Overview of oxidative stress findings in hepatic encephalopathy: From cellular and ammonium-based animal models to human data

Oxidative stress is a natural phenomenon in the body. Under physiological conditions intracellular reactive oxygen species (ROS) are normal components of signal transduction cascades, and their levels are maintained by a complex antioxidants systems participating in the in-vivo redox homeostasis. Increased oxidative stress is present in several chronic diseases and interferes with phagocytic and nervous cell functions, causing an upregulation of cytokines and inflammation. Hepatic encephalopathy (HE) occurs in both acute liver failure (ALF) and chronic liver disease. Increased blood and brain ammonium has been considered as an important factor in pathogenesis of HE and has been associated with inflammation, neurotoxicity, and oxidative stress. The relationship between ROS and the pathophysiology of HE is still poorly understood. Therefore, sensing ROS production for a better understanding of the relationship between oxidative stress and functional outcome in HE pathophysiology is critical for determining the disease mechanisms, as well as to improve the management of patients. This review is emphasizing the important role of oxidative stress in HE development and documents the changes occurring as a consequence of oxidative stress augmentation based on cellular and ammonium-based animal models to human data.


Introduction
Oxidative stress (OS) is believed to play a role in the pathogenesis of hepatic encephalopathy (HE). Under normal physiological conditions reactive oxygen and nitrogen species (ROS and RNS) ( Table 1) are balanced by antioxidants scavenging systems, while a permanent increase in ROS and RNS levels results in OS being harmful on cell-and tissue-homeostasis [1][2][3]. These antioxidants are low molecular weight reductants such as glutathione (GSH) and ascorbate (Asc) which are accompanied by protein antioxidants such as superoxide dismutase (SOD) that reacts with superoxide anion (O ⋅− 2 ), catalase (CAT) and the peroxiredoxins (glutathione peroxidase (GPX)) that catabolize hydrogen peroxide (H 2 O 2 ) [4][5][6]. There is growing evidence that the modulation of antioxidant levels in cells is inextricably linked to intracellular OS levels [2,6].
ROS are small messenger molecules that are normal components of signal transduction cascades during physiological processes, but when in excess are also involved in neurotoxicity and neurodegeneration [7][8][9].
The redox homeostasis is maintained by the balance between ROS generation and elimination by antioxidants (Fig. 1).
Ammonium (NH + 4 ) is produced by metabolism of amino acid obtained from dietary proteins [10]. Liver plays an important role in ammonium detoxification and in oxidative stress regulation. Therefore, when liver fails, blood ammonium levels increase and its impaired clearance by the diseased liver leads to brain glutamine (Gln) accumulation, causing disturbance of central nervous system (CNS) functions [11][12][13]. In parallel, a significant decrease in two major liver antioxidant enzymes activity, CAT and SOD, was detected in cirrhotic liver of HE patients [14].
Albumin, which is synthesized by liver hepatocytes and rapidly excreted into the bloodstream, is a very abundant and important circulating antioxidant, accounting for more than 70% of serum's free radical-trapping activity [15,16]. Albumin has been found to be decreased in cirrhotic patients and has been associated with the development of overt HE as well as risk factors associated with death [17][18][19][20][21]. Epidemiological studies have found that decreased albumin level together with irreversibly oxidized albumin (HNA2) are an independent predictors of mortality [22][23][24][25][26]. The increase of mortality rate of HE patients with hypoalbuminemia has been previously shown and was accounted for 6.7-10.9% for short period of hospitalization and about 42.8-73.9% for the long-term follow up [17][18][19][20][21]. Recent studies demonstrated that long-term human albumin infusion improves the prognosis of cirrhotic patients by lowering the overall mortality and the likelihood of emergent hospitalizations [25,26].
When healthy the environmental/physiological factors, i.e., ROS, trigger immune mechanisms, which positively regulate neuroplasticity and neurogenesis, promoting learning, memory, and cognition [27]. ROS facilitate the response to growth factor activation and the formation of the inflammatory process, as well as playing critical functions in the immune system and microorganisms eradication [28]. Polymorphonuclear neutrophils (PMN) are the most abundant circulating immune cells that participate in immune and inflammatory processes (host defense) by releasing a significant amounts of pro-inflammatory cytokines and ROS/oxidative burst. When in excess, these cause endothelial dysfunction and increase the leakiness of the blood vessels/blood-brain barrier (BBB) [29][30][31].
Clinical studies have revealed that systemic inflammation is linked to neuroinflammation [32] and leads to cognitive deterioration in HE patients [33][34][35]. TNF-, IL-18, IL-6, endotoxins (i.e., NH + 4 ) levels have been shown to be significantly greater in the plasma of cirrhotic patients with HE [36]. In particular high levels of IL-6, which have been found in brains of BDL rat model of type C HE [2], are associated with ageing [27], the injury response of the CNS [37], impairment of the BBB integrity [38], memory and cognitive functions decline [27,39].
Systemic and brain cytokines, in combination with endotoxins (i.e., NH + 4 ), will set up an inflammatory cascade that exacerbates OS [40][41][42][43] and OS related activation of the astrocytes, and microglia [13,44]. Therefore, it is becoming increasingly evident that OS, both systemic and CNS is an important feature in the pathogenesis of HE [2,42,45]. In addition, impaired brain NH + 4 detoxification together with osmotic effect of Gln induces ROS generation [2,7,8,46]. It is worth noting that OS and inflammation are pathophysiological processes that are inextricably linked and interdependent [47]. It is widely accepted that in the presence of OS, inflammatory processes will develop, accelerating the progression of the disease. Similarly, if inflammation is the trigger, an OS response will be activated, contributing to the inflammatory response [11,[48][49][50][51][52][53]. Ammonium, hyponatremia, and inflammatory cytokines have been shown to trigger a self-amplifying cycle between astrocyte osmotic stress and cerebral oxidative/nitrosative stress [41].

Oxidative stress
ROS and RNS are common byproducts of normal aerobic cellular metabolism and serve crucial physiological functions in intracellular cell signaling, homeostasis, cell death, immunological response to infections, and mitogenic response induction [3,[54][55][56][57][58]. All metabolic processes involve the oxidant-antioxidant equilibrium required to perform routine molecular and biochemical functions [29,59]. The outcome of oxidative stress depends on the number of affected cells and functioning of the antioxidants systems (enzymatic and nonenzymatic antioxidants), i.e., concentrations of GSH, CAT, SOD, GPX, vit. C and vit. E ( Fig. 1) [5,60].
Mitochondria respiratory chain form most of the ROS produced in the body. The superoxide anion (O ⋅ − 2 ) is one of the most important free radical in biological systems involved in cell signaling. O ⋅ − 2 is generated at the complexes I and III (Fig. 2), acts as a precursor for the synthesis of many other ROS (i.e., H 2 O 2 and ONOO − ), and lipid peroxides [61,62]. Approximately 80% of O ⋅ − 2 is released into the intermembrane space, with most of remainder going to the mitochondrial matrix [63]. In addition, ROS (ie. O ⋅ − 2 ) can be generated while endo/exogenous toxin detoxification by microsomal cytrochrome-p 450 conciliated hydroxylation. When in pathophysiological states the excess production and accumulation of ROS cause signaling pathways deterioration, overwhelm the antioxidant defense mechanisms capacity, cause the redox homeostasis imbalance, form OS state and became toxic [51,64]. Therefore, ROS can cause nonspecific damage, and participate in degeneration of essential cellular components like lipids, proteins, and DNA, potentially leading to cell senescence and death [51,64].
Elevated OS caused by increased levels of O ⋅ − 2 induce oxidative damage to proteins, lipids, and nucleic acids, compromising cell health, and has been linked to ageing processes and various human pathogeneses such as the development of cancer, neurodegenerative, and cardiovascular diseases [65][66][67]. ROS are short-lived [68], they react rapidly with first line antioxidants, like GSH, Asc, and SOD [4,5,69,70]. GSH is a major antioxidant in the brain and its decline decrease the ability of CNS cells to counteract the OS and is a common sign in patients with neurodegenerative disorders [71]. Astrocytes synthesize more GSH than neurons and serve as a source of precursors for neuronal GSH synthesis [11,71].
ROS are critical for hippocampal long-term potentiation (LTP), a synaptic plasticity for learning and memory as well as in aging/diseaserelated impairment. Therefore, when in excess ROS leads to cellular dysfunction, and long-term depression (LTD) [7,8]. In the CNS the oxidative conditions are essential and play a key role in nerve growth factor (NGF) induced cell differentiation. Despite being the longest-living cell type, CNS cells are more vulnerable to OS-mediated injury because of their physiological and biochemical properties, high energy requirements, and unique redox activities: (i) Neurons generate the highest rate of ROS and utilize ~20% of the oxygen consumed by the bodythe majority of which is used for ATP production (4 × 10 12 molecules/min) to maintain neuronal intracellular ion homeostasis. (ii) Majority of the neuronal cells are nonreplicating therefore more sensitive to OS.
(iii) Neuronal membranes rich in polyunsaturated fatty acids (PUFA) are particularly vulnerable to OS, ie. oxidative damage to myelin. (iv) Modification of ion channels activity, disturbance in Ca 2+ traffic across neuronal membranes and its intracellular concentration increase often leads to OS. the presence of intracellular iron, copper, or manganese ions, which accumulates in brain as a function of age, favors the Haber-Weiss vicious circle/Fenton chemistry, resulting in the generation of OH ⋅ , which is among the most active ROS (Fig. 1). (vii) Antioxidant defense -brain contain low levels of Asc, GSH, CAT, GPX, and vitamin E [29,51,[71][72][73].
Oxidative stress presence may be tested in three ways: (i) direct detection of ROS (in-vivo, in-vitro and ex-vivo living tissue); Direct detection seem to be the preferred method but depends on local antioxidants concentrations and clearance mechanisms [5,91]. Of note in-vivo steady-state concentrations of ROS range from pico-to nanomolar range [92], with the lifetimes span nanoseconds to seconds (Table 1). (ii) detection of resulting damage to biomolecules (in-vitro and exvivo); Due to some of the challenges encountered by the direct detection some scientists prefer to use techniques based on the detection of final oxidation products and measure the damage on proteins, DNA, RNA, lipids, and other biomolecules. (iii) detection of antioxidants concentrations, total antioxidants capacity, antioxidants activity (in-vitro and ex-vivo). This approach measures the activity of specific antioxidant enzymes, like CAT, Fig. 1. Balance between ROS generation and elimination by antioxidants.   [2]. B) Fluorescent microscopy -immunohistochemistry (IHC) of BDL rats brain tissue: Oxo-8-dG -DNA/RNA damage antibody accumulation-sign of elevated HO ⋅ and its interaction with a nucleobase. GPX antibody build-up -sign of an increased production of H 2 O 2 [2]. C) 1 H-MRSin-vivo detection of two main antioxidants GSH and Ascalteration of their concentrations represents an indirect evidence OS [13]. SOD or GPX, and total antioxidant capacity. Furthermore, each individual marker reflects only partially the oxidative status, and therefore an integrative approach is necessary to achieve comprehensive conclusions.
In-vivo and ex-vivo OS detection is a complex task and require probes that are very sensitive, highly selective (i.e., HO ⋅ , O ⋅ − 2 , NO ⋅ , ONOO − ), and fast to react rapidly with ROS/RNS and create a stable secondary radical, which then can be quantified. Most of experimental assays for OS detection provide relative data. The extensively used approaches are electron paramagnetic resonance (EPR) spectroscopy allowing a real time direct detection of ROS, while fluorescence spectroscopy, microscopy, or flow cytometry can detect the final oxidation products or enzymatic activity [50,61,77]. In-vivo antioxidants (GSH and Asc) concentration measurements provide indirect evidence of redox homeostasis imbalance/OS and can be assessed non-invasively using proton magnetic resonance spectroscopy ( 1 H-MRS) [13,93,94].

EPR
Is a method for direct detection of unpaired electrons. Free radicals are chemical molecules that have unpaired electron and are primarily formed from molecular oxygen. EPR in combination with nontoxic cellpermeable and resistant to antioxidants spin-traps (covalent bond with the radical by addition reaction) or spin probes (oxidized by ROS without binding) have been recognized as one of the most powerful and exclusive technique that allow direct and reliable detection of the ROS presence in the system under study being less uncertain as compared to other methods (i.e.: immunoassays or UV-Vis spectroscopy) [50,61,75] (Fig. 3A). EPR of sable secondary radicals (paramagnetic/EPR active) formed by adding exogenous spin-traps/probes (diamagnetic/EPR non-active) provides direct information about a variety of biological samples, i.e., living tissue, blood, and other body fluids (in-vitro, ex-vivo in room temperature and frozen) redox state in an accurate, rapid and quantitative manner by the generation of stable nitroxide radicals [50,65,74,[95][96][97].

Spectrophotometric assays
Involving ultraviolet and colorimetric assays, are based on the interaction of reactive species with redox compounds and change in absorbance [74,99]. Spectrophotometry has found a widespread use in biomedical research, but, like chemiluminescence probes, it does not provide details about ROS/RNS and its localization within the cell (i.e., extracellular or intracellular) unless various probes and/or inhibitors with differing compartmentalization are used [99].
Remarks: Among most extensively applied spectrophotometric assays are: (i) GSH/GSSG as the one of most powerful self-generated antioxidant in the body [100,101] and the (ii) malonaldehyde (MDA) as a marker of lipid peroxidation [102,103].
(i) The major issue with blood ex-vivo GSH and GSSG concentration measurements is the oxidation of GSH after sample collection, and exaggeration of oxidation by increased formation of ROS while the acid deproteinization process. Therefore, leading to an overestimation of GSSG and an underestimating of GSH concentrations and the GSH/GSSG ratio [81]. To address this issue, the N-Ethylmaleimide (NEM) blocking agent, which forms stable, covalent thioether bonds with sulfhydryls, and prevent the formation of disulfide bond, needs to be added immediately to the blood sample after collection [81,100,101]. (ii) MDA is the most commonly used biomarker of OS, which results from lipid peroxidation of polyunsaturated fatty acids [103]. The most often used method for determining MDA in biological fluids is the thiobarbitoric acid reactive substances (TBARS) test [68,104]. TBARS test suffer from numbers of limitations. MDA may be produced by other than lipid peroxidation reactions [105] and the false increase could be generated by the heating step of the assay [68]. EDTA treated samples as well as samples stored at − 20 • C without addition of antioxidants significantly increase the TBARS levels [102,103,106]. Therefore, the main concerns are: non-specificity of TBA reactivity on MDA (other aldehydes cross-reaction produced from lipid peroxidation), poor reproducibility of analytical results, sample preparation/procedural modifications, storage, and stability of MDA standard solutions [102,103]. Therefore, an expert panel should re-evaluate MDA as an OS biomarker and a validated analytical process should be created [102,103].

The fluorescent redox probes
Are based on sensing mechanisms, and are classified as reactionbased selective probes, and reversible probes which can respond to multiple oxidation-reduction cycles [77]. The advanced fluorescence detection gives the possibility for real-time measurements of large number of samples (plate reader) with high sensitivity, spatial resolution, and specificity [77]. The pattern of oxidation/reduction-induced fluorescence change is crucial in determining the biological significance of a probe. Therefore, it is important to evaluate the probe mechanism: signal enhancement during reduction (nitroxide-based probes -probing of hypoxia or antioxidants effectiveness) or oxidation (most other classes of probes -ROS/RNS activities) [77]. The fluorescence microscopy enables probe localization (cytology/histologysub-cellular organelles, proteins, and membrane-components) (Fig. 3B).
However, for measuring in-vitro/ex-vivo living tissue, the intracellular fluorescent probes concentration is critical for correct interpretation of the results. The fluorescent product formation will depend on experimental conditions and probe internalization [107,108], the formation of multiple nonspecific oxidation products, light sensitivity, and probe redox cycling [78]. Therefore, it is critical to recognize these limitations in order to avoid erroneous interpretations.
As shown in Fig. 3C the two main antioxidants GSH and Asc can also be measured by 1 H-MRS in the brain. Of note Asc and GSH are present in low concentration with significant resonance overlap, thus special editing sequences or ultra-high magnetic fields are required [13,70,76,94,[109][110][111]. Changes in these metabolites represent an indirect evidence of OS. Therefore, additional techniques that allows the precise quantification of tissue ROS and identification of different ROS species would contribute significantly to the understanding of the pathophysiology of type C HE and potentially offer insight into therapeutic options.
As shown above, there are several methods and assays for redox biomarkers detection, making it difficult to select the relevant and most appropriate one. Pros and pitfalls of abovementioned methods are provided in Table 2.
Administration of high concentrations of ammonium 5-12 mmol/kg (intraperitoneally) or 60 mM (intrastriatally) in rats caused acute intoxication and increased generation of free radicals [120,121].
Despite the significant increase in cellular GSH levels, but not mitochondrial, both in cultured astrocytes as well as in microdialyzed cerebral cortex, ROS production was not blocked, suggesting that the increased ROS by ammonium reached levels that were beyond the antioxidant capacity [115,120,122]. Therefore, ROS may induce DNA damage mediated by activation of NMDA receptors [123].
Microglial ROS production increased with ammonium time exposure and in a dose-dependent manner (0.25-5 mM). Treatment with apocynin completely abolished the ammonium induced ROS increase, indicating activation of NADPH-oxidase by ammonium, and therefore suggesting that the OS increase is due to ROS not the reactive nitrogen species (RNS) [112,118].

Oxidative stress in type B and C HE animal models
Over the last several years researchers have demonstrated that oxidative stress is a key factor implicated in the pathogenesis of type B and C HE (Table 4) in animal. Indirect detection of OS using biochemical markers on BDL rats brain homogenates [59,127], as well as direct in-vivo brain and ex-vivo living brain tissue [2,13] have shown the presence of a strong correlation between chronic liver disease induced HE and OS.
The BDL rat model of type C HE has been shown to support the important role of oxidative stress in pathogenesis of HE by increased CNS OS (O ⋅ − 2 ) already at week 2 post-surgery [2], and at week 4 post surgery both, CNS and systemic OS [2,13,127]. GSH and ascorbate (Asc) have been shown to be the first line antioxidants which reacts rapidly with ROS to counteract the deleterious effects of OS [4,5]. GSH and Asc decline in brains of BDL rats correlated with OS increase suggesting an increased ROS scavenging in response to increased production [2,13,127]. Elevated levels of SODs have also been observed in response to increased endogenous and exogenous O ⋅ − 2 in the same animal model at 4 weeks post-surgery [2]. The nuclear translocation of SOD1 for the genomic DNA protection together with SOD2 and GPX-1 upregulation demonstrates a direct evidence of increased need to eradicate ROS [2].

EPR
-direct detection of free radicals -in-vitro and ex-vivo realtime fresh or preincubated frozen sample measurements -high quality spin-traps/ probes that react with short-lived radicals and convert them to longlived radicals -cell permeable and nonpermeable spin probing (cyclic hydroxylamines) -intra vs. extracellular OS -react rapidly with ROS/ RNS and create a stable radical -direct detection of ROS -single chemical reaction -no redox cycling = no artificial ROS -very sensitive, highly selective -(i.e., HO ⋅ , nucleic acids rather than nuclear DNA with HO ⋅ [2]. It has been shown that several brain regions of BDL rats had impaired enzymes activity of mitochondrial respiratory chain together with increase in mitochondrial reactive ROS generation, and augmentation of lipids and proteins oxidation [127]. However, the PCA model of type B HE has demonstrated conflicting results. At 4 weeks post-surgery no significant differences in plasma ROS, xanthine oxidase (XO), GSH and GSH/GSSG ratio, and CAT levels or in CSF ROS levels have been found between PCA and PCA-sham rats [129,130]. In contrary, in brains of PCA rats at 6 weeks post-surgery ammonium induced an increased brain ROS generation and promoted carbonylation of proteins, together with increased lipid peroxidation [131]. These different results could be related to the use of different animal strains and the length of the experimental design.
In parallel it has been shown that systemic OS correlated significantly with blood NH + 4 concentrations [2] in BDL rats. Increase of ROS was represented by increased production of O ⋅ − 2 , H 2 O 2 , increased activity of XO and decreased activity of CAT [2,129,130]. Significantly larger amount of O ⋅ − 2 production in BDL PMN and LYM as compared to the healthy sham controls [2] indicated high resting oxidative burst and hypo-responsivity to the bacterial challenge [2,132].
There are few studies on CNS OS in patients with chronic liver disease and chronic liver disease associated HE ( Table 5). Evidence of CNS OS presence in HE arose first from the postmortem EM observation of lenticular nuclei [142]. Alzheimer type II astrocytes, the integral neuropathological aspect of HE [143], contain large amounts of lipofuscin pigment aggregates [142] produced by peroxidation of unsaturated fatty acids [144,145]. Further, the IHC of the cerebral cortex of cirrhotic patients with HE supported the presence of oxidative stress in pathogenesis of HE [112]. As such increased expression of neuronal nitric oxide synthase (nNOS) has been accompanied by increased expression of SOD [112], a direct indication of OS presence. The induced RNA oxidation has been detected by increased accumulation of Oxo-8-dG, a direct sign of HO ⋅ [84,112,128]. Furthermore, recent studies supported the link between the OS related lipid peroxidation [141,142] and brain atrophy/the grey matter volume (GMV) loss [141].
Liver biopsies of NAFLD patients indicated increased OS accompanied by decreased antioxidant capacity (SOD, GSH and CAT) [14,112].
A significant disorder in redox status was also detected in HE patients' blood, indicating systemic OS involvement in the disease progression, evidenced by significant enhancement of plasma/serum lipid peroxidation, protein carbonylation together with reduction in antioxidant capacity, i.e., significant decrease of albumin, SOD, CAT and GPX enzymatic activity [14,112,[134][135][136][137]139,140].
Similarly, to the animal model studies the isolated PMN cells of cirrhotic patients demonstrated significantly higher resting state oxidative burst along with reduced phagocytic capacity [132,136], which was correlated with higher risk of multiple infections and organ failure [136].

Oxidative stress consequences
It is now well accepted that OS is implicated in various neurodegenerative disorders, causes neurofilaments (NfL) phosphorylation and leads to proteins aggregates formation [147,148]. To date little is known about OS involvement in neurodegeneration in patients with HE.
Excessive lipofuscin pigment aggregates, a sign of lipids peroxidation, have been detected in the Alzheimer type II astrocytes of the lenticular nuclei of HE patients [142]. OS accelerate the lipofuscin accumulation [144,145,149,150]. Lipofuscin is composed primarily of protein, lipid peroxides, and transition metals [150]. Lipofuscin aggregates at high concentration undergo Fenton reaction, resulting in the generation of OH ⋅ , one of the most active ROS [144], which in consequence may generate additional lipid, protein and RNA/DNA oxidation [141,142,[151][152][153] and therefore lead to brain atrophy [141].
Astrocytes protect neurons against the ROS toxicity, by supplying them with the GSH precursors [71,154,155]. Therefore, it has been shown that inhibition of cystine uptake into astrocytes under chronic ammonium exposure [156] may lead to reduced astrocyte and neuronal GSH levels [14,59,112,127], and place both at risk for oxidative damage.
The loss of glial filaments and dendritic spines of pyramidal neurons, hippocampal and cortical as shown in the BDL rat models [13,96,157], may be related to aforementioned oxidative modification of RNA and may thereby provide another link between OS and ammonium toxicity [2,45,[112][113][114]118,158]. RNA oxidation interferes with translational machinery, gene expression, thereby providing the link between ammonium-induced OS and cognitive decline through impaired protein synthesis [158] and neurotransmission [67].
Filaments loss can be also induced by OS related protein depolymerization [159]. Moreover, ammonium intoxication has been shown to lead to protein nitrotyrosylation and nitric oxide (NO ⋅ ) overproduction in in-vitro and human brain studies [45,112,114,116,118,123]. NO ⋅ out-compete with SOD and react with O ⋅ − 2 , leading to powerful and toxic peroxinitrite (ONOO − ) formation, which may lead to DNA damage [123], NfL and actin nitration and disrupt filaments assembly [160]. Similarly, the actin cytoskeleton is a primary target of OS in hepatocytes [161]. Therefore, OS-induced cytoskeleton morphological alterations in hepatocytes will have a negative impact on hepatobiliary functioning.
Postmortem IHC of cortex of NAFLD patients have demonstrated an increase expression of a small heat shock protein-27 (Hsp27) [112], a direct sign of OS presence. During oxidative stress, Hsp27 has an important antioxidant role lowering the ROS, raising GSH, lowering the intracellular iron concentrations, actin cytoskeleton remodeling, and is capable of inhibiting OS-induced necrosis [162][163][164]. Because the cytoskeleton is a network of filaments linked together by crosslinking proteins, damage to one component affects the entire cytoskeletal network.   cell protein in medium, the increase is facilitated by activation of the uptake of the GSH precursor cystine -GSH ↑-10 mM NH 4 Cl for 48 h-80% -SOD supplementation → ROS ↓ in astrocytes exposed to NH 4 Cl 5 mM -In vivo, intracerebral administration of ammonium via a microdialysis probe → GSH ↑ in the brain extracellular space → dependent on undisturbed GSH synthesis in astrocytes -basal GSH # concentrations in microdialysates from rat's striatum were about 10-fold higher than from prefrontal  binding adaptor protein-1 -pretreatment of microglia with apocynin completely abolished the ammonium induced ROS increase → activation of NADPH-oxidase by ammonium -oxidative stress ↑ is due to ROS ↑, but not RNS -iNOS ↑ in cultured astrocytes -iNOS was not affected by ammonium acetate treatment in vivo in the cerebral cortex [118] HAhyperammonemia. # Values approximationestimation from graphs.    week-8 +207.42% (p < 0.001) -SOD2 -granular layer: week-4 +12.65% (p < 0.001), week-8 +14.93% (p < 0.001) -SOD2 ↑ -hilus: week-4 +4.9% (ns), week-8 +50.91% (p < 0.001) -Oxo-8-dG ↑ granular layer: week-8 +213.87% (p < 0.001) -Oxo-8-dG ↑ hilus: week-8 +153.06% (p < 0.001) -GPX-1 ↑ granular layer: week-8 +116.39% In-vitro studies of cultured astrocytes exposed to Gln revealed an increase in OS [165]. Increased ammonium and Gln concentration together with Asc decline correlated significantly with OS increase in BDL rats [2,13]. In addition, the same study demostrated an elevated O ⋅ − 2 production and overexpression of SOD1 and SOD2 in CNS accompanied by SOD1 nuclear translocation to protect the genomic DNA [2]. Similarly, an increase of SOD expression was found in postmortem brain tissue of HE patients [112]. Increased IL-6 expression in activated microglia and astrocytes [2] can trigger the proinflammatory cascade as well as increased ROS formation [34,37,49,53,118,166]. SOD2 overexpression [2,112] acts as a switch to modulate microglial activation/inactivation during inflammation [167]. Increase of O ⋅ − 2 production and overexpression of SOD will lead to an increase of H 2 O 2 concentrations having a negative impact on LTP [7,168,169].
OS and inflammation are strongly linked and interdependent pathophysiological processes [47,170]. In the presence of OS, inflammatory processes will occur, hence assisting in the progression of OS. In parallel, if inflammation is the trigger, an OS response will be generated, contributing to the immune response [48,49]. Previous studies have demonstrated the synergistic participation of CNS OS which precedes the systemic OS [2], and inflammation in the progression of HE [2,96,135,138].
Bacterial infection and especially spontaneous bacterial peritonitis is a frequent precipitating factor in HE that contributes to the systemic inflammation [171,172]. Leukocytosis, a hallmark of systemic inflammation and indirect sign of OS, associated with an increased risk of mortality, was identified in both, the BDL rat model [2] and in HE patients [173,174]. Accelerated immune response (increased levels of TNF-, IL-18, IL-6) [24,36] and the oxidative metabolism/oxidative burst of phagocytic cells (neutrophil dysfunction and increased ROS emission) [2,33,129] at the inflamed site will advance tissue injury and immunopathology [175,176]. Furthermore, studies have indicated that high levels of IL-6 have a deleterious impact on the BBB integrity [38]. As a result, high levels of IL-6 in both the peripheral [36] and CNS [2] may increase the BBB permeability, allowing neurotoxic components (ROS/RNS, cytokines, NH + 4 , bile acids, bilirubin) to enter the brain and affect neurological functions.
Bile acids have the potential and ability of altering the gut microbiota, resulting in modifications of the total bile acid pool [177,178]. Increased bile acid concentrations have been shown to disrupt tight junctions, permeabilize the BBB via detergent-like cytolytic actions on cell membranes, gain access to the CNS, and contribute to neurological decline [179,180]. Bile acids have been also recognized as a pro-oxidants causing ROS release (interrupt electron transport chain at complex III), which may lead to depletion of antioxidants, thiol groups oxidation and lipid peroxidation [181,182]. Furthermore, bile acids may cause indirectly increase of OS through resident macrophages activation/oxidative burst [183]. The enteric nervous system connects the gut microbiome to the CNS and acts as a key communication route for the gut-liver-brain axis mediated by the activity of the vagal nerves [184]. Therefore, bacterial infection and systemic inflammation may also impact brain function through afferent vagal nerves activation by cytokines/chemokines release at the inflammatory sites [185,186] and impact the cognitive and motor functions [184]. Therefore, OS being related to innate inflammation is a common denominator and therapeutic target for many neurodegenerative disorders associated with cognitive deficits [11,[50][51][52][53].
Serum albumin accounts for a significant component of total extracellular antioxidant capacity (~70%). The reduced cysteine residue (Cys34) and a thiol group in serum albumin allow to scavenge the HO ⋅ and ONOO − respectively [16,24,171,[187][188][189][190]. The hypoalbuminemia was shown to increase the mortality risk in HE patients [17][18][19][20][21]. Therefore, a significant decrease of plasma albumin concentrations, an important extracellular antioxidant, and increase in the percentage of oxidized albumin [25,26,140,190] in HE patients will lead to decreased systemic antioxidative capacity and contribute to increase of OS. Furthermore, albumin binds several molecules reversibly, allowing for solubilization and transport (i.e., bilirubin, bile acids, hormones, and endotoxins), controls the immune system, and protects the endothelium [16,24,171,[187][188][189][190]. In liver cirrhosis bilirubin and bile acids concentrations rise [191,192], while binding capacity of albumin is impaired [140,190,193]. Therefore, decreased albumin synthesis by the diseased liver together with decreased binding capacity will decrease toxin clearance [190], increase of lipid peroxidation [171] and therefore lead to increase of OS, both systemic and CNS, and patient clinical status deterioration.

Conclusions
Taken together, the presented findings show that OS is a critical   component of HE pathogenesis even at an early stage and that a lack of defense exacerbates CNS status. According to recent studies, CNS OS occurs before systemic OS, suggesting that increased BBB permeability in the latter stage of disease progression may play a substantial role and contribute significantly to the increase of ROS/RNS in the brain. Therefore, further work on cell culture, animal models and postmortem brain tissue of patients with type C HE is required to elucidate the relationship/synergy between OS, ammonium, Gln, inflammation and the pathogenesis of HE. Moreover, additional longitudinal, multiparametric, and multimodal studies combining direct/indirect with invivo/ex-vivo/in-vitro techniques to study OS, brain regional differences in parallel with the relationship between OS, brain metabolic/functional alterations, cellular changes, and neurological manifestations in HE are needed. Because of the complexity of oxidative damage within the CNS, identifying OS biomarkers in clinical samples of HE patients is critical for a better understanding of OS-induced processes (molecular mechanisms) and to develop an appropriate diagnostic strategies. Despite significant research on oxidative stress related damage to the cells biomolecules (RNA/DNA, lipids, and proteins) and antioxidant status in biological samples, the literature on direct measurement of OS in clinical samples is limited, and screening and surveillance for OS biomarkers are not yet routine procedures in the healthcare sector. Furthermore, sample preparation should be done with caution to ensure sample stability and to reduce the possibility of oxidative damage to tissue/cells/biomolecules during collection. Because there is no gold standard for defining redox status, complementary techniques must be used when screening for OS biomarkers to eliminate methodological biases and to obtain clinically comprehensive (diagnostic and prognostic) information with high sensitivity and specificity to pathological alterations. It is critical to conduct a comprehensive panel analysis of both pro-and antioxidants, as well as inflammation biomarkers, which should be defined by the study's goal and provide an overall redox state in specific conditions.

Authors' contributions
KP, CC and DS have participated in conceptualization and manuscript preparation and have read and approved the final manuscript.