2.1.

Fabrication of Cell Array Platform

The cell array platform was fabricated by coating a glass plate with a thin layer of platinum and gold sequentially. The 30-nm gold coating was achieved by physical vapor deposition in an electron beam evaporation system. Before that, the same glass slide was over-layered by an extremely thin 7-nm platinum layer to achieve a satisfactory attachment of gold particles. The gold layer over the substrate functioned as the thermal transducer layer to achieve proper photothermal ablation. Next, a cytophobic layer was formed over the transducer layer by soaking the substrate in a 0.2% ($\mathrm{w}/\mathrm{v}$) solution of 2-methacryloxyethyl phosphorylcholine (MPC) polymer in ethanol. Upon drying, the substrate was ready for use. When the laser was focused, the transducer layer absorbed light energy and produce a local heat of high degree at the laser spot. At a rather high temperature, the cytophobic polymer and transducer layers were depleted, subsequently exposing the cytophilic glass layer.

2.2.

Experimental Setup

The setup of the laser micro-engraving system consists of a diode-pumped Q-switched ${\mathrm{Nd}:\mathrm{YVO}}_4$ laser, piloting diode laser, pair of scanning galvano meter mirrors, and water immersion objective lens with a numerical aperture (NA) of 0.3 [Fig. 1(a)]. Central wavelength of the Q-switched laser is 1064 nm, with a peak power of around 220 mW. The ${\mathrm{Nd}:\mathrm{YVO}}_4$ laser beam was focused on the transducer layer under the guidance of the piloting diode laser of wavelength 653 nm through a $10\times$ water-immersing objective (Olympus, Japan) coupled in a modified Olympus microscope system (Olympus, Japan) [Fig. 1(a)]. The maximum detectable laser power was 120 mW at the focal plane. Based on a computer control by marking mate, laser marking software, the laser beam was vector scanned for pattern drawing by galvano mirrors [Fig. 1(a)]. The patterns were designed on computer graphics as typically shown in Fig. 1(b), where 136 circles (20 μm) are written.

Fig. 1

Laser engraving system for biological applications: (a) schematic and (b) circular array patterns of the laser marking system.

2.3.

Cell Culture

Cervical carcinoma HeLa cells were cultured in Dulbecco’s modified eagle medium supplemented with 10% fetal bovine serum, streptomycin, and penicillin. Next, cells were seeded on the substrates with a concentration of $2000\text{cells}/{\mathrm{cm}}^2$. Culturing was performed in humidified atmosphere at 37°C with 5% ${\mathrm{CO}}_2$.²⁹ For further dynamic alteration of cell adhesiveness, the laser was carefully irradiated on the substrate at a safe distance from the preresiding cells to prevent cell damage.

2.4.

Cell Observation

Cells on the respective micro-domains were observed under a normal wide field microscope (Olympus, Japan) with $20\times$ objective of NA 0.4. The images were taken with an Olympus digital camera “Camedia C-5060” (Olympus, Japan) that is compatible with the microscope.

3.1.

Photothermal Engraving

Cell cultivation micro-domains were fabricated through photothermal ablation by area specific removal of the MPC polymer by spot heating with a focused laser beam on a thin light-absorbing gold layer. Owing to that the polymer fails to absorb light at 1064 nm; the gold coating over the glass substrate functions as a photothermal transducer that transforms light energy into heat. The cytophobic MPC polymer film along with the underneath gold layer ablated at the laser spot exposing the cell-adhering domains. Figure 2 schematically depicts the cell adhesiveness alteration. This work first designed an array of desired pattern in silico, which functioned as the mold for the pattern engraving. Those patterns were engraved on the platform by the focused ${\mathrm{Nd}:\mathrm{YVO}}_4$ laser. The laser scanning speed and frequency were controlled through a pair of scanning galvano mirrors, with a computer system. In this work, two micropatterns were engraved: one was an array of circular domains, each with a diameter of 20 μm. The other one was rectangular, with a length of 1 mm and breadth of 150 μm. Figure 3(a) and 3(b) shows the circular array and rectangular microdomains, respectively, before and after cell seeding.

Fig. 2

(a and b) Schematic diagram of the cell adhesiveness alteration of the platform by photothermal ablation. (c) Trapped cell in cell adhesive domain.

Fig. 3

Micrograph of the laser engraved area on the platform before and after cell seeding: (a) Single cells array in circular microdomains, (b) rectangular microdomain, and (c) connection induction along with the guided patterns.

Experimental results indicated that the ablation in an aqueous condition was more confined than that of air for the same applied laser power. Figure 4(a) displays the scatter plot of the pattern width dependency on laser power and environmental conditions, indicating that the ablation was feasible above 5 mW laser power for both air and water conditions. Upon increase in power, the size of the ablated area increased and, at one point, reached saturation value. For engraving under an aqueous condition, 10 mW was the saturation power above which the pattern sizes remained at 20 μm. Conversely, in air, the saturated size of 30 μm was measured at 35 mW power. To better determine the laser substrate interaction, we used two different concentrations of MPC polymer 0.1% and 0.05% in alcohol to compare the size of the ablated domain with the substrate without the polymer layer when the gold coating was 30-nm thick. Figure 4(b) depicts the scatter plots of the domain size (in diameter) of the engraved pattern in response to different MPC concentration for the same applied laser power. Experimental results show that without the polymer layer, the engraved domains have larger size than that in polymer layered substrates. However, the variation in polymer concentration did not cause detectable differences in domain size. We used gold layers of two different thicknesses, 30 and 50 nm, over-layered with 0.1% MPC polymer. The results indicate that laser ablates a larger area in 50-nm gold-coated surface than the 30 nm one for the same applied laser power. Figure 4(c) depicts the scatter plots of power against the diameter of the engraved domains for the two different thicknesses. This is attributed to the increase in light absorption due to deeper thickness.

Fig. 4

Scatter plots of microdomain diameter as a function of the applied laser power. (a) The substrates were coated with 30-nm gold layer and over-layered by 0.1% MPC polymer. The difference is attributed to variation of thermal mass in air and water, respectively. (b) Comparison of no polymer over-layer and different concentrations of the MPC polymer over-layer under water. The substrate was coated with 30-nm thick gold layer. The polymer over-layer helps to contain the gold coating under ablation. (c) The effects due to different thickness of gold coating over-layered with 0.1% MPC polymer under water. Thicker coating would result in higher absorption and thus larger domain size. The error bars denote the standard error of the mean in diameter measurements.

The surface analysis of the laser engraved domain, in comparison to the unabated surface, was carried out by an atomic force microscope (AFM) and scanning electron microscope (SEM) (Fig. 5). Analysis of the results indicated that the ablated domain become rougher than the unabated gold MPC-coated surface that was observed by the pillars of the AFM histogram inside the engraved domain [Fig. 5(a)]. When the gold only coated surface was ablated, the number of pillars was less [Fig. 5(b)]. Figure 5(c) is the SEM image of the laser engraved area, where the bright area was the unabated region, rich in metallic gold, which gives bright a SEM image due to secondary electrons, in contrast to the dark area that lacks gold after ablation. Higher pillars were formed at the edge surrounding the engraved domain, which may be attributed to the folding of the polymer-coated gold layer as observed from the SEM image in Fig. 5(c). Some bright spots were observed inside the engraved region due to the presence of gold debris. A previous report suggests that a rough gold surface makes a better cell adherent platform than a smooth one.³⁰

Fig. 5

Atomic force microscopy (AFM) and scanning electron microscopy (SEM) surface analysis of laser ablated domain. (a) AFM image of the ablated MPC-gold composite coated domain. (b) AFM image of the ablated gold-coated domain. (c) SEM image of the ablated MPC-gold composite-coated domain.

3.2.

Cell Array

For single cell cultivation, HeLa cells were seeded over the micropatterns. This work attempted to achive a higher number of single cells in a single domain by seeding highly disperse cells with a concentration of $3000\text{cells}/{\mathrm{cm}}^2$. Comparable size of the microdomains with the cells was preferred for obtaining isolated single-cell culture. According to our results, not all the microdomains accommodated a single cell; it could occasionally have more than one cell and a few microdomains occasionally remained empty. Figure 3(a) shows such an array with single cells and empty microdomains denoted by arrows. To induce a cell-to-cell junction on the platform dynamically, engraving was performed in situ under culturing. Channels connecting the microdomains were engraved without affecting the pre-existing cells.

After the cells were cultured for 24 h on the microdomains, the polymer layer between the domains was ablated to form channels of width around 10 μm. The cultured HeLa cells elongated along with the guided channels and connected to each other within 16 h. Figure 3(c) shows the engraved channel and cell-to-cell connection.

For quantitative analysis of single cell accommodation in response to different domain size, we have engraved patterns of three different diameters. If the domain size is approximately 15 μm or less in diameter, a higher percentage of the domains remain empty after cell seeding. In a selected area, only 21% of the domains contain cells; among them 78% are single cells. We have chosen 15 μm as the smallest domain to grow cells, since our method cannot create uniformly circular domains with diameters in the order of 10 μm. For domains of 20 μm diameter, 88% of the domains contain cells, with 43% of them being single cells. For even larger sized domains around 25 μm, 97% of domains are occupied by cells, with 16% of them single cells. Figure 6 shows cells growing on patterns of three different domains.

Fig. 6

Micrograph of cells in domains of diameter: (a) 15 μm, (b) 20 μm, and (c) 25 μm.

3.3.

Effect of Laser Irradiation on a Cultured Cell

In-situ application of the laser was essential for the dynamic alteration of cell adhesiveness of the platform. Laser applications occasionally cause hypertrophy of the cells and can damage the cell if the laser focal spot is too close to the cell.^27,31 Usually, application of an IR laser reduces the probability of cell damage in comparison with using the UV laser.^32–34 However, during in-situ engraving, the safe distance between the laser spot and the pre-existing cells had to be ensured. This study examined how the laser affects the cells, in which the percentage of damaged cells for different distances of the laser spot from the cells was evaluated. Figure 7(a) shows the laser spots along with the preresiding cells at different distances starting from 300, 200, 100, 50, and 0 μm (0 μm: laser application directly over the cell). Figure 7(b) illustrates the bar chart diagram of the percentage of cells after the laser irradiation at different distances. According to this figure, $>80\%$ of the cells survived after 24 h of the laser irradiation unless the laser was directly irradiated on the cells. Upon direct exposition to laser, only 40% of the cells stood in their respective cell adhesive domains after 24 h.

Fig. 7

(a) Micrograph of cells in their respective microdomains after 24 h since laser was irradiation at different distances. (b) Bar chart diagram of cell survivability study showing the percentage of cells lived after 24 h, in which the laser was irradiated at different positions tested for 30 cells each, where “$\alpha$” denotes laser irradiation at an infinite distance from the cells.

3.4.

Cellular Migrational and Morphological Evaluation in Response to Geometrical Constraints

Cellular movement and their collective migration have received considerable interest in some of the physiological and biological processes, including embryonic development,³⁵ tissue morphogenesis,³⁶ wound healing,³⁷ and cancer metastasis.^38,39 Their approach lacked a dynamic in-situ micro-fabrication method. Depending on the individual physiological requirements, cells can move as a wide sheet for wound closing and may travel in a narrow strip for cancer metastasis.⁴⁰ Here, a cell adhesive microdomain of width 150 μm and length 1 mm was fabricated on the nonadhesive substrate. Size and shape of the wound and the cancer thus have controlling effects in cellular migrational speed and modes.^37,38 This study elucidates the role of geometrical constraint on collective cell migration by designing an in vitro assay using the proposed microfabrication approach based on the results of Vedula et al. on collective cell migration.⁴⁰ Their approach lacked a dynamic in-situ micro-fabrication method. The approach used in this study did not use external guiding biomolecules, such as fibronectin or collagen. Here, a cell adhesive microdomain of width 150 μm and length 10 mm was fabricated on the nonadhesive substrate. Figure 3(b) depicts one typical domain before and after cell seeding. HeLa cells were seeded and a mono layer was cultured for 2 days. Once the domain was filled with cells, strips were engraved with three different widths: 150, 100, and 50 μm, each with a length of 550 μm [Fig. 8(a)]. Consequently, a confluent monolayer of cells was grown in the reservoir, which was ready to flow through three guided strips. Figure 8(a) shows the confluent and its branched strips immediately after the formation of the strips; meanwhile, cells were grown inside the main microdomain. Some observable debris of dead cells and loosely attached cells were observed on the nonadhesive region, which disappeared after 2 to 3 days of culturing. Figure 8(b) displays typical micrographs of cellular migrational assay at an instance after 32 h of confluent formation. Analysis results indicated a faster cell sheet movement in a narrower strip than in the wider strips [Fig. 8(c)]. This finding complies with the same order of migration noted by Vedula et al.⁴⁰

Fig. 8

(a) Representative image of a typical confluent monolayer with the guided strips immediately after formation. (b) Micrograph of the cellular displacement at an instant after 32 h of confluent formation. (c) Scatter plot of the cellular displacement assay through the strips. The error bars denote the standard error of the mean for cellular displacement measurements.

Cellular migration can also affect the cell shape and morphology.⁴¹ Three rectangular microdomains were engraved with different widths of 150, 60, and 30 μm, to further elucidate the change in cellular morphology in relation to geometrical constraints. More elongated cells were observed along the length of the narrow domains, while the cells in the wider domains were less elongated (Fig. 9). Here, the ratio of cell length to breadth was defined as the parameter “cell shape factor” ($\beta =L/B$). The cell shape factor ($\beta$) at different micro-domains was quantified and compared with cells in a normal culture dish (defined as microdomain with a width of $\alpha \mathrm{\mu m}$) in the bar chart diagram of Fig. 9(b). According to our results, cells in wider domains have a similar size and shape with those in an open culture. The $\beta$ value of cells in wider domains and normal culture dish was found to be around 1.4, whereas when the domain width was comparable to the cell size, cells took an elongated shape along the length of the domain (where $\beta =2.7$).

Fig. 9

(a) Photograph of the cells with different morphologies in domains with different geometrical configurations. (b) Bar chart diagram of cell shape factor ($\beta =L/B$) for the cells at different geometrical constrains. The error bar denotes the standard error of mean for $\beta$ value measurements.