Acute liver failure (ALF) secondary to massive hepatic necrosis is a rapidly progressive clinical syndrome that frequently leads to multiorgan dysfunction (MOD), and it is associated with high morbidity and mortality rates. Acetaminophen (APAP)-induced hepatotoxicity is currently the most frequent cause of ALF in the United States and Europe.[1, 2]
APAP exerts its harmful effect through glutathione depletion and forms covalent bonds with cellular proteins; this produces massive centrilobular necrosis. A major paradigm shift in understanding the pathophysiology of ALF has occurred through discoveries of immune regulatory pathways contributing to ALF.[3, 4] Resident Kupffer cells and inflammatory infiltrates, the key components of the innate immune system, play a central role in the initiation, progression, and resolution of ALF.[4, 5] Systemic inflammatory response syndrome, independent of infections, is associated with worsening of the encephalopathy score and a poorer prognosis.[6, 7]
Toll-like receptors (TLRs) are transmembrane signaling proteins that recognize pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs), and they are crucial in the regulation of the innate immune response associated with hepatocellular damage (Supporting Fig. 1). Induced by these DAMPs/endogenous ligands produced after hepatocyte necrosis, Sterile inflammation activates the innate immune system and thereby amplifies the inflammatory cascade culminating in liver failure. Among the TLRs, TLR4 plays a pivotal role in the initiation of the immune response after liver injury.[10, 11]
N-Acetylcysteine is the standard of care for patients with an APAP overdose, and it is effective when it is administered at an early stage. Its prolonged use in late presenters is controversial. Its effectiveness in patients with established liver failure and MOD is not known.
Previous studies have shown improved survival for TLR4-knockout (KO) mice after galactosamine/lipopolysaccharide (LPS)-induced ALF and also for APAP animals pretreated with alcohol; both conditions are associated with increased circulating LPS.[13-16] However, the role of TLR4 in modulating end-organ dysfunction following ALF is uncertain. To that end, in this study we sought to investigate the role of TLR4 in ALF and subsequent end-organ dysfunction after APAP toxicity. We chose APAP-induced sterile inflammation without confounding factors such as LPS to define the progression of liver failure and MOD. The TLR4 antagonists eritoran (Eisai, Japan) and TAK-242 (TLR4 antagonist) (Takeda, Japan) have been tested in septic patients. Phase 2 studies have demonstrated safety but not clinical effectiveness, and as far as we know, they are not being developed for liver diseases. Because TLR4 is a multidomain protein, the available antagonists work at different sites of this protein, and their effectiveness may, therefore, differ in various clinical situations. Very recently, our collaborators developed a novel TLR4 specific antagonist, STM28, which is a peptide formed of 17 amino acids that binds directly to the extracellular domain of the TLR4 molecule. We hypothesized that in the mouse model of APAP-induced ALF, TLR4 plays an important role in the development of MOD and that its deletion or inhibition would protect mice from the effects of APAP-induced toxicity.
The aims of this study were to determine whether TLR4-KO mice would be protected from APAP-induced MOD and to explore whether the specific antagonist STM28 could prevent the progression of liver injury and end-organ dysfunction in mice with APAP-induced ALF.
MATERIALS AND METHODS
All experiments were conducted in accordance with local ethics approval and were subjected to the UK Animal Scientific Procedures Act of 1986. A well-characterized model of APAP-induced ALF was used for this study. Male mice were used for studies 1 and 2. APAP (Sigma-Aldrich, United Kingdom) was freshly prepared daily via dissolution in preheated, sterile normal saline at 37°C. APAP (500 mg/kg) was administered intraperitoneally after overnight fasting, and the mice were maintained at the temperature of 37°C thereafter. The mice then had free access to chow and water and were maintained on a 12-hour light/dark cycle.
Two sets of studies were performed consecutively. The first study was performed in TLR4-KO animals; these animals with a Tlr4Lps-del (TLR4 lipopolysaccharide deletion) spontaneous mutation have a 7-kb deletion in the Tlr4 gene (Jackson Laboratory, United States). Two breeding pairs of TLR4-KO mice were purchased from the Jackson Laboratory and bred in house. The littermates were subsequently used for the experiments. Two groups of mice—a wild-type (WT) group (C57BL/6 [C57 black 6]; n = 8) and a TLR4-KO group, (n = 8)—were fasted overnight.
After APAP administration, the animals were followed until they reached either the coma stage or the designated study end at 5 days. The second study was performed to explore the effects of the TLR4 antagonist STM28 on the prevention of ALF in APAP mice. Three groups of CD1 mice (n = 8 for each group) were studied: a naive group, an APAP group, and an APAP+STM28 (TLR4 antagonist) group (Division of Microbiology, National Institute of Health Sciences, Tokyo, Japan). STM28 (20 μg) or saline was given intraperitoneally 1 hour before APAP, and a second dose was given 5 hours after APAP. The untreated APAP animals were sacrificed at a mean of 7.5 ± 2.1 hours after APAP, the point at which they developed signs of coma. Animals in the STM28-treated group were sacrificed 8 hours after APAP.
Sampling and Processing
Blood was collected via cardiac puncture in heparinized tubes before terminating the animals under anesthesia with isoflurane and was centrifuged (at 4°C and 3500 rpm for 10 minutes), and the plasma was stored at −80°C for later analysis. A fraction of the blood samples from the second set of experiments were processed for cell isolation and stained with specific fluorochrome-labeled monoclonal antibodies (mAbs; Supplementary Information). Parts of livers, kidneys, and brains were snap-frozen (supplementary information) in liquid nitrogen for western blot analyses, and the rest were stored in 10% buffered formalin for histopathology and immunohistochemical staining for TLR4. The frontal part of each brain was used for measurements of brain water.
Liver and Renal Biochemistry
Plasma samples were analyzed for alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine, and ammonia (NH3); Cobas Integra 400, Roche Diagnostics, West Sussex, United Kingdom.
Plasma and Tissue Cytokines
Each liver tissue (100 μg) was homogenized in a trishydroxymethylaminomethane/hydrochloride lysis buffer (pH 7.5). The protein concentration was determined with a biuret assay. Tumor necrosis factor α (TNF-α) and interleukin-10 (IL-10) in the plasma and TNF-α and IL-1α in the liver were measured with a commercial cytometric bead assay (Becton Dickinson, United Kingdom), quantified with flow cytometry (FACSCanto II, Becton Dickinson), and analyzed with appropriate software (FCAP Array, Soft Flow Hungary, Ltd).
Western Blot Analysis
Nuclear factor kappa B (NFkB)–p65 protein expression in the organs was determined with western blotting as previously described. The proteins were probed with a primary anti–NFkB-p65 antibody (rabbit anti–NFkB-p65, Signalway, United Kingdom; 1:500 dilution for 16 hours at 4°C), and this was followed by incubation with a peroxidase-conjugated anti-rabbit secondary antibody (ProSci, United Kingdom; 1:5000 dilution). The protein intensity was quantified with a public-domain, Java-based image processing program (ImageJ). The results were expressed as the NFkB-p65/β-actin ratio (rabbit anti–β-actin, Abcam, United Kingdom; 1:1000 dilution).
The histological assessment was performed with hematoxylin and eosin staining.
Immunohistochemistry for TLR4, CD68, and F4/80
Sections from livers and kidneys were deparaffinized with xylene and rehydrated in graded ethanol. After they were washed with a phosphate-buffered solution (×3), they were heated in a microwave oven at 95°C for 10 minutes with a citrate buffer (Dako, United Kingdom; pH 6.0). The slides were then treated with a 3% hydrogen peroxidase solution, and this was followed by overnight incubation with an anti-TLR4 antibody (Lifespan Bioscience, United Kingdom; 1:200 dilution). To characterize the activated Kupffer cells, anti-CD68 (Abcam; 1:100 dilution) and anti-F4/80 (Abcam; 1:200 dilution) were used. A ready-to-use secondary antibody (EnVision, Dako) was used for 30 minutes at room temperature, and subsequently developed using chromogen and the nuclei were counterstained with hematoxylin. In between steps, the slides were washed with a phosphate-buffered solution (×3).
Brain Water Measurement
The frontal part of the brain (5 mm3) was used to measure the brain water. The specified portion was immediately weighed on an electronic scale to obtain the wet weight. The brain samples were dried in an oven at 100°C for 24 hours to obtain the dry weight. The brain water content was calculated according to the following formula:
Peripheral blood was processed and stained with fluorochrome-labeled mAbs specific for myeloid cells, dendritic cells (DCs), resident and inflammatory monocytes, neutrophils, CD4/CD8 T lymphocytes, B cells, and natural killer cells according to the manufacturer's instructions (BD Pharmingen, San Diego, CA). The surface markers used to identify the different subpopulations are shown in Supporting Table 1. Leukocytes were gated according to their size (forward light scatter) and granularity (side light scatter) and were stained with CD45 mAbs. Dead cells were excluded by propidium iodide staining (BD Pharmingen). All antibodies were purchased from BD Pharmingen. Flow cytometry data acquisition and analysis were performed on FACSCanto II with FACSDiva software (Becton Dickinson, San Diego, CA).
Table 1. Biochemical Parameters and Brain Water in the Naive, APAP, and TLR4-KO Groups
Brain Water (%)
NOTE: Plasma biochemistry and brain water values are expressed as means and standard errors of the mean. The statistical analysis was performed with Bonferroni's multiple comparison test for all the groups.
The data were expressed as means and standard errors of the mean. The significance of differences was tested with a 1-way analysis of variance with Bonferroni's multiple comparison test or the Student t test as appropriate. The probability of cumulative survival was analyzed with Kaplan-Meier analysis in combination with a log-rank test. P < 0.05 was considered statistically significant. All statistical analyses were performed with SPSS 14.0 Statistics and GraphPad Prism 4.0 (GraphPad Software, Inc., San Diego, CA).
Study 1. Role of TLR4 in the Modulation of ALF in APAP Mice
Clinical and Biochemical
After APAP administration, the WT mice developed a coma at a mean of 9.4 ± 4.3 hours. The mean time to a coma for the TLR4-KO mice was 68.1 ± 18 hours (P < 0.01; Fig. 1A). A hunched posture, lethargy, or persistent recumbency in conjunction with a rough or unthrifty hair coat and insensibility to external stimuli were considered to be humane endpoint criteria for a coma. TLR4-KO mice were found to have reduced APAP-induced liver injury (as evidenced by a significantly lower ALT level) in comparison with WT mice (P < 0.05). The serum creatinine level was significantly lower in the TLR4-KO group versus the WT animals (P < 0.01). Similarly, the plasma NH3 level in TLR4-KO mice was significantly lower than the level in WT mice (P < 0.05; Table 1). This was associated with a significant reduction in brain water in TLR KO mice versus WT mice (P < 0.01; Table 1).
After APAP administration, WT animals showed features of extensive pericentral hepatic necrosis and vacuolization of hepatocytes in the hepatic parenchyma (Fig. 2A-b). In contrast, the liver damage in TLR4-KO mice was limited to the presence of vacuolization with occasional minimal pericentral necrosis (Fig. 2A-c and Supporting Table 2). There was no increase in inflammatory infiltrates in the WT animals that were administered APAP.
Table 2. Plasma Biochemistry, Brain Water, and Cytokines in the Naive, APAP, and APAP+STM28 Groups
Brain Water (%)
NOTE: Plasma biochemistry, cytokine, and brain water values are expressed as means and standard errors of the mean. The statistical analysis was performed with Bonferroni's multiple comparison test for all the groups.
After APAP administration, TLR4-KO mice showed low expression of activated Kupffer cells according to CD68 and F4/80 staining in comparison with WT control animals that were administered APAP (Fig. 2B-d,h).
Study 2. Effect of the TLR4 Antagonist on the Prevention of APAP-Induced ALF in Mice
Clinical and Biochemical
After APAP administration, the CD1 mice developed a coma at a mean of 7.5 ± 2.1 hours. None of the animals in the APAP+STM28 group developed a coma before the completion of the study at 8 hours (the onset of a coma was established as described for study 1). There was a marked increase in plasma liver enzymes (ALT and AST) in the mice administered APAP versus the naive controls (P < 0.001), and this was significantly reduced in the APAP+STM28 animals (P < 0.01). The administration of APAP led to an increase in the plasma creatinine level in APAP mice versus naive mice (P < 0.05), but this was prevented in the APAP+STM28 group (P = 0.09). Similarly, plasma NH3 levels were significantly higher in the APAP group versus the naive group (P < 0.01), but they were reduced significantly (P < 0.05) in the APAP+STM28 animals. Also, the pretreatment of mice with STM28 prevented an increase in brain water (P < 0.05 for APAP+STM28 versus APAP; Table 2).
The administration of APAP led to extensive centrilobular hepatic necrosis/degeneration in conjunction with microvacuolization and macrovacuolization. There did not appear to be an increase in inflammatory cell infiltrates in the APAP group versus the naive group (Fig. 2A-d,e). Treatment with STM28 ameliorated the injury by limiting the area of necrosis. There was scattered vacuolization of hepatocytes in the periportal areas (Fig. 2A-f and Supporting Table 2).
Plasma and Tissue Cytokines
The administration of APAP was associated with a rise in plasma (P < 0.05) and liver tissue TNF-α levels (P < 0.01) in the APAP group versus the naive animals; this was reduced in the APAP+STM28-treated animals (P < 0.05 for plasma), but the reduction did not reach statistical significance at the tissue level. APAP administration led to a rise in plasma IL-10 versus the naive controls (P < 0.05), and this was reduced in the STM28-treated animals with a trend toward significance (P = 0.09). The administration of APAP led to an increase in liver IL-1α levels in comparison with the naive controls (P < 0.01), and this was reduced significantly in the STM28-treated group (P < 0.05; Table 2).
Western Blot Analysis
After APAP administration, NFkB-p65 expression was significantly higher in the kidneys (P < 0.01) and brain (P < 0.001) in comparison with the naive group, and there was a limited increase in its expression in the liver (P = 0.1; Fig. 1B-D). STM28 treatment significantly reduced the expression of NFkB-p65 in the kidneys (P < 0.05) and brain (P < 0.01; Fig. 1B,C). In contrast, treatment with STM28 was associated with significantly greater expression (P < 0.001) of NFkB-p65 in the liver (Fig. 1D).
Marked up-regulation of TLR4 was evident in hepatocytes surrounding necrotic liver tissue in mice administered APAP in comparison with the constitutive expression of TLR4 in periportal hepatocytes in the naive group (Fig. 3A,B). This was associated with an increase in the intensity of TLR4 staining in the renal tubular cells of mice administered APAP (Fig. 3E). Treatment with the TLR4 antagonist (STM28) led to a reduction in TLR4 staining, which was limited to the pericentral hepatocytes, and a reduction in the intensity of the expression of TLR4 in renal tubules (Fig. 3C,F).
CD68 and F4/80
There was a marked increase in the expression and intensity of CD68- and F4/80-positive cells in the area surrounding the necrotic liver tissue in the group administered APAP (Fig. 2B-b,f). In the APAP+STM28 group, there was a marked reduction in CD68 and F4/80 cells in comparison with the APAP group, and they were confined to the perisinusoidal area (Fig. 2B-c,g).
Peripheral Immune Cells
APAP-induced liver injury led to a significant increase in the percentage of total myeloid cells and their subtypes (monocytes and neutrophils) in comparison with the naive group. After STM28 treatment, there was a marked reduction in the total numbers of myeloid cells (58% for the APAP group versus 39% for the APAP+STM28 group, P < 0.001), neutrophils (37% versus 22%, P < 0.001), monocytes (11% versus 6%, P < 0.01), and the resident monocyte subtype (15% versus 7%, P < 0.05; Fig. 4A). The individual populations of these cells in each group are shown as percentages in Fig. 4B. Total DCs (identified as a CD11c population) constituted approximately 1% of the total leukocyte count. No significant differences were observed in the APAP and APAP+STM28 groups with respect to DCs (56% versus 54% for myeloid cells and 10% versus 15% for plasmacytoid cells), B cells (6% versus 7%), T cells (14% versus 15%), or natural killer cells (1.6% versus 1%; Supporting Fig. 2A-E).
APAP-induced ALF is characterized by massive liver necrosis, which is often rapidly progressive and leads to hepatic encephalopathy culminating in MOD. Approaches to prevent its progression are limited, with the only definitive available treatment being liver transplantation. The data presented in this article provide compelling evidence for TLR4 being an important mechanism in the development of APAP-associated hepatic dysfunction and MOD in ALF, and they suggest that TLR4 expression in peripheral organs may be associated with the development of MOD. Therefore, antagonizing this receptor may provide a therapeutic benefit by preventing the progression of liver injury and the consequent renal and brain dysfunction characteristic of ALF.
In patients with ALF, systemic inflammatory response syndrome, independent of infections, is associated with worsening of encephalopathy, renal failure, and a poorer prognosis.[6, 7, 22, 23] TLRs are involved in mediating immune responses after exposure to either PAMPs or DAMPs, which result in the secretion of a host of cytokines and activation of the inflammatory cascade. APAP administration in our model produced the classic manifestations of ALF, which was characterized by liver cell necrosis and MOD. These were markedly attenuated in the TLR4-KO animals. Our findings are in concordance with previous studies showing improved survival for TLR4-KO mice after galactosamine/LPS- and APAP-induced ALF.[13-16] The data from the present study extend these observations and show that the protective effect of TLR4-KO is true even in animals with sterile inflammation and MOD after APAP. This supports the hypothesis that the expression of TLR4 in extrahepatic organs is associated with a significant amount of damage, and appropriate TLR4 antagonists may protect animals from developing APAP-induced ALF and subsequent end-organ dysfunction. Several targeted therapies against TLR4 have been devised, but the exact mechanisms of action of these TLR4 antagonists have yet to be fully elucidated. Most of those described to date are lipid A mimetic agents that exert their inhibitory effects by binding to the extracellular TLR4/CD14/MD2 complex and serve as competitive inhibitors of TLR4. A second type of TLR4 antagonist (TAK-242) exerts its inhibitory effect in the intracellular domain and is predominantly represented as an LPS inhibitor. Kitazawa et al.[14, 15] used a small-molecule inhibitor of TLR4 (E5564) in a rat model of galactosamine/LPS-induced ALF and showed improved survival. In the present study, we used a novel TLR4 antagonist, STM28, that was developed by our collaborators. STM28, which specifically binds with extracellular TLR4 and dose-dependently inhibits an LPS-induced cytokine surge, also prevents galactosamine-associated lethality in mice. Emerging evidence shows that the sterile inflammation in APAP is associated with the release of high mobility group box 1 (HMGB1) and other DAMPs such as DNA fragments after cell death, and these are potent endogenous ligands of TLR4 that may perpetuate the inflammatory cascade.[9, 24, 25]
The improved outcome of APAP mice treated with the TLR4 antagonist could be attributed to its dampening effect on the overt cytokine surge following APAP-induced hepatocellular damage. Higher concentrations of circulating TNF-α, IL-6, and IL-10 have been observed in patients with APAP-induced ALF requiring transplantation and in nonsurvivors.[26, 27] In our experimental study, there was a surge in the plasma TNF-α level after APAP administration, and this was reduced in the STM28-treated animals. There was a concomitant increase in TNF-α and IL-1α in the liver, which is a feature of APAP hepatotoxicity,[28, 29] and this was ameliorated in the STM28-treated animals; this indicates the role of TLR4 in modulating hepatic inflammation in APAP-induced ALF. STM28 treatment also prevented a rise in the plasma IL-10 level, which, when raised, is associated with an increased risk of infection and a poor prognosis. After liver necrosis, mice that were administered APAP had a limited increase in the expression of hepatic NFkB-p65 in comparison with other organs. Its expression in the liver increased further in the APAP+STM28 group; this observation at first glance seems counterintuitive because its activation can lead to the generation of proinflammatory cytokines. NFkB is a pleiotropic transcription factor that plays a vital role in maintaining liver homeostasis, including cell death and survival. NFκBp65 KO has previously been associated with extensive hepatocyte apoptosis and embryonic death. The inhibition of NFkB by sublethal doses of APAP has been found to sensitize hepatocytes to the cytotoxic actions of TNF-α. Our data support the view that up-regulation of NFkB-p65 in the liver was most likely due to a positive feedback signal following limited liver damage in the STM28 treatment group, which induced a regenerative drive for the viable cells. However, this was not true for the APAP-treated animals, in which extensive necrosis failed to instigate the NFkB signal. It is of particular interest that we noted viable, healthy hepatocytes in mice treated with STM28. It is likely that STM28 specifically inhibits TLR4 but does not interfere with other TLRs, leaves the NFkB pathway partially active, and allows liver regeneration. Further studies are required to ascertain whether the reduction in the severity of liver injury in the TLR4-KO animals and those treated with STM28 was the result of increased liver regeneration due to the activation of early regenerating genes such as NFkB.[33, 34]
After acute liver injury, there is a surge in the circulating inflammatory cells responsible for the initiation and progression of ALF by secreting a host of proinflammatory cytokines and chemokines in patients with APAP-induced ALF. Furthermore, chemokines secreted from necrotic hepatocytes, Kupffer cells, and sinusoidal endothelial cells facilitate the migration of inflammatory infiltrates attributable to the proinflammatory milieu in APAP hepatotoxicity.[4, 5] Although we did not find an increase in inflammatory cell infiltrates, there was a marked increase in CD68- and F4/80-positive Kupffer cells in WT animals administered APAP in contrast to naive and APAP+STM28 animals. Additionally, TLR4-KO mice that were administered APAP showed low expression of activated Kupffer cells. Fisher et al. showed an abrogation in the inflammatory response after Kupffer cell depletion in APAP-induced liver failure. In addition, Chinnery et al. very recently used a dual approach of Fas-mediated macrophage depletion and a bone marrow chimeric mouse model of myeloid-derived donor TLR+/+ cells in TLR4−/− recipient mice to evaluate the role of myeloid cells in LPS-induced corneal inflammation. The authors convincingly showed that the resident myeloid lineage cells, including macrophages, played a significant role in TLR4-mediated corneal inflammation. Our data show an attenuation in the severity of liver injury together with a reduction in the activation of Kupffer cells in the STM28-treated group, and this implies a possible contributory role of these cells in liver injury.
Additionally, the peripheral inflammatory response was abrogated in the STM28-treated animals. The exact mechanism of MOD in ALF is unclear. Anecdotal reports indicate that the removal of a necrotic liver from a patient with ALF will stabilize the clinical condition, and this indicates that the necrotic liver may produce some undefined substances leading to end-organ dysfunction.[37, 38] The mechanism or mechanisms of the relay of interorgan inflammatory crosstalk are ill understood and remain to be elucidated. However, emerging evidence suggests that sterile inflammation and subsequent TLR4 activation across organ systems may be due to liver injury. A modern concept involving TLR activation in the cross-organ involvement of inflammation in the setting of hemorrhagic shock has recently been proposed.[39, 40] More recently, Antoine et al. demonstrated an increase in total and acetylated HMGB1 in the serum of patients with APAP toxicity and its association with a poor prognosis. In this study, we report that TLR4 activation in extrahepatic organs possibly contributes to the pathogenesis of MOD in APAP-induced ALF, and as such, the inhibition of TLR4 may prove to be an effective strategy for the prevention of MOD (Supporting Fig. 3).
Further studies are needed to elucidate whether the up-regulation of TLR4 in the kidneys is secondary to liver injury or sterile inflammation in the kidneys is independent of liver injury because a cytochrome p450–metabolizing enzyme system similar to its hepatic counterpart exists in the kidneys. Additionally, the STM28-treated animals and the TLR4-KO animals were protected against the brain effects of APAP toxicity, possibly because of a combination of liver injury prevention, reduced hyperammonemia, and an ameliorated inflammatory response. As with the kidneys, there was a significant reduction in brain NFkB-p65 expression.
In conclusion, the data described in this study suggest that TLR4 is a distinct target for the prevention of the progression of ALF and associated multiorgan failure. TLR4-KO mice were protected from APAP-induced liver failure and extrahepatic organ dysfunction and had improved survival. Moreover, the TLR4 antagonist STM28 was able to dampen excessive inflammation, reduce Kupffer cell activation, and preserve end-organ function. However, to take STM28 from bench side to bedside, further studies are required to ascertain its effectiveness as a therapeutic modality alone or in combination with N-acetylcysteine after APAP administration in the prevention of ALF.