Hierarchical TiO₂ nanotube arrays enhance mesenchymal stem cell adhesion and regenerative potential through surface nanotopography

Abstract

The concept of preconditioning mesenchymal stem cells (MSCs) under different stress conditions or with bioactive molecules is introduced to optimize their therapeutic potential. This study investigates the physicochemical effect of hierarchical TiO₂ nanotube arrays, a versatile and easy-to-prepare nanosurface, on MSC behaviour. By precisely controlling the nanotopography through anodization, we demonstrate the significant influence of surface properties on MSC adhesion, proliferation and differentiation. Electrostatic interactions between surface charge and proteins play a crucial role in these cellular responses. In addition, preconditioning MSCs under specific conditions enhances their therapeutic potential by optimizing paracrine signalling and homing properties. Higher surface charges and increasing spiky character of surface roughness of titania samples after anodization at 60 V significantly upregulated chemokine receptor type 4 (CXCR4) and vascular endothelial growth factor A (VEGFA), indicating the enhanced migratory and angiogenic potential of MSCs. The study reveals the mechanotransductive effects of nanotopography on MSC differentiation, suggesting that tailored surface features can direct cellular fate. These findings highlight the potential of hierarchical TiO₂ nanotube arrays as a promising platform for regenerative medicine, offering a novel approach to improve tissue engineering and therapeutic outcomes.

1. Introduction

Advances in biomedical sciences have led to new therapeutic approaches such as tissue engineering, regenerative medicine and cell therapy. In this context, stromal cells play a crucial role in these processes as they form and structure connective tissue in the body [1]. Mesenchymal stem cells (MSCs) are preferred for tissue repair due to their differentiation capacity, paracrine signalling and secretion of various molecules, including extracellular vesicles. These mechanisms critically treat inflammatory and degenerative diseases through the ‘homing’ phenomenon [2,3]. In addition, mesenchymal stem/stromal cells and the extracellular matrix (ECM) components they secrete adapt to their environment by influencing cell adhesion, spreading and differentiation [4,5]. An innovation in this field is the concept of preconditioned MSCs. This concept enhances the repair functions of MSCs by exposing them to various stress conditions or bioactive molecules, optimizing their use in cellular therapies [6,7]. These preconditioned cells have the potential to enhance paracrine mechanisms and homing properties, which could significantly increase the success rate of treating inflammatory and degenerative diseases [8–10].

The topographic features of surfaces significantly affect various cellular responses, including cell viability, adhesion, proliferation, differentiation capabilities, gene expression and the secretion of ECM components [11–14]. Nanotopography, a subfield of nanotechnology, is critical to understanding and controlling the interactions between cells and surface structures at the nanometre scale [15]. Relevant literature indicates that surface nano-protrusions, which create roughness, enhance stem cell viability and proliferation compared with flat surfaces [16].

The impact of nanotopography on the preconditioning of MSCs represents a significant domain of investigation, particularly within the context of tissue engineering and regenerative medicine. Nanotopography is defined as the nanoscale features on a substrate that can influence cellular behaviour, including adhesion, proliferation and differentiation. The interaction between MSCs and nanotopographical cues is mediated through various mechanisms, including mechanotransduction, which is the process by which cells convert mechanical stimuli into biochemical signals. The dimensions and configuration of nanotopographical features exert a pivotal influence on the cellular response. In a study conducted by Zouani et al., it was demonstrated that alterations in nanofeature size could dictate the differentiation pathways of stem cells, emphasizing the importance of precise control over nanotopography [17]. Similarly, Rosa et al. reported that nanotopography could induce osteogenic differentiation of MSCs through specific signalling pathways, including the α₁β₁ integrin signalling pathway [18]. This also indicates that the physical attributes of the substrate can markedly impact the biochemical pathways that are initiated within the cells. Moreover, the mechanosensitivity of MSCs to nanotopography has been extensively documented. Salvi et al. observed that MSCs cultured on specific nanotopographies exhibited increased intracellular calcium responses, which are critical for downstream signalling related to proliferation and differentiation [19]. This mechanotransductive response is further supported by the findings of Yim et al., noting that nanotopography could lead to significant changes in focal adhesion dynamics and cytoskeletal organization, ultimately affecting cell behaviour and fate [20]. Nanotopography is a powerful tool for regulating the cytoskeletal structure of cell receptors and transducing mechanical signals to the nucleus. Modulation of cytoskeletal structure occurs by direct mechanotransduction, where the ECM acts as a mechanical tension structure, or by modulating biochemicals that interact with focal adhesion kinase (FAK) behaviour and extracellular signal-related kinase (ERK1/2) [21].

Titanium (Ti) and its alloys are widely utilized biomaterials because of their high biocompatibility, having non-stimulating behaviour for macrophages, no cytotoxic or inflammatory effects, making them the material of choice for dental and bone implants [22]. Implant surfaces play a critical role in cell adhesion by interacting with ECM and cell surface receptors [23]; therefore, surface modifications can enhance titanium surfaces’ biocompatibility, promoting cell adhesion and spreading [24,25]. Among these modifications, electrochemical anodization is a surface modification method that provides a cost-effective and relatively simple approach to creating TiO₂ layers in nanotube arrays. This technique allows for precise control over the nanoscale widths and lengths of the layers. Anodization is typically performed in a two-electrode system with titanium as the anode and platinum as the cathode in the presence of fluoride ions as the electrolyte, resulting in the formation of nanotubular cavities [16]. Under the influence of an applied electric field, the initially formed TiO₂ undergoes field-assisted dissolution, creating nanopores on the surface [26]. Subsequently, the growth of tubular geometries is continuously observed along with the dissolution of hexafluorotitanate [TiF₆]^2- and the transport of TiO₂ into the substrate. The dimensions and geometries of the resulting TiO₂ layer can be adjusted by varying the process parameters, including applied voltage, electrolyte composition, pH and time [27].

The extracellular environment is of critical importance in regulating cellular behaviour. Cells are surrounded by an ECM composed of many proteins, such as collagen, elastin, fibronectin and glycosaminoglycans. Collagens modulate ECM–cell interactions by forming new networks due to their fibrous structure, while elastin and fibronectins modulate ECM–cell interactions by cross-linking with other ECM proteins. Integrins, the primary mediators of cellular interaction with the microenvironment, act as primary mechanosensory by sensing nanotopography during adhesion via thin membrane protrusions called filopodia[28–30]. Integrins are transmembrane proteins that bind to ECM ligands, which comprise various adhesion complexes and can be found in combinations due to their heterodimer structure [29]. Integrins are in a bent, inactive conformation when cells are at rest. The binding of the intracellular adaptor protein talin activates integrins, which change from bent to closed, extended to closed and finally to extended open [30–32]. Thus, a high affinity for ECM ligands is created, and adhesion continues to increase [32]. The integrin–ECM bond is cyclic, and in response to the force from this anchoring bond, they aggregate until a maximum integrin complex is reached, leading to the adhesion’s collapse [33].

Culturing cells on TiO₂ nanotube surfaces enables the study of the impact of nanometre-scale surface arrangements on cell morphology, spreading and differentiation [34]. These studies can be instructive for a better understanding of the effects of nanotopographic surfaces on cell biology and potential biomedical applications. Investigating the applicability of such surfaces in cellular therapies and understanding their effects on MSC behaviour could provide new insights into regenerative medicine and tissue engineering. Recently, our group conducted in vitro studies to investigate the effect of various titanium surface modifications on the therapeutic potential of MSCs. A transcriptomic dataset was generated to study MSC behaviour on different surfaces, followed by an investigation of their behaviour on some surfaces with advanced modifications. The study demonstrated that MSCs preconditioned on one of the developed surfaces can achieve high therapeutic potential and prompted us to undertake this study, which aimed to investigate the impact of diverse hierarchical nanotube surfaces on the assessment of MSC therapeutic efficacy. This article presents a unique perspective on applying hierarchical TiO₂ nanotube arrays with different nanotopographies obtained by anodic oxidation, which is a versatile and flexible approach to easily adjust nano-roughness. Our approach focuses primarily on the results of varying nanotopography including cell adhesion, spreading, mRNA levels responsible for migration and organ remodelling with a holistic view as putting a variety of physicochemical characteristics of the surfaces. Thus, the present study demonstrated that cells are influenced not only by nanotopographical variations but also by alterations in quantities such as wettability, surface free energy and surface charge during preconditioning, as well as by changes in cellular activities.

2. Methods

2.1. Preparation of hierarchical TiO₂ nanotubular surface decorations

Flat Ti and hTiO₂ array surfaces, along with tissue culture plates, were used at the beginning of the study to evaluate nanotopographically different properties. Pure Ti foil was cut into 3 × 1 cm² pieces for use as an anode electrode. The pieces were cleaned with detergent, ethanol (EtOH) and deionized water (DIW) in a sonication bath for 30 min. After washing, the samples were air dried. The electrolyte for the electrochemical anodic oxidation was prepared by dissolving 333 mg of ammonium fluoride (NH₄F) in 2 ml of DIW and 98 ml of ethylene glycol at room temperature using a magnetic stirrer. The anode was positioned adjacent to the platinum mesh cathode in the prepared electrolyte. A 4 cm gap separated the electrodes. Titanium was anodized at a potential of 60 V for 2 h at room temperature. Then, the initial oxide layer formed was removed using tape peeling and ultrasonic bath procedures, thus preparing a patterned surface for the subsequent fabrication of secondary tubes. The presence of macro defects on the surface renders the tube arrays unstable, lacking sufficient mechanical integration. Consequently, these arrays are removed from the surface through processes such as tape peeling or sonication [35,36]. The second step in the preparation of hierarchical nanotubes involved the application of an anodizing voltage of 20 V for 30 min. Upon completion of this process, the samples were cleaned with DIW and/or EtOH and air dried.

2.2. Cell culture

Human bone marrow-derived MSCs and human dermal fibroblasts (DFs) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) (Cat. no: PCS-500-012™, Lot: 63208778) and Gibco™ (Ref: C0135C, Lot: 2325163), respectively. The cells were cultured in growth medium (DMEM-LG, supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin and 1% L-glutamine) at 37°C under 5% CO₂ conditions. Passage 3 cells were used for all subsequent experiments. Cells were seeded onto titanium surfaces at 12 500 cells cm⁻² for the main experiments and 5000 cells cm⁻² for cell adhesion and spread tests.

2.3. Transcriptomics and bioinformatic analysis

After a 48 h incubation period, cells on all surfaces were washed with phosphate-buffered saline (PBS) solution. Total RNA was isolated using the Single-Cell RNA Purification Kit according to the manufacturer’s instructions (Norgen Biotek Corporation). RNA quality was measured using a spectrophotometer (NanoDrop One, Thermo Scientific). The SENSE Library Kit (Cat. no. 001.24, Lexogen) was used for library preparation. Clonal amplification and sequencing were performed using the Ion PI Hi-Q T2 200 Kit (Cat. no. A26434, Thermo Fisher Scientific) and the Ion PI Hi-Q Sequencing 200 Kit (Cat. no. A26433, Thermo Fisher Scientific), respectively. Samples were analysed in duplicates using the Ion Proton Instrument (Thermo Fisher Scientific). FASTQ files were processed and aligned with RaNA-Seq, an interactive RNA-Seq analysis tool, from FASTQ files to functional analysis [37]. Bioinformatics analysis used the TPM (Transcripts Per Kilobase Million) method. Differentially expressed genes (DEGs) were identified with a p‐value ≤ 0.05 and a log2 fold change greater than 1. Enrichr conducted an enrichment analysis [37] to unravel the biological context of the gene set of interest derived from the DEG analysis. The study provided a comprehensive view of pathways enriched according to Reactome pathways to underlying biological processes and pathways associated with our dataset.

2.4. Derivatization of hierarchical nanotube surfaces

A two-step anodic oxidation process was implemented on titanium surfaces to create TiO₂ nanotube arrays with different hierarchical nanotube decorations. The anodizing potential parameter was manipulated to precisely control the inner diameter sizes of the nanotubes. It was demonstrated that the width of the nanotubes can be adjusted in the range of 20−200 nm by changing the applied potential [38]. At the same time, multiple anodization steps determine the hierarchical order of the nanotubes. In contrast to the preliminary experiments, 80 V was applied for 2 h in the first step, the selection of the outer diameter of the nanotubes, in order to make the hierarchical structure more prominent. To produce surfaces with distinct nanotopographies, the second step, electrochemical anodic oxidation, was performed for 30 min at a potential difference of 20−60 V. The resulting samples were cleaned with DIW and/or EtOH and air dried.

2.5. Flow cytometry analysis

After 72 h, the cells were detached using trypsin, and cell death and reactive oxygen species (ROS) levels were analysed. Cell death was assessed using the Annexin V-FITC Kit (Beckman Coulter, Cat. no. IM2375), and ROS levels were measured using the 2', 7'-Dichlorofluorescein Diacetate Kit (Cayman Chemical, Cat. no: 20656, CAS: 2044-85-1), following the manufacturer’s instructions. The analyses were performed using a Novocyte Flow Cytometer (Agilent).

2.6. RNA isolation, cDNA synthesis and quantitative real-time polymerase chain reaction

After a 48 h incubation period, cells on all titanium surfaces (untreated control, 80–20 V, 80–40 V, 80–60 V) were washed with PBS solution. Total RNA was isolated as described above. cDNA was synthesized with the High-Capacity cDNA Reverse Transcription Kit (Thermo Scientific). Gene-specific primers were used for relative quantification of the target mRNAs, with ACTB as the housekeeping gene. Primer pairs are available upon request. Each sample was analysed in triplicate reactions. Relative levels of mRNA gene expression were calculated using the 2^−ΔΔCT method. LightCycler® 480 SYBR Green Master Mix (Roche Diagnostics) was used for all reactions according to the manufacturer’s recommendations, and samples were run in triplicates using the LightCycler® 480 System (Roche Diagnostics).

3. Results and discussion

3.1. Transcriptomic profiling of mesenchymal stem cells on initial materials

We conducted a transcriptomic analysis to understand how MSCs incubated on the surfaces mentioned in §2.1 initially respond and which gene expression changes are associated with this response. The hTiO₂ nanotube surfaces created through electrochemical anodization displayed distinct nanotopographic features that influenced MSC behaviour. The transcriptomic analysis revealed that MSCs on these surfaces upregulated genes associated with ECM production and cell adhesion. This suggests enhanced cell–surface interactions facilitated by the nanotube structures, promoting cell adhesion and spreading. The downregulation of ribosomal proteins indicates an adaptive response where the cells prioritize adhesion and ECM production over protein synthesis, possibly as an energy conservation mechanism during the initial adaptation phase. In the transcriptomic analysis, 119 genes showed increased expression, predominantly associated with collagen and ECM production (e.g. collagen type IV alpha 1 chain (COL4A1), collagen type III alpha 1 chain (COL3A1), collagen type I alpha 2 chain (COL1A2), collagen type VI alpha 2 chain (COL6A2), tenascin-C (TNC), elastin microfibril interface located protein 1 (EMILIN1) and spondin 2 (SPON2)), indicating enhanced cell adhesion and interaction with the surface. Conversely, 99 genes showed decreased expression, notably ribosomal proteins, suggesting downregulation of protein synthesis, possibly as an energy conservation mechanism while the cells adapt to the new surface (hTiO₂). Pathway analysis using Reactome revealed that the upregulated genes were primarily involved in pathways such as Collagen Chain Trimerization, Assembly of Collagen Fibrils and Other Multimeric Structures and Extracellular Matrix Organization. These pathways support the observed increase in collagen production and ECM synthesis. On the other hand, the downregulated genes were mainly associated with pathways related to protein synthesis, including signal-recognition particle (SRP)-dependent Cotranslational Protein Targeting to Membrane, Peptide Chain Elongation and Eukaryotic Translation Elongation, reflecting a reduction in protein synthesis activity (figure 1). Different studies examining stem cell behaviour in nanotopography have revealed that many biological pathways, including protein synthesis, affect differentiation and cell behaviour [39].

Figure 1.

Reactome pathway analysis of differentially expressed genes in MSCs. This figure illustrates the Reactome pathway analysis results of differentially expressed genes in MSCs. The pathways are categorized into those associated with upregulated genes (top section) and downregulated genes (bottom section).

3.2. Derivation and characterization of hierarchical surfaces

After removing the nanotube arrays produced in the first step, voids are left on the titanium surface. This serves as the basis for inner nanotube growth and, hence, the hierarchical organization. Figure 2a–c shows the scanning electron microscopy images used to measure the diameter of the nanotubes and porosity of the titania surfaces manufactured through the second electrochemical anodic oxidation process with applied potentials of 20, 40 and 60 V. The values derived via image processing are exhibited in figure 2a–c. The sample groups are abbreviated as control (Ti), 80−20, 80−40 and 80−60 V, representing samples that underwent 80 V for the first step and 20−40−60 V for the second step.

Figure 2.

Scanning electron microscopy images and the image processing results of 80−20 V (a), 80−40 V (b) and 80−60 V (c). Note that all scale bars are equal and 200 nm.

After the secondary electrochemical anodic oxidation process, the resulting porosity values were 26.6, 18.0 and 42.8% at 20, 40 and 60 V application potentials, respectively. Notably, an increase in electrical charge corresponded to the rise in inner tube radii, with measurements of 28, 47 and 68 recorded for 20, 40 and 60 V, respectively. It has been reported in the literature that the dissolution rate increases with increasing potential, leading to an increase in tube width [40]. Although a decrease was observed at 40 V, the application potential of 60 V resulted in porosity values greater than 40%. Examining the pore diameter distribution plots, it is evident that at 20 V, there are mainly 30 nm pores, which correlates with a significant volume of indentations, including 10 nm and below. In the 40 V samples, over 60% of the pores are distributed between 40 and 60 nm. In the 60 V sample group, the pores are mainly in the 70−90 nm range but with a wide distribution, resulting in an average pore diameter of 68 ± 19 nm with a standard deviation.

Figure 3 shows the surface line profiles obtained after the second step and at z-height. Noticeable differences in the diameters of the hierarchical nanotubes, as well as the sharpness of the primary nanotube walls obtained by the first 80 V anodization, were observed between the maximum profiles. As will be supported by the numerical atomic force microscopy (AFM) results given below, the heights at the top points of the patterns became clearly spiky in the sample groups where the second step anodic oxidation was 60 V.

Figure 3.

Surface profiles of different nanoporous titania surfaces after the second step of anodization: 80–20 V (a), 8–40 V (b) and 80–60 V (c).

Upon examining the wetting properties of the surfaces, a significant increase in hydrophilicity was observed following anodic oxidation, as evidenced by minimal contact angles. The surface wettabilities of the samples were determined to be 77° for the control, 12° for the 80–20 V, 10° for the 80–40 V and 8° for the 80–60 V samples. This change was expected due to the formation of an oxide layer on the surface, consistent with previous research [41]. Similarly, table 1 summarizes the increases in surface free energy and energy components according to the OWKR/Fowkes approach with increasing wettability. The polar component increased with the voltage parameter applied in anodic oxidation.

Table 1.

Tabulated surface free energy values for the control, 80–20, 80–40 and 80–60 V samples after the OWKR/Fowkes approach were applied.

	γtot [mN m−1]	γd [mN m−1]	γp [mN m−1]	WCA (°)
control	43.8 ± 0.6	38.7 ± 0.6	5.0 ± 0.6	77
80–20 V	76.90 ± 1.2	48.6 ± 0.7	28.3 ± 0.7	12
80–40 V	72.9 ± 0.6	40.9 ± 0.5	31.7 ± 0.3	10
80–60 V	76.9 ± 2.1	43.7 ± 1.6	33.2 ± 0.7	8

As the oxidation potential increased, the polar component of the surface free energy also increased in these groups. This phenomenon increases the wettability of the surface and the adsorption of fibronectin, vitronectin and laminin, the major ECM components involved in cell adhesion [42]. The initial phase of cell adhesion is governed by surface charge, which facilitates the establishment of electrostatic interactions [43]. The number of adhesive proteins and surface roughness also impact cell adhesion and spreading through these interactions [44]. An examination of the RMS roughness values of the samples reveals that the nano-roughness values are notably similar, particularly in the anodized samples. RMS values were measured as 137 nm at control, 83 nm at 80–20 V, 78 nm at 80–40 V and 76 nm at 80–60 V. Nevertheless, despite the comparable values, the disparity in the size distribution of the peaks is evident in the line profiles. The size distribution of the 80–60 V sample exhibits a distinctive spiky character, as indicated by an increase in kurtosis. A comparison of surfaces with comparable roughness levels reveals that the 80–60 V sample exhibits distinctive kurtosis values, differentiating it from the other samples. The peaks comprising the roughness of this sample exhibit a spiky distribution in comparison with the other samples. This discrepancy is also evident in the surface charge evaluation conducted by Kelvin probe force microscopy (KPFM) on the top surface of all samples. As illustrated in figure 4, the 80–60 V sample exhibits the highest negative average surface potential values, with a value of −209 mV µm⁻². The untreated control sample exhibited a positive surface potential of +73 mV µm⁻², whereas the 80–20 V and 80–40 V samples displayed negative surface potentials of −122 and −73 mV µm⁻², respectively. KPFM colour maps indicated that the surface charge distributions on the 80–20, 80–40 and 80–60 V samples were homogeneous across the surface areas. The untreated control sample exhibited manufacturing morphological defects on its surface and heterogeneous surface topography, resulting in a heterogeneous surface charge distribution observed in the KPFM colour map. While the KPFM colour maps of the 80–20, 80–40 and 80–60 V samples indicated that the negative potential is more prevalent at the surface nanocavities and pores, the data are insufficient to determine the precise location of charge spreading. Nevertheless, an increase in charge density is observed with an increase in porosity and the presence of spiky walls in the oxide layers. The correlation between surface potential and surface free energy is revealed by the electrostatic forces of intermolecular interactions. The surface charge generates an electric field that exerts an attractive or repulsive force on adjacent molecules, thereby influencing the overall attractive or repulsive forces between the surface molecules and those in the surrounding environment [45,46].

Figure 4.

Kelvin probe force microscopy images of control, 80–20 V, 80–40 and 80–60 V samples and their surface charges per 1 µm² area with atomic force microscopy topography images at the non-contact mode of the investigation zones. Note that all scale bars are equal and 250 nm.

Energy dispersive X-ray (EDS) analysis was employed to ascertain the atomic percentages of titanium (Ti), oxygen (O) and fluorine (F) atoms at five distinct points along the length of each sample. The surface of the control sample, which had not undergone any electrochemical treatment, was found to consist of 81.82% titanium, 16.84% oxygen and 1.34% fluorine, as illustrated in figure 5a. The positive surface potential measured on the control sample by KPFM analysis can be correlated with the number of metallic Ti atoms forming the sample. As anticipated, the oxygen percentages have increased in accordance with the electrochemical anodic oxidation processes observed in other samples. At 80–20 V, the oxygen percentage has reached 45.2%, while at 80–40 V and 80–60 V, the oxygen percentage has reached 54.3 and 53.9%, respectively. This increase can be attributed to the formation of a new oxide layer, which has undergone hierarchical titania nanodecorations. The considerable standard deviations observed for the titanium and oxygen ratios in the second sample can be attributed to the heterogeneity of the nanotubular oxide layer at the sampled points, which is a consequence of its minimal thickness. With the ionization of F⁻ ions in electrolyte and chemical dissolution of TiO₂ as [TiF₆]²⁻ on positively charged electrodes during anodization, nanofeatures had more F atoms as compressive stress applied by negatively charged fluoride form during nanotube formation [47]. Accordingly, the F signals obtained by EDS analysis on the 80–20, 80–40 and 80–60 V samples were 14.7, 13.6 and 13.1%, respectively. However, the increase in applied anodizing potential of 20, 40 and 60 V at 80–20, 80–40 and 80–60 V in the second step of the anodizing process did not result in an increase in atomic fluorine percentages due to the dissolution of the negatively charged [TiF₆]²⁻ salt in water at both the electrolyte and rinsing process. The presence of hexafluorotitanate residue at the depths of the pores has also been found to correlate with the opposing surface potential, as measured by KPFM analysis on the surfaces of the 80–20, 80–40 and 80–60 V samples. Our previous research has demonstrated that this residue interacts electrostatically with charged proteins [45–48].

Figure 5.

Comparison of EDS results as atomic Ti (blue), O (red), F (yellow) percentages (a) and XRD spectra (b) obtained from samples as control (black), 80–20 V (blue), 80–40 V (green) and 80–60 V (red).

The X-ray diffraction (XRD) measurements on the samples in the study indicate that the application of electrochemical anodic oxidation at room temperature without any further thermal annealing step did not result in the formation of crystalline TiO₂, such as anatase or rutile. Conversely, the metallic titanium peaks exhibited a reduction in intensity following the formation of a new amorphous titania layer on the titanium electrode, as evidenced by the comparison of XRD spectra between anodized 80–20, 80–40 and 80–60 V samples and untreated control samples, as depicted in figure 5b. The peak exhibited a reduction in intensity, while the formation of an oxide was observed at the titanium peaks at (002) and (101) [49]. Through anodizing voltage, the amorphous region below 37° shows its expected behaviour as peak shift and region growth by y-axis at 80–60 V with its most amorphous structure and thickest oxide compared with the others. An increase in negative surface potential at KPFM has been correlated with the thickening of the oxide layer on titanium and a reduction in the thickness of metallic titanium, which has caused a positive charge. This phenomenon has been observed when the voltage has been increased from 80–20 to 80–60 V. Concomitantly, the quantities of hexafluorotitanate at the nanotube depths were found to be slightly comparable for F atom percentages at EDS analysis. Moreover, there were significant differences in surface oxide coverages which can change the surface potentials of the samples with different nanoscale hierarchies with similar F atom percentages. As seen in the scanning electron microscope (SEM) images in figure 2, image process porosity percentages can be used to determine coverage of TiO₂ on the surfaces as in equation (3.1).

$${\displaystyle {\mathrm{\%}}_{OxideCoverage}=100-{\mathrm{\%}}_{Porosity}.}$$ (3.1)

In order to elucidate the reduction in negative surface potential observed at the 80–40 V sample in comparison with the 80–60 V sample, it is necessary to adopt a novel perspective when evaluating the AFM, KPFM and SEM results. The oxide coverage mentioned in equation (3.1), which represents the solid surface available to have an electrostatic charge, was calculated as the extraction of porosity percentages from 100. The AFM and SEM images revealed that the surface of the 80–40 V sample lacked the sharp oxide edges observed on the nanotube walls of the 80–60 V sample. The oxide coverage was calculated to be 82%, representing the highest value observed. The soft and smooth oxide edges on the 80–40 V sample facilitate the spread of electrostatic charge on the surface, resulting in a surface potential of −73.0 mV µm⁻², which is the weakest negative potential observed. The oxide coverages of the 80–20 V and 80–60 V samples were also calculated as 73 and 57%, respectively. Furthermore, the surface potentials per unit area obtained by KPFM analysis were found to be correlated with the values of −122.7 and −209.2 mV µm⁻², respectively.

3.3. Cell adhesion, viability and morphology

SEM images were utilized to assess the way cells interacted with the surface following adhesion (figure 6). Figure 6e illustrates that there is no strong interaction at the surface–cell interface in the control group. However, in the anodized groups, actin organizations are evident, indicating the cytoskeleton formation (figure 6f–h). In the first scenario, the cells do not exhibit the characteristic process of spreading, and the boundaries appear to contract. In the surface-modified samples, the extent of cell spreading is such that the nanopattern in the substrate becomes visible, along with the thinning cellular compartment. In particular, the microfilament organizations in the sample groups 80–20 and 80–40 V are linearly concentrated in the spreading directions, which indicates that the intracellular skeleton structure has begun to form and is becoming increasingly robust. In contrast, the filopodia structures of the cells in the final sample group (80–60 V) are active in all directions. The probing behaviour of the microenvironment with filopodia indicates that the cell is in the migratory phase in order to achieve better anchoring at a more optimal site.

Figure 6.

Cell interactions on sample surfaces: control (a), 80–20 V (b), 80–40 V (c) and 80–60 V (d), and on cell-surface interactions (e–h) obtained at high magnification of the dashed yellow squares from a–d, respectively.

Cell spreads were also analysed quantitatively and graphed in figure 7f. Measurements were taken of random individual cells spread on sample surfaces, and ImageJ software was used to measure the area covered by cells (electronic supplementary material, figure S1). It was observed that in all sample groups displaying a hierarchical tubular organization, the area covered by a single cell was increased. The statistical analysis shows a significant rise in 80–20 and 80–40 V samples compared with the control group. The decrease in cell spreading is observed with the change in the thickness and morphology of the oxide layer on 80–60 V group, due to increased roughness and spiky surface profile, which is consistent with the literature [50,51]. On these surfaces, the cells are moving rather than settling down, as supported by the occasional filopodia elongation observed at high magnification. Among the three modified surfaces with different topographic features, the two groups with the most apparent differences in spreading and adhesion (80–20 and 80–60 V) are visually illustrated (figure 8). In a study by Wang et al., the spreading behaviour of MSCs on hydrogel reinforced with five arginine-glycine-aspartate (RGD) nanopatterns of different nanocavities was investigated [52]. It was observed that small RGD gaps induced cytoskeletal organization with strong focal adhesion, while larger RGD nanospacing created weak focal adhesions and a more fragile cytoskeleton. It was concluded that the critical spreading gap was 70 nm, and less cell spreading was observed at nanospacing above 70 nm [52].

Figure 7.

Bar graphs were combined to compare surface properties correlated to cell–matter interactions for control, 80–20, 80–40 and 80–60 V samples: physical surface properties (a), porosity and oxide coverage percentages (b), alteration in surface potential according to fluorine residue and oxide coverage (c), correlation between surface potentials, surface free energies and water contact angles (WCA) (d), and some of the surface parameters affecting cell spread areas, viable and late apoptotic MSC percentages (e). Cell spread areas cultured on sample groups (f).

Figure 8.

This image shows the spreading behaviour of cells as they encounter the surface. (a) Cells exhibit stronger electrical density and adhesion strength towards fibronectin on rough surfaces than flat surfaces, promoting better initial attachment. (b) On the 80–20 V surface, cells display a more linear and organized spread, probably due to the optimal balance between roughness and surface charge, facilitating cytoskeletal organization. (c) In contrast, on the 80–60 V surface, cells predominantly exhibit migratory behaviour with prominent cell extensions (filopodia). This is attributed to the surface’s increased tapering and spiky roughness, which may challenge stable adhesion but encourage exploration and migration.

Flow cytometry assessed cell viability and apoptosis in MSCs cultured on titanium dioxide surfaces with various surface nanodecorations (figure 9). The viability of the cultured cells and the ROS released into the environment were analysed by flow cytometry and compared with relevant literature [53,54]. On the control surface, 68% of the cells remained alive. On the 80–20 V surface, 90% of cells were viable, indicating that most cells survived on this surface. On the 80–40 V surface, 77% of cells were viable, suggesting that cell viability increased relative to the control surface but decreased relative to the 80–20 V surface. On the 80–60 V surface, 57% of cells are alive, indicating cell viability drops significantly at the highest voltage. When necrotic cell ratios were examined, necrosis was observed at the minimum level (0.66%) on the 80–20 V surface. Early apoptotic cell rates increased on high-voltage surfaces and were measured as 6.2% on the 80–60 V surface. Late apoptotic cell rates peaked at 34.78% on the 80–60 V surface.

Figure 9.

This image presents flow cytometry analysis plots for cell death in mesenchymal stem cells (MSCs) across four different conditions: a control (MSC on non-modified titanium and three different experimental surface treatments (80–20, 80–40 and 80–60 V).

The results show hierarchical TiO₂ nanotube surfaces significantly affect cell viability and apoptotic cell rates. In particular, the 80–20 V surface provided the highest cell viability and the lowest necrotic cell rate. However, there was a trade-off between surface roughness and cell viability. Higher pore diameters, particularly the 80–60 V treatment, increased cell death, as evidenced by the higher proportion of cells in late apoptosis or necrosis. This indicates that nanotopography can enhance certain cellular functions but must be optimized to avoid adverse effects on cell viability. Specifically, the highlighted 80–60 V treatment condition showed a significant rise in the proportion of late apoptotic cells (34.78%), suggesting that higher nanotube diameter may correlate with increased apoptosis or cell death [55]. Our data showing high apoptosis rates align with previous findings that MSC apoptosis is crucial for their immunosuppressive effects and therapeutic efficacy in disease models [56].

Overall, the data imply that the physicochemical properties of the incubation surface with different electrostatic and biomechanical interactions at the surface–cell interface can substantially affect MSC viability and potentially induce more significant cell mortality. These findings highlight the importance of optimizing surface treatment parameters to maintain cell integrity during in vitro experiments.

A comprehensive assessment of the investigated structural surface parameters was conducted following the acquisition of pertinent characterizations. The surface properties of the samples were then represented graphically as bar graphs, as illustrated in figure 7. As a result of the observed correlation between the overall surface properties, it can be noticed that the electrochemical anodic oxidation procedure directly reduced the RMS and water contact angle (WCA) values of the control surface as shown in figure 7a. Similarly, anodization process with increasing voltage caused the formation of a porous structure on samples with increasing nanoscale mean pore diameter as expected [40]. In figure 7b, it was directly presented that the surfaces of groups have different porosity and oxide coverage percentages derived as equation (3.1). The negatively charged [TiF₆]²⁻ residue at the bottom of the formed hierarchical pores is slightly dependent on the voltage applied during the second anodization step as in figure 7c [57]. Due to the charged residue, it was clearly said that sign of the surface potential is shifted from positive to negative at anodized samples. Even though the fluorine percentages were quite similar, surface potentials were altered because of the changing oxide coverage percentage. As the applied voltage reached its maximum value at the 80–60 V sample, the highest negative surface potential and lowest oxide coverage were observed at this same sample. The results of the EDS and XRD analyses demonstrated that the oxide layer thickened from the 80–20 V to the 80–60 V sample. The inverse proportionality between oxide coverage and surface potential can be attributed to the limited capacity of the [TiF₆]²⁻ residue to spread at solid material sites in lieu of hollow sites within the pores. In figure 7d, increase in surface free energies and decrease in WCA was correlated with the surface potential of the samples [58]. It has been proposed that negatively charged surfaces exert a greater electrostatic attraction between cells and the surface of anodized samples than control surfaces. The findings of our study indicate that a negative surface charge, which increased in conjunction with elevated oxygen content in comparison with the titanium surface following anodization, exerted a beneficial influence on cell adhesion and spreading. Although the highest surface charge was observed in the 80–60 V sample group, which exhibited a more favourable surface for cells than the control group, the spiky distribution of roughness and topography led to disparate results in cell spreading and mobility. It is notable that the MSC cell spread areas were increased by anodization of Ti when compared with the control as in figure 7e. The cell viability is the lowest at the 80–60 V sample due to the most potent electrostatic attraction occurring at the sharpest oxide walls at the highest porosity. The viabilities of the 80–20 V and 80–40 V samples are significantly higher than those of the control, probably due to the increased area of the oxide layer that can be adhered to by the cells. Due to the compression of the cells on the sharp nanotube walls with the highest electrostatic attraction by the highest negative surface potential, 80–60 V caused a reduction in cell viability and increase in the late apoptose compared with the others. The increase in surface free energy polar components with higher anodization voltages supports better adsorption of ECM proteins like fibronectin, vitronectin and laminin, essential for cell adhesion and spreading. However, the increased roughness and spiky surface profiles at higher voltages may contribute to the increased cell death, suggesting a need for balance between enhancing adhesion and maintaining cell viability.

3.4. Increased expression of key regenerative genes

We examined the effects of surface nanodecoration changes on the mRNA expression levels of genes critical for the regulation and function of MSCs. Figure 10 shows the quantitative expression of target genes chemokine receptor type 4 (CXCR4) and vascular endothelial growth factor A (VEGFA) in MSCs cultured on different surfaces measured by quantitative polymerase chain reaction (qPCR). CXCR4 expression levels show a marked increase with the change of surface properties from 80–20 to 80–40 V and 80–60 V. This increase reveals the significant effect of surface nanodecoration on CXCR4 mRNA levels. Similarly, VEGFA expression levels show the lowest level at 80–20 V treatment, and a substantial increase in expression levels was observed as the voltage increased in the second anodization step. The highest expression was observed in the 80–60 V voltage treatment. Overall, qPCR data reveal that MSCs exhibit gene-specific responses to surface treatments and that there are nanodecoration-dependent effects on gene expression. In particular, the expression levels of both VEGFA and CXCR4 were significantly increased in both 80–40 V and 80–60 V treatments. Higher anodization voltages significantly upregulated CXCR4 and VEGFA, indicating the enhanced migratory and angiogenic potential of MSCs [59,60]. The migratory phase of the MSCs was also correlated with the SEM images, evidenced by the active filopodia around the roundly shaped cells. These findings align with the observed increase in ECM production and cell spreading, suggesting that fabricated surfaces can potentiate the regenerative capabilities of MSCs by modulating key gene pathways. These findings highlight the potential regulatory effect of electrical stimulation on cellular functions and gene expression profiles and may be necessary for tissue engineering and regenerative medicine applications.

Figure 10.

mRNA expression levels of the target genes of MSCs on different surfaces. This figure shows the relative expression levels of CXCR4 and VEGFA genes in mesenchymal stem cells (MSCs) and dermal fibroblasts, determined by real-time PCR.

4. Conclusions

This study underscores the critical role of surface nanotopography in modulating MSC behaviour and highlights the potential of hTiO₂ nanotube arrays in advancing regenerative medicine. The findings provide a foundation for developing optimized biomaterial surfaces that can enhance the efficacy of cell-based therapies for various inflammatory and degenerative diseases. Examining MSCs and fibroblasts is crucial from a stromal cell perspective, as it provides comprehensive insights into the differential roles and mechanisms of various stromal cell types in therapeutic contexts. The hierarchical TiO₂ nanotube arrays present a promising strategy for enhancing the therapeutic potential of MSCs in regenerative medicine. The ability to control surface nanotopography and tailor cell responses offers a versatile approach to optimizing cell-based therapies. However, carefully optimizing surface properties is essential to balance enhanced functionality with cell viability. Further research should concentrate on optimizing the anodization parameters and investigating the long-term impact of these surfaces on MSC behaviour. Additionally, the stiffness of cells cultured on diverse topographies could provide insights into the mechanobiological aspects of cell–surface interactions.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

Data is available from Zenodo [61].

Supplementary material is available online [62].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

N.K.T.: formal analysis, investigation, writing—original draft; B.K.: formal analysis, methodology, visualization, writing—original draft; G.I.: data curation, formal analysis, investigation; B.K.: data curation, formal analysis, supervision, validation; E.Z.T.: project administration, resources, supervision; C.B.: funding acquisition, project administration, supervision, writing—original draft.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This research was financially supported by Hacettepe University Scientific Research Projects (FKB-2022-19844)