Biocompatible Organic Coatings Based on Bisphosphonic Acid RGD-Derivatives for PEO-Modified Titanium Implants

Currently, significant attention is attracted to the problem of the development of the specific architecture and composition of the surface layer in order to control the biocompatibility of implants made of titanium and its alloys. The titanium surface properties can be tuned both by creating an inorganic sublayer with the desired morphology and by organic top coating contributing to bioactivity. In this work, we developed a composite biologically active coatings based on hybrid molecules obtained by chemical cross-linking of amino acid bisphosphonates with a linear tripeptide RGD, in combination with inorganic porous sublayer created on titanium by plasma electrolytic oxidation (PEO). After the addition of organic molecules, the PEO coated surface gets nobler, but corrosion currents increase. In vitro studies on proliferation and viability of fibroblasts, mesenchymal stem cells and osteoblast-like cells showed the significant dependence of the molecule bioactivity on the structure of bisphosphonate anchor and the linker. Several RGD-modified bisphosphonates of β-alanine, γ-aminobutyric and ε-aminocaproic acids with BMPS or SMCC linkers can be recommended as promising candidates for further in vivo research.


Introduction
Traumatology and orthopedics generate high demand in implants for osseosynthesis, and, according to market forecasts, this demand will increase due to the spread of extreme activities and the aging of the population. Currently, various biomimetic approaches that ensure the biocompatibility of the implants are developed [1]. As a rule, they suggest changes in the architecture and composition of the surface layer so that the devices gain the properties of the bone tissue and cell membranes. In current medical practice, preference is given to implants made of titanium and its alloys, due to their bioinert properties and high corrosion resistance [2]. The modification of the titanium surface is achieved both through inorganic coatings that approximate the phase composition and morphology of coating [71]. The high surface area of the PEO coating allows deposition of various functional organic components, and the PEO coating acts as a sublayer increasing adhesion, and as the organic substance carrier [72][73][74][75][76].
Our recent study [15] showed that the combination of the PEO coating and RGD-modified bisphosphonic acid on nano-Ti gives a 45% increase in the number of proliferated cells compared to uncoated nano-Ti, and 66% compared to coarse grain Ti.
In the presented study, we expanded the number of hybrid RGD-functionalyzed molecules in order to uderstand how the structure of the bisphosphonate anchors and linkers (BMPS, EMCS, SMCC) affects the biological activity of the composite coatings based on inorganic porous PEO sublayer on titanium; moreover, the biological effect of the synthesized organic molecules was studied in vitro on a larger set of cell lines (mesenchymal stem cells, fibroblasts and osteoblast-like cells). Therefore, the aim of this work is to synthesize the hybrid molecules by chemical cross-linking of amino acid bisphosphonates with a linear RGD tripeptide through linkers that differ in length and structure for effective use in combination with inorganic porous PEO sublayer on titanium, and to identify the most promising options for the design of the proposed biomimetic coatings.

Results and Discussion
The experimental design concerns the variation of the anchors and linkers to attach the RGD peptide to the titanium implant. For the sake of the clarity, the following sample identification is used: Ti-for the metal substrate; Ti-PEO-for the inorganic PEO coating over the substrate; Ti-PEO-RGD-for the PEO coating soaked in the RGD aqueous solution; Ti-PEO-(number)-RGD-for the PEO coating soaked in the aqueous solution of the RGD-modified bisphosphonic acids, where the (number) indicates the compound in accordance with the reaction Scheme 1 and the corresponding description in Section 3.2. As a result, this sample identification shows the structures of the proposed composite coatings.

Synthesis of Bioactive Bifunctional Molecules
As a base for the design of the RGD-containing self-assembled monolayers on the PEO-modified meal surface, we synthesized aminoacid bisphosphonates 1-3 via the reaction of the corresponding amino acids (β-alanine, γ-aminobutyric and ε-aminocaproic acids) with PCl 3 in methanesulfonic acid (Scheme 1). We found that carrying out the synthesis at elevated temperature (85-90 • C) decreases the reaction time to 4-5 h, and, along with the separation of the products at pH = 6, increases the resulting yield of the amino bisphosphonates up to 85-89%. The N-maleimidosuccinimide linkers (BMPS, EMCS and SMCC) 4-6 were obtained in yields of 60-89% via the reaction of maleic anhydride, N-hydroxysuccinimide and aminocarboxylic acids in DMF [77,78]. The N-maleimido derivatives 7-14 for further modification by the oligopeptide were obtained by interaction of amino bisphosphonates with N-maleimidosuccinimide linkers. The application of acetone instead of dioxane as the solvent used elsewhere [79] helped to avoid the formation of relatively heavy side products whose appearance is an inevitable consequence of the dioxane cycle modification. Target RGD-modified amino acid bisphosphonates 15-22 were obtained by the reaction of the corresponding derivatives 7-14 with RGDC in an aqueous medium at pH = 7 by Michael reaction.
The structure of the RGD-modified aminoacid bisphosphonates was confirmed by mass-spectrometry MALDI TOF/TOF, 1D ( 1 H, 13 C, 31 P) and 2D NMR spectroscopy. In the MALDI TOF/TOF mass spectra of the hybrid molecules, the peaks with m/z corresponding to the molecular ions were detected. The 31 P-NMR spectra exhibited single resonance lines at δ P 17-18 ppm, characteristic to bisphosphonate groups. Signals corresponding to fragments of amino bisphosphonate, linker, and RGDC were observed in 1 H-and 13 C-NMR spectra. The formation of compounds 15-22 was accompanied by the disappearance of the signal of double bond protons of the N-maleimido fragment in derivatives 7-14 at δ H 6-7 ppm in the 1 H-NMR spectrum.  Figure 1 shows the top view and cross-section of the PEO coating. The coating exhibits a porous morphology, with porosity evaluated as 18.2 ± 2.5%. The coating thickness is 19.3 ± 1.5 µm, surface roughness 1.5 ± 0.2 µm. This type of morphology contributes towards cell proliferation, and it supplies a well-developed surface for the introduction of the organic compounds with bisphosphonate anchors into the resulting composite coating [15]. The coating surface porosity is 9.5 ± 0.8%; the real 3D structure of this well-developed network of pores can be seen elsewhere for a similar PEO coating [80]. As follows from Figure 2, the PEO coating consists of crystalline titania in two phase modifications-rutile and anatase (49 and 51% respectively). Also, the XRD pattern shows the peaks belonging to α-Ti substrate. This coating composition and morphology provides sufficient surface passivation for the implant prolonged operation in the corrosive fluids of a human body [81]. Also, this type of coating exhibits exceptional adhesion to the titanium surface as summarized elsewhere [5].

XPS Analysis of the Composite PEO Coating Functionalized with RGD Derivatives
Since the introduction of the RGD derivatives into the inorganic PEO coating was performed at a low concentration of 10 −3 M, it falls below the capabilities of the SEM and XRD techniques to detect organic compounds on the titania surface. Therefore, XPS studies were employed in order to verify the presence of the RGD derivatives within the coating; a typical survey spectrum is shown in Figure 3. In details the XPS spectra are presented in the Supplementary Materials. The survey XPS spectra collected from the PEO-coated titanium samples indicate typical Ti 2p, Ti 3s O 1s, C 1s, P 2p, N 1s, O KLL and Ti LMM Auger peaks. The spectra of the RGD-modified samples confirm changes in the surface layer chemistry compared to the unmodified PEO sample. As follows from the high-resolution XPS spectra, the organic molecules attached to the surface affect the relative content of nitrogen, phosphorus, carbon, oxygen, and sulfur ( Table 1). The most notable differences appear for the Ti2p, C1s and N1s peaks. The Ti2p signal decreased after the RGD modification, and the N1s and C1s signals increased. Also, the RGD-modified surface shows the presence of S2p signals, whereas for the Ti-PEO surface this signal does not appear. The changes in the atomic ratio (Ti2p/C1s) and (Ti2p/P2p) show that the titanium signal decreases for all the RGD modified types of the surface. This is consistent with the observations made elsewhere [27,82,83].

Electrochemical Behavior of the Composite PEO Coating Functionalized with RGD Derivatives
The electrochemical behavior of the composite PEO coating functionalized with RGD derivatives was studied using compounds 15-18 as examples. The polarization curves ( Figure 4) show notable differences among the surface having PEO and various RGD treatments. As expected, after the PEO, the surface passivates, and the E corr becomes nobler; the corrosion current i corr decreases, which is followed by an order of magnitude increase in the polarization resistance R p ( Figure 5). All the RGD functionalized coatings show the surface, nobler than that of the Ti-PEO sample. The cathodic parts of the polarization curves do not show significant differences compared to Ti and Ti-PEO samples. This indicates that cathodic corrosion processes are not influenced by this type of coating since the cathodic Tafel parts show almost similar slopes. However, the anodic behavior is significantly influenced by the coating. Ti-PEO sample shows a passivation region above −0.1 V(AgCl). The introduction of the RGDmodified bisphosphonates depassivates the surface; the corrosion current increases compared to that of Ti-PEO and Ti samples ( Figure 5). The depassivation region belongs to the range −0.1 to 0.1 V (AgCl) and ends with a pitting current tip followed by a passivation region above 0.1 V (AgCl). The changes in the anodic kinetics rise the E corr , and this is followed by the increase of the i corr . This indicates that the RGD functionalization increases the exchange currents, and this i corr behavior is supported by the decrease of R p .

In Vitro Test Results Supporting The Efficiency of Proposed PEO Coating Functionalized with RGD Derivatives
As follows from the literature analysis, the application of RGD-containing bisphosphonates to the titanium surface leads to the increase in cell adhesion, proliferation and mineralization [22,23,51,53]. The cell adhesion appeared to be sensitive to the anchor used [57,59].
The in vitro results presented in Figure 6 show that the different cell lines give dissimilar reactions to the same type of coating. The biological response also depends on the structure of the source amino bisphosphonate and the linker. It was shown that RGDC-oligopeptide itself introduced into the PEO coating without the anchor does not affect the cell life activity on the surface. A significant (up to 30-40%) increase in the fibroblast proliferation was observed when using the organic molecules with relatively short linkers and aminobisphosphonate fragments BMPS-β   Summarizing the in vitro studies, we can conclude that RGD functionalized hybrid molecules 15, 17, 21, 22 appear to be promising organic top coatings for modulation of the biological activity of the PEO modified surface of titanium implants. The differences in the cell behavior depending on the compound structure can be caused by their cytotoxicity and orientation on the PEO surface during adsorption; this requires further studies that are currently conducted in our research group.
Spectroscopic studies were performed by 1 H, 13 C and 31 P-NMR on an AVANCE-500 spectrometer (Bruker, Rheinstetten, Germany; operating frequency 500.17 MHz ( 1 H), 125.78 MHz ( 13 C) and 202.48 MHz ( 31 P)). D 2 O and CDCl 3 were used as internal standards and solvents. 31 P-NMR chemical shifts are given relative to the standard, an 85% solution of H 3 PO 4 in H 2 O (δ P 0 ppm). Samples were prepared in a standard tube with a diameter of 5 mm. The chemical shifts of carbon and hydrogen atoms are given on the scale δ (ppm) with respect to TMS, the KCCB values (J) are given in Hz. One and two-dimensional NMR spectra (COSY HH, HSQC, HMBC, NOESY) were recorded using standard pulse sequences. Mass spectra were obtained on MALDI-TOF/TOF Autoflex III system (Bruker) using 2,5-dihydroxybenzoic acid (2,5-DHB) or α-cyano-4-hydroxycinnamic acid (CHCA) as matrices.

Synthesis of Amino-1-hydroxyalkan-1,1-diyl-bisphosphonic acids 1-3
The synthesis of aminobisphosphonic acids was basd on a literature method [84] somewhat modified by us. A portion of the amino acid (0.025 mol) (β-alanine, γ-butanoic acid, ε-aminocaproic acid) is dissolved in 10 mL of methanesulfonic acid (MSA) in a flask with intensive stirring. Then, 7.0 mL (0.080 mol) of phosphorus (III) chloride is slowly added dropwise to the solution over 15 min. The contents of the flask are stirred under heating (80-85 • C) for 4-6 h, then cooled to room temperature. After cooling, 12 mL (0.67 mol) of water are added, and the reaction mixture is hydrolyzed at 105 • C for 4 h. After cooling, the pH of the solution is adjusted to 6-7 by adding to the mixture 12 mL of 50% aqueous sodium hydroxide solution. Then the reaction mass precipitated with 40 mL of MeOH. The precipitate is separated on the Schott filter. The crude product is dissolved in 20 mL of water, stirred at 50-70 • C for 30 min, then cooled and precipitated with 40 mL of MeOH. The solid phase is separated using a Schott filter.

Synthesis of (N-Maleimidoalkyl)succinimide Esters (4-6)
The synthesis of compounds was carried out according to the method described in [77,78]. A portion of 3 mmol of an amino acid is added to a solution of maleic anhydride (0.3 g, 3 mmol) in 4 mL of DMF, stirred for 2 h at room temperature. After complete dissolution, at a temperature of 0 • C, 0.43 g (3.75 mmol) of N-hydroxysuccinimide (NHS) and 1.24 g (6 mmol) of dicyclohexylcarbodiimide (DCC) are added to the solution. The reaction mass is stirred in an ice bath 5-10 min and then incubated at room temperature for a day. The formation of a white precipitate of dicyclohexylurea is observed. The precipitate is washed with 20 mL of water and 20 mL of methylene chloride. The organic layer is treated with 10 mL of water, extracted with methylene chloride, then 10 mL of 5% NaHCO 3 are added and extracted with methylene chloride (the procedure is carried out three times). The organic layer was dried over Na 2 SO 4 , the solution was filtered and the solvent was evaporated at the reduced pressure. The residue was purified by column chromatography on SiO 2 with CHCl 3 /MeOH (from 1:0 to 100:1) as an eluent to give compounds 4-6.

Synthesis of RGDC Derivatives 15-22
According to [79], 5 mg (0.01 mmol) of RGDC was dissolved in 1.45 mL of bidistilled water, the pH was adjusted to 7 by addition of 0.1 N NaOH (~120 µL). An equivalent amount of compounds 7-14 (0.01 mmol) was added to the solution. The reaction mixture was stirred for 1-2 h at 38-40 • C until a pinkish-violet tint appeared, and then the solvent was removed under reduced pressure.  (16). NMR and mass spectrometry data correspond to those reported in [15].

Metal Sample Preparation and PEO Coating
The metal samples 10 mm in diameter were cut using a wire-EDM out of a commercially pure Cp-Ti sheet, 0.5 mm thick. Further, the disk samples were polished on a SiC grit to achieve the surface roughness Ra < 0.1 µm. Before the plasma electrolytic oxidation, the samples were washed with deionized water and ultrasonically cleaned in isopropanol for 5 min. To achieve a good contact both the for the PEO and consequent electrochemical tests, a copper wire was attached into a small hole drilled at the sample edge. The contact place and the copper wire were coated with an epoxy resin which could withstand the electrochemical treatments. In order to produce the samples for the in vitro tests, a titanium screw holder was used to supply current to the sample. The holder was PEO coated to form a dielectric layer over its surface except the contact points prior to the treatment of the samples.
The plasma electrolytic oxidation was processed in a 10-L glass vessel equipped with a stainless steel heat exchanger arranged around its walls inside the electrolyte which comprised 20 g/L aqueous solution of Na 3 PO 4 ·12H 2 O of puris grade. The electrolyte temperature was maintained at 20 ± 1 • C using microcontroller regulation. The PEO process was run in a pulsed bipolar mode under the voltage regulation; the positive pulse was 470 V, negative 40 V; the frequency was 300 Hz. The positive pulse duration was 1.7 ms, the negative pulse was 0.87 ms; both separated with pauses of 0.38 ms. The PEO treatment duration was 5 min.

Surface Characterization
The PEO coating was studied using a JSM-6490LV scanning electron microscope (JEOL, Tokyo, Japan) at accelerating voltage of 10 kV. Both top view and cross-sections were examined. The coating thickness was measured using a Positector 6000 eddy current gauge (Defelsko, Ogdensburg, NY, USA) of N-type. The surface roughness was measured with TR-220 profilometer. In order to evaluate the coating porosity, the SEM images were processed using ImageJ software using the technique described elsewhere [88]. By adjusting the threshold in the grayscale images, the relative surface area taken by the objects of interest being darker than the neighboring places was calculated by the program in percents. The phase composition of the surface layer was characterised by an Ultima IV X-ray diffractometer (Rigaku, Tokyo, Japan) in Cu Kα radiation at 40 kV and 40 mA using 0.02 deg. step scan with 2 s exposure, from 20 to 80 degrees 2θ. Further, the XRD spectra were processed using the X'Pert Highscore Plus software (Philips) with PDF2 pattern database and a built-in SemiQuant algorithm was employed to quantify the amounts of the crystalline phases in the coating.
XPS spectra were obtained using a JEOL JPS 9010MX spectrometer equipped with an (Mg K α ) X-ray source. The pressure in the analytical chamber during the spectral acquisition was less than 7 × 10 −8 Pa. The survey spectra were collected from 0 to 1000 eV with a pass energy of 50 eV, and high-resolution spectra were collected for each detected element of interest (C, N, O, P, S, and Ti) with a pass energy of 10 eV. The JEOL SpecSurf Program V. 1.9.2 was used to identify the characteristic peaks, to calculate the elemental compositions, and to fit the peaks of the high-resolution spectra.

Electrochemical Tests
The electrochemical tests were carried out in Ringer's solution (0.86 wt% NaCl, 0.03 wt% KCl, 0.033 wt% CaCl2, pH 7.4) at temperature 37.0 ± 0.2 • C using a P-5X electrochemical system (Elins, Moscow, Russia). After the open circuit potential (OCP) was settled within ±20 mV for at least 30 min, the potentiodynamic polarization (PDP) test was run from −350 mV to +350 mV with respect to the settled OCP at a rate of 0.25 mV/s. The reference electrode was a silver chloride electrode filled with 3.5M KCl. The counter electrode was a platinum rod. The PDP results were processed using Tafel analysis [89]. For the PDP curves having linear Tafel regions at ±200 mV overpotentials, both anodic and cathodic slopes were used; for those having only cathodic linear Tafel regions, only cathodic slopes were used. The corrosion current i corr was evaluated at the intersection of cathodic and anodic (if available) tangent lines with the level of E corr . The polarization resistance Rp was evaluated as a slope of the polarization curve at±10 mV overpotentials for all the samples [90].

In Vitro Assessment
The PEO coated Ti samples were ultrasonically cleaned for 10 min in 95% ethanol and finally washed with deionized water, dried on air and sterilized by autoclaving at 134 • C. This temperature does not affect the PEO coating. In order to deposit the organic pore filler, the PEO coated Ti samples were put into a Petri dish with 10 −3 M solutions of the RGD-derivatives, which were preliminarily sterilized by filtration with CA 0.22 µm filter. In 1 h the samples were dried on air in a laminar box. Then, all the samples were put into a plastic 24-well tissue culture plate.
The FLECH-104, MSC or MG-63 cells suspension was placed into each well of the plate with the samples (0.8 mL containing 20·10 3 cells). The cells in the wells with uncoated Ti+PEO samples were treated as a control. The culture plates were incubated for 7 days in the standard conditions (37 • C, 5 vol% CO 2 ). The culture plate itself (polystyrene) was used as a blank.
The cell proliferation was determined by EZ4U assay (Biomedica, Vienna, Austria), a modification of MTT test, which evaluates the cell metabolic activity is proportional to the number of adherent cells. Three samples of each type were transferred after the incubation into another 24-well plate with 0.8 mL of fresh DMEM medium. Then 80 µL of activated EZ4U solution was added to every well and incubated at 37 • C, 5 vol% CO 2 for 3.5 h. Optical absorbance was measured using microplate reader (Spark10M, Tecan, Männedorf, Switzerland) at 450 nm with a reference wavelength of 620 nm. The optical density per mm 2 was calculated as: where a-absorbance of the test sample at 450 nm minus absorbance at 620 nm; A-absorbance of the control sample at 450 nm minus absorbance at 620 nm; a blank -the absorbance of the blank solution of DMEM with no cells at 450 nm and at 620 nm. The mean value and the standard deviation for four measurements of the optical density were calculated with respect to the control.

Conclusions
We have synthesized a set of RGD-derivatives of amino bisphosphonates, obtained from β-alanine, γ-aminobutyric and ε-aminocaproic acids, containing various linkers (BMPS, EMSC, SMCC) and used them as organic top coatings on porous PEO layer on titanium. The compounds can be introduced into the PEO coating pores due to the presence of bisphosphonate groups that facilitate physicochemical adsorption. Electrochemical studies showed that after RGD functionalization, the PEO surface gets slightly nobler, but the corrosion currents notably increase; also, the anodic PDP curves of the PEO surface lose the passivation region after the introduction of the RGD-derivatives, so their presence can be assessed by indirect electrochemical tests. The presence of organic molecules in the coating was confirmed by XPS spectroscopy. The appearance of compounds 15-20 on the surface provided a decrease in the Ti 2p line intensity and an increase of N1s and C1s line intensities; in addition, a significant decrease in the ratio of Ti 2p/C 1s and Ti 2p/P 2p was observed.
In vitro studies on proliferation and viability of fibroblasts, mesenchymal stem cells and osteoblast-like cells showed the dependence of the molecule bioactivity on the structure of the anchor and linker. In particular, RGD derivatives with relatively short bisphosphonate anchors and BMPS linker, as well as molecules containing a linker with a cyclohexyl fragment, increase cell proliferation on the surface of PEO-modified titanium. RGDC without anchor on Ti-PEO does not affect the cell proliferation. RGD-functionalyzed β-BMPS, γ-BMPS, γ-SMCC, ε-SMCC can be recommended as promising candidates for further in vivo research.