Introduction

Acute liver injury (ALI) is a life-threatening clinical syndrome characterized by rapid hepatocyte death, sterile inflammation, and loss of hepatic function, yet its cellular drivers remain incompletely defined (1, 2). ALI arises from diverse etiologies, including infections, toxins, and autoimmune disorders, with acetaminophen (APAP) overdose representing one of the most common causes (1, 2). A consistent clinical feature of patients with severe ALI is thrombocytopenia, manifested as a marked reduction in circulating platelet counts (3, 4). In our previous study, we observed that these “lost” platelets accumulate within the liver, suggesting a previously unappreciated and potentially pathogenic role for platelets in ALI (5).

Platelets are anucleate blood components classically known for their role in hemostasis, but they are increasingly recognized as active modulators of inflammatory and immune responses (6–11). Nevertheless, technical challenges in platelet isolation, culture, and in vivo lineage tracing have limited mechanistic studies of platelet function in many diseases, including acute liver injury. Although platelet involvement in liver pathology has been suggested (5, 12), the precise molecular mechanisms by which platelets exert their effects remain poorly understood. Defining these pathways is essential for the development of targeted platelet-based therapeutic strategies.

Kupffer cells (KCs), the liver-resident macrophages strategically positioned within the hepatic sinusoids, serve as central orchestrators of hepatic immune responses (13, 14). Following liver injury, KCs integrate damage-associated signals and rapidly adopt dynamic functional states that can either exacerbate hepatocellular damage or promote tissue repair (15, 16). How KC fate decisions are regulated within the injured hepatic microenvironment remains a key unresolved question. A defining feature of inflammatory macrophage activation is metabolic reprogramming, in which a shift toward aerobic glycolysis, reminiscent of the “Warburg effect,” fuels biosynthetic and pro-inflammatory programs in classically activated macrophages (17, 18). Whether a similar metabolic rewiring occurs in KCs during ALI, and how such changes might be imposed by non-immune cells such as platelets, has not been systematically explored.

Building on our previous discovery that platelets infiltrate the liver and physically interact with KCs (5), we hypothesized that platelet–KC crosstalk exacerbates ALI by reshaping KC function. In this study, we sought to define the pathological consequences of platelet–KC interactions and to identify the critical molecular mediators involved. We demonstrate that platelets transfer the glycolytic enzyme aldolase A (ALDOA) to KCs via extracellular vesicles (EVs), thereby driving a glycolytic surge that amplifies liver injury. Using platelet-specific Aldoa knockout mice, we establish the essential in vivo role of platelet-derived ALDOA. Moreover, pharmacological inhibition of ALDOA confers significant protection from liver injury, and circulating ALDOA levels are elevated in patients with ALI, underscoring the clinical relevance of our findings. Together, these results identify a novel platelet–KC metabolic axis in ALI pathogenesis and nominate ALDOA as a promising diagnostic biomarker and therapeutic target.

Results

Recruited platelets enhance glycolytic program in Kupffer cells

To investigate the role of platelets in APAP-induced liver injury (AILI), we established an AILI mouse model using wild-type C57BL/6J mice. Mice treated with APAP for 12 hours exhibited a significant increase in serum alanine aminotransferase (ALT) levels and extensive hepatic necrosis compared to PBS-treated controls, confirming the successful induction of the AILI model (Figure S1A, B). Immunofluorescence staining revealed a substantial infiltration of platelets into the liver following APAP injury (Figure S1C). Additionally, co-immunofluorescence staining of KCs and platelets revealed colocalization between these cell types, suggesting their interaction within the liver during AILI (Figure S1D). These findings are consistent with our previous study (5) and demonstrated that platelets are recruited to the liver and interact with KCs during AILI.

We next sought to characterize the response of KCs to APAP injury and determine whether these changes are influenced by platelets interactions. First, we isolated KCs from mice treated with APAP for 0, 6, and 12 hours and performed RNA sequencing (RNA-seq). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed a notable upregulation in glycolysis/gluconeogenesis pathways in KCs from mice treated with APAP for 12 hours compared to 0 hours, indicating a metabolic shift in KCs upon APAP treatment (Figure 1A). Further comparison of glycolysis and oxidative phosphorylation (Oxphos)-related enzymes expression in KCs treated with APAP for 0, 6, and 12 hours demonstrated a time-dependent upregulation in glycolysis-related gene expression but not in Oxphos-related gene expression (Figure 1B and Table S1). The upregulation of glycolysis-related genes (Slc2a1, Pfkfb3, Pkm) expression at 12 hours post-APAP treatment was further confirmed by qRT-PCR (Figure S1E). These gene expression data suggest a progressive increase in glycolytic activity in KCs in response to APAP. Hypoxia-inducible factor 1 alpha (HIF-1α) is a transcription factor that drives glycolysis and amplifies inflammatory responses under hypoxic conditions (19), while interleukin-1 beta (IL-1β) is a pro-inflammatory cytokine typically upregulated in proinflammatory macrophages (19). The results revealed a significant increase in the expression of these genes in KCs in response to APAP (Figure S1E). This suggests that KCs undergo a glycolytic switch during early AILI.

Figure 1.

To determine if the recruited platelets were responsible for this metabolic reprogramming, we depleted platelets using an anti-CD41 antibody (α-CD41) prior to APAP challenge (Figure S2A). Immunofluorescence confirmed effective platelet depletion in the liver (Figure S2B). Consistent with previous studies (5, 12), platelet depletion reduced liver injury, as indicated by a decrease in necrotic area and lower serum ALT levels (Figure S2C and D). To further explore the effects of platelet depletion on the KC glycolytic program, we isolated KCs from both platelet-depleted and control mice for RNA-seq analysis. Gene Set Variation Analysis (GSVA) of metabolic pathways in KCs from APAP-injured mice revealed that platelet depletion significantly downregulated the glycolysis pathway (p=0.038) (Figure 1C). Consistent with this, a heatmap showed that the expression of glycolysis-related genes induced by APAP was suppressed in KCs from platelet-depleted mice, while OXPHOS-related genes were restored (Figure 1D). qPCR analysis further validated that the mRNA expression of glycolysis- and inflammation-related genes in KCs was significantly reduced upon platelet depletion (Figure S2E). This highlights the critical role of platelets in driving glycolytic pathways in KCs. Collectively, these results demonstrate that platelet depletion prevents the metabolic switch in KCs during acute liver injury, underscoring the pivotal role of platelets in regulating the metabolic and inflammatory responses of KCs.

Finally, to establish a direct causal link, we isolated mouse platelets and co-cultured them with isolated KCs in vitro. We measured the extracellular acidification rate (ECAR) in KCs cultured with either conditional medium (Vehicle) or isolated mouse platelets (PLT). Consistent with the observations in KCs in vivo, KCs treated with platelets in vitro exhibited enhanced glycolytic activity (Figure 1E). Additionally, basal and maximal respiration rate was significantly increased in platelets-treated KCs, indicating elevated glycolytic activity (Figure 1F). This suggests that platelets augment glycolytic activity in KCs in vitro. Collectively, these findings demonstrate that platelets drive a metabolic reprogramming in KCs, characterized by enhanced glycolysis.

Kupffer cell glycolysis promotes APAP-induced liver injury

Having established that platelet-induced glycolysis in KCs is a key feature of the injury response, we next sought to determine whether this metabolic reprogramming functionally contributes to liver damage. To directly test this, we designed an experimental strategy to inhibit glycolysis specifically in KCs during APAP challenge using 2-deoxy-D-glucose (2-DG), a glycolytic inhibitor (20) (Figure 2A). To confirm the specificity of the effect to KCs, we included a group where KCs were depleted using clodronate liposomes (CLDN). We first validated the efficacy of our KC depletion protocol. Flow cytometric analysis of non-parenchymal liver cells confirmed that pretreatment with clodronate liposomes, but not control liposomes, effectively ablated the CD11bˡᵒʷ F4/80^hi Kupffer cell population (Figure S3). We then assessed the functional consequences of KC depletion and glycolytic inhibition. In KC-intact mice, administration of 2-DG alongside APAP challenge led to a significant reduction in both hepatic necrotic area and serum ALT levels compared to the PBS-treated APAP control (Figure 2B, C). Crucially, this protective effect of 2-DG was entirely absent in mice that had been depleted of KCs, demonstrating that the target of 2-DG’s action is the Kupffer cell itself (Figure 2B, C). As expected, KC depletion alone was also protective against APAP injury. Finally, we analyzed the inflammatory state of the isolated KCs. qPCR analysis revealed that the mRNA expression of key inflammation-related genes, which were highly upregulated following APAP challenge, was significantly suppressed by 2-DG treatment (Figure 2D). This indicates that inhibiting the glycolytic program in KCs not only mitigates tissue damage but also directly quells their pro-inflammatory response. Together, these results demonstrate that the platelet-enhanced glycolytic program in Kupffer cells is not merely a correlative event but is a critical driver of the inflammatory response and subsequent liver injury following APAP overdose.

Figure 2.

Platelets-derived extracellular vesicles promote glycolysis in Kupffer cells

We next sought to identify the mechanism by which platelets communicate with Kupffer cells to drive glycolysis. We collected supernatant from isolated mouse platelets after a 2-hour culture and co-cultured KCs with either conditional medium (Vehicle) or platelet supernatant (PLT sup) for 2 hours. KCs exposed to platelet supernatant exhibited increased expression of glycolysis-related and inflammatory response-related genes compared to the vehicle controls (Figure 3A), suggesting that platelets secrete soluble factors capable of inducing a metabolic switch in KCs.

Figure 3.

Previous studies have shown that platelets communicate with other cells through the release of EVs (21, 22). To investigate this further, we isolated platelet-derived extracellular vesicles (PEVs) and confirmed their identity by detecting established EV markers, including Tumor Susceptibility Gene 101 (TSG101) and Cluster of Differentiation 63 (CD63). High expression of both TSG101 and CD63 in the isolated PEVs confirmed their identity (Figure 3B). Transmission electron microscopy (TEM) analysis revealed that PEVs were a spherical vesicle with a bilayer membrane structure (Figure 3C), and the nanoparticle tracking analysis (NTA) detection results show that the particle size of PEVs were mainly distributed in the range of 59–288 nm, with the peak at 105 nm (Figure 3D).

To determine if these PEVs could directly modulate KC metabolism, we performed a Seahorse bioenergetic analysis. Kupffer cells co-cultured with isolated PEVs for just 2 hours exhibited a significantly elevated ECAR, indicating a robust increase in glycolytic flux (Figure 3E). Moreover, PEV treatment also enhanced the basal and maximal respiratory capacity of the KCs (Figure 3F), confirming PEV’s effect on KC glycolysis.

Platelet-derived extracellular vesicles promote APAP-induced liver injury

Having established that PEVs can drive glycolysis in KCs in vitro, we investigated their role in vivo. We utilized a rescue approach in platelet-depleted mice (Figure 4A). Immunofluorescence confirmed the efficient depletion of platelets in the livers of all mice treated with the anti-CD41 antibody (Figure 4B). As expected, platelet-depleted mice were protected from AILI, showing minimal necrosis and low ALT levels. However, administration of isolated PEVs to these platelet-depleted mice restored the liver injury phenotype, as evidenced by significantly increased hepatic necrotic area and serum ALT levels (Figure 4C, D).

Figure 4.

Finally, to confirm that PEVs directly target KCs in vivo, we intravenously injected PKH67-labeled PEVs and analyzed their localization. Immunofluorescence staining revealed that a significant number of F4/80-positive KCs had taken up the fluorescently-labeled PEVs within 3 hours of administration (Figure 4E). Collectively, these data identify PEVs as a critical communicator between platelets and KCs. PEVs are efficiently taken up by KCs, where they drive a potent glycolytic and inflammatory response that is essential to promote AILI.

ALDOA from PEVs induces a glycolytic switch in Kupffer cells

To identify the specific component within PEVs responsible for driving glycolysis in KCs, we isolated PEVs and performed a quantitative mass spectrometry analysis of their protein cargo, which identified 121 proteins (Figure 5A, Table S2). To refine this list and select for proteins specifically enriched in PEVs, we performed a comparative analysis with EVs derived from AML12 hepatocytes (Table S3). We excluded 53 proteins that were common to both PEVs and hepatocyte-derived EVs (HEVs), as well as cytoskeletal and coagulation-related proteins. This stringent filtering yielded a shortlist of seven candidate proteins: Aldolase A (ALDOA), Bridging integrator 2 (BIN2), Carboxylesterase 1C (CES1C), Inter-alpha-trypsin inhibitor heavy chain 2 (ITIH2), Murinoglobulin 1 (MUG1), Pleckstrin (PLEK), and Thrombospondin 1 (THBS1) (Figure 5A). Based on its central role as a catalytic enzyme in the glycolysis pathway, we selected ALDOA for further investigation. Western blot analysis confirmed the abundant presence of ALDOA within our preparations of isolated PEVs (Figure 5B). To test its functional capacity, we treated KCs with recombinant ALDOA (rALDOA) protein. Seahorse metabolic analysis revealed that rALDOA potently enhanced the glycolytic capacity of KCs, as shown by a marked increase in the ECAR (Figure 5C), as well as basal and maximal respiration rate (Figure 5D).

Figure 5.

To establish the in vivo relevance of ALDOA during AILI, we measured its serum levels over time in AILI mouse models. Western blot analysis revealed a dynamic increase in circulating ALDOA, which was detectable as early as 3 hours post-APAP treatment and remained elevated at 6 and 12 hours (Figure 5E). To determine if platelets were the source of this serum ALDOA, we depleted platelets in mice prior to APAP challenge. Platelet depletion significantly reduced serum ALDOA levels compared to the control IgG-treated group (Figure 5F), confirming that platelets are the primary contributors to the elevated ALDOA observed during injury. Together, these findings demonstrate that ALDOA is a specific and abundant protein cargo of PEVs. Upon delivery to KCs, ALDOA induces a profound glycolytic switch in KCs.

ALDOA serve as a potential therapeutic target for acute liver injury

To definitively establish the pathophysiological role of platelet-derived ALDOA, we generated platelet-specific Aldoa knockout mice (referred to as PF4^ΔAldoa mice), with successful gene deletion confirmed by genotyping and western blot of platelet lysates (Figure S4A, B). Following APAP challenge, PF4^ΔAldoamice exhibited a significant attenuation of liver injury, as demonstrated by markedly reduced serum ALT levels and a smaller hepatic necrotic area compared to their Aldoa^fl/fl littermate controls (Figure 6A, B). To rule out the possibility that this protection was due to altered APAP bioactivation, we measured the levels of the metabolic enzyme CYP2E1 and the toxic metabolite NAPQI (23). We found no significant differences in hepatic CYP2E1 protein levels or NAPQI-protein adduct formation between the genotypes (Figure S5A), indicating that the protective effect was independent of the initial metabolic insult. Instead, KCs isolated from APAP-injured PF4^ΔAldoa mice showed a significant downregulation of glycolysis- and inflammation-related genes (Figure S5B), confirming that platelet-derived ALDOA is critical for driving the pathogenic metabolic and inflammatory reprogramming of KCs in vivo.

Figure 6.

We next investigated the therapeutic potential of pharmacologically inhibiting ALDOA after the initiation of injury. Administration of the ALDOA inhibitor Aldometanib (24) one hour after APAP overdose (Figure 6C) significantly protected against liver injury, as evidenced by lower serum ALT levels and reduced hepatic necrosis (Figure 6D, E). Again, this protection was not due to impaired APAP metabolism, as CYP2E1 and NAPQI-adduct levels were unaffected by Aldometanib treatment (Figure S6A). To assess the broader applicability of targeting this pathway, we tested its effect in a Concanavalin A (Con A)-induced immune-mediated liver injury model. Similar to APAP injury, KC depletion protected against Con A-induced hepatitis (Figure S7A, B). Importantly, administration of Aldometanib also significantly alleviated Con A-induced liver injury (Figure S7C-E), demonstrating that the ALDOA-driven pathogenic mechanism extends beyond AILI to immune-mediated hepatitis.

Finally, we sought to validate the clinical relevance of our findings in human disease. Re-analysis of a published proteomic dataset confirmed that ALDOA is a constituent of human platelet alpha-granules (25) (Figure 6F). Furthermore, analysis of EVs from patient serum revealed that EV-associated ALDOA levels were significantly higher in patients with polycythemia vera (26), a condition with hepatic involvement, compared to patients with Alzheimer’s Disease (27) (Figure 6G). Most critically, in a cohort of patients with ALI, we found a strong and significant positive correlation between circulating ALDOA protein levels and serum ALT, a key clinical marker of hepatocyte damage (Figure 6H).

Collectively, these data demonstrate that genetic ablation or pharmacological inhibition of ALDOA is protective in multiple models of acute liver injury, operating through a mechanism that disrupt the secondary, inflammation-driven amplification of damage without affecting the initial insult. The presence and correlation of ALDOA with disease severity in human patients underscore its significant translational potential as a biomarker and a novel therapeutic target for acute liver injury.

Discussion

The progression of AILI is a complex process involving not only the initial hepatocyte necrosis caused by NAPQI but also a robust sterile inflammatory response that amplifies the damage (28). While the role of immune cells like KCs has been acknowledged, the specific signals that activate them remain incompletely understood. In this study, we elucidate a novel and critical axis in which recruited platelets, via the release of EVs loaded with the glycolytic enzyme ALDOA, drive a pathogenic metabolic and inflammatory reprogramming of KCs, ultimately exacerbating liver injury. Furthermore, we identify ALDOA as a viable therapeutic target for ALI (Figure 7).

Figure 7.

Our findings first confirm and extend previous observations that platelets are rapidly recruited to the liver following APAP overdose and physically interact with KCs (5). More importantly, we demonstrate that this interaction is not passive but functionally reprograms KCs metabolism. Through transcriptomic and functional analyses, we provide compelling evidence that KCs undergo a metabolic switch towards glycolysis during AILI, a shift that is entirely dependent on the presence of platelets. This platelet-induced glycolytic reprogramming is a key driver of pathology, as pharmacological inhibition of glycolysis within KCs significantly mitigated liver injury. This aligns with the established concept that a glycolytic switch in macrophages promotes a pro-inflammatory phenotype (29, 30), and our data place platelets as the upstream trigger for this switch in the context of sterile liver injury.

The central mechanistic discovery of our study is the identification of PEVs as the primary communicators in this cell-cell interaction. We show that PEVs are both necessary and sufficient to transmit the glycolytic signal to KCs. Adoptive transfer of isolated PEVs fully restored liver injury in platelet-depleted mice, providing direct evidence for their pathogenic role. This finding adds a new dimension to the understanding of platelet function in sterile inflammation, positioning them not only as coordinators of coagulation and inflammation but also as active regulators of cellular metabolism in recipient cells via vesicular communication (21, 22).

Our proteomic and functional investigations pinpointed ALDOA as the key protein cargo within PEVs responsible for inducing glycolysis in KCs. This is a particularly intriguing finding. While ALDOA is a canonical intracellular enzyme in glycolysis, its presence in the extracellular space and within EVs suggests a non-canonical, moonlighting function. We demonstrated that extracellular ALDOA alone can potently enhance glycolytic flux and the expression of inflammatory genes in KCs. The increase in circulating ALDOA during AILI, which originates predominantly from platelets, underscores its role as a damage-associated signal. Since PEVs can be phagocytosed by KCs in vivo, extracellular ALDOA is more likely internalized into KCs via EVs and influenced glycolysis pathway as a glycolytic enzyme.

The translational potential of targeting this pathway is strongly supported by our genetic and pharmacological studies. The protection observed in platelet-specific Aldoa knockout mice (PF4^ΔAldoa) confirms the in vivo relevance of our mechanistic model. Crucially, the protective effect of the ALDOA inhibitor Aldometanib, even when administered after APAP overdose, highlights a promising therapeutic window for intervention that targets the secondary inflammatory wave rather than the initial metabolic insult. This is a significant advantage, as it circumvents the narrow timeframe associated with N-acetylcysteine therapy (31). The efficacy of Aldometanib in both APAP and Con A-induced liver injury models suggests that the ALDOA-driven pathway is a common mechanism in diverse forms of ALI, potentially broadening its therapeutic applicability.

Finally, the clinical relevance of our findings is underscored by the analysis of human samples. The presence of ALDOA in human platelet granules and its elevated levels in serum EVs from patients with liver-involving diseases like polycythemia vera provide a crucial bridge from mouse models to human pathology. Most importantly, the strong positive correlation between circulating ALDOA levels and serum ALT in patients with acute liver injury positions ALDOA not only as a therapeutic target but also as a potential prognostic biomarker for disease severity.

In conclusion, this study unveils a previously unrecognized pathway in the pathogenesis of ALI: Platelets → PEVs → ALDOA → KC Glycolytic Switch → Inflammatory Amplification → Hepatocyte Necrosis. By identifying ALDOA as a critical effector of this axis, our work opens up new avenues for therapeutic intervention. Targeting the ALDOA-mediated metabolic communication between platelets and KCs could represent a novel and effective strategy for treating ALI in patients.

Materials and Methods

Human sera

Serum samples from healthy volunteers and patients diagnosed with acute liver injury were collected at the Affiliated Hospital of Yunnan University (AHYNU), Kunming, China. The study was designed and carried out in accordance with the principles of AHYNU and approved by the Ethics Committee of AHYNU (Approval number 2024351). All patients provided written informed consent. Serum ALDOA levels in acute liver injury patients were quantified by western blot analysis, using 2 ng of recombinant ALDOA (rALDOA) as a standard, and band intensities were analyzed with ImageJ software.

Animal experiments and procedures

C57BL/6J mice (Strain No. N000013) were purchased from GemPharmatech. Aldoa^fl/fl (Strain No. NM-KO-226330) were purchased from Southern Model Biotechnology Co., Ltd. To achieve platelet-specific deletion, Aldoa^fl/fl mice were crossed with PF4-cre transgenic mice (Jackson Laboratory, Stock #008535), in which Cre recombinase expression is driven by the platelet factor 4 promoter. PF4-Cre; Aldoa ^fl/fl (PF4^ΔAldoa) mice and their Aldoa^fl/fl littermates were used as controls. Genotyping was performed by PCR on genomic DNA extracted from tail clips using the following primers: Aldoa ^fl/fl: F-GTTAGGCAAGCGCTCTGTTG, R-GTCAGCTTTATGGCCCCTGT. PF4-Cre: F1-CCAAGTCCTACTGTTTCTCACTC, R1-TGCACAGTCAGCAGGTT, F2-CTAGGCCACAGAATTGAAAGATCT, R2-GTAGGTGGAAATTCTAGCATCATCC. All the primers are from 5’end to 3’ end.

All mouse colonies were maintained at the Laboratory Animal Center of Yunnan university. Animal studies described have been approved by the Yunnan University Institutional Animal Care and Use Committee (IACUC, Approval No. YNU20220262 for C57BL/6J strain, Approval No. YNU20240898 for Aldoa^fl/fl strain, Approval No. YNU20220387 for PF4-Cre strain). For APAP treatment, mice (8-12 weeks old) were fasted overnight (5:00pm to 9:00am) before i.p. injected with APAP (Sigma, Cat #A7085) at a dose of 210 mg/kg for male mice. Male mice have been the choice in the vast majority of the studies of AILI reported in the literature (28, 32). Therefore, we used male mice in the majority of the experiments presented. However, we observed a similar phenotype in female mice as in male mice regarding the interaction between KCs and platelets(5). For Concanavalin A treatment, mice were i.v. injected with Saline or Concanavalin A (Sigma, Cat #C2010) at a dose of 15mg/kg. Liver paraffin sections and sera were harvested at time points indicated in the figure legends. H&E staining and ALT measurement (Teco Diagnostics, Cat #A524-150) were performed to examine liver injury.

Platelet depletion: WT mice were i.v. injected with anti-IgG (BD Pharmingen, Cat #553922, 2mg/kg) or anti-CD41 antibody (BD Pharmingen, Cat #553847, 2mg/kg) to deplete platelets at 12 hours before APAP treatment, followed by 210mg/kg APAP i.p. treatment for additional another 12 hours.

Kupffer cells depletion: WT mice were i.v. injected with 100 μL Clodronate liposomes (CLDN) to deplete Kupffer cells 9 hours before APAP or Con A treatment, followed by 210mg/kg APAP i.p. treatment or Con A treatment.

Inhibition of glycolysis: WT mice were i.p. injected with 250mg/kg 2-DG (Sigma, Cat# 8375, dissolved with PBS at a concentration of 50mg/mL) half an hour before APAP treatment, followed by 210mg/kg APAP i.p. treatment for additional another 12 hours.

Blocking endogenous ALDOA: WT mice were i.p. injected with 2mg/kg Aldometanib (MCE, Cat# HY-148189, dissolved with N-Methyl-2-pyrrolidone (Sangon Biotech, Cat# 872-50-4) at a stock concentration of 50mg/mL, diluted with PBS) one hour after APAP treatment, liver paraffin sections and sera were harvested at 12 hours after APAP treatment.

PEVs labeling and injection: PEVs were isolated from mouse platelets culture supernatant by ultracentrifuge. And stained with PKH67 Green Fluorescent (sigma, Cat #MINI67) in 37°Cfor 30min. PEVs were washed with PBS for 1 time. And then collects PEVs by ultracentrifuge. WT mice were i.v. injected with anti-IgG (BD Pharmingen, Cat #553922, 2mg/kg) or anti-CD41 antibody (BD Pharmingen, Cat #553847, 2mg/kg) to deplete platelets at 12 hours before APAP treatment, followed by 210mg/kg APAP i.p. treatment and 20 μg PEVs (per mouse) i.v. treatment at the same time for additional another 3 hours.

Preparation of Kupffer cells

Wild-type C57BL/6J mice were anesthetized using a CO₂ chamber. To avoid contamination of the peritoneal cavity with fur, the skin and peritoneum were opened separately. The inferior vena cava was clamped with forceps, and a catheter was inserted into the superior vena cava via the right atrium. The portal vein was then cut once the buffer started to perfuse into the liver. The liver was perfused with 1× Hank’s Balanced Salt Solution (HBSS) containing 53.6 mM KCl, 4.4 mM KH₂PO₄, 1367.5 mM NaCl, 3.36 mM NaHPO₄, 55.5 mM D-Glucose, and 100 mL 1 M HEPES (pH 7.2-7.4) for 3-5 minutes until it appeared clear. Subsequently, the buffer was switched to a collagenase digestion solution (1×HBSS with 1.25 mM MgSO₄, 4 mM CaCl₂, Collagenase Type I, and 50 μg/mL DNase) for 15-20 minutes of digestion. Post-digestion, the liver was dissected and placed in chilled 1×HBSS in a Petri dish to keep it cool until further processing. The gallbladder was removed, and the liver was chopped into 2-3 mm pieces and incubated in 10-20 mL digestion buffer at 37°C for 15 minutes. The liver tissue was then mashed using a 70 μm cell strainer placed over a 50 mL tube and rinsed with 1×HBSS to a final volume of 50 mL. The suspension was centrifuged at 50×g for 5 minutes (acceleration/brake = 0) to remove hepatocytes, with the resulting supernatant containing nonparenchymal cells (NPCs). The supernatant was further centrifuged at 450xg for 5 minutes (acceleration/brake ≥ 5), and the pellet, containing hepatic NPCs, was resuspended in 4 mL of 20% OptiPrep solution diluted with 1×HBSS. This suspension was carefully layered with 5 mL of 1×HBSS buffer in a 15 mL tube and centrifuged at 3000 rpm for 17 minutes (acceleration/brake = 0). The cell fraction between the 20% OptiPrep and 1xHBSS layers, containing KCs and liver endothelial sinusoidal cells (LESCs), was collected and washed with 1×HBSS.To isolate KCs, the cell fraction was incubated in a cell culture dish at 37°C for 10 minutes. KCs adhered to the plate, while LESCs remained in suspension. The adhered KCs were then collected for subsequent experiments. The purity and viability of isolated KCs reached 90% and 80%, respectively, which was confirmed by staining with anti-Timd4 antibodies or trypan blue.

Kupffer cells cultivated in vitro

KCs culture medium consisted of DMEM (VivaCell, Cat #C3113-0500) supplemented with 10% (vol/vol) fetal bovine serum (FBS, VivaCell, Cat #C04001-500), 100 IU penicillin, and 100 μg/mL streptomycin. KCs were cultured for 2 hours. For experiments involving KCs, 1 × 10^⁶ cells seeded in one well of a 12-well plate were co-cultured with either 0.5 × 10^⁶ isolated mouse platelets or vesicles isolated from 1 × 10^⁶ isolated mouse platelets for 2 hours. After co-culture, the cells underwent various experiments including Seahorse assay, and RT-PCR. In another study, KCs were treated with platelet supernatant for 2 hours. Platelet supernatant was collected from isolated mouse platelets after a 2-hour culture. In additional experiments, KCs were treated with recombinant ALDOA for 12 hours. In some experiments, cells were treated with either vehicle or 5 μg/mL platelet-derived extracellular vesicles (PEVs) for 2 hours.

Isolation of mouse platelets

Blood was collected from CO₂-sacrificed mice via cardiac puncture using a 1/10 volume of ACD buffer (78 mM citric acid, 97 mM sodium citrate, 111 mM D-glucose, pH 7.4) as an anticoagulant. The collected blood was added to a 1.5 mL tube containing 300 μL of HT buffer (129 mM NaCl, 20 mM HEPES, 12 mM NaHCO₃, 2.9 mM KCl, 1 mM MgCl₂, 0.34 mM Na₂HPO₄, and 0.045 g of glucose per 50 mL of buffer, pH 7.3). Following centrifugation at 250×g for 3 minutes at 25°C, the supernatant (platelet-rich plasma) along with the top one-third of the red cell layer was transferred to a new 1.5 mL tube. This mixture was centrifuged again at 250g for 3 minutes, and the supernatant was collected into another tube. An additional 200 μL of HT buffer was added to the previous tube, which was then centrifuged at 250×g for 3 minutes, and the resulting supernatant was pooled with the previous collection, resulting in approximately 600 μL of platelet-rich plasma. For functional studies, 50 ng/mL of PGI₂ (Sigma, Cat #P6188, 1 mg/mL in TBS, pH 7.3) was added to the platelet-rich plasma and incubated for 30 minutes at 37°C. The mixture was then centrifuged at 2000×g for 5 minutes to pellet the platelets. All centrifugation steps were performed with slow acceleration and deceleration settings.

Preparation of platelet-derived extracellular vehicles (PEVs)

Platelets were cultured in RPMI-1640 medium (VivaCell, Cat# C3001-0500) containing 10% exosome-depleted FBS, 100 IU penicillin, and 100 μg/mL streptomycin at 37°C for 2 hours. The culture medium was then collected and centrifuged at 500 g for 10 minutes at 4°C to remove cells. The supernatant was subsequently centrifuged at 2,000×g for 15 minutes at 4°C to remove cell debris. This process was followed by another centrifugation at 10,000×g for 15 minutes at 4°C to eliminate small cell debris. The remaining supernatant was used to isolate platelet-derived extracellular vesicles (PEVs) by ultracentrifugation at 30,000 rpm for 90 minutes at 4°C. The obtained PEV pellets were resuspended in an appropriate buffer. PEV-check antibodies TSG101 and CD63 were used to confirm marker expression profiles.

Transmission Electron Microscopy (TEM)

Isolated PEVs were resuspended in PBS. A 10 μL droplet of the PEV suspension was applied to a formvar-carbon coated copper grid for 1 minutes. The grid was then negatively stained with 2% uranyl acetate for 30 second, air-dried, and visualized using a Hitachi HT7800 transmission electron microscope operating at 120 kV.

Nanoparticle Tracking Analysis (NTA)

The size distribution and concentration of PEVs were analyzed using a ZetaView (Particle Metrix). PEVs were diluted in sterile PBS to achieve an optimal particle concentration for analysis. Three videos of 60 seconds each were captured for each sample, and the data were processed using ZetaView software (version 9.2).

In vitro cell culture

For experiments involving KCs, 1×10^⁶ cells seeded in one well of a 12-well plate were co-cultured with either 0.5×10^⁶ isolated mouse platelets or vesicles isolated from 1×10^⁶ isolated mouse platelets for 2 hours. After co-culture, the cells underwent various experiments including Seahorse assay, and RT-PCR. In another study, KCs were treated with platelet supernatant for 2 hours. Platelet supernatant was collected from isolated mouse platelets after a 2-hour culture. In additional experiments, KCs were treated with recombinant ALDOA for 12 hours. In some experiments, cells were treated with either vehicle or 5 μg/mL PEVs for 2 hours.

Sample preparation for flow cytometry analysis

Hepatic nonparenchymal cells (NPCs) were isolated as described previously(5). Briefly, mice were anesthetized and perfused according to the procedure for isolating Kupffer cells. After removing hepatocytes, the remaining cell suspension was centrifuged at 450 × g for 5 minutes to collect NPCs. The pellets were resuspended in 10 mL of 35% Percoll (Cytiva, Cat #17089109) and transferred to a new 15 mL tube. This was followed by centrifugation at 450 x g for 15 minutes to remove dead cells. Isolated liver NPCs were incubated with 1 μL of anti-mouse FcγRII/III (Invitrogen, Cat #14-0161-86) in FACS buffer (PBS + 0.2% FBS) to minimize non-specific antibody binding. The cells were then stained with anti-mouse CD45-APC 780 (Invitrogen, Cat #47-0451-82), F4/80-PE (Invitrogen, Cat #12-4801-82), and CD11b-660 (Invitrogen, Cat #50-0112-82) for 30 minutes on ice. All antibodies were diluted 1:100. Flow cytometry was performed on an LSR Fortessa Cell Analyzer from BD Biosciences, and the data were analyzed using FlowJo software V10.

Histology and immunofluorescence staining

H&E Staining: Hematoxylin and eosin (H&E) staining was performed on paraffin-embedded liver sections. Briefly, liver tissues were harvested and fixed in 10% formalin for overnight, then dehydrated gradient alcohol. The tissues were subsequently embedded in paraffin and sectioned. H&E staining was performed using Hematoxylin (Biocare Medical, CATHE-M) and Edgar Degas Eosin (Biocare Medical, THE-MM). The necrosis areas were calculated using ImageJ.

Immunofluorescent Staining: Immunofluorescent staining was conducted on frozen liver sections. Fresh liver tissues were harvested and immediately fixed in 2% paraformaldehyde for 1 hour, washed with PBS three times, dehydrated in 30% sucrose at 4°C overnight, and finally embedded in OCT. Frozen sections (5 μm) were prepared and fixed with precooled acetone for 15 minutes at -20°C. The sections were permeabilized with 0.2% Triton X-100 in PBS and blocked with a blocking buffer (5% goat serum + 0.3% BSA in PBS). Following this, the sections were incubated with the primary antibody Purified Rat Anti-Mouse CD41 (BioLegend, Cat #553847, 1:200) overnight at 4°C in a humidified chamber. After washing with TBST, the sections were incubated with Alexa Fluor 488-conjugated Affinipure Goat anti-Rat secondary antibody (Jackson ImmunoResearch, Cat #112-545-003, 1:800) or Alexa Fluor^TM 568 goat anti-Rat lgG(H+L) antibody (Invitrogen, Cat #A11077, 1:1000) for 1 hour at room temperature. For Kupffer cells co-staining, sections were further incubated with Alexa Fluor 594 anti-mouse F4/80 (BioLegend, Cat #123140, 1:200) or Alexa Fluor 647 anti-mouse CLEC4F (BioLegend, Cat #156804, 1:200) for 1 hour at room temperature. Finally, the sections were stained with DAPI (Beyotime, Cat #C1006) before mounting and imaging on an LSM 900 microscope (Zeiss). All images were collected using consistent parameters.

Sample Preparation for RNA Sequencing and Mass Spectrometry

Mice were treated according to different experimental protocols. KCs were isolated and lysed in 1 mL RNAiso Plus reagent (TaKaRa, Cat #9109). RNA concentration and purity were measured using a NanoDrop 2000 (Thermo Fisher Scientific), and RNA integrity was assessed using the RNA Nano 6000 Assay Kit on the Agilent Bioanalyzer 2100 system (Agilent Technologies). A total of 1 μg RNA per sample was used for RNA sample preparation. Sequencing libraries were generated using the NEB Next Ultra™ RNA Library Prep Kit for Illumina (NEB) following the manufacturer’s recommendations. Index codes were added to attribute sequences to each sample. Clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v4-cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina NovaSeq 6000 platform, generating paired-end reads (the sequencing data for Fig1A-B were performed in BGI, and the sequencing data for Fig1C-D were performed in BMKCloud).

Mass Spectrometry Analysis of EV Cargo

To identify PEV-specific protein cargo, PEVs were isolated as described. For comparison, hepatocyte-derived extracellular vesicles (HEVs) were isolated from the culture supernatant of AML12, a mouse hepatocytes cell line, using an identical ultracentrifugation protocol. Protein extracts from PEVs and HEVs were trypsin-digested and analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) on a Orbitrap Exploris 480 MS (Thermo Fisher Scientific). Proteins identified in HEVs, as well as those annotated to the Gene Ontology (GO) ter ms “cytoskeleton” (GO:0005856) and “blood coagulation” (GO:0007596), were subtracted from the PEV proteome. Candidate proteins were defined as those present in all three biological replicates with more than two unique peptides.

RNA sequencing analysis

Raw data (raw reads) in fastq format were processed using in-house Perl scripts. Clean data (clean reads) were obtained by removing reads containing adapters, reads containing poly-N, and low-quality reads from the raw data. Quality metrics such as Q20, Q30, GC-content, and sequence duplication levels were calculated for the clean data. All downstream analyses were based on high-quality clean data. Reads were mapped against the reference genome and transcript annotation (GRCm38) using Hisat2 software. Further analysis was performed using RStudio (version 4.3). Differentially expressed genes were determined using the limma package, and enrichment analysis was conducted using clusterProfiler.

Western blot

Cells were lysed in RIPA lysis buffer (10mM Tris-HCl pH 8.0, 1mM SDTA, 0.5mM EGTA, 1% Triton X-100, 0.1%Ssodium deoxycholate, 0.1% SDS, 140mM NaCl, 100mM PMSF and 100 ×Cocktail were added freshly when using) on ice for 30 minutes. Protein extracts were obtained from the supernatants following centrifugation and subjected to SDS-PAGE. The membranes were blocked with 5% non-fat dry milk for 1 hour at room temperature. After blocking, membranes were incubated with primary antibodies followed by species-appropriate horseradish peroxidase-conjugated secondary antibodies. Membranes were washed three times with TBS-T and detected using either ECL Select Western Blotting Detection Reagent (Cytiva, Cat #RPN2235) or SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Scientific, Cat #34580). For mouse serum, which will be diluted with PBS for 20 times, and then add SDS-loading buffer, 98℃ for 10min.

Primary antibodies used to detect specific proteins: ALDOA Polyclonal antibody (Proteintech, Cat #11217-1-AP, 1:5000), CD63 antibody (MX-49.129.5) (Santa Cruz, Cat #sc-5275, 1:500), TSG101 antibody(C-2) (Santa Cruz, Cat #sc-7964), Beta Actin Monoclonal antibody (Proteintech, Cat #66009-1-Ig, 1:10,000), Albumin Antibody (Cell Signaling Technology, Cat #4929, 1:1000), Alpha-Tubulin monoclonal antibody (Proteintech, 66031-1-Ig, 1:2000). Secondary antibodies include goat anti-Rabbit IgG (Jackson ImmunoResearch, Cat #111-035-003, 1:10,000), goat anti-Mouse IgG (Jackson ImmunoResearch, Cat #115-035-003, 1:10,000).

Seahorse Analysis

Real-time changes in extracellular acidification rate (ECAR) of KCs were measured using the XFe24 Seahorse extracellular flux analyzer (Agilent) with the Seahorse XF Glycolysis Stress Test Kit (Agilent, Cat #103020-100). KCs were plated onto XFe24 cell culture plates containing 0.3×10^6 cells a well and subjected to specific experimental treatments. The cells were rinsed with Seahorse XF basal medium supplemented with 2 mM glutamine and maintained in 500 µL of the same basal medium. The analysis was performed using the following Seahorse XF running buffers: Buffer A: 100 mM glucose in basal medium;Buffer B: 10 μM oligomycin in basal medium;Buffer C: 500 mM 2-deoxy-D-glucose (2-DG) in basal medium. At the end of the assay, the cells were lysed with RIPA buffer, and protein quantification was performed using the Pierce™ Protein Assay Kit (Thermo Scientific, Cat #23227) based on the bicinchoninic acid (BCA) method. Results were analyzed using Seahorse Wave Desktop Software (Agilent) and were normalized to protein content.

Prokaryotic purification of recombinant ALDOA

For the prokaryotic purification of recombinant Aldoa, the full-length cDNAs encoding Aldoa were cloned into the pET-28a vector and transformed into E. coli strain BL21 (DE3). The transformed cells were induced with 1 mM isopropyl-β-D-thiogalactopyranoside at an OD600 of 1.0 and incubated for an additional 4 hours at 37°C. Following incubation, the cells were collected and resuspended in ice-cold His binding buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1% Triton X-100, 1% glycerol, 5 mM imidazole). The cell suspension was sonicated and then centrifuged at 15,000 g for 1 hour at 4°C. The supernatant was filtered through a 40 μm filter membrane and subsequently purified using Ni-NTA Agarose. The Ni-NTA Agarose was washed sequentially with ice-cold His wash buffers: first with buffer 1 (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1% glycerol, 20 mM imidazole), followed by buffer 2 (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1% glycerol, 40 mM imidazole). Recombinant Aldoa were eluted from the resin using ice-cold His elution buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 1% glycerol, 500 mM imidazole) and its concentration was determined using a BCA assay.

Real-time RT-PCR assay

Total RNA was extracted from cells using RNAiso Plus reagent, followed by chloroform extraction and isopropanol precipitation to isolate RNA. The RNA pellets were washed with 75% ethanol, air-dried, and resuspended in DEPC-treated water. cDNA synthesis was carried out using the PrimeScript II 1st Strand cDNA Synthesis Kit (TaKaRa, Cat #6210A), adhering to the manufacturer’s protocols. Quantitative RT-PCR (qPCR) was performed in duplicate using the PowerUp SYBR Green Master Mix (Applied Biosystems, Cat #A25742) on a StepOne system. The PCR conditions included an initial denaturation at 50°C for 2 minutes, followed by 95°C for 2 minutes, and then 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. A final cycle involved 95°C for 15 seconds, 60°C for 1 minute, and 95°C for 15 seconds. Data were analyzed using the comparative Ct method with 18S rRNA as the internal control. Primer sequences are detailed in Table S4.

Analysis of Human Proteomics Data

For the analysis of human platelet alpha granules (Fig 6F), the raw mass spectrometry data from Maynard et al. were obtained and re-analyzed using Graphpad Prism software (v.8.0.2) to extract the spectral counts for ALDOA. For the analysis of EV proteomes from polycythemia vera and Alzheimer’s disease patients (Fig 6G), processed data tables from Fel et al (PXD013234) and Nielsen et al (PXD024216) were directly obtained from the ProteomeXchange database as provided in the respective publications, and the normalized intensity values for ALDOA were extracted.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism software (v.8.0.2). Data are expressed as mean values ± SEM. To assess significance between two groups, an unpaired two-tailed Student’s t-test was employed. For comparisons among multiple groups, one-way ANOVA was used. Statistical significance was defined as p < 0.05. Detailed statistical information is provided in the figure legends and associated data. Images shown without biological replicates represent results from at least three independent experiments.

Data availability

RNA sequence data were deposited in GEO (accession number GSE274627). All data generated or analyzed during this study are included in the manuscript and supporting files; source data files have been provided for all figures.

Acknowledgements

We thank Dr. Bin Qi (Yunnan University) and Dr. Wenxiang Fu (Yunnan University) for suggestions and discussions. We thank Dr. Hao Yin (Wuhan University) for providing us with AML12 cells. This work was supported by National Natural Science Foundation of China (82570734 to Z.S.), Yunnan Provincial Science and Technology Department (C619300A086 to Z.S.), General Program of the Yunnan University Medical Research Fund (YDYXJJ2025-0040 to J. H. L.).

Additional information

Author Contributions

R. X.Y. conducted the experiments, analysed the data, and wrote the manuscript. J.H.L. provided the patient samples. T.W., Y.H.L., C.S., and L.Y. were responsible for patient sample collection. K.F. performed the transmission electron microscopy analysis of PEVs. K.Q.W. prepared the mass spectrometry samples of HEVs. Z.S. conceived and designed the study, supervised the project, and contributed to the writing and revision of the manuscript.

Funding

MOST | National Natural Science Foundation of China (NSFC) (82570734)

• Zhao Shan

Yunnan Provincial Science and Technology Department (云南省科学技术厅) (C619300A086)

• Zhao Shan

Additional files

Supplementary figures.

Supplementary tables.