Materials Science and Engineering C 33 (2013) 901–908
Contents lists available at SciVerse ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Reduced myofibroblast differentiation on femtosecond laser treated 316LS stainless steel Martin Oberringer a, 1, Erhan Akman b, Juseok Lee c, Wolfgang Metzger a, Cagri Kaan Akkan c, Elif Kacar b, Arif Demir b, d, Hashim Abdul-Khaliq e, Norbert Pütz f, Gunther Wennemuth f, Tim Pohlemann a, Michael Veith c, Cenk Aktas c,⁎, 1 a
Department of Trauma, Hand and Reconstructive Surgery, Saarland University, Homburg, Germany Laser Technologies Research and Application Center (LATARUM), Kocaeli University, Yeniköy/Kocaeli, Turkey c CVD/Biosurfaces Division, INM — Leibniz Institute for New Materials, Saarbrücken, Germany d BEAM Ar–Ge Optic, Laser and Spectroscopy, KOU Technopark, Kocaeli, 41275, Turkey e Clinic for Pediatric Cardiology, Saarland University Hospital, Homburg, Germany f Department of Anatomy and Cell Biology, Saarland University, Homburg, Germany b
a r t i c l e
i n f o
Article history: Received 9 August 2012 Received in revised form 17 October 2012 Accepted 12 November 2012 Available online 21 November 2012 Keywords: In-stent restenosis Stainless steel Femtosecond laser Microvascular endothelial cell Myofibroblast
a b s t r a c t In-stent restenosis is a common complication after stent surgery which leads to a dangerous wall narrowing of a blood vessel. Laser assisted patterning is one of the effective methods to modify the stent surface to control cell–surface interactions which play a major role in the restenosis. In this current study, 316LS stainless steel substrates are structured by focusing a femtosecond laser beam down to a spot size of 50 μm. By altering the laser induced spot density three distinct surfaces (low density (LD), medium density (MD) and high density (HD)) were prepared. While such surfaces are composed of primary microstructures, due to fast melting and re-solidification by ultra-short laser pulses, nanofeatures are also observed as secondary structures. Following a detailed surface characterization (chemical and physical properties of the surface), we used a well-established co-culture assay of human microvascular endothelial cells and human fibroblasts to check the cell compatibility of the prepared surfaces. The surfaces were analyzed in terms of cell adherence, proliferation, cell morphology and the differentiation of the fibroblast into the myofibroblast, which is a process indicating a general fibrotic shift within a certain tissue. It is observed that myofibroblast proliferation decreases significantly on laser treated samples in comparison to non-treated ones. On the other hand endothelial cell proliferation is not affected by the surface topography which is composed of micro- and nanostructures. Such surfaces may be used to modify stent surfaces for prevention or at least reduction of restenosis. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The use of commercial metal stents results in a restenosis rate of 20–30%. Besides general inflammation and thrombosis, one of the main reasons of in-stent restenosis is the abnormal proliferation of smooth muscle cells (SMC) and soft tissue fibroblasts in such a way that they replace endothelial cells (EC). Recently it has been proposed that in-stent restenosis is largely determined by whether EC or SMC first grow on the surface of a stent [1]. Sufficient endothelialization in the absence of any inflammation and fibrosis formation is thus the main requirement in the design of a stent surface.
⁎ Corresponding author at: CVD/Biosurfaces Division, INM — Leibniz Institute for New Materials, Campus D2 2, 66123 Saarbrücken, Germany. Tel.: +49 681 93001 40; fax: +49 681 9300 223. E-mail address: [email protected] (C. Aktas). 1 Both authors contributed equally to this work. 0928-4931/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msec.2012.11.018
Recent studies showed that surfaces composed of micro- and nanotopographies might provoke an optimized cellular behavior in terms of better endothelialization [2]. Although it has been shown that EC seem to be less sensitive to nanostructures in comparison to fibroblasts and SMC [3], there are some studies clearly demonstrating an increased vascular cell density on nanostructured surfaces [4]. Metallic substances such as stainless steel, nickel-titanium alloys (nitinol), cobalt-chromium alloys, platinum or tantalum, which are used as stent materials, can be modified by various methods including blasting, chemical etching, electrochemical treatment and laser treatment [5]. The use of lasers for the modification of metallic surfaces has become one of the most preferred methods due to unique properties of lasers. Common advantages of laser surface modification are controlled thermal penetration, chemical cleanness and remote non-contact processing. In particular, short pulse lasers operating with pulse durations below 10 ps have become increasingly attractive for precise modifications of metals and metal alloys [6].
902
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
Guo et al. showed that fs laser treatment of titanium leads to a large variety of nanostructures (nanopores, nanoprotrusions, etc.) which measure about 20 nm [7]. In addition, the authors observed the formation of multiple parallel patterns with a period on the sub-micron level and a microroughness in the range of 1–15 μm. In addition to the topography modification, fs laser treatment is also an effective method to alter wettability of surfaces, ranging from hydrophilic to ultra-hydrophobic. Zhang et al. presented micro-scale hierarchical structures achieved by fs laser micromachining, exhibiting ultra-hydrophobicity and reminiscent of the hierarchical structures observed on a hydrophobic lotusleaf [8]. On the other hand, Xiaohong et al. showed that oxidation subsequent to fs laser treatment can affect not only the wetting behavior but can also alter roughness and topography [9]. In this current study, we structured 316LS stainless steel substrates (commonly used implant materials thanks to a favorable combination of mechanical properties, acceptable biocompatibility and low cost [10]) by fs laser to explore the interaction of EC and fibroblasts with a view to opening up new advantages for future stent applications. In comparison to some previous works, we used fibroblasts in addition to EC, since they, together with SMC, are a major component of the sub-endothelial connective tissue of the vessel wall [11]. Fibroblasts are also involved in the fibrotic process, due to a release of a specific cytokine spectrum and due to the establishment of a specific inflammation-related ECM [12]. Expression of α-smooth-muscle-actin (α-SMA) is a clear sign of such a fibrotic development as well as a sign of differentiation of a fibroblast into a myofibroblast (MF) type [13]. Although the effect of the MF with respect to in-stent restenosis has not been investigated in detail, there are indications that α-SMA-expressing cells at least partially account for neointimal hyperplasia [14] and accelerated restenosis [15]. In the present approach, we used a co-culture model of normal human dermal fibroblasts (NHDF) and human dermal microvascular endothelial cells (HDMEC). This model offers major advantages compared to mono-culture models and had previously supplied substantial data concerning general aspects of tissue regeneration and fibrosis [16,17]. Using NHDF instead of SMC in this model makes it not only possible to assess alterations in cell adherence, proliferation and morphology, but also enables the analysis of MF differentiation in detail. We fabricated three different surfaces composed of microstructures (spot arrays) with different interspacings (between identical spot arrays) by fs treatment of 316LS stainless steel plates. Following detailed material characterization, the surfaces were compared in terms of cell adherence, proliferation, cell morphology and the differentiation of the fibroblast into the myofibroblast, which is a process indicating a general fibrotic shift within a certain tissue. 2. Materials and methods 2.1. Laser treatment 316LS polished (mirror quality) stainless steel plates (Goodfellow, Bad Nauheim, Germany) with the dimensions of 20 mm× 20 mm× 1 mm were patterned using an ultrafast Ti:Sapphire laser (Quantronix, East Setauket, NY, USA). This commercial Ti:Sapphire laser consists of a laser oscillator, a pulse stretcher, two amplifiers and a compressor. The laser operates at 1 kHz repetition rate with a pulse width shorter than 130 fs and with maximum output energy of 2.5 mJ/pulse. The surface modification of the samples was carried out at a constant laser pulse energy of 800 μJ/pulse. The laser beam was sent to the sample surface by using a galvo mirror and an F-Theta lens with a focal length of 300 mm was used to focus the laser beam to a spot size of 50 μm. The laser beam was moved at a speed of 50 mm/s over the sample surface in order to create circular structures (spots) with three different interspacings in a vertical direction. The spacing of linearly arranged spots was adjusted to 175 μm, 125 μm and 75 μm (measured from spot-center to next spot-center) to obtain three kinds of surfaces
termed as Low Density (LD), Medium Density (MD) and High Density (HD), respectively. Here “density” refers to the structure density created by the laser on the sample surface. Non-treated 316LS stainless steel samples served as control. 2.2. Surface characterization Following the heat-treatment at 200 °C for 3 h (routine procedure followed for the sterilization), the surface topography of prepared samples was examined using a white light interferometer (NewView™ 7000 Series, Middlefield, CT, USA). Scanning electron microscopy (SEM; JEOL-JSM-6400F, Eching, Germany) with a voltage acceleration of up to 15 kV, was employed to conduct a detailed visual analysis of nanoscale protrusions and other surface features formed during the ablation process. Wettability of surfaces was analyzed using a video contact angle system (Kruess G2, Hamburg, Germany). Samples were placed on a flat stage and a drop of double distilled water (H2Odd) was placed on the surface. The average value of four measurements at different positions of each sample was adopted as the mean contact angle. The surface roughness was analyzed with a stylus profilometer (Zeiss Surfcom 1500 SD3, Gottingen, Germany) operating at ambient atmosphere. For each sample, four line profiles were taken at different locations. Then the average roughness (Ra) of the corresponding surface was determined for each surface. X-ray photoelectron spectroscopy (XPS-PHI 5600 spectrometer, Physical Electronics Chanhassen, MN, USA) analysis of the surfaces was conducted to determine the oxygen content. 2.3. Cell culture Prior to cell culture experiments, all samples were washed and sterilized, working under a sterile laminar airflow cabinet. Washing was performed in phosphate buffered saline (PBS) for 5 min, followed by rinsing in 70% ethanol for 20 min. Subsequently, the samples were washed in PBS and twice in H2Odd, for 5 min each time. After air-drying, the samples were heat-sterilized at 200 °C for 3 h and stored under sterile conditions. Co-cultures could be produced from human dermal microvascular endothelial cells (HDMEC, PromoCell, Heidelberg, Germany) and normal human dermal fibroblasts (NHDF, PromoCell). Cell cultures were expanded in the dedicated culture medium: endothelial cell growth medium MV (Promocell) for HDMEC and Quantum333 medium (PAA, Pasching, Austria) for NHDF. For the experiments (n=5) co-cultures of HDMEC (culture passage 6) and NHDF (culture passage 5) were produced by pooling both cell types after trypsinization. Subsequently, 400,000 NHDF and 1,600,000 HDMEC were pooled in 40 ml of pooled medium (three parts endothelial cell medium and one part Quantum333) and distributed homogeneously on the steel samples (see next chapter). The desired cell density of 250 cells per mm2 and the cell and medium proportions applied here were chosen on the basis of preliminary studies [17]. Two sets of the 4 different steel samples (Non-treated, LD, MD, HD) were placed in the cavities of 6-well-plates (Greiner Bio-One, Frickenhausen, Germany). Per cavity, 5 ml of the total 40 ml suspension of co-cultured cells previously prepared was pipetted onto the surface of each sample. One set was cultured for exactly one day (Non-treated1, LD1, MD1, HD1) and one set was cultured for exactly three days (Non-treated3, LD3, MD3, HD3). Cell growth was stopped by fixation in methanol (− 20 °C) and air-drying. Samples were glued onto glass slides using commercially available superglue and were then covered with PBS buffered glycerin until immunocytochemical staining was performed. 2.4. Immunocytochemical staining In order to determine the cell densities among the different cell populations (HDMEC, NHDF and MF) in co-culture, we applied
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
antibodies against von Willebrand factor (vWF, rabbit polyclonal anti-human vWF Ab-1, 1:500, Dianova, Hamburg, Germany) identifying HDMEC and against α-SMA (mouse monoclonal anti-human α-SMA, 1:800, Sigma, St Louis, USA) identifying those fibroblasts, which had been differentiated into MF. Primary antibodies were visualized by a Cy3-labeled antibody (Cy3TM-conjugated goat anti-rabbit, 1:200, Dianova) and a FITC-labeled secondary antibody (FITC-conjugated goat anti-mouse, 1:100, Dianova), respectively. Slides were stored in mounting medium containing 4′,6-diamidino2-phenylindole (DAPI; Vectashield, Vector Laboratories, Burlingame, USA) for nuclear counterstaining. 2.5. Quantitative microscopic analysis and analytical parameters Microscopic evaluation was done with a Zeiss Axioskop 2 (Zeiss, Göttingen, Germany) equipped with suitable fluorescence filter sets. Cell numbers of each cell type present in the co-culture were determined per sample by counting the cells of 5 × 5 = 25 high power fields, the location of which was identical on each sample in order to exclude any manipulation by the examiners. Cell counting of each cell type resulted in the parameter cell density [cell number per mm 2 culture surface], which was identified for each cell type on day one and day three. In addition, the differentiation of fibroblasts in terms of MF development was investigated. The significance (p b 0.05) of any differences in analytical parameters was checked by paired-samples t-test (comparison of day one and day three data within the same surface) and by analysis of variance (ANOVA) including Bonferroni correction (day one data of all surfaces and day three data of all surfaces). 2.6. Scanning electron microscopy Co-cultures for SEM analysis were seeded with the same cell density described before on the different samples (10 × 10 × 1 mm) and the cells were fixed after 24 h. Remaining medium was removed from the samples by rinsing 2 × with PBS (37 °C). After fixation using 2% glutardialdehyde in 0.12 M cacodylate-buffer, incubation
903
in osmium tetroxide (2% H2Odd) for 2 h under movement in the dark, dehydration and washing, the samples were dried by Critical Point Drying (Polaron CPD 7501, Quorom Technologies, Ontario, Canada) and sputtered with gold-palladium (Polaron, Sputter Coater). Afterwards they were analyzed in a FEI XL 30 ESEM FEG SEM device (Hillsboro, U.S.A.). 3. Results 3.1. Laser treatment and characterization of surfaces White light interferometer micrographs (Fig. 1) demonstrate the surface topographies of non-treated (Fig. 1a) and fs laser treated (three different interspacings) surfaces (Fig. 1b–d). In order to eliminate any uncertainty which can result from the sterilization step carried out at 200 °C for 3 h, all non-treated and laser treated samples were characterized afterwards. The size of the laser induced spots was approximately 50 μm on each sample and the spacings between the spot arrays were kept varied (75 μm, 125 μm and 175 μm) to achieve three different surfaces: HD, MD and LD, respectively. It is clear that when the interspacing is kept shorter, the number of spots per unit area increases, which leads to a high density of structures on the surface. In other words, surfaces with a higher structure density are subjected to more intense laser irradiation. Some inclusions, which are much finer, formed around the spots and between the spot arrays. This may happen as a result of the ablation and re-solidification of the surface features where the laser beam is focused. We analyzed laser induced spots using SEM to examine the details of the topographic changes within and around them. Fig. 2 shows high resolution SEM images of a spot array. There are at least three different regions with distinct morphologies (Fig. 2a). In the center of the spot, the formation of dispersed craters with a diameter of a few μm is clearly visible (Fig. 2b). At higher magnification, nanoscale features appear on top of this microtopography. Fig. 2c shows some sub-micron periodic patterns near the borders of the spot. These sub-micron periods are covered with nanoscaled structures and the periodicity is disturbed by a wavy morphology, reminiscent of a
Fig. 1. White light interferometer micrographs. (a) Non-treated, (b) High Density (HD), (c) Medium Density (MD) and (d) Low Density (LD) samples.
904
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
Fig. 2. SEM micrographs of a spot array. (a) Overview, (b) highly magnified image from the inner part of a spot, (c) periodic nanopatterns at the border of a spot and (d) nanoscale inclusions between two spot arrays.
melt flow. Between two parallel spot arrays the surface is composed of nanoscale inclusions which are around 50–100 nm in diameter (Fig. 2d). While laser induced micro spots form the primary structures, the nanoscale structures induced by ultra-short laser pulses as a consequence of ultra-fast heating, ablation and solidification can be counted as the secondary topography. Such inclusion like nano features have been previously observed by Vorobyev and Guo [7]. In addition to the topography analysis, we tried to find out whether there is a correlation between roughness and wettability. Fig. 3a and b shows graphical illustrations of roughness and wetting of non-treated and fs laser treated surfaces. While the contact angle of non-treated samples was 75 +/− 4°, wetting characteristics changed after the laser treatment towards a hydrophobic nature. The contact angle on LD and MD surfaces reached to 110 +/− 5.5° and 147 +/− 7.4°, respectively. Such an increase was not observed in case of the HD surface, where the wetting angle remained at about 83 +/− 4.2°. Our roughness analysis shows that HD has the highest Ra value of 103 +/− 5.2 nm, which was much higher than that of the non-treated steel surface (Ra = 9 +/− 0.5 nm). In comparison to HD, the MD surface has a relatively lower Ra of 87 +/− 4.4 nm. LD has the smoothest surface (compared to the other laser treated surfaces) with an Ra of 61 +/− 3 nm. As shown by Xiaohong et al., the change in surface oxygen content (after the laser treatment) may also account for the alteration of wettability [9]. Similarly Hao et al. showed that there is a clear difference in surface oxygen content after laser treatment of the 316LS stainless steel using a CO2 continuous wave (CW) laser [18]. Our XPS analysis showed similar results: the oxygen content increases with increasing density of the structures (Fig. 3c). Higher density of the surface structures means higher exposure to the laser beam.
3.2. Proliferation capacity The chosen setup not only enables the evaluation of the initial adherence capacity and the proliferation behavior of the cells in total, but also is an ideal tool to differentiate between the proliferation of HDMEC and NHDF on the substrates after immunocytochemical staining (Fig. 4a). Total cell densities on day one were the same for all substrate types, but day three data showed a significant difference in cell densities on non-treated steel and LD: whereas 542+/−147 cells/mm2 grew on non-treated steel, only 308+/−158 cells/mm2 were detectable on HD substrate (Fig. 4b). The impaired growth of cells on HD substrate was further emphasized by the fact that HD was the only substrate not allowing a significant proliferation of the cells between day one and day three, as was the case for all the other substrates (non-treated, LD and MD). Taking the proliferation of HDMEC as an indicator of the general suitability of the substrates for stent material, better adherence of this cell type on non-treated steel on day one (105+/− 17 cells/mm2) was observed, which was of statistical significance in comparison with the laser treated substrates (Fig. 4c). However, both the LD and MD substrates compensated for this difference up to culture day three, with proliferation then no longer being significantly different. This indicates a similar proliferation behavior on LD and MD substrates compared to non-treated substrates. Here again, only HD substrate showed a significantly weaker HDMEC density (58 +/− 32 cells/mm2) compared to non-treated substrate, indicating poor proliferation characteristics up to day three. The analysis of NHDF revealed similar proliferation characteristics on non-treated substrates; LD and MD surfaces, displaying a significant increase up to day three (Fig. 4d). The highest cell density was detected on non-treated steel (399+/− 144 cells/mm2), whereas the lowest
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
905
Fig. 3. (a) Roughness, (b) wettability and (c) surface oxygen content (in mass percentage) of non-treated and laser treated surfaces. (Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples).
density was seen on HD substrate (249+/− 134 cells/mm2). Again, HD substrate led to a different NHDF behavior in terms of not allowing a significant increase up to day three. However, a comparison of the NHDF densities on the different substrates, both on day one and on day three, showed no significant differences. 3.3. Myofibroblast differentiation As one can see in Fig. 5a on both day one (10 +/− 8 cells/mm2) and day three (16 +/− 7 cells/mm2), the highest MF density was observed on non-treated steel in comparison to laser treated surfaces. This reminds of the proliferation of MF on a standard glass slide as shown in various previous studies [16,17]. Fluorescence microscopy indicated that cell surface enlargement of MF was also most prominent on non-treated substrates. This may be a consequence of the proliferation of already present MF or the triggered differentiation of NHDF. The detailed visualization of the largest cells within the cultures, identified as MF, revealed that there were differences with respect
Fig. 4. (a) Proliferation capacity. Representative example of a fluorescence photomicrograph, serving for quantification: human dermal microvascular endothelial cells (HDMEC; red), normal human dermal fibroblasts (NHDF; blue nuclei) and myofibroblasts (MF; green) grown on LD on day one. Total (b) and cell type specific (c, d) cell densities on day one and on day three on the different substrates determined by quantitative microscopy subsequent to immunocytochemical staining; data given as means with sd; n=5. 1=one day incubation; 3=three day incubation. Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples.
906
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
to the structures which might be responsible for the interaction of cells with the surface. We were able to identify two different MF morphologies in general: the first morphology was characterized by the absence of filopodia, when cells exhibited a planar surface. These MF predominantly appeared on LD, MD and HD substrates (Fig. 5b). The second MF morphology was characterized by the expression of numerous fine filopodia mainly occurring on non-treated substrates (Fig. 5c and d). These filopodia often ended with flattened regions, maybe a sign of an attempt to enhance the cell-to-surface contact area (empty arrowheads in Fig. 5d). 4. Discussion
Fig. 5. (a) Myofibroblast (MF) cell densities on day one and on day three on the different substrates; data given as means with sd; n = 5. 1 = one day incubation time; 3 = three day incubation time. (b) Myofibroblasts without filopodia growing on MD, visualized by scanning electron microscopy. (c) Myofibroblasts with extensive filopodia on non-treated steel. (d) Extract from (c): Detailed view on flattened regions at the filopodia endings (arrowheads). Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples.
In the present study a co-culture assay composed of microvascular endothelial cells and fibroblasts was used to characterize the suitability of structured 316LS stainless steel substrates for stent application, aiming to reduce the risk of restenosis. For this purpose, 316LS surfaces were micro- and nanostructured by fs laser treatment, a versatile process, which can be automated by computer controlling and can maintain a high reproducibility. Since pulsed lasers combine the advantage ultra short material-photon interaction and non-contact processing nature, they are accepted as effective tools to create structured materials [19]. Compared to CW, millisecond and nanosecond lasers, fs lasers offer a higher precision due to a reduced heataffected zone [7]. The fast melting and solidification on the surface of a material, caused by ultra-short laser pulses, lead to changes in the topography, generating cavities in the micrometer range, nanospots and a smaller amount of debris around the spots. White light interferometry was used to visualize submicron topographic features effectively: The spot size was kept identical and only the spacing between the spot arrays was altered (Fig. 1). SEM analysis revealed the presence of much smaller structures (nanometer range) between the spot arrays and within every spot. The first kind of nanostructures refers to finer features observed within the spots and ablated particulates around these spots. The second kind of nanostructures refers to nanodots with diameter of 50–100 nm observed between the spot arrays. These nanodots seem to be particles ablated and re-sputtered during repeating fs laser pulses. Our detailed material characterization further revealed that the average roughness of the LD substrates is relatively low compared to MD and HD substrates. This might be due to the lowest spot density there, which is correlated to a low density of nanodots in the interspacing between spot arrays. An increase in spot density on MD and HD samples relates to a decrease of the interspacing between the spots. Wettability seems to depend on the spot density. As it has been known for a long time from the structure of the lotus leaf, the combination of micro- and nanostructures is a key determinant of the hydrophobicity of a certain material. In this context, MD shows the highest water contact angle which may be due to the combination of the micro- and nanostructures. These results show that not only the minimal size of the structures, but also their distribution plays a key role in influencing wetting behavior. In addition, the change in surface oxygen content is an important factor which influences the wettability. Our XPS analysis showed that the oxygen content increased with increasing density of laser induced spots, which was in line with previous result of Hao et al. [18], (Fig. 3c). Parallel thereto, roughness also increased in parallel to the increased spot density (Fig. 3a). However, the hydrophobic nature of the samples, which should be determined to a great extent by both roughness and oxygen content, was not the highest among HD samples, but in fact among the MD surfaces. Since laser treatment induces a gradient like change in the topography and even in surface chemistry due to Gaussian type intensity of the laser beam, it is difficult to estimate the wetting behavior using a theoretical model which assumes the presence of topographic features with well defined size and geometry.
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
The modified response of the cells to these different laser treated surfaces was clearly detectable. Focusing on the overall proliferation characteristics of the different cell types used in this model, HD substrate was the only substrate not allowing a significant proliferation of the cells between day one and day three. All other substrates provoked a general proliferation. Although HDMEC did initially adhere well on the non-treated substrate, by day three there was no longer any difference in the cell number compared to LD and MD substrates. Only EC proliferation was not sufficient on HD substrate. In comparison to previous studies we did not observe a change in the morphology of EC observed during fluorescence and scanning electron microscopy [2,20]. It seems that the size and distribution of the structures produced by our method do not provoke a visible morphology change among HDMEC. With respect to fibroblasts, we know from literature that both adhesion and proliferation are sensitive to the topography of a surface [21–23]. Furthermore, the alignment of fibroblasts can be provoked by nanostructuring [24–26]. We did not observe a significant difference between cell numbers of co-cultured NHDF on non-treated, LD and MD substrates. Generally, proliferation of the fibroblast cell type was detected on all the substrates, again with the exception of HD substrate. The most striking result of the study emerged upon focusing on the behavior of the differentiated fraction of the NHDF, the MF. Apart from SMC, this cell type is one of the main players involved in tissue fibrosis and thus a crucial cell in the restenosis process. Most in vitro research on preventing restenosis is focused on SMC. We show here that the quantification of MF differentiation could prove to be an effective tool in the assessment of the characteristics of a material with regard to the provocation of fibrosis in general and of restenosis in particular. It is observed that the rate of MF was lower on all laser treated substrates, regardless of the density of spot arrays. On day one the initial MF rate on laser treated samples was slightly lower than that observed on non-treated substrate. This effect became predominant on day three and MF rates were significantly lower on all laser treated samples compared to the non-treated control substrate. It is believed that this is related to multi-scale topography composed of micro- and nanostructures (as secondary structures). Up to our knowledge there is no study which shows the MF differentiation by nanostructures directly. On the other hand, Baxter et al., using osteoblasts, has already shown a reduced a-SMA expression in response to 330 nm deep grooves before [27]. Hydrophobicity alone does not seem to have a direct influence on MF development; otherwise we should have observed a clear difference in MF rates while wetting angles change significantly. In general, MF growing on non-treated steel surfaces expressed fine filopodia formation more than observed on any kind of the laser treated substrates. Dalby et al. recently described filopodia as possible sensors for the three-dimensional environment [28]. Previously, we postulated that an enhanced production of such filopodia by fibroblasts is characteristic for a better cell adherence on a certain surface [29]. In the present context, the phenomenon of increased filopodia appears only on non-treated steel surfaces and coincides with the highest cell numbers and the highest MF differentiation, thus supporting previous results. Possibly, the determination of the number of cellular filopodia may serve as a valuable tool for the analysis of the adherence capacity of a certain cell in response to a given surface in future studies. Summarizing the results, it was somehow surprising that the highest proliferation of MF was observed on the polished 316LS. This is a fact which puts the widely accepted use of polished steel in stent surgery into question. LD substrates seem to provide most promising topography compared to others since the lowest MF differentiation was observed on this surface and the HDMEC proliferation was not disturbed. On the other hand, the effect of the increased water contact angle and low oxygen content is the interfering
907
parameter. We believe that the main effect can arise from the presence of nanofeatures formed between dot arrays. In case of MD and HD surfaces, due to closer dot-arrays micro topography is more predominant and isolated nanofeatures seem to be overshadowed. It is believed that the combination of micro- and nanoscale surface features accounts for the cellular behavior. Consequently, future studies should place priority focus on the distribution density of smaller secondary structures (nanoscale) within larger topographic features (microscale). In this context, fs assisted surface structuring can be applied to various metallic substrates (cobalt-chromium, nickeltitanium and other alloys) and even polymers. On the other hand one should take into account that laser treatment of polymeric materials will also induce chemical modifications on the surface in addition to altered topography. 5. Conclusion An effective and easy method is presented to fabricate micro-and nanostructured surfaces using fs lasers. Such topography which is composed of multi-scale features (micro and nano) reduces the MF population without disturbing the proliferation of EC. This may form a basis for developing micro- and nanostructured stent surfaces for preventing stent restenosis, which is partially mediated by repopulation with MF and enhanced inflammation. On the other hand, the effect of the interfering parameters such as altered wetting and oxygen content should be minimized to understand the direct effect of the topography on the cell response. Although it is clear that prepared surfaces do not disturb the EC proliferation, still the optimization of micro-and nano structure combination is needed to improve the endothelialization. Acknowledgments This work is supported by New INDIGO (the Initiative for the Development and Integration of Indian and European Research) and State Planning Organization of Turkey under Medical Electro-Optics Research Laboratory (project number 2011K120330). References [1] J.C. Palmaz, Intravascular stents: tissue-stent interactions and design considerations, Am. J. Roentgenol. 160 (1193) 613–18. [2] A. Ranjan, T. Webster, Nanotechnology 20 (2009) 305102. [3] S.A. Biela, Y. Su, J.P. Spatz, R. Kemkemer, Acta Biomater. 5 (2009) 2460–2466. [4] D.C. Miller, A. Thapa, K.M. Haberstroh, T. Webster, Biomaterials 25 (2004) 53–61. [5] D.M. Dohan Ehrenfest, P.G. Coelho, B.S. Kang, Y.T. Sul, T. Albrektsson, Trends Biotechnol. 28 (2010) 198–206. [6] R. Le Harzic, N. Huot, E. Audouard, C. Jonin, P. Laporte, S. Valette, A. Fraczkiewicz, R. Fortunier, Appl. Phys. Lett. 80 (2002) 3886. [7] A.Y. Vorobyev, C. Guo, Appl. Surf. Sci. 253 (2007) 7272–7280. [8] D. Zhang, F. Chen, G. Fang, Q. Yang, D. Xie, G. Qiao, W. Li, J. Si, X. Hou, J. Micromech. Microeng. 20 (2010) 075029. [9] L. Xiaohong, X. Qin, L. Zhihui, H. Wenhao, Int. Conf. Electron. Optoelectron. 2 (2011) V2109–V2112. [10] C. Schmidt, A.A. Ignatius, L.E. Claes, J. Biomed. Mater. Res. 54 (2001) 209–215. [11] F. Geneser, W. Schwerdtfeger, Histologie (Köln: Deutscher Ärzteverlag) (1990) 317–340. [12] P. Sivakumar, A.M. Das, Inflamm. Res. 57 (2008) 410–418. [13] D.W. Powell, R.C. Mifflin, J.D. Valentich, S.E. Crowe, J.I. Saada, A.B. West, Am. J. Physiol. 277 (1999) C1–C19. [14] A. Forte, A. Della Corte, M. De Feo, F. Cerasuolo, M. Cipollaro, Cardiovasc. Res. 88 (2010) 395–405. [15] H.J.S. Stewart, A.L. Guildford, D.J. Lawrence-Watt, M. Santin, J. Biomed. Mater. Res. 90A (2009) 465–471. [16] M. Oberringer, C. Meins, M. Bubel, T. Pohlemann, Biol. Cell 99 (2007) 197–207. [17] M. Oberringer, C. Meins, M. Bubel, T. Pohlemann, J. Mol. Histol. 39 (2008) 37–47. [18] L. Hao, J. Lawrence, Y.F. Phua, K.S. Chian, G.C. Lim, H.Y. Zheng, Biomed. Mater. Res. B Appl. Biomater. 73B (2005) 148–156. [19] M.H. Hong, S.M. Huang, B.S. Lukyanchuk, T.C. Chong, Sens. Actuators A 108 (2003) 69–74. [20] C.C. Co, Y.C. Wang, C.C. Ho, J. Am. Chem. Soc. 127 (2005) 1598–1599. [21] M. Kononen, M. Hormia, J. Kivilahti, J. Hautaniemi, I. Thesleff, J. Biomed. Mater. Res. 26 (1992) 1325–1341.
908
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
[22] E.R. Takamori, R. Cruz, F. Goncalvez, R.V. Zanetti, A. Zanetti, J.M. Granjeiro, Artif. Organs 32 (2008) 305–309. [23] M. Wang, Y. Lu, Surf. Coat. Technol. 204 (2009) 525–530. [24] M.J. Dalby, S.J. Yarwood, M.O. Riehle, H.J. Johnstone, S. Affrossman, A.S.G. Curtis, Exp. Cell Res. 276 (2002) 1–9. [25] M.J. Dalby, M.O. Riehle, S.J. Yarwood, C.D. Wilkinson, A.S.G. Curtis, Exp. Cell Res. 284 (2003) 274–282. [26] M.J. Dalby, D. Giannaras, M.O. Riehle, N. Gadegaard, S. Affrossman, A.S. Curtis, Biomaterials 25 (2004) 77–83.
[27] L.C. Baxter, V. Frauchiger, M. Textor, I. Gwynn, R.G. Richards, Eur. Cell Mater. 4 (2002) 1–17. [28] M.J. Dalby, M.O. Riehle, D.S. Sutherland, H. Agheli, A.S.G. Curtis, Biomaterials 25 (2004) 5415–5422. [29] M. Veith, O.C. Aktas, W. Metzger, D. Sossong, H. Wazir, I. Grobelsek, N. Pütz, G. Wennemuth, T. Pohlemann, M. Oberringer, Biofabrication 2 (2010) 035001.
Contents lists available at SciVerse ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Reduced myofibroblast differentiation on femtosecond laser treated 316LS stainless steel Martin Oberringer a, 1, Erhan Akman b, Juseok Lee c, Wolfgang Metzger a, Cagri Kaan Akkan c, Elif Kacar b, Arif Demir b, d, Hashim Abdul-Khaliq e, Norbert Pütz f, Gunther Wennemuth f, Tim Pohlemann a, Michael Veith c, Cenk Aktas c,⁎, 1 a
Department of Trauma, Hand and Reconstructive Surgery, Saarland University, Homburg, Germany Laser Technologies Research and Application Center (LATARUM), Kocaeli University, Yeniköy/Kocaeli, Turkey c CVD/Biosurfaces Division, INM — Leibniz Institute for New Materials, Saarbrücken, Germany d BEAM Ar–Ge Optic, Laser and Spectroscopy, KOU Technopark, Kocaeli, 41275, Turkey e Clinic for Pediatric Cardiology, Saarland University Hospital, Homburg, Germany f Department of Anatomy and Cell Biology, Saarland University, Homburg, Germany b
a r t i c l e
i n f o
Article history: Received 9 August 2012 Received in revised form 17 October 2012 Accepted 12 November 2012 Available online 21 November 2012 Keywords: In-stent restenosis Stainless steel Femtosecond laser Microvascular endothelial cell Myofibroblast
a b s t r a c t In-stent restenosis is a common complication after stent surgery which leads to a dangerous wall narrowing of a blood vessel. Laser assisted patterning is one of the effective methods to modify the stent surface to control cell–surface interactions which play a major role in the restenosis. In this current study, 316LS stainless steel substrates are structured by focusing a femtosecond laser beam down to a spot size of 50 μm. By altering the laser induced spot density three distinct surfaces (low density (LD), medium density (MD) and high density (HD)) were prepared. While such surfaces are composed of primary microstructures, due to fast melting and re-solidification by ultra-short laser pulses, nanofeatures are also observed as secondary structures. Following a detailed surface characterization (chemical and physical properties of the surface), we used a well-established co-culture assay of human microvascular endothelial cells and human fibroblasts to check the cell compatibility of the prepared surfaces. The surfaces were analyzed in terms of cell adherence, proliferation, cell morphology and the differentiation of the fibroblast into the myofibroblast, which is a process indicating a general fibrotic shift within a certain tissue. It is observed that myofibroblast proliferation decreases significantly on laser treated samples in comparison to non-treated ones. On the other hand endothelial cell proliferation is not affected by the surface topography which is composed of micro- and nanostructures. Such surfaces may be used to modify stent surfaces for prevention or at least reduction of restenosis. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The use of commercial metal stents results in a restenosis rate of 20–30%. Besides general inflammation and thrombosis, one of the main reasons of in-stent restenosis is the abnormal proliferation of smooth muscle cells (SMC) and soft tissue fibroblasts in such a way that they replace endothelial cells (EC). Recently it has been proposed that in-stent restenosis is largely determined by whether EC or SMC first grow on the surface of a stent [1]. Sufficient endothelialization in the absence of any inflammation and fibrosis formation is thus the main requirement in the design of a stent surface.
⁎ Corresponding author at: CVD/Biosurfaces Division, INM — Leibniz Institute for New Materials, Campus D2 2, 66123 Saarbrücken, Germany. Tel.: +49 681 93001 40; fax: +49 681 9300 223. E-mail address: [email protected] (C. Aktas). 1 Both authors contributed equally to this work. 0928-4931/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msec.2012.11.018
Recent studies showed that surfaces composed of micro- and nanotopographies might provoke an optimized cellular behavior in terms of better endothelialization [2]. Although it has been shown that EC seem to be less sensitive to nanostructures in comparison to fibroblasts and SMC [3], there are some studies clearly demonstrating an increased vascular cell density on nanostructured surfaces [4]. Metallic substances such as stainless steel, nickel-titanium alloys (nitinol), cobalt-chromium alloys, platinum or tantalum, which are used as stent materials, can be modified by various methods including blasting, chemical etching, electrochemical treatment and laser treatment [5]. The use of lasers for the modification of metallic surfaces has become one of the most preferred methods due to unique properties of lasers. Common advantages of laser surface modification are controlled thermal penetration, chemical cleanness and remote non-contact processing. In particular, short pulse lasers operating with pulse durations below 10 ps have become increasingly attractive for precise modifications of metals and metal alloys [6].
902
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
Guo et al. showed that fs laser treatment of titanium leads to a large variety of nanostructures (nanopores, nanoprotrusions, etc.) which measure about 20 nm [7]. In addition, the authors observed the formation of multiple parallel patterns with a period on the sub-micron level and a microroughness in the range of 1–15 μm. In addition to the topography modification, fs laser treatment is also an effective method to alter wettability of surfaces, ranging from hydrophilic to ultra-hydrophobic. Zhang et al. presented micro-scale hierarchical structures achieved by fs laser micromachining, exhibiting ultra-hydrophobicity and reminiscent of the hierarchical structures observed on a hydrophobic lotusleaf [8]. On the other hand, Xiaohong et al. showed that oxidation subsequent to fs laser treatment can affect not only the wetting behavior but can also alter roughness and topography [9]. In this current study, we structured 316LS stainless steel substrates (commonly used implant materials thanks to a favorable combination of mechanical properties, acceptable biocompatibility and low cost [10]) by fs laser to explore the interaction of EC and fibroblasts with a view to opening up new advantages for future stent applications. In comparison to some previous works, we used fibroblasts in addition to EC, since they, together with SMC, are a major component of the sub-endothelial connective tissue of the vessel wall [11]. Fibroblasts are also involved in the fibrotic process, due to a release of a specific cytokine spectrum and due to the establishment of a specific inflammation-related ECM [12]. Expression of α-smooth-muscle-actin (α-SMA) is a clear sign of such a fibrotic development as well as a sign of differentiation of a fibroblast into a myofibroblast (MF) type [13]. Although the effect of the MF with respect to in-stent restenosis has not been investigated in detail, there are indications that α-SMA-expressing cells at least partially account for neointimal hyperplasia [14] and accelerated restenosis [15]. In the present approach, we used a co-culture model of normal human dermal fibroblasts (NHDF) and human dermal microvascular endothelial cells (HDMEC). This model offers major advantages compared to mono-culture models and had previously supplied substantial data concerning general aspects of tissue regeneration and fibrosis [16,17]. Using NHDF instead of SMC in this model makes it not only possible to assess alterations in cell adherence, proliferation and morphology, but also enables the analysis of MF differentiation in detail. We fabricated three different surfaces composed of microstructures (spot arrays) with different interspacings (between identical spot arrays) by fs treatment of 316LS stainless steel plates. Following detailed material characterization, the surfaces were compared in terms of cell adherence, proliferation, cell morphology and the differentiation of the fibroblast into the myofibroblast, which is a process indicating a general fibrotic shift within a certain tissue. 2. Materials and methods 2.1. Laser treatment 316LS polished (mirror quality) stainless steel plates (Goodfellow, Bad Nauheim, Germany) with the dimensions of 20 mm× 20 mm× 1 mm were patterned using an ultrafast Ti:Sapphire laser (Quantronix, East Setauket, NY, USA). This commercial Ti:Sapphire laser consists of a laser oscillator, a pulse stretcher, two amplifiers and a compressor. The laser operates at 1 kHz repetition rate with a pulse width shorter than 130 fs and with maximum output energy of 2.5 mJ/pulse. The surface modification of the samples was carried out at a constant laser pulse energy of 800 μJ/pulse. The laser beam was sent to the sample surface by using a galvo mirror and an F-Theta lens with a focal length of 300 mm was used to focus the laser beam to a spot size of 50 μm. The laser beam was moved at a speed of 50 mm/s over the sample surface in order to create circular structures (spots) with three different interspacings in a vertical direction. The spacing of linearly arranged spots was adjusted to 175 μm, 125 μm and 75 μm (measured from spot-center to next spot-center) to obtain three kinds of surfaces
termed as Low Density (LD), Medium Density (MD) and High Density (HD), respectively. Here “density” refers to the structure density created by the laser on the sample surface. Non-treated 316LS stainless steel samples served as control. 2.2. Surface characterization Following the heat-treatment at 200 °C for 3 h (routine procedure followed for the sterilization), the surface topography of prepared samples was examined using a white light interferometer (NewView™ 7000 Series, Middlefield, CT, USA). Scanning electron microscopy (SEM; JEOL-JSM-6400F, Eching, Germany) with a voltage acceleration of up to 15 kV, was employed to conduct a detailed visual analysis of nanoscale protrusions and other surface features formed during the ablation process. Wettability of surfaces was analyzed using a video contact angle system (Kruess G2, Hamburg, Germany). Samples were placed on a flat stage and a drop of double distilled water (H2Odd) was placed on the surface. The average value of four measurements at different positions of each sample was adopted as the mean contact angle. The surface roughness was analyzed with a stylus profilometer (Zeiss Surfcom 1500 SD3, Gottingen, Germany) operating at ambient atmosphere. For each sample, four line profiles were taken at different locations. Then the average roughness (Ra) of the corresponding surface was determined for each surface. X-ray photoelectron spectroscopy (XPS-PHI 5600 spectrometer, Physical Electronics Chanhassen, MN, USA) analysis of the surfaces was conducted to determine the oxygen content. 2.3. Cell culture Prior to cell culture experiments, all samples were washed and sterilized, working under a sterile laminar airflow cabinet. Washing was performed in phosphate buffered saline (PBS) for 5 min, followed by rinsing in 70% ethanol for 20 min. Subsequently, the samples were washed in PBS and twice in H2Odd, for 5 min each time. After air-drying, the samples were heat-sterilized at 200 °C for 3 h and stored under sterile conditions. Co-cultures could be produced from human dermal microvascular endothelial cells (HDMEC, PromoCell, Heidelberg, Germany) and normal human dermal fibroblasts (NHDF, PromoCell). Cell cultures were expanded in the dedicated culture medium: endothelial cell growth medium MV (Promocell) for HDMEC and Quantum333 medium (PAA, Pasching, Austria) for NHDF. For the experiments (n=5) co-cultures of HDMEC (culture passage 6) and NHDF (culture passage 5) were produced by pooling both cell types after trypsinization. Subsequently, 400,000 NHDF and 1,600,000 HDMEC were pooled in 40 ml of pooled medium (three parts endothelial cell medium and one part Quantum333) and distributed homogeneously on the steel samples (see next chapter). The desired cell density of 250 cells per mm2 and the cell and medium proportions applied here were chosen on the basis of preliminary studies [17]. Two sets of the 4 different steel samples (Non-treated, LD, MD, HD) were placed in the cavities of 6-well-plates (Greiner Bio-One, Frickenhausen, Germany). Per cavity, 5 ml of the total 40 ml suspension of co-cultured cells previously prepared was pipetted onto the surface of each sample. One set was cultured for exactly one day (Non-treated1, LD1, MD1, HD1) and one set was cultured for exactly three days (Non-treated3, LD3, MD3, HD3). Cell growth was stopped by fixation in methanol (− 20 °C) and air-drying. Samples were glued onto glass slides using commercially available superglue and were then covered with PBS buffered glycerin until immunocytochemical staining was performed. 2.4. Immunocytochemical staining In order to determine the cell densities among the different cell populations (HDMEC, NHDF and MF) in co-culture, we applied
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
antibodies against von Willebrand factor (vWF, rabbit polyclonal anti-human vWF Ab-1, 1:500, Dianova, Hamburg, Germany) identifying HDMEC and against α-SMA (mouse monoclonal anti-human α-SMA, 1:800, Sigma, St Louis, USA) identifying those fibroblasts, which had been differentiated into MF. Primary antibodies were visualized by a Cy3-labeled antibody (Cy3TM-conjugated goat anti-rabbit, 1:200, Dianova) and a FITC-labeled secondary antibody (FITC-conjugated goat anti-mouse, 1:100, Dianova), respectively. Slides were stored in mounting medium containing 4′,6-diamidino2-phenylindole (DAPI; Vectashield, Vector Laboratories, Burlingame, USA) for nuclear counterstaining. 2.5. Quantitative microscopic analysis and analytical parameters Microscopic evaluation was done with a Zeiss Axioskop 2 (Zeiss, Göttingen, Germany) equipped with suitable fluorescence filter sets. Cell numbers of each cell type present in the co-culture were determined per sample by counting the cells of 5 × 5 = 25 high power fields, the location of which was identical on each sample in order to exclude any manipulation by the examiners. Cell counting of each cell type resulted in the parameter cell density [cell number per mm 2 culture surface], which was identified for each cell type on day one and day three. In addition, the differentiation of fibroblasts in terms of MF development was investigated. The significance (p b 0.05) of any differences in analytical parameters was checked by paired-samples t-test (comparison of day one and day three data within the same surface) and by analysis of variance (ANOVA) including Bonferroni correction (day one data of all surfaces and day three data of all surfaces). 2.6. Scanning electron microscopy Co-cultures for SEM analysis were seeded with the same cell density described before on the different samples (10 × 10 × 1 mm) and the cells were fixed after 24 h. Remaining medium was removed from the samples by rinsing 2 × with PBS (37 °C). After fixation using 2% glutardialdehyde in 0.12 M cacodylate-buffer, incubation
903
in osmium tetroxide (2% H2Odd) for 2 h under movement in the dark, dehydration and washing, the samples were dried by Critical Point Drying (Polaron CPD 7501, Quorom Technologies, Ontario, Canada) and sputtered with gold-palladium (Polaron, Sputter Coater). Afterwards they were analyzed in a FEI XL 30 ESEM FEG SEM device (Hillsboro, U.S.A.). 3. Results 3.1. Laser treatment and characterization of surfaces White light interferometer micrographs (Fig. 1) demonstrate the surface topographies of non-treated (Fig. 1a) and fs laser treated (three different interspacings) surfaces (Fig. 1b–d). In order to eliminate any uncertainty which can result from the sterilization step carried out at 200 °C for 3 h, all non-treated and laser treated samples were characterized afterwards. The size of the laser induced spots was approximately 50 μm on each sample and the spacings between the spot arrays were kept varied (75 μm, 125 μm and 175 μm) to achieve three different surfaces: HD, MD and LD, respectively. It is clear that when the interspacing is kept shorter, the number of spots per unit area increases, which leads to a high density of structures on the surface. In other words, surfaces with a higher structure density are subjected to more intense laser irradiation. Some inclusions, which are much finer, formed around the spots and between the spot arrays. This may happen as a result of the ablation and re-solidification of the surface features where the laser beam is focused. We analyzed laser induced spots using SEM to examine the details of the topographic changes within and around them. Fig. 2 shows high resolution SEM images of a spot array. There are at least three different regions with distinct morphologies (Fig. 2a). In the center of the spot, the formation of dispersed craters with a diameter of a few μm is clearly visible (Fig. 2b). At higher magnification, nanoscale features appear on top of this microtopography. Fig. 2c shows some sub-micron periodic patterns near the borders of the spot. These sub-micron periods are covered with nanoscaled structures and the periodicity is disturbed by a wavy morphology, reminiscent of a
Fig. 1. White light interferometer micrographs. (a) Non-treated, (b) High Density (HD), (c) Medium Density (MD) and (d) Low Density (LD) samples.
904
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
Fig. 2. SEM micrographs of a spot array. (a) Overview, (b) highly magnified image from the inner part of a spot, (c) periodic nanopatterns at the border of a spot and (d) nanoscale inclusions between two spot arrays.
melt flow. Between two parallel spot arrays the surface is composed of nanoscale inclusions which are around 50–100 nm in diameter (Fig. 2d). While laser induced micro spots form the primary structures, the nanoscale structures induced by ultra-short laser pulses as a consequence of ultra-fast heating, ablation and solidification can be counted as the secondary topography. Such inclusion like nano features have been previously observed by Vorobyev and Guo [7]. In addition to the topography analysis, we tried to find out whether there is a correlation between roughness and wettability. Fig. 3a and b shows graphical illustrations of roughness and wetting of non-treated and fs laser treated surfaces. While the contact angle of non-treated samples was 75 +/− 4°, wetting characteristics changed after the laser treatment towards a hydrophobic nature. The contact angle on LD and MD surfaces reached to 110 +/− 5.5° and 147 +/− 7.4°, respectively. Such an increase was not observed in case of the HD surface, where the wetting angle remained at about 83 +/− 4.2°. Our roughness analysis shows that HD has the highest Ra value of 103 +/− 5.2 nm, which was much higher than that of the non-treated steel surface (Ra = 9 +/− 0.5 nm). In comparison to HD, the MD surface has a relatively lower Ra of 87 +/− 4.4 nm. LD has the smoothest surface (compared to the other laser treated surfaces) with an Ra of 61 +/− 3 nm. As shown by Xiaohong et al., the change in surface oxygen content (after the laser treatment) may also account for the alteration of wettability [9]. Similarly Hao et al. showed that there is a clear difference in surface oxygen content after laser treatment of the 316LS stainless steel using a CO2 continuous wave (CW) laser [18]. Our XPS analysis showed similar results: the oxygen content increases with increasing density of the structures (Fig. 3c). Higher density of the surface structures means higher exposure to the laser beam.
3.2. Proliferation capacity The chosen setup not only enables the evaluation of the initial adherence capacity and the proliferation behavior of the cells in total, but also is an ideal tool to differentiate between the proliferation of HDMEC and NHDF on the substrates after immunocytochemical staining (Fig. 4a). Total cell densities on day one were the same for all substrate types, but day three data showed a significant difference in cell densities on non-treated steel and LD: whereas 542+/−147 cells/mm2 grew on non-treated steel, only 308+/−158 cells/mm2 were detectable on HD substrate (Fig. 4b). The impaired growth of cells on HD substrate was further emphasized by the fact that HD was the only substrate not allowing a significant proliferation of the cells between day one and day three, as was the case for all the other substrates (non-treated, LD and MD). Taking the proliferation of HDMEC as an indicator of the general suitability of the substrates for stent material, better adherence of this cell type on non-treated steel on day one (105+/− 17 cells/mm2) was observed, which was of statistical significance in comparison with the laser treated substrates (Fig. 4c). However, both the LD and MD substrates compensated for this difference up to culture day three, with proliferation then no longer being significantly different. This indicates a similar proliferation behavior on LD and MD substrates compared to non-treated substrates. Here again, only HD substrate showed a significantly weaker HDMEC density (58 +/− 32 cells/mm2) compared to non-treated substrate, indicating poor proliferation characteristics up to day three. The analysis of NHDF revealed similar proliferation characteristics on non-treated substrates; LD and MD surfaces, displaying a significant increase up to day three (Fig. 4d). The highest cell density was detected on non-treated steel (399+/− 144 cells/mm2), whereas the lowest
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
905
Fig. 3. (a) Roughness, (b) wettability and (c) surface oxygen content (in mass percentage) of non-treated and laser treated surfaces. (Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples).
density was seen on HD substrate (249+/− 134 cells/mm2). Again, HD substrate led to a different NHDF behavior in terms of not allowing a significant increase up to day three. However, a comparison of the NHDF densities on the different substrates, both on day one and on day three, showed no significant differences. 3.3. Myofibroblast differentiation As one can see in Fig. 5a on both day one (10 +/− 8 cells/mm2) and day three (16 +/− 7 cells/mm2), the highest MF density was observed on non-treated steel in comparison to laser treated surfaces. This reminds of the proliferation of MF on a standard glass slide as shown in various previous studies [16,17]. Fluorescence microscopy indicated that cell surface enlargement of MF was also most prominent on non-treated substrates. This may be a consequence of the proliferation of already present MF or the triggered differentiation of NHDF. The detailed visualization of the largest cells within the cultures, identified as MF, revealed that there were differences with respect
Fig. 4. (a) Proliferation capacity. Representative example of a fluorescence photomicrograph, serving for quantification: human dermal microvascular endothelial cells (HDMEC; red), normal human dermal fibroblasts (NHDF; blue nuclei) and myofibroblasts (MF; green) grown on LD on day one. Total (b) and cell type specific (c, d) cell densities on day one and on day three on the different substrates determined by quantitative microscopy subsequent to immunocytochemical staining; data given as means with sd; n=5. 1=one day incubation; 3=three day incubation. Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples.
906
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
to the structures which might be responsible for the interaction of cells with the surface. We were able to identify two different MF morphologies in general: the first morphology was characterized by the absence of filopodia, when cells exhibited a planar surface. These MF predominantly appeared on LD, MD and HD substrates (Fig. 5b). The second MF morphology was characterized by the expression of numerous fine filopodia mainly occurring on non-treated substrates (Fig. 5c and d). These filopodia often ended with flattened regions, maybe a sign of an attempt to enhance the cell-to-surface contact area (empty arrowheads in Fig. 5d). 4. Discussion
Fig. 5. (a) Myofibroblast (MF) cell densities on day one and on day three on the different substrates; data given as means with sd; n = 5. 1 = one day incubation time; 3 = three day incubation time. (b) Myofibroblasts without filopodia growing on MD, visualized by scanning electron microscopy. (c) Myofibroblasts with extensive filopodia on non-treated steel. (d) Extract from (c): Detailed view on flattened regions at the filopodia endings (arrowheads). Non-treated, Low Density (LD), Medium Density (MD) and High Density (HD) samples.
In the present study a co-culture assay composed of microvascular endothelial cells and fibroblasts was used to characterize the suitability of structured 316LS stainless steel substrates for stent application, aiming to reduce the risk of restenosis. For this purpose, 316LS surfaces were micro- and nanostructured by fs laser treatment, a versatile process, which can be automated by computer controlling and can maintain a high reproducibility. Since pulsed lasers combine the advantage ultra short material-photon interaction and non-contact processing nature, they are accepted as effective tools to create structured materials [19]. Compared to CW, millisecond and nanosecond lasers, fs lasers offer a higher precision due to a reduced heataffected zone [7]. The fast melting and solidification on the surface of a material, caused by ultra-short laser pulses, lead to changes in the topography, generating cavities in the micrometer range, nanospots and a smaller amount of debris around the spots. White light interferometry was used to visualize submicron topographic features effectively: The spot size was kept identical and only the spacing between the spot arrays was altered (Fig. 1). SEM analysis revealed the presence of much smaller structures (nanometer range) between the spot arrays and within every spot. The first kind of nanostructures refers to finer features observed within the spots and ablated particulates around these spots. The second kind of nanostructures refers to nanodots with diameter of 50–100 nm observed between the spot arrays. These nanodots seem to be particles ablated and re-sputtered during repeating fs laser pulses. Our detailed material characterization further revealed that the average roughness of the LD substrates is relatively low compared to MD and HD substrates. This might be due to the lowest spot density there, which is correlated to a low density of nanodots in the interspacing between spot arrays. An increase in spot density on MD and HD samples relates to a decrease of the interspacing between the spots. Wettability seems to depend on the spot density. As it has been known for a long time from the structure of the lotus leaf, the combination of micro- and nanostructures is a key determinant of the hydrophobicity of a certain material. In this context, MD shows the highest water contact angle which may be due to the combination of the micro- and nanostructures. These results show that not only the minimal size of the structures, but also their distribution plays a key role in influencing wetting behavior. In addition, the change in surface oxygen content is an important factor which influences the wettability. Our XPS analysis showed that the oxygen content increased with increasing density of laser induced spots, which was in line with previous result of Hao et al. [18], (Fig. 3c). Parallel thereto, roughness also increased in parallel to the increased spot density (Fig. 3a). However, the hydrophobic nature of the samples, which should be determined to a great extent by both roughness and oxygen content, was not the highest among HD samples, but in fact among the MD surfaces. Since laser treatment induces a gradient like change in the topography and even in surface chemistry due to Gaussian type intensity of the laser beam, it is difficult to estimate the wetting behavior using a theoretical model which assumes the presence of topographic features with well defined size and geometry.
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
The modified response of the cells to these different laser treated surfaces was clearly detectable. Focusing on the overall proliferation characteristics of the different cell types used in this model, HD substrate was the only substrate not allowing a significant proliferation of the cells between day one and day three. All other substrates provoked a general proliferation. Although HDMEC did initially adhere well on the non-treated substrate, by day three there was no longer any difference in the cell number compared to LD and MD substrates. Only EC proliferation was not sufficient on HD substrate. In comparison to previous studies we did not observe a change in the morphology of EC observed during fluorescence and scanning electron microscopy [2,20]. It seems that the size and distribution of the structures produced by our method do not provoke a visible morphology change among HDMEC. With respect to fibroblasts, we know from literature that both adhesion and proliferation are sensitive to the topography of a surface [21–23]. Furthermore, the alignment of fibroblasts can be provoked by nanostructuring [24–26]. We did not observe a significant difference between cell numbers of co-cultured NHDF on non-treated, LD and MD substrates. Generally, proliferation of the fibroblast cell type was detected on all the substrates, again with the exception of HD substrate. The most striking result of the study emerged upon focusing on the behavior of the differentiated fraction of the NHDF, the MF. Apart from SMC, this cell type is one of the main players involved in tissue fibrosis and thus a crucial cell in the restenosis process. Most in vitro research on preventing restenosis is focused on SMC. We show here that the quantification of MF differentiation could prove to be an effective tool in the assessment of the characteristics of a material with regard to the provocation of fibrosis in general and of restenosis in particular. It is observed that the rate of MF was lower on all laser treated substrates, regardless of the density of spot arrays. On day one the initial MF rate on laser treated samples was slightly lower than that observed on non-treated substrate. This effect became predominant on day three and MF rates were significantly lower on all laser treated samples compared to the non-treated control substrate. It is believed that this is related to multi-scale topography composed of micro- and nanostructures (as secondary structures). Up to our knowledge there is no study which shows the MF differentiation by nanostructures directly. On the other hand, Baxter et al., using osteoblasts, has already shown a reduced a-SMA expression in response to 330 nm deep grooves before [27]. Hydrophobicity alone does not seem to have a direct influence on MF development; otherwise we should have observed a clear difference in MF rates while wetting angles change significantly. In general, MF growing on non-treated steel surfaces expressed fine filopodia formation more than observed on any kind of the laser treated substrates. Dalby et al. recently described filopodia as possible sensors for the three-dimensional environment [28]. Previously, we postulated that an enhanced production of such filopodia by fibroblasts is characteristic for a better cell adherence on a certain surface [29]. In the present context, the phenomenon of increased filopodia appears only on non-treated steel surfaces and coincides with the highest cell numbers and the highest MF differentiation, thus supporting previous results. Possibly, the determination of the number of cellular filopodia may serve as a valuable tool for the analysis of the adherence capacity of a certain cell in response to a given surface in future studies. Summarizing the results, it was somehow surprising that the highest proliferation of MF was observed on the polished 316LS. This is a fact which puts the widely accepted use of polished steel in stent surgery into question. LD substrates seem to provide most promising topography compared to others since the lowest MF differentiation was observed on this surface and the HDMEC proliferation was not disturbed. On the other hand, the effect of the increased water contact angle and low oxygen content is the interfering
907
parameter. We believe that the main effect can arise from the presence of nanofeatures formed between dot arrays. In case of MD and HD surfaces, due to closer dot-arrays micro topography is more predominant and isolated nanofeatures seem to be overshadowed. It is believed that the combination of micro- and nanoscale surface features accounts for the cellular behavior. Consequently, future studies should place priority focus on the distribution density of smaller secondary structures (nanoscale) within larger topographic features (microscale). In this context, fs assisted surface structuring can be applied to various metallic substrates (cobalt-chromium, nickeltitanium and other alloys) and even polymers. On the other hand one should take into account that laser treatment of polymeric materials will also induce chemical modifications on the surface in addition to altered topography. 5. Conclusion An effective and easy method is presented to fabricate micro-and nanostructured surfaces using fs lasers. Such topography which is composed of multi-scale features (micro and nano) reduces the MF population without disturbing the proliferation of EC. This may form a basis for developing micro- and nanostructured stent surfaces for preventing stent restenosis, which is partially mediated by repopulation with MF and enhanced inflammation. On the other hand, the effect of the interfering parameters such as altered wetting and oxygen content should be minimized to understand the direct effect of the topography on the cell response. Although it is clear that prepared surfaces do not disturb the EC proliferation, still the optimization of micro-and nano structure combination is needed to improve the endothelialization. Acknowledgments This work is supported by New INDIGO (the Initiative for the Development and Integration of Indian and European Research) and State Planning Organization of Turkey under Medical Electro-Optics Research Laboratory (project number 2011K120330). References [1] J.C. Palmaz, Intravascular stents: tissue-stent interactions and design considerations, Am. J. Roentgenol. 160 (1193) 613–18. [2] A. Ranjan, T. Webster, Nanotechnology 20 (2009) 305102. [3] S.A. Biela, Y. Su, J.P. Spatz, R. Kemkemer, Acta Biomater. 5 (2009) 2460–2466. [4] D.C. Miller, A. Thapa, K.M. Haberstroh, T. Webster, Biomaterials 25 (2004) 53–61. [5] D.M. Dohan Ehrenfest, P.G. Coelho, B.S. Kang, Y.T. Sul, T. Albrektsson, Trends Biotechnol. 28 (2010) 198–206. [6] R. Le Harzic, N. Huot, E. Audouard, C. Jonin, P. Laporte, S. Valette, A. Fraczkiewicz, R. Fortunier, Appl. Phys. Lett. 80 (2002) 3886. [7] A.Y. Vorobyev, C. Guo, Appl. Surf. Sci. 253 (2007) 7272–7280. [8] D. Zhang, F. Chen, G. Fang, Q. Yang, D. Xie, G. Qiao, W. Li, J. Si, X. Hou, J. Micromech. Microeng. 20 (2010) 075029. [9] L. Xiaohong, X. Qin, L. Zhihui, H. Wenhao, Int. Conf. Electron. Optoelectron. 2 (2011) V2109–V2112. [10] C. Schmidt, A.A. Ignatius, L.E. Claes, J. Biomed. Mater. Res. 54 (2001) 209–215. [11] F. Geneser, W. Schwerdtfeger, Histologie (Köln: Deutscher Ärzteverlag) (1990) 317–340. [12] P. Sivakumar, A.M. Das, Inflamm. Res. 57 (2008) 410–418. [13] D.W. Powell, R.C. Mifflin, J.D. Valentich, S.E. Crowe, J.I. Saada, A.B. West, Am. J. Physiol. 277 (1999) C1–C19. [14] A. Forte, A. Della Corte, M. De Feo, F. Cerasuolo, M. Cipollaro, Cardiovasc. Res. 88 (2010) 395–405. [15] H.J.S. Stewart, A.L. Guildford, D.J. Lawrence-Watt, M. Santin, J. Biomed. Mater. Res. 90A (2009) 465–471. [16] M. Oberringer, C. Meins, M. Bubel, T. Pohlemann, Biol. Cell 99 (2007) 197–207. [17] M. Oberringer, C. Meins, M. Bubel, T. Pohlemann, J. Mol. Histol. 39 (2008) 37–47. [18] L. Hao, J. Lawrence, Y.F. Phua, K.S. Chian, G.C. Lim, H.Y. Zheng, Biomed. Mater. Res. B Appl. Biomater. 73B (2005) 148–156. [19] M.H. Hong, S.M. Huang, B.S. Lukyanchuk, T.C. Chong, Sens. Actuators A 108 (2003) 69–74. [20] C.C. Co, Y.C. Wang, C.C. Ho, J. Am. Chem. Soc. 127 (2005) 1598–1599. [21] M. Kononen, M. Hormia, J. Kivilahti, J. Hautaniemi, I. Thesleff, J. Biomed. Mater. Res. 26 (1992) 1325–1341.
908
M. Oberringer et al. / Materials Science and Engineering C 33 (2013) 901–908
[22] E.R. Takamori, R. Cruz, F. Goncalvez, R.V. Zanetti, A. Zanetti, J.M. Granjeiro, Artif. Organs 32 (2008) 305–309. [23] M. Wang, Y. Lu, Surf. Coat. Technol. 204 (2009) 525–530. [24] M.J. Dalby, S.J. Yarwood, M.O. Riehle, H.J. Johnstone, S. Affrossman, A.S.G. Curtis, Exp. Cell Res. 276 (2002) 1–9. [25] M.J. Dalby, M.O. Riehle, S.J. Yarwood, C.D. Wilkinson, A.S.G. Curtis, Exp. Cell Res. 284 (2003) 274–282. [26] M.J. Dalby, D. Giannaras, M.O. Riehle, N. Gadegaard, S. Affrossman, A.S. Curtis, Biomaterials 25 (2004) 77–83.
[27] L.C. Baxter, V. Frauchiger, M. Textor, I. Gwynn, R.G. Richards, Eur. Cell Mater. 4 (2002) 1–17. [28] M.J. Dalby, M.O. Riehle, D.S. Sutherland, H. Agheli, A.S.G. Curtis, Biomaterials 25 (2004) 5415–5422. [29] M. Veith, O.C. Aktas, W. Metzger, D. Sossong, H. Wazir, I. Grobelsek, N. Pütz, G. Wennemuth, T. Pohlemann, M. Oberringer, Biofabrication 2 (2010) 035001.
