The interaction between the osteosarcoma cell and stainless steel surface, modified by high-fluence, nanosecond laser pulses

The irradiation of metallic surfaces by high- ﬂ uence laser pulses in an oxygen-containing atmosphere inevitably modi ﬁ es the surface topography, chemistry, and wettability. These modi ﬁ cations signi ﬁ cantly in ﬂ uence cell-surface interactions and, consequently, surface biocompatibility. We investigate how surface texturing by high- ﬂ uence nanosecond laser pulses from a Nd:YAG laser (wavelength of 1064nm) in ﬂ uences cell adhesion and morphology with the aim of assessing its impact on initial cell behaviour. Quantitative and qualitative analysis of osteosarcoma cell adhesion, viability, and cell morphology were evaluated after 24-hour exposure to non-treated and laser-textured stainless-steel (AISI 316L) surfaces by ﬂ uorescent and scanning electron microscopy. The results reveal that this, initial interaction between the cells and the laser-textured surfaces leads to round shaped cells with a smaller footprint. Contrarily, on the non-processed stainless-steel and control-glass surfaces the polygonal, highly elongated, and ﬂ attened cells are observed. The cells on the laser-textured surfaces are less dendritic, with short tubular protrusions and an overexpression of extracellular vesicles, which are rarely found on non-treated and control samples. This likely happens due to the formation of nanostructured, high-tem-perature oxides that are induced by laser ablation. The analysis by X-ray photoelectron spectroscopy reveals that the laser-textured stainless-steel surfaces contain Cr hexavalent oxide, which is more toxic than the native oxide layer on the non-processed samples.

S-2  Figure S1. SEM image of the tilted (65°) non-treated sample. Figure S2. SEM image of the laser-textured surface and surface oxide layer, induced by laser texturing. Figure S3. IFM 3D height image of the non-treated and laser-textured surfaces with roughness measurements. Figure S4. SEM images of the non-treated and laser-textured surface with roughness measurements. Figure S5. Cross-section SEM images of the non-treated and laser-textured surface. Table S1. Surface roughness parameter (Ra and Rz) values. Table S2. EDS O element analyses on the non-treated and laser-textured surface (wt. %). Figure S6. Fluorescence Figure S12. SEM images of the MG63-cell distribution and shapes on the non-treated sample at two magnifications. Figure S13. SEM image of the MG63-cell morphology on the non-treated sample. Figure S14. SEM images of the MG63-cell morphology on the non-treated sample. Figure S15. SEM images of the MG63-cell distribution and shapes on the glass sample at two magnifications. Figure S16. SEM images of the different MG63-cell shapes on the glass sample at different magnifications. Figure S17. SEM images of the MG63-cell morphology on the glass sample at different magnifications. Figure S18. SEM images of the MG63-cell distribution on the laser-textured sample at two magnifications. Figure S19. SEM images of the MG63-cell morphology on the laser-textured sample. Cells are more round shaped and covered with extracellular vesicles.

S1 Surface morphology and topography measurements
As explained in the main manuscript, surface morphology, topography, and roughness of the nontreated and laser-textured 316L samples were examined by using SEM (surface morphology imaging) and IFM (3D surface roughness measurements). Surface topography is a threedimensional parameter and describes the morphological pattern of a surface. Commonly used surface roughness parameters usually describe only one aspect of the surface topography, surface height variation (Ra and Rq, Rz, Rmax), or spatial distribution and shape of the surface features (Rsk, Rku, Sds) [S1]. However, surfaces with clearly different structures can have similar Ra values. Additionally, the topography can significantly vary depending on the scale of the analysis. Figure S1 reveals that surface of the non-treated sample exhibits a typical grain-structured-like morphology with very smooth grains (< 20 µm) separated with narrow crevices (< 1.5 µm).
On the surface of the laser-textured sample, a hierarchical micro-and nanostructured hightemperature oxide is observed at higher magnifications. As visible in Figure S2, the oxide appears on the surface as hairy nanostructured layer reminiscent of a dandelion light. When removed due to different causes, native oxide is forming instead.
Surface area roughness (Sa) reveals that the laser-textured sample is much rougher due to the laserinduced morphology. However, on the laser-textured surface due to the desired topography the roughness along laser path is different than the roughness perpendicular to the laser scanning path.
To enable better comparison between the non-treated and laser-textured surfaces, the arithmetic average roughness (Ra) and mean peak to valley height of the roughness profile (Rz) is also evaluated on both surfaces. However, in some cases Ra is not a good measure of roughness due to specific topography of a surface (topography, oriented in one direction). On the non-treated surface, without any preferential topography, Ra was measured perpendicular (Rap) and parallel (Ral) with respect to the sample ( Figures S3-S4), while on the laser-textured sample Ra was measured perpendicular (Rap) and parallel (Ral) to the direction of the laser-texturing.
No differences in surface roughness are observed on the non-treated sample at different directions of measurement. On the other hand, significant difference is observed on the laser-textured sample (Table S1). Laser-textured sample is much rougher perpendicular to the laser-texturing, while parallel Ral is lower. On the laser-textured surfaces, nanoroughnessas additional level of roughnessappears due to formation of specific high-temperature oxides.
A cross-section of both samples is shown in Figure S5 for better visual comparison of the surface profiles on the non-treated and the laser-textured samples.
S-4    S-6  Table S2 provides numerical values of the O element, measured by EDS on the points, shown in Figure 2 in the main text. Lower values on the non-treated sample indicate that the oxide layer is significantly thinner compared to the laser-textured sample.

S3 Fluorescence microscopy imaging (quantification and viability assay)
To quantify the number of cell adhered and their viability, the samples with attached cells were

S4 SEM of cell adhesion pattern and cell gross morphology
Fluorescence microscopy imaging was followed by SEM observations of the attachment pattern and morphology of MG63 cells on the non-treated and laser-textured surfaces and compared to the control glass surface. From the tested samples, we selected additional images of the most representative cell adhesion pattern and morphology. MG63 cells on the non-treated sample exhibited random orientation ( Figure S12) and the majority of cells were polygonal, highly elongated, and flattened ( Figure S13). Cell tubular interconnections known as tunnelling nanotubes (TNT) were observed connecting membranes of neighbouring cells ( Figure S14). Similar results, random orientation, firmly attaching cells occupying big surface area were observed on control glass surface ( Figure S15). Shape of cells and cell surface characteristics are also similar on both surfaces ( Figure S16 and Figure S17).