Abstract
This study investigates the effects of the structured surface on the adhesion, proliferation and alignment of endothelial cells (HUVECs). The chemical state of laser-structured surface was also investigated in comparison with the non-treated surface. A novel design for biomedical applications consisting of parallel chain-like structures was realised on stainless steel surface by laser micromachining. The structures were designed to employ surface topography in the presence of micron and sub-micron features and to avoid intensive surface modification that could compromise the mechanical properties of thin devices like cardiovascular stents. The results showed that the structured surfaces favored the adhesion, proliferation and alignment of HUVECs. The proliferation and the alignment of HUVECs were pronounced when the periodic distance between two consecutive chain-like structures was 25 µm. Moreover, there was no significant difference in chemical composition on the structured surface suggesting that the cell proliferation and alignment were mainly influenced by the surface topography.
Introduction
Endothelialisation, or a lack thereof, plays a critical role in the occurrence of thrombosis following the implantation of blood contact devices such as heart valves, catheters and cardiovascular stents [1]. In a confluent state, the endothelial cell layer produces nitric oxide (NO) contributing to the inhibition of platelet and maintaining smooth muscle cells in a non-proliferative state [2]. The deployment of a cardiovascular stent involves the disruption of the endothelial layer along the artery due to the catheterisation, particularly within the atherosclerotic lesion where the balloon is deployed to force the stent against the arterial wall. For this reason, stent deployment is associated to 30–40% smooth muscle over-proliferation or intimal hyperplasia (IH) [3]. Although IH is successfully overcome by the incorporation of anti-proliferative drugs, its long-term occurrence is unavoidable once the drugs are fully eluted [4]. To this end, surface structuring by laser was proposed as a potential option to restore the pivotal layer of endothelial cells.
Laser surface structuring stands out as a promising option being highly precise and flexible in terms of generating different possibilities to enhance surface morphology and chemistry. The use of different wavelengths (ultraviolet to near infrared) and pulse durations (femtosecond to nanosecond) allow for structuring different types of material [5-7]. The process can be used to modify the surface topography with a flatter geometry [8]or ordered structures [9-11]. Laser-structured surfaces are most commonly studied for orthopaedic and dental implants to promote osteointegration [12 15]. Periodic surface microstructures such as grooves [5,16], dimples [12], squares and triangles [17], as well as random textures, have been proposed [18]. Cell adhesion and proliferation are commonly achieved by many types of surface structures, such as pits and posts, which have been explored on stainless steel surface and various polymers. They have been reported to increase the adhesion, density, monolayer formation and proliferation of endothelial cells [6,7]. Of all the mentioned structures, grooving appears to be one of the most effective, as it promotes endothelial cells alignment in addition to any mentioned benefit of surface structuring [5,17]. However, a grooving dimension as deep as 5 µm could compromise the mechanical properties of thin structure devices such as cardiovascular stents. In fact, this surface feature could constitute a concern, as newer generation of cardiovascular stent bears thinner struts than those of the older ones [19,20]. Therefore, a surface structure appropriately studied for thin structure medical devices, such as cardiovascular stent applications, is desirable.
The current study is aimed at proposing a novel structure that could promote the adhesion, proliferation, and alignment of endothelial cells. Chemical and physical characterisations of the novel structure were investigated in relation to the behaviour of endothelial cells. To this end, a parallel chain-like structure was proposed and printed on stainless steel surface by laser micromachining (1) to introduce surface topography features in the range of micron and sub-micron size; (2) to avoid an intensive surface modification, such as a grooving that could compromise the mechanical properties of thin structured devices like cardiovascular stent. The surface structure was realised by partial melting and ablation through nanosecond-pulsed fibre laser on stainless steel AISI 316L (SS316L). Similar surface structures have been reported in literature for different applications. In this work, their use for manipulating the cell adhesion and proliferation behaviour has been studied by adjusting their spatial distribution. The structure was applied along linear scans over the sample surface. The resulting surface was then analysed for its chemical composition by XPS analysis. Cellular viability, proliferation and alignment of endothelial cells were then assessed comparing laser-structured and electropolished surfaces. Different periodic distances between adjacent chain-like structures were applied to verify if the density of the chains had an influence on the behaviour of endothelial cells.
Materials and method
Laser treatment
SS316L samples (Goodfellow Corporation, PA, USA) with a thickness of 0.2 mm were cut in square sheets measuring 10 mm × 10 mm. Before the laser treatment, all the samples were mechanically polished using SiC abrasive paper (from 800 to 1200 grit) then finished with 50-nm-diameter colloidal silica. The samples were then washed with acetone, deionised water and eethanol in an ultrasonic bath for 15 minutes each and air dried. Samples were then electropolished and acid dipped as previously described elsewhere [21]. The electropolishing solution for SS316L consists of glycerol 99% (50% v/v), phosphoric acid 85% (35% v/v), and deionised H2O (15% v/v). The electrolyte was kept at 90°C during the procedure, which was performed at a constant voltage of 60 V; the cathode was made of the same material as the anode. The electropolished samples were then acid dipped in a mixture of nitric acid 70% (10% v/v), hydrofluoric acid 50% (2% v/v), and deionised H2O (88% v/v). The samples were then rinsed following the previous washing method.
Details of the experimental plan.
In the second stage, samples were prepared with selected laser parameter combinations as illustrated in Figure 1(A). The periodic distance (d) between the scan lines was varied with regards to the average size of endothelial cells (25 µm). Therefore, the d was varied to 25, 75, and 150 µm, suggesting different density of structures on the surfaces. The resulting samples were conventionally labelled as d25, d75, and d150, respectively. They were observed using Scanning Electron Microscope (SEM JSM-840A from Jeol Ltd., Tokyo, Japan) in secondary electron mode with a tungsten filament at an acceleration voltage of 15 kV. Moreover, the surface topography was quantitatively characterised using an Atomic Force Microscope (AFM Veeco Instrument Inc., Woodbury, USA) equipped with a silicon probe (tip radius of 2 nm) and assisted by the built-in software Nanoscope Analysis V1.40r1 (Veeco Instrument Inc., Woodbury, USA).
A schematic representation of an SS316L square sample (A). Grey lines within the square represent the periodic distance between two adjacent laser structures, represented by d. d was in the range 25–150 μm; only two machined lines are represented in the drawings, to put in evidence their distance. Measurement of HUVECs orientation on the laser-structured surface (B). An ellipse (e) was fitted to the cell profile followed by the measurement between the major ellipse axis (l) and the direction of the chain-like structures (y), which is defined as the alignment angle (a).
Characterisation of surface chemistry
X-ray photoelectron spectroscopy (XPS) was used to analyse the surface chemical composition of the electrochemically polished and laser-structured SS316L. The analysis required an extended width of laser surface structured region due to the sampling spot diameter of the XPS equipment of about 125 μm. For this, a laser-structured zone consisting of highly dense chain-liked structured was created to conduct a selective analysis of the laser irradiated area. Before XPS analysis, the samples were cleaned in an ultrasonic bath for 15 minutes repeatedly with acetone, deionised H2O, and 70% ethanol. XPS was carried out on a PHI 5600 ESCA System (Physical Electronics USA, MN, USA) instrument using standard Al K
radiation for survey scan (K
= 1486.6 eV) and Mg K
for high-resolution scan (K
= 1253.6 eV). The survey scans were acquired with a pass energy of 80 eV. The angle between X-ray beam and the analyser plan was 45o and photo-electrons were collected normal to the sample surface. Data processing was analysed using a Matlab-based program with specific macros called Multipak V9.6.0 (Ulvac-Phi Inc., Kanagawa, Japan). The curve fitting for high-resolution peaks were determined by means of the least-squares method using Gauss-Lorentz functions with a Shirley background subtraction.
HUVECs adhesion and proliferation
The primary human vascular endothelial cells (HUVECs) were isolated following the procedure described elsewhere [22]. HUVECs were cultured with M199 supplemented with 10% (v/v) foetal bovine serum (FBS) (Gibco, Thermo Fisher Scientific, Mississauga, ON, Canada), 1% (v/v) Penicillin–Streptomycin (Thermo Fisher Scientific), 2 ng mL−1 recombinant Human fibroblast growth factor (FGF) (Thermo Fisher Scientific), 0.5 ng mL−1 recombinant human epidermal growth factor (EGF) (Thermo Fisher Scientific), 1 µg mL−1 ascorbic acid (Sigma Aldrich, Oakville, ON, Canada), 5 µg mL−1 insulin (Sigma Aldrich), 1 µg mL−1 hydrocortisone (Sigma Aldrich), and 90 µg mL−1 heparin (Sigma Aldrich) within 37°C humidified incubator with 5° CO2. The samples were sterilised for 10 minutes in ethanol 70%, rinsed three times in PBS 1X solution, and dried under the biological safety hood. The samples of d25, d75, and d150 were placed in 24-well cell culture plate individually. Three samples were involved for each condition (n = 3). HUVECs were poured onto each well with the density of 5000 cells cm−2 for 3 and 5-day incubation periods and 10,000 cells cm−2 for 1-day incubation period. The culture medium was changed every second day allowing a fresh supply of nutrients. To measure the metabolic activity of the HUVECs, the medium was removed following the incubation period; 500 µL of fresh medium containing 10% (v/v) resazurin (Sigma Aldrich) was added into each well and incubated for three hours at 37°C with 5° CO2. Subsequently, 100 µL of medium from each well was transferred into a 96-well plate and absorbance measurements were performed on a spectrophotometer Fluoroskan Ascent (ThermoFisher Scientific) at 450 nm. The proliferation rate of HUVECs was measured by cell counting using haemocytometer following each incubation period by harvesting the cells with 0.5% (v/v) trypsin-EDTA (Thermo Fisher Scientific) for 3–5 minutes at 37°C.
Fluorescence microscopy
Following the incubation period of HUVECs on the samples, the cells were stained for immunofluorescence. The samples were washed with PBS 1X three times and then fixed with 3.7% formaldehyde in PBS for 20 minutes. The cells were then rinsed with PBS 1X three times and permeabilised with 3% BSA and 0.1% saponin in PBS 1X for 1 minute. DAPI and rhodamine-phalloidine (1/200 in 3% BSA in PBS 1X) were added for an hour of incubation time. The samples were then washed with 0.05% Tween-20 in PBS 1X three times. The cells were afterward observed under a fluorescence microscope Olympus BX51 (Olympus Corps., Tokyo, Japan).
Image analysis of cell alignment
For cell orientation study, about 150 cells were analysed from at least 3 images for each laser patterned surface. The angle between cell orientation and chain-like structure direction (oriented nominally at 0o) was measured using an image-processing software (ImageJ Version v1.49 Wayne Rasband, National Institutes of Health, USA) as described elsewhere [23]. Briefly, an ellipse was fitted to the cell profile and followed by the measurement of the alignment angle, that is the angle between the major ellipse axis and the image y-axis (the direction of chain-like structures). The measured angles were then clustered into groups (binned into groups of 10°) representing each alignment angle from −90° to + 90°, with 0° as the reference y-axis of the image (chain-like structure direction).
Statistical analysis
Results are presented as mean ± standard error unless specified otherwise. Sample inter-group comparison was performed using one-way analysis of variance (one-way ANOVA) and Tukey's post-test. Statistical significance was set at p < 0.05. Data and statistical analysis were performed using GraphPad Prism version 5.01 for Windows (from GraphPad Software, San Diego, USA).
Results
Laser-structured surface
Figure 2 reports the measured micro-pit width (w) and depth (h) of the chain-like structure as a function of process parameters. It could be observed that w increases with the pulse energy, whereas the increase in scan speed resulted in a moderate decrease for fixed energy levels (Figure 2(A)). On the other hand, h showed a significant increase at 50 μJ pulse energy (Figure 2(B)). Below this energy level, h decreased to values around 1 μm. With ns-long pulses, several mechanisms could contribute to the material removal depending on the pulse energy, pulse duration, and beam dimension [24]. For pulsed-laser micromachining, the ablation threshold was considered commonly to describe the material removal behaviour [25]. While the use of this method for ps- and fs-pulsed lasers was adequate, with ns-pulsed lasers the description of an ablation threshold only falls short for describing the differences observed as a function of process parameters [26]. With ns-pulsed lasers, as observed also in this study at lower pulse energy levels, material removal mechanism became predominantly based on melt displacement [27]. With an increase in energy levels, melt ejection and vaporisation took part, providing a more efficient material removal mechanism. Figure 3(A) shows the micro-pit morphology obtained with E = 20 μJ and v = 97 mm s−1. The morphology was composed of overlapped circles, which were produced by consecutive laser pulses. Each circle exhibited a smaller circle in the centre, which corresponded to the region of effective evaporation. The region between the ablation circle and the border of the wider circle was the product of partial melting and vaporisation [28]. The resultant morphology was the so-called chain-like structure. This structure has been employed as it fits well with the dimensional requirements of the application. The width of the chain-like structure was adequate to fit the single cell, while the depth was shallow enough to avoid cell entrapment in the pits. Moreover, the treatment was relatively non-invasive for the stent applications, remaining at a limited depth and not-reducing significantly the load-bearing part of the implant. Figure 3 shows the three-dimensional morphology of this structure obtained by AFM. It could be seen that the chain-like structure increased the surface area, as well as provided multiple gripping sites for cells. This condition was chosen for further investigation in surface characterisation as well as the adhesion and proliferation of HUVECs. Average w and h were measured as 10.1 ± 0.9 and 0.64 ± 0.07 μm, respectively.
Geometrical characteristics of the laser micromachined micro-pits as a function of process parameters. The width (w) of chain-like structure increased with the pulse energy (A), whereas the increase in scan speed resulted in a moderate decrease for fixed energy levels. Micro-pit depth (h) showed a significant increase at 50 μJ pulse energy (B). Close-up SEM image of chain-like surface structure following laser treatment; the width of chain-like structure is represented by w (A). Topographical characterisation of a chain-like structure by AFM; the depth of chain-like structure is represented by h (B). SEM image of chain-like d75 structure following laser structuring with d defined as the periodic distance between two adjacent chains (C).
Surface chemical composition
The survey spectra of XPS showed that metallic elements such as Fe and Cr were present on both electropolished and laser-structured surfaces with no significant difference, as shown in Figure 4. Their presence was expected, as they were the alloying elements of SS316L. Moreover, non-metallic elements such as C, N, P, and F were also detected on the surface of the samples. The presence of non-metallic elements suggestively came from the processing steps of surface preparation as electropolishing steps involved the use of different acids and solvents. For comparison purpose, S and P were both found on the electropolished surface of SS316L: they could be attributed to the presence of sulphuric and orthophosphoric acids in the electrolyte solution for electropolishing, as reported elsewhere [29]. Moreover, both the surfaces were considered as oxygen-rich. Oxygen binds to chromium forming a passive layer on the surface of SS316L preventing further corrosion. The observed amount of oxygen on both electroformed and laser-treated surfaces is comparable. Moreover, carbon surface contamination was also observed on both surfaces. However, this type of contamination is commonly found since isolating the carbon compounds is not a trivial task [30]. One important thing to be emphasised is that the contamination from hydrocarbon compounds was the same for each sample. All the samples were prepared from the same metal sheet, treated with electrolyte similarly, stored within the same conditions, and then measured at the same time. The only difference is the laser treatment applied to the specific surface area. Assuming that the laser treatment did not affect the carbon contamination on the surface of SS316L, then the measurement error derived from the contamination was the same for each case.
Chemical composition of electropolished and laser-structured SS316L surface by XPS analysis.
High-resolution XPS scan showed that carbon spectra of both electropolished and laser-treated surfaces contained carbon–carbon or carbon–hydrogen peaks with binding energy of 284.8 eV, carbon–sulphur or carbon–nitrogen compounds with binding energy of 286.2 eV, and carbon–oxygen bond with binding energy of 288.6 eV [22]. Deconvolution graphs of oxygen species showed the presence of metal oxide (binding energy 529.7 eV), metal hydroxide (binding energy 531 eV), and others (binding energy 532.4 eV) on both surfaces [31]. Moreover, further analysis of Cr and Fe spectra was performed to determine their oxidation states and formed species in the passive film. For both electropolished and laser-treated surfaces, oxide compounds in chromium spectra were characterised by the peaks at 577.2 and 578.5 eV[22,23] as seen in Figure 5. The presence of Cr2O3 was confirmed by the difference in binding energy of Cr2p1 and Cr2p3 of 9.79 eV [22]. For iron spectra, the presence of oxide compounds was shown by the presence of 709.4 and 711.8 eV peaks [29-34]. The occurrence of FeOOH was confirmed by the difference of Fe2p1 and Fe2p3 binding energies (13.65 eV) [31]. Referring to the deconvolution graphs of high-resolution scans, there was no significant difference between electropolished and laser-treated surfaces. The only exception is found on the graph of iron, which showed a higher amount of iron-hydroxide on the laser-treated surface (higher fourth component). This was reflected on the O1s peak (higher second component). The third component of the O1s peak, probably part of sulphates, nitrates or phosphates, dropped by about half just like the non-metallic elements other than C. The total percentages correspond assuming 2–3 O per other atom: ∼11% O1s(3) down to ∼6% O1s(3) versus ∼6% down to ∼2% for the total of S, N, P and F.
Representative deconvolution graph for carbon, oxygen, chromium, and iron following electropolished and laser-structured SS316L surfaces by high-resolution XPS analysis.
Metabolic activity, proliferation, and alignment of HUVECs
Metabolic activity – HUVECs showed an increased metabolic activity during the first 24-hour incubation period on d25 and d75 surfaces compared to the electropolished surface, as described in Figure 6. The metabolic activity remained significantly higher on the d25 surface after 3-day incubation period. The metabolic activity of HUVECs remained to be notably higher on d25 surface following 5-day incubation period compared to the control and other surfaces. At the same incubation period, the metabolic activity of HUVECs on d150 surface was significantly decreased while d75 surface was comparable to the control.
Relative metabolic activity measurement of HUVECs cultured on electropolished and laser-structured SS316L surfaces. Relative measurement refers to standardisation of the measured values in which control group was set to 1. *p < 0.05 with one-way ANOVA followed by Tukey's post-test (n = 3). Electropolished SS316L was used as a control.
Cell proliferation – Following the first day of incubation period, the proliferation rate of HUVECs grown on the d25 and d75 surfaces tends to increase when compared to the surface of electropolished SS316L (control) as shown in Figure 7. At the end of 3-day incubation period, the proliferation rate of HUVECs on d25 surface was significantly higher compared to those on control, d75, and d150 surfaces. The value was maintained significantly different at the end of 5-day incubation period. Interestingly, the d150 surface showed a significantly lower proliferation rate compared to that of the electropolished samples, d25, and d75 surfaces, which was in accordance with the metabolic activity measurement results.
HUVECs proliferation on electropolished and laser-structured SS316L surfaces. *p < 0.05 with one-way ANOVA followed by Tukey's post-test (n = 3). Electropolished SS316L was used as a control.
Cell alignment study – Immunofluorescence staining permitted the visual observation of cell adhesion behaviour on the laser-structured surfaces. Visual observation showed that structured surfaces gave alignment to the HUVECs parallel to the axis of the chain-like structures, notably for d25 surface as seen in Figure 8. Similarly, the result from the quantitative measurement of cellular alignment showed that the d25 surface gave orientation to the attached HUVECs parallel to the direction of the chain-like structures as seen in Figure 9. Moreover, d75 and d150 surfaces also gave alignment to the endothelial cells and exhibited alignment angle nearby the axis of the chain-like structure although it is not as prominently observed on d25 surface.
Immunofluorescence staining of HUVECs after 5-day incubation period. Confluent layer of HUVECs layer on electropolished SS316 (A), d150 (B), d75 (C), and d25 (D) surfaces. Cells were stained with DAPI to stain nucleus in blue and rhodamine-phalloidin stain cytoskeleton in red. The scale bar equals 20 µm in white. Analysis of the alignment angles of HUVECs. The angles were measured between cell orientation and chain-like structure direction (oriented nominally at 0°). Distribution of HUVECs orientation on electropolished SS316L surface (A), d150 (B), d75 (C), and d25 (D).
Discussion
The adhesion of endothelial cells is influenced by both chemical and physical properties of a material [6,7,35]. In the present study, the adhesion of HUVECs was significantly influenced by the presence of chain-like structures generated by laser structuring. The introduction of chain-like structures would presumably change the chemical composition of the SS316L surface. However, XPS survey scan and high-resolution spectra showed no significant chemical difference between electropolished and laser-structured surfaces except a slightly higher amount of iron-hydroxide and lower amount of impurities on the laser-treated surfaces due to the melting process of the SS316L surface following the laser structuring. As oxygen on the surface has been suggested to attract endothelial cells [35], chain-like structures provided more oxygen available since the structures possess an increased surface area compared to that of the control surface.
Previous study reported that endothelial adhesion and proliferation were increased on surfaces with an elevated amount of iron, but decreased with elevated amounts of chromium [35]. Similarly, integrated nitrogen on the surface of stainless steel, mainly in the form of CrN and FeN, has been related to endothelial adhesion on the surface [35,36]. However, the current study did not observe significant chemical difference that would affect the behaviour of endothelial cells between laser structured and control surfaces. Therefore, the adhesion of HUVECs in the experiment was mainly influenced by the topography of the chain-like structure.
The topography of the chain-like structure allowed mechanical grip within the scanned tracks. The mechanical grip was provided by the presence of micron and sub-micron features on the chain-like structures. This support was minimally found on its counterpart design, grooving. In addition, the dimensional range of the chain-like structure is suitable for stent applications with the average depth (h) of 0.64 ± 0.07 μm. In contrast, the average depth of grooving could be as deep as 5 µm compromising the mechanical properties of newer generations of cardiovascular stents. Thus applying chain-like structure was a promising alternative design to promote endothelialisation for cardiovascular stent applications.
Other than the design, the density of structure apparently played a significant role in the behaviour of endothelial cells. The results are in accordance to the previous reports suggesting that the surface composed of both micro- and nano-topographies might provoke an optimised cellular behaviour in terms of better endothelialisation [37]. It was reported that the optimal endothelial cells adhesion was obtained on the surface of poly(dimethylsiloxane) with periodical grooving of 80 µm. More studies have reported similar size range of grooves (width: 500 nm–50 µm and depth: 50 nm–3.5 µm) on the adhesion of endothelial cells [5,23,35,3]7. The density of chain-like structure was considerably related to the greater surface area. The laser-structured surface constituted 7%, 13%, and 40% of total d150, d75, and d25 surfaces, respectively.
The chain-like structure was shown to give orientation to the endothelial cells because the structure provided a highly ordered pattern. HUVECs were then more likely to be aligned along the direction of the chains. It was shown that d25 surface promoted more HUVECs adhesion and alignment starting from 3-day incubation period compared to other surfaces. With this combination of structured and non-structured surface zones, the adhered HUVECs were easily oriented parallel to the direction of the structured zone. Aligned endothelial cells have been reported to be resistant against vascular inflammation, whereas dis-aligned endothelial cells at a branch or bifurcation expressed biological properties that promoted atherosclerosis [38]. Therefore, it is suggested that the alignment of endothelial cells could provide atheroprotective properties [39]. Consequently, surfaces structured with a chain-like structure applied with a distance comparable to the cell size can be considered as a potential approach to increase endothelialisation for cardiovascular stent applications.
Conclusion
The current work presents a novel surface structure, used for manipulating the cell behaviour on the surface of biomedical implants and consisting of chain-like features produced by partial melting and ablation through pulsed-laser irradiation. It is suggested that topographical changes promoted the behaviour of HUVECs. The results showed that the structured surfaces favored the adhesion, proliferation, and alignment of HUVECs compared to the control surface. There was no significant difference in chemical composition on the structured surface compared to that of the control suggesting that the behaviours of HUVECs were mainly influenced by the physical structure of the surface. Among all the laser-structured surfaces, d25 surface was found to significantly facilitate the adhesion, proliferation, and alignment of endothelial cells compared to both d75 and d150 surfaces. The optimal condition of d25 suggests that in the absence of any significant chemical changes, spacing of laser-structured surfaces close to the cell size is an adequate approach.
Footnotes
Acknowledgements
The authors would like to thank Caroline Loy from LBB for the support concerning the cellular tests, Dr Pascale Chevalier and Dr Stephane Turgeon from LBB for the surface chemical analysis by XPS, Essowe Mouzou and Yingchao Su for the electropolishing preparation of SS316L.
Disclosure statement
No potential conflict of interest was reported by the authors.