Surface phosphonation enhances hydroxyapatite coating adhesion on polyetheretherketone and its osseointegration potential

Accepted Manuscript Surface phosphonation enhances hydroxyapatite coating adhesion on polyetheretherketone and its osseointegration potential Hesameddin Mahjoubi, Emily Buck, Praveena Manimunda, Reza Farivar, Richard Chromik, Monzur Murshed, Marta Cerruti PII: DOI: Reference:

S1742-7061(16)30522-0 http://dx.doi.org/10.1016/j.actbio.2016.10.004 ACTBIO 4470

To appear in:

Acta Biomaterialia

Received Date: Revised Date: Accepted Date:

6 May 2016 28 September 2016 3 October 2016

Please cite this article as: Mahjoubi, H., Buck, E., Manimunda, P., Farivar, R., Chromik, R., Murshed, M., Cerruti, M., Surface phosphonation enhances hydroxyapatite coating adhesion on polyetheretherketone and its osseointegration potential, Acta Biomaterialia (2016), doi: http://dx.doi.org/10.1016/j.actbio.2016.10.004

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Surface

phosphonation

enhances

hydroxyapatite

coating adhesion on polyetheretherketone and its osseointegration potential Hesameddin Mahjoubia, Emily Bucka, Praveena Manimundaa, Reza Farivarb, Richard Chromika, Monzur Murshedc, Marta Cerrutia* a.

Materials Engineering, McGill University, Montreal, Quebec H3A 0C5, Canada

b.

Department of Ophthalmology, McGill University, Montreal, Quebec H3A 1A1, Canada

c.

Division of Experimental Medicine, Department of Medicine, McGill University, Montreal, Quebec H3G 1Y6, Canada

Corresponding author: *Marta Cerruti, Ph.D. Address: Materials Engineering, McGill University, 3610 University St., 2M020 Wong Building, Montreal, QC H3A 0C5, Canada; E-mail: marta.cerruti@ mcgill.ca; Phone: 1 (514) 398-5496; Fax: 1 (514) 398-4492

Abstract Polyetheretherketone (PEEK) has excellent mechanical properties, biocompatibility, chemical resistance and radiolucency, making it suitable for use as orthopedic implants. However, its surface is hydrophobic and bioinert, and surface modification is required to improve its bioactivity. In this work, we showed that grafting phosphonate groups via diazonium chemistry enhances the bioactivity of PEEK. Decreased contact angle indicated reduced hydrophobicity as a result of the treatment and X-ray photoelectron spectroscopy (XPS) confirmed the attachment of phosphonate groups to the surface. The surface treatment not only accelerated hydroxyapatite (HA) deposition after immersion in simulated body fluid but also significantly increased the adhesion strength of HA 1

particles on PEEK. MC3T3-E1 cell viability, metabolic activity and deposition of calcium-containing minerals were also enhanced by the phosphonation. After three months of implantation in a critical size calvarial defect model, a fibrous capsule surrounded untreated PEEK while no fibrous capsule was observed around the treated PEEK. Instead, mineral deposition was observed in the region between the treated PEEK implant and underlying bone. This work introduces a simple method to improve the potential of PEEK-based orthopedic implants. Keywords: diazonium chemistry, adhesion strength, biomineralization, PEEK, in vivo osseointegration

1. Introduction Polyetheretherketone (PEEK) is widely used for orthopedic implants, such as skull plates, spinal cages and arthroscopic suture anchors, [1-4] since it has excellent mechanical strength, biocompatibility, chemical resistance, radiolucency, and can withstand sterilization by steam or gamma radiation [3, 5-7]. However, it does not integrate well with bone so PEEK is typically modified either chemically or physically to enhance its osseointegration [8-10] by overcoming the low surface energy and hydrophobicity and increase the growth rate and adhesion of osteoblasts on the PEEK surface [11-13]. Techniques to modify PEEK include sandblasting [14], surface porosity [15], oxygen plasma treatment [16], and titanium [17] and/or hydroxyapatite (HA; Ca10(PO4)6(OH)2), the mineral component of bone, coatings [18, 19]. Most commercially available PEEK implants are coated with HA or titanium through thermal plasma spray

2

[3], but HA coatings can also be applied with photochemical deposition, radio-frequency magnetron sputtering, aerosol deposition, and electron beam deposition [8-10, 17, 20, 21]. The plasma-sprayed HA coatings suffer from insufficient adhesion to PEEK and tend to delaminate due to low adhesion strengths [22, 23]. Aside from this, plasma spray coatings are expensive and cannot be easily applied to PEEK implants with complex geometry. On the other hand, wet chemical techniques, such as reduction and covalent grafting, can be used to introduce functional groups, such as carboxylates, hydroxyl and amino groups, known for improving cell adhesion [24, 25] and they can be easily applied to implants with complex shapes. Diazonium chemistry can functionalize the surface of a vast range of materials including metals, carbon-based materials and polymers [26-28]. A reactive radical is formed from an aniline-based precursor that attacks the substrate forming a covalent bond with the surface [29, 30]. The grafted aniline layer can then be reactivated and modified with any nucleophilic compound, allowing for the introduction of many different functional groups [31]. Grafting aniline derivatives on silk fibroin protein films was achieved through a diazonium reaction and the modified surfaces had increased hydrophilicity and enhanced human mesenchymal stem cells (hMSCs) attachment, proliferation and differentiation on the films [32]. The immobilization of insulin on layerby-layer films composed of alternating diazoresin and pectin layers was also performed with diazonium chemistry and the modified films exhibited increased hMSC proliferation and osteogenic activity [33]. Our lab has also used diazonium chemistry to incorporate phosphonate groups on the surface of three-dimensional poly(D, L-lactic) acid (PDLLA) scaffolds, which accelerated HA particle deposition and increased the viability, metabolic

3

activity and mineralization of osteoblast and chondrocytes cultured on the modified scaffolds [34]. Here, we modified the surface of PEEK substrates with a combination of physical and chemical treatments. We employed sandblasting and a two-step diazonium chemistry method [34] that formed a phosphonate layer on PEEK. Phosphonate groups attract calcium ions, which enhance not only the nucleation and growth of HA [35, 36], but also osteoblast proliferation and mineralization [34]. Our results show that the treatment increases the adhesion strength of HA particles as well as the metabolic activity and mineralization of MC3T3-E1 cells on PEEK and improves the osseointegration potential of PEEK in vivo.

2. Materials and Methods 2.1.

Materials

PEEK was purchased from McMaster-Carr. 2-aminoethylphosphonic acid (H2NCH2CH2P(O)(OH)2) (AEPA), calcium chloride (CaCl2), di-potassium hydrogen phosphate trihydrate (K2HPO4.3H2O), hydrochloric acid (HCl), magnesium chloride hexahydrate (MgCl2.6H2O), p-phenylenediamine (C6H4(NH2)2), potassium chloride (KCl), sodium chloride (NaCl), sodium hydrogen carbonate (NaHCO3), sodium nitrite (NaNO2),

sodium

sulfate

(Na2SO4),

and

tris-hydroxymethyl

aminomethane

(NH2C(CH2OH)3) (Tris) were purchased from Sigma-Aldrich Canada. Hypophosphorous acid solution (50-52% wt. in H2O) (H3PO2) was purchased from ACP Chemicals Canada.

4

2.2.

PEEK substrate preparation and sandblasting

PEEK substrates with dimensions 20x20x1.5 mm were prepared by cutting discs from a square shaped PEEK bar (2x2cm) using an IsoMet 5000 linear precision saw (BUEHLER, Whitby, ON, Canada). The samples were polished down to 800 with silicon-carbide paper (Paper-c wt, AA Abrasives, Philadelphia, PA). A series of samples were further modified physically by sandblasting with a Pony Sandblaster Complete (Buffalo Dental Manufacturing, Syosset, NY). Finally, all samples were cleaned by sonication in an ultrasonic bath (8893 ultrasonic, Cole-Parmer, Montreal, QC, Canada) with distilled water and ethanol for 15 minutes in each solution at room temperature.

2.3.

Chemical modification

The phosphonation was performed using a two-step method previously developed in our lab [34]. In the first step, 345 mg of sodium nitrite and 540 mg of p-phenylenediamine were dissolved in 50 ml of 0.5 M HCl in a glass beaker to form diazonium cations, which are stable at pH lower than 2.5 [29]. Then, PEEK substrates were introduced into the solution, and 2.5 ml of H3PO2 were added as a reducing agent. The solution was stirred at room temperature for 2 h after which the samples were removed from the solution and sonicated in distilled water for 10 minutes to remove any physiosorbed diazonium cations. This step led to spontaneous grafting of a poly(aminophenylene) layer on the surface of PEEK substrates. In the second step, amino-functionalized samples were reacted for 2 h under constant stirring in a 50 ml solution of 0.5 M HCl containing 100 mM NaNO2, 10 mM AEPA and 2.5 ml of H3PO2 as reducing agent. This step grafted phosphonate groups on the poly(aminophenylene)

5

layer that was produced on the surfaces in the previous step. At the end of the 2h, the samples were sonicated in distilled water for 10 minutes to remove any remaining physiosorbed groups from the surfaces. Both polished and sandblasted PEEK substrates were treated with diazonium chemistry, resulting in four different surface conditions: polished (PEEK-P), polished and phosphonated (PEEK-PP), sandblasted (PEEK-S), and sandblasted and phosphonated (PEEK-SP) samples.

2.4.

Water contact angle measurement

Water contact angles were measured using the sessile drop technique with a video-based contact angle measurement device (OCA 20, Future Digital) equipped with an auto dispenser. A sessile water droplet (30 µl) was placed on the surface of PEEK samples and an image of droplet was taken after 30 seconds. The OCA20 software calculated the contact angle by measuring the tangent angle at the point of contact between water, PEEK and air. The test was conducted on three spots per sample for six samples of each surface condition. The average values and standard deviations are reported.

2.5.

SBF immersion test

Six PEEK samples from each surface condition were immersed in simulated body fluid (SBF) for 10 days at 37°C. The concentrated SBF solution, 1.5x SBF, was prepared according to the protocol developed by Kokubo et al with the composition listed in Table 1 [37], and 1.5x SBF was used to speed up HA deposition. The 1.5x SBF solution was

6

changed every 3 days in order to maintain a constant concentration of ions for better simulation of the exchange of body fluid in vivo. Table 1. Concentration of reagents used to prepare 1.5x SBF

2.6.

Physical and chemical characterization

Characterization of the chemical composition of the PEEK samples after phosphonation and SBF immersion was conducted by X-ray photoelectron spectroscopy (XPS) using a Thermo Scientific Kα spectrometer equipped with an Al Kα x-ray source (1486.6 eV, 0.843nm) and x-ray spot size of 400 µm in diameter. Six points were randomly selected on the surface of each sample and three samples per condition were tested for surface composition analysis. The samples were hit with a flood gun shooting low energy electrons during measurements to prevent charging effects on the polymeric surfaces. The acquired spectral data were analyzed with the Thermo Avantage Software (Version 4.60). XPS peak deconvolution was performed with a fixed LorenztianGaussian mixture (L/G = 30%), smart background and no limitations on the boundary. Fitting was performed for the added peaks with the NLLS (non-linear least squares) function [10] and peak positions were added according to values found in literature for similar chemical groups. The surface morphology of the PEEK samples after immersion in 1.5x SBF was characterized with scanning electron microscopy (SEM, Hitachi SU3500). The samples were coated with a 15-nm Au coating and images were collected under an accelerating voltage of 15 kV. Fourier transform infrared (FTIR) spectra of the samples before and after SBF immersion were measured using a Bruker Tensor FTIR spectrometer in attenuated total

7

reflection (ATR) mode [38]. The spectra were collected in the range of 400 to 4000 cm-1 using a DTGS detector and 512 scans were taken with 4 cm-1 resolution. FTIR measurements were performed for three points per sample and two samples per condition were analyzed. The background was automatically subtracted from all spectra. The peak area calculation used to quantify the mineralization was performed by applying an integration profile to all spectra, which calculated the area between two manually selected points. Only the area above the baseline was included for the measurements.

2.7.

HA adhesion strength measurement

The adhesion strength of the HA particles deposited during SBF immersion was measured using a nanoscratch tester, equipped with Hysitron Triboindentor (Hysitron, Minneapolis USA). Prior to the test, scanning probe microscope (SPM) images of the HA particles were recorded and line profiles were extracted to determine the particle diameter. A diamond cono-spherical tip with tip radius of 20 µm was used for scratch testing. A normal load of 1 mN was applied to the sample and the tip was moved 10 µm laterally. During the process, the tip displaced the particle adhered to the PEEK surface. A second SPM image was then recorded to identify the particle position after the scratch test, and the lateral force experienced by the indenter during the process of particle displacement was used to calculate the adhesion strength. Five scratch tests were performed on each sample and the average of the five tests is presented. The adhesion strength was calculated according to the commonly used stress equation as below,
 ℎ  ℎ 
 =

∆    



(1)

8

where ‘A’ is the HA particle contact area and ∆F is the difference between maximum lateral force and the baseline measured during the particle displacement. A scratch measurement was also performed on neat PEEK surfaces since the diamond tip not only displaces the HA particle but also scratches the PEEK surface. To extract the adhesion force from the particle alone, the lateral force measured from scratching the neat PEEK was subtracted from that measured in the presence of HA particles. Scratch tests were performed on only polished (PEEK-P) and polished and phosphonated (PEEK-PP) samples.

2.8.

MC3T3-E1 cell culture

MC3T3-E1 pre-osteoblasts (subclone 14) were purchased from American Type Culture Collection (ATCC CRL-2594). The cells were cultured in Minimum Essential Medium alpha containing 2mM L-glutamine (MEMα, A10490, Gibco®, Canada) that was supplemented with 10% Fetal Bovine Serum (FBS, 12483-020, Gibco®, Canada) and 100 U/ml penicillin streptomycin (Invitrogen, Canada) at 37°C under 5% CO2 in a humidified incubator. PEEK discs with 18 mm diameter and 1.5 mm thickness were prepared and sterilized by immersion in anhydrous ethanol for 10 minutes. The cells were seeded on the discs at density of 2.5 × 106 and 6.5 × 106 cells/cm2 for proliferation and mineralization assays, respectively. The experiments were performed on three samples per condition and all measurements were taken in triplicates for each sample. To enhance the mineralization process, 100 µg/mL ascorbic acid, 5 mM β-glycerol phosphate, and 10 mM dexamethasone were added to the culture medium.

9

2.8.1.

Alamar blue assay for assessment of MC3T3-E1 cell viability/metabolic activity

The cell viability/metabolic activity of cultures were examined with the addition of Alamar Blue solution (DAL1025, Invitrogen) to the medium after 7 days of culture at a volume equal to 10% of cell culture medium. The fluorescence at 610 nm upon excitation with radiation at 560 nm was measured using a microplate reader (Infinite 200, Tecan) after 1, 2, 3 and 4 h of incubation at 37°C with the dye.

2.8.2. Alizarin red staining for evaluation of MC3T3-E1 cell mineralization After 1 month of incubation in the mineralization medium, 40 mM Alizarin red staining (A5533, Sigma-Aldrich) was added to the MC3T3-E1 cultures, incubated for 5 minutes and washed carefully in distilled water. The bound dye was dissolved in 10% acetic acid glacial and the absorbance of the resulting solution was measured spectrophotometrically at 405 nm using a microplate reader (Infinite 200, Tecan) to quantify the amount of deposited minerals.

2.9.

In vivo experiments

The in vivo biocompatibility and osseointegration of the PEEK samples was evaluated with a rat critical size calvarial defect model [39]. Prior to implantation in the rats, the PEEK discs were sterilized with ethylene oxide, and XPS analysis after sterilization was performed on three discs to ensure that sterilization did not damage the

10

coating. Each rat received two discs (PEEK-S and PEEK-SP) so each animal served as its own control to reduce the number of animals used in the study. Three rats were used for statistical significance. The animal study was conducted in accordance with regulations set by the Canadian Council on Animal Care and the protocol was revised and accepted by McGill ethics committee before beginning the experiments.

Adult wild-type Sprague Dawley rats, each weighing 350 g, were 6 months old at the time of experiment. Immediately before the surgery, the rats received an injection of glycopyyrolate (.01mg/kg). Following the injection, the rats were placed on a stereotaxic device to secure their heads and isofluorane was dosed to effect. Once anesthetized, the scalps were cleaned and an incision in the scalps was made with a scalpel. The soft tissue was removed to expose the skull and two holes of 5-mm diameter were drilled with a circular-hole saw. The underlying bone was carved to a depth of approximately 1-mm with a high-speed drill and one of each PEEK disc (PEEK-S and PEEK-SP) were placed in the skull of each rat and secured with Vetbond. Vetbond was applied only to the edges of the disc where osseointegration was not studied. The area was then rinsed with saline and the scalp was sutured closed. The animals were kept on a heating pad until they recovered from the isofluorane, after which they were transferred to their cages. The animals were closely monitored for signs of irritation or discomfort, but none showed any signs. Three months after the surgery, the rats were sacrificed under deep isofluorane and carbon dioxide followed by decapitation. Their skulls were extracted and stored in 70% ethanol for further study by histology.

2.9.1.

Histology

11

Histology was performed as described previously [40]. Briefly, the extracted skulls were fixed in 4% paraformaldehyde overnight and embedded in poly(methyl methacrylate) for sectioning. At least three undecalcified sections (30 µm) from each skull were cut and stained with von Kossa and counterstained with van Gieson to differentiate mineralized and non-mineralized areas of the tissue.

2.10.

Statistical analysis

The presented results include averages and standard deviations of the mean. We performed statistical analysis by t-test, one-way ANOVA, and p < 0.05 was considered significant.

3. Results 3.1.

Phosphonate coating chemical characterization

The aryldiazonium salt was formed in situ by reacting p-phenylenediamine with one equivalent of NaNO2 in an acidic environment (pH<2.5). These cations were reduced with H3PO2 to generate stable aminophenyl radicals [29, 41] according to the reduction mechanism of aromatic diazonium cations by H3PO2 previously described [42]. Upon introduction of PEEK films into the solution containing aminophenyl radicals, a poly(aminophenylene) (PAP) film with a multilayer structure was formed on their surfaces (Figure 1a) [29]. The aminated PEEK films were then washed, sonicated and transferred into a 0.5M HCl solution. In order to bind the target functional group, AEPA in this case, the PAP layer was diazotized and transformed into poly(diazophenylene) layer (PDP) with the addition of NaNO2 to the solution [29, 43]. Then, H3PO2 was added

12

to the solution to reduce the diazonium cations present on the PDP layer and transform them into radicals. Finally AEPA was introduced into the solution and the reaction between the radicals and AEPA led to the formation of a phosphonate-terminated multilayer film with the structure shown in Figure1b [34]. Evidence of successful grafting of the phosphonate groups on PEEK was provided by X-ray photoelectron spectroscopy (XPS). The atomic percentages of nitrogen, N, and phosphate, P, were quantified from the XPS survey scans and the results are summarized in Table 2. Neither P nor N was detected on PEEK after polishing (PEEK-P) or sandblasting (PEEK-S) but both P and N were identified after phosphonation on PEEK-PP (1.2 ± 0.3%P and 3.5 ± 0.2%N) and PEEK-SP (1.1 ± 0.4%P and 3.2 ± 0.4%N). The small deviations found between measurements performed on different points of the surface of different samples confirmed the homogeneity of the grafted layer and the reproducibility of the diazonium chemistry technique. The high-resolution N1s spectrum for PEEK-PP is shown in Figure 1c. The peak at binding energy (BE) of 399.4 eV was attributed to amino groups and azo bridges (N=N-) [34, 44], both of which are present in the PDP layers [34, 43, 45]. The shoulder at BE of 401.9 eV was attributed to the presence of ammonium groups (-NH3+), which have also been reported in diazonium grafted layers [34, 46]. The peak at BE of 406 eV corresponded to nitro groups, which may result from the use of NaNO2 in the grafting process or from solution contamination [34, 47, 48]. The high-resolution P2p spectrum for PEEK-PP is shown in Figure 1d and the single peak at BE of 134.1 eV was deconvoluted into two peaks that are characteristic for phosphonate groups [49]. Similar results were observed for PEEK-SP.

13

Figure 1. (a) Grafted poly(aminophenylene) (PAP) layer on the PEEK surface and (b) phosphonate-terminated multilayer after reaction of 2-aminoethylphosphonic acid (AEPA) with the PAP layer. Representative high-resolution (c) N1s and (d) P2p XPS spectra of PEEK-PP surfaces.

Table 2. Summary of data acquired from PEEK surfaces before and after diazonium treatment and SBF immersion test (All values indicated with * in the same line exhibit statistically significant differences from each other (p<0.0001), n=3 with six points measured per sample)

After confirming the chemical modifications with XPS, the water contact angles of all PEEK substrates were measured and are summarized in Table 2. The water contact angle of PEEK-P decreased from 76°±1 to 67°±1 after phosphonation (PEEK-PP) and that of PEEK-S decreased from 94.4°±0.8 to 82°±1 after phosphonation (PEEK-SP). PEEK-S was more hydrophobic than PEEK-P due to its rough surface, [50] but both PEEK-P and PEEK-S samples were less hydrophobic after phosphonation.

3.2.

SBF Immersion

To assess if the introduction of phosphonate groups on PEEK enhanced its ability to promote mineralization in a cell-free environment, we immersed the samples in 1.5x SBF for 10 days. The presence of individual flaky particles arranged in an overall spherical morphology was observed on all surfaces as shown by the scanning electron microscopy images in Figure 2. The agglomerates observed on the surface of sandblasted samples (PEEK-S and PEEK-SP) were larger and covered more of the sample surface than those formed on the polished samples (PEEK-P and PEEK-PP). The same trend was observed for the chemical treatment where surfaces with phosphonate groups (PEEK-PP

14

and PEEK-SP) promoted more mineral deposition than surfaces without phosphonate groups (PEEK-P and PEEK-S). The combination of both sandblasting and phosphonation (PEEK-SP) resulted in complete surface coverage by the particles whereas only partial coverage was seen for all other PEEK surfaces. Figure 2. SEM images from surface of (a) PEEK-P, (b) PEEK-PP, (c) PEEK-S and (d) PEEK-SP after 10 days immersion in 1.5x SBF solution

To identify and quantify the composition of the particles found on the PEEK surfaces after SBF immersion, XPS survey spectra were measured and the results are summarized in Table 2. Calcium, Ca, and phosphate, P, were detected on all specimens after immersion in SBF and the Ca/P ratio for all was 1.6 ±0.1, which falls in the range of Ca/P ratio expected for HA [51]. More Ca and P were found on the phosphonated substrates (PEEK-PP and PEEK-SP) than on the non-phosphonated substrates (PEEK-P and PEEK-S).. This confirms the observations from the SEM images that the phosphonate groups introduced by diazonium chemistry promoted an increase in the rate of HA deposition. More Ca and P were found on PEEK-S than on PEEK-P, which was ascribed to the additional nucleation sites for HA deposition provided by the rough morphology of PEEK-S, Overall, the combination of sandblasting and phosphonation significantly increased the amount of HA precipitated on PEEK in comparison with all other conditions. We further analyzed the nature of the precipitates found after SBF immersion with Fourier transform infrared spectroscopy (FTIR). In addition to the characteristic peaks for PEEK, all spectra measured on the samples immersed in SBF (Figure 3)

15

contained peaks at 1034 and 963 cm-1, which were attributed to the ν3 and ν1 stretching modes of the phosphate groups present in HA [34, 52]. This further confirmed that the precipitates found on all samples contain HA. We also quantified the relative amount of HA that deposited on each sample by calculating the ratio of the area of the peaks at 1034 cm-1 and 1227 cm-1, which are attributed to the ν3 phosphate stretching of HA and the asymmetric stretching of diphenyl ether (DE) groups present in PEEK, respectively [53]. The ratios are included in Table 2, and the highest ratio was measured on PEEK-SP, confirming that the combination of phosphonation and sandblasting led to the largest amount of HA precipitation on PEEK. Figure 3. Representative FTIR spectra of (a) PEEK-SP, (b) PEEK-S, (c) PEEK-PP, (d) PEEK-P surfaces after 10 days of immersion in concentrated SBF (1.5 x). The spectrum of PEEK is shown in (e). We normalized all the spectra to the intensity of the peak at 1227 cm-1 for asymmetric stretching of diphenyl ether groups from PEEK.

3.3.

Adhesion of HA coatings on PEEK

A nano-scratch test was performed on the polished PEEK samples (PEEK-P and PEEK-PP) to study the effect of phosphonation on the adhesion of HA on PEEK. We adapted the scratch tests developed by Dickinson and Yamada [54] and Chromik et al [55] to measure adhesion strength of single HA particles on the PEEK surfaces (Figure 4a). We recorded a scanning probe microscope (SPM) image to identify the particle position before (Figure 4c) and after (Figure 4d) the scratch test. We applied a normal load of 1 mN to the sample and moved the diamond tip by 10 µ m laterally. During this process, the tip displaced an HA particle that was adhered to the PEEK surface. Then, we calculated the adhesion strength of the single HA particle by analyzing the lateral force

16

experienced by the indenter during particle displacement (Figure 4b). Finally, the adhesion strength was calculated for PEEK-P and PEEK-PP according to equation 1 and the results are shown in Table 2. The measurement was not performed on PEEK-S and PEEK-SP due to their rough surface morphology, but the effect of roughness on HA adhesion has been previously reported as sandblasting is a common pre-treatment for improving HA adhesion on PEEK [51]. The adhesion strength of HA deposited from SBF immersion on PEEK-P was 15.5±0.5 MPa, and after phosphonation, the HA adhesion strength increased significantly by about 40% to 22 ± 3 MPa. Figure 4. Nano-scratch test: schematic of the setup, FN stands for applied normal force and FT is measured lateral force (a), plot of lateral force versus time recorded during particle displacement (b), and SPM images of the HA particle on PEEK-PP before (c) and after (d) the scratch test.

3.4.

In vitro MC3T3-E1 cell culture

To assess biocompatibility of phosphonated PEEK, the metabolic activity of 7day-old MC3T3-E1 cells was measured fluorometrically after incubation for 1, 2, 3 and 4h at 37ºC with Alamar blue staining (Figure 5a). With longer incubation of the cells with the dye, the fluorescent intensity increased, confirming the viability of the cells on all PEEK substrates. The metabolic activity of the cells cultured on PEEK-PP exhibited higher metabolic activity than cells on PEEK-S, indicating that phosphonation supported the osteoblasts better than sandblasting. Additionally, the metabolic activity of cells grown on PEEK-S and PEEK-SP was higher than that that of cells on PEEK-P and PEEK-PP, which was likely due to the presence of more anchorage sites provided by

17

sandblasting [56, 57]. The combination of phosphonation and sandblasting led to the highest metabolic activity of the cells. We also assessed the mineralization of MC3T3-E1 cells on the PEEK surfaces. The cells were cultured on the PEEK samples in differentiation medium containing ascorbic acid, β-glycerol phosphate, and dexamethasone for a period of 1 month and the amount of calcium-containing mineral deposits at the end of the incubation was measured with Alizarin red staining (Figure 5b). Consistent with osteoblast metabolic activity reported above, sandblasted PEEK (PEEK-S and PEEK-SP) showed higher mineralization than the corresponding polished samples (PEEK-P and PEEK-PP). Similarly, the phosphonated PEEK samples (PEEK-PP and PEEK-SP) promoted increased mineralization compared to the corresponding non-phosphonated PEEK samples. Cells cultured on PEEK-PP induced larger mineralization than those cultured on PEEK-S, which suggests that the phosphonation was more effective than sandblasting at promoting mineralization. Figure 5. (a) Alamar Blue assay for metabolic activity of MC3T3-E1 pre-osteoblast cells cultured on different PEEK substrates for 7 days (n=3). The time on the x-axis refers to incubation period at 37ºC with Alamar Blue dye. (b) Alizarin red assay for mineralization of MC3T3-E1 pre-osteoblasts cultured on various substrates for one month (n=3) (* p<0.05).

3.5.

In vivo

Given the promising results obtained in vitro in terms of biomineralization, cellular activity, and HA adhesion on PEEK-SP, we performed an in vivo study to evaluate the potential of the combined surface treatments to improve the osseointegration of PEEK implants. We used a rat critical size calvarial defect model and implanted PEEK

18

discs (PEEK-S and PEEK-SP) into 5-mm critical-sized defects in the skulls of three adult rats (Figure 6a). Prior to implantation in the rats, the PEEK discs were sterilized with ethylene oxide, and XPS analysis of the sterilized PEEK discs confirmed that no significant changes in the chemical structure and surface composition of the phosphonate coating were caused by sterilization (Figure 6b). The rats were healthy and showed no sign of discomfort during the three months following implantation. After this time, the animals were sacrificed and the skulls were extracted for analysis by histology to evaluate mineralization between the implants and underlying bone. Histology sections from PEEK-S and PEEK-SP implants were stained with von Kossa and counterstained with van Gieson to differentiate mineralized tissue (black color) from non-mineralized collagen (pink/red color) (Figure 6c-d and Figure S1). In the section from the PEEK-S disc (Figure 6c), a pink band of fibrous tissue was seen between the implant and bone, indicating the formation of a fibrous capsule. Since the implant was separated by the fibrous capsule, it most likely was unable to interact with the underlying bone (B). On the other hand, in the section from the PEEK-SP disc (Figure 6d), there was no sign of fibrous capsule formation beneath the implant. Instead, we observed diffused mineral deposition in between the implant and the solid flat bone. This mineralization appears to be organic matrix-independent as almost no osteoid (unmineralized collagen) was detected adjacent to the mineral deposits. Also, no osteoblast-like cells were detected in this region. These findings suggest that phosphonation promoted mineralization in vivo while also preventing the formation of a fibrous capsule.

19

Figure 6. (a) Schematic drawing of rat cranium and photograph taken during surgery illustrating the position of PEEK-S (left) and PEEK-SP (right) implants during the in vivo study (b) percentage of nitrogen and phosphate on the surface of PEEK-SP before and after ethylene oxide sterilization quantified from XPS survey spectra. Representative images of (c) PEEK-S and (d) PEEK-SP implants stained with von Kossa and van Gieson staining three months after surgery showing fibrous capsule (black arrows) formation around PEEK-S and mineralization (red asterisks) beneath PEEK-SP. Mineralized bone is stained black (B).

4. Discussion PEEK has excellent bulk properties, including mechanical strength and chemical resistance, which make it applicable for orthopedic implants. However, it is an inert polymer and suffers from poor bone integration [8]. Surface modification has been shown to improve its surface properties and enhance its adhesion to bone tissue. In this paper, we investigated the effect of sandblasting and phosphonation on improving PEEK’s interactions with bone in vitro and in vivo. The diazonium chemistry technique used in this paper was previously developed in our lab to bind phosphonate groups to the surface of biodegradable polymeric scaffolds [34]. Sandblasting PEEK resulted in increased viability of MC3T3-E1 cells compared to cells grown on smooth PEEK surfaces after one week of culture and significantly increased mineralization of MC3T3-E1 cells after one month of culture in differentiation medium. This can be partially attributed to the fact that the sandblasted surfaces accelerated the deposition of calcium phosphate minerals in a cell-free environment, meaning that the sandblasted surface mineralized more easily than the polished surface. These results are consistent with previous literature which show that sandblasting

20

influences the activity of osteoblasts through increased surface roughness and surface area [58], which provide more sites for attachment of cells on the surface [59]. For example, sandblasting improved the adhesion and proliferation of MG-63 cells on poly(amino acid)/hydroxyapatite/calcium sulfate biocomposites [58] and significantly increased the production and organization of extracellular matrix proteins, such as collagen type I, fibronectin and tenascin, by SaOs-2 cells on titanium substrates [60]. While phosphonation was already known to be able to speed up HA deposition on polymers and carbon nanotubes [34, 61, 62], here we showed for the first time that these groups also increase the adhesion strength of HA particles on PEEK. In fact, while the adhesion of HA that we measured on polished PEEK was almost identical to that found by Hahn et al. (15.5±0.9 MPa) [21], the addition of phosphonate groups significantly increased the adhesion of HA to 22 ± 3 MPa. This value was more than 40% higher than that reported for HA deposited via aerosol deposition [21] and more than 200% higher than the values reported for HA deposited via plasma spraying [22, 23]. Phosphonation also significantly increased the metabolic activity and mineralization of MC3T3-E1 cells, confirming the results previously obtained on PDLLA [36] and on bisphosphonatemodified PEEK [63]. No previous studies have specifically attempted to explain the effect of surface phosphonation on osteoblast activity, but some explanations have been proposed. Since phosphonates are negatively charged under physiological conditions, their negative charge has been suggested to promote the chelation of calcium ions from solution which can initiate the mineralization process [64]. Additionally, the negative charge has been shown to inhibit the adsorption of non-specific proteins, such as albumin, because these proteins are also negatively charged at physiological pH [65].

21

Furthermore, negatively charged surfaces have been shown to specifically adsorb celladhesive proteins like fibronectin for improved cell adhesion [64, 65]. When phosphonates are in solution (i.e. bisphosphonates), they are readily taken up by osteoclasts and inhibit bone resorption by signaling apoptosis in these osteoclasts [66]. Similar findings were observed for phosphonic acid-containing polymer substrates that triggered osteoclast apoptosis [64]. These phosphonic-acid containing polymer substrates were also shown to increase alkaline phosphatase activity, mineralization and collagen type I production of human osteoblasts when compared to the corresponding polymers containing no phosphonic acid [64], indicating that phosphonates play a role in improving the mineralization of osteoblasts as we also saw in this study. Last but not least, phosphonation completely prevented fibrous capsule formation and promoted mineral deposition around PEEK implants in vivo. This result is a significant improvement over previous reports. In fact, even some of the most promising methods, such as the creation of surface porosity to allow osteoblasts to grow inside the pores and form a strong anchor between the implant and the surrounding bone, did not completely prevent fibrous capsule formation [15]. Phosphonate coatings have been previously shown to improve osseointegration and biomechanical stability of titanium dental implants and chitosan scaffolds [67-69], and a small-scale clinical trial indicated that no safety issues were seen for 21 patients who received phosphonated titanium dental implants up to one year after implantation [70]. Phosphonated PEEK would be an excellent alternative to titanium since it has an elastic modulus much closer to that of native bone than titanium and exhibits both chemical resistance and radiolucency [19].

22

The substantial increase in osteoblast mineralization observed on sandblasted and phosphonated PEEK along with significant improvement in its osseointegration potential in vivo imply that these substrates could potentially be implanted directly, without a preliminary HA coating. In addition, the diazonium-based method described in this paper to apply the phosphonate coating on PEEK disks can be used on implants with complex shapes, which cannot be achieved with other reported techniques, such as oxygen plasma treatment [16]. Thus, the method could be combined with recently reported techniques leading to porous PEEK scaffolds [15] to produce implants that would be both mechanically stable and fully osseointegrated with the surrounding tissues. Further investigation will be required to understand the reason for the improved osseointegration potential observed in the present work. The presence of phosphonate groups decreased the hydrophobicity of PEEK (Table 2), which has been suggested as a reason for improved osseointegration potential [16, 38]. Still, phosphonate groups probably have a more specific effect than simply increasing hydrophilicity since previous work in our lab has demonstrated that only decreasing hydrophobicity is not sufficient for improving osteoblast viability and mineralization [34]; in fact, aminated poly(D,L-lactic acid) (PDLLA) surfaces behaved worse than unmodified PDLLA, while phosphonated ones increased osteoblast activity, similar to what we observed on PEEK in this work. It is possible that the high strength of adhesion strength of HA measured in vitro on the phosphonated PEEK surfaces (Figure 4 and Table 2) helped retain minerals formed in the region below PEEK-SP in vivo.

23

5. Conclusions We have shown that phosphonation via diazonium chemistry increased the hydrophilicity of PEEK, strongly enhanced the deposition and the adherence of HA coatings deposited from SBF, increased metabolic activity and mineralization of MC3T3E1 pre-osteoblast cells on PEEK, and improved in vivo osseointegration potential of PEEK implants as shown by the complete lack of a fibrous capsule and presence of minerals beneath the implant. This result is crucial for the development of HA coatings on PEEK that do not delaminate upon implantation, and may be partially responsible for the excellent osseointegration potential observed in vivo for phosphonated PEEK implants. Finally, using the diazonium-based method described in this paper, other chemical groups such as peptides or proteins could be grafted to PEEK surfaces, allowing for further tailoring of the tissue response to PEEK and leading to even better functionality of the implants.

6. Conflict of Interest The authors declare no conflict of interest.

7. Acknowledgements We acknowledge financial support for this project from the Natural Sciences and Engineering Research Council of Canada (NSERC) “Discovery” program, the “Nouveaux chercheurs” program from Fonds de recherché du Quebec-Nature et technologies (FQRNT), and the Canada Research Chair program. We also thank Mr. Stephen Nuara for his assistance during the animal surgery.

24

6. References [1] Camarini ET, Tomeh JK, Dias RR, da Silva EJ. Reconstruction of frontal bone using specific implant polyether-ether-ketone. Journal of Craniofacial Surgery 2011;22:2205-7. [2] Kulkarni AG, Hee HT, Wong HK. Solis cage (PEEK) for anterior cervical fusion: Preliminary radiological results with emphasis on fusion and subsidence. The Spine Journal 2007;7:205-9. [3] Kurtz SM, Devine JN. PEEK biomaterials in trauma, orthopedic, and spinal implants. Biomaterials 2007;28:4845-69. [4] Williams DF, McNamara A, Turner RM. Potential of polyetheretherketone (PEEK) and carbon-fibre-reinforced PEEK in medical applications. Journal of Materials Science Letters 1987;6:188-90. [5] Kitamura E, Stegaroiu R, Nomura S, Miyakawa O. Biomechanical aspects of marginal bone resorption around osseointegrated implants: considerations based on a three-dimensional finite element analysis. Clinical Oral Implants Research 2004;15:40112. [6] Katzer A, Marquardt H, Westendorf J, Wening JV, von Foerster G. Polyetheretherketone—cytotoxicity and mutagenicity in vitro. Biomaterials 2002;23:1749-59. [7] Godara A, Raabe D, Green S. The influence of sterilization processes on the micromechanical properties of carbon fiber-reinforced PEEK composites for bone implant applications. Acta Biomaterialia 2007;3:209-20. [8] Briem D, Strametz S, Schröoder K, Meenen NM, Lehmann W, Linhart W, Ohl A, Rueger JM. Response of primary fibroblasts and osteoblasts to plasma treated polyetheretherketone (PEEK) surfaces. Journal of Materials Science: Materials in Medicine 2005;16:671-7. [9] Henneuse-Boxus C, Dulière E, Marchand-Brynaert J. Surface functionalization of PEEK films using photochemical routes. European Polymer Journal 2001;37:9-18. [10] Noiset O, Schneider Y-J, Marchand–Brynaert J. Surface modification of poly(aryl ether ether ketone) (PEEK) film by covalent coupling of amines and amino acids through a spacer arm. Journal of Polymer Science Part A: Polymer Chemistry 1997;35:3779-90. [11] Boyan BD, Hummert TW, Dean DD, Schwartz Z. Role of material surfaces in regulating bone and cartilage cell response. Biomaterials 1996;17:137-46. [12] Lim JY, Shaughnessy MC, Zhou Z, Noh H, Vogler EA, Donahue HJ. Surface energy effects on osteoblast spatial growth and mineralization. Biomaterials 2008;29:1776-84.

25

[13] Mahjoubi H, Abdollahi S, Cerruti M. Surface modification of biodegradable polyesters for soft and hard tissue regeneration. In: Nazarpour S, editor. Thin Films and Coatings in Biology. Dordrecht: Springer Netherlands; 2013. p. 229-63. [14] Ourahmoune R, Salvia M, Mathia TG, Mesrati N. Surface morphology and wettability of sandblasted PEEK and its composites. Scanning 2014;36:64-75. [15] Evans NT, Torstrick FB, Lee CS, Dupont KM, Safranski DL, Chang WA, Macedo AE, Lin AS, Boothby JM, Whittingslow DC, Carson RA, Guldberg RE, Gall K. Highstrength, surface-porous polyether-ether-ketone for load-bearing orthopedic implants. Acta Biomater 2015;13:159-67. [16] Poulsson AH, Eglin D, Zeiter S, Camenisch K, Sprecher C, Agarwal Y, Nehrbass D, Wilson J, Richards RG. Osseointegration of machined, injection moulded and oxygen plasma modified PEEK implants in a sheep model. Biomaterials 2014;35:3717-28. [17] Han C-M, Lee E-J, Kim H-E, Koh Y-H, Kim KN, Ha Y, Kuh S-U. The electron beam deposition of titanium on polyetheretherketone (PEEK) and the resulting enhanced biological properties. Biomaterials 2010;31:3465-70. [18] Johansson P, Jimbo R, Kjellin P, Currie F, Chrcanovic BR, Wennerberg A. Biomechanical evaluation and surface characterization of a nano-modified surface on PEEK implants: a study in the rabbit tibia. International journal of nanomedicine 2014;9:3903-11. [19] Barkarmo S, Andersson M, Currie F, Kjellin P, Jimbo R, Johansson CB, Stenport V. Enhanced bone healing around nanohydroxyapatite-coated polyetheretherketone implants: An experimental study in rabbit bone. Journal of biomaterials applications 2014;29:737-47. [20] Rabiei A, Sandukas S. Processing and evaluation of bioactive coatings on polymeric implants. Journal of Biomedical Materials Research Part A 2013;101A:2621-9. [21] Hahn B-D, Park D-S, Choi J-J, Ryu J, Yoon W-H, Choi J-H, Kim J-W, Ahn C-W, Kim H-E, Yoon B-H, Jung I-K. Osteoconductive hydroxyapatite coated PEEK for spinal fusion surgery. Applied Surface Science 2013;283:6-11. [22] Ha S-W, Mayer J, Koch B, Wintermantel E. Plasma-sprayed hydroxylapatite coating on carbon fibre reinforced thermoplastic composite materials. Journal of Materials Science: Materials in Medicine 1994;5:481-4. [23] Beauvais S, Decaux O. Plasma sprayed biocompatible coatings on PEEK implants. Thermal Spray 2007: Global Coating Solutions 2007:371-6. [24] Noiset O, Schneider Y-J, Marchand-Brynaert J. Fibronectin adsorption or/and covalent grafting on chemically modified PEEK film surfaces. Journal of Biomaterials Science, Polymer Edition 1999;10:657-77.

26

[25] Noiset O, Schneider Y-J, Marchand-Brynaert J. Adhesion and growth of CaCo2 cells on surface-modified PEEK substrata. Journal of Biomaterials Science, Polymer Edition 2000;11:767-86. [26] Shaulov Y, Okner R, Levi Y, Tal N, Gutkin V, Mandler D, Domb AJ. Poly(methyl methacrylate) grafting onto stainless steel surfaces: Application to drug-eluting stents. ACS Applied Materials & Interfaces 2009;1:2519-28. [27] Griffete N, Lamouri A, Herbst F, Felidj N, Ammar S, Mangeney C. Synthesis of highly soluble polymer-coated magnetic nanoparticles using a combination of diazonium salt chemistry and the iniferter method. RSC Advances 2012;2:826-30. [28] Mitchell CA, Bahr JL, Arepalli S, Tour JM, Krishnamoorti R. Dispersion of functionalized carbon nanotubes in polystyrene. Macromolecules 2002;35:8825-30. [29] Viel P, Le XT, Huc V, Bar J, Benedetto A, Le Goff A, Filoramo A, Alamarguy D, Noel S, Baraton L, Palacin S. Covalent grafting onto self-adhesive surfaces based on aryldiazonium salt seed layers. Journal of Materials Chemistry 2008;18:5913-20. [30] Corgier BP, Marquette CA, Blum LJ. Diazonium−protein adducts for graphite electrode microarrays modification:  Direct and addressed electrochemical immobilization. Journal of the American Chemical Society 2005;127:18328-32. [31] Mahouche-Chergui S, Gam-Derouich S, Mangeney C, Chehimi MM. Aryl diazonium salts: A new class of coupling agents for bonding polymers, biomacromolecules and nanoparticles to surfaces. Chemical Society Reviews 2011;40:4143-66. [32] Murphy AR, John PS, Kaplan DL. Modification of silk fibroin using diazonium coupling chemistry and the effects on hMSC proliferation and differentiation. Biomaterials 2008;29:2829-38. [33] Mikulska A, Filipowska J, Osyczka AM, Nowakowska M, Szczubiałka K. Osteoinductive activity of insulin-functionalized cell culture surfaces obtained using diazonium chemistry. Frontiers in Chemistry 2014;2:117. [34] Mahjoubi H, Kinsella JM, Murshed M, Cerruti M. Surface modification of poly(d,llactic acid) scaffolds for orthopedic applications: A biocompatible, nondestructive route via diazonium chemistry. ACS Applied Materials & Interfaces 2014;6:9975-87. [35] Jiang S-D, Yao Q-Z, Ma Y-F, Zhou G-T, Fu S-Q. Phosphate-dependent morphological evolution of hydroxyapatite and implication for biomineralisation. Gondwana Research 2015;28:858-68. [36] Kim E-H, Yim S-B, Jung H-C, Lee E-J. Hydroxyapatite crystallization from a highly concentrated phosphate solution using powdered converter slag as a seed material. Journal of Hazardous Materials 2006;136:690-7.

27

[37] Kokubo T, Takadama H. How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 2006;27:2907-15. [38] Khoury J, Kirkpatrick SR, Maxwell M, Cherian RE, Kirkpatrick A, Svrluga RC. Neutral atom beam technique enhances bioactivity of PEEK. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 2013;307:630-4. [39] Spicer PP, Kretlow JD, Young S, Jansen JA, Kasper FK, Mikos AG. Evaluation of bone regeneration using the rat critical size calvarial defect. Nature protocols 2012;7:1918-29. [40] Marulanda J, Gao C, Roman H, Henderson JE, Murshed M. Prevention of arterial calcification corrects the low bone mass phenotype in MGP-deficient mice. Bone;57:499508. [41] Pinson J. Attachment of organic layers to materials surfaces by reduction of diazonium salts. Aryl Diazonium Salts: Wiley-VCH Verlag GmbH & Co. KGaA; 2012. p. 1-35. [42] Galli C. Radical reactions of arenediazonium ions: An easy entry into the chemistry of the aryl radical. Chemical Reviews 1988;88:765-92. [43] Laforgue A, Addou T, Bélanger D. Characterization of the deposition of organic molecules at the surface of gold by the electrochemical reduction of aryldiazonium cations. Langmuir 2005;21:6855-65. [44] Adenier A, Cabet-Deliry E, Chaussé A, Griveau S, Mercier F, Pinson J, Vautrin-Ul C. Grafting of nitrophenyl groups on carbon and metallic surfaces without electrochemical induction. Chemistry of Materials 2005;17:491-501. [45] Doppelt P, Hallais G, Pinson J, Podvorica F, Verneyre S. Surface modification of conducting substrates. Existence of azo bonds in the structure of organic layers obtained from diazonium salts. Chemistry of Materials 2007;19:4570-5. [46] Le XT, Viel P, Jegou P, Garcia A, Berthelot T, Bui TH, Palacin S. Diazoniuminduced anchoring process: an application to improve the monovalent selectivity of cation exchange membranes. Journal of Materials Chemistry 2010;20:3750-7. [47] Allongue P, Delamar M, Desbat B, Fagebaume O, Hitmi R, Pinson J, Savéant J-M. Covalent modification of carbon surfaces by aryl radicals generated from the electrochemical reduction of diazonium salts. Journal of the American Chemical Society 1997;119:201-7. [48] Gohier A, Nekelson F, Helezen M, Jegou P, Deniau G, Palacin S, Mayne-L'Hermite M. Tunable grafting of functional polymers onto carbon nanotubes using diazonium chemistry in aqueous media. Journal of Materials Chemistry 2011;21:4615-22.

28

[49] Moulder JF, Stickle WF, Sobol PE, Bomben KD. Handbook of x-ray photoelectron spectroscopy: Physical Electronics, Inc.; 1992. [50] Chau TT, Bruckard WJ, Koh PTL, Nguyen AV. A review of factors that affect contact angle and implications for flotation practice. Advances in Colloid and Interface Science 2009;150:106-15. [51] Robotti P, Zappini G. Chapter 9 - Thermal Plasma Spray Deposition of Titanium and Hydroxyapatite on Polyaryletheretherketone Implants A2 - Kurtz, Steven M. PEEK Biomaterials Handbook. Oxford: William Andrew Publishing; 2012. p. 119-43. [52] Rehman I, Bonfield W. Characterization of hydroxyapatite and carbonated apatite by photo acoustic FTIR spectroscopy. Journal of Materials Science: Materials in Medicine 1997;8:1-4. [53] Nguyen HX, Ishida H. Molecular analysis of the crystallization behavior of poly(aryl-ether-ether-ketone). Journal of Polymer Science Part B: Polymer Physics 1986;24:1079-91. [54] Dickinson ME, Yamada M. A new method for measuring shear adhesion strength of ceramic cold spray splats. Nanoscience and Nanotechnology Letters 2010;2:348-51. [55] Chromik RR, Goldbaum D, Shockley JM, Yue S, Irissou E, Legoux J-G, Randall NX. Modified ball-bond shear test for determination of adhesion strength of cold spray splats. Surface and Coatings Technology 2010;205:1409-14. [56] Orsini G, Assenza B, Scarano A, Piattelli M, Piattelli A. Surface analysis of machined versus sandblasted and acid-etched titanium implants. The International Journal of Oral & Maxillofacial Implants 2000;15:779-84. [57] Li D, Ferguson SJ, Beutler T, Cochran DL, Sittig C, Hirt HP, Buser D. Biomechanical comparison of the sandblasted and acid-etched and the machined and acid-etched titanium surface for dental implants. Journal of Biomedical Materials Research 2002;60:325-32. [58] Fan X, Ren H, Liu P, Wang P, Li H, Yan Y, Lv G. Effects of the surface modification of poly(amino acid)/hydroxyapatite/calcium sulfate biocomposites on the adhesion and proliferation of osteoblast-like cells. Journal of Applied Polymer Science 2015;132:n/a-n/a. [59] Bourges X, Gobin C. Sand-blasting method using biocompatible polymers. Google Patents; 2010. [60] Postiglione L, Domenico GD, Ramaglia L, Lauro AEd, Meglio FD, Montagnani S. Different titanium surfaces modulate the bone phenotype of SaOS-2 osteoblast-like cells. European Journal of Histochemistry 2004;49:213-22.

29

[61] Gemeinhart RA, Bare CM, Haasch RT, Gemeinhart EJ. Osteoblast-like cell attachment to and calcification of novel phosphonate-containing polymeric substrates. Journal of Biomedical Materials Research Part A 2006;78A:433-40. [62] Zhao B, Hu H, Mandal SK, Haddon RC. A Bone Mimic Based on the SelfAssembly of Hydroxyapatite on Chemically Functionalized Single-Walled Carbon Nanotubes. Chemistry of Materials 2005;17:3235-41. [63] Dennes TJ, Schwartz J. A nanoscale adhesion layer to promote cell attachment on PEEK. Journal of the American Chemical Society 2009;131:3456-7. [64] Ghag AK, Gough JE, Downes S. The osteoblast and osteoclast responses to phosphonic acid containing poly(e-caprolactone) electrospun scaffolds. Biomaterials Science 2014;2:233-41. [65] Tan J, Gemeinhart RA, Ma M, Mark Saltzman W. Improved cell adhesion and proliferation on synthetic phosphonic acid-containing hydrogels. Biomaterials 2005;26:3663-71. [66] Bassi AK, Gough JE, Downes S. A novel phosphonate for the repair of critical size bone defects. Journal of Tissue Engineering and Regenerative Medicine 2012;6:833-40. [67] von Salis-Soglio M, Stübinger S, Sidler M, Klein K, Ferguson SJ, Kämpf K, Zlinszky K, Buchini S, Curno R, Péchy P, Aronsson B-O, von Rechenberg B. A novel multi-phosphonate surface treatment of titanium dental implants: A study in sheep. Journal of Functional Biomaterials 2014;5:135-57. [68] Park J-W, Kim Y-J, Jang J-H, Kwon T-G, Bae Y-C, Suh J-Y. Effects of phosphoric acid treatment of titanium surfaces on surface properties, osteoblast response and removal of torque forces. Acta Biomaterialia 2010;6:1661-70. [69] Abueva CDG, Lee B-T. Poly(vinylphosphonic acid) immobilized on chitosan: A glycosaminoglycan-inspired matrix for bone regeneration. International Journal of Biological Macromolecules 2014;64:294-301. [70] Esposito M, Dojcinovic I, Germon L, Lévy N, Curno R, Buchini S, Péchy P, Aronsson B-O. Safety and efficacy of a biomimetic monolayer of permanently bound multi-phosphonic acid molecules on dental implants: 1 year post-loading results from a pilot quadruple-blinded randomised controlled trial. European Journal of Oral Implantology 2013;6:227-36.

30

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Table 1 Reagent

Concentration (ppm)

NaCl

12053

NaHCO3

533

KCl

338

K2HPO4.3H2O

347

MgCl2.6H2O

467

CaCl2

438

Na2SO4

108

Tris

9177

Table 2 PEEK-P

Diazonium

PEEK-PP

PEEK-S

PEEK-SP

P (%)

0.0

1.2 ± 0.3

0.0

1.1 ± 0.4

N (%)

0.0

3.5 ± 0.2

0.0

3.2 ± 0.4

76° ± 1

67°± 1

94.4°± 0.8

82°± 1

Ca (%)

2.4 ± 0.3*

6.4 ± 0.7*

17.1 ± 0.4*

19.4 ± 0.9*

P (%)

1.5 ± 0.2*

3.9 ± 0.5*

10.4 ± 0.3*

11.8 ± 0.6*

0.4± 0.2*

1.5± 0.2 *

3.2± 0.2*

7.2± 0.2*

15.5 ± 0.5*

22.0 ± 3*

N/A

N/A

XPS

modification Water contact angle

XPS SBF immersion test

IR spectroscopy 3 (1034 cm-1)/ DE (1227 cm-1)

HA adhesion strength (MPa)

HA adhesion strength (MPa)

Graphical Abstract

in vitro

30

*

20

10 0 Unmodified PEEK

Phosphonated PEEK

5 μm

Phosphonated PEEK

in vivo

50 μm

Statement of Significance

We have introduced phosphonate groups on the surface of PEEK substrates using diazonium chemistry. Our results show that the treatment not only increased the adhesion strength of hydroxyapatite particles deposited on PEEK in vitro by approximately 40% compared to unmodified PEEK, but also improved the metabolic activity and mineralization of MC3T3-E1 cells. When implanted in cranial defects in rats, the phosphonate coating enhanced the osseointegration of PEEK by successfully preventing the formation of a fibrous capsule and favoring mineral deposition between the implant and the surrounding bone. This work introduces a simple method to improve the potential of PEEK-based orthopedic implants, particularly those with complex shapes.