Gelatin functionalised porous titanium alloy implants for orthopaedic applications E. Vanderleyden, S. Van Bael, Y.C. Chai, J.-P. Kruth, J. Schrooten, P. Dubruel PII: DOI: Reference:
S0928-4931(14)00324-5 doi: 10.1016/j.msec.2014.05.048 MSC 4672
To appear in:
Materials Science & Engineering C
Received date: Revised date: Accepted date:
30 November 2013 25 April 2014 23 May 2014
Please cite this article as: E. Vanderleyden, S. Van Bael, Y.C. Chai, J.-P. Kruth, J. Schrooten, P. Dubruel, Gelatin functionalised porous titanium alloy implants for orthopaedic applications, Materials Science & Engineering C (2014), doi: 10.1016/j.msec.2014.05.048
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.
ACCEPTED MANUSCRIPT Gelatin functionalised porous titanium alloy implants for orthopaedic applications
PT
E. Vanderleydena, S. Van Baelb,c,d, Y.C. Chaib,e, J.-P. Kruthc, J. Schrootenb,f,
SC
RI
and P. Dubruela,*
a Polymer Chemistry & Biomaterials Research group, Department of Organic Chemistry, University of Ghent, Krijgslaan 281 S4, 9000 Ghent, Belgium
NU
b Prometheus, Division of Skeletal Tissue Engineering, Katholieke Universiteit Leuven, O&N 1, Herestraat 49, box 813, 3000 Leuven, Belgium
MA
c Department of Mechanical Engineering, Division of Production Engineering, Machine Design and Automation, Katholieke Universiteit Leuven, Celestijnenlaan 300b, 3001 Leuven, Belgium
D
d Department of Mechanical Engineering, Division of Biomechanics and Engineering Design, Katholieke Universiteit Leuven, Celestijnenlaan 300c, box 2419, 3001 Heverlee, Belgium
TE
e Tissue Engineering Laboratory, Skeletal Biology and Engineering Research Center, Katholieke Universiteit Leuven, O&N 1, Herestraat 49, box 813, 3000 Leuven, Belgium
AC CE P
f Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, Kasteelpark Arenberg 44, bus 2450, 3001 Leuven, Belgium
Corresponding author *P. Dubruel
Polymer Chemistry & Biomaterials Research group, Department of Organic Chemistry, University of Ghent, Krijgslaan 281 S4, 9000 Ghent, Belgium. Tel: 0032-9-2644466 Fax: 0032-9-2644972 E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
PT
Abstract
RI
In the present work, we studied the immobilisation of the biopolymer gelatin onto the surface of
SC
three dimensional (3D) regular Ti6Al4V porous implants to improve their surface bio-activity. The successful immobilisation of the gelatin coating was made possible by a polydopamine
NU
interlayer, a polymer coating inspired by the adhesive nature of mussels. The presence of both
MA
coatings was first optimised on two dimensional titanium (2D Ti) substrates and confirmed by different techniques including X-ray photelectron spectroscopy, contact angle measurements,
D
atomic force microscopy and fluorescence microscopy. Results showed homogeneous coatings
TE
that are stable for at least 24h in phosphate buffer at 37°C. In a next step, the coating procedure was successfully transferred to 3D Ti6Al4V porous implants, which indicates the versatility of
AC CE P
the applied coating procedure with regard to complex surface morphologies. Furthermore, the bio-activity of these stable gelatin coatings was enhanced by applying a third and final coating using the cell-attractive protein fibronectin. The reproducible immobilisation process allowed for a controlled biomolecule presentation to the surrounding tissue. This newly developed coating procedure outperformed the previously reported silanisation procedure for immobilising gelatin. In vitro cell adhesion and culture studies with human periosteum-derived cells showed that the investigated coatings did not compromise the biocompatible nature of Ti6Al4V porous implants, but no distinct biological differences between the coatings were found
Keywords: Ti6Al4V, porous implants, gelatin, in vitro, bone tissue engineering
2
ACCEPTED MANUSCRIPT
RI
PT
1. Introduction
In the field of biomedical research, titanium (Ti) and its alloys are currently still the metals of
SC
choice for orthopaedic implants. They owe this to their appropriate mechanical properties,
NU
excellent corrosion resistance and superior biocompatibility. The advantageous surface properties result from the chemical stability and structure of the surface oxide layer. Despite the
MA
benefits of this oxide layer, the bone binding capacity or the bio-activity of Ti is not sufficient to realise a chemical bond between the implant and the bone tissue. Ti and bone tissue are usually
TE
D
separated by a thin (ca. 10 nm), non-mineralised protein layer.[1] The bond associated with osseointegration is attributed to a mechanical interlocking between the Ti surface
AC CE P
roughness/pores and the newly formed bone tissue.[2] Nevertheless, surface modification of Ti can improve cell adhesion, implant fixation (decrease of micro-movement) and reduce fibrous tissue formation. In this manner, the osseointegration is thus enhanced and bone repair is accelerated. In fact, a higher degree of osseointegration leads to an improved mechanical stability and a reduced risk of implant loosening. Additionally, in orthopaedics, there is an evolution towards personalised porous implants with complex inner structures to adjust for patient specificity and improve osseointegration.[3-4] The combination of a patient specific implant with a high controllable inner structure and tailored surface characteristics can eventually reduce hospitalisation time/costs and improve the quality of life of the patients.[5]
3
ACCEPTED MANUSCRIPT Over the years, researchers have focussed on finding the most suitable surface characteristics to provoke optimal cell and tissue response in order to obtain such an implant with improved
PT
clinical outcomes.
RI
Different methods to improve the initial interaction of the Ti surface with the surrounding bone tissue are being explored worldwide, including the deposition of calcium phosphates[6-10], the
SC
adsorption of proteins like collagen[11-14] and fibronectin[15-18], the attachment of peptides by
NU
the use of self-assembled monolayers (SAM’s) and polymer coatings.[19-22] Binding of molecules to the surface can be physical (through adsorption or incorporation into the top surface
MA
layer), or chemical (through covalent bonds).[1, 23] In the latter case, most of the strategies are based on the introduction of reactive groups onto the Ti surface, such as by silanisation
TE
D
reactions.[24-27]
In a previous study from our research group, gelatin was immobilised onto the Ti surface using
AC CE P
a silanisation reaction.[28] Gelatin is a water soluble, biodegradable protein, derived from collagen, one of the main constituents of the extra cellular matrix, by partial hydrolysis (acidic or alkaline).[29] Due to its cell attractive nature (presence of RGD sequences) gelatin can bioactivate the surface of a Ti implant. Despite the promising results of our previous study, the disadvantages of a silanisation reaction, like the dependence on the number of hydroxyl groups on the Ti surface and the hydrolytical instability of the resulting siloxane layer, urged the search for an alternative immobilisation strategy. In our work, we selected polydopamine as prime layer for the subsequent deposition of gelatin. The idea is based on the composition of mussel adhesive proteins which are known to attach to virtually any surface. Lee et al. identified dopamine [2(3,4-dihydroxyfenyl)ethylamine] as the smallest structural imitation component of these adhesive proteins. Furthermore, dopamine polymerises spontaneously onto any surface (noble metals,
4
ACCEPTED MANUSCRIPT oxides, ceramics, polymers and semi-conductors) into a thin polydopamine film, which can act as a prime layer for depositing functional coatings.[30] Recent insights in the structure of this
PT
polydopamine are still not sufficient to account for its properties. Opposed to the assumed ‘open-
RI
chain polycatechol/quinine model’, on one hand, and the ‘eumelanin model’, on the other hand, Dreyer et al. suggested that polydopamine is not a covalent polymer but a supramolecular
SC
aggregate of monomers.[31] Simultaneously, Hong et al. reported on a physical, self-assembled
NU
trimer entrapped within polydopamine.[32] Even more recent, Della Vecchia et al. reveiled a three-component structure of polydopamine. Nevertheless, they suggest representing
MA
polydopamine as a collection of oligomeric species in which monomer units are linked through different bonding. The coexistence of such structurally diverse components accounts for a unique
TE
D
blend of eumelanin-like and amine containing polycatechol functionalities.[33] Despite the so far unknown structure of polydopamine, the technique has previously been successfully applied on
AC CE P
Ti and Ti alloy surfaces to immobilise biomolecules, such as dextrane, chitosan, hyaluronic acid, heparin, gelatin and growth factors.[34-40] Besides the advantageous characteristics of being a water-based and non-toxic coating strategy, it also presents a straightforward and versatile method to coat a wide variety of surfaces, ranging from the one of a 2D substrate to the one of complex 3D porous implants. Hence, its potential for future use on implants is secured as these final implants can have very customised morphologies which at this moment is a limitation for the present line of sight coating techniques.
In the present paper we describe the application of a stable gelatin coating onto a Ti6Al4V surface by applying an intermediate polydopamine coating. In an attempt to increase the biofunctionality of the surface even more, a final coating of the cell adhesive protein fibronectin
5
ACCEPTED MANUSCRIPT was applied onto the gelatin. Due to the natural binding affinity between fibronectin and gelatin a natural presentation of the immobilised fibronectin on the surface and, hence, a retention of its
PT
cell adhesive activity, is anticipated.[41] The coatings were optimised on two dimensional (2D) Ti
RI
substrates and characterised via static contact angle (SCA) measurements, X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM) and fluorescence microscopy.
SC
In the second part of this study, the optimised gelatin coating technology was applied onto
NU
three dimensional (3D) porous Ti6Al4V implants (height: 3 mm diameter: 6 mm) which were fabricated via additive manufacturing (AM) technology, i.e. selective laser melting (SLM).[42-44]
MA
3D micro-computed tomography (Micro-CT) imaging was used to characterise the average pore and strut size, porosity and surface area of the as-produced porous implants. The as-produced
TE
D
Ti6Al4V porous implants were divided into four categories: (i) uncoated or oxidised (ii) polydopamine coated (Dop) (iii) polydopamine/gelatin coated (Dop-Gel) and (iv)
AC CE P
polydopamine/gelatin/fibronectin coated (Dop-Gel-Fn). Transfer of the coating technology from Ti to Ti6Al4V surfaces was straightforward since the surface composition of the oxidised alloy largely compares to the one of the oxidised, pure Ti surface.[45-46] The effects of the different coatings on in vitro cell seeding efficiency (CSE), proliferation and differentiation of a clinically relevant mesenchymal stem cell-like osteoprogenitors, namely human periosteum-derived cells (hPDCs), were evaluated.
2. Materials and methods
2.1. Materials
6
ACCEPTED MANUSCRIPT Ti foils (0.125 mm thick, 99.6% pure according to the data sheet) were obtained from Chempur (Karlsruhe, Germany). Cylindrical porous Ti6Al4V implants (6 mm diameter x 3 mm
PT
height) were fabricated by an in-house selective laser melting (SLM) machine (figure 1).[44, 47-
RI
49] The average as-produced pore size, strut size, porosity and surface area, as shown in figure 1,
AC CE P
TE
D
MA
NU
SC
were assessed by micro-CT image analysis.
Figure 1. (A) Cylindrical Ti6Al4V porous implants made by selective laser melting (6 mm diameter x 3 mm height). (B) Morphological parameters of the as-produced Ti6Al4V porous implant assessed by micro-CT image analysis. Mean (± standard deviation)
Concentrated hydrochloric acid (HCl; 37%) was provided by Panreac Quimica S.A (Barcelona, Spain), hydrogen peroxide (H2O2; 30 %) and dimethylformamide (DMF) by Aldrich (Bornem, Belgium) and ammonium hydroxide (NH4OH; 25%) by Acros (Geel, Belgium). HPLC grade acetone, HPLC grade pentane and 2-(3,4-dihydroxyphenyl)ethylamine hydrochloride (dopamine hydrochloride) were all purchased from Sigma (Bornem, Belgium). Cyclohexane was obtained from Fiers (Kuurne, Belgium), it was dried with calcium hydride and distilled before use. Gelatin type B, isolated from bovine skins after an alkaline pre-treatment, was a kind gift
7
ACCEPTED MANUSCRIPT from Rousselot (Ghent, Belgium). Gelatin with an isoelectric point of ca. 5, gel strength of 257 Bloom and viscosity (6.67%, 60°C) of 4.88 mPa.s was used. Oregon Green® 488, carboxylic
PT
acid, succinimidyl ester was purchased from N.V. Invitrogen SA (Ghent, Belgium). Fibronectin
RI
from bovine plasma (0.1 % solution, 1 mg/ml in 0.5 M NaCl, 0.05 M Tris-HCl and pH = 7.5) was obtained from Sigma-Aldrich (Bornem, Belgium). The water used throughout this study was
NU
SC
milliQ (double distilled, > 18 MegaOhm/cm).
MA
2.1. Methods
D
2.2.1. Pre-treatment of Ti and Ti6Al4V
TE
Cleaning of the metal substrates and porous implants consisted of sequenced ultrasonic
AC CE P
treatments in (1) cyclohexane (10 minutes), (2) 10 N HCl (30 minutes) and (3) milliQ water (30 minutes). After cleaning, the samples were oxidised for 20 minutes in a 1/1/5 NH4OH/H2O2/H2O-mixture. These steps are described in more detail in a previous paper.[28]
2.2.2. Deposition of polydopamine
The method applied is based on a previous study.[30] Briefly, oxidised metal samples were incubated overnight in a solution of dopamine (2 mg/ml) in Tris-HCl (10 mM, pH = 8.5). Afterwards, the modified samples were rinsed with milliQ water, acetone and dried with pressurised air.
2.2.3. Immobilisation of gelatin and fibronectin coatings
8
ACCEPTED MANUSCRIPT
Polydopamine coated Ti samples were incubated overnight in a 0.5, 1, or 3 w/v% solution of
PT
gelatin in 10 mM Tris-HCl (pH = 8) at 40°C. The gelatin modified samples were rinsed with milliQ water, acetone and dried with pressurised air.
RI
Fluorescent labelled gelatin was obtained by dissolving 15 mg gelatin (5.25 10-3 mmol Ɛ-
SC
amine side groups of lysine and hydroxylysine) in 7.5 ml 0.1 M NaHCO3 buffer. 2 mg of the
NU
succinimidyl ester of Oregon Green® (3.93 10-3 mmol), dissolved in 200 µl DMF, was added and the mixture was stirred for 90 min at room temperature. Purification was done by using PD-
MA
10 columns. In a next step, polydopamine modified Ti samples were incubated in a 2 ml solution of the fluorescent labelled gelatin (1.4 mg/ml). Labelling of fibronectin occurred analogously to
TE
D
the gelatin. A fibronectin solution (700 µl, 1 mg/ml) was diluted with 1.5 ml 0.1 M NaHCO3 buffer and 3 mg of the succinimidyl ester of Oregon Green®, dissolved in 300 µl DMF, was
AC CE P
added. After 1 h reaction at room temperature, the mixture was purified by means of a PD-10 column. Subsequently, (non labelled) gelatin modified Ti samples were dipcoated with the fluorescent labelled fibronectin (200 µg/ml) and rinsed by incubation in a 0.1 M NaHCO3 buffer for 2 h. The buffer was refreshed after 1 h. In case of the 3D porous implants, a gelatin concentration of 1 w/v% was applied. Deposition of fibronectin on the gelatin modified surfaces was realized by a dipcoating step using a fibronectin solution of 0.01 w/v%. After coating, the porous implants were sterilised with ethylene oxide gas (cold cycle) to avoid coating degradation due to the heat induced during sterilisation.
2.2.4. Stability study
9
ACCEPTED MANUSCRIPT
Modified Ti samples were incubated in phosphate buffer (50 mM, pH = 7.4) at 37 °C for
PT
different incubation times. After incubation, the samples were gently rinsed with water and
RI
acetone and dried with pressurised air and analyzed as described below.
NU
SC
2.2.5. Surface analysis
MA
2.2.5.1. X-ray photo-electron spectroscopy (XPS)
The chemical composition of the different metal surfaces was determined using ‘‘FISONS S-
TE
D
PROBE’’, a dedicated XPS (X-ray photoelectron spectroscopy) instrument designed to give high-end analysis performance, while providing a high sample throughput. The fine focus Al-Ka
AC CE P
source with a quartz monochromator, developed by Fisons Instruments Surface Science ensures lower background and higher sensitivity than conventional twin anode sources. All measurements were performed in a vacuum of at least 10-9 Pa. Wide and narrow-scan spectra were acquired at pass energy of 158 and 56 eV, respectively. The binding energy was calibrated by the C 1s peak at 285 eV. The spot size used was 250 µm on 1 mm. Data analysis was performed using S-PROBE software. The measured spectrum was displayed as a plot of the number of electrons (electron counts) versus electron binding energy in a fixed, small energy interval. Peak area and peak height sensitivity factors were used for the quantifications. All surface compositions reported in this work are expressed as atomic percentages (at%).
2.2.5.2. Static Contact Angle (SCA)
10
ACCEPTED MANUSCRIPT
Static contact angle measurements on the 2D Ti surfaces were performed using an optical
PT
contact angle measuring system, OCA 20, from Dataphysics (distributed by Benelux Scientific)
RI
equipped with a 500 µl Hamilton syringe. For each measurement, 1 µl of milli-Q water was placed on the surface. The droplet was imaged using a video camera. The contact angle was
SC
determined on the screen of the monitor using the imaging software provided by the supplier
NU
(SCA 20, version 2.1.5 build 16). Three measurements were made on each sample. Static contact angle measurements could not be performed on the porous Ti6Al4V implants,
MA
due to the immediate penetration of the water drop into the porous implants.
TE
D
2.2.5.3. Atomic Force Microscopy (AFM)
AC CE P
The AFM study was performed with a Nanoscope IIIa Multimode (Digital Instruments, Santa Barbara, California, USA) in ‘tapping mode’ and exposed to air (n = 1). The obtained images were analysed with Nanoscope Image III software. Changes in surface morphology are quantified by root mean square (RMS) roughness values (Rq). Rq is the RMS deviation of the profile and is defined as follows:
(Z i Z ave) 2 Rq N where Zave is the average Z height value within a given area, Zi the current Z value and N is the number of points within the given area.
2.2.5.4. Fluorescence microscopy
11
ACCEPTED MANUSCRIPT Fluorescent microscopy images were recorded with an Olympus IX81 to visualise the presence of the gelatin and fibronectin coating. The chosen filters were an excitation filter with a band
PT
width of 460 – 490 nm and an emission filter with LP510 long path characteristics. Since this
RI
technique is limited to visualise only the surface of the porous implants and not the internal
SC
structure, this technique was only used for the 2D Ti surfaces.
NU
2.2.6. 3D cell seeding and in vitro cell proliferation and differentiation analysis
MA
For the in-vitro cell tests human periosteum-derived cells (hPDCs) were chosen since they play a major role in bone fracture repair[50] and have a higher in-vitro expansion rate compared to
D
osteoblasts.[51] hPDCs (pooled from periosteal biopsies of six donors, age 14.9 + 2.1) were
TE
expanded from the liquid nitrogen cell bank using Dulbecco’s modified eagle’s medium
AC CE P
(DMEM+GlutaMaxTM-1, Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% antibiotics and 1% sodium pyruvate. Upon confluence, cells were harvested by trypsinization and a cell suspension (60 µl) containing 100 000 hPDCs was drop seeded onto each porous implant, incubated statically for 1 h at 37 oC to facilitate cell attachment, before being transferred to a rotator to perform dynamic rotation seeding.[52] After overnight dynamic seeding, the cellseeded porous implants were transferred to a 48-well plate and cultured in growth medium (GM) for 14 days. The medium was refreshed three times a week. Cell seeding efficiency (CSE) was calculated by determining the ratio of the DNA content of the cells successfully seeded on the porous implant to the initially cell seeding number. This, according to Impens et al.[53], results in the most realistic CSE values compared to other calculation methods.
12
ACCEPTED MANUSCRIPT At 1, 7 and 14 days, the proliferation of hPDCs on different coatings was assessed by nondestructive analysis of the metabolic activity using AlamarBlue® (Invitrogen, 10% in GM), and
PT
also by quantifying the DNA content of each sample at defined time points using Quant-iTTM dsDNA measurement kit (Invitrogen, USA). Cell viability on different coatings was characterised
RI
by staining the living and dead cells using the Live/Dead® cell viability kit (Invitrogen).
SC
Osteogenic differentiation of hPDCs was evaluated by the measurement of alkaline phosphatase
NU
(ALP) levels using a BluePhos® microwell phosphatase substrate system (Kirkegaard & Perry Laboratories, USA). The measured absorbance was normalised to the DNA content of the
MA
respective samples.
TE
D
2.2.7. Statistical analysis.
AC CE P
Statistical analysis was performed on the obtained XPS data, using the student t-test. Two data sets (n = 3) were considered to be significantly different when p < 0.05. One-way unpaired ANOVA (n= 3) was used to analyze statistical differences between the cell seeding results. Two-way unpaired ANOVA (n= 3) was used to analyze statistical differences between the DNA, ALP and metabolic activity results. The Tukey test was used to perform post hoc comparisons. All error bars represent 1 σ standard deviation. Two data sets were considered to be significantly different when p < 0.01.
3. Results & Discussion
13
ACCEPTED MANUSCRIPT It is generally accepted that osseointegration requires biofunctionalisation of the implant surface. In this study, we opted to apply the extra cellular matrix biopolymers gelatin and
PT
fibronectin as coatings on Ti6Al4V porous implants. The coating technology was optimised
RI
using 2D Ti substrates and successfully transferred to 3D porous Ti6Al4V implants. Gelatin was immobilised by applying a prime layer on the Ti surface using dopamine. Based
SC
on the research of Lee et al. a polydopamine coating was deposited onto Ti by incubating
NU
oxidised Ti samples overnight in a solution of dopamine in Tris-HCl buffer (pH = 8.5) [30]. The stability of the polydopamine layer was determined by incubating the modified Ti sample in
MA
phosphate buffer at 37 °C for 24 h. As a control, an oxidised sample was incubated in Tris-HCl buffer overnight. The Ti surfaces were characterised by XPS and SCA measurements. For
TE
D
comparison purposes, the theoretical and the measured values of the atomic composition of dopamine and the measured atomic composition of an oxidised surface are also presented in
AC CE P
table 1.
Table 1. Atomic surface compositions and contact angle of dopamine and a polydopamine coated Ti surface before and after 24 h incubation in phosphate buffer at 37 °C (n=3), as determined by XPS and SCA. An oxidised Ti sample, incubated for 24 h in Tris-HCl buffer, served as a control. The data of dopamine (theoretical and measured) and of an oxidised Ti surface served as references. Values with a standard deviation are an average of three measurements. %Ti
%C
%O
%N
N/C
Contact angle
Dopamine (theoretical)
-
73
18
9
0.12
-
Dopamine-HCl (measured)
-
68
20
6
0.09
-
Polydopamine on Ti
-
77 ( 1.2)
18 ( 0.6)
6 ( 1.5)
0.07 (± 0.015)
22° ( 3.0)
Polydopamine on Ti after 24 h incubation in phosphate buffer
-
75 ( 1.2)
18 ( 0.6)
7 ( 0.6)
0.09 (± 0.009)
-
14
ACCEPTED MANUSCRIPT 6
54
40
-
-
-
Oxidised Ti
13
38
49
-
-
12° ( 1.9)
PT
Control
The XPS results indicate that the surface composition of the Ti surface, after modification with
RI
dopamine, is in very good agreement with the atomic composition of the dopamine monomer
SC
which confirms the presence of a polydopamine layer on the surface. Measurements on different
NU
positions on the same sample confirmed that the layer fully covered the Ti surface (data not shown). Furthermore, it can be concluded that the polydopamine layer was stable for at least 24 h
MA
in phosphate buffer. The control, however, showed a higher C content after incubation in TrisHCl buffer. This is probably due the adsorption of contaminants originating from the buffer.
D
In a next step, gelatin was immobilised onto the polydopamine coated Ti surface by incubating
TE
the sample overnight at 40°C in a 0.5, 1 or 3 w/v% solution of gelatin in Tris-HCl buffer
AC CE P
(pH = 8). According to Lee et al., the polydopamine catechols can react with thiols and amines via Michael addition or Schiff base reactions.[30] Gelatin possesses amine functionalities and, thus, should be able to covalently bind to the polydopamine and lead to a stable surface coating. Based on the results in table 2 there appears to be a small change in atomic surface composition in comparison to the one of a polydopamine modified surface (table 1). For each applied gelatin concentration there is a decrease in the C content and an increase in the N content resulting in an increase of the N/C ratio. The values are similar to those of gelatin as such, thus confirming the presence of a gelatin coating on the Ti surface. Moreover, the C contents of the polydopamine and gelatin coated surfaces and their respective contact angle values differ significantly (p < 0.05), indicating two different surface compositions or two different coatings.
15
ACCEPTED MANUSCRIPT
%Ti
%C
%O
N/C
Contact angle
0h
-
69 ( 1.0)
19 ( 1.2) 11 ( 1.0)
0.16 ( 0.017)
-
1h
-
69
19
12
0.17
-
12 h
-
70
19
11
0.16
-
24 h
-
69 ( 0.0)
19 ( 0.6) 12 (0.6)
0.17 ( 0.008)
-
0h
-
68 ( 1.5)
20 ( 0.6) 12 ( 1.0)
0.18 ( 0.018)
10° ( 0.7)
1h
-
69
19
12
0.17
-
12 h
-
69
19
12
0.17
-
24 h
-
MA
NU
0.5%
Gelatin
TE -
AC CE P
3%
D
1%
0h
%N
RI
Incubation time
SC
Gelatin Concentration
PT
Table 2. Atomic surface compositions and contact angle of gelatin coated Ti surfaces, after the deposition of a polydopamine coating, as determined by XPS and SCA. Results from before (0 h) and after (1, 12 and 24 h) incubation at 37 °C in phosphate buffer are shown. Values with a standard deviation mentioned are an average of three measurements.
69 ( 1.0)
19 ( 4.9) 11 ( 0.6)
0.16 ( 0.010)
-
68 ( 0.6)
20 ( 0.6) 12 ( 1.0)
0.18 ( 0.018)
-
1h
-
68
19
13
0.19
-
12 h
-
68
20
12
0.18
-
24 h
-
68 ( 0.0)
20 ( 0.6) 12 ( 0.6)
0.18 ( 0.008)
-
-
69
18
0.19
-
13
During the XPS analyses the C 1s peaks of the polydopamine and gelatin coatings were also compared (figure 2). The spectra show C 1s peaks of the different chemical environments. In case of the polydopamine coated surface, peak deconvolution reveals three distinct peaks at 284.8, 286.0 and 288.8 eV originating from respectively C-C/C-H-, C-N/C-O- and C=O-bonds. In case of the gelatin coated surface, peaks at respectively 284.8, 286.0 and 287.9 eV are
16
ACCEPTED MANUSCRIPT detected. In contrast to the carbonyl function of polydopamine, the carbonyl function of gelatin originates (mostly) from amide bonds.
PT
C-H C-C
C-H C-C
C-O
RI
C-O
SC
C-N
C-N
C=O
288 285 Binding Energy (eV)
282
MA
291
NU
C=O
291
288 285 Binding Energy (eV)
282
TE
D
Figure 2. XPS detailed spectra of the C 1s peaks of polydopamine coated Ti (left) and gelatin coated Ti after deposition of a polydopamine layer (right).
AC CE P
In a next step, the stability of the gelatin coating was studied by incubating the samples in phosphate buffer (50 mM, pH = 7.4) at 37 °C during 1, 12 and 24 h, and subsequently determining the surface composition by means of XPS. Stability longer than 24h was not assessed since initial cell attachment already occurs minutes to a few hours after the seeding step. The values obtained at all time points (table 2) were not significantly different from the control, indicating sample stability for at least one day. Nevertheless, the C and N contents are significantly different from polydopamine coated Ti surface. Based on these data, it can be concluded that the applied immobilisation technique, based on a polydopamine prime layer, leads to a stable gelatin coating on Ti. An additional technique to confirm the presence of a gelatin coating is fluorescent microscopy. For this, gelatin was modified with the fluorescent label Oregon Green® (Gel-OG, figure 3)
17
ACCEPTED MANUSCRIPT before immobilisation on the polydopamine coated Ti surface. The obtained fluorescence microscopy images are presented in figure 4. Images B and C show the surface respectively
PT
before and after 2 h of rinsing in buffer. Both images show visually a homogeneously spread gelatin coating. In a next step, the cell adhesive protein fibronectin, and not the gelatin, was
RI
labelled with Oregon Green® (Fn-OG) and subsequently deposited by dipcoating. This enabled
O
O
F
F C OH
D
O
O C O
+
AC CE P
O
NH2
N O C OG
TE
N
≡
O
O
Gel
MA
HO
NU
SC
us to study possible aspecific binding of fibronectin to polydopamine.
N O C OG O
Gel
NH C OG O
GelB-OG
Figure 3. Chemical structure of Oregon Green® and reaction of gelatin with the succinimidyl ester of Oregon Green®.
A homogeneous distribution of Fn-OG was visually observed on the gelatin coated Ti sample (image D), again confirming the presence of gelatin on the surface. In case of the polydopamine coated sample, almost no Fn-OG was detected (image E). This finding does not only confirm the specific interaction between gelatin and fibronectin[41, 54-55] but also the presence of the gelatin layer in images B and C. A control image of non coated cp Ti (image A) indicated that the polydopamine coating does not bind Oregon Green®.
18
AC CE P
TE
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
200 µm
200 µm
Figure 4. Fluorescent microscopy images (magnification 10x) of (A) cp Ti, (B) polydopamine coated Ti incubated in GelB-OG solution, (C) polydopamine coated Ti incubated in GelB-OG solution after rinsing, (D) gelatin coated Ti dipcoated with Fn-OG after rinsing, and (E) polydopamine coated Ti dipcoated with Fn-OG.
In a final step of the material structural characterisation, the possible influence of the dopamine and subsequent gelatin modification on the Ti surface roughness was studied by AFM. It can be concluded from the left image in figure 5 that polydopamine aggregates are formed during the
19
ACCEPTED MANUSCRIPT deposition of the polydopamine layer, strongly affecting the surface roughness. The Rq value
MA
NU
SC
RI
PT
increases from 154 nm for an oxidised surface to 250 nm for a polydopamine modified surface.
TE
D
Figure 5. Surface topography of a polydopamine coated Ti surface (left image) and a gelatin coated Ti surface (right image), as determined by means of AFM.
AC CE P
From the image on the right side in figure 5, it can be concluded that the modification of a polydopamine coated surface with gelatin leads to a smoother surface. This observation is confirmed by the decrease in Rq value to 173 nm.
In the second part of this study, 3D Ti6Al4V porous implants were fabricated using SLM and coated with polydopamine, gelatin and fibronectin by applying the coating procedures optimised on the 2D Ti samples. Comparison of the surface compositions of polydopamine and gelatin coated 2D Ti (table 1 and 2) and 2D Ti6Al4V surfaces (table 3) confirmed the similarity between both, enabling a successful and straightforward transfer of the coating technology from Ti to the Ti6Al4V alloy. For reason of simplicity, the intermediate 1% gelatin solution was chosen for the modification of the 3D Ti6Al4V porous implants and 2D Ti6Al4V surfaces. Furthermore, stability of the coatings up to 2 weeks (to match the in vitro tests) was studied with still a 20
ACCEPTED MANUSCRIPT significant (p < 0.05) difference in C and N content between polydopamine and gelatin coated
PT
Ti6Al4V surfaces (table 3).
N/C
49 ( 0.5)
1 ( 0.2)
2 ( 0.1)
0.02 ( 0.005)
39 ( 1.2)
47 ( 0.6)
1 ( 0.1)
2 ( 1.2)
0.03 ( 0.004)
37 ( 0.5)
50 ( 0.2)
0 ( 0.2)
2 ( 0.3)
0.00 ( 0.006)
73 ( 0.1)
20 ( 0.2)
7 ( 0.3)
-
0.10 ( 0.005)
72 ( 0.2)
21 ( 0.1)
7 ( 0.2)
-
0.09 ( 0.003)
%C
0h
12 ( 0.3)
36 ( 0.4)
24 h
10 ( 0.2)
2 weeks
10 ( 0.2)
0h
-
24 h
1 % Gelatin on Ti6Al4V-Dop
%Ca
%Ti
MA
NU
%O
D
-
AC CE P
Polydopamine on Ti6Al4V
%N
Incubation time
TE
Oxidised Ti6Al4V
SC
RI
Table 3. Atomic surface compositions of oxidised, polydopamine coated and gelatin coated Ti6Al4V surfaces, as determined by XPS. Results from before (0 h) and after (1h, 24 h and 2weeks) incubation at 37 °C in phosphate buffer are shown. Values with a standard deviation mentioned are an average of three measurements.
2 weeks
-
69 ( 0.1)
23 ( 0.5)
7 ( 0.5)
-
0.10 ( 0.008)
0h
-
67 ( 0.2)
21 ( 0.0)
12 (0.2)
-
0.18 ( 0.005)
24 h
-
68 ( 0.5)
20 ( 0.3)
12 ( 0.6)
-
0.17 ( 0.010)
2 weeks
-
68 ( 0.2)
22 ( 0.2)
10 ( 0.1)
-
0.15 ( 0.001)
Using these surfaces, CSE was determined 2 hours after cell seeding. From the results in figure 6, it can be shown that there was no significant effect of the coating on the CSE. Furthermore, the effect of the coatings on hPDCs proliferation and differentiation after 14 days of in vitro cell culture was evaluated in growth medium.
21
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
Figure 6. Effect of coating on the CSE as determined by dynamic rotation seeding method (n=3).
22
AC CE P
TE
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Figure 7. Representative images of live/dead staining of human periosteum derived cells (hPDCs) seeded on Ti6Al4V porous implants with various coatings. The seeded cells were found viable during 14 days of in vitro culture with growth.
23
ACCEPTED MANUSCRIPT Live/dead staining provided qualitative information on the cell behaviour (figure 7). After 1 day of in vitro cell culture a uniform distribution of hPDCs throughout the fibronectin coated
PT
Ti6Al4V porous implants (Dop-Gel-Fn) was noticed. On the contrary, hPDCs were distributed
RI
mostly at the bottom of the gelatin coated (Dop-Gel), the polydopamine coated (Dop) and the oxidised Ti6Al4V porous implants. After 7 days of in vitro cell culture, cell distribution became
SC
uniform for all different coating groups. Furthermore, hPDCs started to occlude the pores by
AC CE P
TE
D
MA
NU
bridging the corners, resulting in circular cell-filled pores (figure 8).
Figure 8. Representative image of hPDCs bridging the corners of Ti6AL4V porous implant pores after 7 days of in vitro culturing.
Eventually, this phenomenon resulted in almost complete cell-filled pores in the top and side surface of the different Ti6Al4V porous implant coating groups after 14 days of in vitro cell culturing. Based on the live dead staining, it can be stated that the difference in surface chemistry, due to the applied coatings, does not influence this circular pore filling behaviour after 14 days in vitro. This is in line to what has been observed by other researchers on polymer, ceramic and metal porous structures, thus different surface chemistries.[56-60] In a next part of the work, a quantitative DNA assay and a metabolic activity analysis were performed (figure 9 A and B). Similar to what has been visually noticed on the live/dead staining 24
ACCEPTED MANUSCRIPT micrographs, hPDCs were viable and able to proliferate during 14 days of in vitro culturing which resulted in significantly higher levels of measured DNA and metabolic activity after 14
PT
days of in vitro cell culturing. However, on day 14, no significant differences (p > 0.01) were found between the different samples.
RI
Finally, ALP, a marker to indicate early stage of osteogenic differentiation, was measured.[61]
SC
During the 14 days of in vitro culturing, no significant increase in ALP was measured for all
NU
tested samples (not shown in figure since values equal zero). These results imply that the
AC CE P
TE
D
MA
investigated coatings did not influence the in vitro differentiation of the hPDCs.
25
ACCEPTED MANUSCRIPT
PT
Figure 9. DNA assay (A) and metabolic activity (B) of cells on the different coated Ti6Al4V porous implants cultured in GM. Error bars represent standard deviation intervals. Statistical analysis: unpaired t-test (** p
S0928-4931(14)00324-5 doi: 10.1016/j.msec.2014.05.048 MSC 4672
To appear in:
Materials Science & Engineering C
Received date: Revised date: Accepted date:
30 November 2013 25 April 2014 23 May 2014
Please cite this article as: E. Vanderleyden, S. Van Bael, Y.C. Chai, J.-P. Kruth, J. Schrooten, P. Dubruel, Gelatin functionalised porous titanium alloy implants for orthopaedic applications, Materials Science & Engineering C (2014), doi: 10.1016/j.msec.2014.05.048
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.
ACCEPTED MANUSCRIPT Gelatin functionalised porous titanium alloy implants for orthopaedic applications
PT
E. Vanderleydena, S. Van Baelb,c,d, Y.C. Chaib,e, J.-P. Kruthc, J. Schrootenb,f,
SC
RI
and P. Dubruela,*
a Polymer Chemistry & Biomaterials Research group, Department of Organic Chemistry, University of Ghent, Krijgslaan 281 S4, 9000 Ghent, Belgium
NU
b Prometheus, Division of Skeletal Tissue Engineering, Katholieke Universiteit Leuven, O&N 1, Herestraat 49, box 813, 3000 Leuven, Belgium
MA
c Department of Mechanical Engineering, Division of Production Engineering, Machine Design and Automation, Katholieke Universiteit Leuven, Celestijnenlaan 300b, 3001 Leuven, Belgium
D
d Department of Mechanical Engineering, Division of Biomechanics and Engineering Design, Katholieke Universiteit Leuven, Celestijnenlaan 300c, box 2419, 3001 Heverlee, Belgium
TE
e Tissue Engineering Laboratory, Skeletal Biology and Engineering Research Center, Katholieke Universiteit Leuven, O&N 1, Herestraat 49, box 813, 3000 Leuven, Belgium
AC CE P
f Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, Kasteelpark Arenberg 44, bus 2450, 3001 Leuven, Belgium
Corresponding author *P. Dubruel
Polymer Chemistry & Biomaterials Research group, Department of Organic Chemistry, University of Ghent, Krijgslaan 281 S4, 9000 Ghent, Belgium. Tel: 0032-9-2644466 Fax: 0032-9-2644972 E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
PT
Abstract
RI
In the present work, we studied the immobilisation of the biopolymer gelatin onto the surface of
SC
three dimensional (3D) regular Ti6Al4V porous implants to improve their surface bio-activity. The successful immobilisation of the gelatin coating was made possible by a polydopamine
NU
interlayer, a polymer coating inspired by the adhesive nature of mussels. The presence of both
MA
coatings was first optimised on two dimensional titanium (2D Ti) substrates and confirmed by different techniques including X-ray photelectron spectroscopy, contact angle measurements,
D
atomic force microscopy and fluorescence microscopy. Results showed homogeneous coatings
TE
that are stable for at least 24h in phosphate buffer at 37°C. In a next step, the coating procedure was successfully transferred to 3D Ti6Al4V porous implants, which indicates the versatility of
AC CE P
the applied coating procedure with regard to complex surface morphologies. Furthermore, the bio-activity of these stable gelatin coatings was enhanced by applying a third and final coating using the cell-attractive protein fibronectin. The reproducible immobilisation process allowed for a controlled biomolecule presentation to the surrounding tissue. This newly developed coating procedure outperformed the previously reported silanisation procedure for immobilising gelatin. In vitro cell adhesion and culture studies with human periosteum-derived cells showed that the investigated coatings did not compromise the biocompatible nature of Ti6Al4V porous implants, but no distinct biological differences between the coatings were found
Keywords: Ti6Al4V, porous implants, gelatin, in vitro, bone tissue engineering
2
ACCEPTED MANUSCRIPT
RI
PT
1. Introduction
In the field of biomedical research, titanium (Ti) and its alloys are currently still the metals of
SC
choice for orthopaedic implants. They owe this to their appropriate mechanical properties,
NU
excellent corrosion resistance and superior biocompatibility. The advantageous surface properties result from the chemical stability and structure of the surface oxide layer. Despite the
MA
benefits of this oxide layer, the bone binding capacity or the bio-activity of Ti is not sufficient to realise a chemical bond between the implant and the bone tissue. Ti and bone tissue are usually
TE
D
separated by a thin (ca. 10 nm), non-mineralised protein layer.[1] The bond associated with osseointegration is attributed to a mechanical interlocking between the Ti surface
AC CE P
roughness/pores and the newly formed bone tissue.[2] Nevertheless, surface modification of Ti can improve cell adhesion, implant fixation (decrease of micro-movement) and reduce fibrous tissue formation. In this manner, the osseointegration is thus enhanced and bone repair is accelerated. In fact, a higher degree of osseointegration leads to an improved mechanical stability and a reduced risk of implant loosening. Additionally, in orthopaedics, there is an evolution towards personalised porous implants with complex inner structures to adjust for patient specificity and improve osseointegration.[3-4] The combination of a patient specific implant with a high controllable inner structure and tailored surface characteristics can eventually reduce hospitalisation time/costs and improve the quality of life of the patients.[5]
3
ACCEPTED MANUSCRIPT Over the years, researchers have focussed on finding the most suitable surface characteristics to provoke optimal cell and tissue response in order to obtain such an implant with improved
PT
clinical outcomes.
RI
Different methods to improve the initial interaction of the Ti surface with the surrounding bone tissue are being explored worldwide, including the deposition of calcium phosphates[6-10], the
SC
adsorption of proteins like collagen[11-14] and fibronectin[15-18], the attachment of peptides by
NU
the use of self-assembled monolayers (SAM’s) and polymer coatings.[19-22] Binding of molecules to the surface can be physical (through adsorption or incorporation into the top surface
MA
layer), or chemical (through covalent bonds).[1, 23] In the latter case, most of the strategies are based on the introduction of reactive groups onto the Ti surface, such as by silanisation
TE
D
reactions.[24-27]
In a previous study from our research group, gelatin was immobilised onto the Ti surface using
AC CE P
a silanisation reaction.[28] Gelatin is a water soluble, biodegradable protein, derived from collagen, one of the main constituents of the extra cellular matrix, by partial hydrolysis (acidic or alkaline).[29] Due to its cell attractive nature (presence of RGD sequences) gelatin can bioactivate the surface of a Ti implant. Despite the promising results of our previous study, the disadvantages of a silanisation reaction, like the dependence on the number of hydroxyl groups on the Ti surface and the hydrolytical instability of the resulting siloxane layer, urged the search for an alternative immobilisation strategy. In our work, we selected polydopamine as prime layer for the subsequent deposition of gelatin. The idea is based on the composition of mussel adhesive proteins which are known to attach to virtually any surface. Lee et al. identified dopamine [2(3,4-dihydroxyfenyl)ethylamine] as the smallest structural imitation component of these adhesive proteins. Furthermore, dopamine polymerises spontaneously onto any surface (noble metals,
4
ACCEPTED MANUSCRIPT oxides, ceramics, polymers and semi-conductors) into a thin polydopamine film, which can act as a prime layer for depositing functional coatings.[30] Recent insights in the structure of this
PT
polydopamine are still not sufficient to account for its properties. Opposed to the assumed ‘open-
RI
chain polycatechol/quinine model’, on one hand, and the ‘eumelanin model’, on the other hand, Dreyer et al. suggested that polydopamine is not a covalent polymer but a supramolecular
SC
aggregate of monomers.[31] Simultaneously, Hong et al. reported on a physical, self-assembled
NU
trimer entrapped within polydopamine.[32] Even more recent, Della Vecchia et al. reveiled a three-component structure of polydopamine. Nevertheless, they suggest representing
MA
polydopamine as a collection of oligomeric species in which monomer units are linked through different bonding. The coexistence of such structurally diverse components accounts for a unique
TE
D
blend of eumelanin-like and amine containing polycatechol functionalities.[33] Despite the so far unknown structure of polydopamine, the technique has previously been successfully applied on
AC CE P
Ti and Ti alloy surfaces to immobilise biomolecules, such as dextrane, chitosan, hyaluronic acid, heparin, gelatin and growth factors.[34-40] Besides the advantageous characteristics of being a water-based and non-toxic coating strategy, it also presents a straightforward and versatile method to coat a wide variety of surfaces, ranging from the one of a 2D substrate to the one of complex 3D porous implants. Hence, its potential for future use on implants is secured as these final implants can have very customised morphologies which at this moment is a limitation for the present line of sight coating techniques.
In the present paper we describe the application of a stable gelatin coating onto a Ti6Al4V surface by applying an intermediate polydopamine coating. In an attempt to increase the biofunctionality of the surface even more, a final coating of the cell adhesive protein fibronectin
5
ACCEPTED MANUSCRIPT was applied onto the gelatin. Due to the natural binding affinity between fibronectin and gelatin a natural presentation of the immobilised fibronectin on the surface and, hence, a retention of its
PT
cell adhesive activity, is anticipated.[41] The coatings were optimised on two dimensional (2D) Ti
RI
substrates and characterised via static contact angle (SCA) measurements, X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM) and fluorescence microscopy.
SC
In the second part of this study, the optimised gelatin coating technology was applied onto
NU
three dimensional (3D) porous Ti6Al4V implants (height: 3 mm diameter: 6 mm) which were fabricated via additive manufacturing (AM) technology, i.e. selective laser melting (SLM).[42-44]
MA
3D micro-computed tomography (Micro-CT) imaging was used to characterise the average pore and strut size, porosity and surface area of the as-produced porous implants. The as-produced
TE
D
Ti6Al4V porous implants were divided into four categories: (i) uncoated or oxidised (ii) polydopamine coated (Dop) (iii) polydopamine/gelatin coated (Dop-Gel) and (iv)
AC CE P
polydopamine/gelatin/fibronectin coated (Dop-Gel-Fn). Transfer of the coating technology from Ti to Ti6Al4V surfaces was straightforward since the surface composition of the oxidised alloy largely compares to the one of the oxidised, pure Ti surface.[45-46] The effects of the different coatings on in vitro cell seeding efficiency (CSE), proliferation and differentiation of a clinically relevant mesenchymal stem cell-like osteoprogenitors, namely human periosteum-derived cells (hPDCs), were evaluated.
2. Materials and methods
2.1. Materials
6
ACCEPTED MANUSCRIPT Ti foils (0.125 mm thick, 99.6% pure according to the data sheet) were obtained from Chempur (Karlsruhe, Germany). Cylindrical porous Ti6Al4V implants (6 mm diameter x 3 mm
PT
height) were fabricated by an in-house selective laser melting (SLM) machine (figure 1).[44, 47-
RI
49] The average as-produced pore size, strut size, porosity and surface area, as shown in figure 1,
AC CE P
TE
D
MA
NU
SC
were assessed by micro-CT image analysis.
Figure 1. (A) Cylindrical Ti6Al4V porous implants made by selective laser melting (6 mm diameter x 3 mm height). (B) Morphological parameters of the as-produced Ti6Al4V porous implant assessed by micro-CT image analysis. Mean (± standard deviation)
Concentrated hydrochloric acid (HCl; 37%) was provided by Panreac Quimica S.A (Barcelona, Spain), hydrogen peroxide (H2O2; 30 %) and dimethylformamide (DMF) by Aldrich (Bornem, Belgium) and ammonium hydroxide (NH4OH; 25%) by Acros (Geel, Belgium). HPLC grade acetone, HPLC grade pentane and 2-(3,4-dihydroxyphenyl)ethylamine hydrochloride (dopamine hydrochloride) were all purchased from Sigma (Bornem, Belgium). Cyclohexane was obtained from Fiers (Kuurne, Belgium), it was dried with calcium hydride and distilled before use. Gelatin type B, isolated from bovine skins after an alkaline pre-treatment, was a kind gift
7
ACCEPTED MANUSCRIPT from Rousselot (Ghent, Belgium). Gelatin with an isoelectric point of ca. 5, gel strength of 257 Bloom and viscosity (6.67%, 60°C) of 4.88 mPa.s was used. Oregon Green® 488, carboxylic
PT
acid, succinimidyl ester was purchased from N.V. Invitrogen SA (Ghent, Belgium). Fibronectin
RI
from bovine plasma (0.1 % solution, 1 mg/ml in 0.5 M NaCl, 0.05 M Tris-HCl and pH = 7.5) was obtained from Sigma-Aldrich (Bornem, Belgium). The water used throughout this study was
NU
SC
milliQ (double distilled, > 18 MegaOhm/cm).
MA
2.1. Methods
D
2.2.1. Pre-treatment of Ti and Ti6Al4V
TE
Cleaning of the metal substrates and porous implants consisted of sequenced ultrasonic
AC CE P
treatments in (1) cyclohexane (10 minutes), (2) 10 N HCl (30 minutes) and (3) milliQ water (30 minutes). After cleaning, the samples were oxidised for 20 minutes in a 1/1/5 NH4OH/H2O2/H2O-mixture. These steps are described in more detail in a previous paper.[28]
2.2.2. Deposition of polydopamine
The method applied is based on a previous study.[30] Briefly, oxidised metal samples were incubated overnight in a solution of dopamine (2 mg/ml) in Tris-HCl (10 mM, pH = 8.5). Afterwards, the modified samples were rinsed with milliQ water, acetone and dried with pressurised air.
2.2.3. Immobilisation of gelatin and fibronectin coatings
8
ACCEPTED MANUSCRIPT
Polydopamine coated Ti samples were incubated overnight in a 0.5, 1, or 3 w/v% solution of
PT
gelatin in 10 mM Tris-HCl (pH = 8) at 40°C. The gelatin modified samples were rinsed with milliQ water, acetone and dried with pressurised air.
RI
Fluorescent labelled gelatin was obtained by dissolving 15 mg gelatin (5.25 10-3 mmol Ɛ-
SC
amine side groups of lysine and hydroxylysine) in 7.5 ml 0.1 M NaHCO3 buffer. 2 mg of the
NU
succinimidyl ester of Oregon Green® (3.93 10-3 mmol), dissolved in 200 µl DMF, was added and the mixture was stirred for 90 min at room temperature. Purification was done by using PD-
MA
10 columns. In a next step, polydopamine modified Ti samples were incubated in a 2 ml solution of the fluorescent labelled gelatin (1.4 mg/ml). Labelling of fibronectin occurred analogously to
TE
D
the gelatin. A fibronectin solution (700 µl, 1 mg/ml) was diluted with 1.5 ml 0.1 M NaHCO3 buffer and 3 mg of the succinimidyl ester of Oregon Green®, dissolved in 300 µl DMF, was
AC CE P
added. After 1 h reaction at room temperature, the mixture was purified by means of a PD-10 column. Subsequently, (non labelled) gelatin modified Ti samples were dipcoated with the fluorescent labelled fibronectin (200 µg/ml) and rinsed by incubation in a 0.1 M NaHCO3 buffer for 2 h. The buffer was refreshed after 1 h. In case of the 3D porous implants, a gelatin concentration of 1 w/v% was applied. Deposition of fibronectin on the gelatin modified surfaces was realized by a dipcoating step using a fibronectin solution of 0.01 w/v%. After coating, the porous implants were sterilised with ethylene oxide gas (cold cycle) to avoid coating degradation due to the heat induced during sterilisation.
2.2.4. Stability study
9
ACCEPTED MANUSCRIPT
Modified Ti samples were incubated in phosphate buffer (50 mM, pH = 7.4) at 37 °C for
PT
different incubation times. After incubation, the samples were gently rinsed with water and
RI
acetone and dried with pressurised air and analyzed as described below.
NU
SC
2.2.5. Surface analysis
MA
2.2.5.1. X-ray photo-electron spectroscopy (XPS)
The chemical composition of the different metal surfaces was determined using ‘‘FISONS S-
TE
D
PROBE’’, a dedicated XPS (X-ray photoelectron spectroscopy) instrument designed to give high-end analysis performance, while providing a high sample throughput. The fine focus Al-Ka
AC CE P
source with a quartz monochromator, developed by Fisons Instruments Surface Science ensures lower background and higher sensitivity than conventional twin anode sources. All measurements were performed in a vacuum of at least 10-9 Pa. Wide and narrow-scan spectra were acquired at pass energy of 158 and 56 eV, respectively. The binding energy was calibrated by the C 1s peak at 285 eV. The spot size used was 250 µm on 1 mm. Data analysis was performed using S-PROBE software. The measured spectrum was displayed as a plot of the number of electrons (electron counts) versus electron binding energy in a fixed, small energy interval. Peak area and peak height sensitivity factors were used for the quantifications. All surface compositions reported in this work are expressed as atomic percentages (at%).
2.2.5.2. Static Contact Angle (SCA)
10
ACCEPTED MANUSCRIPT
Static contact angle measurements on the 2D Ti surfaces were performed using an optical
PT
contact angle measuring system, OCA 20, from Dataphysics (distributed by Benelux Scientific)
RI
equipped with a 500 µl Hamilton syringe. For each measurement, 1 µl of milli-Q water was placed on the surface. The droplet was imaged using a video camera. The contact angle was
SC
determined on the screen of the monitor using the imaging software provided by the supplier
NU
(SCA 20, version 2.1.5 build 16). Three measurements were made on each sample. Static contact angle measurements could not be performed on the porous Ti6Al4V implants,
MA
due to the immediate penetration of the water drop into the porous implants.
TE
D
2.2.5.3. Atomic Force Microscopy (AFM)
AC CE P
The AFM study was performed with a Nanoscope IIIa Multimode (Digital Instruments, Santa Barbara, California, USA) in ‘tapping mode’ and exposed to air (n = 1). The obtained images were analysed with Nanoscope Image III software. Changes in surface morphology are quantified by root mean square (RMS) roughness values (Rq). Rq is the RMS deviation of the profile and is defined as follows:
(Z i Z ave) 2 Rq N where Zave is the average Z height value within a given area, Zi the current Z value and N is the number of points within the given area.
2.2.5.4. Fluorescence microscopy
11
ACCEPTED MANUSCRIPT Fluorescent microscopy images were recorded with an Olympus IX81 to visualise the presence of the gelatin and fibronectin coating. The chosen filters were an excitation filter with a band
PT
width of 460 – 490 nm and an emission filter with LP510 long path characteristics. Since this
RI
technique is limited to visualise only the surface of the porous implants and not the internal
SC
structure, this technique was only used for the 2D Ti surfaces.
NU
2.2.6. 3D cell seeding and in vitro cell proliferation and differentiation analysis
MA
For the in-vitro cell tests human periosteum-derived cells (hPDCs) were chosen since they play a major role in bone fracture repair[50] and have a higher in-vitro expansion rate compared to
D
osteoblasts.[51] hPDCs (pooled from periosteal biopsies of six donors, age 14.9 + 2.1) were
TE
expanded from the liquid nitrogen cell bank using Dulbecco’s modified eagle’s medium
AC CE P
(DMEM+GlutaMaxTM-1, Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% antibiotics and 1% sodium pyruvate. Upon confluence, cells were harvested by trypsinization and a cell suspension (60 µl) containing 100 000 hPDCs was drop seeded onto each porous implant, incubated statically for 1 h at 37 oC to facilitate cell attachment, before being transferred to a rotator to perform dynamic rotation seeding.[52] After overnight dynamic seeding, the cellseeded porous implants were transferred to a 48-well plate and cultured in growth medium (GM) for 14 days. The medium was refreshed three times a week. Cell seeding efficiency (CSE) was calculated by determining the ratio of the DNA content of the cells successfully seeded on the porous implant to the initially cell seeding number. This, according to Impens et al.[53], results in the most realistic CSE values compared to other calculation methods.
12
ACCEPTED MANUSCRIPT At 1, 7 and 14 days, the proliferation of hPDCs on different coatings was assessed by nondestructive analysis of the metabolic activity using AlamarBlue® (Invitrogen, 10% in GM), and
PT
also by quantifying the DNA content of each sample at defined time points using Quant-iTTM dsDNA measurement kit (Invitrogen, USA). Cell viability on different coatings was characterised
RI
by staining the living and dead cells using the Live/Dead® cell viability kit (Invitrogen).
SC
Osteogenic differentiation of hPDCs was evaluated by the measurement of alkaline phosphatase
NU
(ALP) levels using a BluePhos® microwell phosphatase substrate system (Kirkegaard & Perry Laboratories, USA). The measured absorbance was normalised to the DNA content of the
MA
respective samples.
TE
D
2.2.7. Statistical analysis.
AC CE P
Statistical analysis was performed on the obtained XPS data, using the student t-test. Two data sets (n = 3) were considered to be significantly different when p < 0.05. One-way unpaired ANOVA (n= 3) was used to analyze statistical differences between the cell seeding results. Two-way unpaired ANOVA (n= 3) was used to analyze statistical differences between the DNA, ALP and metabolic activity results. The Tukey test was used to perform post hoc comparisons. All error bars represent 1 σ standard deviation. Two data sets were considered to be significantly different when p < 0.01.
3. Results & Discussion
13
ACCEPTED MANUSCRIPT It is generally accepted that osseointegration requires biofunctionalisation of the implant surface. In this study, we opted to apply the extra cellular matrix biopolymers gelatin and
PT
fibronectin as coatings on Ti6Al4V porous implants. The coating technology was optimised
RI
using 2D Ti substrates and successfully transferred to 3D porous Ti6Al4V implants. Gelatin was immobilised by applying a prime layer on the Ti surface using dopamine. Based
SC
on the research of Lee et al. a polydopamine coating was deposited onto Ti by incubating
NU
oxidised Ti samples overnight in a solution of dopamine in Tris-HCl buffer (pH = 8.5) [30]. The stability of the polydopamine layer was determined by incubating the modified Ti sample in
MA
phosphate buffer at 37 °C for 24 h. As a control, an oxidised sample was incubated in Tris-HCl buffer overnight. The Ti surfaces were characterised by XPS and SCA measurements. For
TE
D
comparison purposes, the theoretical and the measured values of the atomic composition of dopamine and the measured atomic composition of an oxidised surface are also presented in
AC CE P
table 1.
Table 1. Atomic surface compositions and contact angle of dopamine and a polydopamine coated Ti surface before and after 24 h incubation in phosphate buffer at 37 °C (n=3), as determined by XPS and SCA. An oxidised Ti sample, incubated for 24 h in Tris-HCl buffer, served as a control. The data of dopamine (theoretical and measured) and of an oxidised Ti surface served as references. Values with a standard deviation are an average of three measurements. %Ti
%C
%O
%N
N/C
Contact angle
Dopamine (theoretical)
-
73
18
9
0.12
-
Dopamine-HCl (measured)
-
68
20
6
0.09
-
Polydopamine on Ti
-
77 ( 1.2)
18 ( 0.6)
6 ( 1.5)
0.07 (± 0.015)
22° ( 3.0)
Polydopamine on Ti after 24 h incubation in phosphate buffer
-
75 ( 1.2)
18 ( 0.6)
7 ( 0.6)
0.09 (± 0.009)
-
14
ACCEPTED MANUSCRIPT 6
54
40
-
-
-
Oxidised Ti
13
38
49
-
-
12° ( 1.9)
PT
Control
The XPS results indicate that the surface composition of the Ti surface, after modification with
RI
dopamine, is in very good agreement with the atomic composition of the dopamine monomer
SC
which confirms the presence of a polydopamine layer on the surface. Measurements on different
NU
positions on the same sample confirmed that the layer fully covered the Ti surface (data not shown). Furthermore, it can be concluded that the polydopamine layer was stable for at least 24 h
MA
in phosphate buffer. The control, however, showed a higher C content after incubation in TrisHCl buffer. This is probably due the adsorption of contaminants originating from the buffer.
D
In a next step, gelatin was immobilised onto the polydopamine coated Ti surface by incubating
TE
the sample overnight at 40°C in a 0.5, 1 or 3 w/v% solution of gelatin in Tris-HCl buffer
AC CE P
(pH = 8). According to Lee et al., the polydopamine catechols can react with thiols and amines via Michael addition or Schiff base reactions.[30] Gelatin possesses amine functionalities and, thus, should be able to covalently bind to the polydopamine and lead to a stable surface coating. Based on the results in table 2 there appears to be a small change in atomic surface composition in comparison to the one of a polydopamine modified surface (table 1). For each applied gelatin concentration there is a decrease in the C content and an increase in the N content resulting in an increase of the N/C ratio. The values are similar to those of gelatin as such, thus confirming the presence of a gelatin coating on the Ti surface. Moreover, the C contents of the polydopamine and gelatin coated surfaces and their respective contact angle values differ significantly (p < 0.05), indicating two different surface compositions or two different coatings.
15
ACCEPTED MANUSCRIPT
%Ti
%C
%O
N/C
Contact angle
0h
-
69 ( 1.0)
19 ( 1.2) 11 ( 1.0)
0.16 ( 0.017)
-
1h
-
69
19
12
0.17
-
12 h
-
70
19
11
0.16
-
24 h
-
69 ( 0.0)
19 ( 0.6) 12 (0.6)
0.17 ( 0.008)
-
0h
-
68 ( 1.5)
20 ( 0.6) 12 ( 1.0)
0.18 ( 0.018)
10° ( 0.7)
1h
-
69
19
12
0.17
-
12 h
-
69
19
12
0.17
-
24 h
-
MA
NU
0.5%
Gelatin
TE -
AC CE P
3%
D
1%
0h
%N
RI
Incubation time
SC
Gelatin Concentration
PT
Table 2. Atomic surface compositions and contact angle of gelatin coated Ti surfaces, after the deposition of a polydopamine coating, as determined by XPS and SCA. Results from before (0 h) and after (1, 12 and 24 h) incubation at 37 °C in phosphate buffer are shown. Values with a standard deviation mentioned are an average of three measurements.
69 ( 1.0)
19 ( 4.9) 11 ( 0.6)
0.16 ( 0.010)
-
68 ( 0.6)
20 ( 0.6) 12 ( 1.0)
0.18 ( 0.018)
-
1h
-
68
19
13
0.19
-
12 h
-
68
20
12
0.18
-
24 h
-
68 ( 0.0)
20 ( 0.6) 12 ( 0.6)
0.18 ( 0.008)
-
-
69
18
0.19
-
13
During the XPS analyses the C 1s peaks of the polydopamine and gelatin coatings were also compared (figure 2). The spectra show C 1s peaks of the different chemical environments. In case of the polydopamine coated surface, peak deconvolution reveals three distinct peaks at 284.8, 286.0 and 288.8 eV originating from respectively C-C/C-H-, C-N/C-O- and C=O-bonds. In case of the gelatin coated surface, peaks at respectively 284.8, 286.0 and 287.9 eV are
16
ACCEPTED MANUSCRIPT detected. In contrast to the carbonyl function of polydopamine, the carbonyl function of gelatin originates (mostly) from amide bonds.
PT
C-H C-C
C-H C-C
C-O
RI
C-O
SC
C-N
C-N
C=O
288 285 Binding Energy (eV)
282
MA
291
NU
C=O
291
288 285 Binding Energy (eV)
282
TE
D
Figure 2. XPS detailed spectra of the C 1s peaks of polydopamine coated Ti (left) and gelatin coated Ti after deposition of a polydopamine layer (right).
AC CE P
In a next step, the stability of the gelatin coating was studied by incubating the samples in phosphate buffer (50 mM, pH = 7.4) at 37 °C during 1, 12 and 24 h, and subsequently determining the surface composition by means of XPS. Stability longer than 24h was not assessed since initial cell attachment already occurs minutes to a few hours after the seeding step. The values obtained at all time points (table 2) were not significantly different from the control, indicating sample stability for at least one day. Nevertheless, the C and N contents are significantly different from polydopamine coated Ti surface. Based on these data, it can be concluded that the applied immobilisation technique, based on a polydopamine prime layer, leads to a stable gelatin coating on Ti. An additional technique to confirm the presence of a gelatin coating is fluorescent microscopy. For this, gelatin was modified with the fluorescent label Oregon Green® (Gel-OG, figure 3)
17
ACCEPTED MANUSCRIPT before immobilisation on the polydopamine coated Ti surface. The obtained fluorescence microscopy images are presented in figure 4. Images B and C show the surface respectively
PT
before and after 2 h of rinsing in buffer. Both images show visually a homogeneously spread gelatin coating. In a next step, the cell adhesive protein fibronectin, and not the gelatin, was
RI
labelled with Oregon Green® (Fn-OG) and subsequently deposited by dipcoating. This enabled
O
O
F
F C OH
D
O
O C O
+
AC CE P
O
NH2
N O C OG
TE
N
≡
O
O
Gel
MA
HO
NU
SC
us to study possible aspecific binding of fibronectin to polydopamine.
N O C OG O
Gel
NH C OG O
GelB-OG
Figure 3. Chemical structure of Oregon Green® and reaction of gelatin with the succinimidyl ester of Oregon Green®.
A homogeneous distribution of Fn-OG was visually observed on the gelatin coated Ti sample (image D), again confirming the presence of gelatin on the surface. In case of the polydopamine coated sample, almost no Fn-OG was detected (image E). This finding does not only confirm the specific interaction between gelatin and fibronectin[41, 54-55] but also the presence of the gelatin layer in images B and C. A control image of non coated cp Ti (image A) indicated that the polydopamine coating does not bind Oregon Green®.
18
AC CE P
TE
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
200 µm
200 µm
Figure 4. Fluorescent microscopy images (magnification 10x) of (A) cp Ti, (B) polydopamine coated Ti incubated in GelB-OG solution, (C) polydopamine coated Ti incubated in GelB-OG solution after rinsing, (D) gelatin coated Ti dipcoated with Fn-OG after rinsing, and (E) polydopamine coated Ti dipcoated with Fn-OG.
In a final step of the material structural characterisation, the possible influence of the dopamine and subsequent gelatin modification on the Ti surface roughness was studied by AFM. It can be concluded from the left image in figure 5 that polydopamine aggregates are formed during the
19
ACCEPTED MANUSCRIPT deposition of the polydopamine layer, strongly affecting the surface roughness. The Rq value
MA
NU
SC
RI
PT
increases from 154 nm for an oxidised surface to 250 nm for a polydopamine modified surface.
TE
D
Figure 5. Surface topography of a polydopamine coated Ti surface (left image) and a gelatin coated Ti surface (right image), as determined by means of AFM.
AC CE P
From the image on the right side in figure 5, it can be concluded that the modification of a polydopamine coated surface with gelatin leads to a smoother surface. This observation is confirmed by the decrease in Rq value to 173 nm.
In the second part of this study, 3D Ti6Al4V porous implants were fabricated using SLM and coated with polydopamine, gelatin and fibronectin by applying the coating procedures optimised on the 2D Ti samples. Comparison of the surface compositions of polydopamine and gelatin coated 2D Ti (table 1 and 2) and 2D Ti6Al4V surfaces (table 3) confirmed the similarity between both, enabling a successful and straightforward transfer of the coating technology from Ti to the Ti6Al4V alloy. For reason of simplicity, the intermediate 1% gelatin solution was chosen for the modification of the 3D Ti6Al4V porous implants and 2D Ti6Al4V surfaces. Furthermore, stability of the coatings up to 2 weeks (to match the in vitro tests) was studied with still a 20
ACCEPTED MANUSCRIPT significant (p < 0.05) difference in C and N content between polydopamine and gelatin coated
PT
Ti6Al4V surfaces (table 3).
N/C
49 ( 0.5)
1 ( 0.2)
2 ( 0.1)
0.02 ( 0.005)
39 ( 1.2)
47 ( 0.6)
1 ( 0.1)
2 ( 1.2)
0.03 ( 0.004)
37 ( 0.5)
50 ( 0.2)
0 ( 0.2)
2 ( 0.3)
0.00 ( 0.006)
73 ( 0.1)
20 ( 0.2)
7 ( 0.3)
-
0.10 ( 0.005)
72 ( 0.2)
21 ( 0.1)
7 ( 0.2)
-
0.09 ( 0.003)
%C
0h
12 ( 0.3)
36 ( 0.4)
24 h
10 ( 0.2)
2 weeks
10 ( 0.2)
0h
-
24 h
1 % Gelatin on Ti6Al4V-Dop
%Ca
%Ti
MA
NU
%O
D
-
AC CE P
Polydopamine on Ti6Al4V
%N
Incubation time
TE
Oxidised Ti6Al4V
SC
RI
Table 3. Atomic surface compositions of oxidised, polydopamine coated and gelatin coated Ti6Al4V surfaces, as determined by XPS. Results from before (0 h) and after (1h, 24 h and 2weeks) incubation at 37 °C in phosphate buffer are shown. Values with a standard deviation mentioned are an average of three measurements.
2 weeks
-
69 ( 0.1)
23 ( 0.5)
7 ( 0.5)
-
0.10 ( 0.008)
0h
-
67 ( 0.2)
21 ( 0.0)
12 (0.2)
-
0.18 ( 0.005)
24 h
-
68 ( 0.5)
20 ( 0.3)
12 ( 0.6)
-
0.17 ( 0.010)
2 weeks
-
68 ( 0.2)
22 ( 0.2)
10 ( 0.1)
-
0.15 ( 0.001)
Using these surfaces, CSE was determined 2 hours after cell seeding. From the results in figure 6, it can be shown that there was no significant effect of the coating on the CSE. Furthermore, the effect of the coatings on hPDCs proliferation and differentiation after 14 days of in vitro cell culture was evaluated in growth medium.
21
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
AC CE P
TE
D
MA
Figure 6. Effect of coating on the CSE as determined by dynamic rotation seeding method (n=3).
22
AC CE P
TE
D
MA
NU
SC
RI
PT
ACCEPTED MANUSCRIPT
Figure 7. Representative images of live/dead staining of human periosteum derived cells (hPDCs) seeded on Ti6Al4V porous implants with various coatings. The seeded cells were found viable during 14 days of in vitro culture with growth.
23
ACCEPTED MANUSCRIPT Live/dead staining provided qualitative information on the cell behaviour (figure 7). After 1 day of in vitro cell culture a uniform distribution of hPDCs throughout the fibronectin coated
PT
Ti6Al4V porous implants (Dop-Gel-Fn) was noticed. On the contrary, hPDCs were distributed
RI
mostly at the bottom of the gelatin coated (Dop-Gel), the polydopamine coated (Dop) and the oxidised Ti6Al4V porous implants. After 7 days of in vitro cell culture, cell distribution became
SC
uniform for all different coating groups. Furthermore, hPDCs started to occlude the pores by
AC CE P
TE
D
MA
NU
bridging the corners, resulting in circular cell-filled pores (figure 8).
Figure 8. Representative image of hPDCs bridging the corners of Ti6AL4V porous implant pores after 7 days of in vitro culturing.
Eventually, this phenomenon resulted in almost complete cell-filled pores in the top and side surface of the different Ti6Al4V porous implant coating groups after 14 days of in vitro cell culturing. Based on the live dead staining, it can be stated that the difference in surface chemistry, due to the applied coatings, does not influence this circular pore filling behaviour after 14 days in vitro. This is in line to what has been observed by other researchers on polymer, ceramic and metal porous structures, thus different surface chemistries.[56-60] In a next part of the work, a quantitative DNA assay and a metabolic activity analysis were performed (figure 9 A and B). Similar to what has been visually noticed on the live/dead staining 24
ACCEPTED MANUSCRIPT micrographs, hPDCs were viable and able to proliferate during 14 days of in vitro culturing which resulted in significantly higher levels of measured DNA and metabolic activity after 14
PT
days of in vitro cell culturing. However, on day 14, no significant differences (p > 0.01) were found between the different samples.
RI
Finally, ALP, a marker to indicate early stage of osteogenic differentiation, was measured.[61]
SC
During the 14 days of in vitro culturing, no significant increase in ALP was measured for all
NU
tested samples (not shown in figure since values equal zero). These results imply that the
AC CE P
TE
D
MA
investigated coatings did not influence the in vitro differentiation of the hPDCs.
25
ACCEPTED MANUSCRIPT
PT
Figure 9. DNA assay (A) and metabolic activity (B) of cells on the different coated Ti6Al4V porous implants cultured in GM. Error bars represent standard deviation intervals. Statistical analysis: unpaired t-test (** p
