Background
Acute liver failure (ALF) is the clinical manifestation of sudden and severe hepatic injury and arises from many causes. ALF patients develop hepatic encephalopathy (HE) and subsequently symptomatic cerebral edema (CE), which progresses to intracranial hypertension. Although the incidence of clinically overt CE has decreased over the past 20 years, such hypertension still accounts for 20–25% of deaths [
1,
2]. Survival without transplantation is poor in patients with ALF complicated by severe encephalopathy.
Despite several decades of investigation, the mechanisms responsible for HE associated with ALF remain incompletely understood. Ammonia is universally considered the key precipitating factor that drives cerebral dysfunction in HE. In patients with ALF, the plasma concentrations of ammonia are drastically elevated and related to the development of brain edema and herniation [
3,
4]. A longstanding hypothesis proposes that the development of CE is related to the cerebral uptake of ammonia from the blood and subsequent disturbance of the intracellular metabolism [
4]. This raises the question of how ammonia enters the brain, and how the cerebral uptake of ammonia is affected.
The conventional assumption is that unionized ammonia (NH
3) can pass the blood–brain barrier by diffusion, whereas translocation of the ionized species (NH
4+) can be neglected. However, the so-called diffusion hypothesis is controversial [
5,
6]. Evidence suggests that a variable fraction of the ammonia uptake occurs through channel- or carrier-mediated transport. Recently, human nonerythroid Rhesus (Rh) glycoprotein B (RhBG) and nonerythroid Rh glycoprotein C (RhCG), as well as their non-human mammalian homologs (Rhbg/Rhcg), which are members of the Amt/MEP/Rh family, were identified as mammalian ammonium/ammonia transporters [
7]. Rhbg and Rhcg are widely expressed on the plasma membranes of various mammalian organs such as the liver, kidney, gastrointestinal tract, and lung [
8‐
12]. Rhbg transports both NH
3 and NH
4+, whereas Rhcg only transports NH
3 [
13]. Epithelial expression of RhCG mRNA has been reported in the testis, prostate, pancreas, and brain [
8], and Rhcg plays a critical role in ammonium handling and pH homeostasis both in the kidney and in the male reproductive tract [
13,
14]. Whether Rhcg is located at the luminal or abluminal side of brain capillaries and its capacity for brain ammonia transport remain unclear.
In addition, the correlation between the severity of HE and ammonia concentration is not always consistent in patients with ALF or chronic liver failure [
15‐
17]. Compared with placebo/no intervention, ammonia-lowering strategies are associated with beneficial effects on HE in patients with chronic liver diseases [
18,
19], whereas these agents are much less effective in ALF [
20]. These observations have led to a search for additional or alternative causes of brain edema and HE in ALF.
Evidence suggests that, in ALF, ammonia and proinflammatory cytokines act synergistically to affect brain function and cause HE and brain edema [
21‐
25]. TNF-α plays an important role in ALF from various causes. A close association between the severity of HE and the circulating levels of TNF-α is reported in ALF patients and animal models [
24‐
27]. At the cellular level, TNF-α increases ammonia uptake in human cerebrovascular endothelial cells [
28] and causes cell swelling in ammonia-sensitized cultured astrocytes [
29].
The aim of this study was to investigate the expression of the ammonia transporter Rhcg in the brain tissues of an ALF mouse model and the effect of TNF-α on brain Rhcg expression.
Methods
Animals and experimental protocols
Male C57BL/6 mice (6–8 weeks old, Animal Laboratory, China Medical University, Shenyang, Liaoning Province, China) were maintained on a 12:12-h light/dark cycle in a temperature-controlled, air-conditioned room, with free access to food and water. The mice were divided into eight groups. One group received intraperitoneal injections of d-galactosamine (D-GalN, 300 mg/kg body weight, Sigma, St. Louis, MO, USA) and lipopolysaccharide (LPS) (Escherichia coli O111:B4, 5 μg/kg body weight; Sigma) to induce ALF. A second ALF-induction group received intraperitoneal injections of D-GalN at 600 mg/kg body weight and TNF-α (10 μg/kg body weight, Origene, Rockville, MD, USA). Five control groups were intraperitoneally injected with D-GalN at 300 or 600 mg/kg body weight, LPS at 5 μg/kg body weight, TNF-α at 10 μg/kg body weight, or normal saline. One group was treated with anti-TNF-α–IgG antibody (Sigma) at 5 μg/g body weight via vena caudalis injection 30 min before LPS + D-GalN injection. Mice were euthanized 12 h after the administration of LPS/D-GalN. Mice were fasted for 12 h before drug injection. The eight groups were as follows: (1) LPS/D-GalN, (2) TNF-α/D-GalN, (3) D-GalN 300 mg/kg control, (4) LPS 5 μg/kg control, (5) TNF-α 10 μg/kg control, (6) saline control, (7) D-GalN 600 mg/kg 12 h control, and (8) anti-TNF-α–IgG antibody + LPS/D-GalN 12 h control. Mice from groups 1–5 were euthanized at 2, 6, 9, 12, and 24 h, mice from groups 6–8 were euthanized at 12 h. There were five mice at every time point in every group. Mice were sacrificed by decapitation, and serum was isolated from blood samples obtained by eyeball extirpation; brain and liver tissues were stripped for subsequent analyses. The study was approved by the Ethics Committee of China Medical University.
Blood biochemistry and serum TNF-α assay
Serum levels of alanine transaminase (ALT) were determined using an automatic analyzer (Hitachi, Tokyo, Japan). Serum TNF-α levels were determined using an ELISA kit (eBioscience, San Diego, CA, USA) according to the manufacturer’s protocol.
Histopathological analysis of liver tissue and brain tissue
Liver tissues were fixed in 10% phosphate-buffered formalin, and brain tissues (frontal parietal lobe) were fixed in 4% paraformaldehyde solution, embedded in paraffin, and cut into 5-μm serial sections using a rotatory microtome. Sections were stained with hematoxylin-eosin.
Measurement of blood and brain ammonia concentrations
Blood and brain ammonia concentrations were determined using an ammonia assay kit (BioVision, Milpitas, CA, USA) according to the manufacturer’s protocol.
Western blot analysis of Rhcg in brain tissues
To detect Rhcg proteins, brain tissue samples were extracted in RIPA lysis buffer, followed by centrifugation (12,000×g, 30 min). Protein concentration was determined with a BCA protein assay reagent kit (Thermo Fisher Scientific, Waltham, MA, USA). Equal amounts of proteins (100 μg) were separated by 8% SDS-PAGE and electrophoresed. Proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, USA), which was blocked with 5% non-fat dried milk in Tris-buffered saline (TBS) containing 0.05% Tween-20 (TBST) for 1 h at room temperature. After blocking nonspecific antibody binding, the PVDF membrane was incubated with mouse anti-mouse monoclonal Rhcg antibody (1:1000, Sigma) at 4 °C overnight. After washing with a TBST three times to remove unbound antibody, the membrane was incubated with a 1:2000 dilution of goat-anti-mouse IgG conjugated to horseradish peroxidase (HRP) for 1 h at room temperature. The membrane was washed three times with TBST to remove unbound secondary antibody and developed with a luminol chemiluminescent detection kit (Thermo Fisher Scientific). Protein expression was quantified by densitometry. GAPDH was used as an internal loading control. Relative protein levels were calculated based on a densitometric count ratio of Rhcg over GAPDH.
Rhcg RNA isolation and real-time quantitative polymerase chain reaction
Total RNA was isolated from mouse brain tissues using TRIzol reagent (Takara, Dalian, China). The concentration and purity of mRNA were determined by spectrophotometry at 260 and 280 nm, and mRNA was diluted to 500 ng/μL with DEPC water. RNA was incubated at 37 °C for 15 min, followed by 85 °C for 5 s for reverse transcription (RT) (PrimeScript RT Master Mix; Takara) to generate cDNA. The cDNA underwent 45 cycles of PCR (95 °C for 30 s, 95 °C for 5 s, and 60 °C for 30 s) [SYBR Premix Ex Taq II (Tli RNaseH Plus)] in the TP800 Thermal Cycler Dice Real Time System (Takara). The primers used were as follows: Rhcg forward, 5′-GCGCTGTAGGCTTCAACTTC-3′ and reverse, 5′-GGCTGACCTTGCCTAGAACT-3′; GAPDH forward, 5′-TCTCTGCTCCTCCCTGTTCC-3′ and reverse, 5′-GTTCACACCGACCTTCACCA-3′. The length of the PCR products for Rhcg and GAPDH were 183 bp and 91 bp respectively. Relative levels of Rhcg gene expression were calculated based on the amplification of a standard curve after a series of cDNA dilutions. GAPDH was used as a housekeeping control. Each sample was tested in triplicate.
Immunohistochemistry and immunofluorescence analysis for brain Rhcg
Mouse brain sections (frontal parietal lobe) were fixed in polyformaldehyde, embedded in paraffin wax, sectioned, and mounted on coated slides. Immunolocalization of Rhcg was accomplished using immunoperoxidase procedures. Brain sections were dewaxed in ethanol, rehydrated, and then rinsed in phosphate-buffered saline (PBS). To observe Rhcg localization, brain sections were incubated for 3 min in citrate buffer pH 6.0 in a pressure cooker at 100 kPa. Endogenous peroxidase activity was blocked by incubating the sections in 3% H2O2 for 10 min. The sections were blocked for 10 min with 10% normal goat serum (Zhongshan Biotechnology, Beijing, China) and then incubated overnight at 4 °C with a dilution of 1:300 of primary Rhcg antibody (Sigma). The sections were washed in PBS and incubated for 10 min with polymer-linked, peroxidase-conjugated goat anti-rabbit IgG (Zhongshan Biotechnology, Beijing, China); washed with PBS; and then exposed to diaminobenzidine for 5 min. The sections were washed in distilled water and then dehydrated with xylene, mounted, and observed by light microscopy. Image analysis was perfomed using Image J and Rhcg expression was quantified as % of IHC-colored pixels of the cerebral cortex.
To investigate the localization of Rhcg in the cerebral cortex, its colocalization with glial fibrillary acidic protein (GFAP, an astrocyte marker), CD31 (a marker of vascular endothelial cells), and neuronal nuclei (NeuN; a marker of neurons) was examined by immunofluorescence detection. Paraffin-embedded coronal sections were dewaxed with ethanol and rehydrated, and then placed in repair boxes filled with EDTA antigen retrieval buffer (pH 6.0). Antigen retrieval was performed in a microwave oven. After they cooled naturally, the sections were rinsed with PBS, blocked for 30 min with bovine serum albumin (Servicebio, Wuhan, China), and incubated overnight at 4°C in a humidified chamber with mouse monoclonal anti-Rhcg (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:100), rabbit polyclonal anti-CD31 (Abcam, Cambridge, UK; 1:200), rabbit polyclonal anti-GFAP (Proteintech, Wuhan, China; 1:1000), or rabbit polyclonal anti-NeuN (Proteintech; 1:200) antibody diluted in PBS. After washing in PBS, sections were incubated for 50 min at room temperature with a fluorophore-tagged species-specific secondary antibody (Cy3-conjugated goat anti-mouse IgG at 1:300 or AF488-labeled goat anti-rabbit IgG at 1:400; both from Servicebio). After washing with PBS, the sections were stained with 4',6-diamidino-2-phenylindole (DAPI) (Servicebio; 1:200) for 10 min at room temperature. After several washes with PBS, the sections were mounted with fluorescence mounting medium (Servicebio) and protein colocalization was examined by confocal microscopy (Eclipse C1, Nikon, Tokyo, Japan).
Statistical analysis
The software package SPSS, version 20.0, was used for statistical analysis. All data were analyzed using analysis of variance (ANOVA) followed by a least significant difference (LSD) test. P values of < 0.05 were considered significant. All data are presented as the mean ± standard deviation (SD).
Discussion
The present study is the first to investigate the expression of the newly recognized ammonia transporter Rhcg in the murine brain. The murine brain expresses Rhcg at the mRNA and protein levels, and Rhcg expression is detected in brain microvascular endothelial cells, astrocytes, and neurons.
Since 2000, knowledge of the role of Rh glycoproteins in ammonia transport has increased considerably. Multiple experimental models examined their involvement in renal ammonia metabolism, and gene knock-out studies showed that they are critical for normal renal ammonia metabolism. However, the roles of these transporters in the brain remain unknown. A study on cancer-causing genes in glioblastoma identified Rhbg and Rhcg as candidates [
30]. Hillmann et al. [
31] showed that both RhBG and RhCG are expressed at the mRNA level in the NG108-15 cell line, which is a mouse neuroblastoma × rat glioma hybrid cell line that is frequently used as a model for neuronal cells. However, they did not detect RhBG and RhCG mRNA in 1321N1 astrocytoma cells. The authors hypothesized that the expression of NH
3/NH
4+ transporters may be one explanation for the fact that ammonium is not very toxic to neuronal cells, whereas it is highly toxic to astrocytes, which trap ammonia instead of extruding it.
The detection of Rhcg in the murine brain in our experiments differs from the results of Liu et al. [
32]. They found that RhCG was abundantly expressed in the kidneys as well as in the brain, testis, placenta, pancreas, and prostate in human adult tissues. In mouse adult tissues, Rhcg was highly expressed in the kidney and testis but not in other tissues such as the brain [
32]. The most likely explanation is that previous studies used Northern blot analysis, which is less sensitive for detecting mRNA expression than the real-time RT-PCR assay used in the current study. Similar observations regarding the ability to detect low levels of Rhbg and Rhcg mRNA by real-time RT-PCR, but not by Northern blot analysis, were reported previously in the lung [
12] and intestinal tract, including the stomach, small intestine, and colon [
11].
The pathophysiology of HE and brain edema in ALF are multi-factorial. As a potent neurotoxin, the release of ammonia from the failing liver reflects its critical role. To enter the brain, ammonia has to pass the blood-brain barrier. Decades ago, this barrier was thought to function as a semi-permeable membrane; however, this view is currently considered simplistic [
5,
6,
33]. The luminal layer is formed by the endothelium, which is a highly active tissue; the brain endothelium has a higher concentration of mitochondria than any other endothelium in the body. The energy produced is used by a large number of membrane proteins that mediate the transport of ions and organic molecules. Since both the gas (NH
3) and ion (NH
4+) forms are capable of entering the cell, ammonia is considered a complex molecule in comparison with other weak acids and bases. Ammonia as a gas (NH
3) is lipid soluble and therefore enters the brain through diffusion. In the present study, blood ammonia levels began to increase at 6 h, whereas brain ammonia levels began to increase at 2 h in the LPS/D-GalN group. In both the LPS/D-GalN and TNF-α/D-GalN groups, blood ammonia levels at 24 h were still higher than those in the saline controls, whereas brain ammonia levels had decreased to normal. These results suggest that ammonia enters the brain not only through diffusion. The results of the current study demonstrate that Rhcg is expressed in brain microvascular endothelial cells, astrocytes, and neurons, and that brain Rhcg expression is positively associated with blood and brain ammonia levels in ALF mice. The results indicated that Rhcg is involved in brain ammonia transport in brain capillaries in ALF. The specific molecular form of ammonia, NH
3 versus NH
4+, transported by Rh glycoproteins remains to be confirmed. Rhbg appears to transport both NH
3 and NH
4+, whereas its homolog Rhcg only appears to transport NH
3 [
34]. Bakouh et al. showed that RhCG-mediated NH
3 and NH
4+ fluxes vary during NH
4Cl exposure, favoring the hypothesis of uncoupled fluxes or a bi-directional NH
3 flux and unidirectional NH
4+ flux (influx) [
35,
36].
TNF-α is an important pleiotropic proinflammatory cytokine that is involved in many biological and pathological processes. Abundant evidence supports the critical role of TNF-α in the pathogenesis of ALF, especially in LPS/D-GalN-induced ALF [
37‐
39]. In the present study, serum TNF-α levels reached a maximum at 2 h and were gradually restored at 6 h and 9 h, decreasing to normal values at 12 h and 24 h after LPS/D-GalN treatment. In the LPS control group, serum TNF-α levels also reached a maximum at 2 h and decreased dramatically to normal values at 6 h. LPS might induce macrophages and monocytes to release TNF-α. Our results indicated that the enhanced production of TNF-α induced by LPS/D-GalN was responsible for the development of ALF. When TNF-α was blocked with an antibody against TNF-α IgG, both the mortality and liver histopathology were significantly improved. Mice treated with TNF-α instead of LPS in conjunction with D-GalN showed similar histopathological characteristics to those of the LPS/D-GalN group. TNF gene polymorphisms significantly affect the outcome of ALF in humans and animals [
40,
41]. TNF-α may contribute to this process through its receptors TNFR1 [
42] and TNFR2 [
43].
Several lines of evidence indicate that TNF-α could be involved in the pathogenesis of HE in ALF [
24‐
27]. Patients with HE who are in ALF have elevated circulating levels of TNF-α [
44]. A strong correlation between the circulating levels of TNF-α and those of ammonia was identified in these patients, which suggests a significant relationship between TNF-α and ammonia in patients with liver failure and HE. Research shows that TNF-α increases the diffusion of ammonia in central nervous system-derived endothelial cells [
28], increasing the concentration of ammonia in the brain and consequently increasing ammonia toxicity to the central nervous system. TNF-α contributes to cell swelling in ammonia-sensitized cultured astrocytes [
29]. We found that the elevation of serum TNF-α paralleled the increasing concentration of ammonia in the blood and brain tissues in D-GalN/LPS-induced ALF and was correlated with Rhcg expression in brain tissues. When TNF-α was blocked with an antibody against TNF-α IgG, Rhcg expression in ALF brain tissues was significantly downregulated; blood ammonia levels remained higher than in the saline group, but brain ammonia levels decreased to the baseline level. TNF-α is detected between 2 and 6 h from the start of the D-GalN+LPS treatment; blood and brain ammonia reaches a maximum level between 6 and 12 h; and Rhcg expression peaks between 9 and 12 h. The fact that TNF-α expression precedes that of the ammonia transporter Rhcg suggests that it may play a role in transcriptional or posttranslational upregulation of the
Rhcg gene. The fact that the timing of Rhcg expression coincides with the ammonia peak in the brain indicates that Rhcg contributes to the accumulation of ammonia in the brain. These findings provide a likely mechanism for the exacerbation of brain edema or HE in patients with ALF. TNF-α may be involved in the development of brain edema in ALF by affecting brain ammonia through the modulation of Rhcg expression in the brain.