LPS-induced macrophage HMGB1-loaded extracellular vesicles trigger hepatocyte pyroptosis by activating the NLRP3 inflammasome

Extracellular vesicles (EVs) have emerged as important vectors of intercellular dialogue. High mobility group box protein 1 (HMGB1) is a typical damage-associated molecular pattern (DAMP) molecule, which is cytotoxic and leads to cell death and tissue injury. Whether EVs are involved in the release of HMGB1 in lipopolysaccharide (LPS)-induced acute liver injuries need more investigation. EVs were identified by transmission electron microscopy, nanoparticle tracking analysis (NTA), and western blotting. The co-localization of HMGB1, RAGE (receptor for advanced glycation end-products), EEA1, Rab5, Rab7, Lamp1 and transferrin were detected by confocal microscopy. The interaction of HMGB1 and RAGE were investigated by co-immunoprecipitation. EVs were labeled with the PKH67 and used for uptake experiments. The pyroptotic cell death was determined by FLICA 660-YVAD-FMK. The expression of NLRP3 (NOD-like receptor family pyrin domain containing 3) inflammasomes were analyzed by western-blot or immunohistochemistry. Serum HMGB1, ALT (alanine aminotransferase), AST (aspartate aminotransferase), LDH (lactate dehydrogenase) and MPO (myeloperoxidase) were measured using a commercial kit. The extracellular vesicle HMGB1 was detected in the serums of sepsis patients. Macrophages were found to contribute to HMGB1 release through the EVs. HMGB1-RAGE interactions participated in the loading of HMGB1 into the EVs. These EVs shuttled HMGB1 to target cells by transferrin-mediated endocytosis leading to hepatocyte pyroptosis by the activation of NLRP3 inflammasomes. Moreover, a positive correlation was verified between the sepsis serum EVs-HMGB1 level and clinical liver damage. This finding provides insights for the development of novel diagnostic and therapeutic strategies for acute liver injuries.

INTRODUCTION Extracellular vesicles (EVs) have emerged as important vectors of intercellular communication [1,2]. There is evidence that EVs have crucial roles in many liver diseases including hepatitis viral infections, hepatocellular carcinoma, liver fibrosis, non-alcoholic fatty liver disease, alcoholic liver disease, and liver-specific tumor metastasis [3,4]. However, the role of EVs in sepsis-induced acute liver injuries remains unclear and requires further investigation.
Damage associated molecular pattern (DAMP) molecules have key roles in the progression of lipopolysaccharide (LPS)-induced acute liver injuries. HMGB1 is a typical DAMP molecule that is cytotoxic and leads to cell death and tissue injury [5]. HMGB1 release occurs through two major pathways: passive release and active release. Active secretion always requires two steps: the translocation of nuclear HMGB1 to the cytoplasm and the gradual induction of programmed, pro-inflammatory cell death that enables cytoplasmic HMGB1 to reach the extracellular space [6]. One study has shown that amnion epithelial cells release exosomes containing HMGB1 to other uterine tissues to produce parturition-associated inflammatory changes [7]. Remarkably, the neutralization of inflammation by HMGB1-specific antibodies prevents endotoxin-induced lethality [8,9]. It is unknown whether EVs are involved in the release of HMGB1 in LPS-induced acute liver injuries.
LPS has a pivotal role in the pathogenesis of acute liver injuries. Kupffer cells (KCs), resident hepatic macrophages, are well characterized in response to LPS [10]. Through toll-like receptor 4 (TLR4), KCs take up endotoxins and contribute significantly to inflammation and liver damage [10]. It remains unclear whether macrophage-derived EVs contain HMGB1. NLRP3 inflammasomes are mediators of LPS-induced acute liver injury [11]. NLRP3 activation causes the activation of caspase-1, which cleaves pro-IL-1β and pro-IL-18 to their mature forms and induces cell pyroptosis [12]. HMGB1 has a strong correlation with NLRP3 inflammasome activation. A previous study has shown that heatstroke-induced HMGB1 binds to the receptor for advanced glycation endproducts (RAGE), leading to NLRP3 inflammasome activation and subsequent pyroptosis of hepatocytes [13]. Xu et al. revealed that HMGB1 initiates HMGB1 endocytosis through RAGE binding, which in turn induces cell pyroptosis [14]. A recent article has shown that in lethal sepsis, HMGB1 can mediate caspase-11induced pyroptosis [15]. It is unclear whether HMGB1 in EVs activates NLRP3 in recipient cells and causes pyroptosis.
Consequently, we hypothesized that LPS induces the release of HMGB1-loaded EVs from macrophages to trigger the pyroptosis of hepatocytes through the NLRP3 inflammasome pathway. This study provides further insights into the importance of macrophage-derived EVs in LPS-induced acute liver injury.

Macrophages contribute to the HMGB1 release by EVs
We isolated serum EVs (Fig. 1a) from sepsis patients (patient information is shown in Table 1). The common markers of the EVs (CD63, CD9, and CD81) were detected with western blotting (Fig. 1b). Nanoparticle tracking analysis revealed that the diameters of the EVs were approximately 120 nm (Fig. 1c). The counts for the serum EVs in the sepsis patients were higher than that of the healthy controls (Fig.  1d). We divided the serums of sepsis patients into three parts: serum, EVs, and EVs-free serum. HMGB1 was detected by ELISA, and it was found that the EVs contained more HMGB1 than that of the EVs-free serum (Fig. 1e). It is possible that HMGB1 within the EVs could not be detected by ELISA. We conducted further analyses to detect HMGB1 after lysis of the EVs and found no significant difference in the amount of HMGB1 between the EVs and EVs-lysis groups (Fig. 1f). The above results show that HMGB1 in the peripheral blood serum of the sepsis patients may exist in the form of EVs.
Macrophages can release HMGB1 into the extracellular space, therefore leading to cell death and tissue injury [5]. We hypothesized that macrophages might contribute to HMGB1 release through the EVs. We cleaned out mice macrophages with clodronate and exposed the mice to LPS (Fig. 1g). The F4/ 80 staining confirmed that the macrophages were cleaned out (Fig. 1h) in the liver. By measuring the number of HMGB1 + serum-EVs (EVs loaded with HMGB1), we found that the clodronate decreased the number of HMGB1 + serum-EVs compared to the LPS group (Fig. 1i). Next, we isolated EVs from Raw264.7 cells. TEM was performed to reveal the structure of the EVs (Fig. 1j). The common markers (CD63, CD9, and CD81) for EVs and HMGB1 were detected using western blotting (Fig. 1k). LPS increased the number of released Raw264.7 EVs compared to the control group (Fig. 1l). Interestingly, we found that LPS also increased the level of HMGB1 within the EVs derived from the Raw264.7 cells (Fig.  1m-n), and 100 ng/ml of LPS induced the highest expression of HMGB1 + EVs (Fig. 1m). Moreover, HMGB1 was expressed at normal levels within the nuclei while LPS induced the nucleocytoplasmic translocation of HMGB1 (Fig. S1). These results show that macrophages contribute to the release of HMGB1 by the EVs.

LPS-induced macrophage EVs cause liver damage
From the clinical data, we found that the level of HMGB1 + serum-EVs was positively correlated with the level of liver injury in the patient ( Fig. 2a-b) and the inflammation index (C-reactive protein (CRP) and procalcitonin (PCT)) ( Fig. 2c-d). We labeled macrophage EVs with PKH67 and injected the cocktail into mice. We found that the main absorption organ was the liver (Fig. S2). After cleaning out the macrophages with clodronate for 48 h, we injected LPS together with LPS-induced Raw264.7 EVs into the mice for another 8 h. The inflammatory response (myeloperoxidase (MPO) expression) decreased when the macrophages were cleaned out with clodronate when exposed to LPS (Fig. 2e). Interestingly, the inflammatory response was rescued by the LPS-induced macrophage EVs when the mice macrophages were cleaned out. The ALT, AST, and LDH levels showed the same trends ( Fig. 1f-h). These results suggest that LPS-induced macrophage EVs cause liver damage.
HMGB1 absorbed into the EVs is depend on cytoplasmic RAGE Next, we investigated how the HMGB1 was loaded into the EVs and released into the extracellular space. First, we observed that the RAGE was co-localized with the endosome system (EEA1, Rab5, and Rab7). When LPS-induced HMGB1 translocated from the nucleus to the cytoplasm, cytoplasmic HMGB1 co-localized with the RAGE and endosome system rather than with the lysosomal marker (Lamp1) of the Raw264.7 cells (Fig. 3a). Co-IP (co-immunoprecipitation) analysis further confirmed that the interaction between HMGB1 and RAGE could be blocked with FPS-ZM1 (RAGE inhibitor) (Fig. 3b). Meanwhile, FPS-ZM1 did not affect the nuclear translocation of HMGB1 (Fig. 3c). FPS-ZM1 reduced the protein levels of HMGB1 and RAGE in the EVs (Fig. 3d). Collectively, these findings show that HMGB1 is delivered into the EVs which is depend on cytoplasmic RAGE before being released into the extracellular space.
After internalization in the hepatocytes, we found that the EVs co-localized with the endosomal/lysosomal EEA1, Rab5, Rab7, and Lamp1 (Fig. 4c). Interestingly, we found that the EVs expressed transferrin receptors (Fig. 4d) and that the absorbed EVs were colocalized with AML-12 transferrin. Furthermore, the uptake rate of the EVs and cytoplasmic HMGB1 in the target cells decreased significantly in response to transferrin neutralizing antibody (Fig. 4e, f). These results show that EVs secreted from Raw264.7 cells shuttled HMGB1 to AML-12 cells through transferrinmediated endocytosis.

LPS-EVs promote the pyroptosis of hepatocytes
To find the function of LPS-EVs, we focused on the biological effects of the LPS-EVs on hepatocytes because we already found that LPS-induced macrophage EVs cause liver damage. We found that LPS-EVs activated the NLRP3 inflammasomes in AML-12 cells in a dose-and time-dependent manner (Fig. 5a,  b). Immunofluorescence also showed that the NLRP3 inflammasomes were induced by LPS-EVs (Fig. 5c). When the protein or RNA was removed from the EVs, LPS-EVs induced NLRP3 inflammasome activation was blocked (Fig. 5d). This shows that the proteins and RNAs within the LPS-induced macrophage EVs were both involved in the activation of the hepatocyte NLRP3 inflammasomes. From the immunofluorescence, we found that LPS-EVs induced pyroptosis in the AML-12 cells (Fig. 5e). Recent studies have identified pyroptosis as gasdermin-mediated programmed necrotic cell death. Gasdermin D (GSDMD) is a substrate for both caspase-1 and caspase-11/4/5 [16]. We detected GSDMD-N domains in LPS EVs-treated AML-12 cells, which shows that these EVs triggered the pyroptosis in the AML-12 cells (Fig. 5f). These results suggest that LPS EVs activate NLRP3 inflammasome in hepatocytes and cause the pyroptosis of hepatocytes.
HMGB1 + EVs inducing hepatocyte pyroptosis and activate NLRP3 inflammasomes Whether HMGB1 + EVs induced pyroptosis of hepatocytes was not investigated. We isolated primary hepatocytes from wildtype and NLRP3 −/− mice. They were then treated with different EVs (control, LPS, HMGB1-OE, and HMGB1-KD). The number of pyroptotic primary hepatocytes caused by HMGB1overexpressing EVs increased significantly compared to the f EVs HMGB1 and EVs-lysis HMGB1 levels in the sepsis serum, ns: no significance. g Macrophages were cleaned out with clodronate for 48 h and exposed to LPS for another 8 h, serum-derived EVs were collected. h F4/80 staining in all groups; Scale bar = 200 μm. i HMGB1 level in serum-derived EVs in all groups, the proteins were normalized by GAPDH, (**) P < 0.01 versus control-EVs group; ( && ) P < 0.01 versus LPS-EVs group. j TEM of Raw264.7 EVs. Scale bar = 100 nm. k EVs biomarkers and HMGB1 in Raw264.7 derived EVs. Immunoblots are representative of three separate experiments. l Number of EVs analyzed by NTA. m HMGB1 in Raw264.7 derived EVs exposed to different concentrations of LPS protein was normalized to GAPDH protein. Immunoblots are representative of three separate experiments. n EVs HMGB1 and levels in the LPS-induced macrophage culture medium at different time points. (*) P < 0.05 versus 0 h group; (**) P < 0.01 versus 0 h group.
LPS-EVs group in the wild-type mice. However, the number of pyroptotic primary hepatocyte cells decreased in the HMGB1knockdown EVs group compared to the LPS-EVs group (Fig. 6a). The released LDH levels of AML-12 showed a similar trend (Fig. 6b).
An interesting thing was that pyroptosis in the NLRP3 −/− hepatocytes decreased compared to the wild-type hepatocytes. There was no significant change in the NLRP3 −/− hepatocytes pyroptosis and LDH among the HMGB1-OE-EVs group and the LPS EVs or the HMGB1-KD-EVs group and the LPS EVs (Fig. 6c, d). These results show that the pyroptosis of the hepatocytes caused by the HMGB1 + EVs depended on NLRP3 activation in the cells.
For the role of HMGB1 in LPS-EVs toward hepatocytes NLRP3 activation. Our results show that HMGB1-overexpressing EVs promoted the activation of the NLRP3 inflammasome/IL-1β, while HMGB1-knockdown EVs decreased the activation of the NLRP3 inflammasomes/IL-1β axis compared to LPS EVs (Fig. 6e).
When we knocked out the NLRP3 in hepatocytes, the activation of the NLRP3 inflammasome/IL-1β axis in the NLRP3 −/− primary hepatocytes decreased compared to that of the wild-type primary hepatocytes (Fig. 6e), suggesting that HMGB1 in the EVs activated the NLRP3 inflammasomes in the target cells.
HMGB1 + EVs cause liver damage and activate NLRP3 inflammasomes in the liver We further focused on the effects of the HMGB1 + EVs in vivo. Raw264.7-derived EVs (control, HMGB1-overexpressing, and HMGB1-knockdown) were injected into wild-type mice and NLRP3 −/− mice through the tail vein. The results showed that the NLRP3/ASC/caspase-1 level increased in the LPS, CLP (cecal ligation puncture), LPS EVs, and HMGB1-overexpressing EVs groups. However, when HMGB1 + EVs were knocked down, the NLRP3/ASC/caspase-1 level decreased significantly ( Fig. 7a-b). We also found that the expression of ASC and cleaved-caspase-1 decreased in all the groups of NLRP3 −/− mice compared to the WT mice ( Fig. 7a-b). The serum analysis of the ALT, AST, LDH, and MPO levels showed the same trends ( Fig. 7c-f).

Serum-derived EVs of patients with sepsis express high HMGB1 and cause liver injury in vivo
To deduce the function of serum HMGB1 + EVs in sepsis patients, we injected HMGB1-low serum EVs and HMGB1-high serum EVs into wild-type mice. Additionally, other components were different in these two types of EVs. We still found NLRP3 inflammasome activation in the HMGB1-high serum EVs group compared to the HMGB1-low serum EVs group (Fig. 8a, b). The serum analysis showed that the ALT, AST, LDH, and MPO levels in the HMGB1-high serum EVs increased compared to the HMGB1-low serum EVs (Fig. 8c-f). These results show that the sepsis serum-derived EVs express a high level of HMGB1, which causes liver injury and NLRP3 inflammasome activation.

DISCUSSION
The present study demonstrates that LPS-induced HMGB1 + EVs from macrophages trigger hepatocyte pyroptosis by activating the NLRP3 inflammasome. The main findings include the following: (1) HMGB1 released by EVs in the serum and macrophages contribute to EVs-HMGB1 release; (2) LPS induces the sorting of HMGB1 into EVs which depend on RAGE; () Macrophage-derived EVs shuttle HMGB1 to hepatocytes by transferrin-mediated endocytosis and subsequently promote hepatocyte pyroptosis by activating the NLRP3 inflammasome. We report that LPS induces macrophages to release HMGB1 as the cargo of the EVs. Compared to early inflammatory mediators (such as TNF-α, IL-6, and IL-1β), which have limited clinical significance due to their narrow therapeutic window, HMGB1 has a prolonged therapeutic window as a late mediator in lethal systemic inflammation [17]. Circulating TNF-α and IL-6 are elevated for 5 days after the onset of sepsis, and serum HMGB1 levels are elevated from day 7 until at least day 28 [17]. Interestingly, our data showed that EVs in patients with sepsis are important carriers of HMGB1 than free-state HMGB1, and the lysis of EVs or not has no significant effect on the ELISA detection of HMGB1. Due to the stability of the EVs and the development of detection methods, HMGB1 + EVs can be a promising biomarker for sepsis diagnosis. Herein, we found that HMGB1-loaded EVs could be secreted from macrophages at the very beginning of cell injury.
The mechanisms behind the cargo loading of EVs are still poorly understood. Studies have revealed that the ESCRT machinery is essential for controlling specific protein sorting into intraluminal vesicles (ILVs) within the first step of ubiquitination [17,18]. Kang, R [6]. and Yiting Tang [17] et al. have shown that cytoplasmic HMGB1 is modified by posttranscriptional modifications of hyperacetylation instead of ubiquitination after HMGB1 localizes from the nucleus to the cytoplasm. Our results confirm that cytoplasmic HMGB1 is transferred into multivesicular bodies (MVB) by binding to RAGE expressed in the endosome system. We show that RAGE has a vital role during the sorting of HMGB1 into the EVs. Our study suggests that the receptor-ligand pattern is an important mechanism during the cargo sorting of EVs.
Previous reports have revealed that the binding and uptake of EVs involve many pathways [1,2]. The selective uptake of EVs may be critical for them to function properly. For instance, exosomes engineered to express the EGFR ligand on their surfaces preferentially bind to EGFR-expressing breast cancer cells and successfully deliver microRNA to recipient cells [19]. Our results show the participation of the transferrin pathway in the uptake of EVs by hepatocytes. When we blocked transferrin in the hepatocytes with a neutralizing antibody, the percentage of internalized EVs decreased dramatically. Our study thus suggests that the interaction between transferrin and its receptor has a role in the uptake of macrophage-derived EVs by hepatocytes. Wree [20]. et al. reported that continuous NLRP3 inflammasome activation induces hepatocyte pyroptosis and subsequent severe liver damage in mice. Our data suggest a direct effect of HMGB1 + EVs on hepatocyte pyroptosis. Previous studies have highlighted the function of HMGB1 through the binding of receptors to activate pro-inflammatory signaling [6]. Ben Lu et al. discovered that hepatocyte-released high mobility group box 1 (HMGB1) was required for caspase-11-dependent pyroptosis and lethality in endotoxemia and bacterial sepsis. Mechanistically, hepatocyte-released HMGB1 bound LPS and targeted its internalization into the lysosomes of macrophages and endothelial cells via the receptor for advanced glycation end-products (RAGE) [21]. However, our results show that HMGB1 + EVs can induce pyroptosis once absorbed by recipient cells. This function may be responsible for the occurrence of inflammation in LPS-induced acute liver injuries.
From the clinical results, we propose that serum HMGB1 + EVs in sepsis patients can be used as a marker for sepsis diagnosis. Previous research has shown that HMGB1 appears to be an early stage mediator in trauma [22], while HMGB1 is conversely a latestage mediator in sepsis [23]. Wang showed that treatment of FeTPPS (a small molecule selectively inhibits HMGB1-mediated caspase-11 activation) attenuates HMGB1-and caspase-11mediated immune responses, organ damage, and lethality in endotoxemia and bacterial sepsis [24]. Li W report that the exosome was an important pathway for HMGB1 release from hepatocytes; which was dependent on TLR4 [25]. Here, we found that the serum-derived EVs of sepsis patients and macrophages contain high levels of HMGB1 and stimulate the activation of the NLRP3 inflammasome and liver damage. It indicated that both macrophages and hepatocytes could release HMGB1through extracellular vesicles. Numerous animal studies have revealed that HMGB1 is a negative outcome predictor of sepsis [26]. In humans, HMGB1 circulatory levels persistently increase in severe sepsis and septic shock [23,26]. This increase seems to be associated with a poor patient outcome or organ failure. Nevertheless, there is no significant difference in the survival rate [27]. In fact, patients who survive severe sepsis maintain a continuously high HMGB1 circulatory level [27]. Our findings suggest that HMGB1 has the potential as a diagnostic marker for sepsis.
This study had several limitations: (1) These Nlrp3 and other components could also have been included in the macrophage-released EVs, which caused activation of NLRP3 inflammasome; (2) We did not explore whether the HMGB1 + EVs level could be used as a prognostic marker for sepsis; (3) The released EVs might contain LPS itself which cause a similar effect on hepatocyte, which need further investigation.
In summary, we found that LPS induces HMGB1-loaded EVs release from macrophages by binding to RAGE and that these EVs are internalized by hepatocytes by transferrin-mediated endocytosis, which triggers the pyroptosis of the hepatocytes by activating the NLRP3 inflammasome. This study provides novel insights into possible future diagnostic and therapeutic strategies for acute liver injury. Fig. 7 HMGB1 + EVs cause liver damage and activate NLRP3 inflammasomes in the liver. a EVs were collected from Raw264.7 (control, HMGB1-Overexpressing and HMGB1-Knockdown) were injected into WT and NLRP3 −/− mice. The LPS or CLP surgery models served as the sepsis positive models. Expression of NLRP3, ASC and Caspase-1 was evaluated; scale bar = 100 μm. n = 5 mice per group. b The expression levels of NLRP3/ASC/Caspase-1 were analyzed by immunoblotting. Data are expressed as means ± sd. * for p < 0.01 vs control group. # for p < 0 . 05; & for p < 0.05. c-f Serum ALT, AST, LDH and MPO levels. Data are expressed as means ± sd. * for p < 0.01. & for p < 0.05; # for p < 0.01. n = 5 mice per group.

Human studies
Human studies were approved by the Nanfang hospital, Southern Medical College (Institution Review Board No. 0604008488). Human serum samples were obtained from control healthy subjects and sepsis patients with liver injury at Department of Emergency Medicine and Department of Infectious Diseases (Nanfang Hospital, Southern Medical University, China). Human studies were approved by the Nanfang hospital, Southern Medical College (Institution Review Board No. 0604008488). All individuals provided informed consent for serum donation on approved institutional protocols.

Animals
Male Wide Type C57BL/6, NLRP3 knock-out mice (NLRP3 −/− , Jackson laboratory, stock 021302) (6-8 week) were maintained in a specific pathogen-free facility (Central Animal Care Facility of Southern Medical University, Guangzhou, China). Animal experimental procedures were in accordance with the National Institutes of Health guidelines and were approved by the local Animal Care and Use Committee of Southern Medical University.

Acute liver injury models
A single dose of LPS (10 mg/kg; Sigma-aldrich, St. Louis, MO) was administered to mice intraperitoneally and CLP were used as sepsis animal models. Mice were anesthetized with 5% chloral hydrate, after which a 2-cm midline laparotomy was performed under aseptic conditions to expose the cecum. A single through-and-through puncture was performed by a 21-gauge needle between the ligation site and the end of the cecum, and then fecal material was extruded through the puncture. The cecum was repositioned into the peritoneal cavity, and the laparotomy was closed. For macrophage deletion, clodronate (200 μl/mice) was injected into mice intraperitoneally for 48 h. For EVs injection, EVs (40 μg) were injected into mice once through tail vein. All mice in this study were randomly divided into different groups. No blinding was done in this research.

Isolation and identification of extracellular vesicles
Raw264.7-EVs and serum-EVs were isolated and purified as described previously, with added modifications [29]. Briefly, Raw264.7 culture medium and serum were harvested and centrifuged at 300 g for 10 min, 2000 g for 20 min and 10,000 g for 30 min at 4°C. Then supernatant was filtered through 0.22 μm filters (Millipore, MA, USA, SLGP033RB) to remove contaminating apoptotic bodies, microvesicles and cell debris. Clarified culture medium was then centrifuged in a Beckman Coulter Optima TM L-80XP Ultracentrifuge at 100,000 g at 4°C for 70 min to pellet exosomes. The supernatant was carefully removed, and EVs-containing pellets were resuspended in1 mL of ice-cold PBS and pooled. A second round of ultracentrifugation [100,000 g at 4°C for 70 min with a Type 50.2 Ti rotor (kfactor: 157.7)] was carried out, and the resulting EVs pellet resuspended in 100 μL of PBS and stored at −80°C. The protein content of the concentrated EVs was determined using a BCA protein assay kit. Raw264.7-EVs were identified by transmission electron microscopy and NTA (NS3000, Worcestershire, UK), tetraspan molecules CD63 (ab10895, Abcam, Cambridge, UK), CD9 (ab92726, Abcam, Cambridge, UK), CD81 (18250-1-AP, Proteintech, Rosemont, USA) were verified by western blotting.

Co-immunoprecipitation
Raw264.7 were stimulated with 100 ng/ml LPS for 24 h. IP and immunoblotting (IB) were performed, the antibodies for IP was anti-HMGB1; the antibodies for IP was anti-RAGE.

Extracellular vesicles labelling and trafficking in vitro and in vivo
For co-culture experiments, AML-12 cells at a density of 2 × 10 5 cells per ml were cultured in a transwell system with a 0.4-μm porous membrane. For EVs uptake experiments, Raw264.7-derived EVs were labelled with the PKH67 Green Fluorescent Cell Linker Kit (Sigma Aldrich, St. Louis, MO, USA): EVs diluted in PBS were added to 1 ml Diluent C. In parallel, 4 μl PKH67 dye was added to 1 ml Diluent C and incubated with the EVs solution for 4 min. Then 2 ml 0.5% BSA/PBS was added and the labeled EVs were washed at 100,000 g for 1 h, the EVs pellet was diluted in 100 μl PBS and used for uptake experiments. The transferrin neutralized antibody (1:100, ab82411, Abcam, Cambridge, UK) was used to block the transferrin in hepatocytes.

Lentiviral packaging and cell infection
The HMGB1-over-expression (or knock down) lentiviral vector (lenti-HMGB1) and lenti-green fluorescent protein [GFP] (the negative control) were purchased from GenePharma (GenePharma, Shanghai, China). Transfections were performed according to the manufacturer's instructions. Briefly, all lenti vectors were purified to a titer of 1 × 10 9 TU/ml. At a Multiplicity of infection (MOI) of 10:1 and 5 μg/ml polybrene in growth medium for 6 h after transfection, the medium was exchanged for fresh DMEM medium. At 72 h after transfection, the cells were passaged for use in further experiments.
Small interfering RNA (siRNA) siRNA duplexes targeting HMGB1 and negtive control (NC) were designed and synthesized by RiboBio (Guangzhou, China). Transfection was performed using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. Briefly, 24 h before transfection, Raw264.7 cells were seeded in 6-well plates at 40% confluency, and transfected with 100 nM HMGB1 siRNA or NC siRNA for 72 h.

Determination of pyroptotic cell death in vitro
To assess pyroptosis in vitro, we designed the following strategy. Active Caspase 1 was measured in AML-12 cells suspensions with FLICA 660-YVAD-FMK (FLICAR 660 in vitro Active Caspase-1 Detection Kit, Immuno Chemistry Technologie, Bloomington, MN, USA) according to manufactures instructions and with propidium iodide (PI) to mark cells with membrane pores (Life technology, Carlsbad, CA, USA). Cells were then analyzed by flow cytometry FACS CantoII (BD Biosciences, San Jose, CA) and fluorescent images were also obtained by confocal microscopy.

Immunofluorescence
The liver was frozen in liquid nitrogen and embedded in Tissue-TEK OCT compound (4583, Sakura, Japan). For immunofluorescence staining, the tissues were fixed overnight in formalin. Extensive washing with PBS was performed before incubating the slides with DAPI for nuclear staining. Image was observed using with Zeiss LSM 700 laser scanning confocal microscope and was calculated with the ZEN software.

Histopathological and Immunohistochemistry
Liver tissues were processed for paraffin embedding and were sectioned into 4-μm sections. The sections were stained with hematoxylin and eosin according to standard protocols. For immunohistochemical staining, the liver slides were incubated with diluted primary antibodies against antimouse NLRP3 (ab4207), ASC (ab47092), Caspase-1 (ab1872) (Abcam, Cambridge, UK) according to the manufacturer's instructions. The primary antibody was detected using the biotin-conjugated anti-rabbit (or antigoat) immunoglobulin G antibody, and then incubated with streptavidinbiotin. The complex was visualized with the DAB reagent for microscopic examination. Removal of the primary antibody from the procedure provided a negative control.