Forced hemostasis via enhanced erythrocyte adhesion: A procyanidin-based composite sponge for managing hemorrhage in coagulopathic conditions

Abstract

Uncontrolled hemorrhage, particularly under coagulopathic conditions, remains a clinical challenge. Although commercial gelatin sponge (GS) is highly biocompatible, its passive hemostatic mechanism limits efficacy in managing severe hemorrhage. Here, we report a polyphenol-modified gelatin sponge (PG) fabricated via a facile one-step procyanidin conjugation. The optimal sponge formulation, PG3, exhibits enhanced hydrophilicity, rapid fluid absorption, and robust mechanical properties. Mechanistically, the hemostatic mechanism is primarily driven by the synergistic effect between Ca²⁺ and the polyphenol-modified surface, which markedly enhances protein adsorption and erythrocyte adhesion, without altering the intrinsic or extrinsic coagulation pathways. In a rat femoral artery injury model, PG3 significantly reduced bleeding time and blood loss by 62.3% in healthy rats and by 47.0% in heparinized coagulopathy rats, outperforming commercial GS. Additionally, PG3 exhibited excellent cytocompatibility, antioxidant capacity, and biodegradability. This work highlights a simple yet effective strategy for developing composite hemostatic materials, promising for the clinical management of severe bleeding, especially in challenging coagulopathic scenarios.

Graphical abstract

A facilely fabricated gelatin-polyphenol sponge achieves "forced hemostasis' through a synergistic effect between Ca²⁺ and the polyphenol-modified surface that enhances erythrocyte and plasma protein interactions, which maintains high efficacy under coagulopathic conditions and demonstrates good biocompatibility, controllable degradability, and antioxidant activity.

1. Introduction

Hemorrhage resulting from severe trauma in military conflicts, natural disasters, or traffic accidents remains a leading cause of preventable death worldwide [1,2]. Under such conditions, the body's intrinsic coagulation system is often insufficient, necessitating the use of external hemostatic materials to accelerate clot formation and control bleeding [3,4]. Currently available clinical hemostatic agents include microporous inorganic powders (e.g., zeolites and clays) and polymer-based materials (e.g., hydrogels, sponges, and films) [[5], [6], [7], [8]]. Among these, commercial gelatin sponge (GS) is widely employed due to its high porosity, compressibility, and excellent fluid absorption capacity [9]. The porous architecture of GS facilitates rapid water uptake and provides an extensive surface area for concentrating coagulation factors and blood cells, thereby promoting clot formation [10,11]. Moreover, GS is fabricated through freeze-drying crosslinked gelatin solutions, which ensures biocompatibility, safe degradation, and absorption within bodies without causing secondary wound damage [9]. However, the clinical application of GS in severe trauma is limited by its relatively low hemostatic efficiency, which stems from its passive hemostatic mechanism that relies predominantly on fluid absorption and blood concentration, lacking actively procoagulant surface properties [12,13]. This limitation becomes particularly critical in coagulopathic conditions, where impaired coagulation due to underlying diseases or anticoagulant therapy further compromises hemostasis [10,14]. Therefore, developing advanced hemostatic sponges that integrate structural simplicity, high biocompatibility, degradability, and enhanced hemostatic performance—especially for use in challenging coagulopathic scenarios—remains an urgent need.

Natural polyphenols, such as tannic acid, gallic acid, and procyanidin (PC), have attracted significant interest in biomedical materials due to their excellent biocompatibility and multifaceted bioactivities, including antioxidant, anti-inflammatory, antibacterial, and anticancer properties [[15], [16], [17]]. In the design of hemostatic materials, polyphenols can enhance wound adhesion by improving protein and cellular adhesion, which also support therapeutic benefits for post-hemostatic wound healing through their inherent bioactivities [18]. Recent advances further demonstrate the potential of polyphenols in developing hemostatic materials with enhanced blood cell interactions [12,[19], [20], [21]]. However, current polyphenol-based hemostatic materials face two major challenges: (1) complex fabrication processes and structural designs that hinder industrial scalability and clinical translation, or (2) dependence on intact coagulation systems, which reduces their less efficacy in coagulopathic hemorrhage. Despite these limitations, polyphenols' unique ability to regulate biological interfaces and deliver bioactivity presents a promising opportunity for their molecular-level engineering to advance next-generation hemostatic sponges.

In this study, we developed polyphenol-modified gelatin sponges (PG) through a facile one-step conjugation of PC with commercial GS via Michael addition/Schiff base reactions (under mildly alkaline condition of pH = 8.5) [22,23] between phenolic hydroxyl groups (oxidized for reactive quinones) and amino residues of gelatin (Fig. 1). The feasibility of PC incorporation for active procoagulant surface modification was systematically evaluated using in vitro hemostatic assays, including mechanistic studies on blood–material interactions under various conditions (i.e., whole blood, platelet-free blood, blood cell suspension, and heparinized blood). PG sponges with varying PC concentrations were prepared to identify the optimal formulation, which was selected for further analysis based on enhanced surface adhesion toward blood components. The hemostatic performance of the optimal PG sponge was evaluated in femoral artery injury models using both healthy and coagulopathic rats. In addition, comparative studies with unmodified GS were conducted to assess the effects of PC functionalization on cytocompatibility, antioxidant capacity, and controlled degradability, establishing a comprehensive framework for evaluating the translational potential of advanced hemostatic sponges.

Fig. 1.

Schematic illustration of the preparation process of the PG Sponge for potent hemostatic applications, along with the underlying mechanism.

2. Experimental section

2.1. Materials

Gelatin sponge was purchased from Jinling Pharmaceutical (China). Procyanidin was purchased from Wanfang Biotechnology (China). Tris(hydroxymethyl) aminomethane (Tris), DL-methionine, riboflavin (vitamin B2) and nitrotetrazolium blue chloride were purchased from Energy Chemical (China). Normal saline (NS) was obtained from Shijiazhuang No.4 Pharmaceutical (China). 2,2-diphenyl-1-picrylhydrazyl (DPPH) was obtained from TCI Chemical (Japan). Other biochemical kits/reagents are provided in the Supporting Information.

2.2. Preparation of PG sponges

A Tris–HCl aqueous solution (1 mg/mL, pH 8.5) was first prepared and mixed with anhydrous ethanol at a 9:1 (v/v) ratio to obtain a Tris/ethanol solution. PC was dissolved in this solution to achieve concentrations of 1, 3, and 5 mg/mL. GS samples (6 cm × 2 cm × 0.5 cm) were immersed in the PC solutions (40 mL per sponge) under room temperature (RT) for 10 min followed by three rinses with deionized (DI) water to remove unbound PC. The modified sponges were transferred to a culture dish, maintained in a swollen state with DI water, rapidly frozen at −80 °C, lyophilized to produce PG sponges (denoted as PG1, PG3, and PG5, corresponding to the PC concentrations) and subsequently stored under vacuum.

2.3. Physical characterization

Fourier transform infrared (FT-IR) spectra of sponge samples (GS, PG1, PG3 and PG5) and PC were recorded using an FT-IR spectrometer (Scientific Nicolet IS 10, Thermo, USA) with KBr pellets. Spectra were acquired over 4000–400 cm⁻¹ at a resolution of 4 cm⁻¹ with 16 scans. For internal morphological analysis, sponge samples were quenched in liquid nitrogen, sectioned, sputter-coated with platinum, and examined using a scanning electron microscope (SEM; JSM-7500F, JEOL, Japan). The surface potentials of sponge samples were measured with a solid surface potential meter (SurPASSTM3, Anton Paar, Austria) equipped with fixture model 165040. The contact angles of sponge samples were determined using a video-based contact angle instrument (OCA25, Dataphysics, Germany). The compressive properties of sponge samples (0.5 cm × 0.5 cm × 0.5 cm) were assessed using a universal testing machine (Shimadzu EZ-LX 50N) at a compression rate of 5 mm/min, measuring compressive strength at 80% strain. The porosity and liquid absorption capacity of sponge samples were determined according to previously reported methods [13,24]. The PC content in the sponge samples was assessed by absorbance measurements to quantify the amount of PC retained after the preparation/purification process, as well as the stability of PC after incubation in phosphate-buffered saline (PBS). Detailed experimental procedures are provided in the Supporting Information.

2.4. Hemolysis and cytotoxicity assays

For the hemolysis assay, freshly collected citrated whole blood from healthy Sprague-Dawley (SD) rats was used to prepare erythrocyte suspensions. All animal procedures were approved by the Animal Ethics Committee of the China-Japan Friendship Hospital (Beijing, China) and conducted in accordance with institutional guidelines. Erythrocyte suspensions (2%, v/v) were incubated with sponge samples (GS, PG1, PG3, PG5) at a concentration of 3 mg/mL and 37 °C for 3 h. The absorbance of the supernatant was measured to quantify hemoglobin release [25]. For the cytotoxicity assay using the MTT method, leaching solutions from sponge samples (3 mg/mL incubated in cell culture medium at 37 °C for 24 h) were applied to mouse fibroblast (L929) cells. After incubation, cell viability was determined by measuring formazan crystal formation [26]. Detailed experimental procedures are provided in the Supporting Information.

2.5. In vitro blood clotting index (BCI) assay

The hemostatic property of sponge samples (GS, PG1, PG3 and PG5) was assessed using a standard BCI assay, during which citrated whole blood from SD rats was recalcified by adding calcium chloride solution to restore coagulation. The BCI value was determined by absorbance measurement after incubating the samples (5 mg) with the blood (100 μL), in the presence of calcium chloride (CaCl₂, 10 μL, 0.2 M) solution [25]. To investigate the effect of platelets, the BCI assay of sponge samples (5 mg) was measured using platelet-free blood (prepared by centrifugation, where platelets were completely removed, 100 μL), in the presence of CaCl₂ (10 μL, 0.2 M). Additionally, to account for liquid absorption capacity (relevant to the procoagulant mechanism), pre-wetted sponge samples (moistened with PBS) were tested under the same conditions. The BCI assay of dry samples was also assessed using blood from heparinized rats to simulate coagulopathic conditions [13]. Detailed experimental procedures are provided in the Supporting Information.

2.6. In vitro blood-material interaction assays

The erythrocyte adhesion to the dry sponge samples (GS, PG1, PG3 and PG5) was evaluated to assess the erythrocyte-material interactions. The dry sponge samples (5 mg) were incubated with the following liquids (100 μL): citrated whole blood (in the presence and absence of CaCl₂, 10 μL, 0.2 M), Platelet-free citrated blood (in the presence and absence of CaCl₂, 10 μL, 0.2 M), and 50% erythrocyte suspension (in the absence of CaCl₂, 10 μL, 0.2 M). The number of adherent erythrocytes was quantified by absorbance measurement [13,25]. For the platelet-material interaction, platelet adhesion to sponge samples (GS, PG1, PG3, PG5) was assessed by incubating dry samples (5 mg) with platelet-rich plasma (PRP, 50 μL) and quantifying the adherent platelets via biochemical kit [21]. For the plasma-material interaction, the effect of dry samples (GS, PG1, PG3, PG5, 3 mg) on plasma coagulation was determined by activated partial thromboplastin time (APTT) and prothrombin time (PT) assays using platelet-poor plasma (PPP, 100 μL). The fibrinogen (Fgn) adsorption was assessed by incubating the dry sponge samples with PPP (25 μL) in the absence of CaCl₂ (5 mg), and the amount of adsorbed protein was quantified using a biochemical assay kit [12]. The adsorption of whole protein was evaluated by incubating the dry samples with PPP (100 μL) both in the presence and absence of CaCl₂ (5 mg), followed by quantification using the same biochemical method. In addition, pre-wetted sponge samples (moistened with PBS) were tested under identical conditions for erythrocyte adhesion (using citrated whole blood with CaCl₂ and platelet-free citrated blood with CaCl₂) and for whole-protein adsorption (using PPP with CaCl₂). The adhesion of erythrocytes on both dry and pre-wetted sponge samples was further visualized using SEM after exposure to citrated whole blood with CaCl₂. Detailed experimental procedures are provided in the Supporting Information.

2.7. Antioxidant property assays

The antioxidant capacity of sponge samples (GS, PG1, PG3 and PG5) was assessed using DPPH and superoxide anion (•O₂⁻) scavenging assays, according to previously reported methods with slight modifications [12]. The radical scavenging capacity of sponge samples was determined by absorbance measurement. Additionally, intracellular reactive oxygen species (ROS) scavenging was evaluated in L929 cells using a fluorescent probe [27]. Detailed experimental procedures are provided in the Supporting Information.

2.8. In vivo artery injury model in healthy and heparinized rats

The hemostatic performance of sponge samples (GS, PG1, PG3 and PG5) was evaluated using a femoral artery injury model in healthy and heparinized SD rats (male, 190–230 g) [28,29]. All procedures were approved by the Animal Ethics Committee of the China-Japan Friendship Hospital (Beijing, China; Approval No. ZDRWLL240004). In healthy rats, the femoral artery was transected and allowed to bleed freely for 10 s. Pre-treatment blood loss was recorded, and rats with blood loss between 100 and 350 mg were selected for consistency. Sponge samples (2 cm × 2 cm × 0.5 cm) were applied to the wound with a 100-g weight to simulate standard pressure. Bleeding time and post-treatment blood loss were recorded. Detailed experimental procedures are provided in the Supporting Information.

In heparinized rats, coagulopathy was induced by heparin injection (100 IU/kg, circulated for 3 min, confirmed by prolonged APTT values), and the same procedures were followed with those in femoral artery model of healthy rats. The surface and cross-sectional microstructures of GS and PG3 sponges after 3-min pressure in the heparinized femoral artery model were comparatively analyzed using SEM.

2.9. In vitro and in vivo degradation assays

For the in vitro degradation, sponge samples (GS and PG3) were incubated in PBS at 37 °C. At predetermined time points (1, 3, 5, and 7 days), the samples were collected, freeze-dried, and weighed to determine the mass loss [30]. Additionally, the extracts/degradation solutions collected on days 1 and 3 were subjected to BCI and thromboelastography (TEG) assays to evaluate their effects on blood coagulation.

For the in vivo degradation, sponge samples were subcutaneously implanted in SD rats, and harvested on days 3, 7, 14, and 21 for morphological and histological analysis. On day 7, the surrounding tissues of implanted sponges were stained with hematoxylin and eosin (H&E) to evaluate inflammatory responses. Concurrently, organs (heart, liver, spleen, lung, and kidney) and blood samples were collected for H&E staining and TEG analysis, respectively, to assess biosafety. Detailed experimental procedures are provided in the Supporting Information.

2.10. Statistical analysis

All quantitative data are expressed as mean ± standard deviation (SD). Differences between two groups were analyzed using Student's t-test, while multiple groups were compared using one-way ANOVA with post-hoc testing. Statistical significance was set at ∗ P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001, indicating increasing levels of significance.

3. Results and discussion

3.1. Preparation and characterization of PG sponges

A series of proanthocyanidin-modified gelatin sponges (denoted as PG1, PG3, and PG5) were prepared by immersing commercial GS in Tris/ethanol solutions (v/v, 9:1) containing varying concentrations of PC, followed by freeze-drying. The alkaline Tris/ethanol solution was employed to facilitate the Michael addition/Schiff base reactions between the amino groups (of gelatin) and the phenolic hydroxyls (of PC, followed by oxidation for reactive quinones), which also minimizing volume changes during processing [12]. Successful incorporation of PC was first indicated by a visible color change: the PG sponges exhibiting a uniform reddish-brown hue, distinct from the white color of pristine GS (Fig. 2a), which is attributable to the intrinsic color of PC. SEM images revealed that all sponges retained a typical three-dimensional porous network structure after modification (Fig. 2a).

Fig. 2.

(A) Photograph and SEM images of GS and PG sponges. (b) FTIR spectra of GS/PG sponges and PC. (c) Semi quantitative analysis of infrared spectra of PG sponges (mean ± SD, n = 3; one-way ANOVA). (d) Surface potential, (e) water contact angle, (f) porosity of sponges, (g) liquid absorption rate for 30 min, and (h) liquid absorption rate for 10 s of GS and PG sponges (mean ± SD, n = 3; one-way ANOVA).

As shown in Fig. 2b, FT-IR spectroscopy provided further evidence for the successful modification of PC onto PG sponges. While the characteristic peak of gelatin at 2934 cm⁻¹ (C–H stretching vibration) [31] was present in all samples, a new peak emerged at 1110 cm⁻¹ [13,32] in the PG sponges (corresponding to anthocyanins), confirming PC's presence. Semi-quantitative analysis of the absorbance ratio (A₁₁₁₀/A₂₉₃₄) demonstrated a gradual increase in PC content with higher PC feeding concentrations (Fig. 2c). The actual PC content in the three PG sponges were determined to be 19.03, 37.40 and 58.37 mg/g, respectively (Fig. S1a). The introduction of negatively charged phenolic hydroxyl groups from PC significantly altered the surface properties. The surface potential of the sponges became more negative with increasing PC content (Fig. 2d). Consistently, water contact angle measurements revealed enhanced hydrophilicity after modification (Fig. 2e). The original GS was relatively hydrophobic (contact angle of 135.6°), whereas the PG sponges showed a clear trend of increasing hydrophilicity with PC content, aligning with previous reports on polyphenol-modified gelatin [33].

Porosity, which is crucial for liquid absorption and for concentrating blood components in hemostatic applications [34,35], remained high (∼90%) for all sponges and was unaffected by PC incorporation (Fig. 2f). The saturated liquid absorption capacity in PBS also remained largely unchanged (up to 60 times, Fig. 2g). Notably, the water absorption rate of the PG sponges was significantly faster than that of GS (Fig. 2h). An optimal PC content was observed: the medium PC in PG3 achieved the best balance (reaching near-saturation within 10 s) between significantly improved surface hydrophilicity and a well-preserved, open pore-network structure. This balance maximized the capillary driving force for initial liquid uptake, whereas the excessive crosslinking in PG5, despite its similar hydrophilicity, would slightly hinder the very initial diffusion kinetics within the pores.

The stability of the incorporated PC in three PG sponges was evaluated by incubating the sponges in PBS and quantifying the released PC (Fig. S1b). After 12 h, the amounts of released PC were 4.29, 5.71, and 7.50 mg/g, respectively, which were markedly lower than the initial loadings, indicating the high retention stability. Furthermore, the compressive strength of the PG sponges at 80% strain remained robust and comparable to GS (Fig. S2), indicating their structural integrity was sufficient to withstand stresses encountered during clinical handling [36].

3.2. In vitro cytocompatibility and hemostatic performance

Cytocompatibility is a prerequisite for hemostatic materials [36,37]. The hemolysis rate of all PG sponges remained below the critical safety threshold of 5% (Fig. 3a), confirming favorable blood compatibility [37]. Assessment of cytotoxicity using L929 cells via the MTT assay revealed no adverse effects of all PG sponges (Fig. 3b), with cell viability exceeding 100% for all groups, indicating no cytotoxicity and a potential slight promotive effect on cell proliferation [38]. The in vitro hemostatic property of PG sponges was quantitatively evaluated using the standard BCI assay, adopting the recalcified condition of normal-blood (from healthy rats) to restore coagulation [13,28]. All PG sponges exhibited significantly lower BCI values than GS (Fig. 3c), demonstrating the contribution of PC to procoagulant activity. Notably, PG3 showed the lowest BCI value, suggesting an optimal PC content for hemostatic performance.

Fig. 3.

(A) Hemolysis rate and (b) L929 cell viability of GS/PG sponges (mean ± SD, n = 3; one-way ANOVA). (c, d) BCI value of dry GS/PG sponges incubated with (c) normal blood and (d) platelet-free blood in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (e, f) BCI value of wet GS/PG sponges incubated with (e) normal blood and (f) platelet-free blood in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (g) SEM images of erythrocytes adherent on dry and wet GS/PG sponges incubated with normal blood in the presence of Ca²⁺.

Given the excellent hemostatic property of PG sponges in normal blood, we further evaluated their performance under various pathological blood conditions and sample treatments. As shown in Fig. 3d, the procoagulant activity of GS was significantly reduced in recalcified platelet-free plasma. In contrast, PG1, PG3, and PG5 sponges maintained robust hemostatic performance, with PG3 consistently demonstrating the highest property. These results indicate that the procoagulant activity of PG sponges is independent of platelets, indicating their potential application in managing platelet-related coagulation disorders, such as thrombocytopenia and platelet dysfunction [28]. Notably, the hemostatic property of wet PG sponges (pre-saturated with PBS) against recalcified normal blood was reduced compared to that of dry sponges (Fig. 3e). The decline was most pronounced in PG1 and PG5, whereas PG3 maintained relatively strong activity. As shown in Fig. 3f, the BCI values of wet sponges in recalcified platelet-free blood were nearly identical to those observed in recalcified normal blood, further indicating the potential applicability of wet PG sponges in the context of platelet-related coagulation disorders. Furthermore, SEM analysis of pre-saturated sponges revealed less dense clot structures compared to dry conditions (Fig. 3g), which is attributable to the absence of the procoagulant activity endowed by rapid fluid absorption and consequent blood component concentration [39]. Nevertheless, under pre-saturated conditions, PG3 still demonstrated better procoagulant activity than GS and other two PG sponges and shows potential for hemostasis for minimally invasive/interventional deep hemostasis (i.e. hydrated sponge particles can be delivered through a catheter to manage parenchymal hemorrhage and vascular embolism) [40], underscoring the critical role of an optimal PC content.

3.3. Investigation of hemostatic mechanism under varied blood conditions

To elucidate the hemostatic mechanism, we evaluated the APTT/PT, erythrocyte adhesion, and plasma protein adsorption [41,42]. No significant differences in APTT or PT were observed between the GS and PG groups (Fig. S3), indicating that the unaltered endogenous/exogenous coagulation systems after PC incorporation. Erythrocyte adhesion was first performed under the typical Ca²⁺-free condition [13,21,43,44]. In normal citrated whole blood (which is calcium-deficient due to chelation, Fig. 4a), in platelet-free citrated blood (Fig. 4b), and in a simplified erythrocyte suspension (Fig. S4), the overall level of erythrocyte adhesion was below 25% in all GS and PG groups. Nevertheless, the PG3 group consistently exhibited a higher adhesion ratio than GS and other two PG groups, likely due to its optimal PC content. Consistently, whole protein adsorption remained about 35% for all groups under Ca²⁺-free conditions (Fig. 4c), yet the PG3 still demonstrated superior adsorption. Although the polyphenol-modified surface possesses strong intrinsic cellular/protein affinity [45], the moderate PC content in PG3—especially compared with the higher loading in PG5—probably provides the most effective density of phenolic binding sites without excessive crosslinking that could mask these sites or unfavorably alter the surface topography for cell/protein attachment. Furthermore, in the absence of Ca²⁺, the Fgn adsorption and platelet adhesion ratios were similarly low for all sponges (Fig. S5), and PG3 no longer exhibited clear superiority over GS or the other PG groups. These results, together with the similar BCI results of PG3 in normal blood and platelet-free blood (Fig. 3c and d), indicate that the hemostatic effect of the PG3 sponges does not rely on the Fgn-mediated platelet adhesion/activation pathway [46,47].

Fig. 4.

(A,b) Erythrocyte adhesion ratio of dry GS/PG sponges incubated with normal blood and platelet-free blood in the absence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (c) Whole-protein adsorption ratio of dry GS/PG sponges incubated with PPP in the absence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (d,e) Erythrocyte adhesion ratio of dry GS/PG sponges incubated with normal blood and platelet-free blood in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (f) Whole-protein adsorption ratio of dry GS/PG sponges incubated with PPP in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (g,h) Erythrocyte adhesion ratio of wet GS/PG sponges incubated with normal blood and platelet-free blood in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA). (i) Whole-protein adsorption ratio of wet GS/PG sponges incubated with PPP in the presence of Ca²⁺ (mean ± SD, n = 3; one-way ANOVA).

Notably, when erythrocyte and protein adhesion assays were performed in the presence of Ca²⁺, a synergistic enhancement was observed. In recalcified blood (prepared by adding CaCl₂ to citrated whole blood to simulate a free calcium concentration of normal rat blood, 1.1-1.4 mM [21,48]), the erythrocyte adhesion ratio of PG sponges increased markedly, with PG3 reaching 78.4 ± 1.3% (Fig. 4d). Even in platelet-free recalcified blood, PG3 maintained a high adhesion ratio of 73.6 ± 4.1% (Fig. 4e), confirming that its erythrocyte adhesion capability is independent of platelets. Specifically, the presence of CaCl₂ markedly increased the adsorption of whole plasma proteins onto the PG sponges, particularly for PG3 (Fig. 4f), whereas the adsorption ratios were lower under Ca²⁺-free condition (Fig. 4c). These data highlight the synergistic role of surface polyphenol and Ca²⁺ in enhancing protein adsorption; the resulting denser protein layer likely provides a favorable interface for subsequent erythrocyte adhesion [49]. Although previous studies have reported that Ca²⁺-mediated interfacial interactions can promote clotting-factor accumulation and protein/cell adsorption on material surfaces [50], those works typically employed Ca²⁺-containing materials. In contrast, our approach leverages an optimized surface PC modification in combination with endogenous Ca²⁺ present in blood to enhance plasma protein and erythrocyte adhesion.

To further differentiate the respective contributions of rapid fluid uptake and surface interaction, erythrocyte and protein adhesion assays were also performed under pre-saturated condition (which eliminates the rapid absorption capability of PG3). Compared with dry GS and PG sponges, the pre-saturated counterparts exhibited markedly reduced ratios of erythrocyte adhesion and whole protein adsorption (Fig. 4g–i) and the sparser clot structures were observed on their surfaces (Fig. 3g). On one hand, these results indicated attenuated interactions between blood components and the sponge surfaces in the absence of the blood-concentrating effect. On the other hand, even under pre-saturated condition, the adhesion ratios of PG3 remained higher than those of the other groups, reaffirming the PG3 surface acts synergistically with Ca²⁺ for enhance plasma protein and erythrocyte adhesion.

Collectively, the superior hemostatic property of the PG3 can be attributed to three key factors (as illustrated in Fig. 1): firstly, a synergistic effect among the sponge's PC surface, Ca²⁺ ions, and plasma proteins that significantly enhances erythrocyte adhesion (exhibiting a "forced hemostasis' capability); secondly, the inherent strong adhesion of the PC surface for erythrocytes; and thirdly, its rapid fluid absorption capacity.

3.4. In vivo hemostatic performance in healthy and heparinized rat models

To evaluate the hemostatic performance of GS and PG sponges under acute hemorrhage conditions, a femoral artery injury model was established in healthy SD rats [13,21]. Only rats with a pre-treatment blood loss ranging from 100 to 350 mg were included to ensure comparable baseline bleeding levels across groups (Fig. 5a). The in vivo hemostatic efficacy was assessed based on post-treatment blood loss and bleeding time. Blank group exhibited the high amount of post-treatment blood loss of 2049.7 ± 82.3 mg and bleeding time of 19.3 ± 1.5 min, which confirmed the femoral artery injury as one typical hemorrhage scenario (Fig. 5b and c). The PG3 group exhibited a 62.3% reduction in blood loss compared to the GS group (151.6 ± 57.6 mg vs. 401.3 ± 29.6 mg) (Fig. 5b). Representative photographs (Fig. S6) showed that the GS sponge was fully saturated with blood leakage, whereas blood was confined to the central region of PG3, indicating rapid hemostasis. Additionally, the bleeding time was significantly shorter for the PG3 group (4.0 ± 0.6 min) than for GS (8.67 ± 1.4 min) (Fig. 5c), collectively demonstrating the superior in vivo efficacy of PG3.

Fig. 5.

(A) Pre-treatment blood loss, (b) post-treatment blood loss, and (c) bleeding time of GS/PG sponges in femoral artery injury model of healthy rats (mean ± SD, n ≥ 3; one-way ANOVA). (d) PT and (e) APTT values of whole blood drawn from heparinized SD rats (mean ± SD; n = 3). (f) BCI value of GS/PG sponges incubated with heparinized blood; mean ± SD, n = 3; one-way ANOVA). (g) Pre-treatment blood loss, (h) post-treatment blood loss, (i) bleeding time and (j) representative photographs of GS/PG3 sponges in femoral artery injury model of heparinized rats (mean ± SD, n = 3; Student's t-test). SEM images of (k) surface morphologies and (l) section morphologies of GS/PG3 sponges in femoral artery injury model of heparinized rats.

Anticoagulated patients receiving heparin therapy are at high risk of uncontrolled bleeding during surgical procedures, which poses a significant challenge for achieving effective hemostasis [51]. Given the high efficacy of PG3 in healthy rats and mechanism involving plasma protein/erythrocyte interactions, we evaluated its performance in a heparinized coagulopathy model. Heparin administration (100 IU/kg) significantly prolonged APTT (Fig. 5d) without affecting PT (Fig. 5e), confirming a coagulopathic state [13,21]. The BCI of GS was 93.7% in heparinized blood, indicating inadequate hemostasis, whereas PG1, PG3, and PG5 exhibited lower values, with PG3 showing the best efficacy (31.8%, Fig. 5f), suggesting its potential for effective hemorrhage control under coagulopathic conditions. For in vivo performance, the pre-treatment blood loss was similar between GS and PG3 groups (Fig. 5g). After treatment, blood loss decreased from 956.2 ± 95.7 mg (GS) to 507.8 ± 76.3 mg (PG3) (Fig. 5h), and bleeding time was reduced from 30 min to 10.6 ± 1.5 min (Fig. 5i). Photographs of the sponges after hemostasis showed markedly higher residual blood in the GS group compared to PG3 (Fig. 5j). Notably, surface morphology after 3-min pressure (Fig. 5k) showed the PG3 sponge surface covered with a dense layer of adherent plasma proteins and erythrocytes, contributing to effective hemostasis or injury sealing. In contrast, GS adsorbed significantly fewer components. Cross-sectional analysis (Fig. 5l) revealed a relatively thick membrane of densely accumulated proteins and erythrocytes on PG3 (red arrows), whereas GS showed limited adsorption. These results demonstrate that the superior hemostatic performance of PG3 under heparinized conditions is attributable to its strong adhesion capability toward plasma proteins and erythrocytes (exhibiting a "forced hemostasis' capability), facilitating rapid clot formation even when the intrinsic coagulation pathway is compromised.

3.5. Antioxidant properties and degradation behaviors

Excessive ROS in the wound microenvironment can induce oxidative stress, amplify inflammatory responses, and hinder tissue repair [52,53]. To evaluate the antioxidant capacity, the free radical scavenging activities of GS and PG sponges were assessed using DPPH, •O₂⁻, and intracellular ROS scavenging assays. As shown in Fig. 6a, the GS control (without PC) exhibited negligible DPPH scavenging activity. In contrast, PG groups showed a concentration-dependent increase in DPPH radical scavenging efficiency with higher PC content. In the •O₂⁻ scavenging assay (Fig. 6b), PG3 demonstrated moderate activity, while PG5 did not exhibit further enhancement, suggesting that a threshold concentration of PC is required for significant •O₂⁻ scavenging. The antioxidant capacity of GS and PG sponges was further evaluated by intracellular ROS clearance assays (Fig. 6c and d). Both PG3 and PG5 exhibited clearance ratios approaching 100%, confirming their potent ability to eliminate excessive oxidative stress in a cellular environment. These results collectively demonstrate that the incorporated PC endows PG sponges with excellent antioxidant performance.

Fig. 6.

(A) DPPH scavenging ratio and (b) •O₂⁻ scavenging ratio of GS/PG sponges (mean ± SD, n = 3; one-way ANOVA). (c) ROS scavenging ratio and (d) representative fluorescence images of L929 cells incubated with GS/PG sponges (mean ± SD, n = 3; one-way ANOVA). (e) In vitro and (f) in vivo degradation kinetics of GS and PG3 sponges (mean ± SD, n ≥ 3). (g) Photographs of the incisions and GS/PG sponges on Day 0 to Day 14 (white circle indicates the residual sponges). (h) H&E-stained images of harvested skin tissues around the incisions on Day 3 to Day 21 (black box indicates the incision area and black arrow indicates inflammation-associated cells).

An ideal wound dressing should degrade in a timely manner after fulfilling its function to avoid secondary tissue injury [54]. In vitro degradation assays showed that GS degraded completely by day 5. PG3, however, retained about 20% of its mass on day 5 and reached a degradation rate of ∼94% by day 7 (Fig. 6e), which is attributable to cross-linking between PC and gelatin. To evaluate the thrombotic risk associated with degradation products, extracts of PG3 degraded in vitro for 1 or 3 days were incubated with whole blood (v/v, 1:2-1:10) for BCI measurement. No significant differences in BCI values were observed between any extract of PG3 and the PBS (blank) or GS controls, indicating the absence of a prothrombotic response (Fig. S7). Furthermore, TEG analysis of PG3 degradation products revealed no statistically significant changes in key coagulation parameters compared to both GS and PBS controls (Fig. S8). These in vitro results suggest no acute thrombotic risk associated with PG3 degradation.

Subcutaneous implantation in SD rats further validated the in vivo degradation profile: while GS degraded entirely by day 14, PG3 retained 22% of its mass at this time point and achieved complete degradation by day 21 (Fig. 6f and g and Fig. S9). Histological analysis revealed transient inflammatory cell infiltration around PG3 during degradation, which resolved after sponge resorption with no residual inflammation, confirming good biocompatibility and desirable biodegradability (Fig. 6h). The in vivo safety assessment further supported these findings. TEG analysis of blood collected from rats implanted with GS or PG3 sponges for 7 days showed no significant changes in key coagulation parameters compared to the Blank and GS control groups (Fig. S10). Consistently, H&E-stained organ sections from implanted rats exhibited no notable morphological abnormalities compared to those from normal rats, indicating no organ toxicity (Fig. S11). The favorable safety profile of PG3 is further supported by toxicological data: the no observed adverse effect level for PC is 1462 mg/kg bw/day [55]. Based on a PC content of 37.40 mg/g in PG3 and an implanted sponge mass per rat, the administered PC dose was calculated to be below 2 mg/kg, which is well below the established safety threshold. Collectively, PG3 combines potent antioxidant activity, tunable degradation kinetics, and a favorable safety profile, positioning it as a promising degradable dressing for managing complex wound scenarios [56].

4. Conclusions

In summary, we developed a novel composite hemostatic material based on procyanidin-modified gelatin sponges via a facile one-step polyphenol-amino conjugation. The modified PG sponges retained a porous architecture and exhibited enhanced hydrophilicity, with the optimized formulation PG3 (with moderate PC content) achieving rapid liquid saturation. The superior procoagulant activity of PG3 was demonstrated in vitro and attributed to a synergistic effect between the polyphenol-modified surface and Ca²⁺, which significantly enhances protein adsorption and erythrocyte adhesion—a mechanism that remains effective even in platelet-deficient or heparinized coagulopathy conditions. In an artery injury model, PG3 reduced blood loss in both healthy and heparinized coagulopathy rats. PG3 also showed excellent biocompatibility, controlled biodegradation, and antioxidant capacity. The present work highlights the good potential of a simple polyphenol-based functionalization strategy for developing high-performance hemostatic materials, with promising implications for the clinical management of severe bleeding, especially in challenging heparinized coagulopathy scenarios.

CRediT authorship contribution statement

Qun Liu: Data curation, Investigation, Methodology, Validation, Writing – original draft. Jie Lin: Data curation, Investigation, Methodology, Validation, Writing – original draft. Xiaoli Liu: Data curation, Investigation, Methodology, Validation, Writing – review & editing. Guochao Zhang: Data curation, Investigation, Methodology. Rui Lian: Conceptualization, Funding acquisition, Methodology, Supervision, Writing – review & editing. Kai Xiang: Investigation, Methodology. Chen Xu: Data curation, Validation. Yang Hu: Conceptualization, Funding acquisition, Methodology, Supervision, Writing – original draft, Writing – review & editing. Fu-Jian Xu: Conceptualization, Funding acquisition, Supervision, Writing – review & editing.

Declaration of competing 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.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.