Nanobubble-Mediated Oxygen Delivery Mitigates Hypoxia-Induced ROS and HIF-1α Expression in UC-MSCs

TL;DR

This study shows that oxygen nanobubbles help protect stem cells from hypoxia by reducing harmful stress and improving cell survival.

Contribution

The use of Tivida®-stabilized oxygen nanobubbles for mitigating hypoxia in UC-MSCs is a novel oxygen delivery strategy.

Findings

TONBs increased dissolved oxygen levels in culture media up to ~18 ppm.

TONB treatment reduced mitochondrial ROS levels by ~20% under hypoxia.

HIF-1α expression was downregulated by ~8–9 fold in TONB-treated cells.

Abstract

Hypoxia and nutrient-deprived microenvironments pose significant challenges to the survival of transplanted human umbilical cord mesenchymal stem cells (UC-MSCs), necessitating the development of controllable oxygen delivery strategies. In this study, we engineered fluorosurfactant-coated oxygen nanobubbles (Tivida®-stabilized; TONBs) and assessed their cytoprotective effects in a two-dimensional (2D) ischemia-mimetic model (1% O2 and 1% FBS). The TONBs were characterized by nanoparticle tracking analysis and zeta potential, while dissolved oxygen (DO) release was quantified in DMEM culture media. TONBs formed stable sub-200 nm populations with high colloidal stability (−58 mV) and demonstrated elevated DO levels up to ~18 ppm, compared to DMEM control (~ 8 ppm). Under hypoxic stress, TONB treatment preserved metabolic activity and viability, reduced mitochondrial ROS levels by ~20% and resulted in an ~8–9 fold downregulation of HIF-1α expression relative to untreated hypoxic controls. These results indicate that TONBs provide oxygen buffering to mitigate hypoxia-driven metabolic stress, supporting their potential as an oxygen delivery adjunct for regenerative medicine applications and tissue engineering applications.

Figures

Full text

1. Introduction

Mesenchymal stem cells (MSCs) demonstrate significant therapeutic potential through their capacity for multilineage differentiation, including chondrogenic, osteogenic, adipogenic, neural, and cardiomyocyte phenotypes [1,2,3]. Nevertheless, their clinical applications are limited by reduced survival and functionality in hypoxic environments [3,4], which are common in ischemic tissues (e.g., myocardial infarction and stroke) [5], and may also occur following avascular biomaterial implantation. Umbilical cord-derived MSCs (UC-MSCs) retain multilineage capacity in vitro and in vivo and offer practical advantages over other MSC sources, including high proliferative capacity and fewer ethical and legal constraints [6,7]. Despite these attributes, the survival of transplanted MSCs is compromised within a harsh microenvironment characterized by factors such as inadequate blood supply and low oxygen tension or hypoxia, which can predispose MSCs to apoptosis [8].

Hypoxia, defined by inadequate oxygen supply, affects cellular metabolism, prompting a shift towards anaerobic pathways and inducing cellular stress [9]. The hypoxia-inducible factor (HIF), a heterodimeric transcription factor composed of α- and β-subunits, functions as a master regulator of cellular responses to hypoxia. Briefly, HIF α/β translocates to the nucleus, where it facilitates the transcription of numerous genes involved in essential cellular processes, including glycolysis, angiogenesis, and erythropoiesis [10,11,12,13,14]. These adaptive responses initially maintain cell viability and function during periods of oxygen deprivation. Nevertheless, extended or severe hypoxia can exceed these adaptive mechanisms, resulting in tissue damage. This can lead to mitochondrial dysfunction, diminished ATP production, and ultimately cell death [9]. This phenomenon is particularly pronounced in ischemic heart disease, where transplanted MSCs show a rapid loss of viability, with approximately 99% becoming non-viable within the first four days post-transplantation [3]. Rabbit adipocyte-derived stromal cells and UC-MSCs have consistently showed reduced viability and increased apoptosis at 0.1% and 1.5% O_2_, respectively [8,15]. Moreover, ischemic tissues reduce rat neonatal cardiomyocyte survival compared with vascularized tissues [4], underscoring the impact of ischemia on grafted cell survival. In general, hypoxic conditions—especially sustained oxygen levels below 1% O_2_—can contribute to tissue failure, disrupted cellular niches, and impaired MSC function. Therefore, hypoxia and ischemia, often accompanied by limited soluble factors (e.g., serum components and growth factors), remain major challenges for cell therapy, regenerative medicine, and tissue engineering (TE), particularly in cell-laden scaffolds [14].

Multiple strategies have been proposed to enhance tissue oxygenation and alleviate hypoxia [16]. Although approaches such as cell preconditioning have been explored, their efficiency remains under debate, partly due to unpredictable long-term effects and the lack of direct, controlled oxygen delivery [16]. Accordingly, an efficient and targeted oxygen-delivery solution remains an open challenge. In this context, gas nanobubbles (NBs) have emerged as oxygen-delivery carriers with reported utility for improving oxygenation in hypoxic settings [17,18,19]. These gas-filled structures can function as oxygen-releasing nanocarriers, which have been associated with high gas-loading capacity (reported ~4–5-fold higher gas density than dissolved oxygen), enhanced stability (days, weeks, and months) [20], and promising applications for site-specific delivery [21]. The NBs’ size may also facilitate transport through constrained biological environments and access to specific cellular and tissue targets [22]. Collectively, these features motivate investigation of nanobubbles across applications where oxygen limitation is central, including regenerative medicine, neuroscience, oncology, and sports medicine [23,24,25,26].

The clinical translation of this technology is already underway; for instance, charge-stabilized oxygen nanostructures (e.g., RNS60) have demonstrated safety and efficacy in Phase II clinical trials for acute ischemic stroke and amyotrophic lateral sclerosis (ALS) [23,24]. Furthermore, oral formulations of oxygen nanobubbles are currently being evaluated in clinical studies for managing pulmonary fibrosis [26] and improving sprint cycling performance [25], highlighting the versatility of this approach. Although 2D cultures are not inherently hypoxic, they are frequently utilized as models in which hypoxic stress is artificially induced to simulate ischemic conditions. Recent investigations have highlighted the versatility of oxygen nanobubbles (ONBs) in reversing hypoxia in diverse pathological landscapes. For instance, novel PLGA-based ONBs have been shown to downregulate HIF-1α and VEGF-A expression in retinal ischemia models [27], while ONB-laden hydrogels have demonstrated accelerated closure and reduced inflammation in surgical wounds [28]. Furthermore, the therapeutic potential of ONBs has expanded to complex ischemic conditions, including stroke therapy [29] and organ development [30], emphasizing their role in restoring cellular metabolism. However, although the potential of nanobubbles for oxygen delivery has been explored in various contexts, their specific efficacy in mitigating stress within these models remains largely unexplored, particularly in UC-MSCs.

Therefore, this study aimed to assess the in vitro effects of ONBs on UC-MSC survival under 2D culture conditions of (i) low oxygen tension and (ii) nutrient deprivation. The effects of these parameters were investigated separately and in combination. The UC-MSCs were expanded in vitro at 21% O_2_ and exposed to ischemia (low oxygen tension and serum deprivation). Biological parameters, including viability, cell death, cell proliferation, ATP production, and mitochondrial stress, were evaluated at various time points (24–72 h after treatment).

2. Materials and Methods

2.1. Materials

Powdered Dulbecco’s modified Eagle’s medium (DMEM), high glucose, pyruvate (cat 12800017, ThermoFisher Scientific, Waltham, MA, USA), and sodium hydrogen carbonate (cat 401676, Sigma-Aldrich, St. Louis, MO, USA) were dissolved in ultrapure water and filtered through 0.22 µm filters. The ultrapure water (type 1) used in the experiments exhibited a conductivity of 0.055 μS cm^−1^ and pH of 6.7 at a temperature of 20 °C. O_2_ (99.99% purity) was used to prepare the ONBs.

2.2. Generation of Bulk NBs in Cell Culture Medium and Characterization

Bulk NBs (ONBs and TONBs) were generated in DMEM or ultrapure water using a mechanically induced Extension–Compression (EC) technique, adapted from the protocol described by Ferraro et al. [31]. Briefly, 7 mL of DMEM was first pre-saturated with oxygen by sparging the liquid within a sealed 15 mL tube (adapted as a closed chamber). Medical-grade oxygen (99.9% purity) was injected at a flow rate of 0.5 L/min under a constant pressure of 0.05 MPa for 2 min. Immediately following oxygenation, the fluorosurfactant Tivida^®^ FL 2300 (3M, St. Paul, MN, USA) was added to a final concentration of 0.005%. The mixture was then loaded into a syringe, and the tip was hermetically sealed with a Luer lock cap to eliminate headspace and create a closed system. Nanobubble generation was initiated by subjecting the solution to 10, 50, 60, and 100 continuous EC cycles. Each cycle consisted of a rapid extension phase (plunger withdrawal at ~7 cm/s) to generate transient negative pressure and induce bubble nucleation, followed by a compression phase, where the vacuum force drove the plunger back to its initial position. This mechanical oscillation facilitates the formation of stable, surfactant-coated nanobubbles (see Supplementary Materials). In all cases, the NB suspensions formed were stored in 20 mL air-tight glass vials for further analysis.

The physicochemical properties of the suspensions were then characterized. Nanobubble size distribution and concentration were quantified via Nanoparticle Tracking Analysis (NTA) using a NanoSight NS300 (Malvern Panalytical, Malvern, Worcestershire, UK) equipped with a green laser source (65 mW; λ = 510 nm). Additionally, surface charge (zeta potential) was assessed using a Zetasizer Pro (Malvern, PA, USA).

2.3. Dissolved Oxygen Measurement

To measure DO in liquids, we used optical oxygen sensors (Redflash Technology, FireSting^®^-O2 PyroScience, Aachen, Germany). The contact sensor comprised an optical fiber with a 430 µm diameter sensor tip (PyroScience, Aachen, Germany) immersed in DMEM in a closed glass bottle. The sensor emitted oxygen-dependent luminescence when irradiated, and the meter detected luminescence based on oxygen partial pressure. The sensor was calibrated using 2-point measurements: oxygen-free water (0 ppm O_2_) prepared with SO_3_ and air-saturated water. For air-saturated calibration, the flask was filled with water, closed, and agitated for 1 min. Regarding DO measurements, TONBs and their dilutions (1:2 and 1:10) were introduced via a syringe or micropipette in a sealed septum of the vial to prevent atmospheric re-oxygenation or any gas disturbance.

2.4. Cell Culture and Hypoxic In Vitro Model

UC-MSCs were obtained from donated human umbilical cords, as described previously [32], and immunophenotypically characterized according to ISCT guidelines [32]. Experiments were performed using cells from passages three to seven. Briefly, UC-MSCs cryopreserved at P1 (90% FBS; 10% DMSO) were harvested by gently flushing the cells with DMEM containing 10% FBS (, 1% penicillin–streptomycin, and L-glutamine (all obtained from Thermo Fisher Scientific, Waltham, MA, USA). When the UC-MSCs reached 70–80% confluence, they were detached and subcultured (density: 1000 cells/cm^2^). The donated cell codes used in this study were also evaluated for tridifferentiation (osteogenic, adipogenic, and chondrogenic) in a previous study [33].

To strictly evaluate the cytoprotective capacity of TONBs, we utilized a “double-stress” model combining severe hypoxia (1% O_2_) with serum deprivation (1% FBS). This design was chosen to mimic the harsh ischemic microenvironment encountered by MSCs immediately post-transplantation, where cells face both oxygen scarcity and nutrient limitations. The 1% FBS condition (vs. standard 10%) minimizes the confounding effects of growth factors in the serum, ensuring that any observed metabolic rescue is primarily attributable to the oxygen delivery provided by the nanobubbles rather than exogenous signaling molecules. Comparisons were made against a “healthy” baseline (normoxia/10% FBS) and a “severe injury” control (hypoxia/1% FBS). The cells were incubated in a hypoxic chamber (C-chamber, BioSpherix, NY, USA), while the control group was maintained under standard normoxic conditions (21% O_2_ and 5% CO_2_). To assess metabolic plasticity, UC-MSCs were treated for 24 h with metabolic inhibitors: 5 mM 2-Deoxy-D-glucose (2-DG; Sigma-Aldrich, Merck, Germany) to inhibit glycolysis or 1 μg/mL oligomycin (Calbiochem, Merck, Germany) to block mitochondrial oxidative phosphorylation.

Regarding the protocol for nanobubble administration in cell culture, we clarified that TONB treatment was not added during hypoxic incubation (which would disturb the gas equilibrium in the hypoxic chamber) but rather concurrently with the onset of hypoxia. TONBs were generated or mixed with the culture medium and applied to the cells immediately prior to placing the culture plates in the hypoxia chamber (T = 0). This protocol was chosen to simulate the application of nanobubbles as a pretreatment agent at the onset of an ischemic event to mitigate or protect against hypoxic stress.

2.5. Metabolic Activity and ATP Determination

Cells were seeded onto a 96-well plate at several densities (2 × 10^3^ and 10 × 10^3^ cells per well). Metabolic activity was determined using the PrestoBlue colorimetric assay (Thermo Fisher Scientific, Waltham, MA, USA). Cell viability was confirmed by flow cytometry using Ghost Dye 510 Fix (Cat:59863, Cell Signaling Technology, Danvers, MA, USA) at 1 × 10^5^ cells/well in 12-well plates. ATP production was quantified using the CellTiter-Glo^®^ Luminescent Cell Viability Assay (Promega, Madison, WI, USA, G7570) following the manufacturer’s instructions. Isolated MSCs were suspended in 50 µL of medium and seeded in an opaque 96-well plate at 10 × 10^3^ cells/well. Then, 50 µL of CellTiter-Glo^®^ reagent was added and the plate was incubated for 30 min at room temperature. The luminescence was measured using a BioTek FLx800 microplate reader (Thermo Fisher Scientific, Waltham, MA, USA).

2.6. Identifying Reactive Oxygen Species Generated by Mitochondria (Mito-ROS)

Mitochondrial superoxide levels were quantified using MitoSOX™ Red (Thermo Fisher, Waltham, MA, USA). UC-MSCs were cultured at 1.5 × 10^6^ cells/well for 24 h under hypoxic and normoxic conditions. Cells were incubated with 2 μM MitoSOX working solution for 20 min at 37 °C protected from light. After washing twice with PBS, cells were resuspended in flow cytometry buffer. Data acquisition was performed on a BD FACSCanto II cytometer, recording 10,000 events per sample within the living cell gate. Mean Fluorescence Intensity (MFI) was analyzed using FlowJo v10 software. Menadione (40 µM) served as the positive control.

2.7. Cell Death, Apoptosis and Necrosis

We hypothesized that low oxygen levels and insufficient nutrients during ischemia could cause extensive cell death under prolonged hypoxia. Cells were seeded in 12-well plates at a density of 1 × 10^5^ cells/well and exposed to hypoxic conditions for 72 h. Flow cytometry was used to measure Annexin V fluorescence and PerCP (7AAD) in UC-MSCs. Analysis was performed using a BD FACS Canto II Flow Cytometer, with fluorescence acquired on 4-decade logarithmic scales. Each sample contained at least 10,000 cells for statistical analysis.

2.8. RT–qPCR and Western Blot Analysis of HIF-1α Expression

HIF-1α and β-actin mRNA expression levels were measured in UC-MSCs exposed to hypoxic or normoxic conditions. RNA was extracted using the RNeasy Micro Kit (Qiagen, Hilden, Germany) and quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). RT-qPCR was performed using 40 ng of RNA using the Luna Universal One-Step kit (New England Biolabs, Ipswich, MA, USA). Primers were designed using Beacon Designer software version 8.21 with the following sequences: β-actin forward, 5′-GTGGGAGTGGGTGGAGGC-3′, and reverse, 5′-TCAACTGGTCTCAAGTCAGTG-3′; HIF-1α forward, 5′-CATAAAGTCTGCAACATGGAAGGT-3′, and reverse, 5′-ATTTGATGGGTGAGGAATGGGTT-3’. qPCR was performed using 2 µL template [10 ng/µL], 10 µL Brilliant III SYBR Master Mix [2×], and 1 µL primers [10 µM] in a final volume of 20 µL. The temperature profile was 95 °C for 10 s, followed by 40 cycles of 95 °C for 20 s and 72 °C for 5 s. Expression levels were calculated using the modified 2^−ΔΔC^t method, with β-actin as the control.

UC-MSCs (3 × 10^5^ cells) in 30 mm^2^ well plates were treated with TONBs, incubated (hypoxia or normoxia) for 2 h, and lysed with CelLytic M., and the protein concentration was determined using the BCA assay. Proteins (40 µL) were separated by electrophoresis on a 4–12% SDS polyacrylamide gel and transferred to PVDF membranes (Millipore, Burlington, MA, USA). Membranes were probed with primary antibodies against HIF-1α (D1S7W) rabbit (1/1000; Cell Signaling Technology, Danvers, MA, USA, cat36169) and β-actin (1/5000; Cell Signaling Technology, Danvers, MA, USA, cat3700) at 4 °C overnight. After washing with TBS + 0.1% Tween-20, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (1/5000; Cell Signaling Technology, Danvers, MA, USA). Protein bands were visualized using the SuperSignal™ West Pico PLUS substrate and iBright Imaging System (Thermo Fisher Scientific, Waltham, MA, USA). Band intensity was normalized to that of β-actin using ImageJ software version 1.54f. UC-MSCs cultured without TONBs under normoxic conditions were used as controls.

2.9. Statistical Analysis

Statistical analysis was performed using GraphPad Prism 9.0 (GraphPad Software, Boston, MA, USA). First, data distribution was assessed using the Shapiro–Wilk normality test. For normally distributed data, comparisons between multiple groups were performed using One-Way Analysis of Variance (ANOVA) followed by Tukey’s post hoc test. For non-normally distributed data, the Kruskal–Wallis test with Dunn’s multiple comparison post hoc test was utilized. Pairwise comparisons were conducted using the Mann–Whitney U test or Student’s t-test where appropriate. All results are expressed as mean ± standard error of the mean (SEM). Experiments were performed in at least three independent biological replicates (n = 3), and p-values ≤ 0.05 were considered statistically significant.

3. Results

3.1. Characterization of TONBs and Oxygen Release Kinetics

NTA demonstrated that the incorporation of Tivida^®^ fluorosurfactant significantly enhanced the physicochemical properties of the nanobubbles. TONBs achieved a concentration of 3 × 10^9^ bubbles/mL in ultrapure water, representing a significant increase compared to surfactant-free ONBs (p < 0.05). The surfactant coating resulted in a finer, more monodisperse population, significantly reducing the mean diameter from 216 nm (ONBs) to 123 nm (TONBs) (Figure 1A; p < 0.05). Furthermore, zeta potential analysis confirmed superior colloidal stability; TONBs exhibited a highly negative surface charge of −58 mV compared to −32 mV for uncoated ONBs (Figure 1B; p < 0.05), indicating strong electrostatic repulsion that prevents aggregation. Optimization studies determined that 50 EC cycles yielded the maximum nanobubble density (Figure 1C,D).

Regarding oxygen retention, initial gas sparging induced a transient supersaturation peak of ~30 ppm (Figure 1E). However, the stable TONB formulation maintained elevated DO levels ranging between 12 and 18 ppm immediately post-generation, significantly higher than the baseline of 7.5–8 ppm observed in control media (p < 0.05). Dilution assays (Neat, 1:2, and 1:10) revealed a dynamic relationship between particle concentration and gas release. While undiluted TONBs displayed a gradual linear decay, diluted samples (50% and 10%) exhibited a characteristic biphasic profile: an initial rapid release phase followed by a stable plateau as the media reached equilibrium (Figure 1F,G). NTA under normoxic conditions further indicated a correlation between dilution-induced size reduction and increased oxygen release rates, consistent with interfacial gas diffusion mechanisms.

Supplementary characterization confirmed that nanobubble stability was sensitive to environmental factors; deviations in pH and temperature significantly altered production yield and stability over 72 h (Supplementary Figures S1 and S2). Additionally, theoretical modeling of bubble rising velocity (Supplementary Equation (S1)) correlated with the observed stability profiles, confirming that the sub-200 nm size range of TONBs effectively minimizes buoyancy-driven dissipation (Supplementary Figures S3–S6).

3.2. Validation of the Ischemia-Mimetic Model (Hypoxia + Serum Deprivation)

To establish a baseline for therapeutic evaluation, we first characterized the kinetics of UC-MSC death under severe hypoxia (1% O_2_) combined with distinct serum concentrations. At 24 h, flow cytometry analysis using AnnexinV/PI staining revealed that cells cultured in low-serum conditions (1% FBS) maintained a viability of 92.5%, with 7.5% of the population exhibiting apoptotic markers (Figure 2A). In contrast, the high-serum control (10% FBS) maintained significantly higher viability (>98%), confirming that serum deprivation acts as a critical stressor alongside oxygen tension.

Extending the hypoxic exposure to 72 h significantly exacerbated cell death in the low-serum group. This prolonged stress resulted in the total apoptotic/necrotic population exceeding 15% (Figure 2B), accompanied by a qualitative shift from early apoptosis to predominant necrosis. Statistical analysis confirmed that the combination of 1% O_2_ and 1% FBS resulted in a significantly higher proportion of dead/apoptotic cells compared to the 10% FBS group (p < 0.05). Consequently, this “double-hit” condition (hypoxia + 1% FBS) was selected as the specific ischemia-mimetic model for all subsequent TONB efficacy assays, as it provides a quantifiable therapeutic window to measure cytoprotection.

3.3. Cell Density-Dependent Metabolic Preservation

The metabolic efficacy of TONBs was evaluated using the PrestoBlue assay, revealing a critical dependency on initial cell-seeding density (Figure 2C). Under 24 h of hypoxia (1% FBS), TONB treatment failed to rescue metabolic activity in low-density cultures (2000 cells/well). However, at higher seeding densities (10,000 cells/well), TONB treatment (Neat and 50%) significantly elevated metabolic activity compared to the untreated hypoxic control (p < 0.05), restoring values to levels comparable to the normoxic baseline. To confirm that these metabolic changes reflected true bioenergetic preservation rather than transient enzymatic stimulation, intracellular ATP levels were quantified (Figure 3F). TONB treatment also increased ATP production at levels statistically distinguishable from the hypoxic control (p < 0.05) in the “double-stress” model (1% O_2_ + 1% FBS) and “single-stress” model (1% O_2_ + 1% FBS). This correlation between PrestoBlue assay and ATP concentration confirms that TONBs effectively sustain metabolism during a 2D hypoxic model. Furthermore, extending the hypoxic exposure to 72 h demonstrated that all TONB concentrations (10%, 50%, and 100%) maintained metabolic activity significantly above the untreated control (Figure 2D; p < 0.05), especially for TONBs at 10%.

3.4. Metabolic Plasticity in a Double-Stress Model (1% O2 + 1% FBS)

To elucidate the metabolic adaptability of UC-MSCs, cells were treated with metabolic inhibitors under varying oxygen tensions (Figure 3A). When oxidative phosphorylation was blocked by oligomycin, UC-MSCs demonstrated a robust compensatory increase in metabolic activity compared to cells treated with the glycolytic inhibitor 2-DG (p < 0.05). This indicates a high degree of metabolic plasticity, where cells rapidly shift toward glycolysis to sustain bioenergetics when mitochondrial respiration is compromised. Consequently, the 1% FBS condition exposed the cells’ vulnerability, resulting in a drastic > 60% reduction in ATP under hypoxia (Figure 3B; p < 0.05). Furthermore, we compared TONBs’ effect against the most efficient metabolic plasticity (glycolysis induced by oligomycin treatment) and between the “double-stress” model (1% O_2_ + 1% FBS) and the “single-stress” model (1% O_2_ + 10% FBS). As shown in Figure 3C,D, the administration of TONBs under hypoxia induced a metabolic state quantitatively comparable to this oligomycin-driven maximum. While hypoxia typically suppresses bioenergetics, TONB-treated cells maintained ATP levels comparable to the oligomycin-positive control (Figure 3C), suggesting that adequate oxygen delivery allows these cells to achieve their full bioenergetic potential even under stress. This trend remained consistent across normoxia and hypoxia: although the absolute ATP magnitude was significantly higher in the nutrient-rich normoxic/10% FBS controls (Figure 3D; p < 0.05), the relative capacity of TONBs to rescue metabolism to “maximal” levels reflects the plasticity profile observed in the metabolic inhibitor assays.

3.5. TONBs Demonstrate Cytocompatibility and Enhance ATP Synthesis

Cytocompatibility assays confirmed that TONB exposure did not induce cytotoxicity; cell viability remained statistically equivalent to the untreated control across all concentrations (Figure 3E; p > 0.05). Furthermore, we evaluated the capacity of TONBs to rescue the bioenergetic deficit identified in the “double-stress” model. While untreated hypoxic cells suffered a severe drop in ATP (down to ~1200 RLU), treatment with TONBs significantly restored intracellular ATP production (Figure 3F). This pattern of metabolic enhancement was consistent across both the “double-stress” (1% FBS) and “single-stress” (10% FBS) conditions. Although the absolute magnitude of ATP production was significantly higher in the nutrient-rich controls (10% FBS) compared to the starvation model (Figure 3F; p < 0.05), the relative capacity of TONBs to augment metabolism was evident in both hypoxic contexts, suggesting that oxygen delivery functions synergistically with nutrient availability.

3.6. Rapid Hypoxic Response and TONB-Mediated Suppression of Oxidative Stress and HIF-1α Signaling

Exposure to hypoxia triggered an immediate molecular response in UC-MSCs. Within 30 min, mitochondrial ROS (mito-ROS) levels surged, accompanied by a six-fold upregulation of HIF-1α mRNA expression (Figure 4A; p < 0.05). Western blot analysis corroborated this acute activation, revealing characteristic HIF-1α protein bands (120, 110, and 93 kDa) consistent with rapid stabilization and nuclear accumulation prior to full post-translational modification.

We next evaluated the capacity of TONBs to mitigate oxidative stress over a 24 h period. Flow cytometry analysis confirmed that prolonged hypoxia induced a ~2-fold increase in mitochondrial ROS intensity compared to normoxic controls (980 vs. 466 MFI; Figure 4C,D). However, TONB treatment significantly attenuated this oxidative burden in a dose-dependent manner. Specifically, the undiluted formulation (100% TONB) achieved a ~20% reduction in mito-ROS levels (Figure 4E; p < 0.05). Importantly, this scavenging effect was specific to the stress condition; TONBs at concentrations of 100% and 50% did not alter basal ROS levels in normoxic cells (Figure 4F), indicating that the treatment targets hypoxia-induced mitochondrial dysfunction without disrupting homeostatic redox balance.

Finally, we confirmed that this reduction in oxidative stress correlated with the suppression of hypoxic signaling. Consistent with the sustained oxygen delivery observed after 2 h (Figure 4G), TONB treatment resulted in a profound downregulation of HIF-1α protein expression (Figure 4H). Densitometric analysis revealed that the 50% and 100% TONB formulations were the most potent, suppressing HIF-1α expression 8.5- to 9-fold compared to the untreated hypoxic control (p < 0.0001). Even the lowest concentration (10% v/v) yielded a significant reduction (p < 0.001). These findings validate that TONBs effectively disrupt the hypoxic signaling cascade by restoring intracellular oxygen tension.

4. Discussion

A pivotal innovation in this study is the engineering of Tivida^®^-coated oxygen nanobubbles (TONBs), which utilize a fluorosurfactant shell to improve colloidal stability, surface contact per unit area, and yield production. Our results demonstrate that the inclusion of the fluorosurfactant reduced the mean particle size to ~100 nm while dramatically increasing the magnitude of the zeta potential to −58 mV, compared to −32 mV for uncoated bubbles. Mechanistically, this size reduction is likely driven by surfactant-induced lowering of the interfacial tension, which facilitates the kinetic “fission” of gas cavities into smaller radii during shear stress generation. Regarding stability, the shift to −58 mV was attributed to the anionic nature of the surfactant headgroups. While a zeta potential exceeding 30 mV is generally indicative of stability [34,35], the high magnitude observed here suggests a robust electrostatic repulsion barrier that effectively prevents coalescence, thereby selectively preserving the nanoscale distribution. Similar results, but with other ionic surfactants, have demonstrated the role of surfactants in controlling bubble size, reducing dissolution rates, and regulating gas delivery [36,37,38]. These characteristics enable NBs to persist in liquid environments for extended periods of several weeks [39,40]. However, the shell is not impermeable; it also allows oxygen to diffuse from inside the nanobubbles into the media or liquid phase according to Henry’s Law, maintaining a supersaturated state (12–18 ppm) that acts as a “buffer”. Furthermore, we observed that increasing the number of generation cycles—which reduces bubble size and improves surfactant packing—prevents the “runaway” spike in DO often seen with unstable microbubbles. Therefore, unlike conventional lipid- or protein-shelled bubbles that often suffer from rapid gas leakage [36], the fluorocarbon tail of Tivida^®^ likely creates a hydrophobic shield that retards the gas diffusion. This supports the concept that oxygen is effectively compartmentalized within the nanobubble core, serving as a long-lived metastable reservoir that releases oxygen gradually rather than instantaneously. Regarding the limitations, although NTA and DLS confirmed the presence of nanosized populations, these optical techniques cannot strictly distinguish between gas-filled nanobubbles and surfactant micelles. However, the marked increase in dissolved oxygen (up to 18 ppm) provides functional evidence of their gas-carrying capacities.

It is important to note that while fluorosurfactant-only controls confirmed the cytocompatibility of the shell material (Supplementary Figure S6), the efficacy of the surfactant alone under hypoxic conditions was not separately evaluated. However, given that the fluorosurfactant lacks intrinsic oxygen-carrying capacity in the absence of a gas core, the significant metabolic preservation observed in TONB-treated cells is attributed to the oxygen payload delivered by the nanobubbles rather than a biochemical effect of the surfactant shell.

A critical methodological contribution of this study is the rigorous validation of the “double-stress” model (1% O_2_ + 1% FBS) as a true physiological mimic of ischemia. The literature regarding MSC hypoxia is often contradictory, with some studies reporting maintained viability under low-oxygen conditions. Previous findings have shown that exposure of primary MSCs to hypoxic conditions (2% O_2_) does not affect cell viability [41]. Consistent with this observation, Antonina-Lavrentieva et al. reported a significant enhancement in UC-MSC proliferation under 2.5% oxygen compared to that under normoxic conditions [8]. In contrast, hypoxia (2% O_2_) reduces the viability of human osteoblast-like cell lines (MG63) [42], whereas hypoxia (1.5% O_2_) shows low apoptosis levels in 2D UC-MSC cultures [8]. These discrepancies suggest that primary cells, such as MSCs, are inherently less sensitive to hypoxia than immortalized cell lines. However, our data clarify this discrepancy by revealing that 10% FBS effectively masks hypoxic stress. We observed that high serum concentrations provide a “bioenergetic cushion,” allowing cells to maintain ATP production even under oxygen deprivation. By lowering the serum concentration to 1%, we stripped away these exogenous survival factors, revealing the true vulnerability of UC-MSCs to ischemia, as evidenced by the sharp decline in ATP levels and progression towards necrosis after 72 h. This distinction is paramount in regenerative medicine, as post-transplantation microenvironments in infarcted tissues are characterized by both hypoxia and nutrient deprivation. Therefore, the significant rescue effect observed with TONBs in our 1% FBS model carries far greater translational weight than the rescue shown in high-serum models, as it proves efficacy under the harshest physiological constraints.

The restoration of metabolic activity in TONB-treated cells highlights the capacity of these nanobubbles to serve as “bioenergetic enhancers.” Notably, we identified a cell density-dependent efficacy: TONBs were most effective at higher seeding densities (10,000 cells/well) than at lower densities, a finding that can be explained by the non-linear kinetics of cellular oxygen consumption. In high-density cultures, initially the collective oxygen consumption rate (OCR) rapidly depletes local dissolved oxygen, creating a severe diffusion-limited microenvironment known as “pericellular hypoxia.” This aligns with kinetic modeling in tissue engineering, which demonstrates that reaction parameters, such as maximal uptake (Vmax) and Michaelis–Menten constants (Km), are not static but fluctuate dynamically with cell density [43]. Cells sense the increase in cell population and consequently reduce their maximal single-cell oxygen consumption rate (OCR) and affinity towards oxygen (i.e., decrease the efficiency with which oxygen is captured), as indicated by the increase in Km [43].

Evidence suggests that at high densities, cells often engage in “cooperative behavior,” downregulating individual metabolic demand to prevent rapid oxygen depletion and necrotic core formation [43]. Our data suggests that TONBs can disrupt the hypoxic limit by serving as localized high-capacity oxygen reservoirs. We hypothesize that TONBs also elevate interstitial oxygen tension within the pericellular space, thereby overcoming the oxygen diffusion limitations typical of dense cellular constructs. This enhanced oxygen availability could enable cells to sustain high ATP production (metabolic activity) comparable to normoxic controls. Although further research is required, this could have profound translational implications for 3D bioprinting and scaffold engineering, as TONBs can serve as a critical bridge, satisfying the elevated metabolic requirements of densely seeded tissues during the avascular window before neovascularization is established.

The most significant finding of this study is the elucidation of the molecular mechanism driving this rescue: abrogation of the ROS-HIF-1α axis. The role of reactive oxygen species (ROS) in hypoxia has historically been controversial, stemming from the paradoxical increase in ROS production observed under conditions of oxygen depletion [9]. Although it seems counterintuitive that a scarcity of oxygen would drive oxidative stress, experimental evidence confirms that the mitochondrial electron transport chain (ETC) becomes inefficient under low oxygen tension, leaking electrons to form superoxide [44]. Crucially, studies using genetic approaches have provided compelling evidence that mitochondrial ROS (mito-ROS) production, rather than the lack of oxygen per se, is the indispensable signal required for the hypoxic induction of the ROS-HIF-1α axis [9,45]. In strict alignment with this interplay, we observed that exposure of UC-MSCs to hypoxia precipitated a sharp increase in mito-ROS compared with normoxia. However, one of the most noteworthy findings of our study was that pretreatment with TONBs effectively alleviated hypoxia-induced mitochondrial stress in a dose-dependent manner. Although the protective effect was evident across concentrations, 100% TONB was the most effective, establishing a clear relationship between oxygen delivery and stress reduction. Moreover, we established a direct correlation between the reduction in mito-ROS and the downregulation of HIF-1α. This modulation was particularly robust during the initial stages of hypoxia (2 h), where treatment with 100%, 50%, and 10% TONBs significantly blunted the HIF-1α protein spike. Finally, although we observed a reduction in mitochondrial ROS, we acknowledge the inherent limitations of fluorescent probes in quantifying absolute ROS levels [46,47]. However, the consistent downregulation of HIF-1α—a pathway strictly regulated by mitochondrial oxidant signaling—supports the biological relevance of the observed ROS reduction.

TONBs likely restored the efficiency of the electron transport chain (ETC), reducing electron leakage that generates superoxides. This reduction in mito-ROS (~20%) was mechanistically linked to a dramatic 8–9 fold downregulation of HIF-1α expression. This fundamental alteration of the intracellular signaling landscape suggests that TONBs are potential agents for reprogramming metabolism towards efficient oxidative phosphorylation. However, given the robust ATP production compared to the metabolic profile induced by oligomycin, which we previously identified as the inducer of the most active glycolytic state, we do not rule out the possibility of a dual involvement of TONBs in enhancing both glycolytic and oxidative pathways. These findings align with recent mechanistic studies demonstrating that nanobubble-mediated oxygenation effectively reverses the metabolic shifts associated with hypoxia. For example, Lee et al. (2021) reported that ONBs restored glucose uptake and rescued apoptotic tooth germs under hypoxic stress [30]. Similarly, recent studies on retinal cells have confirmed that ONB treatment significantly suppresses the mRNA expression of key hypoxic markers, including HIF-1α, restoring them to normoxic levels [27]. Collectively, these studies support our observation that TONBs do not merely act as a gas source but actively modulate the cellular hypoxic response. Similar results have also been reported in neuronal cells treated with oxygen nanobubbles for myelin maintenance and repair, supporting the role of oxygen nanobubbles from oxygen supply to metabolic optimization of glycolytic-like high-energy state and OXPHOS metabolism [48]. For instance, the literature also demonstrates that exposure to oxygen nanobubbles in normoxic environments can lead to increased mitochondrial activity in neurons and glial cells, marked by increased PGC-1α levels. This enhancement in mitochondrial dynamics translates to amplified respiratory capacity, ATP production, and augmented mitochondrial content in terms of cell size and area [49,50]. These findings pave the way for further exploration of the metabolic effects of TONBs in UC-MSCs and other cell types, confirming their role not only as oxygen carriers but also as metabolic modulators.

Despite these promising findings, the use of 2D culture systems has some limitations. While the “double-stress” model mimics biochemical ischemia, it lacks the physical diffusion barriers of 3D tissue. Future studies should evaluate TONB perfusion and stability within 3D hydrogels or organoids to fully predict in vivo efficacy. Additionally, while we demonstrated short-term cytocompatibility (up to 72 h), the long-term fate of the fluorosurfactant shell and its potential bioaccumulation require further toxicological assessment. Finally, we focused on survival and metabolism; future studies should assess whether TONB-mediated oxygenation affects long-term biocompatibility, potential intracellular accumulation (PFAS-related concerns), and the immunomodulatory or differentiation potency of UC-MSCs, ensuring that the rescued cells retain their therapeutic phenotype.

5. Conclusions

This study demonstrated the efficacy of TONBs as a cytoprotective platform for oxygen delivery in 2D hypoxic environments. Physically, the fluorosurfactant shell refined the nanocarrier population to a monodisperse size (~123 nm) and significantly enhanced colloidal stability (zeta potential ~−58 mV), enabling sustained oxygen delivery (up to 18 ppm) compared to surfactant-free formulations. Biologically, we validated these effects using a rigorous “double-stress” ischemic model (hypoxia + serum deprivation). By decoupling nutrient availability from oxygen tension, we established that TONBs do not merely supplement oxygen but fundamentally reverse the bioenergetic collapse associated with ischemia. Contrary to the severe metabolic deficits observed in untreated controls, TONB-treated UC-MSCs maintained intracellular ATP levels comparable to normoxic baselines, effectively acting as metabolic optimizers. Mechanistically, this rescue was driven by the interception of the pathological hypoxic signaling loop: TONB treatment attenuated mitochondrial ROS generation by ~20% and suppressed HIF-1α protein expression 8- to 9-fold, thereby preventing the activation of stress-induced apoptotic pathways.

In summary, TONBs represent a scalable, highly stable oxygen delivery system capable of preserving stem cell viability and function under severe metabolic stress in 2D models. Future studies should focus on validating these cytoprotective mechanisms in 3D tissue models and in vivo ischemic injuries to fully translate this technology into clinical therapies.