Evaluation of the anti-thrombosis efficacy of MPC-based copolymer coatings in high coagulation risk blood

Abstract

Blood-contacting devices provide dual risks of thrombosis and infection in clinical applications. Conventional anticoagulants cause adverse effects and exhibit inadequate stability in hypercoagulable states. In this study, we broke away from the traditional understanding that zwitterionic hydrophilic groups dominate anti-fouling properties. We synthesized a ternary copolymer (PMLT) from MPC, LMA, and TSMA to investigate hydrophilic group proportion effects on thrombosis resistance, especially in hypercoagulable blood. Simply increasing the density of phosphocholine (PC) groups in the coating resulted in the coating losing its anticoagulant efficacy in hypercoagulable blood. Conversely, the composition-optimized PMLT-12 coating maintained a stable biomimetic bilayer structure. It demonstrated low protein adsorption and high antibacterial activity under normal conditions. Crucially, PMLT-12 retained excellent anti-thrombotic performance in challenging environments, including blood containing elevated levels of calcium ions and lipopolysaccharides (LPS), and blood from a diabetic animal model. The covalently crosslinked network mediated by TSMA concurrently enhanced the mechanical stability of the coating. This study highlights the critical role of hydrophobic–hydrophilic balance in anticoagulant efficacy against high coagulation risk, providing a novel strategy for improving blood-contacting device surfaces.

---

1. Introduction

Blood-contacting catheters, such as central venous catheters and extracorporeal circulation circuits, serve as critical components in modern clinical therapy. Their substantial clinical demand has become a major focus of research.^1,2 Studies indicate that common catheter materials include polyvinyl chloride (PVC), polyurethane (PU), and silicone.^3–6 Although these materials possess excellent mechanical properties and broad applicability, they trigger the coagulation cascade upon blood contact. This leads to thrombosis and distal embolism,^7–9 while increasing infection risks due to protein adhesion.^10,11 Currently, common solutions include the use of anticoagulant drugs and material surface modification. For systemic anticoagulant drugs, long-term use is associated with numerous adverse effects.^12,13 In terms of surface modification strategies, traditional methods mainly rely on hydrophilic polymer coatings, such as polyethylene glycol (PEG)^14,15 and polyvinylpyrrolidone (PVP).¹⁶ However, the immobilization of traditional coatings primarily depends on surface chemical grafting, and their anti-fouling efficacy is highly dependent on chain conformation, surface grafting density, and surface roughness.¹⁷ Furthermore, it is challenging to achieve treatment and uniform modification of the inner wall of catheters,^18,19 which limits the feasibility of such coatings in practical applications. Recently, zwitterionic materials have attracted significant attention due to their ultralow fouling properties, attributed to their uniformly distributed cationic and anionic functional groups at the molecular level.^20–22 Zwitterionic polymers can form a robust hydration layer through strong surface ionic hydration to resist initial protein adsorption,^23–25 and have thus become the preferred component in surface modification strategies for blood-contacting catheter materials.

Zwitterionic coatings can exhibit excellent anti-fouling performance in normal blood within traditional anticoagulant evaluation systems, particularly 2-methacryloyloxyethyl phosphorylcholine (MPC) zwitterionic materials. These materials mimic the structure of phosphatidylcholine headgroups in cell membranes and have become the gold standard for anticoagulant coating design.^26–28 However, the vast majority of studies have focused on blood environments with normal coagulation risk, while neglecting the impact of pathologically hypercoagulable blood on coating functionality, such as in sepsis, diabetes, and post-operative inflammatory states.^29–31 The absence of such an evaluation system raises a critical question regarding blood-contacting materials in clinical environments. When faced with explosive increases in coagulation factors (e.g., sepsis³²) or pathological platelet activation (e.g., in diabetes³³), can traditional MPC-based copolymer designs still maintain excellent anticoagulant performance? Of particular consideration is whether the prevailing view that the density of hydrophilic PC groups determines anticoagulant properties remains valid in hypercoagulable environments. Therefore, further investigation and exploration are required for the design and evaluation of MPC copolymers under pathologically hypercoagulable blood conditions, which is of great significance for the clinical application versatility of blood-contacting catheter materials.

Based on the considerations, to reveal the effect of the regulatory mechanism of MPC copolymer ratios on anticoagulation in pathologically hypercoagulable blood environments, we designed a terpolymer poly (MPC-co-LMA-co-TSMA) (PMLT) by building upon traditional MPC copolymers reported in previous studies (Scheme 1).^34–36 In this terpolymer, the introduced γ-methacryloxypropyl trimethoxysilane (TSMA) containing crosslinking groups enables the internal silane groups to hydrolyze and bind to the substrate, forming an intramolecular crosslinked network to improve the mechanical stability of the polymer coating. This study focuses on evaluating the dynamic synergistic mechanism of hydrophilic–hydrophobic components in MPC polymers under hypercoagulable blood conditions. By systematically regulating the molar ratio of MPC to LMA, a series of coatings were constructed using a one-step dip-coating method, and evaluations were performed using normal blood and diabetic animal blood systems. Notably, when the molar ratio of MPC to LMA was 1 : 2 (PMLT-12), the coating exhibited excellent anticoagulant performance in hypercoagulable blood environments, demonstrating the synergistic effect of hydrophilic–hydrophobic group balance in resisting thrombus formation in hypercoagulable blood. This coating reveals the relationship between PC group density and anticoagulant performance, breaks through the limitations of the theory that hydrophilic groups determine anti-fouling performance, and provides a reference for the design optimization of MPC copolymers in combating high coagulable environments.

Scheme 1 The preparation process of the PMLT copolymer coating with good anti-thrombosis, and antibacterial and anti-protein adhesion properties.

2. Experimental

2.1. Materials

2-Methylacryloxyethyl phosphocholine was purchased from Nanjing Letian Natural Science and Technology Development Research Institute. Methyl acrylate laurate, streptomycin, citric acid, and trisodium citrate were purchased from Shanghai Aladdin Bio-Chemical Technology Co., Ltd. Azobis(isobutyronitrile) and acetone were purchased from Chengdu Kelong Chemical Reagent Co., Ltd. 3-(Trimethoxysilyl) methyl acrylate and FITC-bovine serum albumin were purchased from Sigma Aldrich.

2.2. Synthesis and characterization of the zwitterionic copolymer PMLT

PMLT was synthesized via conventional free-radical copolymerization of the three monomers: MPC, LMA, and TSMA.

Specifically, 1 g MPC, 1.72 g LMA, and 0.75 g TSMA were weighed and added to the reaction vessel. Subsequently, 0.03 g of AIBN was dissolved in a small volume of anhydrous ethanol (deoxygenated by argon) and rapidly transferred to the reaction vessel. The reaction proceeded at 75 °C in the dark for 10 hours. After the reaction was completed, the solution was cooled to room temperature. The solvent was concentrated using rotary evaporation, and the polymer was precipitated with acetone three times. The final product was obtained as a white powder and designated as PMLT.

However, excessive cross-linking groups introduced by TSMA would have an adverse effect on the toughness of the coating. Therefore, the molar fraction of the TSMA monomer was fixed at 7%. Six distinct polymers, PMLT-10, PMLT-41, PMLT-31, PMLT-21, PMLT-11, and PMLT-12, were synthesized by systematically varying only the molar ratio of MPC to LMA. The products were vacuum-dried at room temperature and stored at 4 °C to prevent premature cross-linking.

PMLT was dissolved in methanol-D4 and the polymer was characterized using hydrogen nuclear magnetic resonance spectroscopy (¹H NMR, Bruker, USA). Fourier transform infrared spectroscopy (FTIR, Thermo Fisher, USA) was employed to characterize PMLT in the range of 400–4000 cm⁻¹.

2.3. Preparation and characterization of coatings with different hydrophilic–hydrophobic ratios

2.3.1. Preparation of coatings. PVC substrates were cut into 1 cm × 1 cm squares and cleaned three times with water and ethanol under ultrasonic conditions. Copolymer samples with different ratios were dissolved in ethanol solution (containing 1 wt% water). The substrate was immersed in 10 mg mL⁻¹ copolymer solution, then slowly and uniformly withdrawn after approximately 10–20 seconds, followed by curing at 60 °C for 4 hours. The preparation method for the coating on the inner wall of PVC tubes is the same as above. The coated substrates and tubes were cleaned and dried for subsequent use. The coatings should be stored under dry conditions at room temperature, while the copolymer solutions can be stored at 4 °C with a shelf life of three months.

2.3.2. Characterization of coatings. The cover area of the PMLT coating on the PVC surface was identified by staining with Congo red. Typically, the coated PVC was immersed into a Congo red aqueous solution (10 mg mL⁻¹) for 1 h and then washed with water. The homogeneity of the PMLT coating was observed using an optical microscope. The composition of the coatings was evaluated by measuring the water contact angle (WCA) on the substrate surface using a WCA analyser (Biolin Scientific, SWE). X-ray photoelectron spectroscopy (XPS, Kratos, UK) and FTIR were used to characterize the chemical composition of the coatings. Wide-scan spectra of the samples were recorded in the range of 0–1200 eV, along with narrow-scan spectra for specific elements. The zeta potential changes on the sample surface were characterized using a surface zeta potential meter.

2.3.3. Stability testing. The coated substrate was immersed in four distinct extreme solutions, including NaOH (pH = 12), HCl (pH = 1), 4 M NaCl, and 1% SDS solutions. The sample was gently agitated and immersed at 37 °C for 4 hours. Subsequently, the surfaces were thoroughly rinsed with deionized water. After freeze-drying the samples, their mass was measured using a microbalance. The coating stability was assessed by measuring changes in mass.

2.4. In vitro static whole blood test

The blood compatibility trial received approval from the Sichuan University Medical Ethics Committee (approval no. KS2023396). The adhesion of red blood cells, platelets, and plasma proteins to the materials in a blood environment was evaluated via in vitro whole blood tests. Blood was collected from the ear marginal vein of New Zealand white rabbits. The samples were placed in a 48-well plate, and 500 µL of fresh blood was added to each sample well. After incubation at 37 °C for 1 hour, the surfaces were rinsed three times with saline solution and fixed with 2.5% glutaraldehyde solution overnight. After removing excess glutaraldehyde by rinsing, the samples were subjected to gradient dehydration. Following drying, the adhesion of platelets was observed and photographed using a scanning electron microscope (SEM).

2.5. In vitro protein adsorption test

2.5.1. Determination of albumin adsorption capacity. The albumin adsorption capacity was determined using a Micro-BCA kit (Thermo Fisher Scientific Inc., Waltham). Firstly, the samples in the well plate were soaked overnight in phosphate buffered saline (PBS) solution. An equal volume of 42 mg mL⁻¹ bovine serum albumin (BSA) solution was added dropwise to the surface of the samples, which were incubated at 37 °C for 3 hours and cleaned with PBS three times. Subsequently, the BSA on the sample surface was eluted with 1 mg mL⁻¹ SDS solution for 24 hours. The working solution was prepared by mixing solutions A, B, and C from the Micro-BCA kit, which was mixed with the eluted BSA solution in equal volumes for colour development at 37 °C. A standard curve of BSA solution was prepared using the same method. Finally, a microplate reader was used to measure the absorbance at a wavelength of 562 nm to calculate the albumin adsorption capacity on the surface of different samples.

Fluorescein isothiocyanate-labelled bovine serum albumin (FITC-BSA) was diluted to 1 mg mL⁻¹ with PBS. The above protein solutions were added to the coated PVC substrate and incubated at 37 °C for 1 hour, then washed three times with PBS to remove loosely adherent proteins. A confocal laser scanning microscope (CLSM) was used to observe and photograph the adsorbed proteins on the substrate for qualitative analysis of the proteins adsorbed on the coating surface. The entire experiment was conducted in a dark environment to avoid photobleaching of fluorescence.

2.5.2. Detection of fibrinogen adhesion and denaturation. Enzyme-linked immunosorbent assay (ELISA) is a method that utilizes the specific interaction between antibodies and proteins to determine the protein adsorption on the surface of samples. Human fibrinogen was dissolved in a 42 mg mL⁻¹ BSA solution to a concentration of 3 mg mL⁻¹. The samples were immersed in the fibrinogen (FIB) solution, incubated at 37 °C for 1 hour, and then washed with PBS. Subsequently, an equal volume of 1% BSA blocking solution was added to immerse the sample surface, followed by further incubation at 37 °C for half an hour and washing with PBS. Then, 25 µL of horseradish peroxidase-labelled goat anti-human fibrinogen antibody solution (1/100, bs-1240G-HRP, Bioss) was added and incubated at 37 °C for 1 hour. 120 µL of TMB chromogenic solution was added dropwise to each sample surface. After 10 minutes of chromogenic reaction in the dark, a chromogenic termination solution was added to stop the reaction. Finally, the detection was performed using a microplate reader at a wavelength of 450 nm.

The method for determination of fibrinogen denaturation degree is similar to the above method. Before the chromogenic step in the fibrinogen adsorption assay, 25 µL of HRP-labelled rabbit anti-goat antibody solution (secondary antibody: 1/2000, bs-0294R-HRP, Bioss) was added to cover the sample surface and incubated at 37 °C to bind with the primary antibody. Unbound secondary antibodies were washed away with PBS. Next, the chromogenic solution was added for 10 minutes of chromogenic reaction, followed by the addition of chromogenic termination solution to stop the reaction. Finally, the detection was conducted using a microplate reader at a wavelength of 450 nm.

2.6. Ex vivo thrombogenicity assessment

Making use of arteriovenous shunt models in the New Zealand white rabbit (2.5–3.0 kg), the samples' hemocompatibility was further studied. We prepared coatings on PVC tubes to assess ex vivo thrombogenesis. The coated PVC tubes of the same model were initially connected in series with heparinized catheters, then configured in parallel, and finally integrated with an indwelling needle to construct an extracorporeal circulation loop. This loop was connected to the isolated left jugular vein and right carotid artery of the rabbit to establish a blood flow channel. After 2 hours of blood circulation, the specimens were collected and carefully rinsed with normal saline, followed by cross-sectional analysis to determine the occlusion rate of the tubes. The thrombus inside the tubes was weighed, and the samples were preserved in 2.5% glutaraldehyde overnight at 4 °C, then subjected to gradient dehydration for SEM observation.

2.7. Anti-bacterial adhesion test

Since Staphylococcus aureus is a common microorganism causing infections and diseases associated with medical catheters,³⁷ it was selected in this study to evaluate the anti-bacterial adhesion performance of the coatings. The static antibacterial performance of the materials was assessed by observing the adhesion of bacteria on the material surface after co-cultivation of bacterial solution with the materials. Bacteria were cultured in nutrient broth medium, and the concentration of the bacterial suspension was diluted to 2 × 10⁶ CFU mL⁻¹. 500 µL of the diluted bacterial suspension was added to each well of a 48-well plate, and co-cultivated with sterilized samples at 37 °C for 12 hours. After the co-cultivation, the samples were cleaned, and the bacteria adhering to the sample surface were eluted by repeatedly pipetting the surface and sonicating for 1 minute. 100 µL of the mixed solution was dropwise added onto solid agar medium and immediately spread evenly. Finally, the medium was incubated at 37 °C for 24 hours before counting the colonies.

2.8. Anti-protein adhesion test of the coating surface damage model

The quartz crystal microbalance with dissipation (QCM-D) was used to dynamically monitor the protein adsorption process, aiming to evaluate the anti-fouling effect and reliability of biomimetic cell membrane coatings with different hydrophilic–hydrophobic ratios against plasma proteins.

Bare wafers were placed in the QCM-D module as the control group, with the peristaltic pump speed set to 50 µL min⁻¹ and the temperature maintained at 25 °C. First, PBS (pH = 7.4) was introduced until the baseline stabilized. Subsequently, platelet-poor plasma (PPP) was infused. After reaching adsorption equilibrium, PBS was continuously introduced to wash away loosely adsorbed proteins on the surface until a new equilibrium was achieved.

For the prepared PMLT-10 and PMLT-12 coated wafers, they were placed in the QCM-D module and cleaned by introducing ultrapure (UP) water. 0.1 mM CaCl₂ solution was then infused until adsorption equilibrium was reached, followed by continuous introduction of UP water to remove loosely adsorbed calcium ions. Afterward, PPP was introduced until a new equilibrium was established, and UP water was infused to wash away loosely adherent proteins until the curve stabilized.

The coatings were pre-incubated in PPP for 1 hour, cleaned, and dried before being placed in the module. PPP was introduced to reach adsorption equilibrium, after which PBS was continuously infused to wash away loosely adsorbed proteins on the surface. Data recording was stopped when the curve stabilized, and the protein adsorption amount was finally calculated.

2.9. Evaluation of coating anticoagulant performance under high thrombotic risk conditions

2.9.1. Static whole blood test with calcium ions. Blood was collected from the ear marginal vein of New Zealand white rabbits. CaCl₂ solution was then added to the blood and thoroughly mixed to adjust the calcium ion concentration in the blood to 0.03 M. Subsequent experimental procedures followed those described in Section 2.4.

2.9.2. Dynamic whole blood test with calcium ions. The PVC tube samples were placed in a flow chamber, with both ends of the chamber connected to centrifuge tubes containing 50 mL of fresh whole blood supplemented with calcium ions. The calcium ion concentration in the blood was adjusted to 0.03 M, and a peristaltic pump was used to control the blood flow rate at 3.5 mL s⁻¹ to maintain dynamic circulation. The circulation time of the dynamic whole blood test was 1 hour. After the experiment, normal saline was used to replace the whole blood, and the sample surface was rinsed at the same flow rate (the rinsing solution was directly discharged without recirculation) and the samples were removed. The samples were fixed with 2.5% glutaraldehyde for 12 hours, followed by cleaning, dehydration, drying and observation and analysis using a scanning electron microscope (SEM).

2.9.3. Static whole blood test with lipopolysaccharide. A certain concentration of lipopolysaccharide (LPS) was added to whole blood to construct an in vitro high thrombotic state, adjusting the LPS concentration in the blood to 2 µg mL⁻¹. Subsequent experimental procedures followed those described in Section 2.4.

2.9.4. Establishment and evaluation of a type 2 diabetes mellitus animal model.
2.9.4.1. Establishment of a type 2 diabetes mellitus animal model. In this study, a combination of high-sugar and high-fat diet and a small amount of streptozotocin (STZ) was used to induce an animal model similar to human type 2 diabetes mellitus (T2DM), aiming to evaluate the anti-fouling effect of the coating under the hyperinflammatory blood state in diabetic patients. All animal experimental procedures were conducted in accordance with the regulations of the Committee for the Guidelines for the Care and Use of Laboratory Animals in China.

New Zealand white rabbits with a body weight range of 2.0–2.5 kg (regardless of gender) were selected (provided by the Laboratory Animal Center of Sichuan University). All experimental rabbits were adaptively fed for one week and then randomly divided into two groups. The control group was fed with a normal diet, while the experimental group was fed exclusively with a high-sugar and high-fat diet (37% sucrose, 10% lard, and 53% basic feed). The rabbits were given ad libitum access to sufficient feed and drinking water daily, and were raised for 4 weeks. After 4 weeks, a fresh streptozotocin (STZ) solution at a dose of 40 mg kg⁻¹ was prepared (2% STZ solution was prepared using citric acid-sodium citrate buffer with pH 4.2, freshly prepared and used immediately, and injected within 30 minutes under ice bath conditions) and rapidly injected via the ear marginal vein. The rabbits were fasted for 12 hours before injection but had free access to water. Injections were administered consecutively for 5 days with a 24-hour interval between each injection, followed by continuous feeding for 1 week. During the feeding and modelling process, the food intake, mental state, body shape, and hair of rabbits in each group were observed. Body weight and fasting blood glucose were measured once a week. The criterion for successful modelling was defined as two consecutive fasting blood glucose (FBG) levels ≥8.0 mmol L⁻¹ with sustained hyperglycemia for 1 week.

Diabetes is often accompanied by a hypercoagulable state and secondary thrombosis.³⁸ Therefore, this study tested the activated partial thromboplastin time (APTT) and prothrombin time (PT) to verify the hypercoagulable state of the diabetic rabbit model. Fresh whole blood was centrifuged at 3000 rpm for 15 minutes to obtain platelet-poor plasma (PPP), which was stored at −80 °C. Evaluation was performed using an automatic coagulation analyser.

2.9.4.2. Ex vivo thrombogenicity assessment. The anticoagulant performance of PMLT-10, PMLT-21, and PMLT-12 samples was analysed using the Ex-vivo thrombogenicity assessment of successfully modelled New Zealand white rabbits. For specific experimental operations, the readers are referred to Section 2.6.

2.10. Statistical analysis

Each experiment in this work was replicated no less than three times in parallel, and the results were provided as mean ± standard deviation. Significant differences between groups under the same test were determined using ANOVA. The threshold for statistically significant was established as a probability value (p) below 0.05. For example, the values p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 would be represented by the symbols *, **, ***, and ****.

3. Results and discussion

3.1. Characterization of the zwitterionic copolymer coating PMLT

In this study, the zwitterionic polymer PMLT containing TSMA was synthesized via conventional free radical polymerization, with its synthetic route illustrated in Fig. 1a. MPC, a common hydrophilic group biomimetic monomer, exhibits excellent biocompatibility and antifouling properties. LMA has long-chain alkyl groups, imparting greater flexibility to the polymer. Furthermore, the crosslink monomer TSMA was introduced to promote the formation of a networked crosslinked structure among polymer chains. This enhances the robust bonding stability between the coating and the substrate, thereby improving the stability of the polymeric coating. By adjusting the monomer ratio of MPC to LMA, we investigated the antifouling performance of coatings with varying hydrophilic and hydrophobic ratios. The findings reported by Iwasaki³⁹ demonstrated that a molar fraction of the MPC monomer in the copolymer exceeding 0.3 significantly enhances the polymer biocompatibility. Consequently, copolymers with MPC content below this threshold were not further investigated for the surface modification of PVC materials. The ¹H NMR spectrum of the PMLT polymer is shown in Fig. S1. Characteristic peaks corresponding to MPC, LMA, and TSMA components are observed at 3.25 ppm, 1.40 ppm, and 0.65 ppm, respectively. It confirmed the successful synthesis of the PMLT polymer. Furthermore, as demonstrated in Table S1, the composition of all copolymers closely matches the monomer feed ratios used in synthesis, indicating that the polymerization proceeded as intended. The FTIR spectra of the copolymers (Fig. S2) reveal similar absorption profiles across all samples. Specifically, the PMLT polymer exhibits a C=O stretching vibration at 1725 cm⁻¹, along with asymmetric and symmetric O–P=O stretching vibrations at 1240 cm⁻¹ and 1060 cm⁻¹, respectively. Collectively, the ¹H NMR and FTIR results provide consistent evidence for the successful synthesis of the target polymer.

Fig. 1 (a) Synthesis route of PMLT. (b) XPS results of PMLT coatings. (c) The elemental content of PMLT. (d) XPS spectra of N 1s in the samples. (e) Static and dynamic water contact angles of PMLT coatings. (f) Zeta potential of PMLT coatings. (g) Optical images and microphotograph of the sectional view of PVC: uncoated PVC and PMLT-12 coating (the coating was stained by Congo red).

The coating was primarily prepared using a simple dip-coating process. The substrate was immersed in the polymer solution and subsequently withdrawn at a constant rate, then heated to induce the hydrolysis and crosslinking of the silane coupling agent and facilitating the self-crosslinking and curing of the coating. Concurrently, robust chemical bonds formed between the silane coupling agent and the substrate, ultimately yielding a stable zwitterionic copolymer coating. The stable adhesion of the coating to the substrate is primarily attributed to the hydrolysis of the trimethoxy silyl groups within the polymer chains, generating reactive silanol groups. These silanol groups then undergo condensation reactions, liberating water molecules and forming Si–O–Si crosslinks. Ultimately, under heating conditions, these crosslinking points transform the linear polymer into a network structure, thereby enhancing the stability of the coating (Fig. S3). To determine the chemical composition of the coating, X-ray photoelectron spectroscopy (XPS) was employed for quantitative analysis of the elemental content on the coating surface. As shown in Fig. 1b–d, the XPS results revealed characteristic signals of P, N, and Si elements on the PVC substrate modified by the PMLT polymer. In contrast, these signals were absent on the uncoated PVC substrate. The presence of N and P indicates the existence of the PC groups within the coating, while Si originates from the TSMA monomer. Furthermore, the N elemental content increased proportionally with a higher MPC monomer loading (Fig. 1d). These results collectively confirm the successful fabrication of the PMLT coating.

Surface hydrophilicity and hydrophobicity significantly influence the antifouling performance of biomaterials, including anticoagulant and antibacterial properties. The water contact angle (WCA) was employed to evaluate the surface wettability of the materials, with the results summarized in Fig. 1e. The static water contact angle measurements can be seen that all PMLT-coated materials exhibited contact angles greater than 90°, indicating hydrophobic surface characteristics. This hydrophobicity is likely due to the dominant role played by the hydrophobic alkyl chains within the zwitterionic polymer and the inherent hydrophobicity of the silane coupling agent under static conditions. Dynamic contact angle measurements demonstrated that all PMLT polymer coatings exhibited contact angle hysteresis, signifying favourable surface wettability. This behaviour is ascribed to the high surface energy and hydrophilicity of the PC groups, which promotes a reorientation phenomenon of these groups under continuous water contact. Furthermore, the receding contact angles on the PMLT coating surfaces were nearly identical, and no dependence on the MPC unit composition was observed. Previous studies⁴⁰ have indicated that the advancing contact angle is primarily governed by the lower surface free energy component (the LMA units in the PMLT polymer), while the receding contact angle depends on the higher surface free energy component (the MPC units in the PMLT polymer). In conjunction with this view, it was shown that the PMLT surface may have unique surface properties and that its wettability would be beneficial for its antithrombotic and antibacterial properties. To better elucidate the surface properties of the coatings under hydrated conditions, the zeta potential of the coatings was characterized using a solid surface zeta potential analyser in 0.1 mM KCl solution at pH 7.4. As shown in Fig. 1f, the PMLT-10 coating, possessing the highest phosphorylcholine (PC) group content, exhibited a negative zeta potential of −35.2 mV. This negative value is attributed to the influence of PC groups on the polarization of water molecules during the measurement. The distinct contributions of hydrophilic and hydrophobic groups to the surface energy induce dynamic motion of the coating functional groups upon contact with the KCl solution. Consequently, the surface properties of the coating undergo significant change due to interaction with water. Furthermore, as the MPC composition within the PMLT polymer increased, the zeta potential of the PMLT surface approached zero. This observation suggests that the substrate surface was likely completely covered by MPC units.

Homogeneity, coverage, and stability were essential parameters for surface coating. Therefore, to demonstrate the stability and coverage of the coatings, we selected PMLT-12 as the representative coating. As shown in Fig. 1g, after staining with Congo red, the PMLT-12 coating homogenously covered the inside of the tube surface, and a thin PMLT-12 layer was clearly observed under an optical microscope. To verify the stability of the coating, we immersed the coated samples in extreme solutions and calculated the coating loss rate (Fig. S4). Under all conditions, the mass of the PMLT-12 coating remained largely unchanged, confirming the formation of a highly cross-linked and robust network structure on the substrate. This exceptional stability is directly attributed to the incorporation of the TSMA monomer. During the thermal curing stage, the methoxysilane group (–Si(OCH₃)₃) undergoes hydrolysis to form silanol (–Si–OH), followed by a condensation reaction. This reaction forms stable crosslinks between polymer networks and covalent bonds with –OH groups on the substrate surface, enhancing coating stability.

3.2. Blood compatibility of the zwitterionic copolymer coating PMLT

3.2.1. In vitro static whole blood test. When materials interface with the blood environment, interactions extend beyond single components like platelets and plasma proteins, encompassing other influential factors. To preliminarily assess the hemocompatibility of the samples, in vitro whole blood evaluation was conducted. As shown in Fig. 2a and b, the PVC surface exhibited substantial adhesion of red blood cells (RBCs) and plasma proteins, accompanied by RBC's aggregation and morphological alterations. These observations indicate poor hemocompatibility of the PVC surface. Furthermore, the PMLT-10 surface demonstrated lower levels of adherent plasma proteins and RBCs compared to the PVC control group, but it was less resistant to erythrocyte deposition compared to the other experimental groups, while a small number of erythrocytes had a distorted morphology. In contrast, all other groups exhibited excellent resistance to RBC deposition, with fewer erythrocytes and plasma proteins adhering to the surface and the erythrocyte morphology of biconcave discs, which preliminarily proved that the PMLT polymer coatings had good hemocompatibility. Therefore, considering that the differences between groups were not significant, the PMLT-10, PMLT-21, and PMLT-12 coatings were selected for subsequent experimental evaluations.

Fig. 2 (a) In vitro whole blood evaluation results. (b) Statistical analysis of red blood cell deposition quantity on sample surfaces. (c) Fibrinogen adsorption statistics. (d) Results of fibrinogen denaturation.

3.2.2. In vitro protein absorption. Research indicates that plasma protein adsorption constitutes the initial event triggering platelet adhesion/activation and subsequent coagulation cascade activation.⁴¹ Consequently, this study evaluated the protein adsorption capacity of different coating surfaces using bovine serum albumin (BSA) and fibrinogen (FIB). The adsorption results of BSA protein are shown in Fig. S5; the amount of BSA adsorbed onto all PMLT-coated surfaces was lower than that on the PVC surface. This demonstrates that the incorporation of zwitterionic moieties effectively enhances the resistance to protein adsorption. Among them, the PMLT-12 coatings exhibited significantly reduced BSA adsorption. For the PMLT-12 coating, due to its structural similarity to cell membranes and the inversion phenomenon of hydrophilic groups, the proteins adhered to its surface are not firmly bound, resulting in a lower adhesion amount after the coating is cleaned.⁴²

As evidenced by the results in Fig. 2c, the FIB adsorption levels on the PMLT-coated surfaces were significantly lower than those on the PVC surface. Although there was no significant difference between PMLT-21 and PMLT-12, the amount of FIB adhered to the surface was less than that of the PMLT-10 coating in both cases. In terms of the denaturation of FIB on the material surface (Fig. 2d), the activation degree of fibrinogen on the surfaces of PMLT-10, PMLT-21, and PMLT-12 was significantly lower than that on PVC, with no obvious differences among the three coatings. Fluorescein isothiocyanate-labelled bovine serum albumin (FITC-BSA) was used for qualitative evaluation of non-specific protein adsorption on different sample surfaces. As shown in Fig. S6, the fluorescence intensity on the coating surfaces observed under a CLSM was consistent with the quantitative results determined in Fig. S5. From the above results, it can be concluded that the introduction of zwitterions endows the material with excellent anti-protein adhesion performance. However, simply increasing the content of hydrophilic groups does not yield the optimal anti-protein ability. The PMLT coating, in which hydrophilic and hydrophobic groups reach a state of balanced interaction, exhibits more stable anti-fouling effects.

In conclusion, all the above hemocompatibility experiments have verified that the PMLT copolymer coating exhibits excellent antithrombotic performance, confirming that the introduction of PC groups can endow the material with superior anti-protein adhesion properties. For the PMLT-10 polymer coating, its anti-protein adsorption and antithrombotic properties are slightly reduced. This is because the core of the cell membrane is highly hydrophobic, and the favourable hydrophobic free energy plays a decisive role in maintaining the stability of the cell membrane.⁴³ Therefore, although PMLT-10 has the highest content of PC groups, the instability of the coating leads to a decrease in the anti-fouling effect. In contrast, the PMLT-12 coating, which has a ratio close to that of natural cell membranes, shows stable anti-fouling performance, preliminarily demonstrating the influence of hydrophobic–hydrophilic balance on anticoagulant efficacy.

3.3. The anti-bacterial adhesion properties of PMLT

In addition, Staphylococcus aureus was selected as a representative strain in this study to evaluate the antibacterial ability of PVC, PMLT-10, PMLT-21, and PMLT-12 samples. As shown in Fig. 3a, the results of colony counting on Staphylococcus aureus agar culture revealed that the colony density in the solid medium corresponding to the surfaces of the PMLT-10, PMLT-21 and PMLT-12 samples was significantly lower than in the PVC control group. Statistical analysis via colony counting (Fig. 3b) revealed that the surfaces of samples modified with PMLT exhibited significant antibacterial effects against Staphylococcus aureus. The antibacterial rates of PMLT-10, PMLT-21, and PMLT-12 samples were 81%, 92%, and 94%, respectively. These three groups showed significant differences compared with the PVC control sample, demonstrating good antibacterial effects.

Fig. 3 (a) Culture photos of Staphylococcus aureus on the surfaces of PVC, PMLT-10, PMLT-21, and PMLT-12. (b) Antibacterial rate results (with PVC as the control group).

3.4. Ex vivo thrombogenicity assessment

To explore the anticoagulant effect of the coating in the presence of physiological blood circulation circumstances, the antithrombotic properties of PMLT-10, PMLT-21 and PMLT-12 coatings were evaluated through semi-in vivo blood circulation experiments in practical applications. The PVC sample tubes were connected to the circulatory line, and the rabbit carotid artery and jugular vein were connected to construct an extracorporeal circulation circuit (Fig. 4a). As shown in Fig. 4a and b, the samples formed visible thrombus in the PVC tubes after 0.5 h or 1 h of exposure to blood, and the rate of tube blockage was 52.7% after 1 h of circulation, but no visible thrombus and tube blockage were observed in the PMLT-10, PMLT-21, and PMLT-12 PMLT-coated PVC tubes. Statistical data showed that the mass of thrombus adhering to the surface of the uncoated PVC sample tubes after half an hour of circulation was much higher than that of the PMLT copolymer-coated samples (Fig. 4c). SEM observations (Fig. 4d) showed that after 0.5 hours, a large number of activated red blood cells and fibrin networks appeared on the inner surface of the PVC sample tubes, but no significant platelets or fibrin networks were found in any of the three groups of PMLT-coated samples. The results of the 1 h experiments showed significant thrombus formation on the surface of the PVC samples, and only a small amount of the surface of the PVC tubes coated by PMLT adhesion of platelets no fibrin network was observed. These results indicate that PMLT-modified surfaces can effectively inhibit thrombus formation in blood under normal physiological environments through the hydration and steric repulsion effects of zwitterionic polymers. However, the impact of the hydrophobic–hydrophilic group ratio on antithrombotic efficacy requires further investigation.

Fig. 4 After 0.5-hour and 1-hour ex vivo thrombogenicity assessment. (a) Photographs of sample tubes. (b) Thrombus weight on the surface of sample tubes. (c) Occlusion rate of sample tubes. (d) SEM images of the inner surface of sample tubes.

3.5. Evaluation of anti-fouling performance under in vitro high coagulation risk conditions

Since coatings with various ratios all exhibit good resistance to red blood cell adhesion under normal blood conditions, this study aims to evaluate the differences in anticoagulant properties among coatings with different ratios under more stringent blood conditions. Calcium ions participate in both the intrinsic and extrinsic coagulation pathways throughout the entire coagulation process, playing a crucial role in blood coagulation.⁴⁴ Therefore, this study established a high coagulation risk blood state by adding calcium ions to whole blood. In the static high coagulation whole blood experiment, as shown in Fig. 5a, due to the hypercoagulable state of blood, obvious thrombus was observed on the surface of the PVC control sample. For the PMLT-21 coating, the number of adherent red blood cells was relatively small, but plasma protein adsorption and fibrin network formation were detected. The PMLT-12 coating had fewer adherent red blood cells, and the red blood cells did not exhibit severe morphological deformation. Meanwhile, to further simulate the in vivo blood environment, this study performed a dynamic high-coagulation whole blood experiment using a peristaltic pump to mimic blood flow (Fig. 5b). The results were consistent with those of the static high-coagulation whole blood experiment. A large number of cross-linked fibrins, activated platelets, and deformed red blood cells were observed on the surfaces of PVC, PMLT-10, and PMLT-21. In contrast, only a small amount of red blood cells and plasma proteins were adsorbed on the surface of the PMLT-12 sample, demonstrating that PMLT-12 has relatively better blood compatibility and anticoagulant effect.

Fig. 5 Evaluation results in vitro (a) static hypercoagulability with calcium ions, (b) dynamic hypercoagulability with calcium ions and (c) static hypercoagulability with lipopolysaccharide.

Various inflammatory diseases are accompanied by hypercoagulability, and it has been demonstrated that many signalling pathways are involved.⁴⁵ Lipopolysaccharide (LPS), also known as endotoxin, is a glycolipid composed of lipids and polysaccharides.⁴⁶ It has been proven to induce inflammation, and LPS can also directly affect fibrinogen and blood coagulation.⁴⁷ Therefore, the hypercoagulability risk induced by the pro-inflammatory properties of LPS was further utilized to verify the blood compatibility of PMLT coatings. As shown in Fig. 5c, compared with PVC, the number of red blood cells adhering to the surfaces of PMLT-10 and PMLT-21 was significantly reduced, but severe deformation and massive aggregation occurred. In contrast, the PMLT-12 coating showed fewer adherent red blood cells with relatively intact morphology. It can be seen that PMLT-12 exhibits good anti-thrombotic performance under hyperinflammatory conditions. This is of great significance for research on anticoagulant modification of dialysis tubing or dialyzers used by diabetic patients with hemodialysis who are in a systemic hyperinflammatory state.

3.6. Establishment of the diabetic animal model and ex vivo thrombogenicity assessment

In Section 3.5, the PMLT-12 sample exhibited superior anticoagulant efficacy compared to other groups. To further prove the excellent antithrombotic performance of the coating, a diabetic rabbit model was constructed and ex vivo thrombogenicity assessment was performed, which is of great significance for research on complications such as catheter-related thrombosis in diabetic patients undergoing hemodialysis treatment. Diabetic rabbits were modelled as shown in Fig. 6a. The model rabbits exhibited symptoms such as polyphagia, polydipsia, and polyuria, with a decreasing trend in body weight and a significant increase in blood glucose levels. After administration, their fasting blood glucose levels were ≥8.0 mmol L⁻¹. To observe the differences in the impact of blood from diabetic model rabbits and normal rabbits on the coagulation system, activated partial thromboplastin time (APTT) and prothrombin time (PT) were further tested. APTT is clinically used to evaluate the anticoagulant capacity of the intrinsic coagulation mechanism,⁴⁸ while PT is used to detect the activation of coagulation factors.⁴⁹ As shown in Fig. 6b, the APTT and PT of blood samples from normal rabbits (control group) were 27.8 s and 9.5 s, respectively, whereas those of the experimental group rabbits were shortened compared to the control group. These results indicate that both intrinsic and extrinsic coagulation pathways in the blood of experimental rabbits were affected, presenting a hypercoagulable state in vivo. This may be attributed to polysaccharide molecules, which can induce the production of various pro-inflammatory cytokines and directly act on coagulation pathways through the upregulation of tissue factors, thereby triggering hypercoagulability, promoting early activation of intrinsic and extrinsic coagulation cascades, and facilitating the activation of prothrombin to form thrombin.^50,51 The aforementioned increases in blood glucose levels and changes in coagulation time indicate that the islet β-cell function of the model rabbits was impaired, and their blood presented a hypercoagulable state, which is consistent with the clinical characteristics of T2DM, confirming the successful construction of the diabetic rabbit model.

Fig. 6 (a) The establishment process of the diabetic animal model. (b) APTT and PT detection results of the control group and diabetic model rabbits. After 0.5-hour and 1.5-hour ex vivo thrombogenicity assessment in diabetic model rabbits. (c) Visual photographs of sample tubes. (d) Thrombus mass. (e) Occlusion rate. (f) SEM images of thrombus adhesion.

After successful modeling of diabetic rabbits, ex vivo thrombogenicity assessment was conducted to further verify the antithrombotic performance of the coating. As shown in Fig. 6c, when the samples were exposed to hypercoagulable blood, the PVC sample tube exhibited extensive occlusion, while PMLT-10, PMLT-21, and PMLT-12 remained relatively clean. The statistical data in Fig. 6d and e showed that the thrombus mass and occlusion rate on the surface of PVC samples were significantly higher than those of PMLT-coated samples. As shown in Fig. 6f, after the samples were exposed to blood for 30 minutes, massive thrombus formation was observed on the PVC surface. In contrast, a small amount of red blood cells and platelets adhered to the surfaces of PMLT-10 and PMLT-21 samples. Meanwhile the red blood cells showed aggregation and deformation, and most platelets presented a fully spread activated state. Only a small amount of platelet adhesion was observed on the surface of PMLT-12 samples, and most platelets were round and in an inactivated state. After 1.5 hours, extensive thrombus formation occurred on the surfaces of PVC, PMLT-10, and PMLT-21, while only partial platelet adhesion was observed on the surface of PMLT-12. These results confirm that the PMLT polymer coating has excellent antithrombotic ability. Among them, PMLT-12 exhibited a relatively stable anti-fouling effect in harsh environments and hypercoagulable risk blood environments, further challenging the traditional understanding that anticoagulation is dominated by hydrophilic groups in zwitterions. This proves that the balanced interaction between hydrophobic and hydrophilic groups has a significant impact on the anticoagulant effect of the coating.

3.7. Anti-protein adhesion performance of the coating surface damage model

To further explore the mechanism by which the balance between hydrophobic and hydrophilic groups affects anti-protein adhesion performance, this study pre-treated PMLT coatings with calcium ions and PPP to construct a coating damage model, simulating a real and effective contaminated environment. Subsequently, the quartz crystal microbalance with dissipation (QCM-D) was used to dynamically monitor the protein adsorption process, aiming to further evaluate the intrinsic reasons behind the plasma protein resistance of PMLT coatings with different hydrophilic–hydrophobic group ratios.

As can be seen from Fig. 7b and c, in the coating damage model pre-incubated with PPP for 1 hour, when the adsorption mass of plasma proteins on the coating surface reached saturation, the plasma protein adsorption mass on the surface of the PMLT-12 coating was higher than that on the PMLT-10 coating. However, during rinsing with PBS, a large amount of protein on the surface of the PMLT-12 coating was washed away, resulting in a final surface protein adhesion amount of only 22.3 ng cm⁻². In contrast, the final plasma protein adhesion amount of the PMLT-10 coating was higher than that of the PMLT-12 coating. This is because proteins tend to easily adhere to hydrophobic surfaces. The higher proportion of hydrophobic groups in the PMLT-12 coating leads to more protein adhesion in the early stage of contact with plasma proteins. However, due to the biomimetic cell membrane structure formed by the ratio of hydrophobic to hydrophilic groups in PMLT-12 coating, it exhibits excellent anti-fouling performance. The conformation of non-specifically adsorbed proteins does not change on the surface of PMLT-12, and they are rapidly desorbed, ultimately resulting in a lower adhesion amount. Fig. 7d and e show the effect of the calcium ion damage model on plasma protein adhesion. It can be seen that after the coatings were continuously exposed to plasma proteins, the protein adsorption on the surface of the PMLT-10 coating was higher than that of the control group. This is because calcium ions can cause damage to the PC groups. In contrast, the hydrophobic groups contained in the PMLT-12 coating enhance the stability of the coating, endowing it with reliability. From the results of anti-protein adhesion in the coating surface damage model (Fig. 7f), it can be concluded that the PMLT-12 coating exhibits excellent anti-fouling performance. The content of PC groups must reach a balance with hydrophobic groups to enhance the stability of the impedance layer, thereby achieving better anti-fouling effects. This confirms the dominant role of the balanced interaction between hydrophobic and hydrophilic groups in governing the anticoagulant effect.

Fig. 7 QCM-D characterization of the PPP adsorption amount on material surfaces. (a) Protein adsorption on the surface of bare quartz crystals. After 1-hour of PPP incubation, the protein adsorption on (b) PMLT-12 coating and (c) PMLT-10 coating. In calcium ion-induced coating damage model, the protein adsorption amount on (d) the surface of PMLT-10 coating and (e) the surface of PMLT-12 coating. (f) Statistics of anti-protein adhesion.

4. Conclusions

This study challenges the traditional perception that the antifouling performance of zwitterionic coatings relies on a high content of hydrophilic groups. By designing and synthesizing a MPC-LMA-TSMA terpolymer (PMLT), it explores the synergistic regulatory mechanism of the ratio between zwitterions containing phosphorylcholine groups and long hydrophobic chains in both normal blood environments and hypercoagulable risk blood environments. Under normal physiological conditions, all PMLT coatings exhibit excellent anti-protein adsorption and antithrombotic effects. However, under pathological hypercoagulable conditions, simply increasing the density of PC groups accelerates the disintegration of the hydration layer, leading to reduced antithrombotic efficacy. In contrast, the PMLT-12 coating, which forms a stable biomimetic cell membrane-like bilayer structure through optimized hydrophilic–hydrophobic group ratios, maintains ultra-low protein adsorption and thrombus resistance even under the attack of calcium ions and fibrinogen. Semi-in vivo circulation experiments using hypercoagulable risk blood models further confirmed that the thrombus inhibition rate of the PMLT-12 coating in high shear stress blood perfusion environments is significantly higher than that of traditional coatings with high PC group content. This demonstrates that the hydrophobic–hydrophilic balance plays a decisive role in maintaining hypercoagulable hemocompatibility on material surfaces. In summary, the copolymer coating designed and constructed in this study significantly enhances the antifouling performance of blood-contacting materials through the synergistic effect of dual mechanisms “hydrophilic–hydrophobic biomimetic barrier and covalent cross-linking stabilization”. This not only breaks the oversimplified theory that hydrophilicity determines antifouling properties but also provides a paradigm for molecular design in developing anticoagulant coatings adaptable to pathological conditions. This offers a simple and effective modification strategy for improving the biocompatibility of blood-contacting materials and devices.

Author contributions

Yanan Wang: writing – original draft, visualization, methodology, formal analysis, data curation, and conceptualization. Shiyu Yao: writing – original draft, visualization, methodology, formal analysis, data curation, and conceptualization. Zian Wang: visualization, software, resources, and formal analysis. Lietao Wang: resources and formal analysis. Hui Yan: resources and formal analysis. Fanjun Zhang: resources and formal analysis. Jin Wang: resources and formal analysis. Lu Zhang: writing – review & editing, supervision, resources, and formal analysis. Zhongwei Zhang: resources and formal analysis. Rifang Luo: writing – review & editing, supervision, resources, funding acquisition, and formal analysis. Yunbing Wang: writing – review & editing, resources, and funding acquisition.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The datasets supporting the findings of this study are fully available within the article and its supplementary information (SI). Supplementary information: ¹H NMR spectrum of the PMLT copolymer (Fig. S1). The PMLT polymer composition was calculated from ¹H NMR (Table. S1). The FT-IR spectrum of the PMLT copolymer (Fig. S2). The cross-linking mechanism of TSMA (Fig. S3). PMLT-12 sample residual rate under immersion in different solutions (Fig. S4). Statistical analysis of BSA on sample surfaces (Fig. S5). Fluorescence images of FITC-BSA adsorption on different sample surfaces under a CLSM (Fig. S6). See DOI: https://doi.org/10.1039/d5tb01941a.

Raw data for the characterization of materials, blood compatibility tests and anti-bacterial test are available without restriction from the corresponding authors upon request.

Acknowledgements

This work was supported by the National Key Research and Development Program [2024YFC2419004], the National Natural Science Foundation of China [32271398 and 32371401], the Sichuan Science and Technology Program [2022YFH0092], and the Youth Science and Technology Academic Leaders Cultivation Program of Sichuan University [0082604151410].

References

1. I. H. Jaffer and J. I. Weitz, Acta Biomater., 2019, 94, 2–10.
2. Q. B. Peng, S. Wang, J. G. Han, C. Y. Huang, H. Y. Yu, D. Li, M. Qiu, S. Cheng, C. Wu, M. X. Cai, S. X. Fu, B. H. Chen, X. Y. Wu, S. W. Du and T. T. Xu, Research, 2024, 7, 0339.
3. A. C. de Oliveira, L. Y. C. Madruga, P. Chevallier, F. Copes, D. Mantovani, B. H. Vilsinski, K. C. Popat, M. J. Kipper, P. R. Souza and A. F. Martins, Prog. Org. Coat., 2024, 194, 108629.
4. C. Sun, Y. Zhang, F. Dong, J. Zhao, P. Zhang, S. Li, Y. Gao, Y. Wang and G. Gao, Chem. Eng. J., 2024, 488, 150944.
5. D. U. Lee, M. Kayumov, J. Park, S. K. Park, Y. Kang, Y. Ahn, W. Kim, S. H. Yoo, J.-K. Park, B.-G. Kim, Y. S. Oh, I.-S. Jeong and D. Y. Choi, Bioact. Mater., 2024, 34, 112–124.
6. S. Li, P. Yu, F. Dong, J. Zhang, T. Wang, P. Zhang, G. Gao, Y. Wang and Y. Gao, J. Mater. Chem. A, 2025, 13, 6687–6696.
7. Y. Zhang, L. Zhang, S. Duan, Y. Hu, X. Ding, Y. Zhang, Y. Li, Y. Wu, X. Ding and F.-J. Xu, J. Mater. Chem. B, 2022, 10, 1795–1804.
8. M. H. Li, X. L. Sun, L. Q. Zeng, A. J. Sun and J. B. Ge, Research, 2025, 8, 0679.
9. X. X. Huang, J. P. Huang, P. C. Su and W. B. Li, Research, 2023, 6, 1–10.
10. Y. Zou, Z. J. Shan, Z. P. Han, J. T. Yang, Y. X. Lin, Z. H. Gong, L. Xie, J. Y. Xu, R. L. Xie, Z. F. Chen and Z. T. Chen, Research, 2023, 6, 0225.
11. W. L. Tang, N. G. Fischer, X. Z. Kong, T. Sang and Z. Ye, BMEMat, 2024, 2, e12105.
12. M. Badv, F. Bayat, J. I. Weitz and T. F. Didar, Biomaterials, 2020, 258, 120291.
13. M. Yao, Z. Wei, J. Li, Z. Guo, Z. Yan, X. Sun, Q. Yu, X. Wu, C. Yu, F. Yao, S. Feng, H. Zhang and J. Li, Nat. Commun., 2022, 13, 5339.
14. L. H. Li, Z. J. Cao, C. L. Zhang, L. Li, Q. Y. Li, C. Liu, C. Qu, R. F. Luo, P. Fu and Y. B. Wang, Adv. Funct. Mater., 2024, 34, 2312243.
15. L. H. Li, C. L. Zhang, Z. J. Cao, L. Ma, C. Liu, X. R. Lan, C. Qu, P. Fu, R. F. Luo and Y. B. Wang, Biomaterials, 2024, 305, 122423.
16. K. Li, H. Tang, J. Peng, S. Gao, Z. Du, G. Chen, D. Wu and G. Liu, Adv. Funct. Mater., 2024, 34, 2307760.
17. M. Y. Li, K. P. Liu, W. Q. Liu, N. Y. Chen, Y. A. Wang, F. J. Zhang, Q. F. Luo, L. Yang, R. F. Luo and Y. B. Wang, Appl. Surf. Sci., 2022, 600, 154143.
18. T. Zhang, T. Liang, Q. Pan, S. Zhang, S. Zhang, Z. Geng and B. Zhu, Adv. Sci., 2025, 12, 2502411.
19. H. Guo, C. Wen, S. Tian, X. Zhang, Y. Ma, X. Liu, J. Yang and L. Zhang, ACS Appl. Mater. Interfaces, 2021, 13, 21051–21059.
20. N. Erathodiyil, H.-M. Chan, H. Wu and J. Y. Ying, Mater. Today, 2020, 38, 84–98.
21. Q. Li, C. Wen, J. Yang, X. Zhou, Y. Zhu, J. Zheng, G. Cheng, J. Bai, T. Xu, J. Ji, S. Jiang, L. Zhang and P. Zhang, Chem. Rev., 2022, 122, 17073–17154.
22. Z. Zhang, L. Wang, J. Liu, H. Yu, X. Zhang, J. Yin, S. Luan and H. Shi, Small, 2023, 19, 2304379.
23. Z. Ma, J. Sun, X. Dong, D. Gan, W. Peng, Y. Li, W. Qian, P. Liu and J. Shen, J. Mater. Chem. B, 2022, 10, 3687–3695.
24. Z. Jing, T. Xuying, W. Minmin, F. Jie and Z. Qinggang, J. Mater. Chem. B, 2025, 13, 5948–5956.
25. K. P. Liu, M. Y. Li, F. J. Zhang, Y. N. Wang, C. Chen, Y. Wei, L. Yang, R. F. Luo and Y. B. Wang, Chem. Eng. J., 2021, 426, 131803.
26. S. Shi, M. Hu, X. Peng, C. Cheng, S. Feng, X. Pu and X. Yu, J. Mater. Chem. B, 2024, 12, 3417–3435.
27. J. Liu, J. Wang, Y.-F. Xue, T.-T. Chen, D.-N. Huang, Y.-X. Wang, K.-F. Ren, Y.-B. Wang, G.-S. Fu and J. Ji, J. Mater. Chem. B, 2020, 8, 5361–5368.
28. F. J. Zhang, L. Yang, C. Hu, L. H. Li, J. Y. Wang, R. F. Luo and Y. B. Wang, J. Mater. Chem. B, 2020, 8, 8433–8443.
29. Q. Ma, W. T. Zhang, X. H. Mou, N. Huang, H. M. Wang, H. Y. Zhang and Z. L. Yang, Research, 2024, 7, 0423.
30. W. C. Xiong, J. W. Chai, J. N. Wu, J. L. Li, W. C. Lu, M. L. Tian, M. A. Jmel, J. H. Ippel, M. Kotsyfakis, I. Dijkgraaf, S. W. Liu and X. Q. Xu, Research, 2024, 7, 0381.
31. U. Ahmad, W. N. W. Hanaffi, A. Islam, A. Salman, M. M. Khan, F. Shakeel, Q. Cai, X. J. Cai and Q. T. H. Shubhra, BMEMat, 2025, 70033.
32. A. Aklilu, M. S.-L. Lai, Z. Jiang, S. P. Yip and C.-L. Huang, Int. J. Mol. Sci., 2025, 26, 6114.
33. L. Hu, L. Chang, Y. Zhang, L. Zhai, S. Zhang, Z. Qi, H. Yan, Y. Yan, X. Luo, S. Zhang, Y. Wang, S. P. Kunapuli, H. Ye and Z. Ding, Circulation, 2017, 136, 817–U102.
34. A. L. Lewis, Z. L. Cumming, H. H. Goreish, L. C. Kirkwood, L. A. Tolhurst and P. W. Stratford, Biomaterials, 2001, 22, 99–111.
35. H. Yan, L. Wang, H. Wu, Y. An, Y. Qin, Z. Xiang, H. Wan, Y. Tan, L. Yang, F. Zhang, Q. Jiang, R. Luo and Y. Wang, Chem. Eng. J., 2024, 490.
36. H. Yan, Y. Qin, Y. Li, Y. An, H. Wu, C. Chen, L. Yang, F. Zhang, R. Luo, Q. Jiang and Y. Wang, Composites, Part B, 2023, 267, 111055.
37. T. Singh, A. L. Hook, J. Luckett, M. F. Maitz, C. Sperling, C. Werner, M. C. Davies, D. J. Irvine, P. Williams and M. R. Alexander, Biomaterials, 2020, 260, 120312.
38. A. J. Camm, H. Sabbour, O. Schnell, F. Summaria and A. Verma, Cardiovasc. Diabetol., 2022, 21, 160.
39. Y. Iwasaki, S. Uchiyama, K. Kurita, N. Morimoto and N. Nakabayashi, Biomaterials, 2002, 23, 3421–3427.
40. Y. Inoue, J. Watanabe and K. Ishihara, J. Colloid Interface Sci., 2004, 274, 465–471.
41. Y. Zhao, M. Li, N. Chen, K. Huang, X. Wu, Y. Tan, Q. Hu, R. Luo and Y. Wang, Int. J. Biol. Macromol., 2025, 288, 138741.
42. Z. Zhao, M. Pan, C. Qiao, L. Xiang, X. Liu, W. Yang, X.-Z. Chen and H. Zeng, Adv. Mater., 2023, 35, 2208824.
43. R. Rusinova, C. He and O. S. Andersen, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2113229118.
44. L. Tong, D. Zhang, Z. Huang, F. Gao, S. Zhang, F. Chen and C. Liu, ACS Appl. Mater. Interfaces, 2024, 16, 43244–43256.
45. E. Pretorius, S. Mbotwe, J. Bester, C. J. Robinson and D. B. Kell, J. R. Soc., Interface, 2016, 13, 20160539.
46. P. Lockyer, H. Mao, Q. Fan, L. Li, L.-Y. Yu-Lee, N. T. Eissa, C. Patterson, L. Xie and X. Pi, Arterioscler., Thromb., Vasc. Biol., 2017, 37, 1524.
47. X. Yang, X. Cheng, Y. Tang, X. Qiu, Y. Wang, H. Kang, J. Wu, Z. Wang, Y. Liu, F. Chen, X. Xiao, N. Mackman, T. R. Billiar, J. Han and B. Lu, Immunity, 2019, 51, 983.
48. B. S. Salter, M. M. Weiner, M. A. Trinh, J. Heller, A. S. Evans, D. H. Adams and G. W. Fischer, J. Am. Coll. Cardiol., 2016, 67, 2519–2532.
49. C. U. Nilsson, K. Strandberg and P. Reinstrup, Scand. J. Clin. Lab. Invest., 2018, 78, 358–364.
50. H. Sun, Q. Li, G. Yin, X. Ding and J. Xie, Aging, 2020, 12, 20432–20444.
51. B. Li, J. Fang, Z. Zuo, S. Yin, T. He, M. Yang, J. Deng, L. Shen, X. Ma, S. Yu, Y. Wang and Z. Ren, Infect. Immun., 2018, 86, 10.

---

Footnote

† These authors contributed equally to this work.

---