3C). HS7þve was the strongest inhibitor of VEGF165 with HSpm and HS7ve only able to interfere to modest levels (Fig. 3C).

3.3. HS Interactions with VEGF165 and VEGF receptor 2 To investigate the effect of HS7þve on VEGF165 complex formation with VEGFR, we first examined whether HS7þve was capable of binding VEGFR2. Quantification by GAG-binding ELISA demonstrated that HS7þve had little binding capacity for VEGFR2 (Fig. 4A), with binding at or below control levels (data not shown). To examine whether HS7þve was capable of modulating VEGF165 interaction with VEGFR2, VEGF165 and VEGFR2 were incubated with the various HS variants, co-immunoprecipitated and analysed by Western blot (Fig. 4B). Densitometry analysis was performed and the relative amount of VEGF165 bound to VEGFR2 was normalised to the sample where no GAG was added. In the presence of HS7þve, VEGF binding to VEGFR2 increased with dose, while samples incubated with HS7ve and HSpm showed no significant increase.

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

6783

Fig. 7. HS7þve enhances VEGF165-mediated endothelial tube formation in 3D collagen matrices. HUVECs were seeded in 3D collagen matrices and then treated with VEGF165 preincubated with or without HS for 20 h. (A) Representative images from six separate observations of tube formation in each treatment group (scale bar: 200 mm) demonstrated the increased number of branching points and tube lengths in fractions containing HS7þve. The average number of branching points per image (B) and average tube length per image (C) were analysed using NIH ImageJ software (*p < 0.05; ***p < 0.001). Data obtained from six independent observations.

3.4. Signal transduction

3.6. Tube formation

To examine whether HS mediates VEGF165 cell surface binding, and results in appropriate signal transduction, a combination of flow cytometry and immunoblotting was performed (Fig. 5A and B) [15]. Upon addition of exogenous VEGF165, the level of cell surfacebound VEGF increased to 77.5% from a baseline of 51.6%. However, heparinase pre-treatment reduced this binding by 19.2% to 58.3%, whilst heparinase pre-treatment followed by the subsequent addition of HS7þve complexed with VEGF165, increased surface bound VEGF by 24.5% to 82.8%. By comparison, both HS7ve and HSpm were less effective. As VEGFR2eErk1/2 signalling is known to mediate endothelial cell growth, we next investigated the effect of HS7þve on the phosphorylation of VEGFR2 and Erk1/2 (Fig. 5B). Phosphorylation of VEGFR2 and Erk1/2 could be readily induced (within 5 min) by VEGF165, albeit transiently. This temporal signalling pattern could be markedly reduced by heparinase treatment, a result reversed only by the addition of exogenous HS7þve, and not by either HS7ve or HSpm.

The effect of HS on VEGF165-induced endothelial tube formation was next examined. Our data show that HS7þve could also enhance endothelial tube formation in 3D gel matrices (Fig. 7). After 20 h of treatment in VEGF165-free media, HUVECs failed to form tubular networks, whilst cells treated with VEGF165 readily began to form such networks (Fig. 7A). Addition of exogenous HS7þve further increased this effect (Fig. 7A), most notably in the average number of branching points (Fig. 7B) and average tube length (Fig. 7C), a result partially reproduced by addition of HSpm. Further analysis of branching points and tubular length highlighted the robust effect of HS7þve compared to the other HS variants. Treatment with VEGF165 complexed with HS7þve produced w82% more branch points and w60% longer tubes than VEGF165 treatment alone. HSpm also significantly increased the number of branch points (w50% higher than the VEGF165 group), but was less effective than HS7þve at increasing tube length (w67% longer than the VEGF165-alone group but w56% shorter than the VEGF165eHS7þve complex). By comparison, HS7ve had no effect when combined with VEGF165.

3.5. Cell proliferation

3.7. Blood vessel development

To assess the effect of HS treatment on endothelial cell growth, a BrdU proliferation assay was performed on HUVECs (Fig. 6). This demonstrated that HS7þve significantly increased VEGF165-mediated cell growth by w43% (Fig. 6A) in the presence of its targeted ligand, a result only rivalled by the highest dose of HSpm (36% increase). In contrast, HS7ve had no effect on VEGF165-mediated cell proliferation. We further confirmed that Erk1/2 activation was required for the proliferative effect of VEGF165-HS7þve complex. Inhibition of Erk1/2 phosphorylation by U0126 abolished the proliferative effect of VEGF165 with or without HS7þve (Fig. 6B).

The effects of HS7þve in promoting angiogenesis were also investigated in the organotypic CAM model (Fig. 8). A comparison of blood vessel formation within the allantoic membrane revealed that treatment with HS7þve alone or VEGF165 strongly stimulated blood vessel formation, while HSpm could only stimulate blood vessel growth to a lesser extent (Fig. 8A). Notably, treatment with carrier alone (PBS) failed to stimulate blood vessel formation, a result in line with HS7ve treatment. Vascular coefficient scoring of blood vessel formation was performed as described (Supp. Fig. A) with the data showing that treatment with either VEGF165 or HS7þve resulted in similar levels of blood vessel formation (Fig. 8B).

6784

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

Fig. 8. HS7þve increases blood vessel formation. Chick embryos were cultured ex ovo from day 4 of embryonic development (E4) to E7. At E7, cut filter papers loaded with either 50 ng of VEGF165 or 10 mg of HS were placed between two of the main vessels in the chorioallantoic membrane. Their effect on blood vessel formation was observed after three days. (A) Representative images of blood vessel formation in individual treatment conditions at E7 and E10 (scale bar: 2 mm). A magnified image of the filter paper and surrounding blood vessels obtained at E10 reveals the extensive blood vessel branching in response to VEGF165- or VEGF-targeted HS (arrows). (B) Vascular coefficient scoring was performed at E7 (to obtain baseline values) and at E10 to compare the effect of different treatments on blood vessel growth (**p < 0.01; ***p < 0.001). The results are means of a minimum of sixteen observations in each treatment group.

Most interestingly, HS7þve alone, was sufficient to elicit a vascular response similar to that observed for VEGF165. 4. Discussion In this study we sought to exploit the interaction between HS and VEGF165 in order to develop a new glycotherapeutic agent with angiogenic properties. Our strategy was to utilise affinity chromatography to purify an HS fraction (HS7þve) targeted at binding to VEGF165. This strategy is proving powerful for the isolation of unique, selective HS variants because HS binding to target ligands relies on HS chain domains with specific sulfation patterns and therefore particular disaccharide sequences [18,26]. The data presented here shows that HS7þve (a small fraction (w9% w/w) of the total starting porcine mucosal HS pool) was markedly more potent than the other HS fractions (HS7ve or HSpm) in promoting VEGF165eVEGFR2 interactions, VEGFR2 phosphorylation and

downstream Erk1/2 signalling, endothelial cell proliferation, tube formation in vitro and blood vessel development in the organotypic CAM assay. Disaccharide sequence composition forms the basis for the bioactivity of HS [12,17,27,28]. In particular, the sulfation patterns of the disaccharides within the ligand-binding domains of each growth factor, although subtly different, play pivotal roles in dictating growth factor specificity. For HS7þve, two particular disaccharides are enriched, the trisulfated DUA,2S-GlcNS,6S and the disulfated DUA-GlcNS,6S. Indeed, it is the trisulfated oligosaccharides in heparin that are known to bind VEGF165 with the highest affinity [23]. Notably, HS domains containing the disaccharides DUA,2S-GlcNS,6S and DUA-GlcNS,6S become increasingly abundant in myocardial tissue as it ages that correlates with increased VEGF165-mediated endothelial cell growth [27]. Disaccharides containing the 6-O-sulfate moiety are also known to be enriched on highly proliferative endothelial cells and are involved in binding

C. Wang et al. / Biomaterials 35 (2014) 6776e6786

the pro-angiogenic factors VEGF165 and FGF-2, so augmenting angiogenic responses in vitro and in vivo [29,30]. These two HS disaccharide moieties, enriched in HS7þve, thus appear to be key for the association of HS7þve and VEGF165, so increasing the VEGF165e VEGFR2 interactions that trigger pro-angiogenic signalling [31]. The evidence here supports the idea that the binding of HS7þve to VEGF165 is necessary for the pro-angiogenic activity of the sugar. Support for this contention is provided by the observations that HS alone cannot stimulate endothelial cell growth. VEGF165 affinity assays revealed a graded potency of HS when the starting material (HSpm) is compared to the HS7þve variant isolated from it. This was confirmed by the finding that lower concentrations of HS7þve were able to out-compete the binding of heparin to VEGF165 compared to HSpm (Fig. 3C and Supp. Fig. B). In addition, HS7þve increases the binding of VEGF165 to VEGFR2, so enhancing VEGFR2eErk1/2 signalling. Inhibition of Erk1/2 abolished the increased proliferative activity of the HS7þveeVEGF165 complex, further confirming that HS7þve acts to potentiate the VEGF165eVEGFR2eErk1/2 axis [32,33]. Perhaps the most notable finding was the activity of HS7þve in the organotypic CAM assay. Our data show that HS7þve could promote blood vessel growth in vivo without addition of exogenous VEGF165. We postulate this is because HS7þve was able to bind and sequester sufficient endogenous VEGF to provide a suitable proangiogenic microenvironment. Support for this is also provided by the data showing that HS7ve, with its relatively low affinity for VEGF165, was unable to exert an effect. This glycotherapeutic approach may prove clinically useful in cases of acute myocardial ischaemia where endogenous levels of VEGF increase w7-fold within the myocardium [34], providing a source of VEGF for HS7þve to potentiate. Such a strategy contrasts with a number of comparable peptide-based approaches that utilise mixtures of growth factors such as PDGF-BB and FGF-2 [35] or VEGF165mimicking supramolecular nanostructures [36]. Using such a glycotherapy may also provide an attractive alternative to growth factor-based treatments that, in the case of bone trauma and disease, are presenting with adverse safety profiles [37]. 5. Conclusion Here we have engineered an HS variant, HS7þve, with increased affinity for VEGF165 using a scalable platform technology and then determined its bioactivity in a range of biochemical, in vitro endothelial cell, and in vivo organotypic CAM assays. We observed significantly enhanced VEGF165-mediated pro-angiogenic events when it was complexed with HS7þve. When applied alone to the CAM preparations, HS7þve increased blood vessel formation comparable to exogenous treatment with VEGF165. The results immediately suggest a therapeutic opportunity that warrants further investigation. Acknowledgements The authors acknowledge grant support from the Biomedical Research Council, Agency for Science, Technology and Research (A*STAR) and the Institute of Medical Biology (IMB), A*STAR, Singapore. The authors thank Dr Donna Tillotson, Institute of Medical Biology, A*STAR, Singapore, for her help with technical aspects of this work. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2014.04.084.

6785

References [1] Cause-specific mortality. World Health Organization (WHO) Global Health Observatory Data Repository; 2008. [2] Simons M, Bonow RO, Chronos NA, Cohen DJ, Giordano FJ, Hammond HK, et al. Clinical trials in coronary angiogenesis: issues, problems, consensus: an expert panel summary. Circulation 2000;102:E73e86. [3] van der Laan AM, Piek JJ, van Royen N. Targeting angiogenesis to restore the microcirculation after reperfused MI. Nat Rev Cardiol 2009;6:515e23. [4] Annex BH. Therapeutic angiogenesis for critical limb ischaemia. Nat Rev Cardiol 2013;10:387e96. [5] Yla-Herttuala S, Rissanen TT, Vajanto I, Hartikainen J. Vascular endothelial growth factors: biology and current status of clinical applications in cardiovascular medicine. J Am Coll Cardiol 2007;49:1015e26. [6] Beohar N, Rapp J, Pandya S, Losordo DW. Rebuilding the damaged heart: the potential of cytokines and growth factors in the treatment of ischemic heart disease. J Am Coll Cardiol 2010;56:1287e97. [7] Karvinen H, Pasanen E, Rissanen TT, Korpisalo P, Vahakangas E, Jazwa A, et al. Long-term VEGF-A expression promotes aberrant angiogenesis and fibrosis in skeletal muscle. Gene Ther 2011;18:1166e72. [8] Henry TD, Annex BH, McKendall GR, Azrin MA, Lopez JJ, Giordano FJ, et al. The VIVA trial: vascular endothelial growth factor in ischemia for vascular angiogenesis. Circulation 2003;107:1359e65. [9] Eppler SM, Combs DL, Henry TD, Lopez JJ, Ellis SG, Yi JH, et al. A targetmediated model to describe the pharmacokinetics and hemodynamic effects of recombinant human vascular endothelial growth factor in humans. Clin Pharmacol Ther 2002;72:20e32. [10] Neufeld G, Cohen T, Gengrinovitch S, Poltorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J 1999;13:9e22. [11] Xu D, Fuster MM, Lawrence R, Esko JD. Heparan sulfate regulates VEGF165- and VEGF121-mediated vascular hyperpermeability. J Biol Chem 2011;286:737e45. [12] Cool SM, Nurcombe V. Heparan sulfate regulation of progenitor cell fate. J Cell Biochem 2006;99:1040e51. [13] Dombrowski C, Song SJ, Chuan P, Lim X, Susanto E, Sawyer AA, et al. Heparan sulfate mediates the proliferation and differentiation of rat mesenchymal stem cells. Stem Cells Dev 2009;18:661e70. [14] Krilleke D, DeErkenez A, Schubert W, Giri I, Robinson GS, Ng YS, et al. Molecular mapping and functional characterization of the VEGF164 heparinbinding domain. J Biol Chem 2007;282:28045e56. [15] Ashikari-Hada S, Habuchi H, Kariya Y, Kimata K. Heparin regulates vascular endothelial growth factor165-dependent mitogenic activity, tube formation, and its receptor phosphorylation of human endothelial cells. Comparison of the effects of heparin and modified heparins. J Biol Chem 2005;280:31508e15. [16] Nurcombe V, Goh FJ, Haupt LM, Murali S, Cool SM. Temporal and functional changes in glycosaminoglycan expression during osteogenesis. J Mol Histol 2007;38:469e81. [17] Murali S, Leong DF, Lee JJ, Cool SM, Nurcombe V. Comparative assessment of the effects of gender-specific heparan sulfates on mesenchymal stem cells. J Biol Chem 2011;286:17755e65. [18] Murali S, Rai B, Dombrowski C, Lee JL, Lim ZX, Bramono DS, et al. Affinityselected heparan sulfate for bone repair. Biomaterials 2013;34:5594e605. [19] Keyt BA, Berleau LT, Nguyen HV, Chen H, Heinsohn H, Vandlen R, et al. The carboxyl-terminal domain (111-165) of vascular endothelial growth factor is critical for its mitogenic potency. J Biol Chem 1996;271:7788e95. [20] Cardin AD, Weintraub HJ. Molecular modeling of protein-glycosaminoglycan interactions. Arteriosclerosis 1989;9:21e32. [21] Kanato Y, Kitajima K, Sato C. Direct binding of polysialic acid to a brainderived neurotrophic factor depends on the degree of polymerization. Glycobiology 2008;18:1044e53. [22] Ono S, Hane M, Kitajima K, Sato C. Novel regulation of fibroblast growth factor 2 (FGF2)-mediated cell growth by polysialic acid. J Biol Chem 2012;287: 3710e22. [23] Zhao W, McCallum SA, Xiao Z, Zhang F, Linhardt RJ. Binding affinities of vascular endothelial growth factor (VEGF) for heparin-derived oligosaccharides. Biosci Rep 2012;32:71e81. [24] Kamakura S, Moriguchi T, Nishida E. Activation of the protein kinase ERK5/ BMK1 by receptor tyrosine kinases. Identification and characterization of a signaling pathway to the nucleus. J Biol Chem 1999;274:26563e71. [25] Ribatti D, Nico B, Vacca A, Presta M. The gelatin sponge-chorioallantoic membrane assay. Nat Protoc 2006;1:85e91. [26] Ashikari-Hada S, Habuchi H, Kariya Y, Itoh N, Reddi AH, Kimata K. Characterization of growth factor-binding structures in heparin/heparan sulfate using an octasaccharide library. J Biol Chem 2004;279:12346e54. [27] Huynh MB, Morin C, Carpentier G, Garcia-Filipe S, Talhas-Perret S, BarbierChassefiere V, et al. Age-related changes in rat myocardium involve altered capacities of glycosaminoglycans to potentiate growth factor functions and heparan sulfate-altered sulfation. J Biol Chem 2012;287:11363e73. [28] Sasaki T, Larsson H, Kreuger J, Salmivirta M, Claesson-Welsh L, Lindahl U, et al. Structural basis and potential role of heparin/heparan sulfate binding to the angiogenesis inhibitor endostatin. EMBO J 1999;18:6240e8. [29] Stringer SE. The role of heparan sulphate proteoglycans in angiogenesis. Biochem Soc Trans 2006;34:451e3. [30] Ferreras C, Rushton G, Cole CL, Babur M, Telfer BA, van Kuppevelt TH, et al. Endothelial heparan sulfate 6-O-sulfation levels regulate angiogenic

6786

[31]

[32]

[33] [34]

C. Wang et al. / Biomaterials 35 (2014) 6776e6786 responses of endothelial cells to fibroblast growth factor 2 and vascular endothelial growth factor. J Biol Chem 2012;287:36132e46. Grunewald FS, Prota AE, Giese A, Ballmer-Hofer K. Structure-function analysis of VEGF receptor activation and the role of coreceptors in angiogenic signaling. Biochim Biophys Acta 2010;1804:567e80. Cross MJ, Claesson-Welsh L. FGF and VEGF function in angiogenesis: signalling pathways, biological responses and therapeutic inhibition. Trends Pharmacol Sci 2001;22:201e7. Cross MJ, Dixelius J, Matsumoto T, Claesson-Welsh L. VEGF-receptor signal transduction. Trends Biochem Sci 2003;28:488e94. Miraliakbari R, Francalancia NA, Lust RM, Gerardo JA, Ng PC, Sun YS, et al. Differences in myocardial and peripheral VEGF and KDR levels after acute ischemia. Ann Thorac Surg 2000;69:1750e3. discussion 4.

[35] Lu H, Xu X, Zhang M, Cao R, Brakenhielm E, Li C, et al. Combinatorial protein therapy of angiogenic and arteriogenic factors remarkably improves collaterogenesis and cardiac function in pigs. Proc Natl Acad Sci U S A 2007;104: 12140e5. [36] Webber MJ, Tongers J, Newcomb CJ, Marquardt KT, Bauersachs J, Losordo DW, et al. Supramolecular nanostructures that mimic VEGF as a strategy for ischemic tissue repair. Proc Natl Acad Sci U S A 2011;108: 13438e43. [37] Carragee EJ, Hurwitz EL, Weiner BK. A critical review of recombinant human bone morphogenetic protein-2 trials in spinal surgery: emerging safety concerns and lessons learned. Spine J 2011;11:471e91.