Improving arterial prosthesis neo-endothelialization: Application of a proactive VEGF construct onto PTFE surfaces

ARTICLE IN PRESS

Biomaterials 26 (2005) 7402–7409 www.elsevier.com/locate/biomaterials

Improving arterial prosthesis neo-endothelialization: Application of a proactive VEGF construct onto PTFE surfaces M. Crombeza,b, P. Chevalliera,b, R.C. -Gaudreaulta,c, E. Petitclerca,c, D. Mantovania,b, G. Larochea,b, Unite´ de Biotechnologie et de Bioinge´nierie, Centre de Recherche de l’Hoˆpital Saint-Franc- ois d’Assise, C.H.U.Q., 10 rue de l’Espinay, Que´., Canada G1L 3L5 b De´partement de Ge´nie des Mines, de la Me´tallurgie et des Mate´riaux, Faculte´ des sciences et ge´nie, Universite´ Laval, Que´., Canada G1K 7P4 c De´partement de Me´decine, Faculte´ de Me´decine, Universite´ Laval, Que´., Canada G1K 7P4 a

Available online 11 July 2005

Abstract The formation of a confluent endothelium on expanded polytetrafluoroethylene (PTFE) vascular prostheses has never been observed. This lack of endothelialization is known to be one of the main reasons leading to the development of thromboses and/or intimal hyperplasia. In this context, several efforts were put forward to promote endothelial cell coverage on the internal surface of synthetic vascular prostheses. The goal of the present study was to immobilize the vascular endothelial growth factor (VEGF) onto Teflons PTFE surfaces to generate a proactive polymer construct favoring interaction with endothelial cells. An ammonia plasma treatment was first used to graft amino groups on PTFE films. Subsequent reactions were performed to covalently bind human serum albumin (HSA) on the polymer surface and to load this protein with negative charges, which allows adsorbtion of VEGF onto HSA via strong electrostatic interactions. X-ray photoelectron spectroscopy (XPS) experiments along with surface derivatization strategies were performed between each synthesis step to ascertain the occurrence of the various molecules surface immobilization. Finally, the electrostatic binding of VEGF to the negatively charged HSA matrix was performed and validated by ELISA. Endothelial cell adhesion and migration experiments were carried out to validate the potential of this VEGF-containing biological construct to act as a proactive media toward the development of endothelial cells. r 2005 Elsevier Ltd. All rights reserved. Keywords: Vascular endothelial growth factor; Surface modification; Endothelial cells; Plasma treatment

1. Introduction Arterial prostheses made of microporous Teflons (polytetrafluoroethylene (ePTFE)) are routinely used in vascular surgery as bypasses for small and medium blood vessels. However, the lack of prosthetic endothelialization leads to graft thrombosis and/or intimal hyperplasia which in turn result in the obligation to explant the prosthetic device [1]. Since 1976, graft Corresponding author. De´partement de Ge´nie des Mines, de la Me´tallurgie et des Mate´riaux, Faculte´ des sciences et ge´nie, Universite´ Laval, Que´., Canada G1K 7P4. Tel.: +1 418 656 2131x7983; fax: +1 418 656 5343. E-mail address: [email protected] (G. Laroche).

0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.05.051

seeding with endothelial cells (ECs) has been proposed to solve the afore-mentioned issues. However, the results were inconsistent since the adherence efficacy of the seeded ECs to the graft has been revealed to be unpredictable [2–4]. Even though these attempts have failed, it is well accepted that developing a successful arterial prosthesis requires its ability to promote the formation of an endothelium. Many growth factors have been identified, along with cytokines and other regulatory proteins that stimulate EC either directly or indirectly through various interactions during EC growth process. Among these, basic fibroblast growth factor (bFGF) and vascular endothelium growth factor (VEGF) have received by far the

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

most attention [5–8], primarily because of their high performance (bFGF), specificity and apparent central regulatory role in angiogenesis (VEGF) [9,10]. Consequently, considerable research focused on generating and testing these growth factors. Moreover, recent studies have shown that VEGF accelerates reendothelialization and attenuates intimal hyperplasia in the vascular context [11–17]. The complete signaling pathway of VEGF proceeds through the formation of a chemotactic gradient, directed towards cell receptors, and the stimulation of the secretion of proteases to promote tissue invasion [10]. The action of VEGF requires its internalization within the cell [18]. The development a self-endothelializable vascular graft should therefore take into account these two mitogenic properties of VEGF. In other words, the immobilization of VEGF onto the prosthetic material surface must proceed through non-covalent links strong enough to resist to the blood flow while allowing its eventual internalization by ECs. Such a strategy is in agreement with the observation that VEGF naturally bind to other biological molecules through electrostatic interactions [8,19]. In that context, the aim of the present study was to develop a VEGF-releasing surface to improve its ability to promote reendothelialization of the PTFE vascular prostheses. This stimulation would follow the adhesion, migration, and/or proliferation of activated neighboring ECs. Basically, our experimental approach consisted of immobilizing human serum albumin (HSA) onto the surface of ammonia plasma-treated PTFE previously conjugated with glutaric anhydride. The surface amino groups of HSA were then reacted with cisaconitic anhydride to obtain negative charges that would promote electrostatic interaction with VEGF. The ability of PTFE coated with VEGF biological construct to promote EC adhesion and migration was then evaluated.

2. Materials and methods 2.1. Materials PTFE films were obtained from Goodfellow (Berwyn, PA, USA). Sodium phosphate salts (Na2HPO4 and NaH2PO4), diethylamine (DEA), and 2-(N-morpholino) ethanesulfonic acid (MES) were all purchased from Sigma Aldrich (St Louis, MO, USA). Chemicals required for conjugating or negatively charging HSA (VWR-Canlab, West Chester, PA, USA) on the PTFE surface such as 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), glutaric anhydride, cisaconitic anhydride, trichloroacetic anhydride, and chlorobenzaldehyde were all purchased from Sigma Aldrich (St Louis, MO, USA). Recombinant human VEGF165, was obtained from R&D Systems Inc. (Minneapolis, MN, USA) while molecules needed for semi-quantification of this growth factor

7403

on the polymer surface (antibody goat anti-human VEGF, anti-VEGF, Rabbit anti-IgG-AP (2.1 mg/mL in DEA)), 4-methylumbelliferyl phosphate (MUP), and casein were all obtained from Sigma Aldrich Canada (St Louis, MO, USA). Finally Dulbecco’s Modified Eagle Medium (D-MEM) without glutamine was used to perform the cell migration experiments (Life Technologies, Rockville, MD, USA). 2.2. Method 2.2.1. Plasma treatment Film samples were cut into 3 cm
3 cm from a 250 mmthick commercial PTFE film. Each sample was successively washed in an ultrasonic bath containing high-purity acetone, deionized water, and high-purity methanol for 10 min, and then dried under vacuum. All plasma treatments were performed in a cylindrical plasma reactor described elsewhere under 300 mTorr of high-purity ammonia with 20 W of RF power at a frequency of 13.56 MHz for 100 s [20]. The whole sample was then cut into nine 1 cm2 pieces for the subsequent grafting procedures. Quantification of the surface concentration of the amino groups present on the PTFE samples after the ammonia-plasma treatment was performed through surface derivatization using chlorobenzaldehyde followed by XPS according to a previously published protocol [21]. 2.2.2. Glutaric anhydride grafting The ammonia plasma-treated PTFE films were immersed in 1 mL of a solution of 25 mg/mL of glutaric anhydride in PBS (0.2 M at pH 8.0) in a glove box under dry nitrogen atmosphere. The reaction was performed for 60 min at room temperature under mechanical agitation. After 20 and 40 min, the concentration of glutaric anhydride was raised to 40 and 55 mg/mL, respectively, to ensure completion of the reaction. Following the reaction, the films were thoroughly washed with water, then vacuum-dried overnight at 40 1C. Quantification of the extent of the reaction between glutaric anhydride and the PTFE aminated surfaces was calculated using XPS data as described by Gauvreau et al. [22]. 2.2.3. Activation of the surface carboxylic acid moieties with EDC Prior to the immobilization of HSA, glutaric anhydridegrafted films were activated with a solution containing 10 mg/ mL of EDC in MES buffer (0.1 M, pH 4.75) [23]. Two subsequent additions of 10 mg of EDC were made to this solution every 10 min during 20 min at room temperature and under stirring (for a total reaction time of 30 min) to minimize the effect of the water-induced hydrolysis of the activator on the grafting efficiency. These films were then washed twice with the MES buffer to remove any non-reacted EDC molecules. 2.2.4. Conjugation of HSA on EDC-grafted PTFE surfaces The glutaric anhydride-grafted films activated with EDC were reacted with 1 mL of PBS (0.2 M, pH 7.4) solution containing 1 mg/mL of HSA for 3 h under mechanical agitation [23]. Again, the resulting surfaces were thoroughly washed with water to remove non-covalently bound HSA.

ARTICLE IN PRESS 7404

M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

2.2.5. Loading the HSA-grafted surface with negative charges The free lysine residues on the surface-grafted HSA were reacted with cis-aconitic anhydride therefore allowing to charge negatively the protein surface. Briefly, a 1 cm2 of HSA-grafted film was immersed in 1 mL of PBS (0.2 M, pH 8.0). Three successive additions of cis-aconitic anhydride of 40, 30, and 30 mg were then made every 20 min to reach an anhydride final concentration of 100 mg/mL. The reaction mixture was mechanically agitated at room temperature for 1 h. Unfortunately, the accurate determination of the surface concentration of cis-aconitic anhydride at this stage is rather difficult because the XPS signal of the atoms of this molecule is undistinguishable from that of the underlying protein layer. Therefore, the successfulness of the cis-aconitic anhydride grafting was ascertained by reacting trichloroacetic anhydride with the HSA-grafted surface as both reactions of these anhydrides with the HSA-grafted surfaces proceeds through identical mechanisms. However, the XPS signal coming from the three chlorine atoms on trichloroacetic amido moieties unequivocally confirm the nucleophilic addition of the anhydride to the lysyl residues of HSA. 2.2.6. XPS analyses XPS spectra were recorded using a PHI 5600-ci spectrometer (Physical Electronics, Eden Prairie, MN, USA) using an X-ray source at a power of 400 W. A monochromatic aluminum X-ray source (1486.6 eV) was used to record the survey spectra while high-resolution spectra were recorded using a monochromatic magnesium X-ray source (1253.6 eV). The detection was performed at a take-off angle of 451 with respect to the surface. 2.2.7. VEGF seeding and detection Experiments were performed in triplicate for each of the VEGF-seeded concentrations (0, 0.01, 0.03, 0.1, 0.3, 1, and 3 mg/mL in PBS 0.2 M, pH 7.4) for negatively charged HSAtreated films. A 50 mL drop of each VEGF solution was added onto 4 mm –diameter films cut from the 1 cm2 samples and the immobilization reaction was allowed to proceed between 30 min and 1 h at room temperature. The VEGF solution was then removed from the films, which were exhaustively washed in PBS (0.2 M, pH 7.4). Each VEGF-seeded film as well as various controls were deposited into wells of a polyvinyl microtiter 48-well plate (Falcon cytoplate). Non-specific sites were blocked by incubating the films with 100 mL of 1% casein in PBS at 37 1C for 1 h. The solution was then removed and the plate was repeatedly washed by shaking out with 50 mL PBS per well. The primary antibody, anti-VEGF (100 mL/well), was added at the appropriate concentration (1/500) diluted in 1% of casein in PBS (0.2 M, pH 7.4). The plate was then incubated overnight at 4 1C. It was then shaken out and repeatedly washed with 200 mL/well of PBS. The secondary antibody, an alkaline phosphatase (AP)- conjugated goat anti-mouse, was added (100 mL/well) at the appropriate concentration (1/200) diluted in 1% of casein in PBS. The plate was incubated for 1 h at 37 1C, shaken out and repeatedly washed with 200 mL/well of PBS. 100 mL/well of the substrate solution (MUP at 104 M in DEA buffer at 0.03 M, pH ¼ 9.8) was then added. The amount of substrate converted by the enzyme produced a fluorescent product related to the amount of analyte bound during the

initial reaction. After approximately 30 min of reaction, the AP enzyme reached its highest activity level and fluorimetric measurements were then performed. The resulting fluorescence was determined using a FL600 Microplate Fluorescence Reader (Bio-Tek, Neufahrn, Germany) at excitation and emission wavelengths of 360 and 470 nm, respectively. The fluorescence used in diagrams was calculated by subtracting the fluorescence of one treated film seeded without VEGF obtained following the immunodetection process (background) from the fluorescence of the specimen under study. A standard curve was therefore plotted. The secondary antibody, anti-IgG-AP, was diluted from the stock solution by factors ranging between 7.5
108 and 102 in wells containing 180 mL of DEA buffer. After approximately 1 h at 37 1C, substrate solution was added (20 mL at 103 M). The fluorescence was then measured in each well after a 30 min reaction. The standard curve was obtained by plotting the fluorescence levels against the anti-IgG concentrations. To allow quantitative comparisons between the fluorescence levels of the various samples, fluorescence normalization was performed, with a subtraction of fluorescence obtained for substrate alone from fluorescence relative to each anti-IgG-AP dilution. The resulting fitted model, using only three parameters, enabled us to determine the relationship between the normalized fluorescence and the dilution factors of the secondary antibody, anti-IgG-AP, with a correlation higher than 99% for each standard experiment. To characterize the variations, a standard curve was fitted with the following parametrical Eq. (1): f ¼ a=ð1 þ expððx  x0 Þ=bÞÞ,

(1)

where a and b are the asymptotic values of the sigmoid curve when x tends towards positive and negative infinity, respectively, while x0 is the abscissa of the inflexion point [a/2; x0] with x0 ¼ f 1 ða=2Þ. To test the accuracy of the analytical method, Eq. (1) was solved for the fluorescence for three standard curves (results not shown). Parameters were similar from one standard curve to another. The whole strategy for building up the charged HSA/VEGF construct is illustrated in Fig. 1. 2.2.8. Human ECs adhesion and migration These experiments were performed as previously described except for the difference that the coated surfaces were the PTFE films instead of the bottom of wells [24]. Briefly, PTFE films were prepared with the different substrates (HSA and/or VEGF at a seeding concentration of 3 mg/mL). Cells were seeded at 5
104 cells/well and allowed to attach for 90 min in a cell incubator (37 1C, 5% CO2). At the end of the incubation, the non-adherent cells were washed with PBS and the attached cells were fixed using 1% paraformaldehyde in PBS for 10 min. After several washes, the cell nuclei were labeled using Hoechst 33258 (0.5 mg/mL) for 15 min. The different PTFE films were then visualized using a fluorescence microscope and the cells on 0.25 mm2 areas were numerated. For the migration assays, the lower part of Boyden chambers were coated using gelatin (1% in PBS) for 1 h at room temperature. PTFE films, with or without VEGF seeded on the HSA-coated surface (initial concentration of 3 mg/mL), were placed at the bottom of the wells directly under the

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

7405

HO O

Glutaric anhydride O

O

O

ammonia rf plasma

NH2

1) EDC, MES buffer pH 4.75

NH

O

PBS 0.2M, pH 8.0

2) HSA, PBS 0.2M, pH 7.4

+

+

VEGF

+

+

+ _ O

_

_

O

O

_

O

O

O

O O

NH2 HSA NH

O

NH

NH

HSA

HSA

NH Cic-aconitic anhydride

O

O

NH O

O

O

O

O O

O O

NH

O

C HO

PBS 0.2M, pH 8.0

NH VEGF seeding

NH

PBS 0.2M, pH 7.4

Fig. 1. Schematic representation of the strategy used to build up the charged HSA/VEGF construct, which is attached to a PTFE surface.

Boyden chambers. The medium used was serum-free D-MEM (900 mL/well). In control wells, soluble VEGF (10 mg/mL) was used as the chemoattractant. In the upper chamber, 105 cells were added in 300 mL, and the cells were allowed to migrate for 4 h at 37 1C, 5% CO2. The membranes were cleared of nonmigrating cells using a cotton swab, and the bottom cells were fixed and stained using 20% methanol, 0.5% crystal violet in PBS for 10 min. After several washes, the cells were counted in each well using an inverted microscope. The results are presented as mean7s.e. Multiple comparisons were performed with one-way ANOVA, followed by Dunnett’s test to compare the different surfaces with the control group in each experiment; p valueso0.05 are considered as being significant.

3. Results and discussion 3.1. Plasma treatment and glutaric anhydride grafting Fig. 2 displays the characterization of the first step necessary to build up the VEGF proactive surface construct, i.e., the conjugation of glutaric anhydride on an ammonia plasma-treated PTFE surface. As depicted in Fig. 2b, the ammonia plasma treatment allows the incorporation of 12% of nitrogen-containing species on the PTFE surface. Among them, 30% were shown to be part of primary amines as previously demonstrated by surface derivatization with chlorobenzaldehyde followed

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

7406

a F1s 66.9%

C1s 33.1%

b

F1s 35.9%

c O1s 4.3%

F1s 28.5%

291.8 26.9%

C1s 47.8%

287.2 17.2%

292

290

288

O1s 11.1% N1s 6.4%

1200

1000

800

600

285.0 27.3%

285.9 16.7%

288.8 12.3%

d

294

N1s 12.0%

400

286

284

282

C1s 54.0%

200

0

Binding energy (eV) Fig. 2. XPS spectra of (a) clean PTFE film, (b) ammonia plasmatreated PTFE film, and (c) ammonia -plasma-treated PTFE film conjugated with glutaric anhydride, (d) curve fitting of the C 1s feature of the high-resolution XPS spectrum of sample c.

In addition to detecting the occurrence of the reaction between the aminated PTFE surface and glutaric anhydride, XPS allows the quantification of the reaction yields through calculations similar to those performed in surface derivatization experiments [21,25]. Basically, the extent of the reaction depends on the amino groups surface concentration and is calculated by dividing the experimental relative surface concentration of an atom, as measured by XPS, by the theoretical surface concentration of this atom. As glutaric anhydride grafting leads to the addition of three oxygen atoms among the eight new atoms bound to the polymer surface, it turns out that the O1s XPS signal is very well suited for performing the reaction yield calculations [22]. As aforementioned, the reaction between glutaric anhydride and the aminated PTFE surface leads to an oxygen surface concentration of 11.1% that closely matched with the theoretical value that ranges between 10.7 and 11.6% for surfaces containing 3–3.5% of amino groups, respectively. This clearly demonstrates an almost complete reaction. In other words, this result indicates that approximately 3% of carboxylic acid moieties are available for further surface reaction with primary amino group-containing molecules. This result is further supported by the curve fitting of the C1s highresolution spectra of the glutaric anhydride-grafted PTFE surface. These data reveal that the feature assigned to both CO2H and CONH functionalities (E289 eV), which are formed in equivalent amount upon reaction of glutaric anhydride with the PTFE surface amino groups, accounted for 12.3% of the total carbon contribution on the surface (Fig. 2d). Combining this value with the carbon content of 54% measured from the XPS survey spectrum lead to the conclusion that 6.6% of the carbon atoms are located in COOH and CONH moieties, therefore allowing to conclude to a 3.3% initial amino groups surface concentration on the ammonia plasma-treated PTFE. 3.2. Conjugation and characterization of the negatively charged matrix on the polymer surface

by XPS [21]. Therefore, about 3–3.5% of amino groups are available for nucleophilic reactions. Such reactions were performed using glutaric anhydride. Fig. 2c shows that the presence of this molecule can be easily detected with XPS through an increase of oxygen relative surface concentration that goes from 4.3% for the ammonia plasma-treated surface to 11.1% after the nucleophilic addition of glutaric anhydride. This surface reaction allows the formation of carboxylic acid groups on the polymer surface that, once activated with EDC, are competent for reacting with primary amino groups. This strategy is particularly useful in the context of the present study for conjugation of albumin through its lysyl residues.

Fig. 3a and b shows the XPS spectra of the carboxylic acid-containing surfaces before and after immobilization of HSA, respectively. Unfortunately, these XPS results do not allow the quantification of the surface concentration of HSA. Qualitatively, the important decrease in the fluorine surface content along with the increase in both the oxygen and nitrogen surface concentrations clearly indicate the successfulness of the protein conjugation. The addition of negative charges through the subsequent grafting of cis-aconitic anhydride allows the incorporation of two negative charges for every anhydride addition to an amino group of HSA (Fig. 1). However, the detection of this anhydride grafting on an HSA-grafted PTFE surface was

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

that the multi-step strategy that was put forward to construct a negatively charged matrix for VEGF electrostatic binding has been successful. In other words, amino groups are still available for reacting with the anhydride functionalities despite prior reaction of some of the HSA lysyl residues with the surface activated esters.

F1s 28.5%

O1s 11.1%

C1s N1s 54.0% 6.4%

b F1s 19.2% O1s 13.0%

3.3. VEGF loading and immunodetection

C1s 59.6% N1s 8.2%

c

F1s 13.6%

O1s 16.9%

C1s 63.2%

N1s 6.3%

d

F1s 15.9%

O1s 15.1% N1s 7.4%

C1s 58.3% Cl2s Cl2p 3.3%

1200

1000

800

600

400

200

0

Binding Energy (eV)

Fig. 3. XPS spectra of ammonia plasma-treated PTFE film conjugated with (a) glutaric anhydride, (b) glutaric anhydride+human serum albumin, (c) glutaric anhydride+human serum albumin+cis-aconitic anhydride, and (d) glutaric anhydride+human serum albumin+trichloroacetic anhydride.

somewhat difficult to ascertain due to the complexity of the atomic composition of the surface (Fig. 3c). Indeed, only subtle surface composition changes can be observed as the XPS signal due to the cis-aconitic anhydride conjugation is somewhat marginal as compared to that of the huge albumin protein already present on the surface (11 additional atoms compared to a 66 kDa protein). For this reason, the possibility of adding negative charges on immobilized HSA has been ascertained through the grafting of trichloroacetic anhydride. The conjugation of this later molecule on albumin proceeds exactly through the same mechanism as with cis-aconitic anhydride. However, the three chlorine atoms present in the structure of trichloroacetic anhydride allow to easily confirm the occurrence of the grafting of this molecule on the surface-immobilized HSA. Fig. 3d illustrates the XPS spectrum recorded after conjugating trichloroacetic anhydride on surfaceimmobilized HSA and displays a chlorine surface concentration of 3.3%. This result clearly evidences

Fig. 4 shows the results of VEGF seeding on the PTFE surface layered with the negatively charged HSA matrix. These data clearly indicate the possibility of modulating the VEGF surface concentration by selecting the appropriate seeding concentration. This result is of paramount importance in the context where the interaction between VEGF and ECs is known to be dose dependent [26–28]. Therefore, overloading the biomaterial surface with VEGF may lead to a cellular downregulation in response to an excess of the growth factor. The potential of the VEGF construct to promote EC adhesion has been monitored to validate its biological activity. As epitomized in Fig. 5, virgin PTFE does not promote the ECs adhesion due to the very low surface energy of this polymer. Conversely, the ammonia plasma-treated PTFE is by far the most efficient surface for supporting cell adhesion, in good agreement with data previously published by the Sipehia’s group [29–31]. Despite this behavior, it has been shown in many instances that the adhesion of ECs on ammonia plasma-treated PTFE is a non-specific phenomenon, therefore is non-viable in real in vivo situations. Binding of charged HSA on the polymer surface leads to a somewhat very similar behavior with only few adhering cells. Finally, coating the PTFE surface with the whole VEGF construct (e.g. charged HSA+VEGF) leads to -4.5

Log (Anti-IgG dilution factor)

a

7407

-5.0 -5.5 -6.0 -6.5 -7.0 -7.5 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Seeded VEGF concentration (µg/mL)

Fig. 4. Semi-quantitative measurements showing that the amount of VEGF in the charged HSA/VEGF construct can be controlled through the appropriate selection of amount of seeded VEGF.

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

7408 60

16 14

50

Number of cells

Number of cells

12 40

30

20

10 8 6 4

10

2 0

FE PT

VE G F

in

so

lu tio n

ct V ch EG ar F/n ge e d ga H tiv SA e l co y ns tru

ct

V ch EG ar F/ ge ne d ga H ti SA ve co ly ns tru

N ch eg ar ati ge ve d ly H SA

A p mm PT las on FE ma ia -tr ea te d

PT FE

0

Fig. 5. HUVEC adhesion on the PTFE surface covered with the charged HSA/VEGF construct. This result is compared with the HUVEC adhesion data measured for other PTFE-modified surfaces prepared during the various steps of the preparation of the charged HSA/VEGF construct.

Fig. 6. HUVEC migration induced by the PTFE surface covered with the charged HSA/VEGF construct (VEGF seeded from a 3 mg/mL solution). The positive and negative controls consisted in a VEGF solution (3 mg/mL) and a clean PTFE surface, respectively.

almost no cell adhesion. This result is in agreement with the biological function of VEGF, and confirms that the VEGF cell signaling pathway does not include cellular adhesion. In the present context, where a surface coverage with ECs is required, VEGF requires a prior adhesion signal to exert its effects. The availability of VEGF for inducing a migration signal in the negatively charged HSA/VEGF construct was verified using a Boyden chamber in which HUVEC were separated from their potential cell migration signal by a porous media. In such an experimental setup, positive migration cell signals promote the cells to cross the membrane towards the signaling molecule. As demonstrated in Fig. 6, virgin PTFE, used as negative control, promoted the migration of only a slight amount of cells from one side to the other side of the membrane. Consequently, the very low number of cells measured in this situation could be seen as the statistical probability of cells to migrate through the membrane under no influence. In contrast, a positive control consisting of VEGF in solution led to a concentration gradient through the porous media therefore enabling the ECs to cross it. In the presence of the charged HSA/VEGF construct immobilized on the PTFE surface, ECs clearly demonstrate a tendency to migrate through the porous membrane with a behavior similar to that of the positive control thus indicating that VEGF is released at a rate sufficiently high to promote migration. Taken together, the results on cell adhesion and migration therefore confirmed the well-known role of VEGF to promote cell migration

[10,32]. However, they also demonstrate that a surfaceengineered construct containing only VEGF is not sufficient to allow the self-endothelialization of an artificial blood vessel in vivo as it does not provide the adhesion signal that is necessary at the early stage of the healing process. In this context, the proactive construct developed therein would have to be complemented with cell signaling molecules, either co-immobilized or patterned, specifically targeting the adhesion of ECs.

4. Conclusion The objective of the present study was to verify the ability of a surface-engineered proactive construct to stimulate the in vitro endothelialization. Our data clearly demonstrated the feasibility of using a negatively charged HSA matrix for the subsequent electrostatic binding of VEGF, therefore mimicking the natural interaction of this growth factor with molecules occurring in the physiological environment. On one hand, this construct was clearly shown to have no effect on endothelial cell adhesion. On the other hand, it definitely affected the cell migration behavior, in accordance to the VEGF biological function that is to act as a migration signal. It is therefore concluded that the migrative signal of the charged HSA/VEGF construct should require a prior adhesion cell triggering induced by a molecular organization, made either of peptides, proteins, or more complex assemblies.

ARTICLE IN PRESS M. Crombez et al. / Biomaterials 26 (2005) 7402–7409

Acknowledgements The authors are grateful to Christine Ricard for performing the cell migration experiments. This study was supported by the National Science and Engineering Research Council (NSERC) of Canada and the Fonds pour la Recherche en Sante´ du Que´bec (FRSQ) (G.L.). M. Crombez acknowledges a postgraduate scholarship from the Centre de Recherche de l’Hoˆpital St-Franc- ois d’Assise, Centre Hospitalier Universitaire de Que´bec. References [1] Guidoin R, Chakfe´ N, Maurel S, How T, Batt M, Marois Y, Gosselin C. Expanded polytetrafluoroethylene arterial prostheses in humans: histopathological study of 298 surgically excised grafts. Biomaterials 1993;14:678–93. [2] Criqui MH, Browner D, Fronek A, Klauber MR, Coughlin SS, Barrett-Connor E, Gabriel S. Peripheral arterial disease in large vessels is epidemiologically distinct from small vessel disease. An analysis of risk factors. Am J Epidemiol 1989;129:1110–9. [3] Lundell A, Lindblad B, Bergqvist D, Hansen F. Femoropoplitealcrural graft patency is improved by an intensive surveillance program: a prospective randomized study. J Vasc Surg 1995;21: 26–33. [4] Herring M, Smith J, Dalsing M, Glover J, Compton R, Etchberger K, Zollinger T. Endothelial seeding of polytetrafluoroethylene femoral popliteal bypasses: the failure of low-density seeding to improve patency. J Vasc Surg 1994;20:650–5. [5] Richardson TP, Peters MC, Ennett AB, Mooney DJ. Polymeric system for dual growth factor delivery. Nat Biotechnol 2001;19:1029–34. [6] Ferrara N, Davis-Smyth T. The biology of vascular endothelial growth factor. Endocr Rev 1997;18:4–25. [7] Ferrara N. Molecular and biological properties of vascular endothelial growth factor. J Mol Med 1999;77:527–43. [8] Fairbrother WJ, Champe MA, Christinger HW, Keyt BA, Starovasnik MA. Solution structure of the heparin-binding domain of vascular endothelial growth factor. Structure 1998; 6:637–48. [9] Ferrara N. Role of vascular endothelial growth factor in regulation of physiological angiogenesis. Am J Physiol Cell Physiol 2001;280:1358–66. [10] Ruhrberg C. Growing and shaping the vascular tree: multiple roles for VEGF. Bioessays 2003;25:1052–60. [11] Zisch AH, Schenk U, Schense JC, Sakiyama-Elbert SE, Hubbell JA. Covalently conjugated VEGF-fibrin matrices for endothelialization. J Control Release 2001;72:101–13. [12] Stone D, Phaneuf M, Rohan D, Sivamurthy N, Quist WC, LoGerfo FW. In vitro study of vascular endothelial growth factor for bioengineering improved prosthetic vascular grafts. Curr Surg 2000;57:506–7. [13] Stone D, Phaneuf M, Sivamurthy N, LoGerfo FW, Quist WC. A biologically active VEGF construct in vitro: implications for bioengineering-improved prosthetic vascular grafts. J Biomed Mater Res 2002;59:160–5. [14] Greisler HP. Growth factor release from vascular grafts. J Control Release 1996;39:267–80. [15] Dorafshar AH, Angle N, Bryer-Ash M, Huang D, Farooq MM, Gelabert HA, Freischlag JA. Vascular endothelial growth factor inhibits mitogen-induced vascular smooth muscle cell proliferation. J Surg Res 2003;114:179–86. [16] Wissink MJ, Beernink R, Poot AA, Engbers GH, Beugeling T, van Aken WG, Feijen J. Improved endothelialization of vascular

[17]

[18]

[19]

[20]

[21]

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

7409

grafts by local release of growth factor from heparinized collagen matrices. J Control Release 2000;64:103–14. Asahara T, Bauters C, Pastore C, Kearney M, Rossow S, Bunting S, Ferrara N, Symes JF, Isner JM. Local delivery of vascular endothelial growth factor accelerates reendothelialization and attenuates intimal hyperplasia in balloon-injured rat carotid artery. Circulation 1995;91:2793–801. Li W, Keller G. VEGF nuclear accumulation correlates with phenotypical changes in endothelial cells. J Cell Sci 2000;113: 1525–34. Stauffer ME, Skelton NJ, Fairbrother WJ. Refinement of the solution structure of the heparin-binding domain of vascular endothelial growth factor using residual dipolar couplings. J Biomol NMR 2002;23:57–61. Mantovani D, Castonguay M, Pageau JF, Fiset M, Laroche G. Ammonia RF-plasma treatment of tubular ePTFE vascular prostheses. Plasmas Polym 1999;4:207–28. Chevallier P, Castonguay M, Turgeon S, Dubrulle N, Mantovani D, Mc Breen PH, Wittmann J-C, Laroche G. Ammonia RFplasma on PTFE surfaces: chemical characterization of the species created on the surface by vapor-phase chemical derivatization. J Phys Chem B 2001;105:12490–7. Gauvreau V, Chevallier P, Vallie`res K, Peticlerc E, -Gaudreault RC, Laroche G. Engineering surfaces for bioconjugation: developing strategies and quantifying the extent of the reactions. Bioconjugate Chem 2004;15:1146–56. Hermanson GT. Bioconjugates techniques. New York: Academic Press; 1996. Petitclerc E, Boutaud A, Prestayko A, Xu J, Sado Y, Ninomiya Y, Sarras MP, Hudson Jr. BG, Brooks PC. New functions for non-collagenous domains of human collagen type IV. Novel integrin ligands inhibiting angiogenesis and tumor growth in vivo. J Biol Chem 2000;275:8051–61. Sutherland I, Sheng E, Brewis D, Heath R. Studies of vapourphase chemical derivatisation for XPS analysis using model polymers. J Mater Chem 1994;4:683–7. Chen JH, Wang XC, Kan M, Sato JD. Effect of FGF-1 and FGF2 on VEGF binding to human umbilical vein endothelial cells. Cell Biol Int 2001;25:257–60. Bikfalvi A, Sauzeau C, Moukadiri H, Maclouf J, Busso N, Bryckaert M, Plouet J, Tobelem G. Interaction of vasculotropin/ vascular endothelial cell growth factor with human umbilical vein endothelial cells: binding, internalization, degradation, and biological effects. J Cell Physiol 1991;149:50–9. Pepper MS, Ferrara N, Orci L, Montesano R. Vascular endothelial growth factor (VEGF) induces plasminogen activators and plasminogen activator inhibitor-1 in microvascular endothelial cells. Biochem Biophys Res Commun 1991;181: 902–6. Sipehia R. The enhanced attachment and growth of endothelial cells on anhydrous ammonia gaseous plasma modified surfaces of polystyrene and poly(tetrafluoroethylene). Biomater Artif Cells Artif Organs 1990;18:437–46. Sipehia R. Enhanced attachment and growth of human endothelial cells derived from umbilical veins on ammonia plasma modified surfaces on PTFE and ePTFE synthetic vascular graft biomaterials. Biomater Artif Cells Immobilization Biotechnol 1993;21:455–68. Lu A, Sipehia R. Antithrombotic and fibrinolytic system of human endothelial cells seeded on PTFE: the effects of surface modification of PTFE by ammonia plasma treatment and ECM protein coatings. Biomaterials 2001;22:1439–46. Yoshida A, Anand-Apte B, Zetter BR. Differential endothelial migration and proliferation to basic fibroblast growth factor and vascular endothelial growth factor. Growth Factors 1996;13: 57–64.