Microenvironment regulation of ruthenium complexes based on PDMS microsphere carriers and highly sensitive detection of dissolved oxygen

Accurate monitoring of dissolved oxygen concentration in vivo is crucial for understanding cellular metabolism and pathological changes in the microenvironment. However, traditional oxygen sensing methods such as clark electrodes and magnetic resonance spectroscopy suffer from drawbacks including tissue damage, high cost, or insufficient spatial resolution. This study aims to develop a novel, cost-effective, biocompatible, and non-invasive oxygen sensor based on microfluidic technology for continuous, in situ monitoring of dissolved oxygen in the cellular microenvironment. We fabricated Polydimethylsiloxane flexible microspheres using microfluidic techniques, incorporating Ru (dpp) 3 Cl 2 nanoparticles (a derivative of tris(4,7-diphenyl-1,10-phenanthroline)ruthenium (II) dichloride) as an oxygen-sensitive indicator. By optimizing the flow rate ratio between the dispersed oil phase and continuous aqueous phase, as well as the concentration of Ru (dpp) 3 Cl 2 , we obtained microspheres with the highest fluorescence quenching constant and optimal dissolved oxygen responsiveness. These microspheres were subsequently co-cultured with three cancer cell lines (MDA-MB-231, HeLa, and A549) to dynamically monitor dissolved oxygen levels. To further evaluate their potential as a drug delivery platform, paclitaxel was loaded into the microspheres, demonstrating excellent biocompatibility. This study successfully achieves highly sensitive, in situ , and cost-effective visual monitoring of dynamic Dissolved oxygen fluctuations. The developed microspheres hold significant application prospects in the field of bio-oxygen sensors, providing new insights for biomedical research.

Dissolved oxygen (DO) as a fundamental regulator of biological processes, plays an essential role in sustaining marine ecosystem functions and supporting diverse life activities (Granata et al., 2024; Shaghaghi et al., 2024). At the cellular level, oxygen concentration serves as a critical indicator of cellular metabolic activity, physiological state, and pathological alterations within the microenvironment (Marassi et al., 2022; Raheja et al., 2010). Therefore, accurate monitoring of in vivo DO levels is of great significance for maintaining biological homeostasis and regulating physiological functions (Acker, 1994; Csete, 2005). From the perspective of cellular physiological mechanisms, oxygen imbalance can initiate a cascade of pathological responses. Excessive oxygen levels may promote the generation of reactive oxygen intermediates (ROIs), which can induce oxidative modifications of cellular macromolecules and ultimately result in cell dysfunction or death (Fridovich, 1999; Kazzaz et al., 1996). Conversely, insufficient regulation of reactive oxygen species (ROS) generation, caused by the auto-oxidation of oxidases, components of cellular electron transport pathways, or quinone compounds such as menadione, may also lead to cellular damage and apoptosis (de Groot and Littauer, 1989). Furthermore, hypoxic environments can induce the transcriptional upregulation of hypoxia-sensitive genes in most cell types (Ward, 2008). All these phenomena indicate that the development of oxygen sensors capable of achieving continuous and in-situ monitoring of biological DO concentrations holds significant importance for life science research and clinical applications.

However, existing mainstream oxygen sensing technologies still face substantial challenges in practical applications, particularly in simultaneously achieving continuous, in situ monitoring, cost-effectiveness, and biocompatibility (Majhi et al., 2022). Specifically, the Clark electrode method, a widely used technique for DO measurement, may compromise the structural integrity of organoids or tissues due to the invasive nature of electrode insertion. Moreover, it remains challenging to achieve spatially resolved and accurate in situ oxygen monitoring within different regions of organoids (Clark and Clark, 1987). Fluorescent probe-based approaches, although highly sensitive, are limited by the potential cytotoxicity of certain oxygen-responsive dyes, which may adversely affect long-term cell viability and physiological functions. In addition, the design and optimization of suitable probes remain technically challenging, and these approaches are generally insufficient for long-term in situ monitoring (Jiang et al., 2018). While magnetic resonance spectroscopy (MRS) provides a non-invasive means for oxygen detection; however, its high instrumentation costs and operational complexity restrict its widespread application in practical settings (Wolfbeis, 2015). Microfluidic and chip-based platforms have emerged as powerful tools for investigating cellular microenvironments due to their capability for precise manipulation of cells and integration of multiple analytical functions (Wang and Kelley, 2025; Zhang et al., 2022; Zhang et al., 2026; Zhang et al., 2023). Thus, developing a novel oxygen sensing strategy based on microfluidic technology has become a key direction to overcome the limitations of existing technologies.

Based on these considerations, this study presents a novel strategy for the quantitative monitoring of in vivo DO, using PDMS based oxygen-sensing microspheres as the core component. The microspheres operate through the oxygen-dependent fluorescence quenching of the encapsulated ruthenium complex Ru (dpp) 3 Cl 2, enabling quantitative determination of DO concentrations (Mingoarranz et al., 1995; Zhou et al., 2020). Compared with conventional oxygen-sensing microspheres, PDMS based microspheres possess excellent flexibility, elasticity, and compatibility with oxygen-sensitive probes (Li et al., 2018). Most importantly, they facilitate in situ, real-time oxygen monitoring, overcoming a key limitation of conventional oxygen-sensing techniques. These advantages make PDMS oxygen-sensing microspheres a promising platform for continuous real-time monitoring of DO dynamics in microfluidic systems.

The oxygen-sensitive indicator employed in this work is tris (4,7-diphenyl-1,10-phenanthroline) ruthenium (II) dichloride, which belongs to a class of compounds characterized by their stable coordination structure, typically consisting of a ruthenium ion core and surrounding organic ligands (Meyer and Huynh, 2003). The primary advantage of ruthenium complex-based oxygen-sensitive materials lies in their high responsiveness to variations in oxygen concentration. Upon exposure to oxygen at different concentrations, ruthenium complexes undergo reversible oxygen-induced quenching processes, resulting in a reduction in fluorescence intensity (Yao et al., 2023). By correlating the fluorescence quenching behavior with oxygen concentration, the oxygen content can be quantitatively determined with high accuracy (Pagire et al., 2020). Based on this characteristic, Ru (dpp) 3 Cl 2 was selected as the key sensing component in the oxygen-sensitive indicator developed in this work. Experimental results demonstrated that insufficient doping of Ru (dpp) 3 Cl 2 into the microspheres resulted in poor sensitivity toward oxygen concentration variations. Increasing the Ru (dpp) 3 Cl 2 loading concentration enhanced the oxygen sensitivity; however, excessive incorporation of the indicator reduced the oxygen quenching efficiency of individual indicator molecule due to concentration-related effects (Chen et al., 2013). Therefore, the doping concentration of the Ru (dpp) 3 Cl 2 in the microspheres was systematically optimized to achieve an appropriate balance between indicator loading and oxygen-responsive performance.

In terms of microsphere preparation technology, a cross-flow focusing microspheres was selected as the key microfluidic device. In this system, the flow rates of the dispersed oil phase and the continuous aqueous phase, as well as their flow-rate ratio, are critical factors affecting the stability of microsphere formation, particle size, size distribution, and overall morphology (Mashhadian and Shamloo, 2019). Microfluidic techniques have also been widely used for the controlled fabrication of cell-compatible microspheres and microcapsules because they enable the regulation of particle size, morphology, and structural uniformity (Liu et al., 2021). Accordingly, the present study employed a cross-flow focusing microfluidic device to prepare PDMS oxygen-sensing microspheres with controlled size and uniformity. Generally, the oil-phase flow rate should be maintained at a relatively low level (1–10 μL/min) to ensure the stable droplet generation and enable effective control of microsphere size. In contrast, the aqueous-phase flow rate is typically higher (10–100 μL/min), providing hydrodynamic shear to force and the disperse the oil phase, thereby improving microsphere size uniformity (Gao et al., 2023). To optimize the microsphere preparation process, the oil-phase flow rate was fixed at 200 μL/h. Four oil-phase-to-aqueous-phase flow-rate ratios, namely, 1:10, 1:15, 1:20, and 1:25. Were investigated. The microspheres obtained under these conditions were systematically evaluated with respect to particle size, size distribution, and morphological quality to identify the optimal flow-rate ratio for oxygen-sensing microsphere preparation.

To evaluate the effectiveness and practical application of the proposed monitoring method, a multi-stage validation study was conducted. First, oxygen-sensing microspheres with the desired size and performance characteristics were selected. Their oxygen-sensing stability was assessed by fitting calibration curves, there by establishing a quantitative reference for DO measurement (Zhao et al., 2022). Subsequently, PDMS Ru (dpp) 3 Cl 2 oxygen-sensing microspheres were incorporated into the culture systems of different cell types. Dynamic fluorescence measurements at multiple time points were performed to characterize and compare the oxygen consumption profiles of different cells during culture. Building on these results, the potential of PDMS microspheres as multifunctional drug delivery platforms was further investigated, taking advantage of their suitable size, stable performance, performance, and capability for in situ oxygen monitoring. Paclitaxel (PTX) was loaded into the PDMS microspheres to enable simultaneous drug delivery and oxygen monitoring (Eluu et al., 2024; Zhang et al., 2013). The sustained release of paclitaxel from the PDMS microspheres was evaluated by dynamically monitoring cellular oxygen consumption, providing a real-time indicator of therapeutic efficacy (Seyam et al., 2025). In addition, the application of the proposed oxygen sensor was demonstrated by monitoring oxygen kinetics during cell culture–based biochemical analyses in a microfluidic platform.

Recent advances in organoid-on-chip and microengineered culture platforms have further demonstrated the potential of microfluidic systems to provide controllable and physiologically relevant in vitro microenvironments for the investigation of complex biological processes (Zhang et al., 2025a; Zhang et al., 2025b).

Materials

Negative photoresist SU-8 2050 was purchased from MicroChem Corporation (USA). Silicon wafers were obtained from Tianjin Xilika Crystal Polishing Material Co., Ltd. (China). Ruthenium (II) tris (4,7-diphenyl-1,10-phenanthroline) dichloride Ru (dpp) 3 Cl 2 was acquired from Bide Pharmatech Ltd. Propylene glycol monomethyl ether acetate (developer solution) was purchased from Beijing Bailingwei Technology Co., Ltd. (China). Polydimethylsiloxane (PDMS) and PDMS prepolymer were both obtained from Dow Corning Corporation (USA). Fe 3 O 4 nanoparticles and polyvinyl alcohol (PVA) were purchased from Shanghai Maclin Biochemical Technology Co., Ltd. (China). Other materials included PTX, PBS buffer (pH = 7.4, sterile), DMEM medium (supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin), trypsin (0.25%), as well as MDA-MB-231 cells, HeLa cells, and A549 cells were purchased from Shanghai Jinyuan Biotechnology Co., Ltd. (China).

Preparation of PDMS Ru (dpp) 3 Cl 2 photo-oxygen sensing microspheres

Preparation of the Oil Phase. First, 5 mg of Ru (dpp) 3 Cl 2 was dissolved in 1 mL of absolute ethanol, and magnetically stirred until Ru (dpp) 3 Cl 2 was completely dissolved to obtain a Ru (dpp) 3 Cl 2 -ethanol solution. A 0.5 mL aliquot of the above solution was pipetted and added to 3.5 g of PDMS solution (with a mass ratio of prepolymer to curing agent of 10:6). Subsequently, 50 μL of Fe 3 O 4 nanoparticles were added to endow the PDMS microspheres with magnetic responsiveness for flexible manipulation and spatial localization in future microfluidic research (Zhang et al., 2025c) and the mixture was stirred to achi eve uniform dispersion. The solution was then placed in a vacuum desiccator to remove air bubbles, which eliminated bubble interference on subsequent microsphere formation. Finally, the solution was transferred to a syringe.

Preparation of the Aqueous Phase. First, the temperature of a constant-temperature water bath was set to 85 °C. After the temperature stabilized, 100 mL of purified water was added to a conical flask with a lid, and the flask was placed in the water bath. A total of 10 g of PVA powder was weighed and added to the water in portions (2–3 g per portion). After each addition, gentle stirring was performed to ensure uniform dispersion of PVA and prevent agglomeration; subsequent additions were conducted only after the previously added PVA was completely dissolved. Continuous stirring was maintained throughout the dissolution process until PVA was fully dissolved and the solution became transparent and homogeneous, after which heating was stopped. The solution was filtered through a membrane to remove impurities and finally transferred to a syringe for later use.

Syringes containing the oil phase and aqueous phase were respectively mounted on a Harvard syringe pump. The Harvard syringe pump was activated to initiate fluid flow, and the fluids formed microspheres via shear forces in the cross-flow focusing channel Supplementary Figure S1.

Optimization of flow rate ratio

In the microsphere preparation process, optimizing the liquid flow rate ratio in the cross-flow focusing channel is a crucial step. The adjustment of the flow rate ratio directly affects the morphology, size uniformity, and reaction efficiency of the microspheres; thus, optimizing the flow rate ratio is one of the key steps to regulate the performance of PDMS oxygen-sensing microspheres. In this experiment, four microfluidic chips with different flow rate ratios were designed. A core oil-phase flow rate of 200 μL/h was adopted, and flow rate ratios of 1:10, 1:15, 1:20, and 1:25 were set. The size and uniformity of microspheres generated under different flow rate ratios were compared to screen out microspheres with favorable size and uniformity for subsequent applications.

Detection and optimization of O 2 response for oxygen sensing microspheres

In the experiment, the microsphere suspension was first introduced into a Petri dish. The concentration of DO in the environment was precisely controlled by adjusting the oxygen flow rate, while a DO sensor was used to monitor the real-time oxygen content. To systematically evaluate the response characteristics of the microspheres to oxygen content, a gradient of DO concentrations ranging from 0 to 10 mg/L, with an interval of 2 mg/L, was applied. Under each concentration condition, fluorescence images of the microspheres were acquired using a fluorescence microscope, and then ImageJ image analysis software was employed to quantitatively analyze the fluorescence intensity of the microspheres. For each DO concentration group, the average value of fluorescence intensity was calculated to reduce errors. The fluorescence intensity data under different oxygen concentrations were compared, and a relationship curve between fluorescence intensity and DO concentration was plotted.

The concentration of the oxygen-sensitive material in the oxygen-sensing microspheres directly affects the oxygen-sensing performance of the microspheres. Based on this, four microfluidic chips with different Ru (dpp) 3 Cl 2 concentrations (0.5 mg, 1.0 mg, 1.5 mg, and 2.0 mg) were designed in this experiment. The uniformity and oxygen-sensing performance of the microspheres prepared under different concentrations were compared Figure 1d.

Application of oxygen-sensing microspheres for dissolved oxygen monitoring in multiple cancer cell types

Three types of cancer cells in the logarithmic growth phase were collected and digested with 0.25% trypsin for 2 min. Digestion was terminated by adding DMEM medium, and cells were resuspended into a single-cell suspension by pipetting. Ten microliters of the cell suspension were cultured for 24 h to ensure adherent growth. Under a sterile laminar flow hood, 100 μL of preformed PDMS ruthenium composite microspheres were added to the microchamber. Ten microliters of the cell suspension were introduced into the fabricated microfluidic chip (Supplementary Figure S3) and cultured for 24 h to allow cell attachment.

Fluorescence intensity of microspheres within the MDA-MB-231, HeLa, and A549 cell systems was continuously monitored over an 8 h observation period, with measurements recorded at 0, 2, 4, 6, and 8 h. This monitoring period was selected to capture the short-term dynamic changes in dissolved oxygen associated with cellular respiratory metabolism under stable co-culture conditions. Fluorescence images of the microspheres were acquired using a fluorescence microscope, and the fluorescence intensity was quantified using ImageJ image analysis software. Based on the oxygen-dependent fluorescence response of the microspheres, the dynamic changes in DO concentration over time were subsequently determined and plotted, enabling a comparison of the respiratory metabolic characteristics of the three cancer cell types.

Co-culture of PTX loaded PDMS oxygen-sensing microspheres with different cancer cells

MDA-MB-231, HeLa, and A549 cells were respectively seeded into chips containing DMEM medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Prepared 20 μm PDMS Ru (dpp) 3 Cl 2 oxygen-sensing microspheres were also added to the chips to ensure uniform dispersion of the microspheres and their co-culture with the cells. All chips were incubated in a cell incubator at 37 °C with 5% CO 2.

On this basis, PDMS microspheres were combined with PTX. The chips were gently shaken to disperse the microspheres uniformly in the medium and prevent microsphere aggregation. All chips were returned to the cell incubator to start co-culture. At each monitoring time point (0 h, 2 h, 4 h, 6 h, and 8 h of culture), the chips of the corresponding groups were taken out of the incubator. ImageJ software was used to quantitatively analyze the fluorescence intensity ratio of the microspheres in each visual field.

Detection and optimization of oxygen-sensing performance for oxygen-sensing microspheres

Based on the oxygen-sensing principle of PDMS oxygen-sensing microspheres Figures 2a–c, an oxygen concentration as the abscissa and fluorescence quenching ratio as the ordinate, a linear fitting method was used to plot the oxygen-sensing standard curve, which was based on the Stern–Volmer equation:

I 0 is the fluorescence intensity in the absence of oxygen, I is the intensity in the presence of oxygen, K s ᵥ is the Stern-Volmer constant, and [O 2 ] stands for the oxygen concentration. This relationship enables the determination of oxygen concentration based on changes in fluorescence intensity (Gehlen, 2020; Keizer, 1983).

Microspheres with uniform fluorescence and appropriate size were selected for oxygen-sensing evaluation. The fitting curve results showed that the oxygen-sensing standard curve was y = 0.21072x + 0.94368 with R 2 = 0.997, confirming their suitability for DO monitoring Figures 1a–c. The oxygen quenching process is dynamic, involving energy transfer between oxygen molecules and excited-state fluorescent molecules. This energy transfer causes the excited fluorescent molecules to lose energy and undergo quenching. Using microspheres doped with varying amounts of Ru (dpp) 3 Cl 2 and measuring fluorescence under different dissolved-oxygen concentrations, Stern–Volmer equation (I 0 /I = 1 + K s ᵥ[O 2 ]) was used to determine quenching constants and construct standard curves. The highest sensitivity was observed at a doping level of 1.0 mg, the oxygen-sensing standard curve was y = 0.21572x + 0.94842. At this concentration, the microspheres exhibited the largest fluorescence quenching constant and the highest responsiveness to DO Figure 1d. Lower dopant levels provided insufficient quencher molecules and reduced sensitivity, where as excessive doping impaired performance due to molecular aggregation and structural damage.

Optimization of flow rate ratio for PDMS Ru (dpp) 3 Cl 2 photo-oxygen sensing microspheres

Microsphere were generated using a cross-flow focusing channel in Figure 3a. Excessively large microspheres may increase the risk of physical obstruction within the microfluidic channels and may interfere with stable fluid flow during subsequent experiments, Conversely, excessively small microspheres may provide weaker fluorescence signals and are more difficult to stably visualize and manipulate during microscopic observation. Therefore, optimization of the flow rate ratio was critical. At flow rate ratios of 1:10 and 1:15, the microspheres were excessively large, with diameter ranges of 115–160 μm and 90–130 μm, respectively. Increasing the flow rate ratio produced progressively smaller particles at 1:20, the microspheres exhibited a narrower size distribution and stable morphology, with diameters of 35–75 μm. This size range is appropriate for cellular oxygen sensing because it allows effective contact with cells while minimizing the risk of channel clogging. Accordingly, a flow rate ratio of 1:20 was selected as optimal for this study Figures 3b-e. The microspheres prepared at the flow rate ratio of 1:20 exhibited an average particle size of 56.58819 ± 9.1732 μm (mean ± SD, n = 100).

Monitoring of dissolved oxygen in different cancer cells

PDMS microspheres mitigate limitations of conventional cancer cell oxygen monitoring methods, such as large detection volume and inaccurate oxygen capture (Yu et al., 2013). To quantify oxygen consumption kinetic and validate PDMS microspheres as local oxygen probes, we used them to monitor DO in various cancer cells.

Bright field, imaging showed no significant morphological differences at different culture times (0 h, 2 h, 4 h, 6 h, 8 h), indicating that the microsphere structure remained intact during the experiment. Under dark field, microsphere fluorescence intensity increased progressively with culture time Figure 4a. Because the sensing dye is quenched by oxygen, the observed fluorescence enhancement corresponds to reduced local DO concentration around the microspheres.

For the three cancer cell lines (MDA-MB-231, HeLa, and A549), the DO concentration showed a continuous downward trend as the culture time increased from 0 h to 8 h, and there were differences in the DO consumption rate among different cell lines Figure 4b. At the initial stage, there was no significant difference in DO concentration among the three cancer cell types, all maintaining a relatively high level. This indicates that the metabolic activity of cells was not fully activated at the beginning of culture, resulting in low DO consumption. During the metabolically active stage, the DO concentration decreased rapidly. At the later stage of culture, the DO concentration decreased to a low level, indicating that sustained cellular metabolism continuously consumed DO during prolonged culture, leading to a significant reduction in oxygen content in the environment.

Monitoring of dissolved oxygen during co-culture of different cancer cells with PTX-loaded PDMS microspheres

Mechanism of action of PTX on cancer cells Figure 5a Under bright-field microscopy, PTX-loaded PDMS microspheres were uniformly distributed around MDA-MB-231, HeLa, and A549 cells, and no apparent morphological abnormalities were observed throughout the co-culture period. These observations suggest that the PTX loaded PDMS microspheres exhibit good biocompatibility and do not induce acute physical damage or cytotoxic effects under the experimental conditions.

The dynamic changes in DO concentration were evaluated in the different co-culture systems. In all three cancer cell models, the DO concentration gradually decreased over the 8 h culture period, reflecting continuous oxygen consumption associated with cellular metabolism. Notably, the decline in DO concentration was significantly greater in the PTX loaded PDMS microsphere group than in both the blank PDMS microsphere group and the control group, suggesting that PTX treatment altered the metabolic oxygen consumption profile of the cancer cells Figures 5b–d. In contrast, no significant difference in DO concentration was observed between the blank PDMS microsphere group and the control group, confirming that the PDMS microspheres themselves did not affect cellular oxygen consumption. These findings further validate the reliability of the experimental system and demonstrate that the observed changes in DO were attributable solely to cellular metabolic activity and the effect of PTX.

In conclusion, we developed a method capable of detecting changes in environmental dissolved oxygen during cell growth and metabolism. Ruthenium (II) tris (4,7-diphenyl-1,10-phenanthroline) dichloride [Ru (dpp) 3 Cl 2 ], an oxygen-sensitive material, was doped into PDMS microspheres, enabling the sensing of subtle changes in environmental dissolved oxygen. Meanwhile, the size and performance of these microspheres were optimized; the results showed that when the doping amount of Ru (dpp) 3 Cl 2 was 1.0 mg and the flow rate ratio was 1:20, microspheres with appropriate and uniform size could be prepared.

In addition, changes in the fluorescence intensity of the microspheres could be observed under a fluorescence microscope, realizing real-time imaging of dissolved oxygen changes in the cancer cell growth environment. Furthermore, by combining PDMS microspheres with PTX (an anticancer drug), their drug delivery capability was successfully verified.

The PDMS Ru (dpp) 3 Cl 2 oxygen-sensing microspheres exhibit advantages including in-situ monitoring, controllable size, high sensitivity, simple operation, and easy visualization via fluorescence microscopy, while reducing the overall cost (Lim et al., 2024). In the future, by modifying the stimuli-responsive molecules in PDMS, this method can be extended to detect other molecules and ions in the cell microenvironment, and holds potential applications in biomedical research.