Neuroinflammation in Hepatic Encephalopathy: Mechanistic Aspects

Abstract

Hepatic encephalopathy (HE) is a major neurological complication of severe liver disease that presents in acute and chronic forms. While elevated brain ammonia level is known to be a major etiological factor in this disorder, recent studies have shown a significant role of neuroinflammation in the pathogenesis of both acute and chronic HE. This review summarizes the involvement of ammonia in the activation of microglia, as well as the means by which ammonia triggers inflammatory responses in these cells. Additionally, the role of ammonia in stimulating inflammatory events in brain endothelial cells (ECs), likely through the activation of the toll-like receptor-4 and the associated production of cytokines, as well as the stimulation of various inflammatory factors in ECs and in astrocytes, are discussed. This review also summarizes the inflammatory mechanisms by which activation of ECs and microglia impact on astrocytes leading to their dysfunction, ultimately contributing to astrocyte swelling/brain edema in acute HE. The role of microglial activation and its contribution to the progression of neurobehavioral abnormalities in chronic HE are also briefly presented. We posit that a better understanding of the inflammatory events associated with acute and chronic HE will uncover novel therapeutic targets useful in the treatment of patients afflicted with HE.

Hepatic encephalopathy (HE) is the major neurological complication of severe liver disease. It presents in two forms, chronic HE and acute HE. Chronic HE (portal-systemic encephalopathy) usually occurs in patients with alcoholic liver cirrhosis and is characterized by impaired neurological function, including changes in personality, altered mood, diminished intellectual capacity, and abnormal muscle tone and tremor.¹ Acute HE (AHE, acute liver failure, ALF; fulminant hepatic failure) generally occurs following massive liver necrosis due to viral hepatitis (hepatitis B and C), hepatic neoplasms, vascular causes, or exposure to acetaminophen and other hepatotoxins. AHE is associated with the abrupt onset of delirium, seizures, and coma.

The principal pathological change in chronic HE is characterized by Alzheimer type II astrocytosis,^2,3 which is characterized by astrocytes possessing enlarged, pale nuclei, often occurring in pairs, with the nuclei frequently displaying prominent nucleoli. Cerebral edema and associated increase in intracranial pressure leading to brain herniation are the characteristic features of AHE which occurs in upto 80% of patients with AHE^4–6 and represents the most frequent cause of death in these patients (70% mortality).^4,7 While the basis for the edema in AHE is poorly understood, astroglial swelling (cytotoxic edema) dominates the pathology in experimental animals,^8–11 as well as in humans.¹² Of interest, no significant or consistent morphologic changes have been identified in neurons or other neural cells. Such findings prompted the suggestion that HE fundamentally represents a primary “astrogliopathy”.¹³

While the precise involvement of astrocytes in HE is incompletely understood, it is known that the enzyme responsible for ammonia metabolism in brain, glutamine synthetase, is exclusively found in astrocytes.¹⁴ Such metabolism in the setting of elevated ammonia levels elicits a number of untoward events in astrocytes which exerts negative consequences on other neural cell. For review, see Norenberg¹³; Norenberg et al, 1992.¹⁵

While the molecular basis for the neurological disorder in acute and chronic liver failure remains incompletely understood, elevated blood and brain ammonia levels have been strongly implicated in its pathogenesis. Factors that lead to increased levels of blood or brain ammonia have been shown to worsen HE, whereas reducing blood ammonia levels alleviate HE.¹⁶ Additionally, clinical, pathological, and biochemical changes observed in HE are reproduced by increasing blood or brain ammonia in experimental animals,² while exposure of cultured astrocytes to ammonium salts reproduces the morphological and biochemical findings.^17–20

Recent studies have shown a significant role of inflammation in the mechanism of HE, in particular, the involvement of cytokines and lipopolysaccharide (LPS; endotoxin) in the pathogenesis of acute and chronic HE.^21,22 The focus of this review is to emphasize the involvement of ammonia in stimulating an inflammatory response in brain endothelial cells, microglia and astrocytes, in both acute and chronic HE.

Mechanisms of ammonia neurotoxicity

Impaired bioenergetics, electrophysiological defects, changes in intracellular pH, altered glutamateric and GABAergic neurotransmission, excitotoxicity, involvement of the peripheral benzodiazepine receptor,²³ oxidative/nitrative stress (ONS) and induction of the mitochondrial permeability transition, have all been identified as major mechanistic events triggered by ammonia.^24,25 Additionally, activation of intracellular signaling systems, including c-fos,²⁶ mitogen-activated protein kinases,²⁶ protein kinase G,²⁷ Src kinase family,²⁸ ciliary neurotrophic factor,²⁶ and the transcription factors p53,²⁹ SP-1²⁶ and nuclear factor-kappaB (NF-κB)^30–32 have all been shown to be involved in the mechanism of ammonia neurotoxicity, particularly in neuroinflammation.

Recently, the stimulation of ion transporters, including the Na⁺-K⁺-Cl⁻ cotransporter-1^33,34 and the nonselective cation (NCCa-ATP) channel,³⁵ as well as an increase in the level of the astrocytic plasma membrane protein aquaporin-4,³⁶ have all been shown to be involved in the mechanism of ammonia neurotoxicity, particularly in the development of the cytotoxic brain edema associated with acute HE. While the signaling systems that are stimulated in acute HE are well established, evidence for the involvement of signaling factors in chronic HE remains sparse.

Inflammation in hepatic encephalopathy

A growing body of evidence suggests that inflammation plays an important role in the development of HE. The concept that inflammation is involved in HE was first proposed by Gans et al, 1971,³⁷ based on findings that peripheral infections and inflammation were associated with fulminant hepatic failure (FHF). This aspect was further elaborated upon by Wilkinson et al, 1974; Liehr et al, 1976; Wyke et al, 1982.^38–40 Izumi et al 1995⁴¹ documented increased levels of IL-6 in plasma in both FHF and in chronic HE. Subsequently, Wigmore et al, 1998⁴² and Wright et al 2007⁴³ reported increased levels of serum tumor necrosis factor-alpha (TNF-α), interleukin-1β (IL-1β) and interleukin-6 (IL-6) and their arterio-venous differences, consistent with a brain cytokine efflux in patients with acute HE, supporting the presence of a central nervous system inflammatory component (neuroinflammation) in HE when associated with peripheral infections. Noteworthy, blood-derived cytokines, likely arising from necrotic liver and/or from sepsis, are well-known to freely cross blood–brain barrier under normal conditions.⁴⁴

• a.Role of inflammation in the brain edema associated with acute HE

As noted above, a major complication of acute HE is the development of cytotoxic brain edema. The involvement of neuroinflammation in acute HE was established by Chung et al, 2001,⁴⁵ who showed a protective effect of the anti-inflammatory agent, indomethacin (a potent inhibitor of cyclooxygenase-2), against the development of brain edema in rats with a portacaval anastomosis that were infused with ammonia⁴⁵; indomethacin also normalized the intracranial pressure in patients associated with acute HE.⁴⁶ Subsequently, Jiang et al, 2009⁴⁷ documented that a member of the tetracycline class of antibiotic, minocycline, also attenuated the encephalopathy grade and prevented brain edema formation in experimental ALF. It should noted, however, that minocycline itself induces liver damage,^48,49 which precludes its possible therapeutic use for the treatment of acute HE.

It was further shown that AHE results in a lesser degree of brain edema in transgenic mice deficient in TNF-α, IL-1β and IL-6 receptors, as compared to wild type mice,⁵⁰ and that etanercept, a TNF-α neutralizing molecule, prevented the onset of coma stages of HE as well as liver injury induced by AOM (Chastre et al, 2012).⁵¹ Induction of endotoxemia with LPS in rats was subsequently shown to aggravate the brain edema and encephalopathy associated with ALF.⁴³ Collectively, these findings suggest an important role of “central” inflammation in the development of the brain edema in ALF. For general reviews on neuroinflammation, see references.^21,52

Consistent with the role of inflammation in the brain edema associated with AHE, we previously reported that treatment of cultured astrocytes with inflammatory cytokines (TNF-α, IL-1β, IL-6 and IFN-γ) caused astrocyte swelling.⁵³ Moreover, the swelling caused by these cytokines was markedly potentiated when astrocytes were pre-treated with ammonia for 24 h, and then exposed to cytokines for an additional 24 h. Additionally, astrocyte cultures exposed to a combination of cytokines (TNF-α, IL-1β, IL-6 and IFN-γ) were recently shown to activate NF-κB and that such effect was potentiated by ammonia,⁵³ while blocking the activation of NF-κB prevented the astrocyte swelling.⁵³ Since cytokines are well-known inducers of ONS and activation of NF-κB,⁵⁴ it appears that in addition to ammonia, cytokines also contribute to the ONS and to the exacerbation of the astrocyte swelling acute HE.

Similarly, exposure of cultured astrocytes to a combination of ammonia and recombinant IL-1β resulted in significant additive increases in the expression of both hemoxygenase-1 (HO-1) and inducible nitric oxide synthase (iNOS).⁵⁵ These findings are consistent with the worsening effects of ammonia and inflammation on the CNS complications of ALF.⁵⁵ Together, these studies strongly suggest a synergistic interaction between ammonia, reactive oxygen/nitrogen species and inflammatory cytokines in the astrocyte swelling/brain edema associated AHE.⁵³

• b.Microglial activation and its contribution to the development of astrocyte swelling and brain edema in ALF

While the precise pathophysiological mechanisms responsible for the development of astrocyte swelling/brain edema in AHE have not been fully elucidated, it was recently proposed that microglial activation and the subsequent synthesis and release of pro-inflammatory cytokines are involved.^38,43,47,56 Microglia are well-known to induce inflammatory responses in brain,⁵⁷ as they represent the primary resident immune cells of the CNS. Microglia can be activated by a wide number of factors, including glutamate, pro-inflammatory cytokines (derived as a consequence of liver necrosis and/or sepsis), LPS, as well as by extracellular potassium.⁵⁷ Activated microglia, in turn, can release the pro-inflammatory cytokines IL-1α, IL-1β and TNF-α, all of which play a crucial role in neurodegeneration.⁵⁷

Upregulation of CD11b/c immunoreactivity, consistent with microglial activation, was observed in brains of rats with AHE at the time of coma.⁴⁷ Additionally, increased IL-1β, TNF-α, and IL-6 mRNAs were observed in brains of rats with AHE, and the magnitude of expression of these pro-inflammatory cytokines correlated well with the progression of the encephalopathy and the onset of brain edema.⁴⁷ Microglial activation was also identified in experimental models of hyperammonemia,⁵⁸ as well as in brain tissue obtained at autopsy from patients with ALF.⁵⁹

Increased OX-42 and OX-6 immunoreactivities in microglia (suggestive of their activation), was shown to be accompanied by increased expression of interleukins (IL-1β, IL-6) and TNF-α in the cortex of rats with AHE.⁴⁷ Minocycline treatment prevented both the microglial activation, as well as the upregulation of IL-1β, IL-6 and TNF-α, which was associated with a concomitant attenuation of the encephalopathy and brain edema. These findings are consistent with the involvement of microglial activation in the development of the brain edema in AHE.⁴⁷

We recently identified the synthesis and release of the cytokines IL-6 and TNF-α when primary cultures of microglia were exposed to ammonia (1 mM for 72 h or 5 mM for 24 h).⁶⁰ Additionally, exposure of microglia to ammonia resulted in ONS⁶⁰ and in the activation of the transcription factor NF-κB. Since inflammation is known to influence cell volume in other conditions,⁶¹ we recently examined the potential role of microglia in the mechanism of ammonia-induced astrocyte swelling. Accordingly, conditioned media (CM) derived from ammonia-treated (5 mM, NH₄Cl) cultured microglia, when added to cultured astrocytes, resulted in significant cell swelling (32%).⁶⁰ Such swelling was synergistically increased when astrocytes were additionally exposed to ammonia. We also found that CM from ammonia-treated microglia containing Tempol (a superoxide scavenger), or uric acid (a peroxynitrite scavenger), when added to astrocytes, resulted in a marked reduction in cell swelling as compared to the effect of CM from ammonia-only treated microglia.⁶⁰

In unpublished observations, we found the activation of the toll-like receptor-4 (TLR4; an inflammatory response factor), in ammonia-treated cultured microglia. Additionally, astrocytes exposed to CM from TLR4 gene-silenced microglia that were then treated with ammonia, showed a lesser degree of astrocyte swelling as compared to similarly treated controls, strongly suggesting a contribution of microglial TLR4 to the development of the cytotoxic brain edema associated with AHE.

While the above studies strongly suggest that activated microglia contribute to the cytotoxic edema in acute HE, Rangroo-Thrane et al⁶² reported that while microglia are indeed activated in the azoxymethane-induced animal model of ALF, such activation did not contribute to the cytotoxic brain edema. Instead, these authors noted that the azoxymethane-induced vasogenic edema was associated with a breakdown of the blood–brain barrier (BBB), as well as the activation of microglia. It should be noted, however, that a generalized breakdown of the BBB (as occurs in vasogenic edema) is not a feature of ALF in humans. Further, the liver toxin used in this study, azoxymethane, was recently shown to induce a breakdown of the BBB in vivo.⁶³ Subsequently, a direct toxic effect on brain endothelial cells by azoxymethane was observed in an in vitro model of the BBB, resulting in a severe breakdown of the BBB.⁶⁴ These findings suggest that the use of azoxymethane-induced hepatotoxicity as an experimental model of ALF is not suitable for the study of acute HE.

Rangroo-Thrane et al, 2012⁶² further commented that microglial activation did not occur in ammonia infused mice, despite the mice developing pathologically increased plasma ammonia levels and severe neurological dysfunction. While considerable evidence documents that microglial activation contributes to the pathogenesis of AHE (as noted above), the reason for the absence of microglial activation in their study is not clear. However, these investigators used a relatively high concentration of ammonia (7.5 mmol/kg NH₄Cl, i.p.) which is not pathophysiologically relevant as this concentration of ammonia induces acute deterioration of neurological function, development of myoclonic seizures and death within 5–10 min after ammonia injection,^65–67 thereby precluding the subsequent identification of microglial activation. The basis for such deterioration following treatment with high levels of ammonia is not known, but it suggests the possibility of some generalized negative systemic event(s) (e.g., in heart or lung) occurring as a consequence of the hyperammonemia.

Microglial activation was also reported in animal models of chronic HE due to bile duct ligation (BDL)^68,69 or to portacaval shunts.^70,71 However, a recent report did not observe microglial activation in the BDL model of chronic HE.⁷² While these studies^68,72 employed an identical animal model of chronic HE, identical time point of examination (4 weeks), as well as used the same microglial marker (Iba1), the reason for such discrepancies is unclear.

Neurobehavioral abnormalities are the major clinical consequences of chronic HE.⁵⁶ While significant evidence indicates the involvement of inflammation in the astrocyte swelling/brain edema in acute HE, relatively few studies have examined the potential role of inflammation in the neurobehavioral defects associated with chronic HE. Recently, however, Rodrigo et al, 2010⁵⁶ reported that chronic/moderate hyperammonemia or BDL resulted in microglial activation, increased inducible nitric oxide synthase activity, as well as increased brain levels of interleukin-1β, and prostaglandin E2, and that these changes were associated with impaired cognitive and motor functions. These authors further demonstrated that ibuprofen, a non-steroidal anti-inflammatory agent, reduced the microglial activation and restored cognitive and motor functions in hyperammonemic and in rats with BDL. Additionally, rats that underwent portal vein ligation were shown to exhibit impaired locomotor activity, as well as increased inflammatory events, including the presence of protein tyrosine nitration, RNA oxidation, and IL-6 mRNA increase in brain.⁷³ These authors also reported that prevention of protein tyrosine nitration and RNA oxidation with indomethacin, a non-specific cyclooxygenase-2 inhibitor, prevented brain protein tyrosine nitration, RNA oxidation, as well as disturbances in locomotor activity associated with chronic HE. Taken together, these reports suggest that inflammation also contributes to the neurobehavioral deficits observed in chonic HE.

Astrocytes in the inflammatory response in hepatic encephalopathy

As noted above, swelling of astrocytes (cytotoxic brain edema) is the most prominent neuropathological abnormality in AHE. Additionally, a direct effect of ammonia on the activation of inflammatory factors in cultured astrocytes has been identified. Activation of NF-κB, and the tumor suppressor gene p53 were shown in ammonia-treated cultured astrocytes.^29–31 Additional studies in support of an astrocytic inflammatory response in vitro, is derived from recent reports demonstrating that inflammatory factors such as constitutive nitric oxide synthase (cNOS) and NADPH oxidase (NOX) and the consequent generation of reactive oxygen and nitrogen species (RONS), phospholipase-2 (PLA2) and mitogen-activated protein kinases (MAPKs), were all activated in cultured astrocytes by ammonia.^74,75 Moreover, inhibition of these factors was shown to reduce the ammonia-induced astrocyte swelling in culture.^74,75 Altogether, these findings strongly suggest that ammonia is capable of inducing an inflammatory response in astrocytes, independent of any influence from other neural cells or from any extracellular stimuli, and that such effects may contribute to the cytotoxic brain edema in AHE.

Cerebral endothelial cells (ECs)

ECs are the first resident brain cells exposed to blood-borne ‘‘noxious agents’’ (i.e., ammonia, CKs, LPS). Since ECs are also capable of activating a number of inflammatory factors, including inducible nitric oxide synthase (iNOS), NADPH oxidase (NOX), phospholipase A2 (PLA2), cyclooxygenase-2 (COX2), as well as NF-κB,⁷⁶ it is possible that these inflammatory factors may generate “cell swelling mediators”, including arachidonic acid, RONS, prostaglandins and CKs,⁷⁶ that may ultimately result in astrocyte swelling/brain edema in acute HE. In support of this view, we recently reported that cultured ECs treated with ammonia showed evidence of ONS^77–80 and the activation of NF-κB.^77,79

As noted above, we also found increased TLR4 protein expression and its associated signaling cascades, the myeloid differentiation primary response gene 88 (MyD88), and the TIR-domain-containing adapter-inducing interferon-β (TRIF) protein expression in ammonia-treated cultured rat brain ECs.⁷⁹ Additionally, exposure of cultured ECs to ammonia stimulated the release of cytokines and free radicals, all of which are known to enhance inflammatory responses in other conditions. Further, receptors for CKs, including IL-1β and TNF-α, have been identified in brain ECs in experimental models of sepsis,⁸¹ and that activation of these receptors further stimulated the synthesis and release of additional CKs leading to an exacerbation of the brain edema in acute HE.⁸²

To examine whether ammonia, CKs and LPS stimulate ECs to release inflammatory mediators that may contribute to the astrocyte swelling and brain edema in AHE, primary cultures of rat brain endothelial cells were treated with ammonia, CKs and LPS and the conditioned medium derived from these treatments were added to cultured astrocytes and cell volume was determined. We found that CM from ammonia-treated ECs when added to astrocytes caused significant cell swelling, and such swelling was potentiated when astrocytes were exposed to CM from ECs treated with a combination of ammonia, LPS and CKs.^77,79 We also found an additive effect when astrocytes were exposed to ammonia, along with CM from ammonia-treated ECs. CM derived from ECs treated with ammonia, along with antioxidants or the NF-κB inhibitor BAY 11-7082, when added to astrocytes, resulted in a significant reduction in cell swelling.^77,79 Additionally, astrocytes exposed to CM from TLR4 gene-silenced ECs that were treated with ammonia, cytokines or LPS, alone or in combination, showed less cell swelling as compared to similarly treated controls.^77,79 Altogether, these findings suggest that ECs are indeed capable of activating a plethora of inflammatory factors that play a critical role in the development of the brain edema associated with AHE.

We additionally found increased TLR4 protein expression in brain ECs after treatment of rats with the liver toxin thioacetamide (TAA), and that the administration of TAA to transgenic mice lacking TLR4 exhibited a lesser amount of brain edema, as compared to wild type mice.⁷⁹ These studies strongly suggest that ECs significantly contribute to the astrocyte swelling/brain edema in acute HE, in part, as a consequence of increased TLR4 protein expression triggered by blood-borne noxious agents (e.g., LPS, cytokines). While a major role of brain endothelial cells has been shown in the pathogenesis of AHE, whether endothelial cells also play a significant role in the pathogenesis of chronic HE, remains to be established.

Role of nuclear factor-kappaB in vivo

As noted above, the activation of NF-κB represents a major inflammatory response in brain. We recently examined whether transgenic (Tg) mice that have a functional inactivation of astrocytic NF-κB (NF-κB null mice) are resistant to ALF-associated brain edema. We found that wild type (WT) mice when treated with TAA showed a significant increase in brain water content, along with prominent astrocyte swelling and associated increase in iNOS immunoreactivity.³² By contrast, NF-κB null mice treated with TAA did not exhibit brain edema, histological changes, nor displayed an increase in iNOS immunoreactivity.³² Additionally, astrocyte cultures derived from NF-κB null mice showed no cell swelling nor morphological abnormalities when exposed to ammonia. By contrast, ammonia significantly increased cell swelling in cultured astrocytes from WT mice and displayed cytological abnormalities.³² In aggregate, these findings indicate that the activation of NF-κB represents an important inflammatory factor in the development of astrocyte swelling/brain edema in the setting of AHE.

Summary

Inflammation, in conjunction with ammonia, plays an important role in the pathogenesis of HE. The application of ammonia to cultured astrocytes stimulates a number of inflammatory signaling systems (NF-κB, PLA2, NOS, NOX, COX2) that leads to astrocyte swelling/brain edema in acute hepatic encephalopathy. Additionally, ammonia activates microglia and stimulates brain endothelial cells; some of these events also occur in experimental animals and in humans with HE. The consequences of this inflammatory response (synthesis and release of inflammatory factors from activated microglia and endothelial cells), along with hyperammonemia, leads to severe astrocyte swelling/brain edema in acute hepatic encephalopathy. The presence of both central and peripheral inflammation has also been shown to contribute to the neurobehavioral abnormalities observed in chronic HE. A schematic diagram illustrating mechanisms by which inflammation contributes to the astrocyte dysfunction in HE is presented in Figure 1. Studies focusing on the identification of molecular mechanisms and cell–cell interactions during inflammation and hyperammonia may provide useful therapeutic targets for the treatment of both acute and chronic hepatic encephalopathy.

Figure 1.

Mechanisms of neuroinflammation in hepatic encephalopathy. Inflammatory cytokines and lipopolysaccahride (LPS, endotoxin) activate endothelial cells and microglia that subsequently leads to the stimulation of the Toll-like receptor-4 (TLR4). TLR4 then activates down-stream factors, including NF-κB and myeloid protein88 (MyD88)-dependent pathways, which subsequently leads to the production of inflammatory mediators, including reactive oxygen and nitrogen species (RONS), arachidonic acid, and prostaglandins. These inflammatory mediators ultimately impact on astrocytes, leading to their dysfunction resulting in brain edema in acute HE, and likely contribute to the neurobehavioral deficits in chronic HE.

Conflicts of interest

All authors have none to declare.