Endotoxin-Stimulated Hepatic Stellate Cells Augment Acetaminophen-Induced Hepatocyte Injury

Abstract

Acetaminophen (APAP)–induced liver injury is influenced by inflammatory Gram-negative bacterial endotoxin [lipopolysaccharide (LPS)], mechanisms of which are not completely understood. Because LPS-stimulated perisinusoidal hepatic stellate cells (HSCs) produce cytokines that affect survival of hepatocytes, this study investigated their role in APAP-induced liver injury. Fed (nonstarved) rats were administered 5 mg/kg LPS or phosphate-buffered saline (PBS) vehicle, followed by 200 mg/kg APAP or PBS an hour later, and euthanized at 6 hours. Control rats received PBS at both time points. Both LPS and APAP caused mild hepatocyte injury (apoptosis), as assessed by histopathology, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining, and caspase-3 activation. The liver injury was augmented in rats administered LPS + APAP, in association with increased nuclear translocation of interferon-regulatory factor-1 (IRF1). In vitro, APAP augmented LPS/HSC-conditioned medium–induced inhibition of DNA and protein synthesis, apoptosis, and nuclear IRF1 in hepatocytes. LPS-stimulated HSCs produced interferon-β (IFN-β), and LPS/HSC + APAP-induced hepatocyte apoptosis was inhibited by anti–IFN-β antibody. Finally, HSC-depleted mice produced significantly lower IFN-β and tumor necrosis factor-α, exhibited less oxidative stress, and were protected from excessive injury due to high APAP dose (600 mg/kg), as well as LPS (5 mg/kg overnight) followed by APAP. In co-culture with or without LPS, HSCs increased expression of proinflammatory cytokines by Kupffer cells. These results suggest that HSCs play a critical role in APAP-induced liver injury without or with LPS preconditioning, and it involves INF-β–IRF1 signaling.

Acetaminophen (N-acetyl-para-aminophenol, paracetamol; APAP) is an over-the-counter medicine commonly used for aches and fever, and is generally not toxic.¹ However, about 40% to 50% of acute liver failure cases are estimated to be associated with APAP, making it a leading cause of drug-induced liver injury.2, 3, 4 The liver is the primary site of drug metabolism, including APAP. Cytochrome P450-2E1 (CYP2E1) converts APAP into N-acetyl-p-benzoquinone imine, which is detoxified by irreversible binding to the thiol of reduced glutathione (GSH). However, on exceeding threshold levels, N-acetyl-p-benzoquinone imine depletes cytosolic and mitochondrial GSH, inhibits mitochondrial respiration, and reduces cellular ATP, causing hepatocyte death.^3,4 Rats/mice, used extensively to understand the mechanisms of APAP toxicity, are much less susceptible to APAP-induced liver damage than humans. Therefore, they are rendered highly vulnerable to APAP-induced injury administered generally at high doses (300 to 800 mg/kg) following food deprivation that profoundly depletes glycogen, ATP, and GSH.3, 4, 5 Although hepatotoxicity occurs on APAP overdose in humans, there is also evidence of liver injury at clinically acceptable doses (https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-prescription-acetaminophen-products-be-limited-325-mg-dosage-unit, last accessed February 7, 2018).⁶ Such modest injury may remain undiagnosed, but can progress to overt liver damage in the presence of stresses such as exposure to other stimuli of liver injury. Alcoholics generally have increased portal exposure to Gram-negative bacterial endotoxin [lipopolysaccharide (LPS)]⁷ due to increased gut permeability.^8,9 Indeed, greater sensitivity of alcoholics towards developing acute liver failure due to an accidental dose of ≤4 g (approximately 6 mg/kg) taken over a 24-hour period was reported.¹⁰ Notably, about 37% of the subjects with acute liver failure due to accidental dose of 12 g (ie, approximately 171 mg/kg) were nonalcoholic.¹⁰

Hepatic inflammation plays a major role in perpetuating and accelerating APAP-induced liver damage.¹¹ The liver is the primary site of the uptake and metabolism of gut-derived highly inflammatory LPS delivered via the portal vein. The liver is exposed to higher concentrations of LPS under certain pathophysiological conditions, and its reduced hepatic metabolism increases systemic levels and causes endotoxemia.12, 13, 14 LPS may also enter circulation via the respiratory tract or systemic infection.¹⁵ Mice/rats are at least 1000 times less susceptible to LPS-induced liver injury.¹⁶ The perisinusoidal hepatic stellate cells (HSCs) respond to low (1 to 10 ng/mL) LPS concentration by producing several cytokines and chemokines. LPS-stimulated HSCs induce endoplasmic reticulum stress and autophagy, inhibit DNA and protein synthesis, and cause apoptosis of a subpopulation of hepatocytes.17, 18, 19, 20 Therefore, we hypothesized that the pre-existing stress due to LPS-stimulated HSCs may condition hepatocytes to APAP-induced damage at an otherwise acceptable dose. In the current study, in fed state, LPS-pretreated rats become more susceptible to APAP-induced liver injury, which is mediated by HSC-released interferon-β (IFN-β) and nuclear translocation of interferon-regulatory factor-1 (IRF1) in hepatocytes. HSC-depleted mice produced lower levels of IFN-β and tumor necrosis factor-α (TNF-α), demonstrate less oxidative stress, and are protected from LPS + APAP–induced liver damage, but interestingly also from APAP-induced injury.

Materials and Methods

Reagents

The following were purchased form indicated sources: LPS (Escherichia Coli lipopolysaccharide serotype 0111:B4) and acetaminophen (Sigma-Aldrich, St. Louis, MO); radioimmunoprecipitation assay buffer and anti–β-actin antibody (Ab; Abcam, Cambridge, MA); neutralizing IFN-β Ab (InterferonSource, Piscataway, NJ); anti–caspase-3, anti–c-Jun N-terminal kinase (JNK), anti–phosphorylated JNK, anti-IRF1, and anti–histone H3 Abs (Cell Signaling, Beverly, MA); and recombinant IFN-β (PBL Assay Sciences, Piscataway, NJ).

LPS and Acetaminophen Treatment of Rats

Animal protocols were approved by Institutional Animal Care and Use Committees of the University of Pittsburgh, University of Cincinnati, Cincinnati Children’s Hospital Medical Center, and Cincinnati Veterans Administration Medical Center, where this research was performed according to the NIH guidelines.²¹ Male Sprague-Dawley rats, weighing about 300 g (Charles River Laboratories, San Diego, CA), with access to standard chow diet and water ad libitum, were administered LPS (5 mg/kg) or phosphate-buffered saline (PBS) vehicle intraperitoneally, followed by 200 mg/kg APAP 1 hour later. Control rats received PBS at both time points. Thus, there were four groups of six to eight rats each: control: PBS/PBS; LPS: LPS/PBS; APAP: PBS/APAP; and LPS/APAP. Animals were euthanized 6 hours after APAP administration. Blood was drawn and centrifuged to obtain serum; the liver was excised and washed in cold PBS, and portions of the right hepatic lobe were fixed in buffered formalin and 2% paraformaldehyde or snap frozen in liquid nitrogen. Formalin-fixed liver tissue was embedded in paraffin, and sections were stained with hematoxylin/eosin.

Culture and Treatments of HSCs and Hepatocytes

HSCs were isolated from male Sprague-Dawley rats (450 to 500 g), purified using Nycodenz gradient and cultured, as described previously.^17,18 Medium was renewed after overnight culture, and the cells were used on day 3 of culture.

Hepatocytes were prepared by collagenase digestion of the rat liver (200 to 250 g Sprague-Dawley), purified on Percoll gradient, and cultured, as described previously.^17,18 Medium was renewed after a 3-hour attachment period, and the cells were used after overnight culture.

HSCs were incubated in Dulbecco’s modified Eagle’s medium (DMEM) containing 5% fetal bovine serum, without or with 100 ng/mL LPS for 6 hours. Sterile-filtered HSC-conditioned medium (without or with LPS) was transferred to hepatocytes. After 1 hour, 1 mmol/L APAP was added, and assays were performed at the times indicated in figure legends. For control, hepatocytes were placed in medium preincubated without or with LPS ± APAP.

HSC–Kupffer Cell Co-Culture

Kupffer cells were isolated, as described.²² Briefly, mouse liver was digested with 15 mL Hanks’ balanced salt solution containing 0.05% collagenase (Sigma, St. Louis, MO) and 0.5 mmol/L CaCl₂. Liver cells were released into DMEM using cell scraper. Following filtration through a 100-μm cell strainer, hepatocytes and cell debris were removed by centrifugation at 50 × g for 2 minutes (3×). The supernatant was centrifuged at 1350 × g for 15 minutes; the pellet was suspended in 10 mL DMEM, layered on the isotonic 25%/50% Percoll gradient, and centrifuged at 850 × g for 15 minutes. Enriched fraction of the Kupffer cells was collected from the 25%/50% Percoll interface, washed, and suspended in DMEM containing 20% fetal bovine serum. Mouse HSCs were prepared, as described previously.²⁰

HSCs (4 × 10⁶ cells in 1.5 mL per well) were placed in the Corning Costar Transwell culture inserts (pore size of 0.4 μm diameter; Corning, Inc., Corning, NY) in a 6-well culture plate, and Kupffer cells (4 × 10⁶ cells in 2.5 mL per well) in the culture well. For control, equal number of HSCs or Kupffer cells were seeded separately in 6-well plates. The media were renewed after a 3-hour attachment period. After overnight culture, medium in upper compartment was replaced with DMEM containing 5% fetal bovine serum, without (vehicle) or with LPS (100 ng/mL), whereas medium in lower compartment was replaced with DMEM/5% fetal bovine serum. Similarly, HSC or Kupffer cell monocultures were treated with vehicle or LPS. After 6 hours, cells were collected for mRNA analysis.

Determination of DNA or Protein Synthesis

DNA and protein synthesis was measured, as described previously.^17,18 Cells were incubated in DMEM/0.1% bovine serum albumin containing 1 μCi/mL [methyl-³H]thymidine (3.18 TBq/mmol) or L-[4,5-³H]leucine (5.6 TBq/mmol) (Amersham-Pharmacia, Piscataway, NJ) for 4 hours at 37°C, then washed with ice-cold Hanks’ balanced salt solution/bovine serum albumin, treated with ice-cold 10% trichloroacetic acid for 10 minutes, and washed once with trichloroacetic acid followed by 95% ethanol. Cells were digested with 5% SDS for determination of radioactivity. Leucine-free DMEM was used for [³H]leucine incorporation assay.

Determination of Oxidative Stress

Frozen liver sections were rehydrated in PBS. Slides were incubated for 30 minutes in dark in 10 μmol/L dihydroethidium (Sigma) in PBS prepared from the 20 mmol/L stock dihydroethidium solution in dimethyl sulfoxide, washed with PBS, and mounted with aqueous mounting medium.²³ Images were acquired using an Olympus BX51 fluorescence microscope (Texas red filter; Olympus, Tokyo, Japan) and CellSens Software version 1.18 (Olympus, Hamburg, Germany). Staining was quantified using ImageJ software version 1.50i (NIH, Bethesda, MD; https://imagej.nih.gov/ij, version- 1.53m, last accessed September 28, 2021).

IRF1 Staining

The liver sections were permeabilized with 0.25% Triton X-100 in PBS for 10 minutes, then blocked with PBS containing 1% bovine serum albumin, 10% goat serum, and 0.1% Triton X-100 for 1 hour. Following incubation with rabbit anti-IRF1 Ab overnight at 4°C, the sections were incubated with goat anti-rabbit Alexa Fluor 594 secondary Ab (Invitrogen Molecular Probes, Waltham, MA) for 1 hour in the dark, washed with PBS, mounted with DAPI, and imaged via fluorescence microscopy.

Cell Viability (MTT and Lactate Dehydrogenase Assays)

The culture medium was aspirated, and MTT (100 μL of 5 mg/mL solution) was added to the cells. Following incubation for 3 hours at 37°C, supernatant was aspirated and 200 μL of dimethyl sulfoxide was added to dissolve the formazan crystals formed in the viable cells. The OD was read at 570 and 630 nm. Lactate dehydrogenase was measured with a spectrophotometric assay kit (Stanbeo Laboratory, Boerne, TX).

LPS and Acetaminophen Treatment of HSC-Depleted Mice

Mice were allowed free access to standard chow diet and water ad libitum throughout the experimental period. The 6-week–old HSC-sufficient or HSC-depleted mice, generated from B6.Cg-Tg(Gfap-Tk)7.1Mvs/J (GFAP-TK-Tg) mice (Jackson Laboratory, Bar Harbor, ME), as described previously,23, 24, 25 were administered 5 mg/kg LPS or PBS (intraperitoneally) followed immediately by 1 mL saline (intravenously). After 15 hours (overnight), mice were administered intraperitoneally 600 mg/kg APAP (unless indicated otherwise) or PBS and euthanized 6 hours later. Control mice received PBS at both time points. The liver tissue and blood were acquired and processed for histopathology and biochemical determinations, as described above.

Biochemical Analysis

Serum alanine transaminase (ALT) activity was measured using VetSpec Kits (Catachem, Inc., Bridgeport, CT). Cell death was determined by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay kit (Roche Pharmaceuticals, Nutley, NJ). TUNEL-positive nuclei were detected by fluorescence microscopy. Hepatic glutathione (Cayman Chemical Company, Ann Arbor, MI) and caspase-3 activity (Sigma-Aldrich) were measured using assay kits. Enzyme-linked immunosorbent assay kits were used to measure IFN-β [Biomatic, Wilmington, DE (rat); PBL Assay Science, Piscataway, NJ (mouse)]; TNF-α [Thermo Fisher Scientific, Waltham, MA (rat and mouse)]; interferon-γ [IFN-γ; BioLegend, San Diego, CA (rat); Thermo Fisher Scientific (mouse)]; and IL-10 [Ray Biotech, Peachtree Corners, GA (rat); Thermo Fisher Scientific (mouse)].

CYP2E1 Activity

Liver tissue was homogenized in ice-cold Tris-acetate buffer (pH 7.4) containing 1.15% KCl (1:5 w/v). The homogenate was centrifuged at 10,000 × g for 30 minutes at 4°C. The supernatant was centrifuged at 100,000 × g for 60 minutes at 4°C. The microsomal pellet was resuspended, centrifuged at 100,000 × g, quick frozen, and stored at −80°C. CYP2E1 activity was measured by quantifying hydroxylation of p-nitrophenol to p-nitrocatechol.²⁶ The reaction mixture (0.1 mL) consisted of 250 μg microsomal protein [determined using BCA Protein Assay kit (Pierce, Rockford, IL)] in 50 mmol/L potassium-phosphate buffer, pH 7.4, containing 100 μmol/L p-nitrophenol and 25 μL of NADPH-generating system (26 mmol/L NADP+, 66 mmol/L D-glucose-6-phosphate, 66 mmol/L magnesium chloride, and 0.2 U of glucose-6-phosphate dehydrogenase). The reactions were performed at 37°C for 60 minutes and terminated by addition of trichloroacetic acid (1% final concentration). Precipitated proteins were removed by centrifugation at 10,000 × g for 5 minutes, and the clear supernatants were mixed with NaOH (1 mol/L final concentration). Absorbance of p-nitrocatechol was measured at 546 nm. CYP2E1 activity was calculated from extinction coefficient 9.53 mmol/L⁻¹cm⁻¹ and expressed as pmol/minute per mg of protein or percentage activity. Standard curve was generated using increasing concentrations of p-nitrocatechol.

Western Blot Analysis

Tissue or cell proteins were extracted in radioimmunoprecipitation assay lysis buffer (Santa Cruz Biotechnology, Dallas, TX) containing protease inhibitor cocktail (Sigma-Aldrich), phenylmethylsulfonyl fluoride, and sodium orthovanadate. Following separation via 10% SDS-PAGE, proteins were transferred onto polyvinylidene difluoride membranes (Sigma-Aldrich). The membranes were then blocked, incubated with primary Abs, washed, and treated with appropriate secondary Abs, and proteins were identified using Amersham ECL Select reagent (GE Healthcare UK Limited, Buckinghamshire, UK).

To obtain cytoplasmic and nuclear fractions, the liver was homogenized in 5× volume of ice-cold cytoplasmic extraction buffer [10 mmol/L HEPES, 60 mmol/L KCl, 1 mmol/L EDTA, 1 mmol/L dithiothreitol, 0.075% (v/v) Nonidet P-40, and 1 mmol/L phenylmethylsulfonyl fluoride, pH 7.6] for 20 seconds using Dounce homogenizer. After leaving on ice for 30 minutes, the homogenate was centrifuged at 252 × g in a Sorvall ST 16R centrifuge (Thermo Fisher Scientific) for 4 minutes at 4°C to obtain nuclei. The supernatant was centrifuged at 18,894 × g in a Sorvall ST 16R centrifuge for 10 minutes, and its supernatant (cytosolic fraction) was mixed with glycerol (5:1; v/v) and stored at −80°C. The nuclear pellet was suspended in cytoplasmic extraction buffer without Nonidet P-40 and centrifuged at 112 × g in a Sorvall ST 16R centrifuge for 4 minutes; the resultant pellet was suspended in nuclear extraction buffer [20 mmol/L Tris-HCl, 420 mmol/L NaCl, 25% (v/v) glycerol, 1.5 mmol/L MgCl₂, 1 mmol/L phenylmethylsulfonyl fluoride, and 0.2 mmol/L EDTA, pH 8.0], and the salt concentration was adjusted to 400 mmol/L with 5 mol/L NaCl; the suspension was left on ice for 30 minutes with occasional vortexing, then centrifuged at 18,894 × g in a Sorvall ST 16R centrifuge for 10 minutes at 4°C, and the supernatant (nuclear proteins) was stored frozen at −80°C until use. Nuclear proteins (15 μg) were separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and, after immunoblotting with primary monoclonal anti–IRF-1 or rabbit anti-histone polyclonal Ab (internal control), protein bands were detected using ECL reagent.

mRNA Analysis

RNA was extracted using Trizol Reagent (Life Technology, Carlsbad, CA), according to manufacturers' protocol. Following quantification, 1.5 μg RNA was converted into cDNA using Reverse Transcription Kit (Applied Biosystems, Foster City, CA). The expression of various genes was determined using primers (Integrated DNA Technology, Coralville, IA) (Table 1) and SYBER Green Master Mix in a 7300 Real Time PCR System (Applied Biosystems).

Table 1.

Quantitative PCR Primer Sequences

Mouse primer sequences
β-Actin	Forward	5′-AGAGGGAAATCGTGCGTGAC-3′
	Reverse	5′-CAATAGTGATGACCTGGCCGT-3′
IFN-β	Forward	5′-AGCTCCAAGAAAGGACGAACAT-3′
	Reverse	5′-GCCCTGTAGGTGAGGTTGATCT-3′
IFN-γ	Forward	5′-CTCTTCCTCATGGCTGTTTCT-3′
	Reverse	5′-GGACCCCAGACAATCGGTTG-3′
IL-10	Forward	5′-TTGTCGCGTTTGCTCCCATT-3′
	Reverse	5′-GAAGGGCTTGGCAGTTCTG-3′
IRF1	Forward	5′-CAGAGGAAAGAGAGAAAGTCC-3′
	Reverse	5′-CACACGGTGACAGTGCTGG-3′
TNF-α	Forward	5′-CCCAGGTATATGGGCTCATACC-3′
	Reverse	5′-GCCGATTTGCTATCTCATACCAGG-3′
CYGB	Forward	5′-CCAACTGCGAGGACGTGG-3′
	Reverse	5′-ACTGGCTGAAGTACTGCTTGGC-3′
Rat primer sequences
β-Actin	Forward	5′-GAGACCTTCAACACCCCAGCC-3′
	Reverse	5′-TCGGGGCATCGGAACCGCTCA-3′
IFN-β	Forward	5′-GGTGACATCCACGACTACTTTAG-3′
	Reverse	5′-CCAGGCATAGCTGTTGTACTT-3′
IFN-γ	Forward	5′-CGAATCGCACCTGATCACTAA-3′
	Reverse	5′-TGGATCTGTGGGTTGTTCAC-3′
IL-10	Forward	5′-GTTGCCAAGCCTTGTCAGAAA-3′
	Reverse	5′-TTTCTGGGCCATGGTTCTCT-3′
IRF1	Forward	5′-CTCACCAAGAACCAGAGGAAAG-3′
	Reverse	5′-AGATAAGGTGTCAGGGCTAGAA-3′
TNF-α	Forward	5′-TCCCAACAAGGAGGAGAAGT-3′
	Reverse	5′-TGGTATGAAGTGGCAAATCG-3′

CYGB, cytoglobin; IFN, interferon; IRF1, interferon-regulatory factor-1; TNF-α, tumor necrosis factor-α.

Statistical Analysis

In vivo experiments were performed using six to eight animals per group. In vitro experiments were performed at least three times, each in duplicate or triplicate using cells from different animals. Results are expressed as means ± SD. Statistical significance between groups was determined by t-test using Prism 5 software package (GraphPad, San Diego, CA) or by Kruskal-Wallis one-way analysis of variance on ranks, followed by the Dunn method post hoc test. P < 0.05 was considered statistically significant.

Results

In Vivo Hepatic Effects of LPS and APAP Treatment

Serum ALT levels were similar in control and LPS-treated rats (Figure 1A). Although cellular injury was not noticeable (Figure 1B), ALT increased modestly in APAP-treated rats, and further in LPS + APAP–treated rats. LPS caused only mild inflammatory infiltration that became more pronounced after LPS + APAP treatment (Figure 1B and Supplemental Figure S1A). LPS/APAP-treated rats also showed higher hepatic concentrations of TNF-α, IFN-β, and IFN-γ compared with LPS or APAP alone (Figure 1, C–E). Interestingly, hepatic concentration of a hepatoprotective anti-inflammatory cytokine, IL-10, also increased in LPS/APAP-treated rats compared with only LPS- or only APAP-treated rats (Figure 1F). Concentrations of all of these cytokines were notably greater in LPS- compared with APAP-treated rats. The mRNA expression of all cytokines, except IFN-β, which was near basal level in LPS, APAP, or LPS/APAP-treated rats, showed similar changes as their protein levels (Supplemental Figure S1, B–E). A previous study showed a rapid increase in hepatic IFN-β mRNA in LPS-administered rats that returned to the basal level by 6 hours.²⁷

Figure 1.

Lipopolysaccharide (LPS) ± acetaminophen (APAP)–induced liver injury and cytokine levels in rats. Rats were administered phosphate-buffered saline (PBS) vehicle or 5 mg/kg LPS, followed by PBS vehicle or 200 mg/kg APAP 1 hour later. Blood and liver were harvested at 6 hours and used for various assays. A: Serum alanine aminotransferase (ALT) levels. B: Representative images of hematoxylin/eosin-stained liver sections. C–F: Hepatic concentration of indicated cytokines, as determined by enzyme-linked immunosorbent assay. For all experiments, n = 6 to 8 per group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or between the indicated groups. Scale bars = 50 μm (B). IFN, interferon; TNF-α, tumor necrosis factor-α.

Despite no change in ALT in LPS-treated rats but increased levels in APAP-treated rats, significantly greater hepatocyte apoptosis was observed in LPS-treated rats compared with APAP-treated rats, as indicated by TUNEL staining, caspase-3 cleavage, and caspase-3 activity; all of which increased further in LPS + APAP–treated rats (Figure 2, A–C).

Figure 2.

Acetaminophen (APAP) augments lipopolysaccharide (LPS)–induced apoptosis in rat liver. Rats were administered phosphate-buffered saline (PBS) vehicle or 5 mg/kg LPS, followed by PBS vehicle or 200 mg/kg APAP 1 hour later. Livers were harvested at 6 hours for indicated determinations. A: Sections representative of each group stained with terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL; green) and counterstained with DAPI (blue) to mark nuclei were imaged by fluorescence microscopy. Red boxed area is magnified (inset) to show TUNEL-positive hepatocyte. Bar graph shows quantification of TUNEL-stained cells. B: Western blot analysis shows hepatic procaspase-3 and cleaved caspase 3 (Cas-3) expression. Bar graph shows densitometric quantification of cleaved caspase 3 relative to β-actin. C: Caspase 3 activity. D: Western blot analysis shows hepatic total c-Jun N-terminal kinase (T-JNK; isoforms 1 and 2) or phosphorylated JNK1 and JNK2 (P-JNK1 and P-JNK2, respectively). Bar graph shows relative expression of P-JNK1 versus T-JNK1 and P-JNK2 versus T-JNK2. E: Dihydroethidium (DHE)–treated liver sections demonstrate the magnitude of oxidative stress in the treatment groups relative to control. Bar graph shows quantification of the red fluorescent hydroxylated ethidium produced because of oxidation of DHE by superoxide. F: Hepatic glutathione (GSH) levels in the various groups. For all experiments, n = 6 to 8 per group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or as indicated. Scale bars = 100 μm (A and E). MFI, mean fluorescence intensity.

Because N-acetyl-p-benzoquinone imine, a metabolite produced by the action of CYP2E1 on APAP, is an inducer of oxidative stress and injury, CYP2E1 expression and activity was measured. Interestingly, CYP2E1 protein expression and activity were similar to the controls in all treatment groups (Supplemental Figure S2).

JNK and oxidative stress are implicated in APAP-induced liver injury. Silencing or inhibiting JNK, which is activated by oxidative stress, imparted protection against APAP-induced liver injury.^28,29 Western blot analysis showed JNK1 activation after LPS but not after APAP treatment. However, JNK1 and JNK2 activation increased by LPS + APAP treatment beyond that by LPS or APAP alone (Figure 2D). Hepatic oxidative stress was greater in APAP-treated rats compared with LPS-treated rats and did not change further in LPS + APAP–treated rats (Figure 2E). However, hepatic GSH levels were similar between control and all treatment groups (Figure 2F). It is likely that following the early modest decrease, either GSH levels recovered at 6 hours or oxidative stress was not high enough to cause GSH depletion, and factors other than oxidative stress (eg, mediators released by LPS-stimulated nonparenchymal cells) contributed to hepatic injury.

An important mechanism of hepatocyte injury involves nuclear translocation of a transcription factor, IRF1.^23,25,30, 31, 32 Hepatic IRF1 mRNA expression increased in LPS-treated and to a greater magnitude in LPS + APAP–treated rats, but not in APAP-treated rats (Figure 3A). Similarly, hepatic IRF1 protein expression was unaltered in APAP-treated rats, but increased in LPS-treated and to a greater extent in LPS + APAP–treated rats (Figure 3B). Immunohistochemical analysis demonstrated increased nuclear IRF1 translocation in hepatocytes of LPS-treated rats, which was augmented in LPS + APAP–treated rats (Figure 3, C and D). APAP itself elicited minor nuclear IRF1 translocation.

Figure 3.

Increased hepatic expression and nuclear translocation of interferon-regulatory factor-1 (IRF1) in lipopolysaccharide (LPS)/acetaminophen (APAP)–treated rats. Rats were administered phosphate-buffered saline (PBS) vehicle or 5 mg/kg LPS, and 1 hour later were administered PBS vehicle or 200 mg/kg APAP for 6 hours. A: Hepatic mRNA expression of IRF1 in LPS-, APAP-, or LPS + APAP–treated rats relative to glyceraldehyde-3-phosphate dehydrogenase, as determined via RT-qPCR. B: Immunoblot analysis for nuclear IRF1 in the livers. Bar graph shows densitometric quantification of IRF1 relative to histone H3 (internal control). C: Immunofluorescence detection of IRF1 in livers of indicated groups. Arrowheads indicate IRF1⁺ nuclei. Inset shows IRF1-positive nuclei from the boxed area at greater magnification. D: Quantification of cells expressing the IRF1⁺ nuclei. ∗P < 0.05, ∗∗P < 0.005 versus control or as indicated. Scale bars = 20 μm (C).

Effect of LPS-Stimulated HSCs on Hepatocyte DNA and Protein Synthesis and Injury

LPS-stimulated HSCs release soluble mediators, including INF-β and TNF-α, and strongly inhibit DNA synthesis, but induce apoptosis of a small percentage of hepatocytes.^17,18,27 To examine predisposition of hepatocytes incubated in LPS-conditioned HSC medium (LPS/HSC) to APAP-induced injury, APAP concentration dependence of hepatocyte response was determined first. APAP elicited biphasic response in hepatocytes with stimulation of [³H]thymidine incorporation (DNA synthesis) at 500 and 1000 μmol/L and inhibition at 2000 μmol/L concentration (data not shown). Therefore, 1000 μmol/L APAP was used in these experiments. LPS- as well as HSC-conditioned medium prevented APAP-induced DNA synthesis by hepatocytes (Figure 4A). Consistent with a previous report,¹⁸ LPS/HSC inhibited DNA synthesis by hepatocytes, which was further inhibited in the presence of APAP (Figure 4A).

Figure 4.

Lipopolysaccharide (LPS)–conditioned hepatic stellate cell (HSC) medium inhibits acetaminophen (APAP)–induced DNA and protein synthesis, and induces apoptosis in primary hepatocytes. A and B: HSCs were incubated in medium containing phosphate-buffered saline vehicle (control) or LPS (100 ng/mL) for 6 hours. HSC-conditioned medium was then transferred to hepatocytes, and APAP (1000 μmol/L) was added 1 hour later. After 20 hours of incubation, [³H]thymidine or [³H]leucine was added. Four hours later, specific incorporation was determined by scintillation counting. C–E: Hepatocytes treated as above with control or HSC-conditioned media were incubated in medium containing LPS ± APAP or medium conditioned by HSCs with or without LPS, as described above. After 12 hours, cells were washed and fixed for terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining or used for determination of caspase 3 activity. C: Images show TUNEL⁺ cells. D: Bar graph shows quantification of percentage of apoptotic hepatocytes. E: Caspase 3 activity. ∗P < 0.05. Scale bars = 100 μm (C). CPM, counts per minute.

As measured by [³H]leucine incorporation, APAP or LPS ± APAP modestly increased protein synthesis in hepatocytes (Figure 4B). Interestingly, HSC-conditioned medium also increased protein synthesis, which was not altered by APAP. However, LPS-conditioned HSC medium significantly reduced protein synthesis in the presence of APAP compared with all other conditions (Figure 4B).

Whether hepatocyte viability is affected by the presence of APAP in HSC/LPS medium was examined next. LPS or APAP alone did not affect hepatocyte viability, but a few TUNEL-positive hepatocytes were observed in the presence of LPS + APAP. A significant number of hepatocytes incubated in LPS-conditioned HSC medium were TUNEL-positive, which increased in the presence of APAP (Figure 4, C and D). The loss of hepatocyte viability was predominantly due to apoptosis, as indicated by caspase 3 activation (Figure 4E) and unaltered lactate dehydrogenase release and MTT readings (Supplemental Figure S3).

IFN-β, released by LPS- or concanavalin A–stimulated HSCs, causes nuclear translocation of IRF1 in hepatocytes and consequently their apoptosis.^25,27 Because LPS + APAP treatment increased nuclear IRF1 in hepatocytes in vivo, nuclear translocation of IRF1 was examined in hepatocytes incubated in LPS-conditioned HSC medium containing APAP. Time-course experiments showed strong LPS-induced increase in IFN-β expression in HSCs at 1 to 3 hours that declined subsequently at 6 hours (Figure 5A). Determination of nuclear translocation of IRF1 in hepatocytes showed that control HSC-conditioned medium did not cause significant translocation of IRF1 in hepatocyte nuclei in the absence or presence of APAP, but LPS-conditioned HSC medium significantly increased IRF1 nuclear translocation which was augmented in presence of APAP (Figure 5, B and C). Determination of whether LPS/HSC-released IFN-β sensitizes hepatocytes to APAP-induced apoptosis showed a few TUNEL-positive cells on incubation in the presence of IFN-β or APAP, which increased substantially in the presence of both substances (Supplemental Figure S4). Thus, abrogation of LPS/HSC + APAP–induced apoptosis of hepatocytes (Figure 5D) in association with significantly reduced caspase 3 activity (Figure 5E) by anti–IFN-β Ab is consistent with the role of HSC-derived IFN-β in augmenting APAP-induced hepatocyte injury.

Figure 5.

Interferon (IFN)-β, derived from lipopolysaccharide (LPS)–stimulated hepatic stellate cells (HSCs), induces apoptosis of hepatocytes in the presence of acetaminophen (APAP). A: LPS-induced time-dependent changes in IFN-β mRNA expression in HSCs, as determined by RT-qPCR. B: HSCs were incubated in medium with or without 100 ng/mL LPS for 6 hours. Control [HSC to hepatocyte (HC)] or LPS-conditioned (HSC/LPS to HC) HSC medium was then transferred to HCs, and APAP was added to the indicated wells. At 6 hours, the cells were fixed and nuclear interferon-regulatory factor-1 (IRF1) was determined via immunohistochemistry. Nuclei are shown in pseudogreen color and IRF1 in red. Nuclear localization of IRF1 is shown in yellow. Red boxed areas showing IRF1 positive nuclei are shown at a higher magnification to the right. C: Bar graph shows quantification of percentage of IRF1-positive nuclei in hepatocytes. D and E: Hepatocytes were incubated in LPS-conditioned HSC medium + APAP, as described above, but with IgG or anti–IFN-β. After 6 hours, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining was performed (D) or caspase 3 activity was measured (E) to assess apoptosis. D: Bar graph shows percentage of apoptotic hepatocytes. ∗P < 0.05, ∗∗P < 0.005. Scale bars = 50 μm (B and D). Ab, antibody; CT, control; PBS, phosphate-buffered saline.

HSC Depletion Protects Mice from LPS/APAP-Induced Liver Injury

Although the above data indicated that LPS-induced signaling in HSCs and subsequent release of IFN-β and/or TNF-α condition hepatocytes to APAP-induced liver damage, the results did not prove unequivocally the role of HSCs. Therefore, the mouse model of HSC depletion established in our laboratory was used next.23, 24, 25 Time course of LPS treatment of mice showed increased inflammation and expression of inflammatory cytokines at 1 to 3 hours that decline thereafter until 24 hours (Supplemental Figure S5, A and C). APAP dose-response showed minor injury at 200 to 300 mg/kg that increased at 400 and 600 mg/kg treatment (Supplemental Figure S6A). Time course of APAP-induced liver damage demonstrated a peak at 6 hours, and subsequent repair to normal at 96 hours (Supplemental Figure S6B). Thus, mice were subjected to a much stronger injury stimulus (overnight treatment with 5 mg/kg LPS, followed by 6 hours of treatment with 600 mg/kg APAP). Histopathology and ALT measurement showed only modest liver injury in LPS-treated HSC-sufficient mice (Figures 6, A and B). APAP, as well as LPS/APAP treatments caused strong centrilobular liver damage in HSC-sufficient mice, but the injury was attenuated in HSC-depleted mice (Figure 6, A and B). TNF-α, IFN-β, IFN-γ, and IL-10 protein levels increased in LPS or LPS ± APAP–treated HSC-sufficient mice, and the magnitude of this increase was lower in HSC-depleted mice (Figure 6, C–F). The mRNA expression pattern of these cytokines was consistent with changes in protein concentration. IRF1 mRNA expression also increased in LPS and LPS + APAP–treated HSC-sufficient mice and was significantly reduced in HSC-depleted mice (Supplemental Figure S7). Interestingly, IFN-β mRNA expression also increased in contrast to rat liver in which it was at the basal level in all treatment groups. The reason for this discrepancy may be the species-specific mechanisms, which are not determined yet.

Figure 6.

Lipopolysaccharide (LPS)–, acetaminophen (APAP)–, or LPS/APAP-induced hepatic injury in hepatic stellate cell (HSC)–depleted mice. HSC-sufficient and HSC-depleted mice were administered phosphate-buffered saline (PBS; control) or 5 mg/kg LPS. After 15 hours, PBS (control and LPS) or 600 mg/kg APAP (APAP and LPS/APAP) was administered, and mice were euthanized 6 hours later. A: Representative hematoxylin/eosin-stained sections. B: Serum alanine aminotransferase (ALT) levels. C–F: Hepatic concentrations of tumor necrosis factor (TNF)-α, interferon (IFN)-β, IFN-γ, and IL-10. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or as indicated. Scale bars = 100 μm (A).

Nuclear IRF1 was assessed in LPS ± APAP–challenged HSC-sufficient and HSC-depleted mice. Nuclear IRF1 increased in LPS as well as APAP-treated HSC-sufficient mice, and increased further after LPS + APAP treatment; nuclear IRF1 was much lower in HSC-depleted mice under all conditions (Figure 7, A and C). Interestingly, even though APAP (but not LPS) alone caused strong injury, nuclear IRF1 translocation by APAP was lower in HSC-sufficient and HSC-depleted mice than that by LPS (Figure 7, A and C). These data suggest involvement of (additional) IRF1-independent mechanisms of excessive liver injury due to APAP. In this regard, oxidative stress was much greater in APAP-treated or LPS + APAP–treated mice compared to LPS-treated HSC-sufficient mice (Figure 7, B and D). Oxidative stress was significantly lower in all treatment groups of HSC-depleted mice, and was restricted to centrilobular area as opposed to well spread out in the liver lobule in HSC-sufficient mice (Figure 7, B and D). GSH showed no significant change in LPS-treated, and similar but robust decrease in APAP-treated HSC-sufficient and HSC-depleted mice (Figure 7E). The levels decreased further in LPS + APAP–treated HSC-sufficient mice and, interestingly, improved in HSC-depleted mice (Figure 7E).

Figure 7.

Nuclear translocation of interferon-regulatory factor-1 (IRF1) and oxidative stress in lipopolysaccharide (LPS)/acetaminophen (APAP)–treated hepatic stellate cell (HSC)–sufficient and HSC-depleted mice. HSC-sufficient and HSC-depleted mice were administered phosphate-buffered saline (PBS; control) or 5 mg/kg LPS. After 15 hours, PBS (control and LPS) or 600 mg/kg APAP (APAP and LPS/APAP) was administered, and mice were euthanized 6 hours later. A: Liver sections were immunostained with anti-IRF1 antibody to determine nuclear translocation of IRF1. Arrowheads indicate IRF1-positive nuclei; yellow boxed area showing IRF1 positive nuclei is enlarged to the right. B: Liver sections were treated with dihydroethidium (DHE) to examine oxidative stress. A: Nuclei are shown in pseudogreen color and IRF1 in red. Nuclear localization of IRF1 is shown in yellow. C and D: Quantification of IRF1⁺ nuclei and red fluorescent hydroxylated ethidium produced because of oxidation of DHE by superoxide, respectively. E: Tissue glutathione (GSH) levels. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or as indicated. Scale bars: 20 μm (A); 50 μm (B). MFI, mean fluorescence intensity.

JNK activation in mice showed no effect of LPS (Figure 8), unlike an increase in rats (Figure 2), which may be due to longer time of exposure (approximately 21 hours in mice versus 7 hours in rats). APAP increased the activation of JNK1 and JNK2 in HSC-sufficient as well as HSC-depleted mice, and although further increase in JNK activation occurred in LPS/APAP-treated HSC-sufficient mice, it was reduced in HSC-depleted mice (Figure 8).

Figure 8.

c-Jun N-terminal kinase (JNK) activation in lipopolysaccharide (LPS) ± acetaminophen (APAP)–treated hepatic stellate cell (HSC)–-depleted (Depl) mice. HSC-sufficient (Suff) and HSC-depleted mice were administered phosphate-buffered saline (PBS; control) or 5 mg/kg LPS. After 15 hours, PBS (control and LPS) or 600 mg/kg APAP (APAP and LPS/APAP) was administered, and mice were euthanized 6 hours later. Western blot analysis shows hepatic total JNK (T-JNK; isoforms 1 and 2) or phosphorylated JNK1 and JNK2 (P-JNK1 and P-JNK2, respectively). Bar graphs show relative expression of P-JNK1 versus T-JNK1 and P-JNK2 versus T-JNK2. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or as indicated.

APAP Does Not Stimulate Cytokine Expression by HSCs

Because HSC-depleted mice were protected from APAP-induced liver damage, whether APAP elicits cytokine response in these mice was examined next. Both rat and mouse HSCs responded robustly to LPS challenge by increasing expression of TNF-α, IFN-β, and IL-10, but APAP failed to exert any appreciable effect (Supplemental Figure S8).

HSC Depletion and CYP2E1 Activity

HSC-depleted mice were protected from APAP-induced liver damage, and HSCs were reported to influence cytochrome CYP2E1 activity via O₂ carrier cytoglobin, expressed primarily by HSCs.³³ To determine whether APAP metabolism is altered by HSC depletion, hepatic cytoglobin expression, CYP2E1 protein expression, and its microsomal enzymatic activity was measured next. In HSC-sufficient mice, hepatic cytoglobin expression was similar in all treatment groups; cytoglobin expression was significantly lower in HSC-deleted than in HSC-sufficient mice in all treatment groups (Supplemental Figure S9A). Interestingly, neither CYP2E1 protein nor enzymatic activity was altered in HSC-depleted mice compared with that in HSC-sufficient mice under all treatment conditions (Supplemental Figure S9, B and C). The mechanism of protection from APAP in HSC-depleted mice is complex, and determination of APAP-metabolic products will be required for clarification of these findings. However, serum levels of APAP metabolites are lower in cytoglobin-knockout mice, which show significantly less liver damage.³³

Effect of IFN-β

To examine the direct role of IFN-β, mice were treated with 10³ units of recombinant IFN-β ± 400 mg/kg APAP. Histopathology, as well as serum ALT, showed modest injury by IFN-β in HSC-sufficient and HSC-depleted mice (Supplemental Figure S10). However, although significant damage was observed after APAP administration, and its augmentation in IFN-β + APAP–treated HSC-sufficient mice, the injury was still mitigated in HSC-depleted mice (Supplemental Figure S10). These data suggest that HSC-released IFN-β is only a part of liver damage in LPS + APAP–induced injury.

In Co-Culture, HSCs Induce Proinflammatory Response in Kupffer Cells

Because of the reduced expression of proinflammatory cytokines in LPS- and LPS/APAP-treated mice, we wondered whether a bidirectional interaction between HSCs and Kupffer cells might be responsible for this effect. In noncontact culture, HSCs augmented the LPS-induced expression of TNF-α and IFN-γ, whereas the increase in IFN-β expression was statistically nonsignificant (Figure 9). In contrast, Kupffer cells did not affect LPS-mediated expression of TNF-α or IFN-β in HSCs; however, interestingly in the absence of LPS, Kupffer cells increased IFN-β expression in HSCs (Figure 9).

Figure 9.

Inflammatory cytokine expression by Kupffer cells (KCs) and hepatic stellate cells (HSCs) in co-culture. HSCs were placed in culture in the Transwell insert and Kupffer cells in the culture wells of 6-well plates. After overnight culture, media were changed and cells were stimulated with lipopolysaccharide (LPS; 100 ng/mL). After 6 hours, cellular RNA was extracted, and RT-qPCR was performed to quantify mRNA expression of the indicated cytokines. Data are means from three independent determinations ± SD. ∗P < 0.05, ∗∗P < 0.01 versus –LPS or as indicated.

Discussion

Most investigations to delineate mechanisms of APAP-induced liver damage in mice or rats use high APAP doses (300 to 800 mg/kg or even higher) following overnight or longer period of starvation. The biochemical consequence of such fasting is depletion of hepatic glycogen, ATP, and GSH,^4,5 which renders hepatocytes highly vulnerable to APAP-induced damage. However, liver injury has been documented in humans at clinically acceptable APAP doses (https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-prescription-acetaminophen-products-be-limited-325-mg-dosage-unit, last accessed February 7, 2018).^6,10 The liver is subjected to low levels of LPS physiologically, and its portal concentration increases in several conditions, including alcohol consumption, compromised nutritional state, and gastroenteritis. A previous study observed that LPS increases APAP-induced liver toxicity in starved mice.³⁴ In contrast, others reported that LPS pretreatment protected mice/rats from APAP-induced liver damage.35, 36, 37 This discrepancy may be due to variability in the duration and doses of LPS pretreatment, different doses of APAP, and the duration of starvation before LPS and/or APAP challenges. The first part of the present study used short (1 hour) duration of 5 mg/kg LPS treatment, followed by 6 hours challenge with a modest (200 mg/kg) APAP dose. In the second part, HSC-sufficient and HSC-depleted mice were challenged with much higher dose of APAP (600 mg/kg) for 6 hours, following pretreatment with LPS (5 mg/kg) for 15 hours. All animals had free access to food and water. Results of both treatment conditions demonstrated an important role of HSCs in LPS ± APAP–induced liver injury. These doses of LPS and APAP were selected because rodents are far less susceptible to LPS-induced injury.¹⁶ Whereas 6 to 170 mg/kg APAP can cause liver injury (or failure) in humans (https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-prescription-acetaminophen-products-be-limited-325-mg-dosage-unit, last accessed February 7, 2018),^6,10 strong inflammatory response, but minimal liver injury is observed in rodents with 5 mg/kg LPS (Figures 1 and 6 and Supplemental Figure S5) and minimal injury at up to 300 mg/kg APAP (Figure 1 and Supplemental Figure S6).

In contrast to a previous report in starved rats,³⁴ the current study did not show necrosis in LPS-, APAP-, or LPS/APAP-treated fed (non-starved) rats. The mild liver injury due to short duration of LPS preconditioning before modest APAP dose was primarily apoptotic (TUNEL staining and caspase 3 activity), and was consistent with increased JNK phosphorylation in LPS + APAP–treated compared with LPS- or APAP-treated rats. JNK activation has been previously implicated in APAP-induced²⁸ and other forms of hepatocyte injury, including that by TNF-α and FasL.38, 39, 40, 41 However, increased apoptosis of hepatocytes in LPS + APAP–treated rats in the absence of a significant change in oxidative stress compared with that by APAP suggested an alternate mechanism of injury. Soluble mediators released by LPS-stimulated HSCs promote hepatocyte apoptosis by increasing nuclear translocation of IRF1.²⁷ Among the inflammatory mediators produced by LPS-stimulated HSCs, TNF-α can increase nuclear IRF1 in hepatocytes.30, 31, 32, 31 LPS, but not APAP, treatment increased hepatic TNF-α, but it did not increase further on LPS + APAP treatment. Although nuclear IRF1 in hepatocytes increased significantly in LPS + APAP–treated rats compared with that in LPS-treated rats, APAP alone did not cause increase in IRF1. The role of TNF-α was proposed as serum TNF-α increases rapidly at 2 hours (and declines quickly at 3 hours) in LPS + APAP–treated prestarved mice that showed augmentation of liver damage,³⁴ and mice treated with TNF-α Ab or transgenic mice deficient in the p55-TNF receptor 1 are partially protected from APAP-induced liver injury.⁴² In contrast, more severe liver damage is reported in mice deficient in TNF receptor 1 compared with the wild-type mice,⁴³ and another study found no effect of TNF inhibition on APAP-induced damage.⁴⁴ Thus, whether TNF-α is a major mediator of APAP-induced liver injury is under debate.^34,42, 43, 44

Type I and type II interferons can also damage hepatocytes by inducing nuclear translocation of IRF1.30, 31, 32 HSC-derived IFN-β was recently identified as an important mediator of LPS- or concanavalin A–induced liver injury via nuclear translocation of IRF1.^23,25,27 Thus, greater increase in hepatic IFN-β in LPS + APAP–treated rats compared with that by LPS or APAP provides evidence for IFN-β/IRF1 axis in hepatic injury in this model. In support of this suggestion, LPS stimulated the synthesis of IFN-β by HSCs (Figure 5),²⁷ and LPS-conditioned HSC medium inhibited DNA and protein synthesis in hepatocytes and increased caspase 3 activation and apoptosis. Furthermore, IFN-β + APAP induced apoptosis of a significant population of hepatocytes but individually exerted only minor effect, and LPS/HSC + APAP–induced nuclear IRF1 translocation and IFN-β Ab inhibited hepatocyte apoptosis.

The in vitro evidence for the role of HSCs in APAP-induced hepatocyte injury was confirmed in HSC-depleted mice treated in vivo with much robust LPS + APAP challenge. Depletion of HSCs imparted protection with significantly reduced synthesis of IFN-β and TNF-α. Because LPS-stimulated HSCs also produce these mediators,^24,27 the data indicate that HSCs may be the primary cells regulating hepatocyte injury. LPS also stimulates Kupffer cells and monocytes to produce inflammatory mediators, and APAP-induced liver damage has been attributed to the activation of Kupffer cells and recruitment of circulating monocytes, which then augment liver damage by releasing proinflammatory cytokines.45, 46, 47, 48 Blockade of macrophage function in rats with gadolinium chloride or dextran sulfate inhibited liver injury because of subsequent high-dose APAP challenge.^36,49 However, others reported increased liver injury on elimination of macrophages with clodronate-encapsulated liposomes or even their blockade with gadolinium chloride.⁵⁰

From the above findings, it is apparent that cells other than macrophages can be critical to APAP-induced liver injury. Indeed, significantly reduced levels of TNF-α, IFN-β, and IFN-γ in LPS-, APAP-, and LPS + APAP–treated HSC-depleted mice compared with HSC-sufficient mice suggest that HSCs are able to regulate the expression/release of injury-causing cytokines by immune cells, including Kupffer cells. In fact, HSCs were found to strongly influence the synthesis of cytokines by regulatory T cells²⁰ and dendritic cells.⁵¹ Furthermore, concanavalin A–induced expression of inflammatory mediators, including IFN-β by nonparenchymal cells, is reduced in co-culture with HSCs.²³ The current experiments show that HSCs augmented LPS-induced expression of TNF-α, IFN-γ, and IFN-β in Kupffer cells, whereas Kupffer cells did not influence LPS-induced expression of TNF-α and IFN-β in HSCs. Thus, protection of HSC-depleted mice by APAP-induced injury may involve lack or inadequate generation of such mechanism. Interestingly, however, IFN-β administration did not restore APAP-induced injury in HSC-depleted mice to levels similar to LPS/APAP-treated HSC-sufficient mice, suggesting that the process involves additional mechanisms. However, from the previous work showing protection from LPS- or ischemia-reperfusion–induced injury in HSC-depleted mice²⁴ and the results of the present study, it is apparent that in the absence of HSCs, Kupffer cells are unable to produce significant levels of injurious mediators.

Data from the current study indicate that nuclear IRF1 translocation can be an important mechanism of LPS/APAP-induced liver damage. TNF-α, as well as IFN-β, released by HSCs (and/or Kupffer cells, which additionally produce IFN-γ) can cause IRF1 nuclear translocation. However, there was robust liver damage but relatively less IRF1 nuclear translocation in APAP-treated HSC-sufficient mice, in contrast to minor liver injury in LPS-treated mice despite greater nuclear IRF1. To confirm the role of IRF1, IRF1-knockout mice were challenged to LPS ± APAP. No reliable inference could be drawn because of significant morbidity and mortality. However, oxidative stress was much stronger in APAP- or LPS + APAP–treated HSC-sufficient mice compared with HSC-depleted mice, which suggests that HSCs also play a major role in causing oxidative stress in hepatocytes. The oxidative stress in mice treated with much robust injury stimulus was far greater than in rats subjected to mild LPS/APAP challenge. Because APAP did not stimulate cytokine synthesis by HSCs, it is likely that APAP-induced damaged hepatocytes activate HSCs, which might be critical to the contemporaneous or consequent activation of macrophages.

Protection of HSC-depleted mice from APAP-induced liver injury despite similar activation of JNK and similarly reduced GSH, both of which have been shown to be major factors for the organ damage,²⁹ are two caveats of this study. However, oxidative stress as well as nuclear IRF1 expression were mitigated in APAP-treated HSC-depleted mice. It is likely that the balance between the proinjury and anti-injury mechanisms is skewed toward the latter in HSC-depleted mice. Lower levels of proinflammatory mediators in HSC-depleted compared with HSC-sufficient mice in combination with similarly increased IL-10 levels might be an important reason of such protection. Thus, strongly reduced oxidative stress, proinflammatory mediators, nuclear IRF1 and JNK activation, and relatively less GSH depletion in LPS + APAP–challenged HSC-depleted mice compared with HSC-sufficient mice indicate that HSCs are critical to APAP-induced liver injury. Surprisingly, CYP2E1 activity was similar in microsomes isolated from HSC-sufficient and HSC-depleted mice. Because expression of cytoglobin, which regulates CYP2E1-induced APAP metabolism in hepatocytes and is expressed primarily by HSCs,³³ is reduced in HSC-depleted mice, it is plausible that metabolism of APAP to its toxic metabolites is inhibited in vivo in HSC-depleted mice.

In summary, these data demonstrate that HSCs play an important role in APAP-induced liver damage with or without prior LPS exposure. These data further indicate that the inflammatory mediators produced by HSCs and through their interactions with other cells, such as Kupffer cells, can be a major component of liver damage. Therefore, HSCs are potential targets in treatment of acute liver failure of various etiologies, including APAP.

Supplemental Data

Supplemental Figure S1.

Lipopolysaccharide (LPS) ± acetaminophen (APAP)–induced liver injury and cytokine mRNA expression in rats. Rats were administered phosphate-buffered saline (PBS) vehicle or 5 mg/kg LPS, followed by PBS vehicle or 200 mg/kg APAP 1 hour later. Livers were harvested for indicated determinations. A: Hematoxylin/eosin-stained liver sections. Short black arrow: apoptotic cells; long black arrows: central vein dilation; blue arrows: sinusoidal dilation. B–E: Relative mRNA expression of indicated cytokines, as determined by quantitative RT-PCR. For all experiments, n = 6 to 8 per group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or between the indicated groups. Scale bars = 20 μm (A).

Supplemental Figure S2.

Cytochrome P450-2E1 CYP2E1) protein expression and enzyme activity in lipopolysaccharide (LPS) ± acetaminophen (APAP)–treated rats. Rats were administered phosphate-buffered saline (PBS) vehicle or 5 mg/kg LPS, followed by PBS vehicle or 200 mg/kg APAP 1 hour later. Liver protein lysates (A) or microsomes (B) were prepared for Western blot analysis and CYP2E1 enzyme activity determinations, respectively. The data shown are means of three separate determinations ± SD. n = 6 to 8 per group (A and B).

Supplemental Figure S3.

Viability of hepatocytes incubated in lipopolysaccharide (LPS)–conditioned hepatic stellate cell (HSC) medium with or without acetaminophen (APAP). HSCs were incubated in medium containing phosphate-buffered saline vehicle (control) or LPS (100 ng/mL) for 6 hours. HSC-conditioned medium was then transferred to hepatocytes, and APAP (1000 μmol/L) was added 1 hour later. After 20 hours of incubation, medium was aspirated for determination of lactate dehydrogenase (LDH) activity (A) and the cells were analyzed for viability via MTT assay (B). Ab, antibody.

Supplemental Figure S4.

Effect of interferon (IFN)-β ± acetaminophen (APAP) on apoptosis of hepatocytes in primary culture. Hepatocytes were incubated with 200 U/mL IFN-β for 1 hour, then 1 mmol/L APAP was added to the culture medium. After 24 hours, cells were washed with phosphate-buffered saline and fixed in 2% paraformaldehyde. Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling staining was performed to determine cell death via fluorescence microscopy using an Olympus BX51 fluorescence microscope (fluorescein isothiocyanate filter). Scale bars = 20 μm.

Supplemental Figure S5.

Time course of lipopolysaccharide (LPS)–induced liver injury and inflammation in mice. The 6-week–old C57BL/6 male mice were administered 5 mg/kg LPS (intraperitoneally); control mice received phosphate-buffered saline. Mice were euthanized at indicated time points. A: Hematoxylin/eosin-stained liver sections. Arrows indicate inflammation. B: Serum alanine aminotransferase (ALT) levels. C: Hepatic mRNA expressions of proinflammatory cytokines, as determined by RT-qPCR. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control. Scale bars = 100 μm (A).

Supplemental Figure S6.

Time course and dose dependence of acetaminophen (APAP)–induced liver injury in mice. The 6-week–old C57BL/6 male mice were administered 200, 300, 400, or 600 mg/kg acetaminophen for 24 hours (A) or 400 mg/kg acetaminophen for 6, 24, 48, 72, and 96 hours (B); control mice received phosphate-buffered saline. Hematoxylin/eosin-stained liver sections are shown. Serum alanine aminotransferase (ALT) levels for corresponding conditions are shown in bar graphs. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control. Scale bars = 100 μm (A and B).

Supplemental Figure S7.

Hepatic mRNA expression of cytokines in lipopolysaccharide (LPS) ± acetaminophen (APAP)–treated hepatic stellate cell (HSC)–sufficient and HSC-depleted mice. HSC-sufficient and HSC-depleted mice were administered phosphate-buffered saline (PBS; control) or 5 mg/kg LPS. After 15 hours, PBS (control and LPS) or 600 mg/kg APAP (APAP and LPS/APAP) was administered, and mice were euthanized 6 hours later. Hepatic mRNA expression of tumor necrosis factor-α, interferon (IFN)-β, IFN-γ, IL-10, and interferon-regulatory factor-1, as determined by RT-qPCR. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 versus control or as indicated. WT, wild type.

Supplemental Figure S8.

Acetaminophen (APAP) does not induce cytokine expression in hepatic stellate cell (HSCs). Rat (A) and mouse (B) HSCs were stimulated with 100 ng/mL lipopolysaccharide (LPS) or 0.1 mmol/L APAP for 6 hours, and various mRNA transcripts were measured via RT-qPCR. ∗∗P < 0.01, ∗∗∗P < 0.001 versus control.

Supplemental Figure S9.

Hepatic cytoglobin expression and microsomal ytochrome P450-2E1 (CYP2E1) activity in lipopolysaccharide (LPS) ± acetaminophen (APAP)–treated hepatic stellate cell (HSC)–sufficient (suff) and HSC-depleted (depl) mouse liver. The 6-week–old HSC-sufficient or HSC-depleted mice were administered 5 mg/kg LPS or phosphate-buffered saline (PBS; intraperitoneally), followed immediately by 1 mL saline (intravenously). After 15 hours (overnight), mice were administered 600 mg/kg APAP or PBS and euthanized 6 hours later. Control mice received PBS at both time points. A: Hepatic mRNA expressions of cytoglobin, as determined by RT-qPCR. B: Western blot analysis showing CYP2E1 protein expression. C: Hepatic microsomal CYP2E1 activity in HSC-sufficient and HSC-depleted mice. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. CT, control.

Supplemental Figure S10.

Interferon (IFN)-β ± acetaminophen (APAP)–induced liver injury in mice. The 6-week–old hepatic stellate cell (HSC)–sufficient or HSC-depleted mice were given 103 units of IFN-β, followed by 400 mg/kg APAP 3 hours later, and euthanized at 6 hours. Control mice received phosphate-buffered saline at both time points. A: Hematoxylin/eosin-stained liver sections of control and IFN-β–, APAP-, or IFN-β/APAP–treated mice. B: Corresponding serum alanine aminotransferase (ALT) levels. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Scale bars = 100 μm (A).