Journal of Controlled Release 194 (2014) 122–129

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Fibroblast-derived matrix (FDM) as a novel vascular endothelial growth factor delivery platform Ping Du a,b, Mintai P. Hwang a, Yong Kwan Noh a,c, Ramesh Subbiah a,b, In Gul Kim a, Soon Eon Bae d, Kwideok Park a,b,⁎ a

Center for Biomaterials, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea Dept. of Biomedical Engineering, University of Science and Technology, Daejeon 305-350, Republic of Korea Dept. of Biomedical Science, Kyung Hee University, Seoul 130-79, Republic of Korea d Dept. of Bioengineering, Clemson University, SC 29634, USA b c

a r t i c l e

i n f o

Article history: Received 11 March 2014 Accepted 26 August 2014 Available online 3 September 2014 Keywords: Angiogenesis Fibroblast-derived matrix (FDM) Human umbilical vein endothelial cells (HUVECs) Vascular endothelial growth factor (VEGF) Heparin Alginate capsule (AC)

a b s t r a c t Vascular endothelial growth factor (VEGF) is one of the most important signaling cues during angiogenesis. Since many delivery systems of VEGF have been reported, the presentation of VEGF using a more physiologically relevant extracellular matrix (ECM), however, has yet to be thoroughly examined. In this study, we propose that fibroblast-derived extracellular matrix (FDM) is a novel platform for angiogenic growth factor delivery and that FDM-mediated VEGF delivery can result in an advanced angiogenic response. The FDMs, activated by EDC/ NHS chemistry, were loaded with varying amounts of heparin. Different doses of VEGF were subsequently immobilized onto the heparin-grafted FDM (hep-FDM); 19.6 ± 0.6, 39.2 ± 3.2, and 54.8 ± 8.9 ng of VEGF were tethered using 100, 300, and 500 ng of initial VEGF, respectively. VEGF-tethered FDM was found chemoattractive and VEGF dose-dependent in triggering human umbilical vein endothelial cells (ECs) migration in vitro. When hep-FDM-bound VEGF (H-F/V) was encapsulated into alginate capsules (A/H-F/V) and subjected to release test for 28 days, it exhibited a significantly reduced burst release at early time point compared to that of A/V. The cell proliferation results indicated a substantially extended temporal effect of A/H-F/V on EC proliferation compared to those treated with soluble VEGF. For a further study, A/H-F/V was transplanted subcutaneously into ICR mice for up to 4 weeks to assess its in vivo effect on angiogenesis; VEGF delivered by hep-FDM was more competitive in promoting blood vessel ingrowth and maturation compared to other groups. Taken together, this study successfully engineered an FDM-mediated VEGF delivery system, documented its capacity to convey VEGF in a sustained manner, and demonstrated the positive effects of angiogenic activity in vivo as well as in vitro. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Vasculogenesis is a process of de novo vascular network formation during embryo development via the aggregation of mesodermderived endothelial precursors (angioblasts). Subsequent blood vessels are generated via sprouting from pre-existing vessel networks such as arteries and veins in a process known as angiogenesis. Both vasculogenesis and angiogenesis are typically involved in physiological and pathological processes, and both are regulated by multiple growth factors as well as by the surrounding extracellular matrix (ECM) environment [1,2]. Vascular endothelial growth factor (VEGF) – an endothelial cell (EC)-specific mitogen, angiogenic inducer, and mediator of vascular permeability – is one of the key cytokines implicated in these processes [3]. Not surprisingly, the efficient delivery of VEGF has been an important strategy in the field of vascular tissue engineering [4]. ⁎ Corresponding author at: Center for Biomaterials, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea. Tel.: +82 2 958 5288; fax: +82 2 958 5308. E-mail address: [email protected] (K. Park).

http://dx.doi.org/10.1016/j.jconrel.2014.08.026 0168-3659/© 2014 Elsevier B.V. All rights reserved.

While spatio-temporally controlled supply of VEGF is critical to elicit the development of functional and mature blood vessels, half-life of soluble VEGF is very short when directly injected into the host artery [5]. Dose control of VEGF is another challenge, as high doses of VEGF may lead to undesirable side effects, such as hemangioma and tumor formation. From this perspective, tightly controlled regulation of VEGF release is imperative for therapeutic applications [6]. Researchers have consequently developed a variety of growth factor delivery systems using biocompatible synthetic or natural polymers. In particular, heparin has been widely utilized in designing growth factor delivery system. Heparin, a highly sulfated glycosaminoglycan, is known to bind with many growth factors via electrostatic interactions between negatively charged sulfate groups on heparin and positively charged amino acid groups on proteins. Physical immobilization of VEGF onto heparinized collagen matrices was able to promote EC proliferation and angiogenesis in the chorioallantoic membrane [7]. In addition, co-immobilization of VEGF and angiopoietin-1 onto collagen scaffold further improved angiogenesis as compared to individual immobilization of VEGF and angiopoietin-1 [8]. Sulfated-alginate scaffolds

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

have also been recognized as an alternative for the sustained release of multiple growth factors and the consequent formation of mature blood vessels post-implantation [9]. Other researchers incorporated heparin into photocrosslinked alginate for the sustained release of growth factor without an initial burst release [10]. On the other hand, an injectable hydrogel derived from decellularized natural tissue retains sulfated glycosaminoglycans (GAGs) and such hydrogel injection with basic fibroblast growth factor (bFGF) resulted in an enhanced level of neovascularization in porcine pericardium after myocardial infarction [11]. In this study, we propose a novel platform for angiogenic growth factor delivery using in vitro cultured fibroblast-derived extracellular matrix (FDM). ECM not only provides a physical support of natural tissue architecture but also serves as a natural reservoir of various growth factors and cytokines. In fact, growth factor release from ECM is a tightly regulated process for normal tissue homeostasis [12]. From our previous experiences, FDM is extremely soft in stiffness and mainly composed of fibronectin, type I collagen, and other components. It provides ECs an excellent angiogenic microenvironment for capillary-like structure formation, which is comparable to that of Matrigel®. Fibroblasts are deeply involved in the angiogenesis via the secretion of angiogenic factors and the deposition of ECM [13,14]. Our hypothesis is that FDM-mediated growth factor delivery platform not only acts as a reservoir for the sequestration of growth factor but also provides cell-binding ECM motifs. To our best knowledge, there are few examples of employing naturally derived ECM as a growth factor delivery carrier. In order to fabricate an FDM-mediated growth factor carrier, we covalently conjugate heparin onto FDM (hep-FDM) and then immobilize soluble VEGF onto the hep-FDM. Characterization of hep-FDM and hep-FDM/VEGF has been carried out, along with the examination of the released VEGF effects on in vitro EC behaviors, such as cell adhesion, proliferation, and migration. Both release profile of VEGF and in vitro bioactivity of released VEGF are also investigated in the 3D alginate capsules. Further study assesses in vivo angiogenic effect of FDM-mediated VEGF via subcutaneous implantation of alginate capsules encapsulated with hep-FDM and VEGF into the ICR mice. This study represents another class of innovative growth factor delivery carrier via cell-derived ECM. Matrixmediated delivery platform should find many applications in vascular tissue engineering and regenerative medicine. 2. Materials and methods 2.1. Preparation of fibroblast-derived matrix (FDM) NIH3T3 mouse fibroblasts (ATCC, Manassas, VA) were cultured in Dulbecco's Modified Eagle's Medium (DMEM), supplemented with 10% fetal bovine serum (FBS), 100 IU/ml of penicillin, and 100 μg/ml of streptomycin (Invitrogen, Carlsbad, CA). Cells were seeded at a density of 2 × 104/cm2 onto thermanox plastic coverslips (Thermo Scientific, New York) in 6-well plates, and cultured for 5 or 6 days until confluence. The media was changed every two or three days. Upon confluency, the cells were treated briefly with a detergent solution containing 0.25% Triton X-100 and 10 mM NH4OH (Sigma, St. Louis. MO). After the samples were washed with phosphate buffered saline (PBS), PBS containing 50 IU/ml of DNase I and 2.5 μl/ml of RNase A (Invitrogen) was added to the samples and incubated at 37 °C for 2 h. Finally, the decellularized samples were gently washed twice with PBS. FDMs were either used immediately or stored at −70 °C in PBS. 2.2. Heparin conjugation onto FDM FDMs were washed with PBS and saturated with 0.05 M 2-(Nmorpholino) ethanesulfonic acid hydrate (MES) buffer (pH = 5.6) (M2933, Sigma), in which FDMs were serially immersed in MES buffer (10 min each in 20% and 40%, followed by 30 min in 100% MES buffer). A 0.25% (w/v) heparin working solution is prepared by adding heparin sodium (411210010, Acros Organics, New Jersey) to a freshly prepared

123

solution of 0.05 M N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride) (EDC, E7750, sigma) and 0.06 M N-hydroxysuccinimide (NHS, 130672, sigma) in MES solution; the EDC/NHS/MES solution was vigorously mixed and left for 10 min before the addition of heparin sodium. After a 10 min activation step, 2 ml of heparin working solution was added to FDM samples in a 6-well plate. After a gentle shaking (60–70 rpm) for 24 h at room temperature, the heparin-conjugated FDM (hep-FDM) was washed with PBS at least three times to remove unbound heparin. 2.3. Toluidine blue O staining Heparin-conjugated FDM was visualized by toluidine blue O staining. A 0.005% toluidine blue O (T3260, Sigma) solution was prepared in 0.01 N hydrochloric acid with 0.2% (w/v) sodium chloride (NaCl). The hep-FDM was reacted with 2 ml of 0.2% NaCl and 0.5 ml of toluidine blue solution for 2 h under shaking conditions at room temperature. A standard curve was made by preparing known amounts of heparin under the same conditions, followed by absorption measurements of heparin and toluidine blue O-reacted solutions at 630 nm using a Multiscan micropalte reader (Thermo Scientific, Rockford, IL). The appearance of a purple color on the FDM indicates the presence of heparin. All assays were performed in triplicate. 2.4. Fourier transform infrared (FT-IR) spectroscopy The functional groups in FDM, hep-FDM, and heparin sodium powder were analyzed using an attenuated total reflection (ATR)-4100 FTIR spectrometer (JASCO, Tokyo, Japan). The absorption spectra for heparin sodium powder, freeze-dried FDM, and hep-FDM were screened between 650 and 2000 cm−1. An average of 32 scans, each with a spectral resolution of 4 cm−1 was acquired for each sample; the baseline was automatically corrected using a background scan obtained in the absence of sample. 2.5. VEGF immobilization and in vitro VEGF release profile The effect of heparin dose on VEGF tethering was investigated by loading varying amounts of heparin (0, 0.5, 2.5, and 5 mg) onto FDMs in 6-well plates (n = 3), which were subsequently reacted with 100 ng of recombinant human VEGF165 (293-VE, R&D Systems, Minneapolis, MN) in 1 ml of PBS for 4 h at room temperature under a gentle shaking condition. In addition, varying doses of VEGF (100, 300 and 500 ng) in 1 ml of PBS were also reacted with 5 mg heparinconjugated FDM (n = 3). The samples obtained were abbreviated as VEGF 100, VEGF 300 and VEGF 500, respectively. To determine VEGF loading efficiency, 0.5 M EDTA (pH = 8.0) was used to harvest the immobilized VEGF on hep-FDM. After the samples were shaken with EDTA for 30 min at room temperature, they were collected using a scraper into a 15 ml tube. The mixture obtained was then vigorously shaken and centrifuged; the supernatant was tested using a human VEGF ELISA kit (DVE00, R&D systems) to quantify VEGF. VEGF loading efficiency was calculated as the percentage of immobilized VEGF to the initial loading amounts. VEGF release profiles were also obtained by incubating VEGF 100, VEGF 300, and VEGF 500 (n = 5), respectively in a 6-well plate containing 1 ml of PBS at 37 °C. 1 ml of PBS was collected at multiple time points (days 1, 2, 3, 5, 7, 10, 14, 21, and 28) and stored at − 20 °C until further analysis. Each of the samples was replenished with 1 ml of PBS. The ELISA kit was also used to determine the amount of VEGF released at each time point. 2.6. EC attachment and proliferation test Human umbilical vein endothelial cells (HUVECs, hereafter referred as ECs) (C2517A; Lonza, Walkersville, MD) were cultured in endothelial basal medium (EBM)-2 (CC-3156; Lonza), supplemented with EGM™-2

124

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

Single Quots™ kit (CC-4156; Lonza) at 37 °C under a 5% CO2 humidified atmosphere. ECs between passages 4 to 6 were used in the present study. To investigate the cytocompatibility of modified FDM, ECs were seeded at a density of 1.2 × 104 cell/cm2 in 6-well plate onto one of the four groups – VEGF 0 (hep-FDM without VEGF), VEGF 100, VEGF 300, and VEGF 500 – and cultured in EBM-2 medium supplemented with 2% FBS, 100 IU/ml of penicillin and 100 μg/ml of streptomycin. Under the same conditions, ECs were seeded onto FDM in EBM-2 media supplemented with (positive control, P-con) or without (negative control, N-con) 25 ng/ml of VEGF in 2 ml media (n = 3). After 24 h of culture, a LIVE/DEAD® Viability/Cytotoxicity Kit (03224, Invitrogen) was used to evaluate cell viability. Live and dead cells were visualized as green and red fluorescence, respectively, using a fluorescent microscope (U-RFL-T, CKX-41, Olympus). Meanwhile, cell proliferation assay was carried out using the same groups as above. Medium was changed every two or three days and cultured for 5 days; 25 ng/ml of VEGF in 2 ml media was added as a single dose for the P-con group. A Cell Counting Kit-8 (CCK-8; Dojindo, Japan) assay was employed on days 2 and 5 to evaluate the cell proliferation rate in each group. 10% CCK-8 solution was added to each sample and incubated at 37 °C for 2 h. An aliquot of each sample (200 μl) was transferred to 96-well plates and the absorbance was measured at 450 nm using a Multiscan microplate reader (Thermo Scientific, Rockford, IL). All proliferation assays were performed in triplicate. 2.7. EC migration assay The capacity of hep-FDM-bound VEGF to act as a chemoattractant for ECs was investigated using 6-well-sized transwell inserts (BD Bioscience, Bedford, MA) with a pore size of 8.0 μm as illustrated in Fig. 4A. The lower chamber was loaded with FDM (P-con), VEGF 0 (hep-FDM, N-con), VEGF 100, VEGF 300, and VEGF 500 in 1 ml EBM-2 media, while 1 ml of ECs suspended (1 × 105) in EBM-2 media was added to the upper chamber. A single 50 ng dose of VEGF in 1 ml media was added to the lower chamber for the P-con group. After 24 h culture, ECs that migrated toward the bottom chamber were visualized via calcein AM staining. Migrated cells were quantified by counting the number of cells in five non-overlapping fields (40× magnification) from each sample (n = 3). 2.8. VEGF encapsulation into 3D alginate capsules and in vitro release profile The VEGF tethered on 2D matrix was expanded into a 3D system by encapsulating hep-FDM-bound VEGF into alginate capsules (ACs). 1% (w/v) sodium alginate solution was prepared by dissolving sodium alginate (Wako Pure Chemicals, Osaka, Japan; 500–600 cp) in sodium chloride (0.9 wt.%). Meanwhile FDMs from ten petri dishes (10 cm diameter) were scraped, collected in two different tubes, and subjected to centrifugation to obtain FDM pellets. 10 ml of heparin working solution, as prepared in Section 2.2, was added only to one of the FDM pellets. Heparin-conjugated FDM and non-modified FDM were separately mixed with 0.5 ml of alginate solution using a micro-homogenizer (PowerGen 125, Fisher Scientific, Waltham, UK). The test groups, alginate solution, alginate/FDM, or alginate/hep-FDM were subsequently loaded with 500 ng of VEGF and incubated for 2 h. ACs were fabricated by crosslinking alginate droplets – formed by passing each mixture through a syringe needle (26 G; Kovax, Korea) – in a solution of 0.1 M calcium chloride. ACs formed with VEGF alone (A/V), with both VEGF and FDM (A/F/V), or with VEGF and hep-FDM (A/H-F/V) were loaded equally into the top chamber of transwell inserts (n = 5). 0.5 ml of distilled water (DW) was place in both the top and bottom chambers to allow for VEGF release from the ACs. 0.5 ml of supernatant was collected from the bottom chamber at specific time points (days 1, 2, 3, 5, 7, 10, 14, 21, 28) and stored at −20 °C until further analysis; the top chamber was replenished with fresh DW at each time point. The amount of VEGF

in the ACs was determined: the freshly prepared ACs were washed several times and subjected to dissociation of the ACs in the 0.5 M EDTA (pH = 8.0) solutions. The supernatant was then collected and analyzed using an ELISA kit. The loading efficiency (%) was calculated as the percentage of encapsulated VEGF to initial loading amount of VEGF. 2.9. EC proliferation induced by VEGF released from ACs The samples, A/V, A/F/V, and A/H-F/V, each prepared as described in Section 2.8, were loaded into three transwell inserts. 5 × 104 ECs suspended in EBM-2 media were subsequently seeded into the lower chamber as shown in Fig. 6A. P-con and N-con groups are prepared using the same number of cells cultured with or without a single dose of 50 ng VEGF, respectively. Cells were cultured for 14 days and assessed for proliferation at multiple time points (days 2, 4, 7, 10, and 14) using a CCK-8 assay. Media without soluble VEGF was changed at each time point. 2.10. Examination of EC viability in ACs EC viability inside the ACs was evaluated by mixing ECs at a density of 5 × 106 cells/ml with alginate solution and 500 ng of VEGF during the ACs preparation. Encapsulated ECs were cultured in EBM-2 for 2 and 5 days and were examined via a LIVE/DEAD® Viability/Cytotoxicity Kit. Separately, the encapsulated cells were liberated on days 2 and 5 by dissolving ACs in a 0.5 M EDTA solution (pH = 8.0); live cell numbers were determined using the CCK-8 kit. Cell viability at each time point was calculated as the percentage of live cell number to the initial seeding number. 2.11. Subcutaneous transplantation of ACs The angiogenic effects of A/V, A/F/V and A/H-F/V were compared in a subcutaneous ICR mouse model (8-week old, male). For AC preparation, a total amount of 2 μg VEGF and 0.5 ml alginate solution were applied, which is quite different from the initial amount of VEGF used for in vitro study (Section 2.8). An aggregate of 4 to 5 ACs (1.7 ± 0.2 mm in diameter each) per sample was subcutaneously transplanted into the dorsal area of mice (n = 6), 1 cm distant from the incision site after induced anesthesia via intraperitoneal injection of Zoletil and Rompun (1 ml/kg). The wound was closed using 5-0 sutures and gentamicin (Samu, Korea) was then intramuscular injected in a dose of 0.04 ml per kg body weight to prevent wound infection. The transplants were retrieved 2 and 4 weeks post-op, fixed in 10% buffered formalin, paraffin-embedded, sectioned into 6 μm thick slices, and analyzed via hematoxylin & eosin (H&E) staining and immunofluorescence — CD31 and α-smooth muscle actin (α-SMA). For double immunofluorescence, sections were blocked with 3% bovine serum albumin (BSA) in PBS for 1 h, followed by overnight incubation at 4 °C with a mixture of primary antibodies: rabbit polyclonal anti-CD31 (ab28364, Abcam) diluted 1:50 and mouse monoclonal anti-α-SMA (A 2547, Sigma-Aldrich) diluted 1:1000 in 1% BSA. After washing with PBS three times, the sections were incubated with a mixture of secondary antibodies – goat polyclonal rabbit IgG-H&L (Dylight 488; ab96899, Abcam) and rhodamine (TRITC)-APure F(ab′)2 fragment donkey anti-mouse IgG (H + L) (715-026-150, Jackson Immunoresearch, West Grove, PA) in 1% BSA (1:200) – for 1 h at room temperature. The immunostained specimens were further subjected to DNA-specific 4, 6-diamidino-2- phenylindole (DAPI, 46190, Thermo scientific) staining for nucleus-specific labeling. Fluorescence images were obtained using a fluorescence microscopy (Carl Zeiss, Germany). The capillary density (number/mm2) and percentage area (%) occupied by the new capillaries were calculated from five different fields for each sample (n = 3, each group) using ImageJ software (NIH). The maturation of blood vessels was also evaluated by the ratio of α-SMA positive vessels to CD-31 positive vessels.

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

Fig. 1. Characterization of heparin-conjugated FDM. Initial heparin amounts (0.5, 2.5, or 5 mg) are used each in preparing heparin-conjugated FDMs (hep-FDM) and hep-FDMs are stained by Toluidine blue O in 6-well plate (A). Heparin contents on the hep-FDM are examined using Toluidine blue O assay (B). The columns indicate the average values and the corresponding standard deviations (n = 3). Statistically significant differences among three doses of heparin are indicated as ***(p b 0.001). Further analysis of FDM, hep-FDM and heparin powder is carried out using ATR-FTIR spectra (C).

2.12. Statistical analysis Statistical analysis of current data was conducted using one-way analysis of variance (ANOVA), with Tukey's post hoc multiple comparisons (GraphPad Prism 5). Statistical significance was determined as *(p b 0.05), **(p b 0.01), and ***(p b 0.001). 3. Results 3.1. Characterization of FDM-bound heparin The intensity of purple on hep-FDM surfaces increased with the amount of conjugated heparin as indicated by toluidine blue O staining. Moreover, the entire surface of FDM was homogenously stained, demonstrating an even distribution of heparin (Fig. 1A). The actual amount of grafted heparin was 0.42 ± 0.26, 2.36 ± 0.12 and 5.04 ± 0.98 μg for initial heparin loading amounts of 0.5, 2.5 and 5 mg, respectively (Fig. 1B). Additionally, ATR-FTIR spectra of heparin sodium powder and hep-FDM (5 mg heparin) exhibit an overlap of sulfate (SO3) peaks at 1230 cm− 1 and 1040 cm−1 — peaks inherent in heparin powder (Fig. 1C). These spectra are also present in the FDM. 3.2. Determination of FDM-bound VEGF and in vitro release profile

125

to the amount of heparin. In addition, 5 mg of heparin-grafted FDM yielded the highest level of VEGF retention; the 5 mg of heparinloaded FDMs was used for all subsequent experiments. The VEGF retention capacity of hep-FDM was further evaluated using a constant amount of heparin and various initial VEGF doses. While 19.6 ± 0.6, 39.2 ± 3.2 and 54.8 ± 8.9 of VEGF were immobilized using 100, 300, and 500 ng of initial VEGF, respectively (Fig. 2B), the loading efficiency of VEGF was inversely proportional to the amount of VEGF loaded; a loading efficiency of 19.6% was obtained using an initial amount of 100 ng VEGF compared to 13.1% and 11.0% using 300 and 500 ng of initial VEGF, respectively. Furthermore, individual examination of the release profile of VEGF 100, VEGF 300 and VEGF 500 exhibited very similar release patterns for up to 28 days with a moderate amount of early release (Fig. 2C). This result suggests that they may share a similar release mechanism, irrespective of VEGF contents immobilized. 3.3. VEGF dose effect on EC proliferation in vitro The various substrates produced different levels of EC attachment as indicated via Live & Dead staining upon culture for 24 h (Fig. 3A). While the number of ECs was the lowest on hep-FDM substrates without VEGF tethering (VEGF 0), dead cells (red fluorescence) were barely visible throughout all test groups, indicating an inhibitory effect of heparin on EC adhesion. What's more, ECs exhibited less spreading on hep-FDM and N-con (FDM w/o VEGF). After 2 days of culture, cell number in Pcon – treated with 25 ng/ml of soluble VEGF in 2 ml media – increased by about twice the number initially seeded (1.2 × 105), but decreased with N-con, VEGF 0, and VEGF 100 groups, further confirming the inhibitory role of heparin on EC proliferation (Fig. 3B). However, we were able to observe a counteraction of this inhibition effect at higher VEGF doses. Interestingly, samples cultured for 5 days exhibited a greater level of cell proliferation for VEGF 300 and VEGF 500 compared to P-con that rather declined. Taken together, ECs were proliferated in a VEGF dose-dependent manner. 3.4. Chemoattractive effect of immobilized VEGF EC migration toward hep-FDM bound VEGF was investigated as shown in Fig. 4A. ECs placed in the transwell insert were cultured for 24 h in the presence of hep-FDM/VEGF underneath. Not surprisingly, few cells migrated to the bottom chamber in the absence of VEGF (N-con) (Fig. 4B). VEGF tethered hep-FDM (VEGF 100, VEGF 300 and VEGF 500), however clearly exhibited a positive effect on EC

The amount of VEGF immobilized onto hep-FDM was examined for varying initial amounts of heparin and VEGF. With an initial dose of 100 ng VEGF per sample, the amount of tethered VEGF significantly increased as that of heparin increased; 2.5 ± 0.06, 2.9 ± 0.4, 5.5 ± 1.7 and 19.6 ± 0.59 ng of VEGF were immobilized onto 0, 0.5, 2.5, and 5 mg of heparin-loaded FDMs, respectively (Fig. 2A). Those results indicate that the VEGF retention capacity of hep-FDM is directly proportional

Fig. 2. Determination of hep-FDM bound VEGF and in vitro release profile of VEGF. Initial amount of 100 ng VEGF per sample (n = 3) is incubated with various hep-FDMs for immobilization of VEGF and the level is measured by human Quantikine ELISA kit (A). In addition, different amounts of VEGF (100, 300, and 500 ng) are examined to measure real amount of VEGF immobilized and the corresponding efficiency is determined using hep (5 mg)-FDM (B). In addition, the samples of hep-FDM tethered VEGF (100, 300, and 500 ng) (n = 5) is subjected to a release test in vitro for 28 days (C). Statistically significant differences among the test groups are indicated as **(p b 0.01), and ***(p b 0.001).

Fig. 3. EC attachment and proliferation with varying amounts of VEGF immobilized hepFDM. Various FDM-based substrates are prepared: P-con, N-con, VEGF 0 (hep-FDM), VEGF 100 (hep-FDM/VEGF-100 ng), VEGF 300, and VEGF 500. P-con and N-con represent ECs cultured on FDM with or without 25 ng/ml soluble VEGF in 2 ml media. EC attachment after 24 h is observed using Live & dead staining (A). Scale bar is 200 μm. EC proliferation rate is investigated using CCK-8 assay at day 2 and day 5, respectively (n = 3) (B). Statistically significant differences are indicated as **(p b 0.01), and ***(p b 0.001) (B).

126

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

Fig. 4. Chemoattractive effect of hep-FDM bound VEGF. ECs (1 × 105) are loaded into the top chamber of transwell insert in 6-well plate and FDM-based substrates are placed on the bottom as illustrated in schematic (A). (P-con and N-con: with or without 50 ng VEGF in media). After 24 h culture of ECs, cell migration toward bottom plate is visualized by calcein AM staining (×40) (B). Scale bar is 500 μm. The number of migrating ECs is obtained by counting live cells on the bottom under ×40 magnification at 5 different fields (n = 3). Statistically significant differences are indicated as *(p b 0.05), and ***(p b 0.001) (C).

motility that is highly proportional to the VEGF doses. The number of migrated cells was approximately 3-fold, 15-fold, and 19-fold for VEGF 100, 300, and 500, respectively more than that of N-con (Fig. 4C). The difference was also statistically significant. The chemoattractive effect of VEGF 500 was comparable with that of P-con (50 ng of soluble VEGF added in medium). 3.5. VEGF encapsulation into alginate capsules and in vitro release profile Optical images of alginate capsules (ACs) demonstrate the uniformity in size (1.7 ± 0.2 mm diameter) and morphology, and a homogenous distribution of FDM and hep-FDM within the ACs (Fig. 5A). As indicated by release profiles of VEGF from three AC groups, the incorporation of FDM into the ACs substantially decreased the burst release of VEGF at early time points (Fig. 5B). The presence of hep-FDM within the ACs

Fig. 6. Bioactivity of VEGF released from ACs. Bioactivity of released VEGF from ACs is evaluated by loading ACs in the transwell insert and by seeding ECs on the bottom as illustrated in schematic (A). Five different groups are prepared and evaluated for EC proliferation during 14 days (B): N-con (without VEGF in media) and P-con (with 50 ng VEGF treated once). Statistically significant differences are indicated as *(p b 0.05), **(p b 0.01), and ***(p b 0.001).

further reduced the burst release of VEGF. The percentage of cumulative release for A/V, A/F/V, and A/H-F/V was 78 ± 4.0, 54 ± 3.5, and 31 ± 1.5, respectively, on day 5. By day 7, the release of VEGF from A/V was reduced significantly, whereas that from A/H-F/V was maintained up to 28 days. Furthermore, the co-incorporation of FDM itself and VEGF increased the VEGF loading efficiency from 14.5 ± 2.5% to 21.9 ± 2.1%, and further elevated it to 34.4 ± 4.7% upon incorporation of hep-FDM (Fig. 5C). 3.6. Bioactivity of VEGF released from ACs Bioactivity of the VEGF released from ACs was evaluated by measuring its effect on EC proliferation at different time points (Fig. 6A). For the first 2 days of culture, cells in P-con (25 ng/ml of VEGF initially added to media) showed the highest degree of proliferation. However, on day 4, VEGF delivered via A/V or A/F/V was more effective in inducing cell proliferation than that delivered in P-con; this trend was maintained until 10 days (Fig. 6B). Interestingly, for culture times up to 14 days, VEGF encapsulated in A/H-F was effective in promoting the continuous proliferation of ECs; similarly, cell proliferation in A/H-F/V surpassed all other groups at day 14. These results indicate that VEGF released from ACs remains bioactive for at least 14 days, and that the co-incorporation of VEGF and hep-FDM within ACs promotes a sustained release of VEGF and consequently stimulates EC proliferation. The difference in cell numbers was statistically significant between A/H-F/V and the other groups at day 14. 3.7. Cell viability of ECs encapsulated in ACs

Fig. 5. Alginate capsules (ACs) encapsulation and in vitro release profile of VEGF. Gross and phase contrast images of ACs (1.7 ± 0.2 mm diameter) are obtained from the ACs, with VEGF alone (A/V), with FDM and VEGF (A/F/V), or with hep-FDM and VEGF (A/H-F/V) (A). Scale bar is 500 μm. The release profile of each group (n = 5) is monitored for 28 days and shows that A/H-F/V is very effective in suppressing a burst release of VEGF at the early time point (B). The A/H-F/V also exhibits much better loading efficiency of VEGF (C). Statistically significant differences are indicated as **(p b 0.01), and ***(p b 0.001).

ECs were cultured inside the ACs at different conditions to investigate the effect of encapsulated VEGF on cell viability. Under live and dead staining, more viable cells in green fluorescence are detected in A/H-F/V as compared to A/V and A/F/V (Fig. 7A). Upon quantitative analysis as assessed via CCK-8 assay, the viability of encapsulated ECs cultured in ACs varies significantly across the different test groups, with cells in the A/V group showing the poorest viability on day 2 (Fig. 7B). While cells encapsulated in A/V exhibited a significant decrease in viability on day 5, those encapsulated in A/F/V showed better viability that of A/V. Furthermore, cells co-incorporated with VEGF tethered hep-FDM presented much more improvement in viability and the difference was statistically significant on day 5. 3.8. Histological analysis of subcutaneously implanted ACs Subcutaneously implanted ACs into ICR mice were retrieved 2 and 4 weeks post-op. H&E staining of the sectioned slices exhibited quite

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

127

Fig. 7. Cell viability of ECs encapsulated in ACs. The ACs containing ECs and VEGF or FDM/ VEGF or hep-FDM/VEGF are cultured for up to 5 days and then stained using Live & Dead assay kit (×40) (A). Scale bar is 500 μm. Cell viability of each group (n = 3) is examined at days 2 and 5 using CCK-8 assay (B). The presence of FDM shows an improvement on cell viability compared with EC encapsulated with VEGF alone. A co-delivery of VEGF and hepFDM reveals a further improvement on cell viability. Statistically significant differences among the test groups are indicated as *(p b 0.05), **(p b 0.01), and ***(p b 0.001).

different levels of cellular infiltration and blood vessel development inside the AC implants. For A/F/V and A/H-F/V at 2 weeks, some capillaries with erythrocytes were detected at the center of ACs and these ACs were partially degraded, whereas few cells and capillaries could penetrate inside the A/V (Fig. 8A). At 4 weeks, the angiogenic response was more active as the density of microvessels persistently increased with A/F/V and A/H-F/V, compared to the A/V that found a very poor microvessel formation inside the ACs (Fig. 8B). In a separate study, a simple immunogenic response of FDM was evaluated by measuring cytokine TNF-α level of mouse raw cells. The result clearly shows that the level of TNF-α as stimulated by FDM was significantly low compared to that of LPS-activated group (Fig. S1). And the TNF-α level elevated as the FDM dose increased. Taken together FDM seems to cause a very mild immunogenic response under the given test condition.

Fig. 9. Immunofluorescence of CD31 and quantitative analysis of neovessel formation. Thin sections of three different groups are subjected to immunofluorescent staining of CD31 and representative images (×200) are shown (A). Scale bar is 200 μm. Quantitative analysis of CD31 positive signals results in blood vessel density (B) and vessel occupied area (C). Statistically significant differences are indicated as **(p b 0.01) and ***(p b 0.001).

the vessel occupied area that A/H-F/V showed the largest vessel occupied areas at both 2 and 4 weeks (Fig. 9C).

3.9. Neovascularization

3.10. Blood vessel maturation

CD31-positive blood vessels were easily detected inside the implants with A/F/V and A/H-F/V, whereas for A/V most of blood vessels were found at the interface between implants and host tissue. Apparently, higher blood vessel density, lager lumen diameter, and better EC linings were observed with A/H-F/V compared to A/V and A/F/V at 2 and 4 weeks, respectively (Fig. 9A). When blood vessel density and their occupied area were quantitatively analyzed using CD31-positive images (×200), the average blood vessel densities per mm2 were 1.67 and 2.73-fold larger with A/F/V and A/H-F/V than that of A/V at 2 weeks (Fig. 9B). Moreover the gap became much larger to 3.2 and 5.4 times, respectively at 4 weeks. A similar trend was observed with

Neovessel coverage by pericytes or smooth muscle cells is a prerequisite for the stabilization, maturation, and eventual functionality of blood vessels; an inner lumen stained by EC marker CD31 and an outer layer stained by SMC marker α-SMA is an indicative of mature vessel. Accordingly, blood vessel maturation was further evaluated via double immunofluorescence staining against CD31 and α-SMA. The delivery of VEGF from ACs induced active blood vessels formation at 2 weeks as CD31 (green) staining shows in all the groups but most of which lack the coverage of pericyte (α-SMA, red) (Fig. 10A). However an increase of vessel diameter and pericyte coverage level was noticed at 4 weeks. An examination of the ratio between the number of mature vessels and the total number of vessels demonstrates an increase of mature vessels with time. The highest percentage of them was notable with A/H-F/V compared to the other groups at 4 weeks (Fig. 10B).

Fig. 8. Histological analysis of subcutaneously implanted ACs. Retrieved ACs after 2 (A) and 4 weeks (B) implantation are sectioned into slices, subjected to H&E staining, and photographed at lower (× 50), middle (×100) and high (×400) magnification: Insets (top) are the macroscopic images of retrieved samples. The red arrows indicate the interface between ACs and host tissue (top). The black arrows mark not only cellular infiltration but also the newly formed capillaries inside the ACs (bottom). Scale bars are 500, 200, and 100 μm from top to bottom, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 10. Investigation of blood vessel maturation: immunofluorescence of CD31 and αSMA. Thin sections of each group are subjected to immunofluorescence of both CD31 (green) and α-SMA (red) and the representative images are presented (A). Mature vessel percentage is quantitatively determined via the number of both CD31- and α-SMApositive vessels to only CD-31-positive vessels (B). Statistically significant difference is indicated as *(p b 0.05) and **(p b 0.01). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

128

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

4. Discussion Angiogenic growth factor delivery is one of the most studied approaches for the promotion of neovascularization of engineered tissues. Spatially and temporally controlled release of growth factors is of particular interest to improve the therapeutic efficacy of delivered growth factors. A wide variety of delivery tools have been investigated in numerous literatures. Among the typical formats are polymeric microor nano-particles, micro- or nano-gels, films, fibers, hydrogels, and scaffolds. Compared with the previous studies, the novelty of this study is the use of cell-derived ECM as a growth factor delivery carrier. ECM in vivo not only provides cell-anchoring sites, but also acts as growth factor reservoir for the sequestration and protection of bioactive molecules [15]. In this work, FDM is modified using heparin for VEGF immobilization and alginate was utilized as a delivery carrier of VEGFtethered FDM. Heparin is conjugated to FDM via EDC chemistry and VEGF is then tethered onto the hep-FDM. The efficiency of VEGF immobilization was proportional to the initial amount of heparin. When loaded with 100 ng of VEGF, using 5 mg of heparin per FDM improved the loading efficiency of VEGF by 8-fold than FDM alone. Early studies have reported that VEGF is securely immobilized to collagen scaffolds, matrices, patches, and polycaprolactone scaffolds via an EDC reaction and used to promote EC penetration and proliferation as well as in vivo neovascularization for cardiac repair [4,16–18]. In addition, decellularized tissue-derived ECM that retains native sulfated GAGs, is also utilized as a delivery platform for heparin-binding growth factors [11]. However, some information about loading efficiency and/or actual amount of immobilized growth factor is often unspecified in these literatures, such that direct comparison to current FDM-mediated system is impossible. Instead, conventional ECMs, such as fibronectin and gelatin were processed under the same protocol applied in this study. The results showed that both ECMs were very poor in the efficacy of VEGF immobilization compared to FDM (data not shown). Meanwhile, VEGF bound to heparin-modified FDM in 2D and 3D environments can be released in a sustained manner for up to 28 days (Figs. 2C and 5B), suggesting that the interaction between heparin and VEGF is secure and effective. Especially in alginate capsules (ACs), hep-FDM bound VEGF release profile was significantly distinguished and exhibited a sustained release of VEGF with a low level of initial burst compared to that of A/V. Furthermore, bioactivity of delivered growth factor is another critical issue in judging the efficacy of delivery carrier. Current data support that the bioactivity of VEGF is maintained in 2D and in 3D setting as assessed by EC proliferation. Compared to 50 ng bolus addition of VEGF to culture media, the immobilized VEGF (54.8 ± 8.9 ng in VEGF 500) had a more pronounced effect on EC proliferation on day 5 (Fig. 3B); this difference is likely due to the early loss of VEGF activity when directly added into culture media. Furthermore, VEGF released from ACs also enhanced EC proliferation in a similar manner (Fig. 6B). In agreement with the previous studies – demonstrating that the effectiveness of immobilized growth factors on mitogenesis is greater than that of a soluble form [16,19] – our results indicate a positive effect of hep-FDM in reserving the bioactivity of VEGF for a longer period. In addition, VEGF-loaded hep-FDM also proves chemoattractive in EC migration (Fig. 4). This phenomenon is of particular interest because endothelial progenitor cell homing and EC migration are critical steps in neovascularization. Earlier report has shown that PEGfibrinogen-loaded VEGF induces significant EC migration compared to that induced with gel alone [20]. Taken together, these results demonstrate the bioactivity of VEGF delivered from hep-FDM on the proliferation and migration of ECs, both are important biological processes for angiogenesis. For more practical approach of FDM-mediated VEGF delivery, current system was tailored to a 3D platform using alginate, a widely used hydrogel for growth factor encapsulation [21]. In fact, the previous studies have documented sulfated alginate hydrogel as a growth factor

delivery system [9,22]. However, the purification of sulfated alginate is time-consuming and the yield of modified alginate is very low. Cellbinding sites are also not available with sulfated alginate. The advantage of present system is that a simple mix of hep-FDM and alginate is possible without modifying the mechanical and chemical properties of alginate. Basically, a particular interest of this study is the feasibility of cell-derived matrix-mediated growth factor delivery system, along with the use of alginate simply as a carrier material. This study is not intended to directly compare the effects between sulfated alginate/ VEGF and A/hep-FDM/VEGF. The rationale of incorporation of heparinFDM into alginate is that based on our previous works, FDM itself acts as not only cell-anchoring motifs but also an excellent angiogenic platform we have thus attempted to upgrade naked FDM into VEGFtethered FDM in order to find angiogenesis-related applications. Regarding the size of current alginate capsules, they seem quite larger compared with the conventional ones. In fact the control of AC size is not a major issue at this time, because a primary interest is to show the possibility of matrix-mediated VEGF delivery. Meanwhile, when ACs were subcutaneously implanted into ICR mice to assess the angiogenic effects in vivo, we observed significantly more new blood vessels invasion 2 and 4 weeks post-implantation in the group of ACs encapsulating H-F/V. One of the reasons is likely that compared to A/V and A/F/V, A/H-F/V is more efficient in reserving VEGF securely and/or in suppressing the loss of VEGF during the preparation of delivery system. Based on the loading efficiency in vitro (Fig. 5), estimated VEGF amounts of in vivo samples suggest that A/H-F/V holds much more VEGF (data not shown). The data can also be supported by the chemoattractive effect of VEGF-tethered hep-FDM as shown in Fig. 4. Accordingly, more blood vessel-related cells may be recruited around ACs and then infiltrated inside the ACs. At this point, it is notable that which route the neighboring cells take to penetrate inside the ACs. Actually, alginate hydrogel is not a closed system, rather contains many pores on the surface. While the alginate stability in vivo is not very sure, it is plausible that crosslinked alginate becomes loose with time. Another chance is that degradable FDM fragments can serve as a cell invasion route in the alginate. In light of research done by others indicating that newly formed vessels in a subcutaneous area – normally not under physiological demand – easily regress upon an absence of VEGF supplementation [23,24], we believe that our system effectively provides a robust local supply of VEGF in the form of hep-FDM for the development of mature vessels. The vessel maturation we observed was achieved via single growth factor delivery (VEGF), and is in agreement with a previous study demonstrating a four-fold increase of α-SMApositive blood vessels over those in a saline group when delivering VEGF via a PEG-fibrinogen system [20]. Nonetheless, it should be noted that significantly lower amount of VEGF was employed in the current animal study compared to the other reports (1–10 μg). In fact, the concentration of delivered VEGF plays a critical role in determining the nature and persistence of vasculature. Long-term delivery of higher dose of VEGF (1500 ng/day) as compared to low dose VEGF (150 ng/day) initially elevated scaffold vascularization but tended to regress after 20 days [25]. In addition, it is notable that we believe that the investigation of VEGF concentration effect on FDM-mediated delivery is another major work. Besides the effect of VEGF itself, the fate of ECM with time is also a particular interest, because FDM is destined to degrade by various proteolytic enzymes in vivo and the resultant ECM fragments have been implicated in active or passive angiogenic activity [26]. Although current data lack such evidences about the role of those ECM fragments, future study is worthwhile to address this issue. In addition, we observed no obvious hemorrhaging at the implantation site, leading us to assure that side effect of heparin conjugated to FDM is minimal. Taken together, the capacity of hep-FDM as a convenient and highly efficient approach to the delivery of VEGF is a major finding of this study. VEGF tethered by hep-FDM retained its biological activity to promote EC migration and proliferation in a 2D platform. The angiogenic activity in vivo as assessed via subcutaneous implantation of 3D

P. Du et al. / Journal of Controlled Release 194 (2014) 122–129

hydrogels suggests the potential of this carrier system as a therapeutic vehicle for angiogenesis. 5. Conclusions In this study, FDM is modified by heparin not only to enhance the loading efficiency of an angiogenic growth factor (VEGF), but also to yield a sustained release of VEGF in both 2D and 3D environments. Results on various modified 2D FDM substrates strongly indicate that the bioactivity of VEGF is retained by heparin and that EC proliferation exhibits a dose-dependent response to the amount of tethered VEGF. It is also notable that more ECs migrate toward hep-FDM/VEGF, mainly due to the chemoattractive effect of immobilized VEGF on the FDM. Additionally, cell viability within the ACs is greatly better with the inclusion of hep-FDM bound VEGF. Furthermore, outcomes from the subcutaneous implantation of ACs demonstrate significantly improved neovascularization with VEGF-tethered group (A/H-F/V) after 2 and 4 weeks of implantation, respectively. Taken together, this work suggests a huge potential of using hep-FDM as an efficient delivery tool of angiogenic growth factors for vascular tissue engineering applications. Acknowledgments This work was supported by the intramural grants 2V03350 (KIST) and 2E25113 (KIST) from the Ministry of Science, ICT and Future Planning, Republic of Korea. This study was also funded by a grant from the Korean Health Technology R&D Project (A120216), Ministry of Health & Welfare. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2014.08.026. References [1] M. Potente, H. Gerhardt, P. Carmeliet, Basic and therapeutic aspects of angiogenesis, Cell 146 (2011) 873–887. [2] W. Risau, Mechanisms of angiogenesis, Nature 386 (1997) 671–674. [3] N. Ferrara, H.-P. Gerber, J. LeCouter, The biology of VEGF and its receptors, Nat. Med. 9 (2003) 669–676. [4] Y. Miyagi, L.L.Y. Chiu, M. Cimini, R.D. Weisel, M. Radisic, R.-K. Li, Biodegradable collagen patch with covalently immobilized VEGF for myocardial repair, Biomaterials 32 (2011) 1280–1290. [5] D.F. Lazarous, M. Shou, M. Scheinowitz, E. Hodge, V. Thirumurti, A.N. Kitsiou, J.A. Stiber, A.D. Lobo, S. Hunsberger, E. Guetta, S.E. Epstein, E.F. Unger, Comparative effects of basic fibroblast growth factor and vascular endothelial growth factor on coronary collateral development and the arterial response to injury, Circulation 94 (1996) 1074–1082. [6] A.B. Ennett, D. Kaigler, D.J. Mooney, Temporally regulated delivery of VEGF in vitro and in vivo, J. Biomed. Mater. Res. A 79A (2006) 176–184.

129

[7] G.C.M. Steffens, C. Yao, P. Prevel, M. Markowicz, P. Schenck, E.M. Noah, N. Pallua, Modulation of angiogenic potential of collagen matrices by covalent incorporation of heparin and loading with vascular endothelial growth factor, Tissue Eng. 10 (2004) 1502–1509. [8] L.L.Y. Chiu, M. Radisic, Scaffolds with covalently immobilized VEGF and angiopoietin-1 for vascularization of engineered tissues, Biomaterials 31 (2010) 226–241. [9] I. Freeman, A. Kedem, S. Cohen, The effect of sulfation of alginate hydrogels on the specific binding and controlled release of heparin-binding proteins, Biomaterials 29 (2008) 3260–3268. [10] O. Jeon, C. Powell, L.D. Solorio, M.D. Krebs, E. Alsberg, Affinity-based growth factor delivery using biodegradable, photocrosslinked heparin-alginate hydrogels, J. Control. Release 154 (2011) 258–266. [11] S.B. Seif-Naraghi, D. Horn, P.J. Schup-Magoffin, K.L. Christman, Injectable extracellular matrix derived hydrogel provides a platform for enhanced retention and delivery of a heparin-binding growth factor, Acta Biomater. 8 (2012) 3695–3703. [12] S.A. Eming, J.A. Hubbell, Extracellular matrix in angiogenesis: dynamic structures with translational potential, Exp. Dermatol. 20 (2011) 605–613. [13] F. Berthod, L. Germain, N. Tremblay, F.A. Auger, Extracellular matrix deposition by fibroblasts is necessary to promote capillary-like tube formation in vitro, J. Cell. Physiol. 207 (2006) 491–498. [14] A.C. Newman, M.N. Nakatsu, W. Chou, P.D. Gershon, C.C. Hughes, The requirement for fibroblasts in angiogenesis: fibroblast-derived matrix proteins are essential for endothelial cell lumen formation, Mol. Biol. Cell 22 (2011) 3791–3800. [15] G.S. Schultz, A. Wysocki, Interactions between extracellular matrix and growth factors in wound healing, Wound Repair Regen. 17 (2009) 153–162. [16] Y.H. Shen, M.S. Shoichet, M. Radisic, Vascular endothelial growth factor immobilized in collagen scaffold promotes penetration and proliferation of endothelial cells, Acta Biomater. 4 (2008) 477–489. [17] S. Singh, B.M. Wu, J.C.Y. Dunn, The enhancement of VEGF-mediated angiogenesis by polycaprolactone scaffolds with surface cross-linked heparin, Biomaterials 32 (2011) 2059–2069. [18] S.T. Nillesen, P.J. Geutjes, R. Wismans, J. Schalkwijk, W.F. Daamen, T.H. van Kuppevelt, Increased angiogenesis and blood vessel maturation in acellular collagen-heparin scaffolds containing both FGF2 and VEGF, Biomaterials 28 (2007) 1123–1131. [19] Y. Ito, Covalently immobilized biosignal molecule materials for tissue engineering, Soft Matter 4 (2008) 46–56. [20] A.J. Rufaihah, S.R. Vaibavi, M. Plotkin, J. Shen, V. Nithya, J. Wang, D. Seliktar, T. Kofidis, Enhanced infarct stabilization and neovascularization mediated by VEGFloaded PEGylated fibrinogen hydrogel in a rodent myocardial infarction model, Biomaterials 34 (2013) 8195–8202. [21] F. Gu, B. Amsden, R. Neufeld, Sustained delivery of vascular endothelial growth factor with alginate beads, J. Control. Release 96 (2004) 463–472. [22] I. Freeman, S. Cohen, The influence of the sequential delivery of angiogenic factors from affinity-binding alginate scaffolds on vascularization, Biomaterials 30 (2009) 2122–2131. [23] F. Baffert, T. Le, B. Sennino, G. Thurston, C.J. Kuo, D. Hu-Lowe, D.M. McDonald, Cellular changes in normal blood capillaries undergoing regression after inhibition of VEGF signaling, Am. J. Physiol. Heart Circ. Physiol. 290 (2006) H547–H559. [24] M. Ehrbar, S.M. Zeisberger, G.P. Raeber, J.A. Hubbell, C. Schnell, A.H. Zisch, The role of actively released fibrin-conjugated VEGF for VEGF receptor 2 gene activation and the enhancement of angiogenesis, Biomaterials 29 (2008) 1720–1729. [25] N. Davies, S. Dobner, D. Bezuidenhout, C. Schmidt, M. Beck, A.H. Zisch, P. Zilla, The dosage dependence of VEGF stimulation on scaffold neovascularisation, Biomaterials 29 (2008) 3531–3538. [26] J.D. Mott, Z. Werb, Regulation of matrix biology by matrix metalloproteinases, Curr. Opin. Cell Biol. 16 (2004) 558–564.