Home
Search
Collections
Journals
About
Contact us
My IOPscience
Genetically engineered bacteriophage delivers a tumor necrosis factor alpha antagonist coating on neural electrodes
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Biomed. Mater. 9 015009 (http://iopscience.iop.org/1748-605X/9/1/015009) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.93.16.3 This content was downloaded on 29/05/2016 at 21:58
Please note that terms and conditions apply.
Biomedical Materials Biomed. Mater. 9 (2014) 015009 (13pp)
doi:10.1088/1748-6041/9/1/015009
Genetically engineered bacteriophage delivers a tumor necrosis factor alpha antagonist coating on neural electrodes Young Jun Kim 1,5 , Young-Hyun Jin 2,5 , Georgette B Salieb-Beugelaar 3,4 , Chang-Hoon Nam 1 and Thomas Stieglitz 2 1 Laboratory of Nanomedicine, Korea Institute of Science and Technology Europe (KIST-Europe) Forschungsgesellschaft mbH, Campus E 7 1, Saarbruecken, Germany 2 Laboratory for Biomedical Microtechnology, Department of Microsystems Engineering (IMTEK), University of Freiburg, Georges-Koehler-Allee 102, Freiburg, Germany 3 Medical Intensive Care Unit (MIPS), Research Group Nanomedicine, University Hospital Basel, Petersgraben 4, CH-4031 Basel, Switzerland 4 European Foundation for Clinical Nanomedicine (CLINAM), Alemannengasse 12, CH-4016 Basel, Switzerland
E-mail: [email protected], [email protected] and [email protected] Received 9 July 2013, revised 6 December 2013 Accepted for publication 16 December 2013 Published 21 January 2014 Abstract
This paper reports a novel approach for the formation of anti-inflammatory surface coating on a neural electrode. The surface coating is realized using a recombinant f88 filamentous bacteriophage, which displays a short platinum binding motif and a tumor necrosis factor alpha antagonist (TNF-α antagonist) on p3 and p8 proteins, respectively. The recombinant bacteriophages are immobilized on the platinum surface by a simple dip coating process. The selective and stable immobilization of bacteriophages on a platinum electrode is confirmed by quartz crystal microbalance with dissipation monitoring, atomic force microscope and fluorescence microscope. From the in vitro cell viability test, the inflammatory cytokine (TNF-α) induced cell death was prevented by presenting recombinant bacteriophage coating, albeit with no significant cytotoxic effect. It is also observed that the bacteriophage coating does not have critical effects on the electrochemical properties such as impedance and charge storage capacities. Thus, this approach demonstrates a promising anti-apoptotic as well as anti-inflammatory surface coating for neural implant applications. Keywords: platinum, neural electrode, filamentous bacteriophage, TNF-alpha antagonist, anti-inflammatory effect (Some figures may appear in colour only in the online journal)
bioelectrical signals can be recorded to obtain information from the sensory system and electrical stimulation can be delivered to target structures, for example to activate muscles, deliver sensory input to restore hearing or vision or to modulate brain functions [1, 2]. Neural electrodes, in general, consist of conducting interconnects and an electrode site, either capacitor or metal, that acts as a bidirectional transducer with respect to the bioelectric signals, and a substrate which supports and selectively insulates the conducting part. Various
Introduction A neural implant is a technical system, which interfaces parts of the human nervous system to restore sensor and motor functions that have been lost due to trauma or diseases. In neural implants, the technical system generally interfaces the biological system by miniaturized electrodes. By them, 5
Both authors contributed equally.
1748-6041/14/015009+13$33.00
1
© 2014 IOP Publishing Ltd
Printed in the UK
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Figure 1. Schematic illustration of the platinum neural electrode coated with the bacteriophage. The immobilized rf88 phage is genetically modified to display five copies of platinum binding peptide on p3 protein and 100 to 150 copies of TNF-α antagonists on a p8 protein. The insets of optical (upper row) and fluorescence microscope images (lower row) confirm the binding of the bacteriophage on the platinum surface.
Active biomolecules and anti-inflammatory drugs have also been introduced directly at the implant–tissue interface. Common strategies include injection of drugs through microchannels fabricated into the neural electrode [15] and release of drugs from the electrode surface [16, 17]. Despite the successful modulation of inflammation responses using such approaches, the long-term usability may be restricted due to the limited drug release duration. As an alternative, active biomolecules such as an alpha-melanocytestimulating hormone or interleukin-1 receptor antagonist were immobilized on silicon-based neural electrodes [18, 19] using surface chemistry for reduction of inflammation. In vivo experiments on rats showed reduced glial scarring in close proximity to these implants. The reported immobilization methodology for organic–inorganic hybridization, however, is based on surface chemistry using specific functional groups. This fact means that it is necessary to develop new and specific interfacing and binding methods for every single implant material. Recently, genetically engineered filamentous bacteriophage offered an attractive alternative to achieve various organic–inorganic hybrid immobilizations of biomolecules, especially for biomedical applications [20–22]. Among them, the f88 phage is an fd-derived phage of which a gene cassette encoding recombinant p8 protein is inserted into the fd vector. It is a member of the Ff filamentous phage family consisting of a circular single-stranded DNA (ssDNA) and a surrounding protein sheath (figure 1). The protein sheath is mainly composed of the major coat protein, p8 (∼2700–3000 copies), aligned along the ssDNA. Furthermore,
materials, such as platinum [3, 4], iridium oxide [5] and polyethylenedioxy-thiophene (PEDOT) [6], have been used as a conducting layer. For the substrate materials, polymers such as polyimide [7], polydimethylsiloxane (PDMS) [8] and parylene [9] as well as silicon (Si) [10, 11] have been widely employed. Although the first neural electrode was introduced in the medical field several decades ago, the requirements for functional and long-term stable nerve interfaces have not yet been accomplished completely. One of the limitations of the neural electrode for chronic implantation comes from the adverse tissue response elicited by the materials of neural electrodes and micro-movements between the implant and the target tissue during physiological movements due to a mechanical mismatch. This inflammatory response causes encapsulation of the neural electrode by non-conductive cell layers, which results in electrical isolation of the implant from the target neurons. Various approaches have been vigorously pursued to inhibit the inflammatory tissue response around the implanted neural electrodes. In a material science point of view, modifying the dimension and shape of the electrodes [12] and using alternative materials have been tried. However, the effectiveness of these approaches has not been considerable. Passive biomolecules such as laminin [13, 14] have been coated on the surface of neural electrodes to attenuate the foreign body responses. These coatings have shown improved tissue integration. Laminin is one of the extra-cellular matrix proteins and seems to be involved in early immune response processes such as inflammation [54]. 2
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
five copies of minor coat protein, p3, are located at one end of the fd filament and five copies of p7 and p9 proteins are expressed at the other end [23]. All of these coat proteins can be used to display diverse foreign proteins and peptides on the surface of bacteriophage. According to these facts, filamentous bacteriophages have several positive aspects for hybrid immobilization of biomolecules. By selecting proper short peptide motifs on coat proteins, we can immobilize the bacteriophage on various materials. The immobilization of bacteriophages is reported in previous studies on metals [24], metal oxides [24], metal ions [25], polymeric materials [26], carbon materials [27] and semiconductors [28]. Moreover, together with material binding peptides, other active biomolecules can be displayed on a single bacteriophage. Therefore, a genetically engineered bacteriophage mediates a coating of various active biomolecules on diverse engineering materials. Regarding the biocompatibility of bacteriophages for the medical implant applications, current in vivo clinical trials [29–32] support that the use of phages at even high doses does not have an acute toxic effect and does not provoke an immune response. In addition, Rong and colleagues [33] reported that thin sheets of aligned bacteriophages could guide the direction of mammalian cell growths without any cytotoxic effects. Previous reports [25, 34, 35] showed that self-assembled bacteriophages display functionalities including cell adhesion, proliferation and differentiation. For example, Merzlyak et al [36] reported that bacteriophage assemblies were capable of supporting both proliferation and differentiation of neural progenitor cells. Additionally, the direct orientation of the growth in three dimensions of the neural progenitor cells was supported by these bacteriophage assemblies [36]. As another example, Zhu et al [35] described proliferation and differentiation of mesenchymal stem cells on the grooved bacteriophage films. In this study, we have chosen to investigate the capabilities of modified f88 phage (fd phage) as an anti-inflammatory surface coating on platinum neural electrodes. For the anti-inflammatory functionality, tumor necrosis factor alpha antagonist (TNF-α antagonist) peptides are displayed on the recombinant p8 coat protein of the fd phage. TNF-α is a potent cytokine and the main key regulator for downstream signals that exerts pleiotropic functions, including immunity, inflammation and apoptosis. TNF-α induced cell death has been extensively studied as a model of TNF-α receptormediated cell signaling. For example, TNF-α produced by activated macrophages binds to TNF receptors and triggers one of the intracellular events that result in apoptosis induced cell death. TNF-α antagonist, the WP9QY peptide, is a nonsteroidal anti-inflammatory cyclic peptide widely used for mild to moderate pain relief and in the treatment of osteo- and rheumatoid arthritis [37, 38]. The TNF-α antagonist displayed phage was immobilized on the platinum electrode using platinum binding peptides displayed on the p3 protein. The schematic illustration in figure 1 shows the platinum neural electrode coated with recombinant f88 phage (rf88 phage). Compared to the previous drug injecting and drug releasing approaches, our approach may provide a longer duration of the antiinflammatory effect with high stability. Different from the
previous surface chemistry-based immobilization methods, the proposed approach has a high potential to be used on various electrode materials due to the diverse kinds of binding peptides of bacteriophages, as reported in the literature [39]. The selectivity of the binding peptides might offer a good opportunity for the simple patterning of the peptide without conventional patterning processes like photolithography or embossing (stamping) technique. In addition, another kind of anti-inflammatory biomolecule such as IL-1β antagonist can be displayed on phage as well. We investigated the immobilization of the rf88 phage on the platinum surface using quartz crystal microbalance with dissipation monitoring (QCM-D) measurement and atomic force microscopy (AFM). The anti-inflammatory effect of the rf88 phage was studied by observing TNF-α induced apoptotic cell death. Furthermore, in vitro biocompatibility of the rf88 phage and influence of the coating on the electrochemical properties of the neural electrode were investigated. Materials and methods Fabrication of neural microelectrodes
The neural electrodes used in this study were manufactured using standard micromachining techniques [3, 7] in a class 1000 cleanroom. The schematic of the fabrication process is shown in figure 2. Silicon wafers were cleaned first by immersion in acetone for 10 min with sonication, removed with tweezers and rinsed with deionized water, then sonicated in isopropanol for 10 min followed by rinsing with high purity deionized water and dried with N2. The fabrication process starts with the spin coating of a 5 μm-thick polyimide (U-Varnish-S, UBE Industries Ltd, Tokyo, Japan) layer on a silicon wafer (figure 2, step 1). A 300 nm-thick platinum layer was sputter deposited (Leybold Univex 500, Leybold GmbH, Germany) and patterned using a lift-off process (figure 2, step 2). Before the platinum deposition, oxygen plasma treatment was carried out to enhance the adhesion of the platinum to the polyimide layer. Following the lift-off process, the wafer was again exposed to oxygen plasma and coated with a second layer of polyimide (5 μm), as shown in step 3 of figure 2. Then, the active platinum sites were opened and device perimeters were etched using a reactive ion etching (RIE) process (figure 2, step 4). A two-step etching technique with a single masking layer was applied in this step. The radio frequency (RF) power of the RIE process was reduced after the opening of active sites to protect the exposed platinum layer and the reduced RF power was maintained until the device perimeters were etched down to the supporting wafer. After the micromachining process, single devices were peeled off from the supporting wafer. The electrodes were stored in a clean environment for at least three days before bacteriophage coating. The bacteriophage coating was carried out by submerging the fabricated electrode in 1 ml of bacteriophage solution (figure 2, step 5) for 1 h at room temperature. Then the electrode was washed three times with deionized water and dried in a vacuum chamber. We used electrodes with site diameters of 80 and 120 μm 3
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Figure 2. Microfabrication and bacteriophage coating process of the neural electrode. Step 1: deposition of 5 μm-thick polyimide substrate
layer; step 2: sputtering and patterning of 300 nm-thick platinum layer; step 3: deposition of 5 μm-thick polyimide insulation layer; step 4: two-step RIE process to remove the insulation layer on electrode openings and create the perimeter; step 5: dipping the fabricated electrode into the solution for the immobilization of the bacteriophage; step 6: finished bacteriophage-immobilized neural electrode.
for electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) measurements.
Each of two complementary oligonucleotide mixtures was boiled for 10 min and cooled down to room temperature to produce an adaptor DNA. Each of the annealed DNA fragments was used to clone into f88 plasmids including the appropriate cohesive end for restriction sites followed by overnight ligation at 16 ◦ C with T4 ligase (New England Biolabs, Germany). The ligated plasmid was transformed into Escherichia coli MC1061 cells. The nucleotide sequence of the inserts was confirmed by DNA sequencing (MWG operon, Germany). The Escherichia coli K91BluKan (K91BK) strain (kindly provided by Prof. Dr Georg P Smith, University of Missouri, USA) was used as host for the amplification of phage particles. Isolation of phages from the host cells was performed using the polyethyleneglycol/sodium chloride (PEG/NaCl) precipitation method as described elsewhere [56] after growing the transformed K91BK cells in Luria broth supplemented with tetracycline (20 μg ml−1) and kanamycin (100 μg ml−1) at 37 ◦ C with vigorous shaking (260 rpm) overnight. Final phage pellets were dialyzed against a 10 mM phosphate buffered saline solution (PBS) at pH 7.4 overnight to remove the remaining PEG. The concentration (colony-forming unit per milliliter, CFU ml−1) of fd phages was determined by phage titration as previously described [40]. Phage suspensions were stored at 4 ◦ C.
Preparation of bacteriophage constructs
The rf88 plasmid (kindly provided by Prof. Dr Georg P Smith, University of Missouri, USA), the N-terminus of recombinant p8 and p3, was used to introduce two genes encoding YCWSQYLCY and CPTSTGQAC for TNF-α antagonist and platinum binding peptide, respectively [37–39]. The sequences of adaptor DNA fragments were constructed using the complementary oligonucleotides as follows: 1. TNF-α antagonist oligonucleotides 5 -AGCTTTTATTGTTGGTCTCAGTATCTGTGTTAT GGTAC-3 5 -CATAACACAGATACTGAGACCAACAATAAA-3 2. Platinum binding oligonucleotides 5 -CGTGCCCGACCTCCACCGGCCAGGCCTG TGC-3 5 -GGCCGCACAGGCCTGGCCGGTGGAGGTCGGG CACGGCT-3 . As a control, Pt binding peptide (CPTSTGQAC) was introduced in f88 plasmid (f88-Pt), subsequently, TNF-α antagonist (YCWSQYLCY) was introduced for the rf88 phage. The rf88 plasmid was digested by HindIII/KpnI and SfiI/NotI (New England Biolabs, Germany) for N-terminus of recombinant p8 and p3 genes, respectively. 4
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Analysis of the mass and viscoelastic properties of the deposited layers
Cell culture and viability assay
The SH-SY5Y human neuroblastoma cells, kindly provided by Dr Bernd Fiebich at University Medical Center, University of Freiburg, were grown in Dulbecco’s modified essential R medium (DMEM) (Sigma , St. Louis, USA) containing R 10% fetal calf serum (FCS) (Cultilab , Campinas, Brazil) −1 and 50 μg mL of streptomycine (Sigma, Germany) and incubated at 37 ◦ C in a 5% CO2 atmosphere. The cells were seeded into 24-well plates (VWR, Germany) containing a round glass cover slip at a concentration of 1 × 1012 CFU ml−1 of phage mixture in a complete medium for 24 h. Next, the SH-SY5Y cells at a concentration of 1 × 104 cells ml−1 were seeded on the phages immobilized glass cover slip and platinum coated cover slip in the DMEM medium supplemented with 10% FCS with varying times (6, 12, 24 and 36 h). To evaluate the cell viability with the phages, a WST-1 assay (Roche Applied Science, Switzerland) was performed according to the manufacturer’s instructions. Different phage samples at a concentration of 1 × 1012 CFU ml−1 were treated with cell solutions for 24 h. 10 μl of TNF-α agent (0.5 μg ml−1) were mixed with 0.1 ml of growth medium and added to the cell and phage mixtures and the 200 μl of each supernatant without cells was collected and transferred to the 96-well ELISA plate for measuring cell viabilities. The difference between the absorbances at 450 nm of the medium was read on an ELISA reader (TECAN, Germany). The mixture without phages was used as a negative control. To verify the effects of rf88 phage on TNF-α induced cell apoptosis, the cells on the rf88 phage-immobilized Pt electrode were treated with 50 ng ml−1 of TNF-α and grown in a cell culture dish with a diameter of 10 cm for 24 h. Subsequently, the genomic DNA fragmentation was analyzed by apoptotic DNA ladder kit (Roche, Germany) according to the manufacturer’s instruction. Shortly, 15 μl extracted DNA samples from cells (2 × 106 cells ml−1) were loaded onto 1% agarose gels containing ethidium bromide. Gels were run at 70 V for 1 h and DNA fragmentations were visualized using UV illumination.
To monitor the immobilization of f88-Pt and rf88 phages on a platinum electrode, an increase in mass and the viscoelastic properties of the phage films were measured by a QCM-D (Q-sense, Sweden) as previously described [41]. A standard 5 MHz AT-cut Pt coated quartz crystal was induced at the resonance frequency. Information about the absorption process is obtained from changes in the resonance frequency ( f ) and the dissipation factor (D, 1 × 10−6) at overtone numbers 3, 5 and 7 for the crystals. Measurements were performed at a flow rate of 100 μl min−1 at 25 ◦ C. Phage samples were injected into the chamber after equilibration with a stable baseline buffer. To perform parallel QCM-D and AFM analysis, quartz crystals at different phage concentration were collected and visualized by AFM during the injection process. Topography visualization of phage samples was carried out with a Mobile S AFM (Nanosurf AG, Switzerland) having a 110 μm by 110 μm scanning area at the non-contact mode. Measurements were carried out using an ArrowTM Si AFM probe (Nanoworld AG, Switzerland) with a spring constant of 48 N/m and a resonance frequency of around 190 kHz. Phase and color enhanced topography images were made with XE-100 (Park Systems) at the true-non-contact mode. Image analysis was performed using Gwyddion image processing software (free software developed by Czech Metrology Institute, downloaded at http://gwyddion.net).
Electrochemical characterization
EIS and CV were performed using a three-electrode configuration, which consists of a working electrode (i.e. the device under test), a platinum counter electrode (PT 1800, Schott Instruments, Germany) and an Ag/AgCl reference electrode (3M KCl bridging electrolyte, B2820+, Schott Instruments, Germany). The configuration was operated using a potentiostat combined with a frequency response analyzer (Solartron 1260 and 1287, Solartron Analytical, UK). The software packages ZPlot (v3.1c, Scribner Associates Inc.) and CorrWare (v3.1c, Scribner Associates Inc.) were used for EIS and CV, respectively. Impedance spectra were recorded from 1 Hz to 100 kHz with an excitation amplitude of 10 mV in 0.9% NaCl solution. CV was carried out in 0.01 M PBS solution (pH 7.4) from −0.6 V to 0.8 V versus reference electrode with a scan rate of 100 mV s−1.
Fluorescence microscopy
Pt binding f88 phage (f88-Pt phage), rf88 phage and cell mixtures were imaged with an Olympus IX51 inverted microscope under the phase contrast mode with a 100 × oil objective. The images were acquired using a CCD camera (Olympus soft imaging system F view). Following an incubation time of 24 h at 25 ◦ C, cell and phage mixtures were observed by immunofluorescent staining. Each phage sample (20 μl) was first incubated in 2% bovine serum albumin/PBS solution for 1 h at 25 ◦ C. Biotin conjugated anti-fd phage antibodies (1:5000, Sigma) were added and incubated for 1 h at 37 ◦ C. After the antibody incubation, phage samples were mixed with cyanine dye-3 (Cy3)-conjugated streptavidin solution (1:1000, Sigma) and incubated at 25 ◦ C for 1 h. Cy3labeled samples were washed and immediately visualized with the Olympus IX51 inverted microscope with an excitation wavelength of 540 nm and an emission wavelength of 570 nm.
Statistics
The results of cell viability are represented as the mean and standard deviation ( ± SD) from three independent experiments, with significance of differences evaluated using ANOVA with Fisher’s post hoc comparisons. A probability of P values of 0.05 were considered significant for all tests. 5
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
Figure 3. Representative QCM-D response graphs (left side) during the bacteriophage immobilization process and AFM images (right side) of the resulting platinum electrode surface. The numbers in the QCM-D response graphs indicate the injection of solution with different bacteriophage concentrations and the corresponding surfaces are shown in the left AFM images with the same indication number. The numbers (1), (2) and (3) indicate the injection of solutions with 1 × 1010 CFU/ml, 1 × 1011 CFU/ml and 1 × 1013 CFU/ml rf88 phage concentration, respectively. In the AFM images, each scale bar indicates 2 μm.
Results
Cell cytotoxicity effect of the rf88 phage
The cell cytotoxicity effect of the rf88 phage was investigated by the observation of the cells’ behavior on the bare and rf88 phage-immobilized Pt electrodes (figure 4). In the absence of TNF-α, cells cultured on the bare Pt electrode and rf88 phageimmobilized Pt electrode showed the typical healthy fibroblast morphology with a homogeneous distribution as shown in figures 4(a) and (b). After 24 and 36 h exposure to TNF-α, cells on the bare Pt already underwent typical morphologic changes of apoptotic cell deaths such as shrunken shape (figures 4(c) and (d)). Apoptotic cell death was also clearly observed with f88-Pt phage (figures 4(e) and ( f )). In contrast, the cells on the Pt surface coated by rf88 phage did not show significant apoptotic morphological change during the same time period (figures 4(g)–(i)). In line with these results, the cell cultured with different phage samples and phage-immobilized Pt surface showed only subtle changes of cell viabilities shown in figures 5(a) and (b). Compared to the cell viability of the untreated control group (defined as 100%), the cell viability incubated with f88-Pt phage, rf88 phage, Pt electrode and rf88 modified Pt electrode was measured as 94.1 ± 2.8%, 93.4 ± 3.7%, 95.4 ± 5.3% and 92.5 ± 3.2%, respectively. The cells in both groups were treated with TNF-α, a well-known pro-inflammatory cytokine, which leads to cell apoptosis and typical oligonucleosomal DNA fragmentation. The anti-apoptotic effects against TNF-α induced cell death were also analyzed by the apoptotic DNA fragmentation kit in figure 5(c). In lanes 4 and 5 of figure 5(c), a distinct DNA ladder ranging from 150 bp (base pair) to 1 kbp was observed from the cell extract. The DNA of rf88 DNA was visible around 10 kbp in lanes 3 and 5. In the absence of rf88 phage (lane 4 in figure 5(c)), exposure of cells to TNF-α
Immobilization of rf88 bacteriophage on the Pt surface
Fluorescence microscopy images (inset in figure 1) show that the rf88 phages were successfully immobilized on Pt electrodes. As expected, selective binding of rf88 phages to the Pt surface results in high fluorescence intensity contrast between the Pt surface (circular shape) and the polyimide surface indicating that rf88 phages were selectively bound to the Pt surface in contrast to polyimide. The change in frequency ( f ) and dissipation energy (D) in the QCM-D analysis (figure 3) shows the concentration-dependent transient and saturated immobilization of the rf88 phage on the Pt surface. We injected 100 μl of rf88 phage solution with 1 × 1010 CFU/ml concentration into the flow channel (figure 3, arrow 1), followed by subsequent injection of another 100 μl rf88 phage solution with concentration of 1 × 1011 CFU/ml (figure 3, arrow 2) and concentration of 1 × 1012 CFU/ml (figure 3, arrow 3). An immediate decrease in frequency along with the increase in dissipation energy was observed, which proves the transient immobilization of the rf88 phage on the Pt surface. The deposition of rf88 phages on the surfaces was further confirmed by the corresponding AFM images (figure 3, right panel). Low phage concentration (1 × 1010 CFU/ml) leads to the non-uniform shape of phage assembly. When the phage concentration was increased, the phage assemblies became uniform, straight and thicker, which is in line with binding of rf88 to Pt surface shown in figure 1. We also observed that wild type f88 phage also bound to Pt electrode; however, the QCM-D results showed negligible frequency changes (data not shown). 6
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
(c)
(d )
(e)
(f)
( g)
(h)
(i)
Figure 4. Cytotoxicity effects of the SH-SY5Y cell on the bare platinum surface as a control after 0 h (a) and 24 h (b) and in the presence of TNF-α (50 ng ml−1) for 24 h (c) and 36 h (d). The f88-Pt phages on the SH-SY5Y cell in the presence of TNF-α (50 ng ml−1) (e) and ( f ) incubation for 24, 36 h and rf88 (g), (h), (i) incubation for 0, 24, 36 h in the presence of TNF-α (50 ng ml−1), respectively. Each scale bar indicates 100 μm size scale.
average value of the ten independent measurements (gray lines). In all cases, the results showed typical impedance spectra of thin film platinum electrodes: resistive response at a high frequency region and capacitive behavior at the low frequency region. The impedance of both the 80 μm-diameter and 120 μm-diameter electrodes increases after the phage coating process. At 1 kHz, the impedance of the 120 μmdiameter electrode, for example, increased to 131.4 k (average value) after the phage immobilization, compared to the impedance of the bare electrode (62.1 k, average value). Cyclic voltammograms of the bare Pt electrode and rf88 phage-immobilized electrode were shown in figures 7(a) and (b), respectively. As a whole, the shape of CV did not change significantly after the rf88 phage immobilization. However, the area enclosed by CV was slightly reduced and characteristic peaks also were decreased. The cathodal charge storage capacities (CSCC, time integral of the negative current in CV) were calculated as 6.07 ± 0.17 mC cm−2 for the bare Pt electrodes and 5.89 ± 0.33 mC cm−2 for the rf88 phage-coated electrodes, which yielded less than 3% reduction after the rf88 phage immobilization. Even if we exclude the electrode 1 in figure 7(b), which showed larger CV current than the other electrodes in the phage-immobilized group, the CSCC was calculated to 5.72 ± 0.12 mC cm−2, which yields around 6% reduction.
Table 1. Effect of different phage solutions on the cell viability in SH-SY5Y cells at constant TNF-α concentration.
Samples
% Cell viability
TNF-α (50 ng ml−1) alone f88-Pt (1 × 1012 CFU ml−1) + TNF-α (50 ng ml−1) rf88-Pt (1 × 1010 CFU ml−1) + TNF-α (50 ng ml−1) rf88-Pt (1 × 1012 CFU ml−1) + TNF-α (50 ng ml−1)
68 67 76 90
± ± ± ±
7 4 3 2
Cells were exposed to TNF-α (50 ng ml−1) and different phage concentrations (see also figure 5(d)). Data represent means ± SD from three independent experiments, each performed in triplicate, expressed as percentage compared to untreated cells.
resulted in well-pronounced DNA fragmentation. In contrast, the DNA fragmentation was significantly diminished in the rf88 phage-immobilized electrode as shown in lane 5, which is in line with cell viability and morphological changes. As can be seen in figure 5(d) and table 1, TNF-α treatment reduced approximately 30% of cell viability on the bare electrode and f88-Pt phage-immobilized electrode, indicating that wild type phage failed to prevent the TNF-α induced cell death. In contrast, 10–20% recovery of cell viability was observed on 1 × 1010 and 1 × 1012 CFU ml−1 of rf88 phage-coated electrode in a dose-dependent manner, respectively. Effects of rf88 phage on electrochemical properties
Discussion
Electrochemical impedance spectra were analyzed prior to the bacteriophage immobilization and after the rf88 phage coating process (figure 6). In figure 6, dark black lines indicate the
Surface modifications and synthetic material coatings of medical devices are clinically used in a variety of applications 7
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
(c) (d)
Figure 5. Comparison of cell viabilities in the absence (a) and presence of the Pt electrode (b). DNA fragmentation from the cell extracts was analyzed by 1% agarose gel electrophoresis after 24 h cell culture (c). The cell viability was determined by WST-1 analysis (d). Bars show means ± SD from three independent experiments. Data were analyzed statistically by ANOVA at a P
Search
Collections
Journals
About
Contact us
My IOPscience
Genetically engineered bacteriophage delivers a tumor necrosis factor alpha antagonist coating on neural electrodes
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Biomed. Mater. 9 015009 (http://iopscience.iop.org/1748-605X/9/1/015009) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.93.16.3 This content was downloaded on 29/05/2016 at 21:58
Please note that terms and conditions apply.
Biomedical Materials Biomed. Mater. 9 (2014) 015009 (13pp)
doi:10.1088/1748-6041/9/1/015009
Genetically engineered bacteriophage delivers a tumor necrosis factor alpha antagonist coating on neural electrodes Young Jun Kim 1,5 , Young-Hyun Jin 2,5 , Georgette B Salieb-Beugelaar 3,4 , Chang-Hoon Nam 1 and Thomas Stieglitz 2 1 Laboratory of Nanomedicine, Korea Institute of Science and Technology Europe (KIST-Europe) Forschungsgesellschaft mbH, Campus E 7 1, Saarbruecken, Germany 2 Laboratory for Biomedical Microtechnology, Department of Microsystems Engineering (IMTEK), University of Freiburg, Georges-Koehler-Allee 102, Freiburg, Germany 3 Medical Intensive Care Unit (MIPS), Research Group Nanomedicine, University Hospital Basel, Petersgraben 4, CH-4031 Basel, Switzerland 4 European Foundation for Clinical Nanomedicine (CLINAM), Alemannengasse 12, CH-4016 Basel, Switzerland
E-mail: [email protected], [email protected] and [email protected] Received 9 July 2013, revised 6 December 2013 Accepted for publication 16 December 2013 Published 21 January 2014 Abstract
This paper reports a novel approach for the formation of anti-inflammatory surface coating on a neural electrode. The surface coating is realized using a recombinant f88 filamentous bacteriophage, which displays a short platinum binding motif and a tumor necrosis factor alpha antagonist (TNF-α antagonist) on p3 and p8 proteins, respectively. The recombinant bacteriophages are immobilized on the platinum surface by a simple dip coating process. The selective and stable immobilization of bacteriophages on a platinum electrode is confirmed by quartz crystal microbalance with dissipation monitoring, atomic force microscope and fluorescence microscope. From the in vitro cell viability test, the inflammatory cytokine (TNF-α) induced cell death was prevented by presenting recombinant bacteriophage coating, albeit with no significant cytotoxic effect. It is also observed that the bacteriophage coating does not have critical effects on the electrochemical properties such as impedance and charge storage capacities. Thus, this approach demonstrates a promising anti-apoptotic as well as anti-inflammatory surface coating for neural implant applications. Keywords: platinum, neural electrode, filamentous bacteriophage, TNF-alpha antagonist, anti-inflammatory effect (Some figures may appear in colour only in the online journal)
bioelectrical signals can be recorded to obtain information from the sensory system and electrical stimulation can be delivered to target structures, for example to activate muscles, deliver sensory input to restore hearing or vision or to modulate brain functions [1, 2]. Neural electrodes, in general, consist of conducting interconnects and an electrode site, either capacitor or metal, that acts as a bidirectional transducer with respect to the bioelectric signals, and a substrate which supports and selectively insulates the conducting part. Various
Introduction A neural implant is a technical system, which interfaces parts of the human nervous system to restore sensor and motor functions that have been lost due to trauma or diseases. In neural implants, the technical system generally interfaces the biological system by miniaturized electrodes. By them, 5
Both authors contributed equally.
1748-6041/14/015009+13$33.00
1
© 2014 IOP Publishing Ltd
Printed in the UK
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Figure 1. Schematic illustration of the platinum neural electrode coated with the bacteriophage. The immobilized rf88 phage is genetically modified to display five copies of platinum binding peptide on p3 protein and 100 to 150 copies of TNF-α antagonists on a p8 protein. The insets of optical (upper row) and fluorescence microscope images (lower row) confirm the binding of the bacteriophage on the platinum surface.
Active biomolecules and anti-inflammatory drugs have also been introduced directly at the implant–tissue interface. Common strategies include injection of drugs through microchannels fabricated into the neural electrode [15] and release of drugs from the electrode surface [16, 17]. Despite the successful modulation of inflammation responses using such approaches, the long-term usability may be restricted due to the limited drug release duration. As an alternative, active biomolecules such as an alpha-melanocytestimulating hormone or interleukin-1 receptor antagonist were immobilized on silicon-based neural electrodes [18, 19] using surface chemistry for reduction of inflammation. In vivo experiments on rats showed reduced glial scarring in close proximity to these implants. The reported immobilization methodology for organic–inorganic hybridization, however, is based on surface chemistry using specific functional groups. This fact means that it is necessary to develop new and specific interfacing and binding methods for every single implant material. Recently, genetically engineered filamentous bacteriophage offered an attractive alternative to achieve various organic–inorganic hybrid immobilizations of biomolecules, especially for biomedical applications [20–22]. Among them, the f88 phage is an fd-derived phage of which a gene cassette encoding recombinant p8 protein is inserted into the fd vector. It is a member of the Ff filamentous phage family consisting of a circular single-stranded DNA (ssDNA) and a surrounding protein sheath (figure 1). The protein sheath is mainly composed of the major coat protein, p8 (∼2700–3000 copies), aligned along the ssDNA. Furthermore,
materials, such as platinum [3, 4], iridium oxide [5] and polyethylenedioxy-thiophene (PEDOT) [6], have been used as a conducting layer. For the substrate materials, polymers such as polyimide [7], polydimethylsiloxane (PDMS) [8] and parylene [9] as well as silicon (Si) [10, 11] have been widely employed. Although the first neural electrode was introduced in the medical field several decades ago, the requirements for functional and long-term stable nerve interfaces have not yet been accomplished completely. One of the limitations of the neural electrode for chronic implantation comes from the adverse tissue response elicited by the materials of neural electrodes and micro-movements between the implant and the target tissue during physiological movements due to a mechanical mismatch. This inflammatory response causes encapsulation of the neural electrode by non-conductive cell layers, which results in electrical isolation of the implant from the target neurons. Various approaches have been vigorously pursued to inhibit the inflammatory tissue response around the implanted neural electrodes. In a material science point of view, modifying the dimension and shape of the electrodes [12] and using alternative materials have been tried. However, the effectiveness of these approaches has not been considerable. Passive biomolecules such as laminin [13, 14] have been coated on the surface of neural electrodes to attenuate the foreign body responses. These coatings have shown improved tissue integration. Laminin is one of the extra-cellular matrix proteins and seems to be involved in early immune response processes such as inflammation [54]. 2
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
five copies of minor coat protein, p3, are located at one end of the fd filament and five copies of p7 and p9 proteins are expressed at the other end [23]. All of these coat proteins can be used to display diverse foreign proteins and peptides on the surface of bacteriophage. According to these facts, filamentous bacteriophages have several positive aspects for hybrid immobilization of biomolecules. By selecting proper short peptide motifs on coat proteins, we can immobilize the bacteriophage on various materials. The immobilization of bacteriophages is reported in previous studies on metals [24], metal oxides [24], metal ions [25], polymeric materials [26], carbon materials [27] and semiconductors [28]. Moreover, together with material binding peptides, other active biomolecules can be displayed on a single bacteriophage. Therefore, a genetically engineered bacteriophage mediates a coating of various active biomolecules on diverse engineering materials. Regarding the biocompatibility of bacteriophages for the medical implant applications, current in vivo clinical trials [29–32] support that the use of phages at even high doses does not have an acute toxic effect and does not provoke an immune response. In addition, Rong and colleagues [33] reported that thin sheets of aligned bacteriophages could guide the direction of mammalian cell growths without any cytotoxic effects. Previous reports [25, 34, 35] showed that self-assembled bacteriophages display functionalities including cell adhesion, proliferation and differentiation. For example, Merzlyak et al [36] reported that bacteriophage assemblies were capable of supporting both proliferation and differentiation of neural progenitor cells. Additionally, the direct orientation of the growth in three dimensions of the neural progenitor cells was supported by these bacteriophage assemblies [36]. As another example, Zhu et al [35] described proliferation and differentiation of mesenchymal stem cells on the grooved bacteriophage films. In this study, we have chosen to investigate the capabilities of modified f88 phage (fd phage) as an anti-inflammatory surface coating on platinum neural electrodes. For the anti-inflammatory functionality, tumor necrosis factor alpha antagonist (TNF-α antagonist) peptides are displayed on the recombinant p8 coat protein of the fd phage. TNF-α is a potent cytokine and the main key regulator for downstream signals that exerts pleiotropic functions, including immunity, inflammation and apoptosis. TNF-α induced cell death has been extensively studied as a model of TNF-α receptormediated cell signaling. For example, TNF-α produced by activated macrophages binds to TNF receptors and triggers one of the intracellular events that result in apoptosis induced cell death. TNF-α antagonist, the WP9QY peptide, is a nonsteroidal anti-inflammatory cyclic peptide widely used for mild to moderate pain relief and in the treatment of osteo- and rheumatoid arthritis [37, 38]. The TNF-α antagonist displayed phage was immobilized on the platinum electrode using platinum binding peptides displayed on the p3 protein. The schematic illustration in figure 1 shows the platinum neural electrode coated with recombinant f88 phage (rf88 phage). Compared to the previous drug injecting and drug releasing approaches, our approach may provide a longer duration of the antiinflammatory effect with high stability. Different from the
previous surface chemistry-based immobilization methods, the proposed approach has a high potential to be used on various electrode materials due to the diverse kinds of binding peptides of bacteriophages, as reported in the literature [39]. The selectivity of the binding peptides might offer a good opportunity for the simple patterning of the peptide without conventional patterning processes like photolithography or embossing (stamping) technique. In addition, another kind of anti-inflammatory biomolecule such as IL-1β antagonist can be displayed on phage as well. We investigated the immobilization of the rf88 phage on the platinum surface using quartz crystal microbalance with dissipation monitoring (QCM-D) measurement and atomic force microscopy (AFM). The anti-inflammatory effect of the rf88 phage was studied by observing TNF-α induced apoptotic cell death. Furthermore, in vitro biocompatibility of the rf88 phage and influence of the coating on the electrochemical properties of the neural electrode were investigated. Materials and methods Fabrication of neural microelectrodes
The neural electrodes used in this study were manufactured using standard micromachining techniques [3, 7] in a class 1000 cleanroom. The schematic of the fabrication process is shown in figure 2. Silicon wafers were cleaned first by immersion in acetone for 10 min with sonication, removed with tweezers and rinsed with deionized water, then sonicated in isopropanol for 10 min followed by rinsing with high purity deionized water and dried with N2. The fabrication process starts with the spin coating of a 5 μm-thick polyimide (U-Varnish-S, UBE Industries Ltd, Tokyo, Japan) layer on a silicon wafer (figure 2, step 1). A 300 nm-thick platinum layer was sputter deposited (Leybold Univex 500, Leybold GmbH, Germany) and patterned using a lift-off process (figure 2, step 2). Before the platinum deposition, oxygen plasma treatment was carried out to enhance the adhesion of the platinum to the polyimide layer. Following the lift-off process, the wafer was again exposed to oxygen plasma and coated with a second layer of polyimide (5 μm), as shown in step 3 of figure 2. Then, the active platinum sites were opened and device perimeters were etched using a reactive ion etching (RIE) process (figure 2, step 4). A two-step etching technique with a single masking layer was applied in this step. The radio frequency (RF) power of the RIE process was reduced after the opening of active sites to protect the exposed platinum layer and the reduced RF power was maintained until the device perimeters were etched down to the supporting wafer. After the micromachining process, single devices were peeled off from the supporting wafer. The electrodes were stored in a clean environment for at least three days before bacteriophage coating. The bacteriophage coating was carried out by submerging the fabricated electrode in 1 ml of bacteriophage solution (figure 2, step 5) for 1 h at room temperature. Then the electrode was washed three times with deionized water and dried in a vacuum chamber. We used electrodes with site diameters of 80 and 120 μm 3
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Figure 2. Microfabrication and bacteriophage coating process of the neural electrode. Step 1: deposition of 5 μm-thick polyimide substrate
layer; step 2: sputtering and patterning of 300 nm-thick platinum layer; step 3: deposition of 5 μm-thick polyimide insulation layer; step 4: two-step RIE process to remove the insulation layer on electrode openings and create the perimeter; step 5: dipping the fabricated electrode into the solution for the immobilization of the bacteriophage; step 6: finished bacteriophage-immobilized neural electrode.
for electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) measurements.
Each of two complementary oligonucleotide mixtures was boiled for 10 min and cooled down to room temperature to produce an adaptor DNA. Each of the annealed DNA fragments was used to clone into f88 plasmids including the appropriate cohesive end for restriction sites followed by overnight ligation at 16 ◦ C with T4 ligase (New England Biolabs, Germany). The ligated plasmid was transformed into Escherichia coli MC1061 cells. The nucleotide sequence of the inserts was confirmed by DNA sequencing (MWG operon, Germany). The Escherichia coli K91BluKan (K91BK) strain (kindly provided by Prof. Dr Georg P Smith, University of Missouri, USA) was used as host for the amplification of phage particles. Isolation of phages from the host cells was performed using the polyethyleneglycol/sodium chloride (PEG/NaCl) precipitation method as described elsewhere [56] after growing the transformed K91BK cells in Luria broth supplemented with tetracycline (20 μg ml−1) and kanamycin (100 μg ml−1) at 37 ◦ C with vigorous shaking (260 rpm) overnight. Final phage pellets were dialyzed against a 10 mM phosphate buffered saline solution (PBS) at pH 7.4 overnight to remove the remaining PEG. The concentration (colony-forming unit per milliliter, CFU ml−1) of fd phages was determined by phage titration as previously described [40]. Phage suspensions were stored at 4 ◦ C.
Preparation of bacteriophage constructs
The rf88 plasmid (kindly provided by Prof. Dr Georg P Smith, University of Missouri, USA), the N-terminus of recombinant p8 and p3, was used to introduce two genes encoding YCWSQYLCY and CPTSTGQAC for TNF-α antagonist and platinum binding peptide, respectively [37–39]. The sequences of adaptor DNA fragments were constructed using the complementary oligonucleotides as follows: 1. TNF-α antagonist oligonucleotides 5 -AGCTTTTATTGTTGGTCTCAGTATCTGTGTTAT GGTAC-3 5 -CATAACACAGATACTGAGACCAACAATAAA-3 2. Platinum binding oligonucleotides 5 -CGTGCCCGACCTCCACCGGCCAGGCCTG TGC-3 5 -GGCCGCACAGGCCTGGCCGGTGGAGGTCGGG CACGGCT-3 . As a control, Pt binding peptide (CPTSTGQAC) was introduced in f88 plasmid (f88-Pt), subsequently, TNF-α antagonist (YCWSQYLCY) was introduced for the rf88 phage. The rf88 plasmid was digested by HindIII/KpnI and SfiI/NotI (New England Biolabs, Germany) for N-terminus of recombinant p8 and p3 genes, respectively. 4
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
Analysis of the mass and viscoelastic properties of the deposited layers
Cell culture and viability assay
The SH-SY5Y human neuroblastoma cells, kindly provided by Dr Bernd Fiebich at University Medical Center, University of Freiburg, were grown in Dulbecco’s modified essential R medium (DMEM) (Sigma , St. Louis, USA) containing R 10% fetal calf serum (FCS) (Cultilab , Campinas, Brazil) −1 and 50 μg mL of streptomycine (Sigma, Germany) and incubated at 37 ◦ C in a 5% CO2 atmosphere. The cells were seeded into 24-well plates (VWR, Germany) containing a round glass cover slip at a concentration of 1 × 1012 CFU ml−1 of phage mixture in a complete medium for 24 h. Next, the SH-SY5Y cells at a concentration of 1 × 104 cells ml−1 were seeded on the phages immobilized glass cover slip and platinum coated cover slip in the DMEM medium supplemented with 10% FCS with varying times (6, 12, 24 and 36 h). To evaluate the cell viability with the phages, a WST-1 assay (Roche Applied Science, Switzerland) was performed according to the manufacturer’s instructions. Different phage samples at a concentration of 1 × 1012 CFU ml−1 were treated with cell solutions for 24 h. 10 μl of TNF-α agent (0.5 μg ml−1) were mixed with 0.1 ml of growth medium and added to the cell and phage mixtures and the 200 μl of each supernatant without cells was collected and transferred to the 96-well ELISA plate for measuring cell viabilities. The difference between the absorbances at 450 nm of the medium was read on an ELISA reader (TECAN, Germany). The mixture without phages was used as a negative control. To verify the effects of rf88 phage on TNF-α induced cell apoptosis, the cells on the rf88 phage-immobilized Pt electrode were treated with 50 ng ml−1 of TNF-α and grown in a cell culture dish with a diameter of 10 cm for 24 h. Subsequently, the genomic DNA fragmentation was analyzed by apoptotic DNA ladder kit (Roche, Germany) according to the manufacturer’s instruction. Shortly, 15 μl extracted DNA samples from cells (2 × 106 cells ml−1) were loaded onto 1% agarose gels containing ethidium bromide. Gels were run at 70 V for 1 h and DNA fragmentations were visualized using UV illumination.
To monitor the immobilization of f88-Pt and rf88 phages on a platinum electrode, an increase in mass and the viscoelastic properties of the phage films were measured by a QCM-D (Q-sense, Sweden) as previously described [41]. A standard 5 MHz AT-cut Pt coated quartz crystal was induced at the resonance frequency. Information about the absorption process is obtained from changes in the resonance frequency ( f ) and the dissipation factor (D, 1 × 10−6) at overtone numbers 3, 5 and 7 for the crystals. Measurements were performed at a flow rate of 100 μl min−1 at 25 ◦ C. Phage samples were injected into the chamber after equilibration with a stable baseline buffer. To perform parallel QCM-D and AFM analysis, quartz crystals at different phage concentration were collected and visualized by AFM during the injection process. Topography visualization of phage samples was carried out with a Mobile S AFM (Nanosurf AG, Switzerland) having a 110 μm by 110 μm scanning area at the non-contact mode. Measurements were carried out using an ArrowTM Si AFM probe (Nanoworld AG, Switzerland) with a spring constant of 48 N/m and a resonance frequency of around 190 kHz. Phase and color enhanced topography images were made with XE-100 (Park Systems) at the true-non-contact mode. Image analysis was performed using Gwyddion image processing software (free software developed by Czech Metrology Institute, downloaded at http://gwyddion.net).
Electrochemical characterization
EIS and CV were performed using a three-electrode configuration, which consists of a working electrode (i.e. the device under test), a platinum counter electrode (PT 1800, Schott Instruments, Germany) and an Ag/AgCl reference electrode (3M KCl bridging electrolyte, B2820+, Schott Instruments, Germany). The configuration was operated using a potentiostat combined with a frequency response analyzer (Solartron 1260 and 1287, Solartron Analytical, UK). The software packages ZPlot (v3.1c, Scribner Associates Inc.) and CorrWare (v3.1c, Scribner Associates Inc.) were used for EIS and CV, respectively. Impedance spectra were recorded from 1 Hz to 100 kHz with an excitation amplitude of 10 mV in 0.9% NaCl solution. CV was carried out in 0.01 M PBS solution (pH 7.4) from −0.6 V to 0.8 V versus reference electrode with a scan rate of 100 mV s−1.
Fluorescence microscopy
Pt binding f88 phage (f88-Pt phage), rf88 phage and cell mixtures were imaged with an Olympus IX51 inverted microscope under the phase contrast mode with a 100 × oil objective. The images were acquired using a CCD camera (Olympus soft imaging system F view). Following an incubation time of 24 h at 25 ◦ C, cell and phage mixtures were observed by immunofluorescent staining. Each phage sample (20 μl) was first incubated in 2% bovine serum albumin/PBS solution for 1 h at 25 ◦ C. Biotin conjugated anti-fd phage antibodies (1:5000, Sigma) were added and incubated for 1 h at 37 ◦ C. After the antibody incubation, phage samples were mixed with cyanine dye-3 (Cy3)-conjugated streptavidin solution (1:1000, Sigma) and incubated at 25 ◦ C for 1 h. Cy3labeled samples were washed and immediately visualized with the Olympus IX51 inverted microscope with an excitation wavelength of 540 nm and an emission wavelength of 570 nm.
Statistics
The results of cell viability are represented as the mean and standard deviation ( ± SD) from three independent experiments, with significance of differences evaluated using ANOVA with Fisher’s post hoc comparisons. A probability of P values of 0.05 were considered significant for all tests. 5
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
Figure 3. Representative QCM-D response graphs (left side) during the bacteriophage immobilization process and AFM images (right side) of the resulting platinum electrode surface. The numbers in the QCM-D response graphs indicate the injection of solution with different bacteriophage concentrations and the corresponding surfaces are shown in the left AFM images with the same indication number. The numbers (1), (2) and (3) indicate the injection of solutions with 1 × 1010 CFU/ml, 1 × 1011 CFU/ml and 1 × 1013 CFU/ml rf88 phage concentration, respectively. In the AFM images, each scale bar indicates 2 μm.
Results
Cell cytotoxicity effect of the rf88 phage
The cell cytotoxicity effect of the rf88 phage was investigated by the observation of the cells’ behavior on the bare and rf88 phage-immobilized Pt electrodes (figure 4). In the absence of TNF-α, cells cultured on the bare Pt electrode and rf88 phageimmobilized Pt electrode showed the typical healthy fibroblast morphology with a homogeneous distribution as shown in figures 4(a) and (b). After 24 and 36 h exposure to TNF-α, cells on the bare Pt already underwent typical morphologic changes of apoptotic cell deaths such as shrunken shape (figures 4(c) and (d)). Apoptotic cell death was also clearly observed with f88-Pt phage (figures 4(e) and ( f )). In contrast, the cells on the Pt surface coated by rf88 phage did not show significant apoptotic morphological change during the same time period (figures 4(g)–(i)). In line with these results, the cell cultured with different phage samples and phage-immobilized Pt surface showed only subtle changes of cell viabilities shown in figures 5(a) and (b). Compared to the cell viability of the untreated control group (defined as 100%), the cell viability incubated with f88-Pt phage, rf88 phage, Pt electrode and rf88 modified Pt electrode was measured as 94.1 ± 2.8%, 93.4 ± 3.7%, 95.4 ± 5.3% and 92.5 ± 3.2%, respectively. The cells in both groups were treated with TNF-α, a well-known pro-inflammatory cytokine, which leads to cell apoptosis and typical oligonucleosomal DNA fragmentation. The anti-apoptotic effects against TNF-α induced cell death were also analyzed by the apoptotic DNA fragmentation kit in figure 5(c). In lanes 4 and 5 of figure 5(c), a distinct DNA ladder ranging from 150 bp (base pair) to 1 kbp was observed from the cell extract. The DNA of rf88 DNA was visible around 10 kbp in lanes 3 and 5. In the absence of rf88 phage (lane 4 in figure 5(c)), exposure of cells to TNF-α
Immobilization of rf88 bacteriophage on the Pt surface
Fluorescence microscopy images (inset in figure 1) show that the rf88 phages were successfully immobilized on Pt electrodes. As expected, selective binding of rf88 phages to the Pt surface results in high fluorescence intensity contrast between the Pt surface (circular shape) and the polyimide surface indicating that rf88 phages were selectively bound to the Pt surface in contrast to polyimide. The change in frequency ( f ) and dissipation energy (D) in the QCM-D analysis (figure 3) shows the concentration-dependent transient and saturated immobilization of the rf88 phage on the Pt surface. We injected 100 μl of rf88 phage solution with 1 × 1010 CFU/ml concentration into the flow channel (figure 3, arrow 1), followed by subsequent injection of another 100 μl rf88 phage solution with concentration of 1 × 1011 CFU/ml (figure 3, arrow 2) and concentration of 1 × 1012 CFU/ml (figure 3, arrow 3). An immediate decrease in frequency along with the increase in dissipation energy was observed, which proves the transient immobilization of the rf88 phage on the Pt surface. The deposition of rf88 phages on the surfaces was further confirmed by the corresponding AFM images (figure 3, right panel). Low phage concentration (1 × 1010 CFU/ml) leads to the non-uniform shape of phage assembly. When the phage concentration was increased, the phage assemblies became uniform, straight and thicker, which is in line with binding of rf88 to Pt surface shown in figure 1. We also observed that wild type f88 phage also bound to Pt electrode; however, the QCM-D results showed negligible frequency changes (data not shown). 6
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
(c)
(d )
(e)
(f)
( g)
(h)
(i)
Figure 4. Cytotoxicity effects of the SH-SY5Y cell on the bare platinum surface as a control after 0 h (a) and 24 h (b) and in the presence of TNF-α (50 ng ml−1) for 24 h (c) and 36 h (d). The f88-Pt phages on the SH-SY5Y cell in the presence of TNF-α (50 ng ml−1) (e) and ( f ) incubation for 24, 36 h and rf88 (g), (h), (i) incubation for 0, 24, 36 h in the presence of TNF-α (50 ng ml−1), respectively. Each scale bar indicates 100 μm size scale.
average value of the ten independent measurements (gray lines). In all cases, the results showed typical impedance spectra of thin film platinum electrodes: resistive response at a high frequency region and capacitive behavior at the low frequency region. The impedance of both the 80 μm-diameter and 120 μm-diameter electrodes increases after the phage coating process. At 1 kHz, the impedance of the 120 μmdiameter electrode, for example, increased to 131.4 k (average value) after the phage immobilization, compared to the impedance of the bare electrode (62.1 k, average value). Cyclic voltammograms of the bare Pt electrode and rf88 phage-immobilized electrode were shown in figures 7(a) and (b), respectively. As a whole, the shape of CV did not change significantly after the rf88 phage immobilization. However, the area enclosed by CV was slightly reduced and characteristic peaks also were decreased. The cathodal charge storage capacities (CSCC, time integral of the negative current in CV) were calculated as 6.07 ± 0.17 mC cm−2 for the bare Pt electrodes and 5.89 ± 0.33 mC cm−2 for the rf88 phage-coated electrodes, which yielded less than 3% reduction after the rf88 phage immobilization. Even if we exclude the electrode 1 in figure 7(b), which showed larger CV current than the other electrodes in the phage-immobilized group, the CSCC was calculated to 5.72 ± 0.12 mC cm−2, which yields around 6% reduction.
Table 1. Effect of different phage solutions on the cell viability in SH-SY5Y cells at constant TNF-α concentration.
Samples
% Cell viability
TNF-α (50 ng ml−1) alone f88-Pt (1 × 1012 CFU ml−1) + TNF-α (50 ng ml−1) rf88-Pt (1 × 1010 CFU ml−1) + TNF-α (50 ng ml−1) rf88-Pt (1 × 1012 CFU ml−1) + TNF-α (50 ng ml−1)
68 67 76 90
± ± ± ±
7 4 3 2
Cells were exposed to TNF-α (50 ng ml−1) and different phage concentrations (see also figure 5(d)). Data represent means ± SD from three independent experiments, each performed in triplicate, expressed as percentage compared to untreated cells.
resulted in well-pronounced DNA fragmentation. In contrast, the DNA fragmentation was significantly diminished in the rf88 phage-immobilized electrode as shown in lane 5, which is in line with cell viability and morphological changes. As can be seen in figure 5(d) and table 1, TNF-α treatment reduced approximately 30% of cell viability on the bare electrode and f88-Pt phage-immobilized electrode, indicating that wild type phage failed to prevent the TNF-α induced cell death. In contrast, 10–20% recovery of cell viability was observed on 1 × 1010 and 1 × 1012 CFU ml−1 of rf88 phage-coated electrode in a dose-dependent manner, respectively. Effects of rf88 phage on electrochemical properties
Discussion
Electrochemical impedance spectra were analyzed prior to the bacteriophage immobilization and after the rf88 phage coating process (figure 6). In figure 6, dark black lines indicate the
Surface modifications and synthetic material coatings of medical devices are clinically used in a variety of applications 7
Y J Kim et al
Biomed. Mater. 9 (2014) 015009
(a)
(b)
(c) (d)
Figure 5. Comparison of cell viabilities in the absence (a) and presence of the Pt electrode (b). DNA fragmentation from the cell extracts was analyzed by 1% agarose gel electrophoresis after 24 h cell culture (c). The cell viability was determined by WST-1 analysis (d). Bars show means ± SD from three independent experiments. Data were analyzed statistically by ANOVA at a P
