Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Supported by the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich SFB 575 “Experimental Hepatology” Düsseldorf.
Astrocytes play an important role in the pathogenesis of hepatic encephalopathy (HE) and ammonia toxicity, whereas little is known about microglia and neuroinflammation under these conditions. We therefore studied the effects of ammonia on rat microglia in vitro and in vivo and analyzed markers of neuroinflammation in post mortem brain tissue from patients with cirrhosis with and without HE and non-cirrhotic controls. In cultured rat microglia, ammonia stimulated cell migration and induced oxidative stress and an up-regulation of the microglial activation marker ionized calcium-binding adaptor molecule-1 (Iba-1). Up-regulation of Iba-1 was also found in the cerebral cortex from acutely ammonia-intoxicated rats and in the cerebral cortex from patients with cirrhosis who have HE, but not from patients with cirrhosis who do not have HE. However, ammonia had no effect on microglial glutamate release, prostaglandin synthesis, and messenger RNA (mRNA) levels of inducible nitric oxide synthase (iNOS), cyclooxygenase-2 (COX-2), and the proinflammatory cytokines interleukin (IL)-1α/β, tumor necrosis factor α, or IL-6, whereas in cultured astrocytes ammonia induced the release of glutamate, prostaglandins, and increased IL-1β mRNA. mRNA and protein expression of iNOS and COX-2 or mRNA expression of proinflammatory cytokines and chemokine monocyte chemoattractive protein-1 in cerebral cortex from patients with liver cirrhosis and HE were not different from those found in patients with cirrhosis who did not have HE or control patients without cirrhosis. Conclusion: These data suggest that microglia become activated in experimental hyperammonemia and HE in humans and may contribute to the generation of oxidative stress. However, HE in patients with liver cirrhosis is not associated with an up-regulation of inflammatory cytokines in cerebral cortex, despite microglia activation. (HEPATOLOGY 2011;)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepatic encephalopathy (HE) defines a neuropsychiatric syndrome associated with acute or chronic liver disease. It is characterized by impaired motor functions, cognitive dysfunction, and emotional/affective and behavioural disturbances.1 It is generally accepted that HE represents a primary gliopathy due to astrocyte swelling and oxidative/nitrosative stress, which disturbs astrocytic/neuronal communication, synaptic plasticity, and oscillatory networks in the brain, which finally trigger the clinical HE symptoms.1-3
Studies on cultured astrocytes and HE-relevant animal models suggest that ammonia intoxication triggers a self-amplifying cycle between oxidative and osmotic stress.2-4 Here, astrocyte swelling promotes prostanoid-dependent glutamate exocytosis, which triggers an oxidative/nitrosative stress response by way of N-methyl-D-aspartate receptors.5 The latter induces tyrosine nitration of astrocytic key proteins and RNA oxidation in astrocytes and neurons with potential impact on local postsynaptic protein synthesis, and affects gene transcription due to an interference with intracellular zinc homeostasis.6-8 Most importantly, increased protein tyrosine nitration and RNA oxidation were shown in post mortem brain tissue from patients with liver cirrhosis and HE, but not from patients with cirrhosis who did not have HE.9
Whereas astrocytic and neuronal dysfunction has been studied extensively in HE and hyperammonemia, the role of microglia in the pathobiology of HE is less clear. Recently, microglia activation has been shown in the rat brain after hyperammonemic diet intake and following bile duct ligation10 or hepatic devascularization with acute liver failure,11 but not after portal vein ligation.12
Microglia activation has been shown in cerebral infections or in neurodegenerative diseases such as Alzheimer disease.13, 14 Here, microglia experience a change in functional phenotype, which is reflected at the morphological level by the transition from a ramified into an ameboid appearance.15, 16 However, microglia activation can result in a broad spectrum of phenotypic and functional diversity, and resting microglia can adopt an alerted phenotype before becoming a fully activated, so-called reactive cell.16
Reactive microglia can release large amounts of proinflammatory and cytotoxic mediators such as nitric oxide derived from inducible nitric oxide synthase (iNOS), prostanoids, or inflammatory cytokines, thereby promoting further tissue damage and neuronal dysfunction.15, 16 However, HE is not characterized by neurodegeneration, and HE symptoms are potentially reversible.1, 17 We therefore studied the effect of ammonia on microglia activation in vivo and in vitro and tested for markers of microglia activation and neuroinflammation in post mortem brain tissue from patients with cirrhosis with and without HE. The findings suggest that microglia become activated in response to ammonia and in patients with cirrhosis who have HE, but is not reactive with regard to cytokine formation.
Detailed information about materials used in this study can be found in the Supporting Information.
Acute Ammonium Acetate Intoxication of Rats.
Information about experimental animal treatment in this study can be found in the Supporting Information.
Preparation and Cultivation of Rat Brain Astrocytes and Microglia.
Cells were prepared from cerebral hemispheres of newborn male Wistar rats (P1-P3) as described recently6 and in the Supporting Information.
Post Mortem Human Brain Tissue.
Post mortem human brain tissue was obtained at autopsy of eight control subjects and eight patients suffering from liver cirrhosis and accompanying HE. Controls were free from hepatic or neurological disorders at the time of death. Informed written consent was given either by the patients or by their relatives or had been included in the body donor program of the Department of Anatomy, University of Düsseldorf, Germany. In addition to these brain samples from European patients, brain samples from four Australian controls without cirrhosis, four Australian patients with cirrhosis who did not have HE, and five Australian patients with cirrhosis who had HE were analyzed. These brain samples were obtained from the Australian Brain Donor Programs New South Wales Tissue Resource Centre, which is supported by the University of Sydney, National Health and Medical Research Council of Australia, Schizophrenia Research Institute, National Institute of Alcohol Abuse and Alcoholism, and New South Wales Department of Health. Details on the medical history of the European and Australian patients were published recently.9
Immunofluorescence analysis was performed as described recently6 and in the supplemental materials.
Analysis of Cytokine Messenger RNA Expression by Quantitative Reverse-Transcription Polymerase Chain Reaction.
Real-time polymerase chain reaction (PCR) was performed as described recently8 and in the Supporting Information.
Western Blot Analysis and Densitometric Analysis.
Western blot analysis was performed as described recently6 and in the Supporting Information.
Measurement of Reactive Oxygen and Nitrogen Species.
Reactive oxygen and nitrogen species were detected using the reactive oxygen species (ROS)-sensitive fluorescence dye CM-H2DCFDA and fluorescence microscopy as described in the Supporting Information.
Measurement of Cell Migration and Phagocytosis.
Cell migration was assessed using a commercial colorimetric cell migration assay (Chemicon QCM Colorimetric Cell Migration Assay) according to the manufacturer's instructions. Phagocytosis of microglia was assessed by detection of latex beads using fluorescence microscopy. For further details, see the Supporting Information.
Determination of Cell Diameter and Filopodia Length.
For measuring cell diameter and filopodia length real-time differential interference contrast microscopy was performed as described in the Supporting Information.
Measurement of L-Glutamate and Prostanoids in Culture Medium.
L-Glutamate in culture medium was measured as described.5 Concentrations of 6- keto-prostaglandin F1α, which is a stable metabolite of prostaglandin I2, and prostaglandin E2 (PGE2) were determined in culture medium by using commercial enzyme immunoassay kits (Cayman Chemicals, Hamburg, Germany) according to the manufacturer's protocol.
Data processing was performed using Excel and Graph Pad Prism (4.0) for Windows. Data are presented as the mean ± SEM. Descriptive statistics were performed using a Student t test or one-way analysis of variance followed by Tukey's or Dunnett's multiple comparison post hoc test, where appropriate. P ≤ 0.05 was considered statistically significant.
Effect of Ammonia on Microglia Migration, Morphology, and Phagocytosis.
Microglia activation is associated with increased cell migration, which depends on a polarized cell morphology and local protrusion formation.18 As shown by brightfield microscopy and photometric quantification, ammonia treatment significantly increased microglia migration through a porous membrane within 3 hours by approximately 40% when compared with untreated controls (Fig. 1A,B).
Quiescent, ramified microglia cells continuously monitor their surrounding environment through filopodia expansion and retraction.19 As shown by time-lapse microscopy, primary microglia in culture expresses fine-structured filopodial processes which continuously reorganize while probing their microenvironment (Fig. 1C). However, upon addition of NH4Cl (5 mmol/L) the cells retract within seconds their filopodia to a length of about 50% of that found in untreated control cells (Fig. 1D). This retraction was accompanied by a reduction of the cell diameter by about 30% as compared with the control condition (Fig. 1D).
Microglia are a major phagocytosing cell type that removes cell debris and pathogens in the brain.15, 16 Brain dysfunction in neurodegenerative diseases is frequently associated with increased phagocytotic activity, which may represent another surrogate marker for microglia activation.15, 16, 19 As shown in Fig. 2, ammonia inhibited phagocytosis in a subset of microglial cells, thereby reducing overall phagocytosis to approximately 65% of the control condition (Fig. 2A). However, in phagocytosing microglia cells, the number of phagocytosed fluorescent latex beads per cell was not significantly affected (Fig. 2B,C). These findings may reflect the known heterogeneity of microglia in the brain.
Ammonia and Microglia Activation.
Upon activation, microglia increase the expression of the ionized calcium-binding adaptor molecule-1 (Iba-1).18 This protein transduces calcium signals for reorganization of the cytoskeleton, thereby allowing for morphology changes and cell migration.18
As shown by immunofluorescence analysis (Fig. 3A), NH4Cl induced Iba-1 up-regulation in cultured microglia in a time- and concentration-dependent manner (Fig. 3A-D). Significant Iba-1 up-regulation occurred 6 hours after NH4Cl (5 mmol/L) treatment (Fig. 3A) but not at 1 or 3 hours of exposure (data not shown). Ammonia concentrations of 5 mmol/L (6 hours) and 1 mmol/L (20 hours), respectively, were sufficient to induce a significant Iba-1 up-regulation to approximately 1.25 or 1.5-fold of control. However, Iba-1 messenger RNA (mRNA) expression was not up-regulated by NH4Cl (5 mmol/L) after 6 hours (0.65 ± 0.12-fold of control, n = 3) or 20 hours after treatment (0.70 ± 0.07-fold of control, n=4 [P = 0.02]).
Microglia activation is frequently accompanied by an increased formation of ROS and an induction of iNOS.13, 14, 16 As shown in Table 1, NH4Cl increased microglial ROS production significantly in a time- and dose-dependent manner as measured by DCF-fluorescence. Pretreatment of microglia with apocynine (300 μmol/L, 30 minutes pretreatment) completely abolished the NH4Cl (5 mmol/L, 6 hours)-induced DCF-fluorescence increase suggestive for an activation of the reduced form of nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase by ammonia (0.96 ± 0.05-fold of apocynine-treated controls, n = 4). This finding also suggests that the NH4Cl-induced increase of DCF-fluorescence is due to ROS, but not nitrogen species.
Table 1. Ammonia Induces Production of ROS in Cultured Rat Microglia
NH4Cl (1 mmol/L)
NH4Cl (5 mmol/L)
Cultured rat microglia were either exposed to NH4Cl (1 and 5 mmol/L) or left untreated for 3, 6, or 20 hours. ROS production was detected and quantified in H2DCFDA-loaded cells by way of epifluorescence microscopy. Relative DCF fluorescence in NH4Cl-exposed microglia is expressed relative to the fluorescence found under the respective control condition. Data are presented as the mean ± SEM. Results are based on three to six independent experiments.
Statistically nonsignificant compared with untreated microglia.
Statistically significant compared with untreated microglia.
Statistically significant compared with NH4Cl (1 mmol/L)–treated microglia.
Increased iNOS protein and mRNA expression was found in ammonia-treated cultured rat astrocytes6, 20 and in brains of portocaval-shunted rats in vivo.21 However, iNOS mRNA expression was significantly reduced by approximately 50% in ammonia (5 mmol/L, 6 hours)-treated microglia. As for control, lipopolysaccharide (LPS, 1 ng/mL, 6 hours) strongly increased iNOS mRNA expression (Supporting Information Fig. 1A). In addition, NH4Cl (5 mmol/L) exposure for 20 hours had no effect on iNOS protein expression as detected by immunofluorescence (data not shown).
Activated microglia can express cyclooxygenase-2 (COX-2) and can be a powerful source of proinflammatory prostanoids, such as PGE2. However, NH4Cl (5 mmol/L) had no significant effect on COX-2 mRNA expression in microglia and astrocytes (Supporting Information Fig. 1B) and even lowered the PGE2 and 6-keto prostaglandin F1α (PGF1α) content of microglial supernatants after 6 hours and 20 hours of ammonia exposure (Fig. 4B). On the other hand, and in contrast to microglia, NH4Cl (5 mmol/L) stimulated the release of PGE2 and 6-keto PGF1α from cultured astrocytes (Fig. 4A). LPS treatment (1 ng/mL, 6 hours or 20 hours), which served as a positive control, increased PGE2 and 6-keto PGF1α release from both cultured microglia and astrocytes (Fig. 4A,B).
Extracellular glutamate can promote neuroinflammation by overactivation of N-methyl-D-aspartate receptors.22 In order to analyze the effect of ammonia on glutamate release, cultured microglia and astrocytes were incubated with NH4Cl (5 mmol/L) for 6 hours and 20 hours, and the glutamate content was measured in the supernatant. As shown in Supporting Information Fig. 2, no increase in extracellular glutamate was detected in the supernatant of cultured microglia exposed to NH4Cl (5 mmol/L) for 6 hours or 20 hours, whereas ammonia triggered glutamate release from cultured astrocytes as described recently.5
Activated microglia can promote neuroinflammation through the release of proinflammatory cytokines,23 which may play a role in the pathobiology of HE.1, 10, 11, 24 As shown in Fig. 5, treatment of cultured microglia or astrocytes with NH4Cl did not significantly change tumor necrosis factor α (TNF-α), interleukin (IL)-1α, or IL-6 mRNA expression. Interleukin-1β mRNA expression was reduced by NH4Cl treatment in microglia, but increased in astrocytes. LPS, which served as a positive control, strongly increased cytokine mRNA expression (Fig. 5) and prostanoid synthesis (Fig. 4) in both cell types.
Markers of Neuroinflammation in Cerebral Rat Cortex After Ammonia Intoxication In Vivo.
As shown by immunofluorescence (Fig. 6A) and western blot analysis (Fig. 6B,C), intraperitoneal injection of ammonium acetate (4.5 mmol/kg) increased Iba-1 expression in the cerebral cortex within 6 hours, suggestive for in vivo microglia activation after ammonia intoxication. However, cerebrocortical Iba-1 mRNA levels were not significantly altered in ammonium acetate (NH4Ac)-treated rats (0.84 ± 0.07-fold of control, n = 4). Moreover, expression levels of the microglial activation marker proteins CD74 and CD6812 remained unchanged after NH4Ac treatment (Fig. 6B,C), and ramified microglial morphology was preserved in NH4Ac-treated rats (Fig. 6A). This contrasts the in vitro finding depicted in Fig. 1 and may be due to different ammonia concentrations in rats in vivo.7
As shown by real-time PCR and western blot analysis, neither iNOS nor COX-2 mRNA and protein expression in the cerebral cortex were affected by ammonium acetate treatment in vivo (Supporting Information Fig. 3A-D). In addition, mRNA expression of the proinflammatory cytokines TNF-α, IL-1α/β, or IL-6 in the cerebral cortex was not significantly affected after acute ammonium acetate challenge (Supporting Information Fig. 4).
Markers for Neuroinflammation in Cerebral Cortex from Patients Who Have Cirrhosis With and Without HE and Controls.
As shown by western blot analysis (Fig. 7A,B), expression of the microglial activation marker Iba-1 was significantly increased in post mortem cortical brain tissue from patients with liver cirrhosis and HE, but not from patients with cirrhosis who did not have HE. This indicates that HE, but not cirrhosis per se, is associated with microglia activation.
As shown recently for iNOS protein,9 iNOS mRNA levels in the cerebral cortex were not significantly different between controls without cirrhosis and patients with cirrhosis, regardless of whether HE was present or not (Supporting Information Fig. 5A). Similar findings were obtained for the expression of COX-2 protein and mRNA (Supporting Information Fig. 5B-D). There were also no significant differences in the mRNA expression levels of the proinflammatory cytokines TNF-α, IL-1α/β, or IL-6 (Fig. 8A) or the chemokine monocyte chemoattractive protein-1 (MCP-1) (Supporting Information Fig. 6) in the cerebral cortex in patients with liver cirrhosis and HE when compared with controls or patients with cirrhosis who do not have HE. In these human brain samples, protein levels for TNF-α and cleaved IL-1β protein were below the detection limit, whereas the IL-1β precursor protein was detectable. In contrast, IL-1β precursor as well as TNF-α proteins were both up-regulated in the cerebral cortex of a patient with multiple sclerosis that served as a positive control (Fig. 8B)
Ammonia-Induced Microglia Activation.
It is widely accepted that HE represents a primary gliopathy in which ammonia, cell swelling, and oxidative/nitrosative stress play key roles. Studies on ammonia effects in cultured rat astrocytes suggest that astrocytes may contribute to cerebral neuroinflammation in HE through the release of glutamate, prostanoids, and reactive oxygen/nitrogen species due to ammonia-induced up-regulation of iNOS and NADPH-oxidase activation.5, 6, 25 Impaired neurotransmission associated with microglia activation and increased cerebral cytokine synthesis has been shown in different animal models for chronic HE.10, 26, 27 However, the role of microglia in the pathogenesis of acute ammonia toxicity and HE is largely unknown. Microglia activation in the rat brain has been reported to occur in response to a hyperammonemic diet and bile duct ligation10 and in acute liver failure due to hepatic devascularisation.11 However, it remained unclear whether microglia activation is triggered by ammonia directly or represents a secondary event.
As shown in the present study, ammonia directly activates primary rat microglia as shown by the induction of the microglial activation marker protein Iba-1. Iba-1 serves as an actin–cross-linking adaptor that facilitates membrane reorganization required for migration and phagocytosis.18 Ammonia stimulated microglia migration, which is also characteristic for the activated phenotype. On the other hand, microglial phagocytosis was significantly inhibited by ammonia, as shown in the present study. An impairment of phagocytosis has also been observed in neutrophils treated with ammonia in vitro or isolated from hyperammonemic patients with liver cirrhosis.28 Although the underlying mechanisms remained unclear, an ammonia-induced activation of the p38MAPK pathway was shown to mediate phagocytosis inhibition.28
In acute liver failure due to hepatic devascularisation, microglia activation is associated with an increased synthesis of proinflammatory cytokines,11, 29 which were suggested to contribute to the development of brain edema.30 These findings raise the possibility that microglia are a source for proinflammatory cytokines in hyperammonemia.11, 29 However, as shown in the present study, ammonia failed to increase IL-1α/β, IL-6, or TNF-α mRNA expression in cultured microglia and microglia activation in brains from acutely ammonia-intoxicated rats was not accompanied by increased cytokine mRNA levels.
This is in line with a recent report that found no release of proinflammatory cytokines in astrocyte or microglia cultures in response to NH4Cl treatment.31 Therefore, the reported increase of cerebral cytokine formation in acute liver failure11, 29 is probably not explained by direct ammonia effects on microglia. However, one has to keep in mind that mRNA levels need not necessarily reflect the behavior of cytokine protein expression. Whereas NH4Cl treatment up-regulated IL-1β mRNA level in astrocytes, a significant down-regulation was observed in microglia. Transcription of the IL-1β gene is controlled by nuclear factor κB, which becomes activated in ammonia-treated astrocytes.6 Therefore, one may speculate that nuclear factor κB is differently regulated by ammonia in astrocytes and microglia, respectively, with potential impact on IL-1β mRNA synthesis and/or stability.32
Activated microglia can produce high amounts of reactive nitrogen and oxygen species through activation of NADPH-oxidase and iNOS-derived nitric oxide, which may contribute to neuronal dysfunction in neurodegenerative diseases.15, 23 Increased iNOS protein and mRNA expression has been shown in ammonia-treated astrocytes in culture6, 20 as well as in chronic animal models of hepatic encephalopathy.21 In primary microglia cultures, ammonia up-regulated the synthesis of ROS in a time- and dose-dependent manner, which was sensitive to apocynine. These findings suggest that microglia participates in the generation of ammonia-induced oxidative stress through activation of NADPH-oxidases. However, microglial iNOS mRNA or protein expression remained unchanged after ammonia treatment. Also, synthesis of proinflammatory prostaglandin E2 was up-regulated in cultured astrocytes, but decreased in NH4Cl-treated microglia. This is in line with findings showing that prostanoid synthesis is differently regulated in astrocytes and microglia as exemplified by somatostatin treatment.33 In contrast to astrocytes, microglia are well known to express high levels of COX-2 protein constitutively,16 which may be reflected in our study by higher PGE2 concentrations at baseline. Given the pH-dependence of the enzyme, COX-2 activity in microglia may decrease in response to ammonia due to an alkalinization-induced inhibition.34 These results (summarized in Supporting Information Fig. 7) suggest that ammonia triggers a transition from a resting state into an early activation state of microglia, which may be characterized by an increased alertness, but does not reflect the fully reactive microglia phenotype.16
Neuroinflammation and HE in Patients with Cirrhosis.
Neuroinflammation, which was formerly termed reactive gliosis, has been defined as an acute or chronic activation of glial cells in response to brain injury.35 Microglia, which represent the innate immune cells of the central nervous system, are key players in neuroinflammatory processes. Their activation can be associated with increased synthesis or release of proinflammatory signaling molecules such as cytokines and chemokines. Additional factors that contribute to inflammation are ROS and prostanoids. With respect to this, iNOS-derived nitric oxide and COX-2–mediated PGE2 synthesis have been implicated in neuroinflammation in several neurodegenerative diseases.13-16, 19, 35-37
The results of the present study suggest that ammonia directly activates rat microglia as assessed by Iba-1 and isolectin-B412 expression, morphology, migration, and ROS formation, but has no effect on glutamate release, induction of iNOS and COX-2 and synthesis of prostaglandins, proinflammatory cytokines, and the chemokine MCP-1. These findings indicate that microglia were activated but not reactive. Microglia activation was also found in the cerebral cortex of acutely ammonia-challenged rats and post mortem brain tissue from patients with liver cirrhosis and HE. Interestingly, microglia activation as detected by increased Iba-1 expression was not observed in the cerebral cortex from patients with cirrhosis who do not have HE. This suggests that microglia activation is a feature of HE, but not of cirrhosis itself. Microglia activation in the cerebral cortex from patients who have cirrhosis with HE was not associated with increased mRNA and protein levels of inflammatory cytokines. Increased synthesis of MCP-1 is a consistent finding in many neuroinflammatory disorders (for a review, see Conductier et al.35). MCP-1 mRNA expression was not significantly changed in the cerebral cortex from patients who have cirrhosis with HE; however, chemokine mRNA levels need not parallel respective protein levels. If an increased synthesis of inflammatory cytokines is used to define neuroinflammation, as has been done by many investigators,13-16, 35-39 our study suggests that neuroinflammation is absent in the cerebral cortex of patients who have cirrhosis with HE, but the possibility is not excluded that neuroinflammation is present in other brain areas, as has been shown in animal models of chronic HE.10, 26, 27 However, our findings do not rule out a major contribution of neuroinflammation in acute liver failure, a condition in which the relationship between neuroinflammation and HE has mostly been studied. The present findings also do not rule out the possibility that systemic or cerebral infections can trigger and worsen HE episodes in patients with cirrhosis.
It is important to note that microglia activation not only mediates neuronal dysfunction, but can also confer neuroprotection, depending on the pathological stimulus.40 For example, impaired astrocytic glutamate uptake during neuroinflammation has been shown to be accompanied by de novo expression of glial glutamate transporters in activated microglia,41 and increased GLAST protein expression has been shown in cerbrocortical post mortem brain biopsies of HE patients.9 Therefore, microglia activation in patients who have cirrhosis with HE could also confer neuroprotection against glutamate toxicity.41 Further studies are required to clarify the role of microglia and their interaction with other cell types in the pathogenesis of hepatic encephalopathy.
Expert technical assistance was provided by Torsten Janssen, Brigida Ziegler, and Stefanie Winandy. We are grateful to the Australian Brain Donor Programs New South Wales Tissue Resource Centre, Sydney, for tissue support.