Long-term biostability and bioactivity of “fibrin linked” VEGF₁₂₁in vitro and in vivo

Abstract

Despite major advances in understanding angiogenesis over the last few years, the ability to induce angiogenesis in ischemic wounds or larger tissue-engineering constructs remains elusive. Serious risks and limited control over dose, duration, and localization of growth factor delivery make materials-based approaches viable alternatives. In an effort to minimize passive diffusion and control the release profile of delivered growth factors, matrix properties have been engineered with regard to pore size, growth factor affinity or stable growth factor binding. Recently, fibrin or biomimetic hydrogels have been engineered towards the covalent immobilization of vascular endothelial growth factor (VEGF). Most of the studies pertaining to VEGF delivery by fibrin gel constructs have focused on characterizing release profiles, receptor activation, and the angiogenic response in vitro and in vivo. Herein we demonstrate that gels containing covalently-linked VEGF (α₂PI_1–8-VEGF₁₂₁), compared to diffusible VEGF, elicit stronger and longer-lasting angiogenic responses in subcutaneous implants of mice. This superior angiogenic response was due to both the sustained release and significant retention of bioactivity (80%) of the delivered engineered VEGF over a 12-day period. To the best of our knowledge, this is the first report to characterize long-term matrix liberated α₂PI_1–8-VEGF₁₂₁ bioactivity, important for future efforts in angiogenesis research.

---

1. Introduction

Therapeutic vascularization and generation of intricate vascular networks needed to sustain the viability of tissue engineered constructs of more than a few cubic millimeters in volume remain significant challenges in regenerative medicine.^1,2 Physiological processes such as wound healing, the formation of new vessels during the menstrual cycle and pregnancy, or the adaption of the vascular network to metabolic demands rely on angiogenesis, the formation of new vessels from preexisting ones.^1–3 Angiogenesis is controlled by the tight regulation of soluble- or matrix-associated pro- and anti-angiogenic signals, resulting in a complex morphogenetic process consisting of (i) disintegration of the basement membrane, (ii) sprouting of the endothelial tip cells, (iii) lumen formation, and (iv) vessel maturation.^4–7

The isolation of potent pro-angiogenic growth factors such as fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), or angiopoietin-1 (Ang-1) has not only advanced our understanding of their molecular and cellular mechanisms but also prompted their use as therapeutic agents.^8–10 Numerous preclinical and clinical studies have investigated the effects of these pro-angiogenic growth factors in therapeutic settings and have confirmed their potential benefits in re-perfusing ischemic tissues.⁹ Treatment regimens have capitalized on viral-mediated gene therapies, the transplantation of engineered growth factor-secreting cells, delivery of plasmids, and administration of the growth factor protein from polymeric scaffolds or hydrogels.^11–15 These clinical studies and other basic science efforts elucidated the need for growth factor-based therapeutic applications to match a narrow therapeutic window. In nature, the distribution and availability of VEGF is tightly regulated by expression, binding to the extracellular matrix (ECM), and its release by cell-mediated proteolytic activity.^16,17 Thus, it is not surprising that the most frequently isolated human VEGF isoforms are either freely diffusible (VEGF₁₂₁) or bind heparan sulfate proteoglycans (HSPG) of the ECM with intermediate (VEGF₁₆₅) or very high affinity (VEGF₂₀₆) (reviewed in ref. 18). While Ozawa et al. demonstrated that excessive microenvironmental concentrations of VEGF could result in the formation of malformed and dysfunctional vessels,^19,20 others show the importance of the localized matrix immobilization by different VEGF isoforms in vitro and in vivo.^21,22 Additionally, studies in Keshet's laboratory indicate that newly formed capillaries of the vascular system require a sustained, long-term exposure to VEGF until their association with mural cells materializes.^23–25

The relatively high risk of cell- and gene therapy-based treatments, along with the difficulty in controlling the distribution and duration of the produced signals, has made the administration of growth factor proteins an attractive alternative. Despite its high potency, VEGF has been shown to rapidly diffuse from the site of treatment. It also undergoes high clearance and proteolytic degradation in vivo, all of which necessitate the use of supra-physiological doses of this highly potent growth factor.²⁶ So far, outcomes of clinical trials (when compared to preclinical efforts) have been disappointing. Considering VEGF's relatively low in vivo efficiency and potentially adverse side effects, a novel slow-delivery regimen that utilized naturally-derived or synthetic biodegradable hydrogels such as fibrin, collagen, alginate, and hyaluronic acid (HA) or (polyethylene) glycol (PEG), poly(N-vinyl-2-pyrrolidone), and poly(acrylamide) were sought.^27–31

In an attempt to mimic the naturally-occurring growth factor immobilization within the extracellular matrix (ECM), our group developed α₂PI_1–8-Pla-VEGF₁₂₁, a variant of VEGF₁₂₁ that is designed to covalently bind to the forming fibrin hydrogel, a matrix with limited natural binding capacity for growth factors.^32,33 During fibrin polymerization, α₂PI_1–8-VEGF₁₂₁ readily cross-links to fibrin via the transglutaminase activity of factor XIII. Unlike soluble VEGF₁₂₁ admixed with fibrin, fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ is protected from diffusion and is liberated in minute, biologically effective levels into the surrounding tissue via local, cell-mediated fibrinolysis.^15,34 Experiments performed on chick chorioallantoic membranes and subcutaneously in mice demonstrated a higher number of structurally-normal vessels with fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ compared to those of the non-fibrin conjugated VEGF₁₂₁.¹⁵ Longitudinal studies (utilizing transgenic mice expressing luciferase controlled by a VEGF-receptor promoter) have indicated that continuously delivered fibrin-conjugated VEGF₁₂₁ promotes the formation of mature vessels at three weeks of application. However, these vessels were prone to regression in the absence of a metabolic demand.³² By this effort, having strived to provide VEGF₁₂₁ stimulation over longer periods, with the goal to establish mature, fully functional vascular networks, the long-term bioavailability and stability of α₂PI_1–8-VEGF₁₂₁ was not formally demonstrated. In this longitudinal study, we sought to correlate the long-term bioactivity and bioavailability of α₂PI_1–8-VEGF₁₂₁ with the angiogenic response it induces at different times of implantation.

2. Materials and methods

2.1 Fibrin gel formation

Fibrin gels of 20 μL were made by mixing human fibrinogen (Sigma-Aldrich, Switzerland) solution with thrombin (Sigma-Aldrich, Switzerland) to final concentrations of 10 mg mL⁻¹ human fibrinogen, 10 U mL⁻¹ factor XIIIa (Fibrogammin, CSL Behring, Switzerland), 0.02 U mL⁻¹ human thrombin, and 2.5 mM Ca²⁺ in 20 mM Tris, pH 7.5, and 150 mM NaCl (TBS). Gels containing 10 μg mL⁻¹ aprotinin (Sigma-Aldrich, Switzerland) and 2 μg of soluble VEGF₁₂₁ (PeproTech, UK), aprotinin and 2 μg covalently-bound α₂PI_1–8-VEGF₁₂₁ ¹⁵ or aprotinin alone (negative control) were obtained by adding specific concentrations of each to the constituents. Fibrin gels were allowed to polymerize in a 5% CO₂ incubator at 37 °C for 30 minutes before being transferred into medium or implanted in mice.

2.2 Stability and bioreactivity in vitro assays

In order to determine the influence of fibrin immobilization on biostability in vitro, fibrin-conjugated α₂-Pl_1–8-VEGF121 was incubated in proteinaceous solutions (0.1% BSA in TBS) at 37 °C. VEGF liberation by gel solubilization was performed for 24 hours at 37 °C using 3 mU μL⁻¹ human plasmin (Sigma, Switzerland) in 20 μL TBS per 20 μL fibrin gel.³⁴ Residual quantities of plasmin-liberated α₂PI_1–8-VEGF₁₂₁ were quantified via ELISA at 7, 14, and 21 days.

2.3 Subcutaneous fibrin gel implantation

All animal procedures were approved by the ethical committee of the Cantonal Veterinary Office Zurich. Female mice (CD1nu:Crl1><No><Crl:CD1-Foxn1<nu>(ICRnu) (Charles River, Germany)) of about 4–5 weeks of age were used for this study. All materials were prepared prior to surgery.

Mice were anesthetized with 5% isoflurane until their breathing rate was measured to be approximately 16 to 20 breaths per minute. Isoflurane was then maintained at 2%. The dorsal skin was disinfected with 70% ethanol and allowed to air dry. Two incisions of approximately 1.5 cm were made with a scalpel along the left and right sides of the dorsum orthogonal to the spine. A 10-0 Dafilon suture (Braun, Germany) was placed through fibrin gel pellet to fasten the gel to the subcutaneous area just beneath the skin. Different groups of gels (VEGF₁₂₁, α₂PI_1–8-VEGF₁₂₁, or negative control without growth factor), all containing 10 μg mL⁻¹ aprotinin, were implanted in a randomly assigned order. 7-0 Prolene sutures (Ethicon, USA) were then used to close the wound. Two knots were made with special care to avoid pinching any excess skin within the knots. Mice were placed in a recovery zone and pain medication (Temgesic® 2 Amp, 0.3 mg mL⁻¹, Essex Chemie, Switzerland) was administered on demand once per day subcutaneously. Animals were euthanized after 4, 8, and 12 days via carbon dioxide asphyxiation. Gels were then excised with some surrounding skin and connective tissue. If the samples were to be assessed histologically, the skin and connective tissue were maintained. Otherwise, they were extricated and the remaining gel was stored at −80 °C for further analysis. Gels were completely digested using plasmin as described above and analyzed via ELISA and bioactivity assays.

2.4 Assessment of VEGF-mediated signaling

Human umbilical vein endothelial cells (HUVEC) (PromoCell, USA) were seeded into 6-well plates (200 000 cells per well) in complete EGM-2 medium (Lonza, Switzerland) and allowed to adhere for 24 hours. Cells were then serum-starved for 6 hours in EBM-2 medium containing 1% FCS (Lonza, Switzerland) and subsequently stimulated for 10 minutes using 50 ng mL⁻¹ matrix-liberated VEGF₁₂₁, pervanadate was used as a positive control and plasmin digested fibrin gels with no growth factor as a negative control. Afterwards, the HUVEC were washed twice with ice cold PBS and lysed using 1× SDS sample buffer containing 100 mM sodium orthovanadate. Cell lysates were run on 7.5% or 12% SDS-PAGE gels and transferred onto nitrocellulose membranes (GE Healthcare, Switzerland) using a wet transfer method. Membranes were blocked for 2 hours at room temperature in blocking buffer containing 10% non-fat milk in TBS with 0.1% TWEEN®-20 (Sigma-Aldrich, Switzerland) and incubated with the primary antibodies (1 : 3000 dilution) against pERK Thr202/Tyr204, ERK1/2, pVEGFR-2 Tyr1175, VEGFR-2, and VEGF (Cell Signaling, USA) at 4 °C for another 12 hours. After three 5-minute washes in blocking buffer, blots were incubated with secondary IgG donkey anti-rabbit-HRP antibody (GE Healthcare, Switzerland, 1 : 5000 dilution) for 2 hours. Membranes were washed again, as before, and once with TBS. The membranes were then developed using ECL solution (GE Healthcare, Switzerland) and exposed to autoradiography films.

2.5 Quantification of VEGF (ELISA)

Human-specific VEGF Enzyme-Linked Immunosorbent Assay (ELISA) (PeproTech, UK) was performed at room temperature and according to the manufacturer's instructions. Between the various incubation steps, wells were washed four times with 400 μL washing buffer (PBS, 0.1% Tween-20). Triplicates of each sample (diluted in 0.1% BSA in PBS to a total volume of 100 μL) were incubated for at least 2 hours in capture antibody-coated and blocked ELISA plates. Incubation with detection antibody was performed for 2 hours, followed by Avidin-HRP conjugate incubation for 30 minutes. 100 μL of the substrate solution was added to each well for color development. The enzymatic reaction was stopped by adding 100 μL of stop solution to each well. Plates were immediately measured with ELISA plate reader (BioTek, Switzerland) at 405 nm with the wavelength correction set to 650 nm.

2.6 Quantification of VEGF bioactivity by endothelial cell proliferation

HUVEC were cultured in 48-well plates (5000 cells per well) containing EGM-2 and stored under a humidified atmosphere (37 °C, 5% CO₂). Cells were subsequently starved for 6 hours with EBM-2 containing 1% FCS. After this, HUVEC were stimulated with varying concentrations of VEGF₁₂₁ or α₂PI_1–8-VEGF₁₂₁ and allowed to incubate for 72 hours. Afterwards, 30 μL of Cell Proliferation Reagent (WST-1, Roche, Switzerland) was added to each well; the cells were then incubated for 4 hours. Subsequently, the plates were thoroughly shaken for approximately 1 minute. Three 100 μL aliquots from each 48-well plate were transferred to a 96 well plate and absorbance of the samples was measured against a background control as a blank using a micro-plate reader at 420–480 nm with the reference wavelength set to 650 nm.

2.7 Histological and immunohistochemical analysis

Tissue specimens were fixed in 10% paraformaldehyde and embedded in paraffin. 5 μm thick histological sections were cut perpendicular to the skin using a rotary microtome. Double staining was performed on BondMax slides (Leica, Germany). Leica reagents were utilized for the entire procedure according to the manufacturer's guidelines. Primary antibodies for CD31 (1 : 200, Dianova GmbH, Germany) and smooth muscle actin (1 : 1000, Epitomics, Austria) in PBS containing 5% bovine serum albumin (BSA) and 0.3% Triton-X were applied after 20 minutes of pre-treatment with Epitope-Retrieval-Buffer 2 (Leica, Germany) at 100 °C. For smooth muscle actin, which appeared brown, a rabbit anti-rat secondary antibody linked with horseradish peroxidase (1 : 1000 dilution, Abcam, Switzerland) was used and detection was performed with the Refine-HRP-Kit (Leica, Germany). For CD31, which appeared red, a rabbit anti-rat secondary IgG linked with alkaline phosphatase (1 : 200, Abcam, UK) was used and detection was performed with the Refine-AP-Kit (Leica, Germany). Hematoxylin (Leica, Germany) was used as a counterstain. Fold-increase in vessel density was determined by using a computer-generated grid from ImageJ that was laid over each micrograph. Vessels intersecting the grid area were counted.

2.8 Statistical analysis

All experimental data are depicted as mean ± standard deviation. A p value <0.05 was considered to be statistically significant. One-way ANOVAs were performed to determine statistical significance using SPSS (Version 19) statistical analysis software. Post hoc multiple comparisons followed, if warranted (Holm–Sidak method). Graphs were generated using Graph-Pad Prism 4.0.

3. Results

3.1 α₂PI_1–8-VEGF₁₂₁ stability and bioactivity in vitro

Though many studies have investigated the biological effects of covalently-bound growth factors, such as VEGF, in vitro and in vivo,^15,32,33,35 there is insufficient knowledge about the long-term stability of matrix-immobilized, engineered growth factors. In an effort to strengthen our understanding and give precedence to future biomaterial research endeavors, we sought to characterize the biostability and bioactivity of α₂PI_1–8-VEGF₁₂₁ immobilized in fibrin matrices. In order to determine the influence of fibrin immobilization on biostability in vitro, fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ was incubated in proteinaceous solutions (0.1% BSA in TBS) at 37 °C prior to gel solubilization and VEGF liberation by plasmin digestion. However, residual quantities of plasmin-liberated α₂PI_1–8-VEGF₁₂₁, as determined by ELISA, decreased upon fibrin gel immobilization. There was no statistical significant reduction at 7 (p = 0.979), 14 (p = 0.965) or 21 (p = 0.665) days of incubation when compared to time point 0 (Fig. 1A). Furthermore liberated VEGF did not show any unspecific degradation as assessed by Western blot analysis (data not shown).

Fig. 1 In vitro stability and bioactivity of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁. Fibrin gels containing immobilized α₂PI_1–8-VEGF₁₂₁ were incubated at 37 °C in TBS containing 0.1% BSA. Fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ was liberated at days 0, 7, 14 or 21 by plasmin degradation of the fibrin carrier. (A) In vitro stability of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ as determined by the amount of liberated VEGF using ELISA. (B) Residual bioactivity of liberated α₂PI_1–8-VEGF₁₂₁ by HUVEC proliferation. 50 ng mL⁻¹ of liberated α₂PI_1–8-VEGF₁₂₁ was used to stimulate proliferation of HUVECs (5000 cells per well) for 72 hours before metabolic activity was determined by WST-1 (NC: digested fibrin gel without VEGF; PC: fibrin gel plus freshly-thawed, untreated α₂PI_1–8-VEGF₁₂₁; n = 5; mean ± SD). (C) VEGF specific downstream signaling by liberated α₂PI_1–8-VEGF₁₂₁. HUVEC cells were stimulated with 50 ng of liberated α₂PI_1–8-VEGF₁₂₁, protein from cell extracts were separated by SDS-PAGE, blotted, and probed for either VEGFR-2, pVEGFR-2, ERK1/2, and pERK1/2. Experiments were repeated at least twice. (PC: pervanadate was used as a positive control; NC: un-stimulated HUVECs were used as a negative control.)

Perhaps more important than the long-term stability of VEGF is its long-term residual bioactivity. HUVEC proliferation assays demonstrated that 50 ng mL⁻¹ matrix-liberated α₂PI_1–8-VEGF₁₂₁ had virtually the same bioactivity as 50 ng mL⁻¹ of freshly-thawed α₂PI_1–8-VEGF₁₂₁, resulting in a 175% induction of proliferation when compared to the control (no stimulation; endothelial cell base medium) (Fig. 1b). However, there was a moderate but continuous drop in long-term bioactivity of α₂PI_1–8-VEGF₁₂₁, which resulted in a reduced induction proliferation of 150% at day 14 (p = 0.002) and 140% on day 21 (p = 0.006) respectively. Furthermore, to demonstrate that matrix-liberated α₂PI_1–8-VEGF₁₂₁ activity maintains its ability to activate VEGFR-2 mediated signaling, the phosphorylation of VEGFR-2 and the downstream signaling kinase ERK1/2 (extracellular-signal regulated kinase) was studied. Western blots demonstrate that both VEGFR-2 and ERK activity were increased for all time points, including day 21 (Fig. 1C).

3.2 In vivo angiogenic potential of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁versus soluble VEGF₁₂₁

In order to test the fibrin gels localized angiogenic potential, fibrin implants containing no growth factor (negative control), soluble VEGF₁₂₁ or fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ were subcutaneously implanted in immunocompromised mice. Whereas tissues exposed to empty fibrin matrices showed no grossly visible signs of new vessel formation, those exposed to matrices carrying native VEGF₁₂₁ exhibited an induction of angiogenesis after four days of treatment. To contrast the temporary angiogenic induction exhibited by VEGF₁₂₁-releasing fibrin, the slow releasing fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ constructs stimulated a robust formation of blood vessels that persisted for up to twelve days (Fig. 2).

Fig. 2 Skin tissue response to fibrin implants. Representative photographs of skin regions that were exposed for 4, 8 and 12 days to 20 μL fibrin gels containing either 2 μg soluble VEGF₁₂₁ (n = 4), 2 μg α₂PI_1–8-VEGF₁₂₁ (n = 5) or no VEGF (negative control) (n = 4).

3.3 In vivo stability of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁

The ability to provide a continuous and localized release of low, therapeutically-relevant levels of angiogenic molecules is essential to sustained angiogenic therapy for ischemic tissues. As the release of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ is coupled to the fibrin plug stability, the amount of fibrin present at different times of implantation was demonstrated by haematoxylin- and eosin-stained histological sections collected from the middle of the gel implantation site (Fig. 3). For all conditions, the fibrin implants were continuously degraded over the course of the experiment.

Fig. 3 Histological sections from fibrin gel implant sites. Representative haematoxylin- and eosin-stained histological sections of fibrin gels containing either 2 μg soluble VEGF₁₂₁ (n = 4), 2 μg α₂PI_1–8-VEGF₁₂₁ (n = 5) or no VEGF (negative control) after 4, 8, and 12 days of subcutaneous implantation. The fibrin gel (marked by a star) can be seen as an amorphous plug that disappears over time.

3.4 Determination and quantification of neo-vascular density via immunohistochemistry

In order to determine the number and quality of newly formed blood vessels by the fibrin-mediated delivery of VEGF, collected histological sections from the site of implantation were immunohistochemically stained using anti-CD31 (which detects endothelial cells, red) and anti-SMA (anti-smooth muscle actin, brown) antibodies (Fig. 4A). Comparable to earlier studies where fibrin gels were transplanted in silicone-based cups,³² sites treated with soluble VEGF₁₂₁ contained significantly more vessels compared to those treated with empty fibrin matrices after 4 days of implantation (Fig. 4B). While the number of vessels produced by soluble VEGF₁₂₁ treatment was more than two-fold higher compared to controls on day 8, a significant reduction of vessels became apparent by day 12. Impressively, while all vessels induced by soluble VEGF were unstable and regressed by 12 days, ultimately bringing the vessel density back to the levels exhibited by the controls, α₂PI_1–8-VEGF₁₂₁ yielded a significant fraction of stable vessels, ensuring a greater than 2-fold increase in vessel density compared to controls by 12 days.

Fig. 4 Quantification of microvessel density in subcutaneous tissues exposed to VEGF-releasing fibrin gels by histomorphometry. (A) Representative micrographs of skin regions exposed to fibrin gels containing either 2 μg soluble VEGF₁₂₁ (n = 4), 2 μg α₂PI_1–8-VEGF₁₂₁ (n = 5) or no VEGF (negative control) after 4, 8, and 12 days of implantation. Endothelial cells appear red (CD31; indicated by arrow heads) and smooth muscle cells appear brown (SMA; indicated by stars). (B) Quantification of the angiogenic response. Changes of CD31-positive microvessel numbers of tissue areas from the same animal treated by fibrin gels containing either 2 μg VEGF₁₂₁ (n = 4) or 2 μg α₂PI_1–8-VEGF₁₂₁ (n = 5) in comparison to negative control (fibrin gels without VEGF) at 4, 8, and 12 days per grid unit area. (p < 0.01).

3.5 In vivo stability and bioactivity of fibrin-conjugated α₂PI_1–8-VEGF₁₂₁

In vivo growth factors maintain a half-life and bioavailability on the order of hours largely due to the effects of proteolytic cleavage and processing via clearance. For VEGF₁₂₁, the half-life was noted to be between two and three hours.^39,40 To determine whether immobilizing VEGF to the fibrin network (i) prevented passive diffusion of the factor from the treatment site and (ii) protected the growth factor from proteolytic degradation, residual growth factor quantity and bioactivity were determined using explanted fibrin gels. Fibrin gels containing immobilized α₂PI_1–8-VEGF₁₂₁, soluble non-bound VEGF₁₂₁, or negative control were harvested at 4, 8 and 12 days post subcutaneous implantation and digested with plasmin. ELISAs revealed that fibrin gels originally containing soluble VEGF₁₂₁ exhibited no residual growth factor by day 4. In stark contrast, 88% of the fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ initially present in the gels was recovered (p < 0.001) at the same time point. The amount of residual α₂PI_1–8-VEGF₁₂₁ dropped to 54% by day 8 (p < 0.001) and 15% (p = 0.003) by day 12 (Fig. 5A).

Fig. 5 Long-term in vivo stability and bioactivity fibrin-conjugated α₂PI_1–8-VEGF₁₂₁. Fibrin gels containing 2 μg VEGF₁₂₁ (n = 4), 2 μg α₂PI_1–8-VEGF₁₂₁ (n = 5) or no VEGF (negative control) were explanted from subcutaneous pockets of mice 4, 8 and 12 days after implantation. The explanted fibrin gels were subsequently digested in plasmin (3 mU μL⁻¹) for 24 hours. (A) The amount of VEGF from plasmin-digested fibrin gels was determined in triplicate by ELISA. (B) The residual bioactivity of 50 ng liberated α₂PI_1–8-VEGF₁₂₁ was determined by the stimulation of HUVEC (5000 cells per well) proliferation. Fresh α₂PI_1–8-VEGF₁₂₁ was used as a positive control. Data are given as mean ± SD (statistical significance (*) denotes p < 0.05).

As not only protein stability but also bioactivity might be compromised during in vivo implantation, the bioactive potential of VEGF₁₂₁ and α₂PI_1–8-VEGF₁₂₁ recovered from mice was determined in a way similar to the in vitro process described above. HUVEC proliferation was measured in response to stimulation by 50 ng mL⁻¹ of liberated growth factor (Fig. 5B). When compared to fresh (i.e., not implanted into a mouse) α₂PI_1–8-VEGF₁₂₁, matrix-liberated α₂PI_1–8-VEGF₁₂₁ from explanted gels at day 4 retained 90% bioactivity; 80% bioactivity was still preserved at day 12.

Discussion

This study demonstrates the improved in vivo angiogenic response of engineered α₂PI_1–8-VEGF₁₂₁, compared to native VEGF₁₂₁. This enhancement is the result of the engineered factor's efficient, covalent binding to clinically relevant fibrin gels, sustained delivery, and residual bioactivity after liberation by plasmin.

Soluble, non-matrix-bound growth factors, such as VEGF₁₂₁, are prone to fast clearance and proteolytic degradation, resulting in an in vivo half-life in the range of two to three hours.^39,40 The development of a relevant delivery regimen for VEGF will critically depend on the preservation of the factor's bioactivity within the delivery device in vivo. Various delivery modalities for VEGF, a potent regulator of angiogenesis, have been developed based on controlled diffusion from carriers, altering its affinity to matrices, and inducing covalent binding to matrices via chemical or enzymatic strategies. Using these different strategies, induction of vascularization has been demonstrated in various in vivo models.^32,36–38

Our group has engineered a VEGF variant that, by means of factor XIII-mediated transglutaminase activity, is efficiently bound to the fibrin matrix during polymerization.³² We have also previously demonstrated that engineered VEGF was released in an active, proteolysis-dependent manner in vitro.¹⁵ Data presented herein indicated that the bioactivity of the immobilized growth factors incubated in aqueous solution at 37 °C in vitro remained almost unchanged for up to three weeks (Fig. 1).

To correlate the VEGF₁₂₁ release kinetic with the angiogenic response, fibrin gels were implanted subcutaneously into the backs of mice. By explanting fibrin hydrogels at various time points after implantation, we demonstrated that native VEGF₁₂₁ diffused out of the hydrogel. Further, less than 5% of the initial quantity was available by days 4 or 8 (Fig. 5A). In stark contrast, approximately 88% of engineered fibrin-conjugated α₂PI_1–8-VEGF₁₂₁ was retained by day 4; thereafter, it linearly decreased through day 12. The concomitant histological evaluation of the fibrin hydrogels revealed that the release profile was as expected due to the proteolytic degradation of the fibrin matrix (Fig. 3). This clearly confirms earlier studies where fibrin-conjugated slow-release VEGF implanted in subcutaneous pouches induced the formation of more patent and leakage-resistant vessels as compared to native fast-release VEGF.¹⁵

In earlier longitudinal studies utilizing biophotonic monitoring of VEGFR-2 gene activation in transgenic mice, VEGFR-2 expression was up-regulated in response to both wild-type VEGF₁₂₁ and α₂PI_1–8-VEGF₁₂₁; no detectable difference in the angiogenic response was seen with exposure to either growth factor.³² The robust increase in microvascular density in the absence of up-regulated VEGFR-2 expression pointed towards a better bioavailability of α₂PI_1–8-VEGF₁₂₁ compared to VEGF₁₂₁. Since (a) the in vivo release kinetics of the growth factors were only correlated to remaining fibrin mass and (b) use of silicone cups as carriers for the biomaterial could interfere with both the consumption of the fibrin, final conclusions of the duration of active growth factor release could not be drawn. Thus it could only be surmised that the amount of VEGF released over a prolonged period of time was not sufficient to induce a robust increase in microvascular density. The operator-induced induction of luciferase expression additionally led to large variations of the measured signals, such that precise conclusions were not possible. Also, as previously mentioned, the interference of silicone cups with both the consumption of the fibrin and the host response towards the implant could not be excluded. Importantly, by harvesting of fibrin plugs after various times of subcutaneous implantation we were able to demonstrate, by endothelial cell proliferation experiments, that fibrin-bound α₂PI_1–8-VEGF₁₂₁ retained most of its bioactivity (80%) after 12 days in vivo (Fig. 5B).

The fibrin gels utilized here were almost degraded within 12 days, presumably due to the fast release of aprotinin. This indicates that improving the in vivo stability of the fibrin would likely result in a system which allows enhanced sustained low-level delivery over an extended period of time. A recent study showing that the fibrin hydrogel stability can be significantly improved by the covalent incorporation of the plasmin inhibitor aprotinin⁴¹ holds great promise for optimized delivery of fibrin-bound α₂PI_1–8-VEGF₁₂₁. Alternatively, materials with predictable and tunable proteolytic properties could be employed to deliver VEGF for context-dependent applications.^27–31 Both of these strategies are viable routes to achieve a delivery regimen that presents sufficient amounts of bioactive VEGF over a period of at least four weeks, which is considered to be critical for the induction and stabilization of new vessels by peri-endothelial cells.^23–25

Conclusions

The induction of therapeutically relevant, patent vessels requires the prolonged delivery of angiogenic factors such as VEGF. We have provided evidence that an active, slow release of VEGF results in more, functionally and morphologically superior blood vessels compared to passive, fast-release formulations. Although engineered VEGF was assumed to possess a superior long-term bioavailability, conclusive experimental data were not available. In response, we have successfully demonstrated that in vivo tissue exposure to VEGF followed fundamentally different kinetics (burst versus slow release) dependent on whether the native or engineered forms were employed. Since the bioactivity of the matrix-immobilized VEGF only modestly decreased within the observed period, we concluded that the observed enhanced angiogenic potency resulted from a superior bioavailability. The maintenance of newly formed vessels by the long-term administration of the growth factor remains an unsolved problem. Further improvement of the delivery devices, via in situ stability modulation, will be required.

Acknowledgements

We would like to thank Ms Esther Kleiner and Silvia Behnke (Sophistolab AG) for their help with histology. Support was provided by the European Community's Seventh Framework Programme contract (Angioscaff) and the Max & Hedwig Niedermayer Foundation.

References

1. J. Folkman and M. Hochberg, Self-regulation of growth in three dimensions, J. Exp. Med., 1973, 138(4), 745–753.
2. E. A. Phelps, N. Landázuri, P. M. Thulé, W. R. Taylor and A. J. García, Bioartificial matrices for therapeutic vascularization, Proc. Natl. Acad. Sci. U. S. A., 2010, 107(8), 3323–3328.
3. P. Carmeliet, VEGF gene therapy: stimulating angiogenesis or angioma-genesis?, Nat. Med., 2000, 1102–1103.
4. P. Carmeliet and R. K. Jain, Angiogenesis in cancer and other diseases, Nature, 2000, 407(6801), 249–257.
5. R. Blanco and H. Gerhardt, VEGF and Notch in tip and stalk cell selection, Cold Spring Harbor Perspect. Med., 2013, 3(1), a006569.
6. L. Jakobsson, K. Bentley and H. Gerhardt, VEGFRs and Notch: a dynamic collaboration in vascular patterning, Biochem. Soc. Trans., 2009, 37(Pt 6), 1233–1236.
7. H. Gerhardt, VEGF and endothelial guidance in angiogenic sprouting, Organogenesis, 2008, 4(4), 241–246.
8. S. B. Freedman and J. M. Isner, Therapeutic angiogenesis for coronary artery disease, Ann. Intern. Med., 2002, 136(1), 54–71.
9. S. B. Freedman and J. M. Isner, Therapeutic angiogenesis for ischemic cardiovascular disease, J. Mol. Cell. Cardiol., 2001, 33(3), 379–393.
10. T. D. Henry, Therapeutic angiogenesis, BMJ, 1999, 318(7197), 1536–1539.
11. J. S. Choi and H. S. Yoo, Chitosan/pluronic hydrogel containing bFGF/heparin for encapsulation of human dermal fibroblasts, J. Biomater. Sci., Polym. Ed., 2013, 24(2), 210–223.
12. M. L. Fabiilli, C. G. Wilson, F. Padilla, F. M. Martín-Saavedra, J. B. Fowlkes and R. T. Franceschi, Acoustic droplet-hydrogel composites for spatial and temporal control of growth factor delivery and scaffold stiffness, Acta Biomater., 2013, 9(7), 7399–7409.
13. D. Dyondi, T. J. Webster and R. Banerjee, A nanoparticulate injectable hydrogel as a tissue engineering scaffold for multiple growth factor delivery for bone regeneration, Int. J. Nanomed., 2013, 8, 47–59.
14. P. S. Lienemann, M. P. Lutolf and M. Ehrbar, Biomimetic hydrogels for controlled biomolecule delivery to augment bone regeneration, Adv. Drug Delivery Rev., 2012, 64(12), 1078–1089.
15. M. Ehrbar, V. G. Djonov, C. Schnell, S. A. Tschanz, G. Martiny-Baron and U. Schenk, et al., Cell-demanded liberation of VEGF121 from fibrin implants induces local and controlled blood vessel growth, Circ. Res., 2004, 94(8), 1124–1132.
16. K. A. Houck, D. W. Leung, A. M. Rowland, J. Winer and N. Ferrara, Dual regulation of vascular endothelial growth factor bioavailability by genetic and proteolytic mechanisms, J. Biol. Chem., 1992, 267(36), 26031–7.
17. J. E. Park, G. A. Keller and N. Ferrara, The vascular endothelial growth factor (VEGF) isoforms: differential deposition into the subepithelial extracellular matrix and bioactivity of extracellular matrix-bound VEGF, Mol. Biol. Cell, 1993, 4(12), 1317–1326.
18. N. Ferrara, Binding to the extracellular matrix and proteolytic processing: two key mechanisms regulating vascular endothelial growth factor action, Mol. Biol. Cell, 2010, 21(5), 687–690.
19. G. Degenfeld von, A. Banfi, M. L. Springer, R. A. Wagner, J. Jacobi and C. R. Ozawa, et al., Microenvironmental VEGF distribution is critical for stable and functional vessel growth in ischemia, FASEB J., 2006, 20(14), 2657–2659.
20. C. R. Ozawa, A. Banfi, N. L. Glazer, G. Thurston, M. L. Springer and P. E. Kraft, et al., Microenvironmental VEGF concentration, not total dose, determines a threshold between normal and aberrant angiogenesis, J. Clin. Invest., 2004, 113(4), 516–527.
21. C. Ruhrberg, H. Gerhardt, M. Golding, R. Watson, S. Ioannidou and H. Fujisawa, et al., Spatially restricted patterning cues provided by heparin-binding VEGF-A control blood vessel branching morphogenesis, Genes Dev., 2002, 16(20), 2684–2698.
22. I. Stalmans, Y. S. Ng, R. Rohan, M. Fruttiger, A. Bouche and A. Yuce, et al., Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms, J. Clin. Invest., 2002, 109(3), 327–336.
23. R. Hlushchuk, M. Ehrbar, P. Reichmuth, N. Heinimann, B. Styp-Rekowska and R. Escher, et al., Decrease in VEGF expression induces intussusceptive vascular pruning, Arterioscler., Thromb., Vasc. Biol., 2011, 31(12), 2836–2844.
24. L. E. Benjamin, I. Hemo and E. Keshet, A plasticity window for blood vessel remodelling is defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B and VEGF, Development, 1998, 125(9), 1591–1598.
25. A. Lazarus and E. Keshet, Vascular endothelial growth factor and vascular homeostasis, Proc. Am. Thorac. Soc., 2011, 8(6), 508–511.
26. N. Davies, S. Dobner, D. Bezuidenhout, C. Schmidt, M. Beck and A. H. Zisch, et al., The dosage dependence of VEGF stimulation on scaffold neovascularisation, Biomaterials, 2008, 29(26), 3531–3538.
27. D. Seliktar, A. H. Zisch, M. P. Lutolf, J. L. Wrana and J. A. Hubbell, MMP-2 sensitive, VEGF-bearing bioactive hydrogels for promotion of vascular healing, J. Biomed. Mater. Res. A, 2004, 68(4), 704–716.
28. A. H. Zisch, M. P. Lutolf, M. Ehrbar, G. P. Raeber, S. C. Rizzi and N. Davies, et al., Cell-demanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth, FASEB J., 2003, 17(15), 2260–2262.
29. J.-Y. Sun, X. Zhao, W. R. K. Illeperuma, O. Chaudhuri, K. H. Oh and D. J. Mooney, et al., Highly stretchable and tough hydrogels, Nature, 2012, 489(7414), 133–136.
30. K. Y. Lee, K. H. Bouhadir and D. J. Mooney, Controlled degradation of hydrogels using multi-functional cross-linking molecules, Biomaterials, 2004, 25(13), 2461–2466.
31. M. Ehrbar, R. Schoenmakers, E. H. Christen, M. Fussenegger and W. Weber, Drug-sensing hydrogels for the inducible release of biopharmaceuticals, Nat. Mater., 2008, 7(10), 800–804.
32. M. Ehrbar, S. M. Zeisberger, G. P. Raeber, J. A. Hubbell, C. Schnell and A. H. Zisch, The role of actively released fibrin-conjugated VEGF for VEGF receptor 2 gene activation and the enhancement of angiogenesis, Biomaterials, 2008, 29(11), 1720–1729.
33. A. H. Zisch, U. Schenk, J. C. Schense, S. E. Sakiyama-Elbert and J. A. Hubbell, Covalently conjugated VEGF–fibrin matrices for endothelialization, J. Controlled Release, 2001, 72(1–3), 101–113.
34. M. Ehrbar, A. Metters, P. Zammaretti, J. A. Hubbell and A. H. Zisch, Endothelial cell proliferation and progenitor maturation by fibrin-bound VEGF variants with differential susceptibilities to local cellular activity, J. Controlled Release, 2005, 101(1–3), 93–109.
35. M. Ehrbar, S. C. Rizzi, R. Hlushchuk, V. Djonov, A. H. Zisch and J. A. Hubbell, et al., Enzymatic formation of modular cell-instructive fibrin analogs for tissue engineering, Biomaterials, 2007, 28(26), 3856–3866.
36. C. Del Gaudio, S. Baiguera, M. Boieri, B. Mazzanti, D. Ribatti and A. Bianco, et al., Induction of angiogenesis using VEGF releasing genipin-crosslinked electrospun gelatin mats, Biomaterials, 2013, 34(31), 7754–7765.
37. J. Chen, L. Yang, L. Guo and X. Duan, Sodium hyaluronate as a drug-release system for VEGF 165 improves graft revascularization in anterior cruciate ligament reconstruction in a rabbit model, Exp. Ther. Med., 2012, 4(3), 430–434.
38. S. M. Jay and W. M. Saltzman, Controlled delivery of VEGF via modulation of alginate microparticle ionic crosslinking, J. Controlled Release, 2009, 134(1), 26–34.
39. S. Suga, Y. G. Kim, A. Joly, E. Puchacz, D. H. Kang and J. A. Jefferson, et al., Vascular endothelial growth factor (VEGF121) protects rats from renal infarction in thrombotic microangiopathy, Kidney Int., 2001, 60(4), 1297–1308.
40. Z. Li, Y. Zhang, J. Ying Ma, A. M. Kapoun, Q. Shao and I. Kerr, et al., Recombinant vascular endothelial growth factor 121 attenuates hypertension and improves kidney damage in a rat model of preeclampsia, Hypertension, 2007, 50(4), 686–692.
41. K. M. Lorentz, S. Kontos, P. Frey and J. A. Hubbell, Engineered aprotinin for improved stability of fibrin biomaterials, Biomaterials, 2011, 32(2), 430–438.

---

Footnotes

† Remo Largo, Venkat M. Ramakrishnan and Jeffrey S. Marschall contributed equally to this work.

‡ Daniel Eberli and Martin Ehrbar both serve as last authors.

---