A. MDA-MB-231 cell culturing on microstructured substrates

Figure 1(a) shows the SEM micrograph of a microstructured silicon area, surrounded by unprocessed flat silicon. The microstructured area is shown in higher magnification in Fig. 1(b), as the side (45°) view. A quasiordered array of conical microspikes covers the structured surface. The average height of the spikes is 35 ± 4 μm, the average half-height width is 11 ± 2 μm, and the average distance between neighboring spikes is 16 ± 2 μm center-to-center, which results in a spike density of 8.6 × 105 spikes/cm2. This morphology was obtained when irradiating silicon in SF6 ambient with a fluence of 1 J/cm2 and translating the silicon surface with respect to the laser beam with a speed such that each spot on the silicon surface was irradiated by 1000 laser pulses.59,6059. M. Kanidi, A. Papagiannopoulos, A. Matei, M. Dinescu, S. Pispas, and M. Kandyla, Appl. Surf. Sci. 527, 146841 (2020). https://doi.org/10.1016/j.apsusc.2020.14684160. G. Chatzigiannakis, A. Jaros, R. Leturcq, J. Jungclaus, T. Voss, S. Gardelis, and M. Kandyla, ACS Appl. Electron. Mater. 2, 2819 (2020). https://doi.org/10.1021/acsaelm.0c00492 At this fluence, this is the result of melting, ablation, and interference effects that occur upon the nanosecond laser irradiation of silicon in a SF6 environment.55,5655. D. H. Lowndes, J. D. Fowlkes, and A. J. Pedraza, Appl. Surf. Sci. 154, 647 (2000). https://doi.org/10.1016/S0169-4332(99)00369-456. B. R. Tull, J. E. Carey, E. Mazur, J. P. McDonald, and S. M. Yalisove, MRS Bull. 31, 626 (2006). https://doi.org/10.1557/mrs2006.160The results of MDA-MB-231 cell attachment on the microstructured silicon surface, relative to a flat silicon surface [Fig. 1(c)] and a control glass coverslip [Fig. 1(d)], are shown in Fig. 2. Figure 2(a) shows a fluorescent microscope image of cell attachment on the coverslip. We observe that cells were uniformly spread throughout the surface, which was identical to the behavior of the cells on the flat silicon surface [Fig. 2(b)]. On the contrary, surface microstructuring through the introduction of anisotropic microcones seemed to act as a repellent for the TNBC cells [Fig. 2(c)]. For a direct comparison, Fig. 2(d) shows the results of cell adherence on the border between a flat (right) and a microstructured (left) silicon surface. We clearly observe that the flat silicon side is cell-adhesive and the microstructured silicon side is cell-repellent. These observations were further confirmed by a quantitative analysis, which showed 4028 ± 110.7 (N = 3) cells on the glass coverslip (control surface), 4639 ± 196.8 (N = 3) cells on flat silicon, and 1498 ± 69.04 (N = 3) cells on the microstructured silicon. When the cells on the microstructured silicon were compared to those on the flat silicon on the same sample [Fig. 2(d)], there was a significant difference in the adhesion of cells between them [222 ± 18.19 (N = 3) to 1797 ± 99.51 (N = 3), respectively].Similar structures have been employed for the growth of neuronal cells, where they have been shown to promote and direct neuronal outgrowth;6161. C. Simitzi, E. Stratakis, C. Fotakis, I. Athanassakis, and A. Ranella, J. Tissue Eng. Regen. Med. 9, 424 (2015). https://doi.org/10.1002/term.1853 however, in the case of the TNBC MDA-MB-231 cells, they hinder cell attachment. This is probably due to the inherent hydrophobicity of these surfaces,6262. C. Simitzi, P. Efstathopoulos, A. Kourgiantaki, A. Ranella, I. Charalampopoulos, C. Fotakis, I. Athanassakis, E. Stratakis, and A. Gravanis, Biomaterials 67, 115 (2015). https://doi.org/10.1016/j.biomaterials.2015.07.008 which act as a barrier for migratory cells, a finding which is further confirmed by other studies where microstructures and their effect on cell behavior were investigated.2626. H. Miyoshi, J. Ju, S. M. Lee, D. J. Cho, J. S. Ko, Y. Yamagata, and T. Adachi, Biomaterials 31, 8539 (2010). https://doi.org/10.1016/j.biomaterials.2010.07.076 Laser-structured silicon microspikes in SF6 have been covered by a surface layer with a sulfur concentration of 0.5%–0.7%.63,6463. C. H. Crouch, J. E. Carey, M. Shen, E. Mazur, and F. Y. Genin, Appl. Phys. A 79, 1635 (2004). https://doi.org/10.1007/s00339-004-2676-064. C. H. Crouch, J. E. Carey, J. M. Warrender, M. J. Aziz, M. Shen, E. Mazur, and F. Y. Genin, Appl. Phys. Lett. 84, 1850 (2004). https://doi.org/10.1063/1.1667004 Additionally, because the samples used in this work were exposed to air for several days before being used as cell culture substrates, a native oxide layer was also formed on the surface—thus, the MDA-MB-231 cells interact with the oxidized surface. Therefore, the presence of sulfur underneath the oxide layer is not expected to affect the results of the cell behavior. Indeed, preliminary data for MDA-MB-231 cell culture on nanosecond laser-structured silicon microspikes in vacuum (not shown here) also demonstrated a similar cell-repellent behavior.Increasing the nanosecond laser fluence, we created crater structures on the surface of silicon. Figure 3 shows an area on the silicon surface, which has been irradiated by 1000 pulses with a laser fluence of 2 J/cm2 in the SF6 ambient. The employment of a high laser fluence led to the formation of a crater at the center of the irradiated area, with the appearance of various other surface topographies as we move towards the periphery of the crater, due to the Gaussian profile of the laser beam fluence. The major diameter of the elliptical crater is 870 μm and its depth is 530 ± 20 μm. As the laser fluence decreases radially, the central crater is surrounded by a zone of silicon microspikes, similar to the ones generated with a lower laser fluence (Fig. 1), which is succeeded by a zone of microripples, followed by a zone of resolidified silicon after melting.6565. A. Cavalleri, K. Sokolowski-Tinten, J. Bialkowski, M. Schreiner, and D. Von der Linde, J. Appl. Phys. 85, 3301 (1999). https://doi.org/10.1063/1.369675The examination of cell attachment in and around these craters revealed that the cells tend to concentrate more in the circumference of the crater (Fig. 4). Indeed, the cell population density appears slightly lower inside the crater rather than on the circumference and the flat silicon area surrounding the crater.The crater effect on cell behavior becomes more pronounced for another crater morphology, which can be obtained by changing the environment in which laser irradiation takes place. Figure 5 shows an area on the silicon surface, which has been irradiated in vacuum, while the rest of the experimental conditions were kept identical to the conditions employed for the crater shown in Fig. 3 (1000 pulses, fluence 2 J/cm2). For this crater, too, we observe the same effect as in Fig. 3, i.e., an elliptical crater with a major diameter of 860 μm and depth 130 ± 20 μm at the center of the irradiated area, followed sequentially by microspikes, microripples, and finally, molten and resolidified silicon toward the edges. However, the fact that the crater shown in Fig. 5 was fabricated upon laser ablation in vacuum, results in a strikingly different morphology at the center, relative to the crater fabricated in the SF6 ambient, as the former is shallower with vertical walls surrounding a central area which is decorated with microripples.Figure 6 shows the behavior of MDA-MB-231 cells when seeded in the crater presented in Fig. 5. We observe that the cells prefer to adhere to specific sites around the crater, avoiding the zone with the microspikes and developing efficiently on a zone with submicrometer roughness, as we show below. The adhesion of cells was similar to the flat silicon surface and the glass coverslip used as the positive control surface. This effect was further verified by quantitative analysis, which showed that 652 ± 15.63 (N = 3) cells were located around the crater (on the microripple zone), whereas 600 ± 19.43 (N = 3) cells were attached inside the crater. Taking into account the different areas of the crater and the microripple zone, the density of cells on the microripple zone is higher, indicating the preference of cells toward this topography.We studied the behavior of the cells around the crater shown in Fig. 5 in more detail by creating a series of four such craters side-by-side, as shown in Fig. 7. The four craters present a similar morphology to the single crater shown in Fig. 5, with a central flat cavity, surrounded by a zone with microspikes, followed by a zone of microripples. The circumference of the craters is shown in more detail in Fig. 7(c), which shows that the microspikes protrude above the surface of flat silicon, while the microripples have a significantly smaller height than the microspikes and a smaller curvature. Figure 7(d) shows a magnification of the microripples, where we observe the shape of the microripples and the fact that they are decorated with submicrometer roughness on the surface.Cell adhesion on the series of four craters is demonstrated in Fig. 8, which shows again that the cells developed increased adherence capacity toward the areas with the microripples. Furthermore, the difference between microripples and conical microspikes in terms of cell preference becomes evident, as cells congregated in the rippled topography (indicated by rectangles), with a submicrometer surface roughness and a different aspect ratio than the microspikes, rather than the spiked one (indicated by circles). Similar to Fig. 6, quantification analysis showed that 994 ± 67.87 (N = 3) cells adhered around the crater, while 691 ± 22.59 (N = 3) cells localized inside the crater. The fact that different types of microstructures can affect cancer cell adherence, and consequently, proliferation and growth, in various ways is further confirmed by the results obtained by other groups.6666. M. Ermis, D. Akkaynak, P. Chen, U. Demirci, and V. Hasirci, Sci. Rep. 6, 36917 (2016). https://doi.org/10.1038/srep36917 Therefore, the next step is to study cellular attachment on substrates with nanomorphology, as the above observations indicate that a substrate morphology exhibiting roughness at the nanoscale is significantly more favorable for cell adherence.

B. MDA-MB-231 cell culturing on nanostructured substrates

We obtained nanostructured silicon substrates by employing a femtosecond laser system. Femtosecond laser irradiation of silicon in water, as described in , resulted in the formation of columnar nanopillars on the surface, as shown in Fig. 9.67,6867. M. Kanidi et al., J. Phys. Chem. C 123, 3076 (2019). https://doi.org/10.1021/acs.jpcc.8b1035668. D. G. Kotsifaki, M. Kandyla, and P. G. Lagoudakis, Appl. Phys. Lett. 107, 211111 (2015). https://doi.org/10.1063/1.4936600 This is the result of ultrafast melting and interference effects, which occur only with femtosecond laser irradiation and not with nanosecond irradiation.69,7069. M. Shen, J. E. Carey, C. H. Crouch, M. Kandyla, H. A. Stone, and E. Mazur, Nano Lett. 8, 2087 (2008). https://doi.org/10.1021/nl080291q70. D. G. Georgiadou, M. Ulmeanu, M. Kompitsas, P. Argitis, and M. Kandyla, Mater. Res. Express 1, 045902 (2014). https://doi.org/10.1088/2053-1591/1/4/045902 The nanopillars were monolithically formed on silicon with a mean pillar diameter of 168 ± 33 nm, a mean height of 476 ± 68 nm, and a mean distance between neighboring nanopillars 257 ± 56 nm center-to-center.Figure 10 shows the results of cell attachment on the nanostructured silicon surface. Two different rectangular areas of laser-processed silicon, similar to the one shown in Fig. 9(a), coexist on the same wafer and are surrounded by a flat silicon area. The MDA-MB-231 cells showed a remarkable preference for the nanostructured areas and accumulated on the nanopillars to a great extent. Specifically, 493 ± 3.844 (N = 3) [Fig. 10(a)] and 465 ± 2.517 (N = 3) cells [Fig. 10(b)] were attached to the nanostructured areas, while 3 ± 1.155 (N = 3) and 12 ± 2.082 (N = 3) cells, respectively, were found on the flat areas aside. This is in stark contrast to the behavior of the cells on the microspikes shown in Fig. 1, which proved to be cell-repellent compared to the surrounding flat silicon area. Therefore, the nanoscale proves to be a naturally favorable topographical scale for MDA-MB-231 cell adherence. Nanoroughness is considered to be the closest to natural tissue morphology with a positive effect on cell adhesion, growth, and maturation. In human venous endothelial cells, it has been shown that increasing the roughness of the biomaterial surface at the nanometer scale can enhance cell adhesion and growth.2222. S. Cai, C. Wu, W. Yang, W. Liang, H. Yu, and L. Liu, Nanotechnol. Rev. 9, 971 (2020). https://doi.org/10.1515/ntrev-2020-0076 Mesenchymal MDA-MB-231 cells show maximal response to the influence of substrate morphology for cue dimensions close to the focal adhesion size.1414. C. Leclech and C. Villard, Front. Bioeng. Biotechnol. 8, 551505 (2020). https://doi.org/10.3389/fbioe.2020.551505