Male C57Bl/6 mice were subjected to the ASH regimen, which consisted of feeding a solid Western diet high in cholesterol and saturated fat (HCFD) or regular chow ad libitum for two weeks. This initial acclimation period was followed by intragastric (ig) feeding of ethanol and a high-fat liquid diet (36%Cal corn oil) at 60% of their total caloric intake plus ad libitum intake of HCFD for the remaining 40% of calories. The ethanol dose was increased to 27 g/kg/day over an eight-week period, and pair-fed (PF) control mice were fed an isocaloric high-fat liquid diet. Additionally, mice were subjected weekly to an alcohol binge (5 g/kg) from the second week of ig feeding as described(31). After eight weeks of feeding, mice were euthanized between 10:30am to 1pm one day after the final binge. All experimental protocols were approved by the University of Southern California Institutional Animal Care and Use Committee.

Hepatocyte-specific Mkp1 knockout mice were generated using a Cre-lox mechanism as described (32). Male and female hepatocyte-specific Mkp1^-/- knockout (abbreviated as liver-specific knockout, LSKO) and their littermate control (Mkp1^+/+ “f/f”) mice were subjected to the chronic plus binge model (33). Mice were housed in a pathogen-free, temperature-controlled animal facility with 12 h light/12 h dark cycles. Mice were pair-fed the Lieber-DeCarli liquid diet containing either isocaloric maltose dextrin (PF) or 5% (w/v) ethanol (AF) ad libitum for ten 10 days. Food consumption was monitored daily. f/f and LSKO mice on alcohol diet consumed in average 8 ml/day. Pair-feeding was done by giving the same amount of control diet (8 ml) to mice in PF group. On day 11, mice on the alcohol-fed diet (AF) were given 31.5% (v/v) ethanol via oral gavage, while mice on the isocaloric maltose dextrin control diet (PF) were gavaged with water. Mice were euthanized 9 hours after gavage. There were 8-10 mice in each experimental group. Whole blood was collected, and plasma samples were stored at -80°C until processing. Tissues were snap-frozen and stored at -80°C before analysis. Additional tissues were fixed in 10% neutral-buffered formalin and cryo-preserved for immunohistochemical analysis. All experimental protocols were approved by the University of Louisville Institutional Animal Care and Use Committee in accordance with the National Institutes of Health Office of Laboratory Animal Welfare Guidelines.

Hepatocytes were isolated from male and female hepatocyte-specific Mkp1 ^-/- knockout (LSKO) and their littermate control (“f/f”) mice by collagenase perfusion as described (34). Cells were plated in 6 well culture plates (cat# 92006, Midwest Scientific, Valley Park, MO) at 0.5x10⁶cells/well. After attachment, media was replaced with DMEM supplemented with 5% FBS and 1% P/S (cat# 15140-122, Gibco). Ethanol (50 mM) was added to ethanol-treated cells for indicated times.

Liver cytokines were measured using the MesoScale Discovery (MSD) platform. Total liver tissue samples were lysed with a buffer of 200 mM NaCl, 5 mM EDTA, 10 mM Tris, 10% glycerin, and protease/phosphatase inhibitors (cat#78445, ThermoFisher Scientific). The incubation of liver tissue lysates with the V-Plex MULTI-SPOT Assay System from MSD was carried out overnight at 4°C to increase sensitivity of the assay. Proinflammatory Panel 1 (mouse) and Cytokine Panel 1 (mouse) kits were used (MesoScaleDiscovery, Rockville, MD, cat # K15048D and K152245G, respectively). The plate was read with a MESO QuickPlex SQ 120 imager and analyzed using Discovery Workbench v4.0 software. The assay was performed according to the manufacturer’s instructions.

Plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were measured using colorimetric enzymatic assay kits (cat# MAK052 (ALT) and cat# MAK055 (AST), Millipore Sigma, St. Louis, MO), according to the manufacturer’s instructions.

Plasma glucose levels were measured using glucose colorimetric detection kit (Invitrogen, Carlsbad, CA; Cat# EIAGLUC). Plasma was diluted 20 times before measurement. Data are presented in mg/dL.

Total RNA from mouse liver tissue and cells was isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). RT-qPCR was performed as described previously (35, 36). Specific primers were purchased from Integrated DNA Technologies (IDT, Coralville, IA, USA). Primer sequences are presented in Table 1 .

Liver tissue was lysed using RIPA buffer (50mM Tris HCl at pH 8, 150mM NaCl, 1mM EDTA, 1% NP40, 0.5% Na-deoxycholate, and 0.1% SDS) containing protease and phosphatase inhibitors (ThermoFisher Scientific). Proteins (25 µg of cell lysate and 50 µg of liver tissue lysate) were analyzed by SDS-polyacrylamide gel electrophoresis using a BioRad electrophoresis system (Hercules, CA, USA). Membranes were stripped using Restore™ PLUS Stripping Buffer (cat# 46430, Thermo Scientific, Rockford, IL, USA) and reprobed using another antibody. Immunoreactive bands were visualized using enhanced chemiluminescence light detection reagents (cat# WBLUF0500, Millipore, Burlington, MA, USA). Quantification was performed with Image Lab™ software (Bio-Rad, version 5.2.1). MKP1 antibody was purchased from Santa Cruz Biotechnology (sc-373841) and used at 1:500 dilution. CYP2E1 antibody was from Abcam (ab28146, dilution 1:2500). Antibodies to GAPDH (#5174), phospho-JNK (Thr183/Tyr185, #9251), JNK (#9252), CHOP (#5554), ATF3 (#18665), ATF4 (#11815), phospho-Akt (Ser473) and Akt (#9272) were purchased from Cell Signaling Technology.

5 µm of paraffin sections were de-paraffinized in xylene and rehydrated in a series of ethanol and water, then PBS. Sections were then stained with Chemicon^® ApopTag^® Peroxidase In Situ Apoptosis Detection Kit (Sigma-Aldrich, Burlington, MA) until equal staining was achieved across samples. The sections were then dehydrated and mounted.

Total RNA was extracted from liver tissue by TRIzol and a NanoDrop 2000 spectrophotometer (ThermoFisher) was used to determine RNA concentration. mRNA was enriched with magnetic beads containing oligo(dT), and Fragmentation Buffer was added to the mRNA to cut fragments into short fragments. Then, short fragments were used as templates to synthesize the first cDNA strand with random hexamers. Buffer, dNTPs, RNase H, and DNA polymerase I were added to synthesize the second cDNA strand. After purification (QiaQuick PCR kit, cat # 28104), cDNAs were repaired and were added with base A, then were recovered with the target size fragments by agarose gel electrophoresis. PCR amplification was conducted to complete the entire library preparation. HiSeq2000 (Illumina) was used for the sequencing of the constructed library. Quality control was performed using FastQC (v.0.10.1) (37). Sequenced reads were of good quality, and no sequence trimming was necessary. The sequences were aligned to the mouse reference assembly (mm10) using the TopHat2 aligner (v.2.0.13) (38) with alignment rates above 97.5%. Gene read counts were generated using Cuffnorm (39) with FPKM normalization. Differential expression analysis was performed using Cuffdiff2 (v.2.2.1) (39).

Statistical analysis was performed using GraphPad Prism version 9 for Windows (GraphPad Software, Inc., La Jolla, CA). Data were analyzed, as appropriate, by the Student’s t-test (for two groups) or by two-way analysis of variance (ANOVA) with Bonferroni post-test analysis (for greater than two groups). Data points with a distance greater than two standard deviations from the mean were considered outliers and removed from the data set. Differences were considered statistically significant at p< 0.05.

Several murine models of alcohol exposure have been developed to mimic different stages of ALD to study its mechanisms (for review see (40)). Tsukamoto’s group developed intragastric chronic alcohol feeding model of ASH, which recapitulates several features of human ASH (steatosis, inflammation and pericellular fibrosis). Liver injury is more severe in this model, as indicated by high ALT levels with some mice reaching 500 U/L (31). In this model, a hybrid feeding involving 40% of total calories from solid chow, and the remaining intake in the form of an intragastric liquid diet was administered as shown in Figure 1A . As reported previously, this approach led to severe liver injury as demonstrated by high serum ALT levels, immune cell infiltration, and lipid accumulation in the liver ( Figures 1B, C ). Investigation of differential changes in gene expression through RNA sequencing analysis showed that alcohol-fed (AF) mice had significantly elevated levels of ER stress markers such as activating transcription factor 3 (ATF3) and 4 (ATF4), DNA damage-inducible transcript 3, also known as C/EBP homologous protein (CHOP), DNA damage-inducible alpha (GADD45a) and beta (GADD45b) compared to pair-fed counterparts ( Figure 1D ). RNA sequencing analysis also showed that alcohol feeding resulted in significant changes in the expression of several dual specificity phosphatases (Dusps or Mkps) ( Figure 1E ). Notably, while Dusp3, Dusp4, Dusp5, Dusp6, Dusp8, Dusp10, and Dusp14 were upregulated, only Dusp1 (Mkp1) was significantly downregulated, which was confirmed to be significantly decreased as shown by real time qPCR analysis ( Figure 1F ). This result shows that exposure to alcohol led to decreased liver expression of Dusp1 mRNA levels and increased ER stress-associated genes.

To show a causal role of hepatocyte MKP1 in alcohol-induced liver injury, we subjected wild type (f/f) and hepatocyte-specific Mkp1 knockout (LSKO) male and female mice to a NIAAA chronic plus binge model of ALD ( Figure 2A ). Alcohol feeding did not alter hepatic Mkp1 mRNA levels in male f/f mice, while alcohol-fed female f/f mice showed significantly lower levels when compared to pair-fed counterparts ( Figure 2B ). To examine the direct effect of alcohol on hepatocyte MKP1, we isolated hepatocytes from male and female mice and treated them with 50 mM ethanol for 24 and 48 hours. mRNA and protein expression analyses of MKP1 showed that ethanol had a much more pronounced effect on female hepatocytes ( Figures 2C, D ). Hepatocytes from LSKO mice showed no expression of MKP1 protein as expected ( Figure 2D ).

Alcohol feeding in vivo resulted in a modest increase in ALT levels in male f/f mice, while it led to a significantly higher levels of ALT in LSKO mice ( Figure 3A ). AST levels were also much higher in LSKO male mice ( Figure 3B ). In contrast, alcohol feeding of female f/f mice led to a significant increase in ALT levels, which was further elevated in LSKO mice ( Figure 3A ). AST levels were also increased in AF fed f/f and LSKO female mice, however there was no difference between the genotypes ( Figure 3B ).

To evaluate the degree of apoptosis in the livers, tissue samples were stained for terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) ( Figure 3C ). Five images were gathered randomly from each liver, and stained TUNEL-positive cells were counted ( Figure 3D ). A response similar to that of ALT levels was observed in apoptosis indices for each group, although there were some differences between male and female mice. Specifically, in LSKO-PF female mice, fewer apoptotic cells were found; nevertheless, alcohol increased apoptotic cell numbers similar to their f/f counterparts. Taken together, these observations show that Mkp1 deletion augmented alcohol induced liver injury in both males and females. However, alcohol had a more pronounced effect on liver injury and Mkp1 expression in f/f females, which suggests that alcohol mediated decrease in Mkp1 expression might be a contributing factor to their susceptibility to alcohol induced liver injury compared to males.

Steatosis and altered lipid metabolism are hallmarks of ALD. To determine whether Mkp1 deletion influenced pathways involved in alcohol mediated lipid accumulation in the liver, we performed real time qPCR for genes involved in de novo lipogenesis, fatty acid synthase (Fasn) and acetyl-CoA carboxylase (Acaca). Alcohol-fed f/f male mice showed a modest increase in Fasn mRNA levels and no change in Acaca ( Figure 4A ). In LSKO mice, alcohol feeding had no effect on either Acaca or Fasn. Carnitine palmitoyl transferase I (Cpt1a), which is essential for fatty acid β-oxidation, was significantly decreased in male f/f and LSKO mice following alcohol feeding ( Figure 4A ). Expression of peroxisome proliferator activated receptor alpha (Pparα), a critical transcription factor for Cpt1a, was also decreased in both genotypes ( Figure 4A ). Baseline mRNA levels of Acaca and Fasn were significantly higher in PF LSKO females in comparison to f/f ( Figure 4C ), suggesting that Mkp1 deletion increases lipogenesis in female mice. Alcohol feeding led to a modest increase of both Acaca and Fasn mRNA levels in f/f females but decreased their levels in LSKO females ( Figure 4C ). Both alcohol-fed f/f and LSKO female mice showed decreased levels of Cpt1 and Pparα when compared to PF mice. Examination of hematoxylin and eosin-stained slides showed similar lipid accumulation in alcohol fed male mice of both genotypes ( Figure 4B ). In comparison, alcohol fed f/f females had more steatosis when compared to male mice, however LSKO female mice tended to show less steatosis when compared to their alcohol fed f/f counterparts ( Figure 4D ). This observation is consistent with changes in mRNA levels of Fasn and Acaca ( Figure 4C ). Since CYP2E1-mediated alcohol metabolism plays a major role in alcohol induced liver steatosis and injury, we also examined whether sex and Mkp1 deletion modulate CYP2E1 protein levels. Our data show that CYP2E1 protein expression was significantly increased by alcohol feeding in both f/f males and females ( Figure 4E, F ). However, this increase was significantly higher in LSKO mice when compared to their PF controls. These results suggest that Mkp1 could play a role in ethanol metabolism.

Recent studies have shown that chronic alcohol consumption is associated with dysregulated glucose metabolism and insulin resistance in the liver (41, 42). Hence, we examined whether Mkp1 deletion influenced glucose metabolism in male and female mice. Baseline plasma glucose levels were lower in PF LSKO male mice when compared to their f/f counterparts ( Figure 5A ). Alcohol feeding decreased glucose levels in male f/f mice but not in LSKO mice ( Figure 5A ). Baseline mRNA levels of Glycogen synthase (Gys2), Pyruvate kinase (Pklr), and Glucose-6-phosphatase catalytic subunit 1 (G6pc1) were not statistically different in f/f and LSKO mice ( Figure 5 ). Alcohol did not change mRNA levels of G6pc1, but decreased Gys2 in both f/f and LSKO mice ( Figures 5B, G ). Pklr mRNA levels were significantly decreased only in f/f mice by alcohol feeding ( Figure 5C ). pAkt levels increased in alcohol fed male f/f and LSKO mice, however the levels were slightly higher in AF LSKO mice ( Figure 5E ). In females, alcohol feeding resulted in decreased plasma glucose levels in both f/f and LSKO mice, however in f/f mice it did not reach significance ( Figure 5F ). Baseline levels of Pklr mRNA were lower and G6pc1 higher in LSKO females ( Figures 5G, H ). Alcohol had no effect on the expression of G6pc1 in either genotype, but decreased Gys2 levels in both f/f and LSKO females. Pklr levels were slightly increased in alcohol fed f/f but decreased in LSKO females ( Figures 5H ). pAkt levels did not increase with alcohol feeding in f/f females, while we saw higher pAkt levels in alcohol fed LSKO females suggesting that Mkp1 deletion influenced insulin sensitivity ( Figure 5J ). Overall, these observations indicate that hepatocyte Mkp1 deletion influences glucose metabolism and insulin sensitivity in male and female mice differently.

Endoplasmic reticulum stress is strongly associated with ALD and has been shown to contribute to alcohol-induced liver injury (43). To evaluate whether hepatocyte-specific Mkp1 deletion influenced ER stress, we examined the mRNA expression levels of genes associated with ER stress. Real-time qPCR analyses showed that alcohol feeding significantly increased mRNA levels of Chop and Gadd45b in f/f male mice. Levels of Atf3, Atf4, and Gadd45a also increased with alcohol exposed f/f males, though the changes did not reach a statistical significance ( Figure 5A ). Notably, Mkp1 deletion resulted in a statistically significant increase in the ER stress response to alcohol as evidenced by a significantly higher mRNA levels of Atf3, Atf4, Gadd34, Gadd45a, and Gadd45b in male LSKO mice ( Figure 6A ). In comparison, Mkp1 deletion led to additional increases in mRNA levels of Atf4, Chop and Gadd45b in alcohol fed LSKO female mice when compared to their f/f counterparts ( Figure 6C ). Western blot analyses of ER stress proteins ATF3, ATF4 and CHOP confirmed that hepatocyte Mkp1 deletion increased the ER stress response in both male and female mice ( Figures 6B, D ).

Hepatic levels of cytokines and chemokines were analyzed in all treatment groups by using MesoScale Discovery V-PLEX assays. No difference was found between sexes or genotypes in the expression of MCP-1 (Ccl2), IL-1β, IL-4, IL-5, IL-6, IL-10, and IL-12p70 (data not shown). Interestingly, male mice had lower baseline TNFα, IL17 and IL9 protein levels and higher levels of KC-GRO than those from females ( Figure 7 ). Alcohol further decreased TNFα levels in males but increased them in female mice. This increase was much more pronounced in LSKO female mice ( Figure 7 ). KC-GRO, IL-9, IL-15, IL27p28/IL-30, and IL-17 expression did not change among males of all genotypes and feeding groups. MIP-2, IL-9, IL-15, IL-17, and IL27p28/IL-30 levels were significantly increased with alcohol feeding in LSKO female mice. Expression of IL-33 was unchanged in male mice of all groups but was significantly decreased in female alcohol-fed mice of both genotypes compared to their pair-fed counterparts ( Figure 7 ). Overall, we did not observe significant changes in inflammatory response to alcohol in male mice of either genotype in this model. However, there was a modest increase in the response of female mice, which was significantly enhanced by Mkp1 deletion.