Chronic-plus-binge alcohol intake induces production of proinflammatory mtDNA-enriched extracellular vesicles and steatohepatitis via ASK1/ p38MAPK α -dependent mechanisms

induces role of the kinase apoptosis signal–regulating kinase 1 (ASK1) and p38 mitogen-activated protein kinase (p38) in chronic-plus-binge ethanol–induced steatohepatitis and mtDNA-enriched EV release. revealed the greatest hepatic upregulation of metallothionein 1 and 2 ( Mt1/2 ), which encode 2 of the most potent antioxidant proteins. Genetic deletion of the Mt1 and Mt2 genes aggravated ethanol-induced liver injury, as evidenced by elevation of serum ALT, neutrophil infiltration, oxidative stress, and ASK1/p38 activation in the liver. Inhibition or genetic deletion of Ask1 or p38 ameliorated ethanol-induced liver injury, ROS levels, and expression of phagocytic oxidase and ER stress markers in the liver. In addition, inhibition of ASK1 or p38 also attenuated ethanol-induced mtDNA-enriched EV secretion from hepatocytes. Taken together, these findings indicate that induction of hepatic mtDNA-enriched EVs by ethanol is dependent on ASK1 and p38, thereby promoting alcoholic steatohepatitis. found that pharmacologic inhibition of ASK1 could attenuate liver cell death and hepatic fibrosis driven by inflammasome activation using gain-of-function NOD-like receptor protein 3 ( Nlrp3 ) mutant mice, providing mechanistic insight into the antifibrotic mechanisms of ASK1 inhibition (41). In the current study, we demonstrated that ASK1 plays an important role in promoting oxidative stress and liver injury in chronic-plus-binge ethanol–fed mice. Inhibition or genetic deletion of ASK1 ameliorated injury in the and macrophage infiltration. Our data suggest the possible beneficial roles of ASK1 inhibitors in the early stages of alcoholic steatohepatitis (as indicated by our chronic-plus-binge ethanol–fed model) but not in the severe form of the disease with advanced fibrosis/cirrhosis, inflammation, and bacterial infection (as those who were enrolled in the aforementioned phase II clinical trial) (40). ROS can activate not only ASK1 but also other downstream kinases that may accelerate the progression of alcoholic steatohepatitis. The p38 MAPKs are known to be activated by MEK3 (or MKK3) and MEK6 (or MKK6); both MEK2 and MEK6 can be activated by one of the several upstream MAP3Ks, such as ASK1, dual-leucine zipper-bearing kinase 1 (DLK1), and TFG- β –activated kinase 1 (TAK1) (18). Once activated by MKK3 or MKK6, p38 MAPKs can potentiate the downstream signals or activate other proteins directly (42). The p38 family comprises 4 members, p38 α (19), p38 β (43–45), p38 γ (46), and p38 δ (47, 48), which have been implicated in various cellular responses controlling cell proliferation, differentiation, apoptosis, embryogenesis, inflammatory responses, and stress responses in human reproductive tissues. A large body of evidence indicates that p38 activity is critical for normal immune and inflammatory responses (42). In our study, we focused on specific roles of hepatic p38 α and its implication in the progression of alcoholic steatohepatitis. Our data revealed that hepatocyte-specific deletion of p38a or administration of p38 inhibitors markedly ameliorated serum ALT levels, oxidative stress, as well as ER stress, suggesting the role of hepatic p38–induced liver injury and inflammation in the chronic-plus-binge ethanol–feeding model. In conclusion, our findings indicate that activation of ASK1 and p38 mediates alcoholic liver injury and inflammation by promoting proinflammatory mtDNA-enriched EV release in hepatocytes induced by ROS stress and ER stress. Despite the failure of anti-ASK1 therapy for patients with severe alcoholic hepatitis in a recent clinical trial (40), our study suggested that ASK1 and p38 activation plays an important role in the pathogenesis of the early stages of alcoholic steatohepatitis. Therapeutic intervention targeting this pathway in patients with a less severe form of alcoholic hepatitis deserves further investigation.


Introduction
Alcohol-associated liver disease (ALD) is the leading cause of liver-related morbidity and mortality worldwide, contributing to approximately half of liver cirrhosis cases in Western countries (1)(2)(3). ALD encompasses a broad spectrum of hepatic pathologies, ranging from simple steatosis to severe forms of liver injury such as alcoholic steatohepatitis, progressive fibrosis, end-stage cirrhosis, complications from portal hypertension, and superimposed hepatocellular carcinoma (HCC) (1)(2)(3). The short-term mortality of patients with severe alcoholic hepatitis is very high (20%-50% at 3 months) (2). Although significant scientific progress has been made in understanding the disease process, the pathogenesis of ALD remains elusive, with significant impact on the discovery of mechanism-based therapies (1)(2)(3). The development of such therapies requires translational studies using human samples and suitable animal models that reproduce clinical and histological features of alcoholic hepatitis. One of the major advances in the field of ALD research was the introduction of binge ethanol in chronically ethanol-fed mice (4,5), which causes neutrophilia and hepatic neutrophil infiltration, simulating some features of Alcohol-associated liver disease is a spectrum of liver disorders with histopathological changes ranging from simple steatosis to steatohepatitis, cirrhosis, and hepatocellular carcinoma. Recent data suggest that chronic-plus-binge ethanol intake induces steatohepatitis by promoting release by hepatocytes of proinflammatory mitochondrial DNA-enriched (mtDNA-enriched) extracellular vesicles (EVs). The aim of the present study was to investigate the role of the stress kinase apoptosis signal-regulating kinase 1 (ASK1) and p38 mitogen-activated protein kinase (p38) in chronic-plus-binge ethanol-induced steatohepatitis and mtDNA-enriched EV release. Microarray analysis revealed the greatest hepatic upregulation of metallothionein 1 and 2 (Mt1/2), which encode 2 of the most potent antioxidant proteins. Genetic deletion of the Mt1 and Mt2 genes aggravated ethanol-induced liver injury, as evidenced by elevation of serum ALT, neutrophil infiltration, oxidative stress, and ASK1/p38 activation in the liver. Inhibition or genetic deletion of Ask1 or p38 ameliorated ethanol-induced liver injury, inflammation, ROS levels, and expression of phagocytic oxidase and ER stress markers in the liver. In addition, inhibition of ASK1 or p38 also attenuated ethanol-induced mtDNA-enriched EV secretion from hepatocytes. Taken together, these findings indicate that induction of hepatic mtDNA-enriched EVs by ethanol is dependent on ASK1 and p38, thereby promoting alcoholic steatohepatitis.
human alcoholic hepatitis. By use of this chronic-plus-binge ethanol feeding model, many novel mechanisms underlying ALD have been identified (6,7). For example, we have recently demonstrated that binge ethanol intake activates liver mitochondrial DNA-enriched (mtDNA-enriched) extracellular vesicle (EV) release in chronically ethanol-fed mice and in patients with alcohol use disorder, leading to neutrophilia and liver injury (8). However, the molecular mechanisms by which binge ethanol intake induces hepatic mtDNA-enriched EV release remain unknown.
Previous studies have identified several mechanisms involved in the pathogeneses of ALD, including ethanol metabolism-associated oxidative stress, glutathione depletion, abnormal methionine metabolism secondary to malnutrition, ethanol-mediated induction of endotoxins from the impairment of intestinal permeability and subsequent activation of Kupffer cells (9)(10)(11)(12). Oxidative stress is driven by molecules such as ROS and reactive nitrogen species (RNS), formed during ethanol metabolism, which can lead to activation of MAPKs. Yet the mechanisms by which ROS activates these kinases are unclear. It has been reported that ROS can activate MAPK pathways via the activation of growth factor receptors in several cell types (13) or via the oxidative modification of intracellular kinases (e.g., MAPK kinase kinase [MAP3K]) that are involved in MAPK signaling cascade (14). Another potential mechanism for MAPK activation by ROS may include the inactivation and degradation of the MAPKs that maintain the pathway in an inactive state (15). Activation of MAPKs, which are involved in the transduction of most extracellular signals, is one of the major signal transduction mechanisms by which the cell adapts to changes in the surrounding environment (e.g., oxidative stress, ER stress, calcium overload, ATP, Fas ligand, as well as inflammatory signals such as TNF and LPS) (16). The apoptosis signal-regulating kinase 1 (ASK1)/p38 MAPK (p38) signaling pathway is one of the major stress MAPKs that mediates diverse biological signals leading to cell death, differentiation, survival, and senescence as a response to cellular stress factors (16). The ASK1/p38 pathway has been described in C. elegans as a defense mechanism against bacterial attack, suggesting that this MAPK signaling cassette represents an ancient, evolutionarily conserved component in innate immunity (17). There are 4 isoforms (α, β, γ, and δ) of p38 MAPKs; however, the α-isoform has been postulated to be the most important in inflammation (18). Most of the studies have focused on p38α and identified an important function of p38α in tissue homoeostasis and pathologies from inflammation to cancer, heart, neurodegenerative diseases and metabolic syndrome, making p38α an interesting pharmaceutical target (19)(20)(21).
Multiple studies have demonstrated that the activation of hepatic ASK1/p38 signaling pathways contributes to regulation of hepatic lipid metabolism, inflammation, fibrosis, and systemic insulin resistance in mice (22)(23)(24)(25)(26). However, how the ASK1/p38 signaling pathway contributes to liver inflammation remains largely unknown. In this study, we provide evidence that the ASK1/p38 pathway, which is regulated by the antioxidant protein metallothionein (MT), plays an important role in controlling hepatic mtDNA-enriched EV release in chronic-plus-binge ethanol-fed mice, thereby promoting liver inflammation and injury.

Results
Hepatic Mt1/2 expression is highly elevated in chronic-plus-binge ethanol-fed mice, playing an important role in ameliorating ALD. Oxidative stress is a prominent mediator of alcoholic hepatitis (1). To investigate the regulation of oxidative stress-associated genes after chronic-plus-binge ethanol feeding, microarray analyses of liver samples from mice fed ethanol for 8 weeks (E8w) and mice fed ethanol for 8 weeks plus 1 binge (E8w+1B) were performed as previously described (27). Microarray analysis revealed a significant modulation of 46 oxidative stress-associated genes in E8w+1B mice compared with E8w mice. Among the most highly upregulated genes, Mt1 and Mt2 were elevated by 33-fold and 20-fold, respectively ( Figure  1A). This result was further confirmed by reverse transcription quantitative PCR (RT-qPCR), showing Mt1 and Mt2 upregulation of 60-fold and 40-fold in E8w+1B mice compared with E8w mice, respectively. No modulation in Mt1 and Mt2 expression was found between pair-fed and E8w mice ( Figure 1B). In addition, RT-qPCR analysis demonstrated that hepatic expression of Mt1 and Mt2 was upregulated 6-fold after a single ethanol binge alone (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.136496DS1). Hepatic expression of Mt1 and Mt2 was also markedly upregulated in short-term chronic plus 1 binge (E10d+1B) compared with pair-fed mice ( Figure 1B). Elevated hepatic expression of Mt1 and Mt2 was also observed in mice fed high-fat diet (HFD) or chow plus 1 binge, and their expression was significantly greater in the HFD+1B group (Supplemental Figure 1B). Interestingly, a significant induction of expression of Mt1 and Mt2 after 1 binge was only observed in the liver and not in kidney, heart, spleen, or pancreas (Supplemental Figure 1C).
To determine the role of MT1/2 in ALD, we subjected Mt1/2 -/and corresponding WT mice to chronic-plus-binge feeding. As illustrated in Figure 1C, Mt1/2 -/mice had higher levels of serum ALT and a higher degree of hepatic steatosis as indicated by hepatic triglyceride levels compared with WT mice. Hepatic neutrophils were examined by IHC and RT-qPCR analysis. As showed in Figure 1D and . Liver tissues were collected 9 hours after binge in the E8w+1B group and subjected to microarray analysis. A heat map of oxidative stress-related genes is shown. (B) C57BL/6N mice were pair-fed with control diet (pair-fed, n = 4) or fed an ethanol diet for 8 weeks (E8w), or E8w+1B, and were euthanized 9 hours after binge. C57BL/6N mice were fed a control diet (pair-fed) or ethanol diet for 10 days plus 1 binge (E10d+1B) and were euthanized 3, 6, or 9 hours later. Hepatic mRNA levels of Mt1 and Mt2 were examined by RT-qPCR. (C-F) WT and Mt1/2 -/mice were pair-fed or fed an ethanol diet for 10 days, followed by gavage of a single dose of maltose or ethanol, respectively and were euthanized 9 hours later. Serum ALT levels and hepatic TG levels were measured (C). Hepatic neutrophils were examined by IHC staining with an anti-MPO antibody or RT-qPCR analysis of Ly6g mRNA (D). Hepatic macrophages were examined by immunostaining with an anti-F4/80 antibody or RT-qPCR analysis of F4/80 mRNA (E). Liver tissues were also subjected to TUNEL staining, and the TUNEL + hepatocytes were counted (F). Values represent mean ± SEM (each dot represents 1 mouse sample). Statistical evaluation was performed by Student's t test or 1-way ANOVA with Tukey's post hoc test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001. Figure 2, the number of myeloperoxidase-positive (MPO-positive) neutrophils in the liver was elevated in both WT and Mt1/2 -/mice after chronic-plus-binge feeding, with a higher number in Mt1/2 -/mice. Hepatic expression of Ly6g was also significantly increased after ethanol feeding of Mt1/2 -/mice but not WT mice. Interestingly, the number of F4/80 + macrophages and hepatic expression of F4/80 were decreased in both WT and Mt1/2 -/mice after chronic-plus-binge feeding ( Figure  1E and Supplemental Figure 2). Finally, the number of apoptotic TUNEL + hepatocytes was higher in ethanol-fed Mt1/2 -/compared with WT mice ( Figure 1F).

Supplemental
Deletion of Mt1/2 exacerbates oxidative stress and activation of stress kinases in ethanol-fed mice. To examine the role of MT1/2 in the control of oxidative stress and its associated kinase activation, we first measured hepatic levels of lipid peroxides including 4-hydroxynonenal (4-HNE) and malonaldehyde (MDA) and activated ASK1 and p38 MAPK in ethanol-fed and pair-fed mice. As shown in Supplemental Figure 3A, hepatic expression of 4-HNE and MDA was higher in mice subjected to E10d+1B treatment than in pair-fed mice. ASK1 is a MAPK that mediates ROS-induced cell death through activation of stress kinases such as p38 MAPK and JNK (28,29). To understand whether MAPK signal is activated in ALD, we performed immunostaining of p-ASK1 and p-p38 in ethanol-and pair-fed mice. As shown in Supplemental Figure 3B, higher levels of hepatic pASK1 (Thr838) and p-p38 expression were observed in ethanol-fed compared with pair-fed mice.
To further confirm the role of MT1/2 in the activation of MAPK induced by oxidative stress in ALD, we examined hepatic oxidative stress and MAPK activation in ethanol-fed Mt1/2 -/and WT mice. Hepatic levels of 4-HNE and MDA were higher in ethanol-fed Mt1/2 -/mice than in ethanol-fed WT mice, as illustrated by Western blot analysis ( Figure 2A) and thiobarbituric acid-reactive substance (TBARS) assays ( Figure 2B). Higher levels of hepatic MDA and 4-HNE in ethanol-fed Mt1/2 -/versus WT mice were also evidenced by IHC staining analyses ( Figure 2C). Correspondingly, stronger phosphorylation of ASK1 and p38 was detected in ethanol-fed Mt1/2 -/mice than in WT mice ( Figure 2D).
Deletion of the Ask1 gene or administration of ASK1 inhibitors ameliorates ALD in mice. To further confirm the role of ASK1 in ALD, we performed chronic-plus-binge ethanol feeding (E10d+1B) of WT and Ask1 -/mice. First, downstream signaling of ASK1 pathway was examined by Western blot analysis. As shown in Supplemental Figure 4A, activation of hepatic p38 and JNK1/2 was dramatically lower in Ask1 -/versus WT mice after E10d+1B treatment. In contrast, activation of hepatic STAT3 was higher in Ask1 -/mice compared with WT controls.
Serum levels of ALT were lower in ethanol-fed Ask1 -/mice than in WT mice, as was the percentage of circulating neutrophils ( Figure 3A). Absolute neutrophil counts in the peripheral blood were also decreased; however, this difference was not statistically significant (Supplemental Figure 4B). In addition, compared with ethanol-fed WT mice, ethanol-fed Ask1 -/mice showed a lower number of MPO + neutrophils and F4/80 + macrophages, and lower expression of hepatic Ly6g and F4/80 and several other inflammatory mediators (Supplemental Figure 4, C and D). Moreover, hepatic levels of 4-HNE and MDA ( Figure 3B), as well as expression of phagocytic oxidase units, including p47 phox , p22 phox , p40 phox , and gp91 phox (Supplemental Figure 4E), were lower in ethanol-fed Ask1 -/mice than in WT mice. In addition, hepatic expression of several ER stress-associated genes, including Ero1a, Ero1b, Bim, Gadd34, Pdi, Bax, and Bak, were also markedly lower in Ask1 -/compared with WT mice after E10d+1B treatment (Supplemental Figure 4F). Finally, a decreased number of apoptotic TUNEL + hepatocytes was detected in ethanol-fed Ask1 -/mice compared with WT mice (Supplemental Figure 4G).
Further evaluation on the role of ASK1 in ALD was conducted using a pharmacological approach by treating chronic-plus-binge ethanol-fed mice with 2 ASK1 inhibitors (NQDI-1 and GS-4997). As shown in Supplemental Figure 5A, hepatic activation of ASK1 and p38, as measured by their phosphorylated forms, was strongly attenuated after treatment with these 2 inhibitors, with stronger inhibition by GS-4997. Treatment with NQDI-1 or GS-4997 ameliorated hepatic injury, as demonstrated by a significant reduction in serum ALT without an affect on the percentage or absolute count of circulating neutrophils ( Figure 4A). The numbers of MPO + neutrophils and F4/80 + macrophages in the liver were lower in ethanol-fed mice treated with NQDI-1 or GS-4997 compared with those treated with vehicle ( Figure 4B). Hepatic expression of F4/80 mRNA but not Ly6g was lower in mice treated with ASK1 inhibitors compared with those with treated with vehicle ( Figure 4B). Moreover, treatment with ASK1 inhibitors also reduced hepatic expression of some inflammatory mediators (e.g., Tnfa, Vcam1, and Mip1a) (Supplemental Figure 5B). ASK1 inhibitor treatment also ameliorated hepatic levels of 4-HNE and MDA ( Figure 5A). To understand the mechanism underlying the effects of ASK1 inhibitors on oxidative stress, we examined hepatic expression of NADPH-associated molecules. Five phagocytic oxidase (phox) units -p47 phox , p67 phox , p40 phox , gp91 phox , and p22 phox -were found to be downregulated after treatment with ASK1 inhibitors (Supplemental Figure 5C). Next, we examined hepatic ER stress by detecting hepatic expression of ER stress-associated genes and proteins. RT-qPCR revealed that hepatic expression of Ero1a, Ero1b, Gadd34, Pdi, Bax, and Bak was significantly decreased after treatment with ASK1 inhibitors (Supplemental Figure 5D). Western blot analysis also showed downregulation of several ER stress-related proteins (i.e., CHOP, p-JNK1/2) after Values represent mean ± SEM (each dot represents 1 mouse sample). Statistical evaluation was performed by Student's t test or 1-way ANOVA with Tukey's post hoc test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001. treatment with ASK1 inhibitors. Unexpectedly, hepatic expression of BCL2 was elevated in the ASK1 inhibitor groups ( Figure 5B). Finally, a reduced number of apoptotic TUNEL + hepatocytes was observed in ethanol-fed mice after ASK1 inhibitor compared with vehicle treatment (Supplemental Figure 5E).

Genetic deletion of the p38a gene in hepatocytes ameliorates ALD in chronic-plus-binge ethanol-fed mice.
To further examine the role of hepatic p38 MAPK, downstream of ASK1, in the pathogenesis of ALD, we generated mice with hepatocyte-specific deletion of p38α (p38a Hep-/mice) (p38α is the major form of p38 MAPKs in the liver) (18). As illustrated in Figure 6A, serum ALT levels as well as circulating neutrophil counts were markedly lower in p38a Hep-/compared with WT mice when they were treated with E10d+1B. The number of MPO + neutrophils in the liver as well as hepatic expression of Ly6g were significantly reduced in p38a Hep-/mice ( Figure 6, B and C). Similarly, F4/80 + macrophage infiltration decreased significantly; however, hepatic expression of F4/80 was similar in p38a Hep-/and WT mice ( Figure 6, B and C). Hepatic expression of several proinflammatory mediators is shown in Figure 6C. Expression of Il6 and Mip2 was significantly lower in ethanol-fed p38a Hep-/mice than in ethanol-fed WT mice. There was a trend toward a decrease in the expression of Tnfa, Mip1a, and Mcp1 in p38a Hep-/mice, although this difference was not statistically significant ( Figure 6C). Furthermore, a significant reduction in oxidative stress was observed in p38a Hep-/compared with WT mice after ethanol feeding, as evidenced by the expression of hepatic 4-HNE and MDA based on IHC staining ( Figure 6D). In contrast, there was no significant difference in hepatic expression of several phagocytic oxidase units and ER stress-associated genes in p38a Hep-/mice compared with WT mice after E10d+1B treatment, except for the significant downmodulation of Ero1b in p38a Hep-/mice (Supplemental Figure 6, A and B).
Treatment with p38α inhibitors ameliorates ALD in chronic-plus-binge ethanol-fed mice. To further dissect the role of p38 in ALD, we employed a pharmacological approach by treating ethanol-fed mice with p38 inhibitors: LY2228820, PH797804, and SB239063. A significant reduction in hepatic activation of p38 was observed after p38 inhibitor treatment (Supplemental Figure 7A). Interestingly, inhibition of p38 also resulted in decreased phosphorylation of ASK1 (Thr838) (Supplemental Figure 7A). In agreement with the observation in p38a Hep-/mice, a significant reduction in serum ALT levels and the percentage of circulating neutrophils was observed in C57BL/6J mice treated with LY2228820 and PH797804 compared with controls ( Figure 7A). Yet no significant change in serum ALT levels and circulating neutrophil counts was found in mice treated with SB239063 ( Figure 7A). Furthermore, hepatic infiltration of MPO + neutrophils was lower in all groups treated with p38 inhibitors compared with those with vehicle treatment ( Figure  7B), while hepatic expression of Ly6g was only significantly reduced in the PH797804-treated group (Figure 7C). Approximately 2-and 1.5-fold reductions in infiltrated F4/80 + macrophages were found in the LY2228820 and SB239063 treatment groups, respectively ( Figure 7B); hepatic expression of F4/80 mRNA was significantly (2-fold) downregulated in the LY2228820 treatment group but not in the PH797804 and SB239063 treatment groups ( Figure 7C). We next examined whether the decrease in neutrophil and macrophage infiltration in the liver was secondary to downregulation of inflammatory mediators in the liver. RT-qPCR analysis demonstrated that LY2228820 treatment downregulated most of the hepatic inflammatory cytokines, except for Il1b and Icam1. PH797804 treatment only led to the reduction of Il6 and Il1b, and SB239063 treatment had no effect on inflammatory cytokine gene expression ( Figure 7C). We also evaluated the effect p38 inhibitors on oxidative stress in the liver of E10d+1B-fed mice. As shown in Figure  8A, levels of hepatic 4-HNE and MDA expression markedly decreased in mice treated with p38 inhibitors. In particular, LY2228820 was the most potent oxidative stress inhibitor, as demonstrated by 12-and 13-fold reductions in the expression of 4-HNE and MDA, respectively ( Figure 8A). Additionally, an apparent decrease in hepatic expression of NADPH-associated molecules units -p47 phox , p67 phox , p22 phox , p40 phox , and gp91 phox -was observed only in LY2228820-treated mice (Supplemental Figure 7B). We also found differential effects of each p38 inhibitor on the expression of hepatic Mt1 and Mt2. LY2228820 treatment led to the induction of Mt1 and Mt2 by 2-and 1.7-fold, respectively, compared with vehicle treatment, while PH797804 treatment led to a reduction in their expression (Supplemental Figure 7B). Last, treatment with SB239063 had no effect on either Mt1 or Mt2 expression (Supplemental Figure 7B). Moreover, LY2228820 treatment led to a significant decrease in hepatic ER stress, as demonstrated by a reduction in protein and mRNA levels of CHOP ( Figure 8B and Supplemental Figure 7C). LY2228820 treatment attenuated levels of p-eIF2α protein without affecting total eIF2α protein ( Figure 8B).
Deletion or inhibition of ASK1 and p38 decreases mtDNA-enriched EVs in chronic-plus-binge ethanol-fed mice. We have previously demonstrated that chronic-plus-binge ethanol exposure elevates serum levels of mtDNA-enriched EVs in an ER stress-dependent manner, leading to liver inflammation and hepatocyte injury (8). We therefore reasoned that the amelioration of alcohol-induced liver injury by inhibition of ASK1 or p38 may be secondary to the reduction in mtDNA-enriched EV release. Compared with WT mice, Ask1 -/and p38a Hep-/mice had significantly lower levels of serum EVs and lower mtDNA load in these EVs, as determined by the quantification of cytochrome c oxidase ( Figure  9A). Similar observations were made in mice treated with either an ASK1 or p38 inhibitor ( Figure  9B). Next we asked whether ASK1/p38 pathway-mediated neutrophilia and liver injury are indeed mediated by mtDNA-enriched EVs. To test this hypothesis, we purified mtDNA-enriched EVs from the serum of E10d+1B-treated WT and Ask1 -/mice and injected these EVs into chronically ethanol-fed C57BL/6J mice (E10d). As illustrated in Supplemental Figure 8, injection of E10d+1B-treated WT mouse EVs induced a greater number and percentage of circulating neutrophils than injection of E10d+1B-treated Ask1 -/mouse EVs. However, the 2 types of EVs caused similar elevations in serum ALT levels in ethanol-fed mice. The lack of difference in serum ALT levels between these 2 groups may be because multiple pathways contribute to chronic-plus-binge ethanol-induced liver injury in addition to neutrophils.
To confirm that deficiency in Ask1 and p38a led to the reduction in the release of EVs and EV-containing mtDNA from hepatocytes, we isolated primary hepatocytes from WT, Ask1 -/-, or p38a Hep-/mice and treated them with either PBS or ethanol (100 mM for 24 hours). The quantity of EVs and EV-containing mtDNA was then determined from the supernatants. We found a significant reduction in the quantity of EVs released from the primary Ask1 -/or p38a Hep-/hepatocytes treated either with PBS or ethanol compared with WT hepatocytes ( Figure 9C). Moreover, treatment with ASK1 inhibitors (GS-4997 or NQDI-1) or p38 inhibitors (LY2228820 or PH797804) markedly reduced the number of EVs and EVs-containing mtDNA in the supernatant of primary WT hepatocytes (Figure 9, D and E).
Inhibition of ASK1 or p38 downregulates expression of EV biogenesis-and secretion-related genes in chronicplus-binge ethanol-fed mice. To understand how the ASK1/p38 pathway affects EV release and biogenesis in ethanol-fed mice, we performed RT-qPCR analyses of multiple genes related to EV biogenesis and trafficking processes (30). EV formation involves endosomal sorting complex required for transport (ESCRT) machinery, either in a dependent or independent manner. In the ESCRT-dependent pathway, 4 different protein complexes, ESCRT-0, -I, -II, and -III, and the associated AAA ATPase Vps4 complex act with other accessory proteins to drive multivesicular body (MVB) formation by invagination and scission of endosomal membranes (31). Certain ESCRT components, such as TSG101 (ESCRT-1) and ALIX (an ESCRT accessory protein), are commonly found in exosomes themselves. EVs can be  also formed in an ESCRT-independent manner. In this pathway, ceramide is generated from sphingomyelin via the action of neutral sphingomyelinase 2 (NSMase2), and inhibition of this enzyme or gene knockdown has been shown to block exosome release or exosome-mediated signaling in some cell types (32,33). Additionally, many small Rab GTPases play important roles in exosome production because they regulate intracellular trafficking of endosomes (from which MVBs originate) as well as transport of vesicles to the plasma membrane and their subsequent release from the cell (34). Rab27a and Rab27b are essential for exosome secretion (35).
Furthermore, we performed Western blot analyses to confirm the changes in protein expression related to the biogenesis or transportation of EVs. As illustrated in Supplemental Figure 9, treatment with ASK1 or p38 inhibitor markedly reduced expression of NSMase2, ALIX, and Rab27A proteins in AML12 cells (Supplemental Figure 9).

Discussion
Neutrophil infiltration is the pathological hallmark of alcoholic steatohepatitis; however, the mechanisms underlying hepatic neutrophil infiltration have not been fully elucidated. In this study, we examined oxidative stress-associated genes after chronic-plus-binge ethanol feeding by microarray analysis and found Mt1 and Mt2 genes were the 2 most highly elevated oxidative-stress-related genes. MT1 and MT2 are small, cysteine-rich proteins that play an important role in the intracellular stress response (36). Zhou et al. previously reported that ethanol intake-induced liver injury was exacerbated in Mt1/2-KO mice but reduced in Mt-transgenic mice (37,38). Here we found that deletion of the Mt1 and Mt2 genes aggravated alcohol-induced liver inflammation, injury, and oxidative stress after ethanol feeding, and subsequently enhanced activation of the stress kinases ASK1 and p38 in the liver. We further provide evidence suggesting that activation of the ASK1/p38 pathway induced mtDNA damage and release via EVs, and that these mtDNA-enriched EVs contribute to neutrophil infiltration in the liver.
The ASK1 pathway is thought to coordinately regulate gene expression, mitosis, metabolism, motility, survival, apoptosis, and differentiation (39). ASK1 has been considered to be the potential therapeutic target for patients with severe alcoholic hepatitis, and an ASK1 inhibitor, selonsertib (GS-4997), was administered to patients with severe alcoholic hepatitis in a recent phase II clinical study (40). Unfortunately, this study did not demonstrate favorable outcomes, as patients developed bacterial infection, and no improvement in mortality or liver function was observed (40). However, the pivotal role of ASK1 cannot be neglected; indeed, researchers remain committed to ASK1-targeted therapies in human disease. Schuster-Gaul et al.
found that pharmacologic inhibition of ASK1 could attenuate liver cell death and hepatic fibrosis driven by inflammasome activation using gain-of-function NOD-like receptor protein 3 (Nlrp3) mutant mice, providing mechanistic insight into the antifibrotic mechanisms of ASK1 inhibition (41). In the current study, we demonstrated that ASK1 plays an important role in promoting oxidative stress and liver injury in chronicplus-binge ethanol-fed mice. Inhibition or genetic deletion of ASK1 ameliorated liver injury in the E10+1B model as evidenced by decreased serum ALT, oxidative stress, and ER stress and reduced hepatic neutrophil Hep-/mice were subjected to E10+1B ethanol feeding and were euthanized 9 hours after gavage. Serum EVs were isolated and quantified by exosome quantitation assay. Cytochrome c oxidase (Cyto c ox) DNA levels in EVs were measured. (B) C57BL/6J mice were subjected to E10d+1B feeding and received i.p. injection of ASK1 or p38 inhibitors 30 minutes before gavage. Mice were euthanized 9 hours after gavage. Serum EVs were isolated and counted by exosome quantitation assay. Cyto c ox DNA levels in EVs were measured. (C) Hepatocytes were isolated from WT, Ask1 -/-, or p38a Hep-/mice and treated with ethanol (100 mM) or PBS for 24 hours. EVs were isolated from culture medium and counted by exosome quantitation assay. (D and E) Hepatocytes were isolated from WT mice and treated with ethanol (100 mM) or ethanol plus ASK1 or p38 inhibitors for 24 hours. EVs were isolated from culture medium and counted by exosome quantitation assay. Cyto c ox DNA levels in the EVs were measured. Values represent mean ± SEM (each dot represents 1 mouse sample). Statistical evaluation was performed by Student's t test or 1-way ANOVA with Tukey's post hoc test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001. Figure 10. Inhibition of ASK1 or p38 activation attenuates expression of the genes related to EV biogenesis and secretion in the E10d+1B model. (A and B) WT and Ask1 -/or p38a Hep-/mice were subjected to E10d+1B feeding and euthanized 9 hours after gavage. Liver tissues were subjected to RT-qPCR analysis of EV biogenesis-and secretion-related genes. (C and D) C57BL/6J mice were subjected to E10d+1B feeding and received i.p. injection of ASK1 inhibitor (3 mg/ kg GS-4997 or 4 mg/kg NQDI-1) or p38 inhibitor (3 mg/kg LY2228820 or 12 mg/kg PH797804) 30 minutes before gavage. Mice were euthanized 9 hours after ethanol gavage Liver tissues were subjected to RT-qPCR analysis of EV biogenesis-and secretion-related genes. Values represent mean ± SEM (each dot represents 1 mouse sample). Statistical evaluation was performed by Student's t test or 1-way ANOVA with Tukey's post hoc test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001. and macrophage infiltration. Our data suggest the possible beneficial roles of ASK1 inhibitors in the early stages of alcoholic steatohepatitis (as indicated by our chronic-plus-binge ethanol-fed model) but not in the severe form of the disease with advanced fibrosis/cirrhosis, inflammation, and bacterial infection (as those who were enrolled in the aforementioned phase II clinical trial) (40).
ROS can activate not only ASK1 but also other downstream kinases that may accelerate the progression of alcoholic steatohepatitis. The p38 MAPKs are known to be activated by MEK3 (or MKK3) and MEK6 (or MKK6); both MEK2 and MEK6 can be activated by one of the several upstream MAP3Ks, such as ASK1, dual-leucine zipper-bearing kinase 1 (DLK1), and TFG-β-activated kinase 1 (TAK1) (18). Once activated by MKK3 or MKK6, p38 MAPKs can potentiate the downstream signals or activate other proteins directly (42). The p38 family comprises 4 members, p38α (19), p38β (43)(44)(45), p38γ (46), and p38δ (47,48), which have been implicated in various cellular responses controlling cell proliferation, differentiation, apoptosis, embryogenesis, inflammatory responses, and stress responses in human reproductive tissues. A large body of evidence indicates that p38 activity is critical for normal immune and inflammatory responses (42). In our study, we focused on specific roles of hepatic p38α and its implication in the progression of alcoholic steatohepatitis. Our data revealed that hepatocyte-specific deletion of p38a or administration of p38 inhibitors markedly ameliorated serum ALT levels, oxidative stress, as well as ER stress, suggesting the role of hepatic p38-induced liver injury and inflammation in the chronic-plus-binge ethanol-feeding model. Interestingly, although ethanol-fed p38a Hep-/mice had lower serum ALT levels than ethanol-fed WT mice, hepatic steatosis was greater in p38a Hep-/mice than in WT mice (our unpublished data). These findings suggest that hepatic p38α plays distinct roles in controlling steatosis and liver injury in ALD, which is consistent with our recent studies indicating that hepatocyte-specific deletion of p38a increases simple steatosis but ameliorates oxidative stress-driven nonalcoholic steatohepatitis (NASH) in the HFD +Cxcl1 -induced nonalcoholic fatty liver disease (NAFLD) model (49). In addition, treatment with p38 inhibitors induced greater amelioration of ethanol-induced liver injury than genetic deletion of the p38a gene in hepatocytes. This is likely due to the effect of p38 inhibitors on other cell types in addition to hepatocytes that may contribute to alcohol-induced liver injury and inflammatory processes. Indeed, a recent study reported that macrophage p38α promoted the progression of steatohepatitis by inducing proinflammatory cytokine secretion and M1 polarization (26).
Although p38 has been implicated in various cellular responses including cell death and oxidative stress (42), the exact mechanisms by which p38 activation affects hepatocytes remain obscure. Wang et al. (50) reported that p38 was involved in cyclophilin D stimulation of permeability transition pore excessive opening, subsequently causing ER stress and hepatic steatosis. In the current study, we demonstrated that p38 activation promoted mtDNA damage in hepatocytes, which was then released via EVs and subsequently induced neutrophil activation. EVs are emerging as new players in the investigation of intercellular communication. Our mechanistic studies found that inhibition and genetic deletion of Ask1 and p38a ameliorated ethanol-induced liver injury by downregulating mtDNA-enriched EVs -one of the key factors in promoting ethanol-induced hepatic neutrophil infiltration and liver injury in the chronic-plus-binge ethanol feeding model (8). It is conceivable that attenuated ER stress and mitochondrial damage contributed to the decreased production of mtDNA-enriched EVs that occurred after ablation of ASK1 and p38 function. In addition, we also found that ethanol-mediated induction of EVs in circulation either in vivo or in hepatocytes in vitro was decreased after inhibition of ASK1 or p38, suggesting that the ASK1/p38 pathway plays an important role in the biogenesis of hepatocyte-derived EVs. Previous studies reported that inhibition of p38 in human aortic endothelial cells resulted in a 50% reduction in TNF-α-induced EVs (51). Additionally, p38 is essential for P2X7-dependent NSMase translocation to the plasma membrane, and its inhibition blocks capsaicin-induced EV shedding from microglia in vitro and in cortical brain slices (52,53). The ability of EVs to transfer cargo from donor to acceptor cells, thereby triggering phenotypic changes in the latter, has generated substantial interest in the scientific community. We and others have previously demonstrated that mtDNA-enriched EVs released by hepatocytes are able to induce neutrophil activation, resulting in liver inflammation (8,54). In addition, mtDNA can also induce NLRP3 inflammasome activation through engagement of Toll-like receptors in macrophages (55) and promote neutrophil extracellular trap formation, contributing to the progression of diseases (56)(57)(58). It will be interesting to examine whether these mechanisms also contribute to liver inflammation and injury induced by mtDNA-enriched EVs after chronic-plus-binge ethanol feeding.
In conclusion, our findings indicate that activation of ASK1 and p38 mediates alcoholic liver injury and inflammation by promoting proinflammatory mtDNA-enriched EV release in hepatocytes induced by ROS stress and ER stress. Despite the failure of anti-ASK1 therapy for patients with severe alcoholic hepatitis in a recent clinical trial (40), our study suggested that ASK1 and p38 activation plays an important role in the pathogenesis of the early stages of alcoholic steatohepatitis. Therapeutic intervention targeting this pathway in patients with a less severe form of alcoholic hepatitis deserves further investigation.

Methods
Mice. C57BL/6J, Mt1/2 -/-, and Albumin-Cre mice were obtained from the Jackson Laboratory. C57BL/6N mice were obtained from Charles River. Ask1 -/mice were originally obtained from Hidenori Ichijo (University of Tokyo, Tokyo, Japan) (59). Heterozygous mice were bred to generate Ask1 -/mice and their corresponding WT littermate controls. Mt1/2 -/mice with deletion of both Mt1 and Mt2 genes were backcrossed to a C57BL/6N background for more than 15 generations, and C57BL/6N mice were used as controls. p38a-floxed mice were provided by Yibin Wang (UCLA, Los Angeles, California, USA) and were crossed with Albumin-Cre mice via several steps to generate hepatocyte-specific p38a deletion. All animals were cared for in accordance with NIH guidelines. Mice were housed in a temperature-controlled room with a 12-hour dark/12-hour light cycle.
Chronic-plus-binge ethanol feeding protocols. The chronic-plus-binge ethanol feeding model was described previously (4,27). Briefly, 10-to 12-week-old female mice were initially fed a Lieber-DeCarli liquid diet (F1259SP, Bio-Serv) ad libitum for 5 days during acclimatization. Subsequently, the ethanol-fed groups were allowed free access to an ethanol diet (F1258SP, Bio-Serv) containing 5% (v/v) ethanol for 10 days or 8 weeks, then received a single dose of ethanol (5 g/kg body weight) via oral gavage in the early morning and were sacrificed 3, 6, or 9 hours later. In addition, mice were fed an ethanol-containing Lieber-DeCarli diet for 10 days as described for chronic feeding (designated as E10d) or received a single dose of ethanol (5 g/kg body weight) via gavage in the early morning (designated as 1B) and were sacrificed 9 hours later. In some experiments, mice received ASK1 inhibitor GS-4997 at 3 mg/kg or NQDI-1 at 4 mg/kg by i.p. injection; or p38 inhibitor LY2228820 at 3 mg/kg, PH797804 at 12 mg/kg, or SB239063 at 10 mg/kg by i.p. injection approximately 30 minutes before oral gavage. All inhibitors were obtained from Selleck Chemicals.
Microarray analyses of mouse liver samples. Dye-coupled cDNAs were purified with a MiniElute PCR purification kit (QIAGEN) and hybridized to a 44K mouse 60-mer oligo microarray (Agilent Technologies). Data were processed and analyzed using GeneSpring GX software (Agilent Technologies). Full microarray data were deposited in the NCBI's Gene Expression Omnibus database (GEO GSE67546).
Biochemical assays. Serum ALT levels were determined from serum collected from mouse retro-orbital plexus using a Catalyst Dx Chemistry Analyzer (IDEXX Laboratories) according to the manufacturer's instructions.
TBARS assay. Approximately 25-mg liver tissue samples were homogenized or lysed in RIPA buffer containing a cocktail of protease inhibitors (Santa Cruz Biotechnology). The samples were then vortexed and centrifuged at 1600 g at 4°C for 10 minutes. The supernatant was used for analysis according to the TBARS assay kit manufacturer's instructions (Cayman).
Liver histology and IHC staining. Liver specimens were fixed in 10% buffered formalin and embedded in paraffin. Then, 4-μm sections were cut and used for IHC staining. TUNEL staining was performed with an ApopTag Peroxidase In Situ Apoptosis Detection Kit (Millipore). For IHC, after heat-induced epitope retrieval, paraffin-embedded sections were incubated in 3% H 2 O 2 and blocked in 3% normal serum buffer. Sections were incubated with primary antibodies overnight at 4°C. Vectastain Elite ABC Staining Kit and DAB Peroxidase Substrate Kit (Vector Laboratories) were used to visualize the staining according to the manufacturer's instructions, as described previously (60). The following primary antibodies were used: MPO (PP023AA, Biocare Medical), F4/80 (30325, Cell Signaling Technology), MDA (1F83, Genox), 4-HNE (HNEJ-2, Genox), p-ASK1 (PA5-64541, Thermo Fisher Scientific), and p-p38 (4511s, Cell Signaling Technology). Positive cells and positive areas in 10 randomly selected high-power fields were analyzed.
Total RNA isolation and RT-qPCR. Total RNA was purified from liver tissues or cell cultures using TRIzol Reagent (Thermo Fisher Scientific) according to the manufacturer's instructions. RNA (1 μg of total) was reverse transcribed into cDNA using a High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Expression levels of mRNA were measured by RT-qPCR with an ABI7500 RT-PCR system (Applied Biosystems). 18S rRNA was used as the invariant housekeeping control gene. The 2 −ΔΔCt method was used to calculate relative gene expression. The primers used for RT-qPCR are listed in Supplemental Table 1.
Isolation and culture of mouse primary hepatocytes and AML12 cells. For hepatocyte isolation, mice were anesthetized with 30 mg/kg pentobarbital sodium i.p. The portal vein was cannulated and the liver perfused with ethylene glycol tetraacetic acid and subsequently with 0.075% collagenase (Worthington) as previously described (61). Primary hepatocytes were collected after centrifugation at 50 g for 5 minutes and cultured in DMEM containing 10% EV-free FBS and penicillin-streptomycin. In some experiments, primary hepatocytes were cultured with either ethanol at 100 mM or ethanol plus 30 μM NQDI-1, 30 μM GS-4997, 2 μM LY2228820, or 4 μM PH797804, respectively, for 24 hours. All inhibitors were obtained from Selleck Chemicals. AML12 cells (mouse hepatocyte cell line) were cultured in a 1:1 mixture of DMEM and Ham's F12 medium with 5 μg/mL insulin, 5 μg/mL transferrin, 5 ng/mL selenium, 40 ng/mL dexamethasone, 10% FBS, and penicillin-streptomycin.
EV isolation and cell treatment. EVs were isolated as previously described (62). Briefly, after centrifugation of serum or cultured medium at 3000 g for 15 minutes at 4°C, the supernatant fraction including EVs was incubated with an appropriate volume of ExoQuick or ExoQuick-TC exosome precipitation solution (System Biosciences), depending on the type of samples for EV isolation procedure, according to the manufacturer's manual. The size of EVs was assessed using a NanoSight NS300 System (Malvern Instruments) equipped with a fast video capture and nanoparticle tracking analysis, and the concentration of EVs was measured using a Pierce BCA protein assay Kit (Thermo Fisher Scientific) and EXOCET Exosome Quantitation Kit (System Biosciences).
Injection of serum EVs. EVs were isolated from 1 mL serum from E10d+1B-fed WT and Ask1 -/mice as described above and then measured using an EXOCET Exosome Quantitation Kit (System Biosciences). The EV pellets (~1 × 10 12 particles) were resuspended in 150 μL PBS and injected through the tail vein into E10d-fed C57BL/6J mice.
Quantification of mtDNA in EVs. After quantification of EVs using a BCA protein assay, an equal amount of EVs was subjected to DNA extraction using a QIAmp DNA Mini Kit (QIAGEN) as previously described (62). The extracted DNA was further quantified using NanoDrop spectrometer (Thermo Fisher Scientific), and an equal amount of DNA was subjected to real-time PCR reaction using the primers for mitochondrial genome-specific genes (e.g., cytochrome c oxidase). The primer sequences used for PCR reactions are listed in Supplemental Table 1.
Statistics. Results are expressed as mean ± SEM. All statistical analyses were performed using Graph-Pad Prism software (v. 7.0a; GraphPad Software). To compare values obtained from 2 groups, 2-tailed Student's t test was performed. Data from multiple groups were compared with 1-way ANOVA followed by post hoc analysis with Tukey's test, as appropriate. P values less than 0.05 were considered significant.
Study approval. Animal care and experiments were conducted under guidelines and protocols approved the National Institute on Alcohol Abuse and Alcoholism (NIAAA) Animal Care and Use Committee.

Author contributions
JM and HC designed and conducted the experiments and wrote the manuscript; RMR and SL helped with the experiment design, performed data analysis and edited the manuscript; MX, TR, YH, SH, DF, RR, and PY conducted some experiments; JS and BG are co-advisors for JM's PhD thesis, supervised the entire study, and edited the manuscript.