The composite nanoparticles (Ag@TiO₂NPs) were obtained by loading AgNPs particles on the surface of TiO₂NPs using photocatalytic reduction, as shown in Figure 1. Briefly, P25-TiO2NPs (Sigma, United States) were ultrasonically dispersed in 5% alcohol solution to obtain a suspension of 1 mg/ml, and then the TiO₂NPs suspension was mixed with 1 mg/ml of silver nitrate solution in equal volume. The AgNO₃ & TiO₂ suspensions were irradiated with UV light (UV light intensity = 10 mW/cm², wavelength λ = 365 nm) at 2 mm liquid depth to obtain the Ag@TiO₂NPs suspensions. Then the suspension was centrifuged, washed, and dried at 60°C, obtaining Ag@TiO₂NPs. Photocatalytic reduction times were 1, 5, 20, and 60 s to obtain the composite nanoparticles labeled as “1#”, “2#”, “3#”, and “4#”. Samples storage away from light.

Ultrasonically resuspend Ag@TiO₂NPs with 5% alcohol in the same volume as the Ag@TiO₂NPs suspensions in Section 2.1. subsequently, 1 × 1 cm² PU sheets were immersed in the suspension and deposited for 3 hours at room temperature to form a coating. Medical-grade silicone rubber (SR) catheters were peristaltic pump circulating with Ag@TiO₂NPs suspension for 12 h to obtain the coating for the ex-vivo antithrombogenicity test in Section 2.5.2. After coating, samples were rinsed with RO water, dried under nitrogen, and stored away from light.

Photo-functionalization of the samples was achieved by UV irradiation (UV light intensity = 10 mW/cm², wavelength λ = 365 nm) for 1 h.

The untreated coatings of TiO₂, 1#, 2#, 3#, and 4# nanoparticles are labeled as “UNT-TiO₂C”, “UNT-1#C”, “UNT-2#C”, “UNT-3#C”, “UNT-4#C”, and the corresponding UV photo-functionalized nanoparticles coatings are labeled as “UV-TiO₂C”, " UV-1#C”, " UV-2#C”, " UV-3#C”, “UV-4#C”.

The microstructure and the element distribution of Ag@TiO₂NPs were examined by field-emission transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan) and the energy dispersive spectroscopy (EDS, JEM-2100F, JEOL, Japan).

Water contact angles (WCA) of Ag@TiO₂NPs coatings were obtained by analyzing images using ImageJ software (National Institutes of Health, United States).

Photocatalytic degradation of methylene blue assay was performed to examine the photocatalytic activity of Ag@TiO₂NPs coating. In brief, methylene blue powder was prepared as a 5 mg/L solution in deionized water. A sample was immersed in 1 ml of the solution, and UV irradiated (UV light intensity = 10 mW/cm², wavelength λ = 365 nm) for 1, 3, and 5 h. 200 μL of the solution was collected at each time point, and the absorbance (A) at a wavelength of 664 nm was determined using a microplate reader (BIO-TEK Instruments, United States). The relationship between A and the degradation rate (G) was calculated as G = [ ( A 0 − A t ) / A 0 ] × 100 % (1) where A₀ is the original absorbance of the undegraded methylene blue, and A_t is the absorbance value after t hours of degradation.

In the Ag⁺ release test, atomic absorption spectroscopy (AAS, TAS-990F, Beijing Purkinje General Instrument Co., Ltd., China) was performed to measure the release of Ag⁺ from different samples into phosphate-buffered saline (PBS; pH = 7.4). Briefly, PU samples coated with Ag@TiO₂NPs were immersed in 1 ml of PBS in the dark. The PBS was collected and replaced with 1 ml of fresh PBS every 2 days. This process was repeated for a total of 14 days. All collected PBS solutions were analyzed for their content of Ag⁺, and an Ag⁺ time-release curve was plotted.

In this study, Staphylococcus aureus (S. aureus) and Pseudomonas aeruginosa (P. aeruginosa) were used to evaluate the antibacterial ability of the Ag@TiO₂NPs coating. Briefly, S. aureus and P. aeruginosa obtained from Sichuan Provincial People’s Hospital were inoculated onto blood agar plates and cultivated in an incubator at 37°C. After the appearance of multiple colonies, S. aureus and P. aeruginosa were collected and dispersed in an F12 medium containing 10% fetal bovine serum (FBS, Sigma, United States). The density of S. aureus and P. aeruginosa was adjusted to 1 × 10⁶ colony forming units (CFU)/mL, and then 1 ml of bacterial suspension was added to the wells of a 24-well plate to soak the samples. After incubation at 37°C for 6 hours, the samples were washed three times and then transferred to a new 24-well plate containing a mixture of F12 FBS medium and cell counting kit-8(CCK-8, APExBIO Ltd., Houston, United States) agent. After 3 h of incubation, the activity of adherent bacteria was tested by detecting the absorbance of the medium at a wavelength of 450 nm using a microplate reader. The samples were then fixed with 2.5% glutaraldehyde for 12 h, after which they were dehydrated, and the adherent bacteria on the samples were observed and analyzed with an optical microscope (OM, DM4000M; Leica, Germany).

To confirm that Ag@TiO₂NPs were successfully prepared, TEM was used to characterize the microstructure of the nanoparticles, and EDS was used to perform elemental analysis of the nanoparticles. From the TEM results, additional spherical secondary structures with diameters of 5–10 nm were observed on the surface of 1# nanoparticles compared to TiO₂NPs (Figure 2A). the signal of elemental Ag appeared in the EDS data for 1# nanoparticles (Figure 2B), which indicates that the spherical secondary structures appearing on the surface of 1# nanoparticles are AgNPs. These AgNPs should be the products of Ag⁺ reduction by electrons generated on the surface of TiO₂NPs during the photocatalytic reduction process (Lu et al., 2013; Chen et al., 2021).

The results of TEM and EDS indicate that Ag@TiO₂NPs were successfully prepared. Furthermore, the color of the suspensions of Ag@TiO₂NPs deepened with the prolongation of the photocatalytic reduction time, which implied that the AgNPs content in Ag@TiO₂NPs might increase with the photocatalytic reduction time prolongation (Lu et al., 2013). The color of the coatings deposited by Ag@TiO₂NPs onto the PU sheets also showed the same trend (Figure 2C).

The photo-induced hydrophilicity of TiO₂ has been reported in many pieces of literature (Carp et al., 2004). Excellent hydrophilicity is considered to facilitate biofouling resistance (He et al., 2021). Compared with the non-photo-functionalized UNT-Ag@TiO₂NPs coatings, the photo-functionalized UV-Ag@TiO₂NPs coatings showed a significant decrease in WCA and exhibited strong photo-induced hydrophilicity (Figure 2D), which means that photo-functionalized nanoparticle coatings may be somehow more anticoagulant and antibacterial than their counterparts in the group. Among all UV-Ag@TiO₂NPs coatings, 2# coating had the highest water contact angle, perhaps related to the content of AgNPs on the surface of Ag@TiO₂NPs, the exact mechanism of which needs to be further investigated.

The photocatalytic oxidation activity of TiO₂ is another manifestation of the photocatalytic activity of TiO₂, which is closely related to the mechanism of photo-induced anticoagulant properties of TiO₂ (Chen et al., 2014; Chen et al., 2015). Although it has been reported in the literature that loading AgNPs on the surface of TiO₂ can enhance the photocatalytic activity of TiO₂ by forming Schottky energy barriers (Gomathi Devi and Kavitha, 2016). However, in this study, through the photocatalytic oxidation decomposition of methylene blue, we found that the loading of AgNPs did not significantly affect the photocatalytic oxidation activity of Ag@TiO₂NPs coating, compared to TiO₂NPs coating (Figure 2E).

The Ag⁺ release rate is closely related to AgNPs’ biosafety and antimicrobial properties (Marchioni et al., 2018). 1# nanoparticles were used to analyze their Ag⁺ release in PBS after coating on PU, and it was found that the Ag⁺ release rate of UNT-1#C was stable around 0.35 μg cm⁻² day⁻¹ (Figure 2F), which was considered to be safe (Liu et al., 2018).

Here, Gram-positive Staphylococcus aureus (S. aureus) and Gram-negative Pseudomonas aeruginosa (P. aeruginosa) were used to examine the broad-spectrum antimicrobial resistance of the Ag@TiO₂NPs coatings. The results of optical microscopy showed that morphologically intact S. aureus and P. aeruginosa could be seen on the surface of all PU and TiO₂NPs coatings, in contrast to all Ag@TiO₂NPs coatings with a large amount of debris, probably necrotic bacterial fragments (Figure 3A).

The results of the cellular activity analysis (Figure 3B) were consistent with the observation under optical microscopy, where the bacterial activity on the surface of the TiO₂NPs coating was similar to that of the PU. In contrast, all Ag@TiO₂NPs coatings exhibited significant antimicrobial activity compared to the PU substrate and the TiO₂NPs coating. Notably, there was no significant difference between the antimicrobial properties of UNT-Ag@TiO₂NPs coatings and UV-Ag@TiO₂NPs coatings, suggesting that the photo-functionalization treatment did not affect the AgNPs portion of Ag@TiO₂NPs to perform the antimicrobial function. Another noteworthy point is that the antimicrobial performance of 1# coatings with the lowest AgNPs loading is comparable to that of 4# coaitngs with the highest AgNPs loading, indicating a powerful antimicrobial performance.

In vitro platelet adhesion experiments were performed using deposited Ag@TiO₂NPs coatings on PU sheets, for PU is widely used in various blood contact devices. The results of fluorescence images (Figure 4A) and adhesion density (Figure 4C) showed that all nanoparticles coatings significantly reduced the number of platelet adhesions after photo-functionalization compared with that before treatment. Besides, the platelet adhesions density on the photo-functionalized nanoparticle coatings was significantly lower than that of PU substrates. Notably, the platelet adhesion density of nanoparticles coatings tended to increase with the prolongation of photocatalytic reduction time in the composite nanoparticles preparation stage. Considering the issue of blood compatibility of AgNPs, this result is similar to the phenomenon that the color of Ag@TiO2NPs suspension deepens with increasing photocatalytic reduction time (Figure 2C), supporting the idea that the AgNPs content in Ag@TiO2NPs increases with increasing photocatalytic reduction time.The SEM results of platelet adhesion (Figure 4B) showed that the platelets on the PU substrate were in a partially spreading dendrites state, with the degree of spread much higher than that of platelets on the UV-photo-functionalized nanoparticles coating, which was between spherical to partially spread. In contrast, the degree of platelet spreading on all nanoparticles coatings showed a tendency to increase with increasing photocatalytic reduction time at the composite nanoparticles preparation stage. This trend is consistent with the results of platelet adhesion density, which could be similarly influenced by the content of AgNPs in the composite nanoparticles.

The results of platelet adhesion experiments suggested that the Ag@TiO₂NPs coatings can acquire anti-platelet adhesion ability after photo-functionalization. However, this photo-induced anticoagulant properties of the UV-Ag@TiO₂NPs may be weakened by increasing the content of AgNPs, this implies that the photo-induced anticoagulant properties of TiO₂ have a limited or dose-dependent improvement on the hemocompatibility of AgNPs. Therefore, proper loading of AgNPs is essential to impart UV-Ag@TiO2NPs with excellent antimicrobial properties without causing deterioration of hemocompatibility.

Here, combined with the results of the antimicrobial experiments and in vitro platelet adhesion assays, the UV-1# sample with the lowest silver loading may be an ideal candidate capable of obtaining balanced antimicrobial and anticoagulant properties.

The Antithrombogenicity test by ex-vivo blood circulation provides a more comprehensive test of the anticoagulant capacity of the UV-Ag@TiO₂NPs coating compared to the in vitro blood test. SR is a material commonly used for central venous catheters, which is a typical medical device that requires both anticoagulation and antimicrobial activity. Considering the in vitro platelet adhesion and Antibacterial Assay results, SR catheters deposited with 1# nanoparticles were selected for the ex-vivo blood circulation assay. After 30 min of ex-vivo blood circulation, the surface of the UNT-1# coating showed patches of sizable thrombus layer that blocked nearly 8% of the catheter lumen compared to the UV-1# coating with only a small amount of thrombus on the surface (Figures 5B,D). The analysis of the weight of the formed thrombus showed a significant increase in UNT-1# coating compared to both UV-1# coating and PU substrate (Figure 5C). At the same time, there was no significant difference in UV-1# coating compared to PU substrate.

SEM images of the inner wall of the catheter revealed that UNT-1# coating had fully initiated the coagulation mechanism, and the fibrinogen network captured a large number of red blood cells, forming a thrombus layer that completely covered the substrate. In contrast, the UV-1# coating and the PU substrate were similar, with only some platelets adhering to the surface (Figure 5E). It is worth noting that most of the erythrocytes in the thrombus layer remain in typical form, and combined with the data on the rate of silver ion release from the coating (Figure 2F), hemolysis may not be a problem that the coating would cause.

Here, UV-Ag@TiO₂NPs coating did not show significantly better results than PU substrates in the ex-vivo blood assay as in the in vitro platelet adhesion assay. This may be due to the dynamic environment and more comprehensive blood composition in the ex-vivo blood circulation assay (Courtney et al., 1994). Nevertheless, the results of in vitro platelet adhesion and ex-vivo blood experiments were sufficient to suggest that loading AgNPs onto the surface of TiO₂NPs and performing UV- Photo-functionalization treatment was an effective way to improve their hemocompatibility.