Fat-1 transgenic mice that have been engineered to express the C. elegans n3-fatty acid desaturase gene (fat-1), and therefore have increased tissue n3-PUFAs without the need for dietary intervention, were obtained from J.X. Kang and have been described previously (Kang et al., 2004). These mice were bred in the Association for Assessment and Accreditation of Laboratory Animal Care-accredited animal facility at the University of Louisville, maintained as heterozygotes, and genotyped as previously described (Warner et al., 2019) using primer sequences to the fat-1 transgene (with Gdf5 as an internal control, Table 1). Tissue n3-PUFA enrichment and a decrease in the n6/n3-PUFA ratio in fat-1 mice has been confirmed in our previous reports (Warner et al., 2019; Hardesty et al., 2021), and is consistent with that reported by J.X. Kang for these mice (Kang et al., 2004). 10–12 week old wild type (WT) and fat-1 ^+/- male littermates were placed on an all liquid diet that contained 5% (v/v) EtOH for 10 days followed by a single binge of 5 g/kg EtOH on day 11 [the NIAAA, or 10 + 1 model (Bertola et al., 2013),]. Pair-fed (PF) control mice were also provided an all-liquid diet, but with maltose dextrin supplying the same number of calories as EtOH (diets are the Lieber-DeCarli Rodent Liquid Diets (control [PF], F1259SP; EtOH, F1258SP) purchased from Bio-Serv, Flemington, NJ). According to the manufacturer, the PUFA content of both diets was 9.7 g/L, divided between 9.3 g/L linoleic acid (n6) and 0.3 g/L alpha linolenic acid (n3), indicating an n6:n3 ratio of 31:1. Animals were euthanized 9 h following the EtOH binge, and liver tissue was collected and snap-frozen for future analyses. Blood was collected from the inferior vena cava using heparinized 25 g needles, and plasma was separated by centrifugation at 7,000 g for 5 min. Daily food consumption and beginning/ending body weights were recorded. There were four experimental groups in total: WT-PF, WT-EtOH, fat-1-PF, and fat-1-EtOH (n = 8–14 per group). This study was performed under IACUC protocol number 15423 to I.A.K. Note: for simplicity, “fat-1 ^+/- ” mice are referred to as “fat-1”.

Liver samples (∼50 mg) were homogenized using water at a ratio of 1 mg sample per 10 µl solvent. 200 µl ethanol was mixed with 200 µl of each homogenized sample to extract the PUFAs. After vortexing, the mixture was centrifugated at 14,000 rpm for 15 min. Solid phase extraction was used to purify and concentrate PUFAs using a Waters Oasis HLB cartridge (1 ml/30 mg). The cartridge was conditioned with 1 ml methanol followed by 1 ml water. After loading all supernatants, the cartridge was washed with 1 ml 5% methanol (v/v). The PUFAs were eluted with 1 ml methanol/acetonitrile (10/90, v/v). The extract was dried by evaporation under a gentle nitrogen gas stream. The dried sample was then redissolved in 75% ethanol for liquid chromatography-mass spectrometry (LC-MS) analysis using a Waters Acquity H-class UPLC system (Milford, MA, United States) coupled with a Waters Xevo TQ-S micro triple quadrupole mass spectrometer (Milford, MA, United States). The chromatographic separation was carried out on a Waters Acquity UPLC BEH C8 2.1 × 100 mm, 1.7 µm column (Milford, MA, United States) equipped with a guard column. The mobile phase consisted of A: 0.1% formic acid in water (v/v) and B: 0.1% formic acid in acetonitrile (v/v). LC-MS conditions are the same as in Yuan et al. (2020). Briefly, the column temperature was held at 40°C. Linear gradient elution was performed at 0.4 ml/min starting at 30% B for 3 min (0–3.0 min), increased to 99% B over 20 min (3.0–20 min), then continued at 99% B over 25 min (20–25 min), and finally returned to 30% B at 25.1 min for column re-equilibration (25.1–28 min). The MS detection was performed using electrospray ionization in negative mode. The relative quantification of each PUFA compound was achieved by multiple reaction monitoring by measuring peak area. Our relative measurement does not allow direct calculation of the n6:n3-PUFA ratio, instead, n3-and n6-PUFAs are reported separately. The data are reported as average fold-change between the two most highly abundant PUFAs for each class (arachidonic acid [AA] and linoleic acid [LA] for n6 and eicosapentaenoic acid [EPA] and docosahexaenoic acid [DHA] for n3) relative to WT PF (set as 1). PUFAs were measured in all mice (n = 8–14 per group).

Plasma alanine aminotransferase activity (ALT, an indirect measure of hepatocyte cell death) was determined using the ALT/GPT reagent as per the manufacturer’s instructions (Thermo Fisher Scientific, Waltham, MA). Liver tissue hematoxylin and eosin (H&E) staining for gross tissue morphology and Oil Red O staining for neutral lipids were performed as described previously (Hardesty et al., 2021). From H&E-stained sections, lipid accumulation was quantified by steatosis scoring, where a score of 0 indicates 0–5% tissue coverage by lipid droplets, 1 = 5–33% coverage, 2 = 33–66% coverage, and 3 = 66–100% coverage (Bedossa and Consortium, 2014). Oil Red O staining was quantitated by capturing 10 random digital photomicrographs, then using a macro in ImageJ to measure the % of the area within the field that was colored red. Total hepatic triglycerides (TGs) were measured using the triglycerides reagent from Thermo Fisher Scientific. Hepatic neutrophils were determined using an ELISA to myeloperoxidase, an enzymatic marker for neutrophils (LifeSpan BioSciences, Inc. Seattle, WA), and by immunohistochemistry with anti-MPO antibodies (R&D Systems, Minneapolis, MN). MPO immunohistochemistry was quantitated by counting MPO+ cells in ten random digital photomicrographs at 200X magnification. The thiobarbituric acid-reactive substances (TBARS) assay was performed on liver tissue to evaluate oxidative stress (Cayman Chemical, Ann Arbor, MI).

Total RNA from liver samples was purified with Trizol (Thermo Fisher Scientific), contaminating DNA was removed with DNase I (Thermo Fisher Scientific), and cDNAs were synthesized and quantitated using reagents from Quanta Biosciences (Beverly, MA). For gene expression analysis, 10 ng cDNA was assayed (0.01 ng for 18 s) on an Applied Biosystems StepOne real-time PCR instrument (Thermo Fisher Scientific). Data were reduced using the ΔΔCt method (Livak and Schmittgen, 2001). Primer sequences are provided in Table 1.

Liver immune cell isolation was conducted as previously described (Chu et al., 2020). Briefly, livers were homogenized by mechanical disruption with a rubber-tipped syringe plunger and then passed through a 70 µm strainer to obtain a single cell suspension. Immune cells were labeled using the FOXP3/Transcription Factor Staining Buffer Set per the manufacturer’s protocol (Thermo Fisher Scientific, Waltham, MA), followed by labeling with antibodies obtained from Thermo Fisher, as listed below. To label T regulatory cells (Tregs, CD4⁺ FOXP3⁺), APC-labeled anti-CD4 (RM4-5) and FITC-labeled FOXP3 (FJK-16s) were used. To label Kupffer Cells [KCs, F4/80⁺ CD11b^lo LY6C⁻ (Alisi et al., 2017; Lynch et al., 2018)] and determine the M1 [CD11c⁺ CD206^-) and M2 (CD11c⁻ CD206⁺ (Triantafyllou et al., 2021)] KC abundance, APC-labeled anti-CD11c (Bu15), PerCP-Cy5-labeled anti-CD11b (ICRF44), FITC-labeled anti-LY6C (RB6-8C5), PE-labeled anti-F4/80 (BM8), and Pacific Blue-labeled anti-CD206 (19.2) were used. Natural killer cells (NK1.1⁺) were identified with PE-labeled anti-NK1.1 (PK136). Finally, conventional cytotoxic T lymphocytes (TCRβ⁺ CD8a⁺) were identified with APC-labeled TCR beta (H57-597) and PerCP-Cy5-labeled anti-CD8a (53–6.7). Flow cytometry data was collected on a BD FACSCanto II Flow Cytometer and analyzed with FlowJo Software (v10.7, BD Biosciences, Franklin Lake, NJ). Gating strategy is summarized in Supplementary Figure S1. n = 3–4 mice per group were used.

Liver tissue was homogenized by sonication in 20 mM Tris (pH 7.5), 2 mM EDTA, 10 mM EGTA, 1% Triton X-100, and protease/phosphatase inhibitors (Thermo Fisher Scientific). Insoluble material was removed by centrifugation at 10,000 × g for 10 min, and protein concentrations were measured (Bicinchoninic Acid Assay, Pierce Chemical Company, Rockford, IL). Samples (50 μg protein) were separated by SDS-PAGE, electroblotted onto nylon membranes (PVDF), and then probed with primary antibodies overnight at 4°C followed by a 1 h incubation with HRP-conjugated secondary antibodies (Thermo Fisher Scientific). Signals were visualized using Clarity Max™ Western ECL substrate and images were collected with the ChemiDoc™ imaging system and quantitated with Image Lab software, version 6.0.1 (Bio-Rad Laboratories, Hercules, CA). Anti-CYP2E1 antibodies were obtained from Abcam (Cambridge, MA, catalog number 28146), and anti-GAPDH antibodies from Cell Signaling Technology (Danvers, MA, catalog number 5147). n = 6 mice per group were chosen randomly of the 8–14 total mice for this analysis.

Liver tissue was homogenized by sonication in 20 mM TRIS (pH 7.5), 2 mM EDTA, 10 mM EGTA, 1% Triton X-100, and 250 mM sucrose with HALT™ Protease and Phosphatase Inhibitor (Thermo Fisher Scientific), then centrifuged for 10 min at 10,000× g. Supernatants were collected, protein concentration was measured by BCA method (Thermo Fisher Scientific). CXCL2 was detected using the MesoScale Discovery Mouse MIP-2 V-Plex Kit as per manufacturer’s instructions (MesoScale Discovery, Rockville, MD), read on the MESO Sector S 600 Instrument, and analyzed with Discovery Workbench Software, v 4.0 (Meso Scale Discovery). PAI-1 was detected using the Mouse PAI-1 ELISA Kit according to the manufacturer’s instructions (Catalog No. EMSERPINE1, Thermo Fisher Scientific).

Bone marrow was flushed from the femurs and tibias of WT and fat-1 mice, dissociated with a 21-gauge needle, and then passed through a 70 μm filter. Single cell suspensions were then plated in RPMI 1640 containing 10% FBS, penicillin, and streptomycin (Thermo Fisher Scientific). Cells were differentiated for 7 days into macrophages in the presence of 20% conditioned medium collected from cultured L929 cells (ATCC, Manassas, VA). Macrophage identity was verified by flow cytometry using PE-labeled anti-F4/80 (BM8) and PerCP-Cy5-labelled anti-Cd11b (ICRF44) antibodies (Thermo Fisher Scientific). Macrophages were then trypsinized and re-plated at 3.5 × 10⁵ cells/well in a 24-well plate for treatment. Cells were incubated in the presence of 100 mM EtOH for 24 h or 100 ng/ml LPS for 4 h before harvesting for RNA isolation and cDNA synthesis. Treatments were performed in triplicate. Each condition was performed in two independent experiments with similar results.

Blood alcohol concentration were determined in plasma using the EnzyChrom™ ethanol assay kit (San Jose, CA) according to the manufacturer’s instructions.

Formalin-fixed, paraffin-embedded liver sections were deparaffinized and re-hydrated through graded EtOH solutions. Sections were then incubated in 20% goat serum and 0.2% Triton-X100 for 1 h at room temperature followed by an overnight incubation with a 1:100 dilution of anti-PAI-1 antibody (MA5-17171, Thermo Fisher Scientific). Sections were thoroughly washed with Tris-buffered saline + 0.1% Tween-20, incubated with HRP-conjugated anti-mouse antibodies and then developed with DAB (Dako/Agilent, Santa Clara, CA). Sections were de-hydrated and mounted under Cytoseal XYL mounting media (Thermo Fisher Scientific). Photomicrographs were taken with an Olympus BX43 microscope equipped with CellSens version 2.3 software (Olympus Life Science, Waltham, MA).

Differences among multiple groups were analyzed by unpaired one-way analysis of variance (ANOVA) followed by Sidak’s multiple comparison tests. Differences between two groups were analyzed by Student’s t test. All data are presented as the mean ± standard error of the mean. Analyses were performed using GraphPad Prism (GraphPad Software, version 9.0.1, San Diego, CA). p values < 0.05 were considered significant.

Male fat-1 and WT littermates were placed on a diet containing EtOH for 10 days, followed by a bolus of EtOH delivered by oral gavage on the 11^th day to recapitulate human AH, an advanced stage of ALD (Bertola et al., 2013) (see Figure 1A for experimental design). Average food consumption and body weights for both WT and fat-1 mice at the beginning and end of the experiment were similar, with both genotypes exhibiting a minor decrease in final body weights (Table 2). Both liver/body weight and epididymal fat/body weight ratios were also similar between WT and fat-1 mice at the beginning and end of the experiment. There were no differences in blood alcohol concentration between WT and fat-1 EtOH-treated mice (Table 2). Lastly, hepatic levels of n3-PUFAs were significantly higher in fat-1 mice compared to WT mice both in PF and EtOH-fed groups (Figure 1B), whereas there were no differences in n6-PUFAs between genotypes. Interestingly, EtOH feeding resulted in an increase in both n6-and n3-PUFAs in WT and fat-1 mice.

We observed a significant EtOH-induced increase in liver injury in WT mice, as demonstrated by elevation of plasma ALT levels, that was not evident in fat-1 mice (Figure 1C). Analysis of H&E-stained liver sections revealed a similar overall morphology between WT and fat-1 mice following EtOH treatment and demonstrated a similar level of microvesicular steatosis in both WT and fat-1 mice (staining in Figure 1D and quantitation in Figure 1E). To better characterize the extent of hepatic steatosis, we performed Oil Red O staining for neutral lipids, which also demonstrated a similar degree of EtOH-induced steatosis in WT and fat-1 mice (Figures 1F,G), further confirmed by a biochemical analysis of total liver TGs (Figure 1H). Interestingly, fat-1 PF mice had significantly less steatosis than WT mice as measure by both Oil Red O and total TGs.

We next determined if fat-1 mice were protected from EtOH-associated oxidative stress, a key contributor to ALD pathogenesis (Leung and Nieto, 2013). EtOH consumption leads to oxidative stress, in part, through its metabolism by CYP2E1, a cytochrome P450 that produces reactive oxygen species as a byproduct of detoxification. EtOH also induces CYP2E1expression, thus leading to further reactive oxygen species production and oxidative stress. Western blotting analysis demonstrated that EtOH increased the expression of CYP2E1 by 6-7-fold in both WT and fat-1 mice (Figures 2A,B). To determine resulting reactive oxygen species generation, we performed an assay for hepatic lipid peroxidation (TBARS) as an indirect a measure of oxidative stress and found a similar pattern as that for CYP2E1 expression. (Figure 2C). These data suggest that EtOH induction of oxidative stress was similar between WT and fat-1 mice.

Another pathological feature of ALD recapitulated by our acute-on-chronic EtOH treatment model is increased liver neutrophil infiltration (Bertola et al., 2013). To assay liver neutrophil accumulation, we measured liver myeloperoxidase (MPO) expression by both immunohistochemistry and ELISA (Figures 3A–C, respectively). While MPO immunohistochemistry showed no significant differences between groups, ELISA analysis of liver tissue lysates showed a significant increase in MPO levels in EtOH-fed vs PF WT mice which was not observed in fat-1 mice. Neutrophils are recruited to the liver following injury by several chemokines, including CXCL2. Although the expression of whole-liver Cxcl2 was increased (but not significantly) by EtOH in both WT and fat-1 mice, there were no differences between the two genotypes (Figure 4A). CXCL2 protein in the liver was also modestly induced by EtOH, although again we observed no significant differences between genotypes (Figure 4B). Another mediator that can contribute to neutrophil chemoattraction is PAI-1, which is commonly elevated in both human ALD patients and in mouse models of ALD (Mukamal et al., 2001;Bergheim et al., 2006). PAI-1 is a pleiotropic acute phase response protein that has been classically associated with fibrin deposition but has been more recently identified to have a role in inflammation and/or neutrophil recruitment in some disease models (De Taeye et al., 2006;Praetner et al., 2018). In our study, the hepatic expression of Pai-1 mRNA was significantly induced by EtOH in WT but not in fat-1 mice (Figure 4C). At the protein level, EtOH had a limited effect on PAI-1 expression (Figure 4D). Importantly, fat-1 mice expressed significantly less PAI-1 than WT mice (Figure 4D). Under certain conditions (e.g., liver regeneration), hepatocytes can be a large source of PAI-1 (Thornton et al., 1994). Immunohistochemistry of liver sections revealed increased PAI-1 staining in hepatocytes of the pericentral region in WT EtOH vs WT PF mice, an induction which was not observed in fat-1 EtOH-fed mice (Figure 4E). We also assayed the expression of additional markers of hepatic inflammation, namely IL-6 and TNF-α, but found no differences between PF or EtOH-treated fat-1 and WT mice (data not shown).

We next sought to determine other cell-specific contributions to the expression of neutrophil chemokines, Cxcl2 and Pai-1, which are expressed by multiple cell types including macrophages (De Filippo et al., 2013; Fang et al., 2018). To determine the expression of Cxcl2 and Pai-1 expression in macrophages, we isolated bone marrow cells from WT and fat-1 mice and differentiated them into macrophages (BMDMs). Notably, baseline expression of Pai-1 was lower, albeit not significantly, in fat-1-derived BMDMs than in WT-derived BMDMs (Figure 5B). Incubation with EtOH had no significant effect on either Cxcl2 (Figure 5A) or Pai-1 (Figure 5B) expression in both fat-1 and WT BMDMs. However, LPS stimulation led to a large increase in both Cxcl2 and Pai-1 expression in WT and fat-1-derived BMDMs, but with significantly less induction in fat-1 BMDMs.

To investigate the effects of n3-PUFA enrichment on liver immunity, we analyzed several liver immune cell populations by flow cytometry. First, we found that fat-1 EtOH-treated mice had decreased populations of M1 KCs relative to WT EtOH-treated mice with no change in M2 KCs, indicating a shift away from the pro-inflammatory M1 KC phenotype (Figure 6A). We also noted an increase in liver anti-inflammatory Tregs in fat-1 EtOH-treated mice vs WT EtOH-treated mice (Figure 6B), a cell type which is critical for maintaining immune homeostasis and self-tolerance, as well as blocking pro-inflammatory signals (Corthay, 2009). There were also trending increases in additional liver immune cells including natural killer cells as well as conventional cytotoxic T cells, indicating multiple effects of n3-PUFA enrichment on liver innate and adaptive immunity (Figures 6C,D).