Healthy adult volunteers were recruited for median cubital vein blood collection. These individuals had abstained from any antiplatelet medications for at least 2 weeks prior to donation and reported adequate rest. All blood samples collected were exclusively for scientific research purposes, with informed consent obtained in advance. The study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Soochow University.

The BALB/c wild-type (WT) mice were procured from Changzhou Laboratory Animal Co., Ltd. (China). Soochow University provides specific pathogen free (SPF) grade animal housing facilities for laboratory animals. The experimental animals utilized in this study were housed in the SPF grade facility at Soochow University, and all animal experiments were conducted in strict compliance with the guidelines of animal ethics, utilizing pentobarbital sodium or isoflurane gas anesthesia for both experimentation and euthanasia. These procedures have been approved by the Medical Ethics Committee of the First Affiliated Hospital of Soochow University. Male BALB/c mice, weighing approximately 20 g and aged around 6 weeks, were selected for the study. The ADR mouse model was established by administering a single intravenous injection of doxorubicin hydrochloride at a dose of 7.5 mg/kg via the tail vein. Doxorubicin hydrochloride was dissolved in 0.9% sodium chloride solution. Two weeks post-injection, kidney histopathological staining and plasma protein concentration were evaluated to confirm successful model establishment. The ADR Albumin mouse model was further developed by injecting mouse albumin at a dose of 30 mg once a week through the intraorbital venous plexus, repeated twice.

Human plasma albumin (70024-90-7), thrombin (10602-40-001), U46619 (5685-40-1), ADP (58-64-0), fibrinogen (9001-32-5), glycine methyl ester hydrochloride (5680-79-5), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) (25952-53-8), and doxorubicin hydrochloride (25316-40-9) were obtained from Sigma-Aldrich (St. Louis, MO, United States). FITC-labeled anti-human P-selectin (clone AK4, 304904) was purchased from BioLegend Corporation (San Diego, CA, United States). FITC-labeled anti-human PAC-1 (340507) was sourced from BD Biosciences (San Jose, CA, United States). FITC-conjugated rat anti-mouse P-selectin (clone Wug.E9, M130-1) and PE-conjugated rat anti-mouse Integrin αIIbβ3 (M023-2) were obtained from Emfret Analytics (Eibelstadt, Germany). Collagen (385) and Chrono-Lume (395) were provided by Chrono-Log Corporation (Havertown, PA, United States). PKC activator PMA (HY-18739), PKC activator PDBu (HY-18985), Akt activator 740Y-P (HY-P0175), and Akt activator SC79 (HY-18749) were acquired from MedChemExpress (Monmouth Junction, NJ, United States). Mouse serum albumin (SEKM-0297), Tris-Base, bovine serum albumin (BSA), tissue fixative solution, loading buffer solution, and glycine were purchased from suppliers in Beijing, China. Alexa Fluor 488-phalloidin (A12379) and the Slip-Alyzer dialysis kit with a 20 kDa molecular weight cut-off (MWCO) (66003) were obtained from Thermo Fisher Scientific (Rockford, IL, United States). An SDS-PAGE adhesive kit (10%) was purchased from Epizyme Biotech Co., Ltd. (Shanghai, China). Rabbit antibodies against phospho-PKC substrate (2261), phospho-Akt (Ser473) (11E7), Akt (pan) (11E7), PKC alpha (2056), GAPDH (5174), cell lysate buffer (9803), and PAR4-AP (2328) were sourced from Cell Signaling Technology, Inc. (Beverly, MA, United States). HRP-conjugated goat anti-mouse IgG (A0216), anti-rat IgG (A0192), and anti-rabbit IgG (A0208) were obtained from Beyotime Biotechnology (Shanghai, China). Mouse albumin, prealbumin, total protein, and globulin ELISA testing kits were procured from Aifang Biotechnology Co., Ltd. (Hunan, China). The biochemical reagents utilized for the preparation of modified tyrode’s buffer (MTB) and acid-citrate-dextrose (ACD) blood preservative solutions included citric acid, glucose, sodium citrate, sodium chloride, magnesium chloride, and EDTA. Magnesium chloride and EDTA were sourced from suppliers in Suzhou, China.

Male WT, ADR, and ADR Albumin mice weighing 20∼25 g and approximately 8 weeks old were selected for the experiment. Blood was collected using ACD anticoagulant at a ratio of 1:7 (ACD: whole blood). The ACD solution contained 2.5% trisodium citrate, 2.0% D-glucose, and 1.5% citric acid. The mixture was centrifuged at 1,100 rpm for 8 min to separate platelet-rich plasma (PRP). PRP was then subjected to a second centrifugation at 3,500 rpm for 2 min. The supernatant was discarded, and platelets were washed with CGS buffer. Following another centrifugation at 2,500 rpm for 2 min, the supernatant was discarded, and the washing process was repeated with 1 mL of CGS buffer. After a final centrifugation the supernatant was removed, and the platelets were resuspended in MTB buffer. 1 M CaCl₂ and 1 M MgCl₂ were added at a ratio of 1:1000.

Following the selection of resting platelets, thrombin, U46619, PAR4-AP, and Trap-6 were introduced as platelet agonists and incubated at 37 °C for 30 min. Subsequently, 10 μL of the treated platelets were transferred to a flow cytometry tube, followed by the addition of 1.5 μL of fluorochrome-conjugated monoclonal antibody. The mixture was then incubated in the dark for 15 min. To terminate the reaction, 350 μL of MTB was added, and flow cytometry analysis was performed as soon as possible.

The platelet aggregometer was preheated to 37 °C prior to the experiment. A volume of 250 μL of washed platelets was introduced into an aggregation tube, followed by the addition of various platelet agonists including thrombin, U46619, collagen, PAR4-AP, Trap-6, and ADP. The platelet aggregation rates elicited by these different agonists were recorded at a stirring speed of 1,200 rpm.

A total of 250 μL of washed platelets were introduced into the aggregometer tube. Lume and ATP standard solution were sequentially added to achieve a maximum ATP release value between 40% and 80%. The volumes of lume added and the corresponding ATP release percentages were meticulously recorded. Following this step, various agonists including thrombin, U46619, collagen, PAR4-AP, Trap-6, and ADP were selected for further analysis.

A slide coated with 30 μg/mL fibrinogen was utilized. Blocking was performed using 5% BSA at room temperature for 2 h. Rested platelets at a concentration of 2 × 10⁷/mL were stimulated with platelet agonists and promptly transferred to the slide at approximately 20 μL per well, spreading at 37 °C for 40 min. The slide was then washed three times and fixed with 4% PFA for 10 min. Permeabilization was achieved using 0.1% Triton-X100. Blocking was repeated with 5% BSA for 1 h. F-actin in the platelets was labeled with Alexa Fluor 488-phalloidin in a light-protected environment. Platelet spreading was observed under a Leica confocal microscope using a 63 × oil immersion lens.

WT, ADR, and ADR Albumin mice weighing 20∼25 g and approximately 8 weeks old were selected for intraperitoneal injection of 1% pentobarbital sodium for anesthesia. The fluorescent indicator calcein was added at a ratio of 1:300, and platelets were incubated in the dark for 30 min. Four-week-old WT mice underwent gas anesthesia followed by light anesthesia, after which calcein-labeled washed platelets were transfused via the medial orbital canthus venous plexus. Recipient mice were placed on their side on a clear glass plate and a midline abdominal incision was made. An appropriate arterial vessel was selected, and a 1 mm² filter paper soaked in 6% FeCl₃ was placed on the vessel. A fluorescence microscope was used to observe thrombus formation at the FeCl₃ induced injury site, and the time of thrombus formation was recorded.

The mice were positioned horizontally on the operating table. The distal 5 mm of the mouse tail was transected perpendicular to the workbench, and the tail was gently submerged in the preheated normal saline. The bleeding time from the tail was meticulously recorded.

The standard substances in the ELISA test kit were serially diluted to prepare the standard curve. Standard substances with gradient concentrations were added to the 96-well plate respectively. Each standard substance should be set up in duplicate or triplicate and gently mixed to avoid cross-contamination. Measured the OD value within 15 min after adding the stop solution using an enzyme-linked immunosorbent assay reader. The standard curve was used to calculate the final concentration values.

Added EDTA to the mixture of platelets and albumin at a ratio of 1:500, the supernatant was removed by centrifugation at 2,500 rpm for 2 min, after which the cells were lysed using 1 × cell lysis buffer. Subsequently, 4 × protein loading buffer was added to the lysed platelets at a volume ratio of 1:4. After thorough mixing by vortexing, the samples were heated in a metal bath at 100 °C for 5 min and then stored at −80 °C after cooling to room temperature.

A total of 300 mg albumin were dissolved in 2.25 mL of 0.133 M glycine methyl ester hydrochloride solution (pH 4.75). Subsequently, 750 μL of 0.04 M EDC solution (pH 4.75) was added to the albumin solution. Over the following 30 min, 1 mL aliquots of the reaction mixture were withdrawn every 5 min and immediately quenched with 1 mL of 4 M acetate buffer (pH 4.75) to halt the reaction. The resulting solution was then incubated at room temperature for an additional 30 min. Following this, the solution was transferred to a dialysis cassette and subjected to dialysis against 1 L of PBS at room temperature for 2 h under gentle stirring conditions. The dialysis was subsequently continued overnight at 4 °C. After dialysis, the solution was removed from the cassette and concentrated using an ultrafiltration tube. The isoelectric point of the protein was determined by two-dimensional electrophoresis, and its secondary structure was characterized by circular dichroism spectroscopy.

Statistical analysis of the experimental data was conducted using GraphPad Prism 9.0 software. Differences between groups were assessed by one-way ANOVA and followed by Dunnett’s post hoc test. Statistics are presented as mean ± SEM. *P < 0.05, **P < 0.01 and ***P < 0.001. The two-tailed Student’s t-test was utilized to compare data between two groups. Each experiment was replicated at least three times. All animal experiments were randomized, and no animals or samples were excluded from the study or data analysis.

In clinical medicine, it is observed that patients with portal vein thrombosis exhibit lower plasma albumin levels compared to healthy controls (Zhou and Fang, 2020). To investigate the effect of albumin on platelet activation, we incubated human washed platelets with different concentrations of albumin, then stimulated platelets with agonists and detected the effect of albumin on platelet activation (Figure 1A). The results showed that, albumin significantly inhibited platelet aggregation induced by thrombin (Figure 1B), U46619 (Figure 1C), collagen (Figure 1D) and ADP (Figure 1E). However, albumin had no effect on platelet aggregation induced by Trap-6 (Figure 1F). Based on the results from the platelet aggregation experiment, we selected thrombin and U46619, which exhibited significant inhibitory effects, as well as Trap-6 with no inhibitory effect, for further experimentation. Both thrombin and U46619-induced platelet integrin αIIbβ3 activation were inhibited by albumin in a concentration-dependent manner (Figures 1G,H). However, consistent with the results of platelet aggregation, platelet integrin αIIbβ3 activation induced by Trap-6 also showed no significant differences at each albumin concentration (Figure 1I). Upon platelet activation, P-selectin is rapidly translocated to the platelet surface (Yan et al., 2024). Therefore, we subsequently investigated the release of P-selectin. The results showed that albumin significantly inhibited thrombin (Figure 1J) and U46619 (Figure 1K) induced P-selectin release from human washed platelets. However, Trap-6 induced P-selectin release did not show significant differences in each group (Figure 1L).

When platelets are stimulated, granules can fuse with the platelet open canalicular system (OCS) and be released into the extracellular environment (Chung et al., 2021). We then investigated the effect of albumin on platelet ATP release. Surprisingly, at the concentrations up to 2.5 mg/mL, albumin could significantly inhibited platelet ATP release induced by all agonists, including Trap-6 (Figures 2A–D). The above results indicated that albumin could significantly inhibit platelet granule release. We then used transmission electron microscopy to observe the distribution of granules in platelets (Figures 2E,F). The results showed that resting platelets had rounded edges and contained a large number of granules and OCS in their interior. Activated platelets, on the other hand, formed a large number of pseudopods that aggregated together. Compared to resting platelets, the numbers of α-granules and dense granules were significantly reduced in activated platelets (Figures 2G,H), whereas the number of mitochondria was not different (Figure 2I). Surprisingly, when we incubated albumin with platelets for 15 min before stimulated the platelets with agonists, the number of α-granules and dense granules inside the platelets increased significantly (Figures 2G,H), which was consistent with our observation that albumin had a strong inhibitory effect on the release of ATP from platelets. We also counted the percentages of α-granules (Supplementary Figure S1A), dense granules (Supplementary Figure S1B) and mitochondria (Supplementary Figure S1C), the results were consistent with the statistics of their absolute counts. By morphological observation, we found that platelets incubated with albumin had fewer pseudopods and a reduced spreading area in the 500 nm field of view. Interestingly, by measuring the absolute and relative distance (Figures 2J,K) of the granules from the center of the platelet, we also found that after incubation with albumin, the granules of activated platelets are crowded in the central region of the cell surrounded by the OCS. In the 2 μm field of view, the aggregation of activated platelets incubated with albumin was significantly improved. Stimulation of second-phase platelet aggregation by U46619 requires integrin outside-in signaling (Yan et al., 2024). To investigate whether albumin is involved in this signaling pathway, the effect of albumin on platelet spreading was examined. The results demonstrated that, albumin inhibited platelet spreading in a concentration-dependent manner after stimulation with U46619 (Figure 2L). By marking the cytoskeletal proteins of platelets, it was found that albumin could suppress the spreading area of activated platelets (Figure 2M) and the proportion of spread platelets (Figure 2N) in a concentration-dependent manner.

Cyclic guanosine monophosphate (cGMP) regulates various signaling pathways that inhibit platelet activation, albumin can indirectly regulate cGMP by regulating the duration of NO (Gansevoort et al., 2025). Consequently, this study aims to investigate whether albumin has a direct effect on platelet signaling. Members of the PKC family serve as crucial effector molecules for second messengers such as calcium ions and diacylglycerol, while Akt functions as a serine and threonine kinase (Zhang et al., 2024; Sun et al., 2023). To assess whether albumin can directly regulate PKC and Akt phosphorylation, we activated platelets after co-incubation with albumin and observed the phosphorylation levels of PKC and Akt. The results showed that albumin was able to significantly inhibit U46619 induced phosphorylation of PKC and Akt in a dose-dependent manner (Figures 3A,B). When we used PMA, an activator of PKC, the inhibitory effect of albumin on PKC and Akt phosphorylation was restored (Figures 3C,D). The same phenomenon was observed when we used PDBu, the other activator of PKC (Figures 3E,F). The inhibitory effect of albumin on PKC and Akt phosphorylation was similarly restored when we used 740Y-P, the activator of Akt (Figures 3G,H), but not as dramatically as with the PKC activator. When the other Akt activator, SC79 was used, the results were not different from those obtained with 740Y-P (Figures 3I,J). The aforementioned findings demonstrated that albumin exerted a significant inhibitory effect on PKC and Akt phosphorylation. Furthermore, it appeared that PKC and Akt did not exhibit a clear upstream-downstream relationship in the context of albumin’s inhibition of platelet activation. Subsequently, we incubated platelets with 20 mg/mL albumin, followed by treatment with four kinase activators, and stimulated the platelets with U46619. We then measured platelet aggregation, P-selectin release, and integrin αIIbβ3 activation levels. The results indicated that PKC activators, PMA and PDBu, could significantly reversed the inhibitory effects of albumin on platelet aggregation, P-selectin release, and integrin αIIbβ3 activation. In contrast, Akt activators, 740Y-P and SC79, failed to restore these inhibitory effects (Figures 3K–M).

In the peripheral blood environment, albumin typically exhibits a negative surface charge (Sleep, 2015). Literature also indicates that the internal granules of platelets carry a negative charge (Chung et al., 2021). Based on these background studies and our previous observations, we hypothesized that albumin could suppress platelet activation via its charge properties. We observed that adjusting the pH of the solvent equalized the acidic and alkaline dissociations of the albumin molecules, thereby neutralizing the positive and negative surface charges (Supplementary Figure S2). Consequently, the isoelectric point shifted from 5.0 to 5.4 to 6.8∼7.8 (Figure 4A). Additionally, we analyzed the secondary structure of the neutralized albumin using circular dichroism spectroscopy and found that the structures remained unchanged compared to those of native albumin. These findings indicated that the neutralization of surface charges was achieved without compromising the secondary structure of the albumin. We then tested the effect of charge stripped albumin (CS Albumin) on platelet aggregation and ATP release. The results showed that the inhibitory effect of albumin on platelet aggregation induced by thrombin (Figure 4B), U46619 (Figure 4C), collagen (Figure 4D) and ADP (Figure 4E) could be all reversed by neutralising the surface charge. The inhibitory effects of albumin on platelet ATP release were also reversed (Figures 4F–H). Subsequently, the release and arrangement of platelet granules were examined using transmission electron microscopy (Figure 4I). The results showed that there was no significant difference in the number of platelet granules in the resting state after co-incubation with albumin or CS albumin. However, when platelet activation was induced, the effect of albumin on the number of platelet granules was reversed by charge neutralisation (Figures 4J,K). The absolute counts of mitochondria were not significantly different in any of the groups (Figure 4L). The relative count statistics for the three organelles were consistent with the absolute count results (Supplementary Figures S3A–C). The effect of albumin on the degree of platelet granule discrete was also restored with both the relative distance (Figure 4M) and the absolute distance (Supplementary Figure S3D) of granules from the platelet center. Next, we used a cationic surface activator, a type of quaternary ammonium salt, hexadecyltrimethylammonium chloride (CTAC) for validation (Figure 4N). We directly induced the platelet aggregation with CTAC without using any platelet agonist and determined whether albumin could directly inhibit platelet aggregation induced by the cationic surface activator (Figure 4O). The results demonstrated that the CTAC induced platelet aggregation was progressively inhibited as the albumin concentration increased (Figure 4P). These findings suggested that the surface charge of albumin exerted a significant influence on inhibiting platelet aggregation, suppressing ATP release, preventing granules release, and altering the internal structural organization of platelets.

Following the in vitro demonstration of albumin’s inhibitory effects, we sought to validate these findings in vivo using a hypoproteinemia model. We employed the well-established adriamycin (ADR)-induced acute kidney injury model in BALB/c mice, which reliably induces hypoproteinemia with low mortality and high reproducibility (Li et al., 2020; Yu et al., 2024; Cha et al., 2024). We established a renal hypoproteinemia mouse model by administering adriamycin hydrochloride to BALB/c mice and supplemented the model mice with albumin to create a treatment model for hypoproteinemia (Figure 5A). The results demonstrated that there were no significant differences in platelet count, mean platelet volume (MPV), platelet distribution width (PDW), and plateletcrit (PCT) among WT, ADR, and ADR Albumin mice (Figures 5B–E). ELISA analysis of plasma protein concentrations revealed that compared to WT and ADR Albumin mice, the concentration of plasma albumin in ADR mice was significantly reduced (Figure 5F), while the total plasma protein (TP) level remained unaffected (Figure 5G). Notably, the concentration of plasma prealbumin (PA) was markedly decreased in ADR mice (Figure 5H). Additionally, the globulin levels in ADR Albumin mice were slightly elevated compared to those in WT and ADR mice (Figure 5I). Histological examination of mouse kidney tissues using HE, Masson, and PAS staining indicated that, relative to WT mice, the kidneys of ADR mice exhibited moderate glomerular sclerosis, mesenchymal fibrosis, thickening of the glomerular basement membrane, and renal matrix hyperplasia. Cellular crescent formation was also observed within the glomerular sac lumen (Figures 5J,K). In conclusion, the mouse model of adriamycin-induced acute kidney injury leading to hypoproteinaemia had been successfully established.

To directly assess the in vivo state of platelet activation in this model, we analyzed the functional responsiveness of isolated platelets. We found that hypoproteinemic (ADR) mice exhibited a significantly heightened platelet activation profile. The results demonstrated that, compared with WT mice, ADR mice exhibited significantly elevated levels of platelet aggregation (Figures 6A–E), integrin αIIbβ3 activation (Figures 6F–H), P-selectin release (Figures 6I–K), ATP release (Figures 6L–O), platelet spreading (Figures 6P–R). This comprehensive hyper-reactivity suggested that platelets circulating in ADR mice existed in a primed or pre-activated state. Crucially, supplementation with human albumin (ADR + Alb group) restored these functional parameters to levels indistinguishable from WT controls. In addition to establishing a renal hypoproteinemia model, we utilized CCl₄ to induce chemical liver injury and construct a hepatogenic hypoproteinemia model (Supplementary Figure S4A). However, this modeling process was time-consuming, and the use of corn oil as a solvent for CCl₄ could potentially cause hepatic steatosis. Notably, platelet aggregation did not differ significantly among WT, Oil, and CCl₄ groups (Supplementary Figures S4B–F).

The observed increase in platelet activation biomarkers,we tested this directly using two complementary in vivo assays (Figure 7A). The results indicated that the tail bleeding time in ADR mice was significantly shorter compared to WT mice, while the tail bleeding time in ADR Albumin mice was prolonged (Figure 7B). Platelets from these mice were isolated, labeled with calcein, and transfused into 4 weeks WT mice via the orbital venous plexus. A suitable mesenteric artery was selected and placed flat on a glass slide moistened with 0.9% saline. The artery was injured using filter paper soaked in 6% FeCl₃, and thrombus formation at the injury site was observed. The findings revealed that, compared to WT mice, the time to platelet thrombosis in ADR mice was significantly reduced, whereas the thrombosis time in ADR Albumin mice was significantly prolonged (Figures 7C–F). Based on the aforementioned experimental results, using mouse models, we can conclude that: (1) Hypoproteinemia causes circulating platelets to enter an over - reactive state. This was verified by extracting single - cell suspensions of mouse platelets for comprehensive in vitro functional experiments. (2) Hypoproteinemia results in accelerated thrombosis in vivo. These results provided direct in vivo evidence that hypoproteinemia promoted a pro-thrombotic state characterized by elevated platelet activation biomarkers and enhanced thrombus formation, both of which were effectively counteracted by albumin replenishment.