Effects of calcium-modified titanium implant surfaces on platelet activation, clot formation and osseointegration. Eduardo Anitua1,2, Roberto Prado2, Gorka Orive2, Ricardo Tejero2,3* 1

Private practice in implantology and oral rehabilitation in Vitoria, Spain

2

Biotechnology Institute (BTI), Vitoria, Spain.

3

Department of Biochemistry and Molecular Biology, University of the Basque Country

(UPV-EHU), Leioa, Spain

*

Main and corresponding author:

Ricardo Tejero Biotechnology Institute (BTI), Vitoria, Spain. C/ Leonardo Da Vinci, 14B 01510 Miñano (Álava), Spain Phone: +34 945 297 030; Fax: +34 945 297 031; Email: [email protected]

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences   between this version and the Version of Record. Please cite this article as an ‘Accepted Article’, doi: 10.1002/jbm.a.35240

 

Abstract The clinical success of load bearing dental and orthopaedic implants relies on adequate osseointegration. Because of its favourable properties, titanium is generally considered as the material of choice. Following implant placement, titanium surfaces establish an ionic equilibrium with the surrounding tissues in which calcium plays major roles. Calcium is a cofactor of the coagulation cascade that mediates plasma protein adsorption and intervenes in a number of other intra and extracellular processes relevant for bone regeneration. In this study, titanium surfaces were modified with calcium ions (Ca2+ surfaces) and their responses to in vitro and in vivo models were analysed. Unlike unmodified surfaces, Ca2+ surfaces were superhydrophilic and induced surface clot formation, platelet adsorption and activation when exposed to blood plasma. Interestingly, in vivo osseointegration using a peri-implant gap model in rabbit demonstrated that Ca2+ surfaces significantly improved peri-implant bone volume and density at 2 weeks and bone implant contact at 8 weeks as compared to the unmodified controls. The combination of Ca2+ surfaces with plasma rich in growth factors (PRGF) produced significantly more bone contact already at 2 weeks of implantation. These findings suggest the importance of the provisional matrix formation on tissue integration and highlight the clinical potential of Ca2+ titanium surfaces as efficient stimulators of implant osseointegration. Key

words:

titanium

surface

modifications,

calcium

ions,

coagulation,

osseointegration, plasma rich in growth factors (PRGF).

 

2

 

Introduction Titanium-based endosseous implants have been widely employed since the discovery of osseointegration because they allow predictable and long-lasting solutions for a variety of orthopaedic and dental reconstructions [1]. Although the processes occurring at the bone-titanium oxide interface are still not completely understood, research at this level has allowed the production of increasingly preforming implants. The most established modifications concern the implant physical characteristics [2,3] but more recently, researchers have introduced elements of the extracellular matrix (ECM) in the quest of an active interplay between the implant surface and the surrounding biomolecules and tissues [4]. Some of the inorganic elements of the ECM are metal ions essential for bone formation and repair [5]. Indeed, implant biocompatibility is believed to be determined, to a large extent, by the establishment of proper surface-ion equilibrium [6]. Calcium stands out among these bioinorganics because it has principal roles at almost every stage of the bone repair process: from the formation of the provisional matrix to biomineralization and bone remodelling [7-9]. It is believed to be also crucial in the osseointegration of implants: calcium alters the zeta potential of the titanium oxide layer and mediates protein adsorption, platelet adhesion and activation and cell signalling [10-12]. Calcium establishes intersurface electrostatic bridges between the negatively charged implant surface and the negatively charged residues of several proteins, modifying thereby the composition, orientation, and/or conformation of the proteins at the TiO2 interface [8,12-15]. Upon implantation, blood is the first tissue contacting the implant surface. The platelets contained within are pivotal in the processes of haemostasis and blood coagulation that precede tissue repair. For platelets to fulfil these roles, they have to become activated.

 

3

 

Platelet activation can be induced mechanically or by the addition of agonists such as thrombin, thromboxane A2, collagen, adenosine diphosphate (ADP) and also, calcium [7,15,16]. Calcium (Factor IV) mediates the binding of the platelet membrane phospholipids to Factor Xa and Factor IXa, which are required for the tenase and prothrombinase complexes to operate. These complexes convert prothrombin into thrombin (Factor IIa), which further triggers fibrin polymerization [17]. It has been shown that calcium ions at TiO2 surfaces can produce the activation of platelets and the exocytosis of the alpha and dense granules [11,15]. A clinical derivative of these principles has been exploited via the application of plasma rich in growth factors (PRGF) to wound healing and tissue regeneration [18-20]. PRGF is a type of platelet-rich plasma obtained from patient’s own blood that is devoid of leukocytes and erythrocytes. Upon combination with calcium, PRGF forms a threedimensional fibrin scaffold and releases a pool of biologically active proteins that influence and promote a range of biological process, including cell proliferation, migration and differentiation, and protection against microbial contamination [21-25]. PRGF technology provides different therapeutic formulations including a growth factorenriched liquid, a biomimetic scaffold and an autologous fibrin membrane [20]. In this work we have evaluated a novel implant surface modification consisting on calcium ions (Ca2+ surface) that are released to the implantation site to accelerate the clotting process at the bone-implant interface. The establishment of a surface-bound provisional matrix is intended to configure a scaffold for the recruitment and differentiation of osteogenic cells. Osseointegration has been evaluated in vivo in a rabbit model. We have additionally used plasma rich in growth factors (PRGF), first, to examine the effects of the modified surfaces on platelet activation and fibrin polymerization and second, to test the bone forming ability of the PRGF biomimetic

 

4

 

approach in combination with this new implant surface.

 

5

 

Materials and Methods Unless otherwise stated, all reagents were purchased from Scharlab S.L., Barcelona, Spain. Nanopure water used in this study was obtained by purification with a Milli-Q Direct water system (Millipore, Madrid, Spain). Preparation of rough surfaces: rough surfaces were prepared on three different geometries machined from CP titanium grade IV: screw shaped implants 5 mm diameter and 10 mm height, tubes 6 mm outer and 5 mm inner diameter and 8 mm height and discs 12.7 mm diameter and 1 mm height. All these substrates were subjected to the same proprietary process of acid etching, cleaning and conditioning in a clean room class A (BTI Biotechnology Institute S.L., Vitoria, Spain). When no further chemical modification was done (Control), samples were ß-ray sterilized and stored until use. The roughness was measured by optical profilometry (3D Sensofar Plµ, Terrasa, Spain). We employed a Gaussian filter of 50 x 50 µm of cut-off on the primary surface to split the information between roughness and waviness and selected the parameters of the primary and the roughness surfaces Sq (root mean square value of the ordinate values), Sdr (ratio of the increment of the interfacial area of the surface within the definition area), and Vvc (core void volume of the surface) as representative. The areas scanned were 249 x 187 µm. The roughness of the tubes was measured in their external facet, assuming that the internal part had the same roughness. Calcium-ion modified surfaces (Ca2+ surface) were prepared by sonication during 30 s in a bath containing 5 wt.% CaCl2 in a clean room class A and before sterilization. NaCl modified surfaces (NaCl surface) were prepared by immersion and storage in a NaCl 0.9 wt.% solution under clean room class A conditions before sterilization and until use.

 

6

 

Preparation of smooth surfaces: smooth surfaces were prepared by coating 25 mm round microscope coverslips (Menzel-Gläser, Braunschweig, Germany) with a ~ 20 nm thick layer of TiO2 by magnetron reactive sputtering according to the procedure described in [26]. Prior to experimentation, surfaces were cleaned by sonication in 2% sodium dodecyl sulphate solution (Fluka Analytical, Sigma-Aldrich GmbH, Germany) that was filtered through 0.2 mm pore diameter syringe filter (Millipore, Madrid, Spain) and rinsed under stream of Nanopure water. Water was eliminated with a filtered nitrogen stream. Dry surfaces were further treated with UV-Ozone for 30 min in a UV/Ozone cleaner (BioForce Nanosciences, USA) that was pre-heated for 30 minutes immediately before use (Control). Calcium-ion modified surfaces (Ca2+ surface) were additionally coated with 10 µl 5 wt% CaCl2 solution before use. Calcium quantification and release kinetics: we used EDTA complexometric titration with a murexide dye indicator to measure the amount of calcium ions released over time from calcium-modified substrates. NaOH 2M was added to keep the pH values between 12 and 13. For each repetition, we immersed 25 implants in 50 ml of nanopure water at 37 ± 0.5 ºC under a relative humidity of 25 ± 2 %. Then we transferred the samples to a new 50 ml nanopure water solution for the next measurement and repeated this procedure at the different evaluation times until no more calcium release could be detected. We registered at the end of each time interval the total amount of calcium released. To model the release mechanics we applied a linear fit regression y = A + B·x using Origin V7.5 (OriginLab Corporation, Northampton, MA, USA). Contact angle measurements: we used rough discs unmodified (Control), modified with calcium ions (Ca2+ surface) or with NaCl (NaCl surface) to evaluate surface hydrophilicity as measured by the angle formed by water on the surface with an optical tensiometer Theta T200 (KSV Attension, Helsinki, Finland). We used 3 discs per

 

7

 

surface type and registered the left and right contact angles of 3 drops per disc, which made a total of 18 contact angles per surface type. Scanning electron microscopy and microanalysis: we used a Scanning Electron Microscope (SEM) equipped with an Energy Dispersive X-Ray (EDX) Spectroscopy unit

(JEOL JSM-6490LV, Tokio, Japan) to sample morphology and analysis the

surface chemistry of rough discs either unmodified (Control) or modified calcium ions (Ca2+ surface). The samples were scanned at 20 kV acceleration voltage and 11 mm working distance. For imaging the PRGF coatings, the samples were fixed for 1 h in 2 wt. % glutaraldehyde in 0.1 M sodium cacodylate buffer (pH = 7.4) at room temperature, washed 3 x 10 min with the 6.5 wt. % sucrose in the same cacodylate buffer, stained with 1 wt. % OsO4 in 0.1 M cacodylate buffer for 1 h at 4 ºC in the dark, and finally washed 3 x 10 min with cacodylate buffer. Fixed samples were dehydrated in a series of solutions of increasing ethanol concentrations (30, 50, 70, 96, 3 x 100 vol. %).

Each

step

took

10

min.

Dehydrated

samples

were

immersed

in

hexamethyldisilazane for 2 x 10 min and allowed to dry. Dry samples were coated with gold by sputtering for 180 s immediately before observation in argon atmosphere in a JFC-1000 ion sputter (Jeol Ltd., Japan). Blood collection and PRGF preparation: PRGF was prepared according to the Endoret Dental protocol by BTI (BTI Biotechnology Institute S.L., Vitoria, Spain), which is described in detail elsewhere [27]. Briefly: blood was collected into 3.8 % (wt./vol.) sodium citrate containing tubes (BTI Biotechnology Institute S.L., Vitoria, Spain). Samples were centrifuged at 580 g for 8 min in a PRGF-Endoret System IV centrifuge (BTI Biotechnology Institute S.L., Vitoria, Spain) to obtain 2 to 3 times concentrated platelet-rich plasma (Table 2). Blood was obtained from healthy volunteers with consent for thrombogenicity studies and from the jugular vein of the

 

8

 

rabbits for the in vivo study of osseointegration. Blood cells were counted with two Micros60 hematologic counters (Horiba, Japan), the models ABX and ABC were used to count human and animal blood cells respectively. The most relevant hematologic data is listed in Table 2. UV-Vis coagulation kinetics: tubes unmodified (Control), modified with calcium ions (Ca2+ surface) or NaCl (NaCl surface) were filled with 150 µl of PRGF of the same donor in order to test the coagulation kinetics. A microplate reader IEMS MF Type 1401 (Thermo Labsystems, Madrid, Spain) registered the changes over time in the absorbance at 452 nm through the open sides of the tubes. The start of the experiment was set upon contact of the PRGF with the test tubes. Unlike the negative control and the other surfaces tested, PRGF in the positive control was previously activated with 15 µl CaCl2 5 wt. % solution, immediately before filling the tubes and starting the experiment. Three tubes per experimental surface were employed. Confocal fluorescence microscopy evaluation of platelet activation: we employed a scanning laser confocal fluorescence microscope Zeiss LSM equipped with a 63× oil immersion objective. Unmodified (Control) and modified (Ca2+ surface) TiO2-sputtered glass coverslips were placed in a Teflon holder and incubated with 100 µl plasma solution for 10 min at 37 ºC before washing and examination. Immunostaining for quiescent and activated platelets was done by incubating the plasma samples 1 h before coating with fluorescently labelled mouse-anti-human conjugated antibodies antiCD42b (FITC, emission: 519 nm) and anti-CD62P (PE, emission: 578 nm), respectively (BD Biosciences, Franklin Lakes, NJ, USA). CD42b stains the glycoprotein GPIb present in all platelet membrane surfaces, irrespective of their activity, and CD62p or PSelectin, translocates from the granules to the membrane surface upon activation. Therefore, the signalling CD42b is related to the presence of platelets and that of P-

 

9

 

Selectin, to their activation. Antibodies were diluted 1:25 prior to use. Two laser lines at 488 nm (13 % power) and 561 nm (15 % power) excited the fluorochromes. Emission was filtered with band-pass filters 505-550 nm (FITC) and 575-615 nm (PE). Three samples per experimental surface were analysed. We used autoflurescence conditions to visualize the formation of fibrin fibres and platelet aggregates at the coverslips’ surfaces over time. Autofluorescence in tissues is associated to the cross-links formed between structural proteins, as detected in fibrin, collagen and elastin [15,28]. Briefly, the laser line used excited unstained samples at 633 nm (11 % power). A low-pass filter at 575 nm was employed and the pinhole, master gain, contrast and brightness adjusted to avoid artefacts as much as possible. We recorded a 1h time-lapse series and took captions at 5, 15 and 25 min from contact of the plasma solution with the coverslips’ surface. Two samples per experimental surface were analysed. In vivo implantation testing in a rabbit femoral condyle model: 36 screw-shaped 5 x 10 mm implants were inserted bilaterally in the medial femoral condyle of 18 New Zeeland White female rabbits. Six implants (n=6) were used per surface type (Control, Ca2+ surface, Ca2+ PRGF) and time of evaluation (2 and 8 weeks). All experiments were carried out in conditions that minimized pain or discomfort and under Spanish laws and regulations. The research ethics committee of the Universidad Politécnica de Valencia (Spain) authorized the protocols and facilities used (No. 1656), in accordance with the European Directive (86/609/EEC). The rabbits were skeletally mature, aged 22 ± 2 weeks and weighing 3.8 ± 0.2 kg. Following sedation and anaesthesia of each animal, a preoperative antibiotic was administered.

The incision was made through the skin, the muscular fascia and

Sartorius muscle, exposing the superior distal quadrant of the medial condyle, which

 

10

 

corresponded to the zone of implantation. Perforation was limited in the inferior part by the medial collateral ligament. To prepare the implant sites, drills of growing diameter (2, 3.5 and 5.5 mm) were employed under thorough saline irrigation (Fig. 1a). The drill stop was set at 10 mm. The implant was placed by press fit into the implant site, which was previously clean from drilling remnants (Fig. 1b). The differences in diameters between the last drill and the implants left a gap of around 250 μm from the host bone to the implant surface (Fig. 1d). When PRGF was employed, the implant site was filled with a non-activated liquid PRGF fraction just before implant insertion. Upon insertion, PRGF rose by capillarity along the implant length, covering the surface completely. PRGF was obtained from the jugular vein of each of the rabbits in which the PRGF protocol was employed, in order to respect the autologous principle of the technique. Tissues were sutured in layers (Fig. 1c). Post surgery, the rabbits received analgesia (Meloxican, 0.2 mg/kg - subcutaneous) and anti-inflammatories. This routine was kept for 3 days and the animals were evaluated daily. Weight, behaviour and health conditions were evaluated during the whole study period. After 2 and 8 weeks of implantation, the animals were euthanized. Histological staining and histomorphometric analysis: after sacrifice, the condyles were harvested and immediately fixed in 4 % formaldehyde, dehydrated in graded series of ethanol and embedded in a light-curing acrylic resin (Technovit 7200 VLC, HeraeusKulzer, Wehrheim, Germany) according to the manufacturer´s instructions. Following polymerization, the blocks were cut to a thickness of 300 μm and polished to their final thickness. At least three non-decalcified 20 µm-thick sections of the implants following their longitudinal axis were obtained using a diamond microtome saw (Exakt

 

11

 

Technologies, Oklahoma City, USA). Sections were stained with Harris haematoxylin and Wheatley’s trichromatic stain. The examination at different magnifications with a Leica DMLB light microscope (Leica Microsystems, Wetzlar, Germany) coupled to a Leica DFC300FX digital camera, allowed identifying the different structures and the degree of calcification (Fig. 1e,f). Histomorphometrical analysis was performed in a blinded fashion in order to quantify the bone response and osseointegration around the implants. A 2.5x objective was the magnification chosen for the measurements. The digitalized images were analysed using the software ImageJ (version 1.47, National Institutes of Health, Bethesda, MD, USA). Bone to Implant Contact (BIC) and Bone Volume Density (BVD) percentages were measured. BIC is defined as the contour of direct bone-implant contact without interposition of fibrous tissue. BVD is defined as the area between threads occupied by bone tissue. Quantitative measurements were conducted in at least two different sections per sample (condyle). The entire implant contour was used for the measurements. Statistical analyses: the data is expressed as means ± Standard Deviation or Standard Error of the Mean (histomorphometries). Differences between the means were determined by one-way analysis of variance (ANOVA) after confirmation of data normality (Shapiro-Wilk) using Originv7.5. Statistical significance was accepted for p