Femtosecond laser nano/micro patterning of titanium influences mesenchymal stem cell adhesion and commitment

Ti squares with dimensions of 23   ×   23 mm² were prepared from 1 mm thick sheets of titanium 6-aluminium 4-vanadium alloy. The surfaces were mechanically polished until a mirror image was obtained, after which 3 different topographies were processed using a femtosecond laser. Specific laser parameters, including the mean power (P), beam diameter, fluence (F), and speed, were adapted for each topography. The laser system was a Ti : Sa laser chain (Bright Thales), which delivers 120 fs, 800 nm pulses at a repetition rate of 5 kHz. The Ti plate's position was controlled with XYZ stages connected to a computer. The laser power P was adjusted using a half wave plate and a polarizer. The beam was focused on the sample using an 88 mm f-theta lens and the displacement of the beam was ensured with a pair of galvanometric mirrors (Scanlab). That way, an arrangement of laser-induced micropits was generated following a hexagonal array on the surface of the sample with a well-controlled gap (sub-micrometric precision) between the cavities. The number of light pulses was computer-controlled for each cavity, and we refer to it as N in the following. A circular diaphragm of diameter D was positioned on the laser path in front of the focusing unit. For the surface structuration of the micropits, the parameters were D  =  3 mm, N  =  9 pulses, P  =  35 mW (texture A), and P  =  100 mW (texture B). We obtained 30 μm diameter pits (textures A and B) distributed in staggered rows with a distance of 35 μm between the centers of 2 pits (figure 1(a)). The laser parameters for the surface with nano-ripples only (texture C, figure 1(a)) were D  =  5 mm, N  =  2 pulses, and P  =  100 mW.

Zoom In Zoom Out Reset image size

According to the type of experiment, we prepared plates of 23 mm  ×  23 mm that were totally textured for quantitative analyses, and multi-textured plates comprising the 3 textures (squares of 7 mm  ×  7 mm) as well as a polished mirror zone for imaging and video microscopy experiments. After irradiation, the titanium samples were ultrasonically cleaned in distilled water to remove metal debris and sterilized for 2 h at 180 °C for cell culture.

The measurements were performed at room temperature using distilled water as the testing medium. After cleaning, the remaining water was removed and the samples were dried at room temperature overnight. Water droplets of 20 μL were deposited on the horizontal surfaces, and the evolution of the droplet shape was recorded with a microscope camera (dnt^® Digimicro 2.0). The droplet profile was analyzed using the angle tool of ImageJ. It was evaluated in terms of the contact angle corresponding to the angle between the substrate surface and a tangent from the edge to the drop contour. The values presented are averages of at least five measurements at 10 min.

An optical 3D surface metrology system, InfiniteFocus (ALICONA), was used to image and measure the geometries (diameter and depth) of the micropits. SEM observations were obtained by a TESCAN VEGA3 SB scanning electron microscope at 20.0 kV. No sputter coating of the samples was required. 3D images of the micro pits were created from SEM images. Before SEM imaging, cellularized plates were immersed in 4% paraformaldehyde for 1 h and then washed with PBS. Dehydration was performed by immersing the plates in increasing concentrations of ethanol in distilled water, (60%, 70%, 80%, 90%, and 100%).

The fetal bovine serum (FBS) was provided by PromoCell, (Heidelberg, Germany); the rhodamine-phalloidin and bicinchoninic acid (BCA) protein assay kit were purchased from Interchim (Montluçon, France). The goat anti-rabbit Alexa Fluor 488 antibody (Molecular Probes) and PicoGreen DNA quantitation kit were obtained from Invitrogen. Unless otherwise specified, all other materials were purchased from Sigma Corp.

The mouse pluripotent MSCs line C3H10T1/2 (clone-8, American Type Culture Collection; LGC Promochem, Molsheim, France) was chosen to avoid donor-to-donor variation and cell population heterogeneity encountered with primary cell cultures. This cell line represents a homogeneous cell population without spontaneous differentiation under basal culture conditions and yet possesses many MSCs characteristics. Under the same inductive conditions as those applied to BM-MSCs, C3H10T1/2 cells have been found to show comparable or similar differentiation potentials, including osteogenic differentiation [17]. C3H10T1/2 cells were maintained in a standard medium: DMEM (Sigma D5796) supplemented with 10% FBS, 2 mM L-glutamine, and antibiotics (50 U ml⁻¹ penicillin and 50 μg ml⁻¹ streptomycin) in a humidified atmosphere of 5% CO2 at 37 °C and passaged before confluence. The cells were used between passages 13 and 16, and the plates were immersed in the culture medium for 1 h before cell seeding. The cells were seeded at a density of 4000 cells cm⁻².

For the differentiation analyses, the cultures were maintained for 2 d in a standard medium, and then switched to the specific medium. For the analysis of genic expression, the cultures were grown in a mixed medium, permissive for adipogenic and osteoblastic differentiation, composed of alpha MEM supplemented with 10% FBS, 50 μg ml⁻¹ ascorbate phosphate, 10⁻⁶ M beta-glycerophosphate, 10⁻⁸ M all-trans retinoic acid, 10⁻⁸ M dexamethasone, 5   ×   10⁻⁵ M isobutyl-1-methylxanthine. For adipogenic differentiation, 10⁻⁷ M dexamethasone, 1% insulin, and 5   ×   10⁻⁵ M isobutyl-1-methylxanthine were added. For osteogenic differentiation, the cultures were supplemented with 50 μg ml⁻¹ ascorbate phosphate, 10⁻⁵ M beta-glycerophosphate, 10⁻⁸ M all-trans retinoic acid. The culture medium was changed every 2 d.

The cells were fixed in 4% paraformaldehyde, and permeabilized with 0.1% Triton X-100 in PBS (3 min). The cells were pre-incubated in blocking solution (PBS, 1% BSA) for 15 min, and incubated with rhodamine-conjugated phalloidin (45 min) or a vinculin antibody (2 h). After incubation, they were washed and incubated with Alexa Fluor 488 conjugated goat anti-rabbit antibody for 1 h. The nuclei were stained with 1 μg ml⁻¹ DAPI. Fluorescence spectroscopy of the MSCs was carried out using a Leitz (Wetzlar, Germany) DMRB fluorescence microscope and the images were acquired with a Roper Scientific (Trenton, NJ) CoolSnapfx camera using the Meta Imaging series 4.6.6 software (Universal Imaging, Ypsi-lanti, MI).

For quantification, the freeware image analysis software ImageJ was used. Morphometric measurements and FA analysis of the cells were obtained by the 'analyze particle' program. The raw image was converted to an 8-bit file, and the background was removed. The image was then converted to a binary image. The shape of the cells was quantitatively described by the expression 4π (area/perimeter²), which assesses the deviation from circularity: a circle has a value of 1, the branched shape is defined as the ratio area/perimeter, and the cell elongation is expressed as the ratio of the major/minor axis.

For real-time fluorescence visualization, cells were labeled before seeding with 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocynanine perchlorate (DiIC₁₂; Molecular Probes). A cell suspension was prepared and labeled for 1 h with 0.4 μg ml⁻¹ DilC in a standard culture medium at 37 °C. After washing, the DiI-labeled cells were seeded on the Ti plates at a density of 4000 cells cm⁻². The cells were allowed to attach and spread for 30 min before the time-lapse recording was started on the microscope equipped with an incubator 5% CO₂, 37 °C. (AxioObserver Z1 SIP51353). For cell adhesion, images were acquired every 3 min for 3 h. For migration analysis, images were recorded every 15 min for 12 h, beginning 7 h post seeding.

RNA extraction and RT-PCR were performed on C3H10T1/2 cells to analyze the cell commitment. C3H10T1/2 cells were harvested after 2 d of preculture and 12 d in a permissive medium. The cells were lysed with 600 μl of TRIzol (Ambion, TX) and stored at  −80 °C. The total RNA was isolated (RNeasy Mini Kit, Qiagen, CA) and its quality and concentration were determined (NanodropND-1000, Thermo Scientific, NY). Upon reverse transcription (High Capacity RNA to cDNA kit, Applied Biosystems, CA), RT-PCR was performed (Step-One Plus, Applied Biosystems, CA) using Taqman primer probes (Applied Biosystems). The expression level was quantified with the delta-delta CT method and the results were reported and compared to polished titanium.

Complementary DNA (cDNA) was synthesized from 1.5 μg of the total RNA with the 1st strand of the cDNA Synthesis Kit for RT-PCR (AMV, Roche, Hague Road, Indianapolis). For quantitative Real Time PCR, 8 μl of a cDNA mixture diluted to 1 : 10 in water was subjected to real-time PCR using SYBR Green I dye (Lightcycler faststart DNA master SYBR green I, Roche, Penzberg, Germany). The reactions were performed in 20 μl of a PCR mixture containing 4 μl 5  ×  Master Mix (dNTP mixture with dUTP, MgCl2, SYBR Green I dye, Taq DNA Polymerase and reaction buffer) and 2 μl of 10 μM primers. β-actin was used as a housekeeping gene.

The ALP activity was measured in MSCs cultured for 7 d in an osteogenic medium. The cells were harvested in a lysis buffer (Tris-NaCl-0.1% TritonX100, pH9) and sonicated. The assay consisted of hydrolyzing p-nitrophenyl, and calorimetrically determining the product (p-nitrophenol) at 405 nm. The results were expressed as absorbance values recorded per minute and they were normalized by the DNA content.

The PicoGreen DNA quantitation kit was used to measure double-stranded DNA concentrations in solution. All the reagents (dsDNA reagent, TE buffer: 200 mM Tris–HCl, 20 mM EDTA, pH7.5 and lambda DNA standard) were obtained from the kit and the assay was performed as outlined in the protocol from the manufacturer. The data were corrected for cell-free values. The samples were placed in a black 96-well plate and excited at 485 nm. The emission was measured at 538 nm using a fluorometer.

The cells were cultured in an adipogenic medium for 7 d, after which they were incubated for 10 min with Nile Red (0.25 mg ml⁻¹), a fluorescent marker for intracellular lipid droplets. After washing, the cultures were observed under a Leitz (Wetzlar, Germany) DMRB fluorescence microscope. The images were acquired with a Roper Scientific (Trenton, NJ) CoolSnapfx camera using the Meta Imaging series 4.6.6 software (Universal Imaging, Ypsi-lanti, MI).

The data were analyzed with R software. One-way analysis of variance ANOVA followed by the Tukey post hoc test were used for statistical evaluation of the effect of the different topographies (p  <  0.05). For smaller samples size (PCR data, n  =  6), group comparisons were performed with Kruskal–Wallis analysis of variance. Post hoc comparisons between individual samples were performed with the Mann–Whitney test with Bonferroni correction.

The processing parameters of the FS laser were adjusted to control the micro- and nano-topography, as detailed in the Materials and methods section. Three patterns were created: 2 surfaces with both micropits and nano-ripples (textures A–B), and one texture with only parallel nano-ripples (texture C). On texture C, a higher magnification of the SEM image showed periodic waves that were about 600 nm wide with directional orientation (figure 1(a)).

The 3D imaging reconstruction of SEM images illustrated the location of the nano-ripples at the bottom of the pits on texture A, and around the pits on texture B (figure 1(c)). The 3D geometry of the micropits was measured with the optical 3D surface metrology system. The micropits have an average depth of 800 nm and a diameter of 30 μm (figure 1(b)). The average depth of the nano-ripples was 200 nm (figure 1(b)).

Contact angle measurements were used to characterize the wettability of the different surfaces by following the shape of a water drop resting on a horizontal surface. On surfaces with micropits (textures A–B), the wettability was similar to the mirror-polished surface. The surface featured with ripples (texture C) was slightly more hydrophilic than the others (figure 1(d)). Because of its anisotropy, texture C guided the spreading of liquid in the direction of the nano-grooves formed by the ripples, leading to lower contact angle values in this direction. Such an orientation-dependent wettability was also observed by Cunha et al [18].

24 h after seeding the MSCs, cell spreading was achieved, and specific morphologies could be observed both by SEM and fluorescent microscopy after actin labeling. The cells were largely spread on the polished surface, whereas they looked smaller and branched on textures A and B, displaying a large polygonal morphology as compared to cells cultured on texture C, which exhibited a more elongated, spindle-like morphology (figure 2(a)).

Zoom In Zoom Out Reset image size

Morphometric parameters were obtained on cells labeled with phalloidin-rhodamine (figure 2(b)). Cell area measurements indicated that the cells grown on all the textures were significantly smaller in size than their counterparts grown on polished plates. The area/ perimeter ratio was significantly decreased on all the textured materials, reflecting a more stellate cell shape. On texture C, the cells were the smallest, and the elongation factor, expressed as major axis /minor axis, was increased compared with all the other surfaces.

Cytoskeletal organization after actin labeling could be observed at a higher magnification (figure 2(c)). On the polished surfaces, actin stress fibers were sparsely arranged throughout the cell body. On texture A, actin was reinforced all around the pit periphery, underlining the design of the pits (see figure 2(c), texture A insets). On texture B, the actin fibers were thin and less abundant, substantially underlining the cell periphery and the extremities of the cytoplasmic extensions. On the nano-ripples (texture C), the network was poorly organized across the cell body and only cortical actin was reinforced. These results showed that cytoskeletal organization was influenced by the surface topography.

Measurements of the location of nuclei related to the pits on textures A and B were performed after DAPI staining. It was found that nuclei were randomly distributed, without any preference for pits (data not shown).

Close examination using SEM imaging (figure 2(a)) rendered it possible to visualize the cell adhesion in relation to the nano-ripples that were differently located from the pits. On texture A, the extremity of the cytoplasmic extensions was preferentially in contact with the smooth zones separating the pits. These observations were confirmed by analyses of FAs immunostained for vinculin (figure 3(a)). On 200 analyzed FAs, a manual quantification of the FA distribution related to pattern A demonstrated that 75% of these FAs were located on the area devoid of nano-ripples. On texture B, the ends of the cytoplasmic protrusions were found to be frequently settled inside the pits (figure 3(a)).

Zoom In Zoom Out Reset image size

A quantification of the size and number of FAs from the images after vinculin staining (figure 3(b)) exhibited significant differences: on TiL, the FAs were distributed in small patches all around the cell, whereas on the textured surfaces, the FAs were located at the edges of the extensions. On texture A, the FAs were organized in clusters: the measurement of the area of focal contacts showed a 25% increase in their size compared to all the other surfaces (figure 3(c)). A quantification of the number of FAs per cell indicated that their number decreased on all the textures compared to the control, and that the amount differed significantly between textures. On texture C, the number of FAs was significantly lower than on the other two surfaces (figure 3(c)).

For these experiments, all the patterns were located on a single plate, allowing simultaneous time-lapse microscopy analysis, thus reducing experimental variability. In an attempt to better describe the cell behavior observed on the textured surfaces, time-lapse microscopy was performed to track cell geometry changes during early cell adhesion, between 30 min and 3h30 post seeding.

For the cell morphology, only representative images with 30 min spacing are presented. A different evolution of the cell morphology was observed during the first 3 h, depending on the substrata (figure 4(a)). On the smooth surface, the cells maintained a round shape (circularity factor  =  0.55) and the cell area grew slowly (figure 4(b)). On texture A, the MSCs rapidly acquired a stellate shape. Cytoplasmic elongations were visible at 2h30, i.e. the development and retraction of cells resembling spider protrusions on smooth zones, externally to the pits. On texture B, the aspect ratio (major/minor axis) was poorly modified between 30 min and 3h30, but the cellular periphery acquired small irregularities. The circularity factor reflected this shape, with the lowest value on texture B at 2h30. On texture C, a noticeable elongation was observed as early as 2h30 with a significant increase in the aspect ratio versus the control. Moreover, the aspect ratio continued to increase over time, reaching a value 1.6-fold that of the control at 3h30.

Zoom In Zoom Out Reset image size

The spreading dynamic is expressed here by the delta of the area at time T with a reference area at 30 min. We considered that 30 min was necessary for the cell to attach via physical cues (figure 4(b)). Globally, the speed of early spreading adhesion was higher on the textured substrates. The presence of pits was felt by the cells: the MSCs on textures A or B with pits spread 2 times faster than their counterparts on a surface with only nano-ripples (texture C).

Seven h after seeding, the cells cultured on both plane and patterned substrates spread widely, as shown in figure 5(a). At this time, cell tracking was used to quantify the migration speed of the cells seeded on the different textured biomaterials. On texture B, the motility behavior was the same as the polished control. On textures A and C, the cells moved more rapidly (approximately 40%) than on the smooth surface. The migration was not influenced by the presence of pits, considering that the cells moved equally on textures A and C. However, the cells took into account the localization of the nano-ripples: cell migration was faster in the case where the nano-ripples were at the bottom of the pits (texture A) rather than at the pit periphery (figure 5(b)). The tortuosity, measured by the total distance/ straight distance ratio, was lower on texture C; in this case, the parallel orientation of the nano-ripples guided the cell migration (figure 5(c)).

Zoom In Zoom Out Reset image size

The next part of the study explored how the topographies drove the fate of the MSCs toward either osteogenesis or adipogenesis. C3H10T1⁄2 cells were grown in the different surfaces for 12 d in the presence of a permissive culture medium. To investigate the osteoblastic commitment, we examined the gene expression of RUNX2 and OCN. The key bone markers expressed at different stages of osteoblast differentiation were evaluated by qPCR analysis (figure 6(a)). The master gene RUNX2 was up-regulated for textures A and B compared to the control. An up-regulation of osteocalcin was induced by all the textured surfaces at day 12. This late osteoblastic marker increased 2-fold on texture A and B, and 1.4-fold on texture C. In addition, the ALP activity normalized by the DNA level was enhanced after 7 d of culture in an osteogenic medium in all textures (figure 7(a)).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The levels of the expression of adipogenic markers were evaluated by qPCR analysis at day 12 under a permissive culture medium (figure 6(b)). We noted that adipogenesis was equally inhibited for all the textured surfaces: approximately a decrease by half for PPARϒ2 and by more than 3 for C/EBPα. The results were confirmed by measuring the lipid accumulation in the MSC cultures grown in an adipogenic medium. Wide-field epifluorescence images after Nile Red staining demonstrated a lipidic production under all conditions at day 5 (not shown), which increased at day 7 (figure 7(b)). The amount of lipids was largely decreased on the textured surfaces compared to the smooth one. This cannot be attributed to a difference in cell number, as the numeration of nuclei at this time was the same for all the surfaces (data not shown). The down-regulation of lipid production was higher on the multi-scale textures (A and B) than on texture C (figure 7(b)).