Amino-modified silica surfaces efficiently immobilize bone morphogenetic protein 2 (BMP2) for medical purposes

Acta Biomaterialia 7 (2011) 1772–1779

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Amino-modified silica surfaces efficiently immobilize bone morphogenetic protein 2 (BMP2) for medical purposes Nina Ehlert a,b, Andrea Hoffmann c,1, Tammo Luessenhop a,b, Gerhard Gross c, Peter P. Mueller c, Martin Stieve d, Thomas Lenarz d, Peter Behrens a,b,⇑ a

Institut für Anorganische Chemie, Leibniz Universität Hannover, Callinstraße 9, 30167 Hannover, Germany Center for Solid-State Chemistry and New Materials, Leibniz University Hannover, 30167 Hannover, Germany Helmholtz-Zentrum für Infektionsforschung GmbH, Inhoffenstraße 7, 38124 Braunschweig, Germany d Hals-Nasen-Ohren-Klinik, Medizinische Hochschule Hannover, Carl-Neuberg-Straße 1, 30625 Hannover, Germany b c

a r t i c l e

i n f o

Article history: Received 25 August 2010 Received in revised form 6 December 2010 Accepted 21 December 2010 Available online 25 December 2010 Keywords: Bone morphogenetic protein 2 Immobilization Nanoporous silica BioveritÒ II Mesenchymal progenitor cells

a b s t r a c t Due to its ability to induce de novo bone formation the differentiation factor bone morphogenetic protein 2 (BMP2) is often used to enhance the integration of bone implants. With the aim of reducing possible high dose side-effects and to lower the costs, in order to produce affordable implants, we developed a simple and fast method for the immobilization of BMP2 on silica-based surfaces using silane linkers which carry amino or epoxy functions. We put an especial emphasis on the influence of the nanoscale surface topography of the silica layer. Therefore, we chose glass (for control experiments) and BioveritÒ II (as a typical implant base material) as support materials and coated these substrates with unstructured or nanoporous amorphous silica layers for comparison. Immobilized BMP2 was quantified by two different methods: by ELISA and by a cell-based assay for active BMP2. These tests probe for immunologically and biologically active BMP2, respectively. The results show that the amino functionalization is better suited for immobilizing the protein. Strikingly, a considerably higher amount of BMP2 could be immobilized on coated BioveritÒ II surfaces compared with coated glass substrates, which was presumably due to the macroscopic roughness of the BioveritÒ II substrates. In addition, it was found that the nanoporous silica coatings on BioveritÒ II substrates were able to bind more BMP2 than the unstructured ones. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Enhancement of the surface properties of medical implants for improved integration into their biological environment is a major focus of current biomaterial research. In particular, improvements in the cell–material surface interactions of bone replacement implants by surface functionalization appears highly promising to facilitate the healing process. Due to the fact that bone morphogenetic proteins (BMPs) induce mesenchymal progenitor cells to differentiate into osteoblasts and chrondrocytes they have highly potent osteoinductive properties and play a key role in bone formation and repair. BMPs can be used to improve the integration of bone replacement prostheses and to speed up the healing process. In particular, BMP2 can induce bone formation in both ectopic

⇑ Corresponding author at: Institut für Anorganische Chemie, Leibniz Universität Hannover, Callinstraße 9, 30167 Hannover, Germany. Tel.: +49 511 762 3697; fax: +49 511 762 3600. E-mail address: [email protected] (P. Behrens). 1 Present address: Unfallchirurgie, Medizinische Hochschule Hannover, Carl-Neuberg-Straße 1, 30625 Hannover, Germany.

[1–4] and orthotopic sites [5–15]. Although BMP2 must be present in a certain minimum local concentration to induce new bone formation, excess free BMP2 is not practical in surgical praxis because BMP2 possibly has unwanted side-effects in other regions of the body, including induction of immune responses [16–18]. In addition, BMP2 is costly and loses its bioactivity in solution after a short time in vivo, which makes high dose clinical treatments with BMPs unfeasible [19,20]. Two approaches have been developed in order to achieve a low but sufficient and sustained local supply of active BMP2. The protein may be incorporated into either a polymer or mineral carrier phase which may be non-degradable or degradable within a certain period of time [2,5,7,13,14,21–29]. The delivery is then controlled either by the diffusion of BMP2 within the carrier or by degradation of the carrier. The other approach consists of connecting the BMP2 firmly and permanently to the implant surface, which has the advantage of preventing BMP2 activity in untargeted tissues. This approach can be realized by chemically attaching the protein to the implant surface, either by covalent or by strong ionic bonding. Several techniques are based on modification of the surface by an aminosilane, such as 3-aminopropyl-triethoxysilane.

1742-7061/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2010.12.028

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

The reactive amino function is then used to couple the protein to the surface using succinimidyl 4-(N-maleidomethyl)cyclohexane1-carboxylate (SMCC) [30,31], hexamethylene diisocyanide [32] or carbonyldiimidazole (CDI) [8,9,33,34] as linking agent. Another complex procedure starts with a phosphonic acid monolayer on a titanium surface and applies polymer chemistry for fixation of the protein [35]. BMP2 can also be bound to dextran-coated titanium surfaces by a reductive amination method [36]. Additionally, strategies have been described based on the strong physisorption of BMP2 on nanocrystalline diamond [37,38] and immobilization of the protein on plasma-activated polystyrene [39]. Recently, controversially, whether strongly bound BMP2 can be biologically effective has been discussed. The Shiba group showed that BMP2 activity as assessed by the induction of BMP2-dependent signalling cascades and the induction of cellular differentiation is enhanced when binding to the surface is reversible [39]. However, their system involved a genetically altered, non-naturally occurring BMP2 variant that had been created by molecular biological techniques. In contrast, some promising work has been presented showing that BMP2 fixed on a surface can still be biologically active [9,22,30,32,36,39,41]. In this study we describe the immobilization of recombinant human BMP2 on structurally different silica surfaces by means of a silane linker without the need to use a further linking agent like SMCC or CDI (Fig. 1). These silica surfaces are either an unstructured amorphous silica layer or an amorphous silica layer containing regular nanopores, both on plain glass disks or BioveritÒ II substrates. We were especially interested in the properties of the nanoporous layer with regard to its capability to bind BMP2. Such nanoporous layers (or, as they are also often called in materials chemistry, mesoporous layers) are currently being studied with regard to their applications as biomaterials [42–47]. Our former studies on the immobilization of the enzyme alkaline phosphatase [48] on such nanoporous layers had shown promising results, allowing the immobilization of large amounts of enzymatically active immobilized protein. These layers can also be coated onto BioveritÒ II substrates [49]. BioveritÒ II is a glass–mica ceramic with various applications as a non-resorbable bone substitute material

1773

[50–53]. The present study was undertaken with special regard to the modification of BioveritÒ II middle ear prostheses to locally induce bone formation in order to provide a strong fixation of the prosthesis to remaining bone structures [42]. The amount of immobilized BMP2 was quantified by two different methods: a standard ELISA and a cell-based test. The ELISA measures immunologically active BMP2, whereas the cell test quantifies biologically active BMP2 only. This cell test relies on the signalling cascade initiated by biologically active BMP2. 2. Materials and methods 2.1. Nanoporous and unstructured silica films In this study two different types of base materials were used, namely BioveritÒ II (3di GmbH, Jena, Germany) and glass (Glasbearbeitung Henneberg and Co., Martinroda, Germany), both in the form of square sheets (10  10 mm) with a height of 1.0– 1.3 mm for the BioveritÒ II and 0.95 mm for the glass. The substrates were coated with unstructured or nanoporous silica layers. Prior to use all specimens were cleaned in absolute ethanol (Merck, Darmstadt, Germany) and acetone. All chemicals except for the ethanol were purchased from Sigma–Aldrich Chemie GmbH (Munich, Germany) and were used without further purification. The solution used for the preparation of nanostructured silica coatings contained ethanol, water, hydrochloric acid, tetra-ethoxysilane (TEOS) as a silica source and poly(ethylene glycol)– poly(propylene glycol) block co-polymer, PEG-PPG-PEG, (Sigma– Aldrich, EO20PO70EO20, average Mn 5800, similar to PluronicÒ P-123, BASF) as the structure-directing agent [54]. A solution with the molar composition TEOS:EtOH:H2O:HCl:EO20PO70EO20 = 1:48.9:26.9:0.06:0.0135 was prepared by adding the TEOS to EO20PO70EO20 dissolved in a mixture of ethanol, water and hydrochloric acid and was stirred for about 10 min before coating the specimens. The unstructured silica coatings were prepared using similar solutions, but without the EO20PO70EO20. The BioveritÒ II and the glass squares were coated using a dip coating procedure, employing a d.c. small dip-coater with 75 mm

Fig. 1. Scheme of the strategies for the immobilization of BMP2 on silica surfaces.

1774

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

travel from NIMA (Coventry, UK), operated in a climate box at constant humidity, adjusted using a 50 wt.% glucose solution. The samples were immersed in the coating solution and then withdrawn perpendicular to the surface of the solution at a speed of approximately 1 mm min1. The samples were then left at constant humidity for 5 min, followed by a drying step at 60 °C for 30 min. For the BioveritÒ II samples this procedure was repeated twice, resulting in three layers of silica. Multiple coating of the BioveritÒ II substrates was necessary to fill the cavities present on the material surface and to create a continuous layer. For the glass substrates only a single coating was needed. Afterwards the specimens were dried at 60 °C overnight, followed by calcination at 415 °C for 4 h (rate of heating/cooling 1 °C min1). 2.2. Characterization The presence of the nanostructured layers was confirmed by Xray diffraction (XRD) and scanning electron microscopy (SEM). The samples were measured in a Stoe (Darmstadt, Germany) – diffractometer in reflection geometry. A secondary beam monochromator (graphite) was applied to produce CuKa radiation. SEM images were collected in a field emission scanning electron microscope type JSM-6700F from JEOL (Eching, Germany) with an acceleration voltage of 2 keV and a working distance of 15 mm. The thickness of the mesoporous silica film was measured with a stylus profiler DETAK6 M from Veeco Instruments Inc. (Plainview, NY) with a force of 9 mg, duration of 100 s and length of 3000 lm per measurement. 2.3. Surface modification The calcined samples were modified by a chemical functionalization step. After cleaning with absolute ethanol the specimens were immersed for 2 min in 10 wt.% aqueous solutions of 3-aminopropyl trimethoxysilane or glycidyloxypropyl trimethoxysilyl silane, respectively. They were then rinsed with water and dried under a flowing argon atmosphere for 10 min. 2.4. Immobilization of the BMP2 After surface modification the substrates were covered with a solution of recombinant human BMP2 (250 lg ml1 in 50 mM 2(N-morpholino)ethanesulfonic acid (MES) buffer, pH 5.0) and left overnight at 4 °C under gentle shaking. The BMP2 was produced by and purified from genetically engineered Escherichia coli bacteria [55]. The washing procedure consisted of eight washings with 0.125 M sodium tetraborate buffer (pH 10.0) containing 0.066% (w/v) sodium dodecyl sulphate (SDS), followed by one washing step with phosphate-buffered saline (PBS), resembling the washing procedure decribed by Jennissen et al. for in vivo experiments [56]. For negative control experiments the procedure was identical except that MES buffer solution was used instead of protein. 2.5. Stability testing of the silica coating The stability of the nanoporous film during the coupling procedure was evaluated by XRD. For this purpose, XRD patterns of standard glass slides coated with the nanoporous silica film were taken as reference. Afterwards, in analogy to the BMP2 immobilization strategy, the glass slides were put into a solution of MES buffer overnight, followed by treatment with 0.125 M sodium tetraborate buffer (pH 10.0) containing 0.066% (w/v) SDS. A second XRD pattern was then collected for comparison. To imitate the protecting surface coverage due to the immobilized protein the surface of the nanoporous silica film was additionally coated with the enzyme alkaline phosphatase following a similar procedure [48]. To determine the stability of the nanoporous silica film under cell cul-

ture conditions we recorded XRD patterns before and after incubation in 10 vol.% foetal calf serum (FCS) for different time periods between 6 h and 2 days. Ten glass slides were analysed for each time interval and representative diffraction patterns were chosen. 2.6. Quantification of the amount of immobilized BMP2 Two quantification methods were applied. An indirect enzymelinked immunosorbent assay (ELISA) was applied according to published procedures [35]. In brief, non-specific protein binding sites were blocked by incubation with 10 vol.% FCS in PBS. Monoclonal mouse anti-human BMP2 antibodies were added (R&D, Wiesbaden, Germany, catalogue No. MAB3551), followed by washing with Tris-buffered saline containing 0.1% (w/v) Tween-20 and incubation with goat anti-mouse antibody–horseradish peroxidase conjugate. After removal of unbound antibody detection was performed with 3,30 ,5,50 -tetramethylbenzidine (TMB Plus, KemEnTec, Taastrup, Denmark) and stopped with 2 M sulphuric acid. Absorbance was read at 450 nm versus 620 nm. In addition, the so-called BRE-luc (BMP2 responsive element– luciferase) assay for the determination of biologically active BMPs was used. This was carried out according to Logeart-Avramoglou et al. [57]. The assay uses a mouse muscle satellite cell line, C2C12, transfected with an inhibitor of differentiation promoter– luciferase construct [58], resulting in a BMP2-induced dosedependent increase in luciferase activity in the cell lysates. C2C12 mesenchymal progenitor cells were stably transfected with the luciferase reporter plasmid (kindly provided by Peter ten Dijke, Leiden) using DOSPER™ according to the manufacturer’s protocol (Roche, Mannheim, Germany). A selection plasmid conferring resistance to the antibiotic G418 (geneticin, pAG60) was co-transfected and cells were selected with 750 lg ml1 G418. Individual clones were picked, propagated and tested for the presence of the reporter by stimulation with BMP2 and detection of luciferase activity (cf. below). For the test C2C12-BRE-luc cells were seeded at 35,000 cells well1 of 24-well plates in medium containing 10% FCS without G418 directly onto the surface-modified substrates. 2 h after seeding the medium was removed and each well was washed once with 500 ll of medium containing 2% FCS, 4 mM glutamine and penicillin/streptomycin (‘‘test medium’’). Thereafter, 500 ll of test medium was added. 2 ll of BMP2 diluted in test medium was added for the standards to give final amounts of 200, 50, 20, 5 and 1 ng BMP2 per well. 2 ll of test medium was added for the negative control (0 ng BMP2). To correct for any putative influence of the different surface coatings on cell behaviour, these standards were performed in parallel for all surface-modified materials tested. Cells were incubated for 2 days and harvested by washing once with PBS, then frozen at 70 °C. 70 ll of CAT lysis buffer (from a CAT ELISA kit, Roche, Mannheim, Germany) with protease inhibitors (Roche, Mannheim) was added. The lysates were centrifuged for 10 min at 20,000g, 4 °C and 7.5 ll of the supernatants were used for detection of luciferase activity (Promega, Mannheim, Germany, E1500, 25 ll per sample). The amounts of BMP2 bound to the surfaces of glass or Bioverit were calculated from the standards by linear regression analysis. For both methods all experiments were carried out in triplicate. The values observed were corrected for those values obtained on blanks, which again were an average of three measurements. The samples for the blank values were prepared by the same procedure, but without BMP2. 2.7. BMP2 release experiment To determine the amount of BMP2 released from the surface into the medium under the cell culture conditions of the BRE-luc test a special release experiment was conducted. BMP2 was

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

immobilized on BioveritÒ II squares coated with a nanoporous layer as described above and the substrates incubated in 550 ll of the ‘‘test medium’’ of the BRE-luc test (cf. above). The samples were left in the medium for 5 min (0 days, negative control), 1, 2, 3 and 5 days and the supernatants were collected (three samples per time point). Then the amount of BMP2 which was released into these supernatants was determined using the BRE-luc test as described above by adding 500 ll of supernatant to each well of a 24-well plate containing 35,000 cells. A standard curve obtained with soluble BMP2 was performed in parallel to estimate BMP2 release.

3. Results 3.1. Stability testing of the nanoporous silica coating XRD diffraction measurements gave results similar to those described in detail elsewhere [48]. The nature of the nanoporous silica coating can be described as an irregular pore system with a layer thickness of from 40 to 150 nm [59]. The porosity of the film was determined by Krypton sorption measurements as presented elsewhere [59] and gave a value of 11.2 cm2 per cm2 of substrate. As recent publications have indicated a limited stability of nanoporous silica coatings in biologically relevant media [60,61] we carried out corresponding investigations using XRD and SEM. To ascertain that the nanoporous silica coating is stable throughout the whole procedure for coupling BMP2 and especially during the washing steps at pH 10 we measured standard glass slides with a nanoporous coating before and after incubation in MES buffer plus washing with 0.125 M sodium tetraborate buffer (pH 10.0) containing 0.066% (w/v) SDS (Fig. 2). The original calcined nanoporous film exhibits one peak at 1.5° 2h and a broad reflection at 2.8° 2h, corresponding to a disordered arrangement of nanopores [48]. The intensity of the main peak at 1.5° 2h is only slightly decreased after the washing procedures. The minimal intensity reduction of this reflection is not significant and we conclude that the silica film is stable under these conditions. In addition, SEM pictures were taken before and after the treatment described above. This was performed for the glass and BioveritÒ II specimens. As shown in Fig. 3, differences between the two films on the glass surface can be recognized. The glass surface be-

Fig. 2. X-ray diffraction patterns of a nanoporous silica layers on glass. Comparison of samples before (black) and after (grey) the immobilization treatment (incubation in MES buffer and washing with borate buffer).

1775

fore treatment is highly planar with few small defects. After treatment some larger areas of damage appear. Nevertheless, it can be concluded that the silica layer as a whole is still present on the surface of the glass substrate. The surfaces of the BioveritÒ II-based samples look similar before and after treatment. Here, both images demonstrate the presence of a covering silica layer, as can be inferred from the observable small cracks which result from calcination. Combining the results from XRD and SEM, we conclude that the nanoporous silica layer is still present after the treatment analogous to the immobilization step. Further stability tests were carried out with glass substrates in 10 vol.% FCS in PBS at 37 °C, mimicking the medium of the cell culture experiments in the BRE-luc test. XRD patterns measured on the films after different exposure times revealed that the nanoporous structure of the film is destroyed, showing different successive phases (Fig. 4). Within the first 6 h of incubation the intensities of the reflections at 1.5° 2h and at 2.8° 2h decreased somewhat. This reduction is observable up to an incubation time of 12 h. Afterwards the reflection at 2.8° 2h is no longer present, but a broad reflection at 2.1° 2h appears. This development can be explained by a structural rearrangement of the pore system. When the nanoporous silica film was modified with 3-aminopropylsilane and the enzyme alkaline phosphatase was immobilized as a substitute protein (according to Ref. [48]) clear reflections from the nanopore structure can be observed up to 14 h incubation. Afterwards, up to a time period of 24 h, the reflection at 1.5° 2h is reduced and a very broad reflection at 2.1° 2h remains. For longer time periods no diffraction peaks can be observed. Interestingly, profilometer measurements and microscopic controls on glass substrates showed that a silica film of reduced thickness is still present on the surface for at least 24 h, either with or without the substitute protein on the surface. In the light of these results it can be concluded that the nanoporous silica layer deteriorates under conditions resembling cell culture medium. This process involves partial dissolution of the coating and destruction of the pore system, e.g. by Ostwald ripening-type processes. Nevertheless, a non-porous, non-structured silica coating persists. 3.2. Quantification of the amount of immobilized BMP2 To quantify the amounts of immobilized BMP2 two complementary tests were carried out: an ELISA [35] and a BRE-luc [57] test. The results are shown in Fig. 5. The results depicted there reflect the data obtained under different experimental regimes on glass- and BioveritÒ II-based samples. Results from individual experiments are given in the text. Both detection methods show that it was possible to immobilize small amounts of BMP2 on the glass substrates using an aminosilane linker and different silica coatings. According to the ELISA test, a native glass surface modified with the aminosilane linker bound about 5 ng cm2 BMP2. Use of the unstructured or nanoporous silica coating increases these values to about 13–15 ng cm2 BMP2. In contrast, without the amino modification no BMP2 was bound. In comparison, the amounts of biologically active BMP2 detected by the BRE-luc test were considerably lower, about 2 ng cm2 for the amino-modified unstructured and 5 ng cm2 for the amino-modified nanoporous glass surface. No BMP2 could be detected on the uncoated but amino-modified glass surface using this method. Also, in the absence of aminosilane-derived linking groups no notable amounts of BMP2 were observed in the BRE-luc experiments, in line with the results obtained by the ELISA test. In similar experiments with glass substrates, but using epoxy instead of amino modification, conflicting results were obtained. Whereas in the BRE-luc test the amount of BMP2 was below the detection limit of < 1 ng cm2 in all cases, the ELISA detected

1776

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

Fig. 3. Scanning electron microscopy images of glass disks and of BioveritÒ II substrates coated with nanoporous silica, reflecting the state before and after the washing and incubation treatment involved in the immobilization of BMP2. White arrows mark cracks in the nanoporous silica layer.

experiment was repeated twice for the nanoporous surface. In both cases even higher amounts of 156 and 160 ng cm2 bound BMP2 were detected. All tests carried out for comparison (omitting the silane or the silica layer or both) gave values below 4 ng cm2 BMP2. In contrast, the ELISA test was unable to detect any BMP2, except for the uncoated amino-modified surface, on which 24 ng cm2 was observed. Obviously, this test is affected by the presence of BioveritÒ II, for unknown reasons. 3.3. BMP2 release from BioveritÒ II surfaces

Fig. 4. X-ray diffraction patterns of a nanoporous silica layer on glass. Samples were exposed to 10 vol.% FCS for different time periods.

58 ng cm2 in the case of the unstructured and 46 ng cm2 in the case of the nanoporous silica surface when these were epoxy modified. We were able to reveal by further experiments that this discrepancy is possibly caused by a direct binding of antibody molecules to the epoxy functions of the silanized surface, despite previous blocking with FCS. Due to the low biological activity detected by the cellular test system this chemical modification strategy was abandoned. Considerably higher amounts of immobilized BMP2 could be achieved on BioveritÒ II substrates compared with the glass surfaces. The bioactivity test detected about 67 ng cm2 for the amino-modified unstructured and more than 100 ng cm2 for the amino-modified nanoporous surface. To validate this result, the

As described above, the nanoporous silica coating was stable during the treatments used to immobilize BMP2. However, in cell culture medium the pore system in the silica layer is destroyed and the layer is in part removed. In order to clarify whether BMP2 is released from the support during this deterioration process release experiments were carried out with BioveritÒ II samples with an aminosilane-modified nanoporous layer. Samples were placed in cell culture medium and aliquots of the supernatants removed after defined time periods. The amount of free BMP2 in these supernatants was then determined in the absence of the supports. While no free BMP2 could be detected up to 2 days, after 3 and 5 days about 5% of the originally bound BMP2 was recovered in the supernatants. In order to demonstrate correct functioning of the test even in the presence of silica layer components in the medium some supernatants were spiked with 100 ng BMP2. These values were correctly detected. These results show that BMP2 immobilized by the method presented here will be released from the surface in small amounts, starting around day 3 of incubation in medium. 4. Discussion The present study has investigated the immobilization of BMP2 on ceramic bone replacement materials. In this regard we have addressed two main topics, developing a simple method of immobilizing BMP2 on the surface of silica-based materials and

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

Fig. 5. Amounts of BMP2 detected on different supports based on glass (top) and on BioveritÒ II (bottom) substrates. Plain surfaces, surfaces with an unstructured and with a nanoporous silica layer, either with no modification or with a modification derived from 3-aminopropylsilane, were employed. The results from ELISA and BRE-luc tests are given. ⁄, No detectable BMP-2 compared with the blank value; note the different scales of the ordinates.

determining whether a layer of nanoporous (or mesoporous) silica material applied to a ceramic surface enhances the amount of BMP2 immobilized. We decided to use a simple silanization method for the immobilization procedure and tested two different functional groups in this respect. Tests on silica-coated glass and ceramic surfaces showed that functionalizing the surface by aminopropyl silanization is a successful method to bind BMP2 to a silica surface. This result is in line with those of Jennissen and co-workers [9]. They were able to immobilize BMP2 on titanium surfaces using a rather elaborate and laborious method. Titanium surfaces treated first with chromosulfuric acid were then refluxed in a solution of aminopropyl triethoxysilane in toluene under an inert gas atmosphere for several hours. BMP2 was coupled to this surface either directly or employing carbonyldiimidazole (CDI) as an intermediate agent. In comparison with this procedure, our route for the modification of the silica surface consists of a simple dipping method carried out in aqueous solution, taking 2 min. With this method we were able to consistently immobilize 100–160 ng cm2 BMP2. In the work of Jennissen and co-workers 596 ng cm2 bound BMP2 could be achieved without the use of CDI as a coupling agent, and 819 ng cm2 when CDI was employed [9]. In this work determina-

1777

tion of the amount of BMP2 immobilized was based on radiolabelled [125I]-BMP2. This method detects all BMP2 molecules on the surface, whether they are immunologically and biologically active or not. In contrast, we can safely conclude that the values given for the amount of immobilized BMP2 on our samples denotes biologically active protein, since the BRE-luc test employed relies on the activation of intracellular protein signalling cascades, which are only induced by biologically active BMP2. In fact, our samples may contain more immobilized BMP2 than detected by the BRE-luc test, but these BMP2 molecules are biologically inactive. Jennissen and co-workers further showed that at least some of the immobilized BMP2 on their samples was biologically effective in vivo. In animal experiments on dogs they found that samples with immobilized BMP2 could induce bone formation around a dental implant [9]. As another option for the immobilization of BMP2 on biomaterials we cite the work of Matzuzaka et al., who detected 44.2 ng cm2 BMP2 on plasma-activated polystyrene [39], determined by a frequency shift method, and that of Park et al., who bound 800 ng cm2 BMP2 to a chitosan matrix using the crosslinking agent SMCC [31], detected by radiolabelling. The type of binding between the substrate and BMP2 is not clear. We first consider the interaction between the silica layer and the aminosilane. In order to covalently bind aminopropylsilyl functions by establishing siloxane bonds the material has to be heated [62,63], which is not part of our dipping procedure. It is more probable that a layer of intermolecularly condensed aminopropylsilane is formed and connected to the silanol groups on the silica surface by hydrogen bonds to its amino residues [63]. With regard to the interactions between the aminopropylsilyl linkers and BMP2, several binding modes are possible. In neutral or slightly acidic solutions (the pH of the MES buffer used for coupling is 5.0) the amino groups will be partially protonated and electrostatic interactions can occur with negatively charged or polarized elements in the BMP2 molecules. Alternatively, it has also been proposed that BMP2 interacts strongly with alkyl residues placed on a surface by hydrophobic interactions [33]. Such interactions could occur with the propyl residues of our linking agent or with hydrophobic parts of the silica surface (i.e. areas which contain only fully interlinked [SiO4/2] tetrahedra). The formation of amide bonds between the surface amino groups and the carboxylic groups of BMP2 appears improbable. Jennissen and co-workers postulated that such true covalent binding occurs after activation with CDI, so that a linkage is formed between surface amino groups and e-amino groups of lysine residues within the protein [8]. The question of whether surface-immobilized BMP2 can be biologically effective is currently under debate. Although some positive results have been reported for immobilized BMP2 [9,30,32,39,41], a recent study found that the BMP2 has to be reversibly coupled to the surface and needs to be released into solution to be biologically effective [40]. The Shiba group conducted experiments on the release of immobilized BMP2 from titanium and found more prominent activation of BMP signalling and induction of differentiation in cells in vitro on samples with reversibly bound BMP2. However, the BMP2 used was genetically modified. In contrast, Shi et al. [36] were able to immobilize BMP2 on dextran-grafted titanium surfaces which not only showed significantly higher osteoblast spreading, alkaline phosphatase activity and calcium mineral deposition, but also reduced bacterial adhesion due to the dextran-grafting. In our case we were able to demonstrate that the deposited BMP2 is slowly released from the BioveritÒ II surface: 5% of the BMP2 was found in the supernatant after 3 and 5 days. It is not clear whether BMP2 release is caused by dissolution of the nanoporous silica coating or by direct detachment of the BMP2, with or without the linker. Apparently the BMP2 is strongly fixed to the support in our case, as otherwise it

1778

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779

would no longer be present after the extensive washing procedures, and in all cases where no aminosilane linker was used no significant BMP2 could be detected. On the other hand, the stability of the nanoporous silica layer under cell culture conditions appears to be quite low, as has been discussed in the recent literature [61]. However, the stability of nanoporous silica films crucially depends on the preparation and calcination conditions and on further modifications. Our own results described above indicate a somewhat higher stability to that described by Bass et al. [61]. Especially, the nanostructural features evident by XRD may disappear before the silica layer has been dissolved. In this context we note that our studies relating to the stability of the nanoporous silica layer were carried out on glass samples which were coated only once. In contrast, the BioveritÒ II substrates were coated three times with a nanoporous film. This thicker layer should show longer persistence. This assumption is in good agreement with the results from the release experiments, where BMP2 was found in solution starting after 3 days incubation. As the BRE-luc assay lasts only 2 days we only detect immobilized BMP2 on BioveritÒ II surfaces. For longer time periods deterioration of the support itself can liberate BMP2 molecules into solution. The system developed here would then function as a controlled release system where disappearance of the pore system (but persistence of the silica coating layer) will also have a decisive influence on release. With regard to this topic, we can also state that for BMP2 immobilization on BioveritÒ II surfaces using the method developed here a nanoporous silica coating (to which 100–160 ng cm2 can be bound) appears superior to an unstructured silica coating (on which ca. 67 ng cm2 were immobilized). These results are in line with an investigation with ALP on similar silica surfaces [48] and can, on the one hand, be traced back to the higher surface area of the nanoporous film and to the larger number of silanol groups present on this material [64]. However, breakdown of the pore system, which may be accompanied by the inclusion of protein molecules into the remaining silica coating, could induce specific release kinetics. With regard to the structural influence of the substrate, our study also shows that the amounts of BMP2 bound to BioveritÒ II-based samples are much larger than those on glassbased samples, provided a silica layer and an aminosilane linker are applied. This could be for several reasons. Firstly, the effective surface area per square centimetre is larger for the BioveritÒ II material compared with glass due to its roughness in the micrometre range (cf. Fig. 3). Secondly, the applied silica coating is thicker, as it was prepared by three subsequent dipping procedures in order to cover the whole surface. ELISA tests are a well-established method to determine the concentrations of proteins in solution with high specificity. However, our results show that ELISA results have to be evaluated carefully when solid surfaces (of implant materials) come into play. When the ELISA for BMP2 and the BRE-luc test results conflict, the latter are likely to be more reliable. Also, the BRE-luc test provides evidence for biologically active BMP2 and is very sensitive, since concentrations as low as 1 ng BMP2 per sample could be detected. This contrasts favourably with other methods for the detection of BMP2-induced biological activity, such as, for example, determination of ALP activity. Other methods which have been employed so far detect either immunologically active BMP2 (ELISA), all BMP2 present (radiolabelling methods) [9] or simply a mass increase (frequency shift method) [39]. By adapting the originally developed BRE-luc assay [56] for the detection of BMP2 on implant surfaces it is even possible to reduce the number of animal experiments. Nevertheless, to clarify whether the immobilized BMP2 can fulfil its role in living systems and can induce bone formation in vivo the samples have to be tested in animal experiments. These are currently being carried out, in both mice and rabbits.

5. Conclusions We have developed a simple, fast and effective binding system for the immobilization of BMP2, based on the established bone reconstruction material BioveritÒ II. Coating of BioveritÒ II with a nanoporous silica layer and subsequent modification with 3-aminopropyl-trimethoxysilane reliably gives BMP2-functionalized surfaces which carry 100–160 ng cm2 of biologically active BMP2. The immobilized amount of biologically active BMP2 is judged by the highly BMP2-specific BRE-luc test which helps to reduce the number of animal experiments. Small amounts of the immobilized BMP2 are set free after about 3 days of incubation in cell culture medium. It is noteworthy that the BMP2 coupling procedure employed is so simple that it could potentially also be carried out by a surgeon in the operating theatre. This could be advantageous to avoid long-term storage of coupled BMP2 or when implants have to be reshaped during operation, making it necessary to carry out the BMP2 functionalization on the ‘‘fresh’’ implant surface.

Acknowledgements This work was supported by the DFG within the Collaborative Research Program SFB 599 ‘‘Sustainable Bioresorbable and Permanent Implants Based on Metallic and Ceramic Materials’’. We thank our colleagues in work packages D1 (‘‘Functionalized Middle Ear Prostheses’’) and D7 (‘‘Implant surfaces’’) for valuable discussions.

References [1] Hosseinkhani H, Hosseinkhani M, Khademhosseini A, Kobayashi H. Bone regeneration through controlled release of bone morphogenetic protein-2 from 3-D tissue engineered nano scaffold. J Control Release 2007;117:380–6. [2] Jeon O, Song SJ, Kang SW, Putnam AJ, Kim BS. Enhancement of ectopic bone formation by bone morphogenetic protein-2 released from heparin-conjugated poly(L-lactid–co-glycolic acid) scaffold. Biomaterials 2007;28:2763–71. [3] Saito N, Okada T, Horiuchi H, Murakami N, Ota H, Miyamoto S, et al. Biodegradable poly-d, l-lactic acid–polyethylene glycol block copolymer as a BMP delivery system for inducing bone. J Bone Joint Surg 2001;83:S92–8. [4] Sotome S, Uemura T, Kikuchi M, Chen SI, Tanaka J, Tateishi T, et al. Synthesis and in vivo evaluation of a novel hydroxyapatite/collagen–alginate as a bone filler and a drug delivery carrier of morphogenetic protein. Mater Sci Eng C 2004;24:341–7. [5] Suzuki A, Terai H, Toyoda H, Namikawa H, Yokota Y, Tsunoda T, et al. A biodegradable delivery system for antibiotics and recombinant human bone morphogenetic protein-2: a potential treatment for infected bone defects. J Orthop Res 2006;24:327–32. [6] Hashimoto Y, Yoshida G, Toyoda H, Takaoka K. Generation of tendon-to-bone interface ‘‘enthesis’’ with use of recombinant BMP-2 in a rabbit model. J Orthop Res 2007;25:1415–24. [7] Chen B, Lin H, Wang J, Zhao Y, Wang B, Zhao W, et al. Homogenous osteogenesis and bone regeneration by demineralized bone matrix loading with collagen-targeting bone morphogenetic protein-2. Biomaterials 2007;28:1027–35. [8] Voggenreiter G, Hartl K, Assenmacher S, Chatzinikolaidou M, Jennissen HP. Biologische Beschichtung von Implantaten mit rhBMP-2. Chirurgisches Forum 2001;30:453–5. [9] Becker J, Kirch A, Schwarz F, Chatzinkolaidou M, Rothamel D, Lekovic V, et al. Bone apposition to titanium implants biocoated with recombinant human bone morphogenetic protein-2 (rhBMP-2). A pilot study in dogs. Clin Oral Invest 2006;10:217–24. [10] Lichtinger TK, Müller RT, Schürmann N, Oldenburg M, Wiemann M, Chatzinikolaidou M, et al. Osseointegration of titanium implants by addition of recombinant bone morphogenetic protein-2 (rhBMP-2). Mat-wiss Werkstofftechn 2001;32:937–41. [11] Vögelin E, Jones NF, Huang JI, Brekke JH, Liebermann JR. Healing of a criticalsized defect in the rat femur with use of a vascularized periosteal flap, a biodegradable matrix, and bone morphogenetic protein. J Bone Joint Surg 2005;87:1323–31. [12] Seeherman HJ, Azari K, Bidic S, Rogers L, Jian X, Hollinger JO, et al. RhBMP-2 delivered in a calcium phosphate cement accelerates bridging of critical-sized defects in rabbit radii. J Bone Joint Surg 2006;88:1553–65. [13] Toriumi DM, O’Grady K, Horlbeck DM, Desai D, Turek TJ, Wozney J. Mandibular reconstruction using bone morphogenetic protein 2: long-term follow-up in a canine model. Laryngoscope 2009;109:1481–9.

N. Ehlert et al. / Acta Biomaterialia 7 (2011) 1772–1779 [14] Bouxsein ML, Turek TJ, Blake CA, D’Augusta DD, Li X, Stevens M, et al. Recombinant human bone morphogenetic protein-2 accelerates healing in a rabbit ulnar osteotomy model. J Bone Joint Surg 2001;83:1219–30. [15] Bragdon CR, Doherty AM, Rubash HE, Jasty M, Jian Li X, Seeherman H, et al. The efficacy of BMP-2 to induce bone ingrowth in a total hip replacement model. Clin Orthop 2003;417:50–61. [16] Wahl SM. Transforming growth factor b: the good, the bad, and the ugly. J Exp Med 1994;180:1587–90. [17] Reddi AH. Role of morphogenetic proteins in skeletal tissue engineering and regeneration. Nat Biotechnol 1998;16:247–52. [18] Schmidmaier G, Wildemann B, Ostapowicz D, Kandziora F, Stange R, Haas NP, et al. Long-term effects of local growth factor (IGF-I and TGF-b1) treatment on fracture healing: a safety study for using growth factors. J Orthop Res 2004;22:514–9. [19] Takahashi Y, Yamamoto M, Tabata Y. Enhanced osteoinduction by controlled release of bone morphogenetic protein-2 from biodegradable sponge composed of gelatine and b-tricalcium phosphate. Biomaterials 2005;26:4856–65. [20] Ziegler J, Anger D, Krummenauer F, Breitig D, Fickert S, Guenther KP. Biological activity of recombinant human growth factors released from biocompatible bone implants. J Biomed Mater Res A 2007;86A:89–97. [21] Oest ME, Dupont KM, Kong HJ, Mooney DJ, Guldberg RE. Quantitative assessment of scaffold and growth factor-mediated repair of critically sized bone defects. J Orthop Res 2007;25:941–50. [22] Yamachika E, Tsujigiwa H, Shirasu N, Ueno T, Sakata Y, Funkunaga J, et al. Immobilized recombinant human bone morphogenetic protein-2 enhances the phosphorylation of receptor-activated smads. J Biomed Mater Res A 2009;88A:599–607. [23] Yuan Q, Lu H, Tang S, Liu K, Pan Z, Pan H, et al. Ectopic bone formation in vivo induced by a novel synthetic peptide derived from BMP-2 using porous collagen scaffolds. J Wuhan University Technology Mater Sci Ed 2007;22:701–5. [24] Susuki Y, Tanihara M, Suzuki K, Saito A, Sufan W, Nishimura Y. Alginate hydrogel linked with synthetic oligopeptide derived from BMP-2 allows ectopic osteoinduction in vivo. J Biomed Mater Res 2000;50:405–9. [25] Saito A, Suzuki Y, Ogata SI, Ohtsuki C, Tanihara M. Prologed ectopis calcification induced by BMP-2derived synthetic peptide. J Biomed Mater Res 2004;70A:115–21. [26] Saito A, Suzuki Y, Kitamura M, Ogata SI, Yoshihara Y, Masuda S, et al. Repair of 20-mm long rabbit radial bone defects using BMP-derived peptide combined with an a-tricalcium phosphate scaffold. J Biomed Mater Res 2006;77A:700–6. [27] Niedhart C, Maus U, Redmann E, Schmidt-Rohlfing B, Niethard FU, Siebert CH. Stimulation of bone formation with an in situ setting tricalcium phosphate/ rhBMP-2 composite in rats. J Biomed Mater Res 2003;65A:17–23. [28] Liu Y, Huse RO, de Groot K, Buser D, Hunziker EB. Delivery mode and efficacy of BMP-2 in association with implants. J Dent Res 2007;86:84–9. [29] Aebli N, Stich H, Schawalder P, Theis JC, Krebs J. Effects of bone morphogenetic protein-2 and hyaluronic acid on the osseointegration of hydroxyapatitecoated implants: an experiment study in sheep. J Biomed Mater Res 2005;73A:295–302. [30] Seol YJ, Park YJ, Lee SC, Kim KH, Lee JY, Kim TI, et al. Enhanced osteogenic promotion around dental implants with synthetic binding motif mimicking bone morphogenetic protein (BMP)-2. J Biomed Mater Res A 2006;77A:599–607. [31] Park YJ, Kim KH, Lee JY, Ku Y, Lee SJ, Min BM, et al. Immobilization of bone morphogenetic protein-2 on a nanofibrous chitosan membrane for enhanced guided bone regeneration. Biotechnol Appl Biochem 2006;43:17–24. [32] Klee D, Böing J, Höcker H. Surface modification of titanium for improvement of the interfacial biocompatibility. Mat-wiss Werkstofftechn 2004;35:186–91. [33] Chatzinikolaidou M, Laub M, Rumpf H, Jennissen HP. Biocoating of electropolished and ultra-hydrophilic titanium and cobalt chromium molybdenum alloy surfaces with proteins. Mat-wiss u Werkstofftechn 2002;33:720–7. [34] Jennissen HP. Ultra-hydrophilic transition metals as histophilic biomaterials. Macromol Symp 2005;225:43–69. [35] Adden N, Gamble LJ, Castner DG, Hoffmann A, Gross G, Menzel H. Phosphonic acid monolayers for binding of bioactive molecules to titanium surfaces. Langmuir 2006;22:8197–204. [36] Shi Z, Neoh KG, Kang E-T, Poh C, Wang W. Titanium with surface-grafted dextran and immobilized bone morphogenetic protein-2 for inhibition of bacterial adhesion and enhancement of osteoblast function. Tissue Eng A 2009;15:417–26. [37] Steinmüller-Nethl D, Kloss FR, Najam-Ul-Haq M, Rainer M, Larsson K, Linsmeier C, et al. Strong binding of bioactive BMP-2 to nanocrystalline diamond by physisorption. Biomaterials 2006;27:4547–56. [38] Kloss F, Gassner R, Preiner J, Ebner A, Larsson K, Hächl O, et al. The role of oxygen termination of nanocrystalline diamond on immobilisation of BMP-2 and subsequent bone formation. Biomaterials 2008;29:2433–42.

1779

[39] Matzuzaka K, Yoshinari M, Kokubu E, Shimono M, Inoue T. Behavior of osteoblast-like cells on fibronectin or BMP-2 immobilized surface. Biomed Res 2004;25:263–8. [40] Kashiwagi K, Tsuji T, Shiba K. Directional BMP-2 for functionalization of titanium surfaces. Biomaterials 2009;30:1166–75. [41] Voggenreiter G, Hartl K, Assenmacher S, Chatzinikolaidou M, Rumpf HM, Jennissen HP. Assessment of the biological activity of chemically immobilized rhBMP-2 on titanium surfaces in vivo. Mat-wiss Werkstofftech 2001;32:942–8. [42] Turck C, Brandes G, Krueger I, Behrens P, Lenarz T, Stieve M. Histological evaluation of novel ossicular chain replacement prostheses: an animal study in rabbits. Acta Otolaryngology 2007;27:801–8. [43] Vogt JC, Brandes G, Krueger I, Behrens P, Nolte I, Lenarz T, et al. A comparison of different nanostructured biomaterials in subcutaneous tissue. J Mater Sci Mater Med 2008;19:2629–36. [44] Vogt JC, Brandes G, Ehlert N, Behrens P, Nolte I, Mueller PP, et al. Free Bioverit Ò II implants coated with a nanoporous silica layer in a mouse ear model–a histological study. J Biomater Appl 2009;24:175–91. [45] Sanchez C, Boissière C, Grosso D, Laberty C, Nicole L. Design, synthesis, and properties of inorganic and hybrid thin films having periodically organized nanoporosity. Chem Mater 2008;20:682–737. [46] Izquierdo-Barba I, Ruiz-González L, Doadrio JC, González-Calbet JM, ValletRegí M. Tissue regeneration. A new property of mesoporous materials. Solid State Sci 2005;7:983–9. [47] Izquierdo-Barba I, Colilla M, Vallet-Regí M. Nanostructured mesoporous silicas for bone tissue regeneration. J Nanomater 2008. doi: 10.1155/2008/106970. [48] Ehlert N, Mueller PP, Stieve M, Behrens P. Immobilization of alkaline phosphatase on mesoporous silica. Microporous Mesoporous Mater 2010;131:51–7. [49] Krueger I. Auf dem Weg zu neuartigen Biomaterialien: nanoporöse Beschichtungen und organisch-anorganisch Kompositwerkstoffe. PhD thesis, Universität Hannover, Germany, 2006. [50] Vallet-Regí M. Revisiting ceramics for medical applications. Dalton Trans 2006:5211–20. [51] Arcos D, Izquierdo-Barba I, Vallet-Regí M. Promising trends of bioceramics in the biomaterials field. J Mater Med 2009;20:447–55. [52] De Aza PN, De Aza AH, Pena P, De Aza S. Bioactive glasses and glass-ceramics. Bol Soc Esp Ceram V 2007;46:46–55. [53] Beleites E, Schneider G, Fried W, Schumann D, Linß W. 3-DReferenzimplantate für den Gesichts- und Hirnschädel. Deutsches Ärzteblatt 2001;98:A244–8. [54] Yamada T, Zhou H, Uchida H, Honma I, Katsube T. Experimental and theoretical NOx physisorption analyses of mesoporous film (SBA-15 and SBA-16) constructed surface photo voltage (SPV) sensor. J Phys Chem B 2004;108:13341–6. [55] Vallejo LF, Brokelmann M, Marten S, Trappe S, Cabrera-Crespo J, Hoffmann A, et al. Renaturation and purification of bone morphogenetic protein-2 produced as inclusion bodies in high-cell-density cultures of recombinant Escherichia coli. J Biotech 2002;94:185–94. [56] Jennissen HP, Chatzinikolaidou M, Rumpf HM, Lichtinger T, Müller R. Modification of metal surfaces and biocoating of implants with bone morphogenetic protein 2 (BMP-2). In: DVM-Bericht 313. Berlin: DVM; 2000. p. 127–40. [57] Logeart-Avramoglou D, Bourguignon M, Oudina K, Ten Dijke P, Petite H. Assays for the determination of biologically active bone morphogenetic proteins using cells transfected with an inhibitor of differentiation promoter–luciferase construct. Anal Biochem 2006;349:78–86. [58] Korchynskyi O, ten Dijke P. Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter. J Biol Chem 2002;277:4883–91. [59] Andersson J, Rosenholm J, Areva S, Linden M. Influence of material characteristics on ibuprofen drug loading and release profiles from ordered micro- and mesoporous silica matrices. Chem Mater 2004;16:4160–7. [60] Ehlert N, Badar M, Christel A, Lohmeier SJ, Luessenhop T, Behrens P. Mesoporous silica coatings for controlled release of the antibiotic ciprofloxacin from implants. J Mater Chem 2011. doi: 10.1039/C0JM01487G. [61] Bass JD, Grosso D, Boissiere C, Belamie E, Coradin T, Sanchez C. Stability of mesoporous oxide and mixed metal oxide materials under biologically relevant conditions. Chem Mater 2007;19:4349–56. [62] Howarter JA, Youngblood JP. Optimization of silica silanization by 3aminopropyltriethoxysilane. Langmuir 2006;22:11142–7. [63] Howarter JA, Youngblood JP. Surface modification of polymers with 3aminopropyltrimethoxysilane as a general pretreatment for controlled wettability. Macromolecules 2007;40:1128–32. [64] Heinroth F, Münnekhoff R, Panz C, Schmoll R. The Sears number as a probe for the surface chemistry of porous silicas: precipitated, pyrogenic and ordered mesoporous silicas. Microporous Mesoporous Mater 2008;116:95–100.