Polyurethane modification with acrylic acid by Ce(IV)-initiated graft polymerization

Abstract This paper presents a method for polyurethane surface functionalization for tissue engineering applications. Functionalization has been carried out by grafting acrylic acid to the polyurethane surface with the use of radical polymerization with a Ce4+ initiator. Contrary to other papers suggesting that the presence of hydroxyl groups are essential for successful grafting via ceric ions, we propose a method with the omission of the surface hydroxylation step. The influence of reaction conditions: reaction time, reaction temperature and monomer concentration on carboxyl groups surface density has been analyzed and described. The quantity of carboxyl groups on the surface was determined with the use of the TBO method. Materials grafted with acrylic acid have been subjected to conjugation with a peptide using sulfoNHS/ EDC chemistry. Successful incorporation of the peptide has been confirmed by an ELISA assay. Additionally, for better characterization, after each step of modification materials were subjected to SEM, FTIR-ATR, XPS and contact angle measurement analysis.


Introduction
Polyurethanes (PUs) have been commonly used to fabricate blood-contacting implantable devices since the 1960s [1] because of good mechanical and physiochemical properties and acceptable hemocompatibility [2]. However, it has been known that like any other artificial materials, a patient's body rejects cardiovascular devices made of PUs. Small ions and protein adsorption takes place immediately after implantation, then blood cell adhesion causes chronic inflammation and fibrous tissue formation [3,4]. Another issue is long-term thrombosis that is likely to occur as an effect of constant blood-biomaterial interaction [5]. Thus, there is a need for biomaterial modification to improve its features and enhance its interactions with tissues and blood.
One of the approaches assumes superficial modification of commercially available PUs. In this way, the modification process would not affect the bulk of the polymer and thus would not influence mechanical properties of the material. We have reported in previous works the increase of hydrophobicity of the surface by incorporation of hydrophobic agents [6] or hydrophilization of the surface by grafting poly(vinylpirrolidone) [7]. Nevertheless, these methods concern covering the surface of PU with a layer of another biocompatible polymer which leaves mechanical properties of a final product in doubt. Taking this into consideration, we have gained interest in grafting functionalized spacer molecules onto the PU surface. We focused on acrylic acid (AA) grafting since it is soluble in water, which enables water-based processes, free from toxic organic solvents. Carboxyl groups provided by AA could be subsequently utilized to bond biomolecules such as collagen [8] and adhesive peptides, e.g. RGD [9], that induce anchorage of endothelial or smooth muscle cells. In this way, it is expected that modified PU would act as a natural tissue.
In this paper, we present a method of acrylic acid (AA) grafting to a polyurethane (PU) surface via radical polymerization induced by ceric ion reduction. The method has been reported for chitosan [10], PET [11] or cellulose [12] surface modification with other compounds. There were also some trials with PU grafting triggered by Ce4+ reduction [14,15]. Significantly, many works claim that in order to successfully carry out this approach, superficial -OH groups are required [15,16] due to the proposed mechanism of preferable ceric ion reaction with primary -OH groups. The reaction is often described as -CH 2 OH+ Ce 4+ = -C*HOH + H+ + Ce 3+ , where the asterisk stands for free carbon radicals [15]. In contrast with this statement, we present a method of successful AA introduction onto the PU surface with the omission of the surface hydroxylation step. The AA-grafted PU can be subsequently conjugated with adhesive peptides, e.g. REDV, that induce anchorage of endothelial cells. It has to be mentioned that PAA is often used in biomaterials synthesis. Biocompatibility of AA is well-known [16]. A possibility of its polymerization in situ predestines AA to be used in obtaining polyelectrolytes for biomedical applications [17]. Furthermore, polymerized AA -poly(acrylic acid) (PAA)not only provides COOH functional groups, but according to a recent study, also exhibits antimicrobial properties. It was demonstrated that PAA containing diblock copolymers prevents biofilm formation [18]. This is certainly a desired feature for a biomaterial.

PU films preparation and grafting
PU pellets were purified in aqueous ethanol, dried at 37 °C and dissolved in DMAC (20% w/v). After complete dissolution of the granules the solution was poured onto glass surfaces and dried at 40 °C to completely evaporate the solvent. The sample discs were cut off from the film and modified with AA according to the following procedure: PU discs were immersed in a 0.4 M solution of HNO 3 in distilled water and heated to 25,35  2 H 2 O (0.1% w/v) and AA (1%, 3% or 5% v/v) were added to the solution. The mixture was left covered, stirring continuously for 0.5, 1, 1.5 or 2 h. The quantity of carboxyl groups was determined colorimetrically (see below). For further experiments, PU-COOH materials showing the largest amount of COOH groups were chosen. PU-COOH discs were washed with a 0.1% (w/v) SDS aqueous solution, rinsed 3 times with distilled water and placed in 0.05 M MES buffer (pH=6.0) for 1 h. Materials were transferred to a sulfo-NHS (5mM) and EDC (2 mM) solution in MES (pH=6.0) for 15 min, washed with MES buffer and placed in a GSGREDVGSG solution (0.1mM in PBS, pH=10.5) for 1 h, RT. Afterwards, discs were rinsed with PBST, followed by rinsing with PBS. As a control material samples physically coated with peptide (without activation with EDC and sulfo-NHS) were prepared.

Carboxyl groups determination
Effect of different process parameters (temperature, time, monomer concentration) on AA grafting yield was tested. Surface density of COOH groups on PU-COOH was measured using colorimetric reaction with toluidine blue (TBO) as described elsewhere [19]. The discs were put in TBO solution, pH=11, for 3 h, RT. Next, discs were dipped in distilled water and placed in 50% acetic acid solution, pH=1.5. Absorbance of the obtained solutions at 633 nm was measured using a plate reader. Each measurement was performed 10 times for each PU-COOH material. PU-COOH discs with the biggest amount of superficial COOH groups (PU-COOH max ) were submitted to REDV grafting.

Peptide presence confirmation
To confirm successful introduction of the peptide on PU-COOH max an ELISA assay was performed. As a control material samples physically coated with peptide (without activation by EDC and sulfo-NHS) were used. The samples were placed in 24-well polystyrene plates, equilibrated with PBS and blocked with PBST (1 h, 37 °C). After blocking, samples were incubated with primary antibodies (1 h, 37 °C, rabbit monoclonal anti-REDV antibodies, dilution 1:1000) and secondary antibodies (1 hour, 37 °C, antirabbit IgG conjugated to peroxidase, dilution 1:20000) and transferred to fresh plates to eliminate the influence of the antibodies adsorbed to the well walls. After each assay step samples were rinsed with PBST (6 x 5 minutes on a shaker). In the last assay step samples were incubated with a peroxidase substrate solution -SigmaFast OPD (o-phenylenediamine dihydrochloride) in the dark at RT for 30 minutes. After the reaction, an aliquot of the solution (200 μl) from each well was transferred to a 96well plate; the optical density of the solution was read at the 450 nm. Each measurement was performed 8 times for each type of material.

SEM imaging
Scanning electron microscope (SEM) imaging was performed for PU, PU-COOH max and PU-REDV in order to visualize differences in surface topography after each modification step. SEM images were obtained with the use of a Zeiss Ultra Plus microscope equipped with a secondary electron detector SE2. Materials were fixed by dipping in a 2% (v/v) glutaraldehyde aqueous solution for 24 h rinsing with distilled water and dehydrating in ethanol solutions (ethanol concentrations form 50% to 100%). After fixation, materials were sputter-coated with carbon. The imaging was performed in at least 5 randomly selected spots on the analyzed material. Each material was prepared in triplicate. Representative images for each material were selected.

Wettability measurement
Surface wettability was measured for PU, PU-COOH max and PU-REDV. Materials were glued to a glass slide and a drop of distilled water (5 µl) was placed on a clean and dry surface. The contact angle was measured automatically using Kruss DCA 100 software. The measurement was performed in at least 5 randomly selected spots on the analyzed material and each material was prepared in triplicate.

FTIR-ATR analysis
FTIR spectra were recorded with a Nicolet™ 6700 spectrometer (Thermo Scientific). All samples were detected in attenuated total reflection (ATR) mode. Spectra were analyzed with the use of OMNIC 8.3 software. Spectra were collected in at least 5 randomly selected spots and each material was prepared in triplicate. One representative spectrum for each material was selected.

XPS analysis
The changes of chemical composition during surface modification were determined by XPS on a VG Scientific ESCALAB photo-electron spectrometer. Spectra fitting and determination of atomic composition was performed with software provided by VG Scientific.

Modification mechanism
The aim of this research was to develop a simple and effective method for the chemical attachment of AA to PU. The method may be utilized to improve the wetting properties of PU, which is essential as we discuss biomedical purposes. Also, AA provides carboxyl groups that can be used to couple biomolecules such as peptides. AA was chosen as a spacer molecule that provides connection and proper distance between the surface and bioactive molecule (peptide). PU belongs to a group of polymers with no surface functional groups that could be useful for chemical grafting. In order to attach biomolecules, the process of surface activation (which enriches surface in functional groups) must be carried out. It should be also kept in mind, that peptides or proteins immobilized on a polymer should posses appropriate conformation (orientation in space). This ensures that the biological activity and recognition by specific cell receptors of the peptide and/or protein is maintained. Therefore, some free space around them should be provided. In addition, when a peptide molecule and the surface of the polymer are too close to each other, unwanted electrostatic repulsion can occur. Consequently, biomolecules should be placed at the end of spacer molecules bonded to the surface of the polymer.
In many cases, photooxidation is a method of choice for PU surface activation [20][21][22]. The surface hydroxyl groups seemed to be essential especially for ceriuminduced grafting [15]. However, our initial experiments demonstrated that the PU surface treated with UV was physically changed (in terms of topology and roughness) which negatively aff ected surface adsorption properties. In order to minimize the invasiveness of the modifi cation process, we have decided to abandon the oxidation step. The mechanism of graft ing has been schematically shown in Figure 1, details have been described elsewhere [23]. Briefl y, cerium initiator abstracts hydrogen from urethane groups and generates nitrogen radicals, which react with AA. In order to provide an acidic environment, which is necessary in this reaction, nitric acid has been added to the solution. As a polymerization initiator ammonium cerium(IV) sulphate has been used, which, apart from ammonium cerium nitrate, is one of the most common initiators utilized in cerium-induced graft ing.
In the next modifi cation step, AA-graft ed materials with the largest amount of COOH (PU-COOH max ) were reacted with the peptide solution. Aft er AA graft ing, materials were subjected to reaction with sulfo-NHS/EDC (PU-NHS). First, activation of carboxyl groups with sulfo-NHS/EDC was conducted. Then, materials were immersed in the peptide solution, where activated carboxyl groups reacted with amine groups present in peptide chain (PU-NHS-REDV). In order to optimize the AA graft ing yield the following parameters were tested: reaction temperature, reaction time and AA concentration. The materials preparation scheme is presented in Figure 2.

Effect of temperature on AA grafting yield
As presented in Figure 3a) the concentration of surface carboxyl groups increases with the increase of temperature. That trend can be observed for lower concentrations monomer (1% and 3%). The increase in graft ing effi ciency may be due to: (i) increased decomposition of Ce 4+ -PU complex and (ii) increased diff usion of monomer to the active sites. For the highest AA concentration analyzed (5%) the density of COOH increases with temperature and reaches a maximum at 35 °C. Further increase in the temperature resulted in a rapid decrease in graft ing effi ciency. This decrease may be due to: (i) increased monomer hydrolysis, (ii) acceleration of chain termination processes and (iii) acceleration of chain transfer processes leading to increased homopolymerization [24,25].

Effect of time on AA grafting yield
AA graft ing yield strongly depended on the reaction time ( Figure 3b). It has been observed that with an increase in reaction time from 30 to 60 minutes the graft yield increases. The further course of the curve was diff erent depending on monomer concentration. For small AA concentration (1%) the increase does not signifi cantly infl uence the graft ing density. In the case of 3% AA, graft ing yield increases with increasing time, reaching a maximum graft ing yield aft er 90 minutes. A further increase in time did not alter the graft ing density. The course of the curve was diff erent in the case of the higher monomer concentration analyzed (5%). Density of surface carboxyl groups reached a maximum in 60 minutes. A further increase in time leads to a decrease in the graft ing yield. This trend has been observed in similar studies [13,25,26]. As other authors suggest, this eff ect can be due to termination or destruction of growing chains by active radicals, which decreases graft ing yield [24,27].

Effect of the monomer concentration on AA grafting yield
The eff ect of monomer concentration on graft ing density is presented in Figure 3c). The graft ing yield increases together with the increasing AA concentration from 0.14 M to 0.70 M. This trend can be simply explained: the more AA molecules in solution, the bigger chance for growing polyacrylic chains to encounter more monomer units.

Selection of PU-COOH material for further experiments
For REDV coupling and characterization PU-COOH with the largest volume of COOH groups (PU-COOH max ) was selected. PU-COOH max was obtained for the following parameters of AA graft ing: 35 °C, 1.5 h and 5% (v/v) AA. For this set of parameters, the procedure seemed to be the most effi cient and repeatable.

Peptide presence confirmation
The aim of the presented surface modifi cations was the introduction of peptide sequences onto the PU surface. For peptide coupling, PU with the maximum volume of COOH groups (PU-COOH max ) was used. In order to confi rm peptide presence, an ELISA assay was carried out. The results are shown in Figure 3d). As expected, materials aft er AA graft ing (PU-NHS) gave negligible responses. Similar results were observed for materials graft ed with AA and immersed in the peptide solution without previous activation of carboxyl groups (PU-REDV). This result proves that creating a stable peptide coating requires a chemical reaction between the amine and carboxyl groups. Peptides that were physically adsorbed on the surface, without a chemical bond, were washed away during washing step. On the other hand, materials marked as "PU-NHS-REDV" gave a higher response. In this case, carboxyl groups present on the materials surface were activated with sulfoNHS. This step allows the chemical reaction between COOH and NH 2 groups to occur. The positive result of the ELISA assay confi rmed that the proposed surface modifi cation method can be applied to chemically conjugate peptide molecules to spacer-graft ed PU. What is more, the presence of surface hydroxyl groups is not essential for the reaction.

SEM imaging
Scanning electron microscope (SEM) imaging was performed for PU, PU-COOH max and PU-REDV in order to visualize changes in topography aft er each modifi cation step. Figure 4a-c shows SEM images for these surfaces. It can be noticed that the PU-COOH max surface presented more irregularities than the PU surface. There were some particles on the PU-COOH max surface (Fig. 4b) seen as sharp and bright points which were caused by monomer (AA) polymerization [14]. Figure 4c presents PU-REDV surfaces on which peptide aggregates are clearly seen. There are similar SEM images of PU graft ed with peptides in other works [29], which suggests that such clusters might represent peptide molecules. We assume that peptide molecules bonded to PU in the course of chemical graft ing are not seen on the SEM images. Those visible aggregates might be formed as a result of crosslinking between additional peptides adhered to the PU surface and glutaraldehyde used for materials fi xation. This is probably the result of insuffi cient rinsing of materials aft er the peptide-coupling step.

Wettability measurement
The contact angle (CA) values of the surface of various types of materials are presented in Figure 5. AA graft ing resulted in a decrease in CA value: from average value CA = 101. 4±2.8° (PU) to CA = 66. 4±5.0° (PU-COOH max ). This result was expected since polymerized AA (PAA) graft ed to the PU surface is completely dissociated at neutral pH. Aft er peptide coupling (PU-REDV) the hydrophilicity of the surface was further increased (CA = 52.6 ±2.6°). The increase in hydrophilicity is due to the coupling of the peptide chain, which consists of highly hydrophilic amino acids: arginine, aspartic acid and glutamic acid. All three of them have polar side chains that are dissociated at neutral pH. Such an increase in hydrophilicity aft er REDV incorporation on the modifi ed PU surface was reported in our previous work [28]. Figure 6 shows the comparison of FTIR-ATR spectra obtained for PU, PU-COOH max and PU-REDV surfaces. A characteristic stretching vibration band at 3700-3400cm -1 that relates to NH 2 , -OH and -C=O groups is shown [18]. The signal was more intense for PU-REDV than for other materials, which indicates the presence of peptide containing amine groups. On the other hand, there was barely any diff erence in this range for PU and PU-COOH max . Nevertheless, it should be noted that according to the manufacturer of the Nicolet 6700 spectrometer, the depth of laser beam penetration for ATR mode is approx. 2 µm. This means that the band at 3700-3400 cm -1 came from the PU surface more than from graft ed AA. This is why bands assigned to C=O and -OH groups present in the PU hid bands assigned to C=O and -OH in COOH groups from AA. Nevertheless, the presence of COOH on the PU-COOH max surface was confi rmed by other methods used in this work.

XPS analysis
The atomic composition of PU, PU-COOH max and PU-REDV was determined with the use of XPS (Table 1). As expected, aft er AA graft ing the percentage of oxygen atoms increased form 23.7% in PU to 30.2% in PU-COOH max due to the introduction of COOH groups on the surface. It proves the assumed course of the AA graft ing reaction. The decrease of nitrogen content on PU-COOH max is a result of AA presence on the surface. Since AA does not contain N atoms, nitrogen percentage dropped from 4.3% in PU to 0.8% in PU-COOH max . However, nitrogen could still be detected in PU-COOH max , thus the PU surface was not completely covered with PAA. In PU-REDV nitrogen percentage rose to 3.5%. This confi rms that peptide molecules of relatively high nitrogen content (amine groups and guanidine group in arginine) were successfully coupled.

Conclusion
In this work PU fi lms were modifi ed with peptides in order to prepare specifi c scaff olds for tissue engineering applications. AA has been graft ed to PU using a cerium   initiator. On the contrary to other papers suggesting that the presence of hydroxyl groups is essential for successful grafting with ceric ions, we propose a method with the omission of the PU hydroxylation step. The process of AA grafting has been analyzed in terms of reaction time, temperature and AA concentration. Successful peptide conjugation has been confirmed with ELISA assays. PU, PU-COOH max and PU-REDV materials were additionally characterized with SEM, XPS, FTIR-ATR and contact angle measurements. All these methods indicated that PU was efficiently grafted with AA and that it was possible to attach peptides to the PU-COOH surface. In continuation of this work, prepared materials will be contacted with human endothelial cells in order to assess their potential as scaffolds for in situ endothelialization.

Conflict of interest:
The authors declare that they have no conflict of interest.