Decreased astrocytic thrombospondin-1 secretion after chronic ammonia treatment reduces the level of synaptic proteins: in vitro and in vivo studies

Authors


Abstract

Chronic hepatic encephalopathy (CHE) is a major complication in patients with severe liver disease. Elevated blood and brain ammonia levels have been implicated in its pathogenesis, and astrocytes are the principal neural cells involved in this disorder. Since defective synthesis and release of astrocytic factors have been shown to impair synaptic integrity in other neurological conditions, we examined whether thrombospondin-1 (TSP-1), an astrocytic factor involved in the maintenance of synaptic integrity, is also altered in CHE. Cultured astrocytes were exposed to ammonia (NH4Cl, 0.5–2.5 mM) for 1–10 days, and TSP-1 content was measured in cell extracts and culture media. Astrocytes exposed to ammonia exhibited a reduction in intra- and extracellular TSP-1 levels. Exposure of cultured neurons to conditioned media from ammonia-treated astrocytes showed a decrease in synaptophysin, PSD95, and synaptotagmin levels. Conditioned media from TSP-1 over-expressing astrocytes that were treated with ammonia, when added to cultured neurons, reversed the decline in synaptic proteins. Recombinant TSP-1 similarly reversed the decrease in synaptic proteins. Metformin, an agent known to increase TSP-1 synthesis in other cell types, also reversed the ammonia-induced TSP-1 reduction. Likewise, we found a significant decline in TSP-1 level in cortical astrocytes, as well as a reduction in synaptophysin content in vivo in a rat model of CHE. These findings suggest that TSP-1 may represent an important therapeutic target for CHE.

image

Defective release of astrocytic factors may impair synaptic integrity in chronic hepatic encephalopathy. We found a reduction in the release of the astrocytic matricellular proteins thrombospondin-1 (TSP-1) in ammonia-treated astrocytes; such reduction was associated with a decrease in synaptic proteins caused by conditioned media from ammonia-treated astrocytes. Exposure of neurons to CM from ammonia-treated astrocytes, in which TSP-1 is over-expressed, reversed (by approx 75%) the reduction in synaptic proteins. NF-kB = nuclear factor kappa B; PSD95 = post-synaptic density protein 95; ONS = oxidative/nitrative stress.

Abbreviations used
CHE

chronic hepatic encephalopathy

CM

conditioned media

GFAP

glial fibrillary acidic protein

NH4Cl

ammonium chloride (ammonia)

TAA

thioacetamide

TSP-1

thrombospondin-1

Hepatic encephalopathy (HE) is a major neurological complication in patients with severe liver disease. It is characterized by impaired neurological function and occurs in acute and chronic forms. The encephalopathy associated with acute HE (acute liver failure) generally occurs following massive liver necrosis, generally because of viral hepatitis, hepatic neoplasms, vascular causes, or exposure to various hepatotoxins. It presents with the abrupt onset of delirium, seizures, and coma, and has an extremely poor prognosis (70% mortality) (Lee 2012; Shawcross and Wendon 2012). HE in the setting of chronic liver disease [chronic hepatic encephalopathy (CHE)] generally occurs as a consequence of cirrhosis of the liver, usually secondary to alcoholism (Wakim-Fleming 2011). CHE is characterized by confusion, disorientation, behavioral changes, impaired cognition, inverted sleep–wake cycles, and motor disturbances. The molecular basis for the neuropsychiatric disorder in CHE remains elusive (Mullen and Prakash 2012; Patel et al. 2012).

Neuronal dysfunction is a well-established finding in CHE (Cauli et al. 2009). Increased synthesis of gamma-aminobutyric acid (GABA) and subsequent alterations in GABA-ergic neurotransmission, as well as an increase in endogenous benzodiazepines (which modulate GABA-mediated neurotransmission), have been proposed to be involved in CHE (Llansola et al. 2012). Inhibition of cortical function and the subsequent behavioral defects from altered GABA-ergic signaling have also been postulated as major mechanisms leading to CHE (Bismuth et al. 2011).

HE is one neurological disorder in which, early on, astrocytes were suggested to play a vital role (Norenberg 1986). A major factor in the pathogenesis of HE is elevated blood and brain ammonia levels because of the inability of the injured liver to detoxify ammonia by the synthesis of urea. Once in brain, ammonia is metabolized to glutamine by the action of glutamine synthetase, a process that only occurs in astrocytes (Norenberg 1979). It is therefore of interest that the principal histopathological finding in CHE is the presence of Alzheimer type II astrocytosis. These astrocytes are characterized by the presence of pale and enlarged nuclei (often found in pairs), along with margination of nuclear chromatin and the presence of prominent nucleoli (Norenberg 1979). Although the significance of this astrocytic change is incompletely understood, abundant data strongly suggest that Alzheimer type II astrocytes are dysfunctional cells (Norenberg 1987; Norenberg et al. 1998). This has led to the concept that HE fundamentally represents a primary astrogliopathy (Norenberg 1987; Norenberg et al. 1998).

Astrocytes play a crucial role in the central nervous system by regulating a number of critical processes, including synaptogenesis, synaptic function, modulation of neurotransmission, regulation of pH, ion and water homoeostasis, energy metabolism, defense against oxidative stress and the detoxification of ammonia, metals, and other toxins (Norenberg 1987; Wang and Bordey 2008; Bélanger and Magistretti 2009). Astrocytes are also involved in the provision of growth factors and nutrients to neurons and other neural cells (Wang and Bordey 2008), as well as in the formation and maintenance of the blood–brain barrier (Abbott et al. 2006).

One potential mechanism by which defective astrocytes may impact neuronal integrity is through a reduction in thrombospondin-1 (TSP-1) levels. Astrocytes are known to synthesize and secrete TSP-1 (Christopherson et al. 2005; Tran and Neary 2006), and a reduction in TSP-1 expression by siRNA silencing was reported to decrease neuronal synaptophysin protein expression (Yu et al. 2008). Such reduction in TSP-1 and synaptophysin protein levels was associated with behavioral abnormalities in experimental models of stroke (Lin et al. 2003; Liauw et al. 2008), Alzheimer's disease (Buée et al. 1992), and Down's syndrome (Garcia et al. 2010).

This study examined intra- and extracellular levels of TSP-1 in ammonia-treated cultured astrocytes, and the effect of conditioned media (CM) from ammonia-treated astrocytes on synaptic protein levels in cultured neurons. A significant decrease in TSP-1 protein level was observed in ammonia-treated astrocyte cultures. We also found a reduction in synaptophysin protein levels when cultured neurons were exposed to CM from ammonia-treated astrocytes, and that enhancing intra- and extracellular levels of astrocytic TSP-1 reversed the synaptophysin loss. We further observed a significant decline in astrocytic TSP-1, and in neuronal synaptophysin levels in a rat model of CHE. Our findings suggest that dysfunctional astrocytes resulting from ammonia treatment negatively impacts neuronal synaptic integrity, thereby contributing to the neurological abnormalities associated with CHE.

Materials and methods

Astrocyte cultures

Primary cultures of cortical astrocytes were prepared from brains of 1- to 2-day-old rat pups by the method of Ducis et al. (1990). Briefly, cerebral cortices were freed of meninges, minced, dissociated by trituration and vortexing, and were seeded onto 35-mm culture dishes in Dulbecco's modified Eagle's medium containing penicillin, streptomycin, and 15% fetal bovine serum. The culture plates were incubated at 37°C with 5% CO2 and 95% air. Culture media were changed twice weekly. On day 10 post seeding, fetal bovine serum was replaced with 10% horse serum. After 14 days, cultures were treated with 0.5 mM dibutyryl cAMP (Sigma, St. Louis, MO, USA) to enhance cellular differentiation (Juurlink and Hertz 1985). Cultures consisted of at least 95% astrocytes as determined by glial fibrillary acidic protein (GFAP) immunohistochemistry. All cultures used were 21–23 days old.

All animal procedures followed the guidelines established by the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by our Institutional Animal Care and Use Committee (IACUC).

Neuronal cultures

Cortical neuronal cultures were prepared by a modification of the method described by Schousboe et al. (1989). Briefly, cortices were removed from 16- to 18-day-old rat fetuses and placed in Dulbecco's modified Eagle's medium (30 mM glucose) with 25 mM KCl and 10% horse serum. The tissue was minced and mechanically dissociated with a pipette. Approximately 1–2 × 106 cells per mL were seeded onto poly-d-lysine-coated 35-mm culture dishes. To prevent the proliferation of astrocytes, cytosine arabinoside (10 μM) was added to the culture medium 48 h after seeding. These cultures consist of at least 90% neurons as determined by immunohistochemical staining for neurofilament protein; the remaining cells were chiefly astrocytes. Experiments were performed on cultures that were 7–8 days old.

TSP-1 over-expression in cultured astrocytes

To examine whether the exposure of cultured neurons to CM from ammonia-treated astrocytes in which TSP-1 is over-expressed diminishes or prevents the reduction in synaptic proteins, we over-expressed TSP-1 in cultured astrocytes. Briefly, astrocytes were transfected with TSP-1 cDNA (tagged with the pIRES2-AcGFP1 vector; GENEWIZ Inc, South Plainfield, NJ, USA). Cultures were exposed to different concentrations of TSP-1 cDNA (50, 100, and 250 ng/2.5 × 105 cells) for 72 h. Mirus TransIT-TKO transfection reagent was used to transfect TSP-1 following the manufacturer's instructions (#MIR 2150; Mirus, Madison, WI, USA). At the end of transfection, the culture media were replaced with normal media. These cultures were treated with and without ammonia for 10 days, and at the end of treatment, intra- and extracellular levels of TSP-1 were measured by western blots. The empty vector, as well as the transfection reagent alone, was used as controls for all experiments.

Immunohistochemistry of TSP-1 and synaptophysin in CHE

Control and thioacetamide (TAA)-treated rats (four animals each) were anesthetized at the end of TAA treatment with a mixture of ketamine (80 mg/kg) and xylacine (20 mg/kg), and were then transcardially perfused with heparinized saline for 1 min, followed by fixation in 4% paraformaldehyde for 15 min. After decapitation, heads were left in the same fixative for an additional 24 h at 22°C. Brains were cryoprotected (12–24 h) in 30% sucrose. Cortical sections (20 μm thick) were obtained with a cryostat; sections were blocked with 10% goat serum and incubated overnight at 4°C with anti-synaptophysin antibody (rabbit monoclonal, YE269, 1 : 150 dilution; cat# 32127; Abcam, Cambridge, MA, USA), BA24 anti-thrombospondin 1 antibody (mouse monoclonal A6.1, 1 : 100 dilution; Millipore, Billerica, MA, USA), and anti-GFAP antibody (astrocyte marker, 1 : 150 dilution, cat# ab7260; Abcam), as described previously (Jayakumar et al. 2014). Following incubation with primary antibodies, sections were washed with tris-buffered saline containing 1% Tween-20 and incubated with fluorescent horseradish peroxidase-conjugated secondary antibodies [(1 : 500; Alexa Flour-546 goat anti-rabbit IgG (H + L) (Life Technologies, Grand Island, NY, USA) for GFAP, Alexa Flour-488 goat anti-mouse IgG (H + L) (Life Technologies) for TSP-1, and Alexa Flour-488 goat anti-rabbit IgG (H + L) (Life Technologies) for synaptophysin)], for 2 h. Sections were then covered with commercial mounting media (Vector Laboratories, Burlingame, CA, USA) containing 4′,6-diamidino-2-phenylindole (DAPI) (nuclear stain). Immunofluorescent images were acquired with a Zeiss LSM510/UV Axiovert 200M confocal microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY, USA) with a plan apochromat 40× objective lens, and a 2× zoom resulting in images 125 × 125 μm in area and 1.0 μm optical slice thickness (1.0 Airy units for Alexa Fluor 546 or 568 emission channel). Random collection of images from sections of control and TAA-treated rats was achieved by systematically capturing each image in a ‘blinded’ manner by moving the microscope stage approximately 5 mm in four different directions. At least 17 fluorescent images were captured per rat, and the images merged to localize astrocytic TSP-1 protein. TSP-1 and synaptophysin levels in cultured neurons and in brain sections were quantified using the Volocity 6.0 High Performance Cellular Imaging Software (PerkinElmer, Waltham, MA, USA) as described previously (Rick et al. 2013; Jayakumar et al. 2014), and normalized to the number of DAPI-positive cells, as well as to the intensity of DAPI.

Statistical analysis

All experiments were performed and repeated four to six times using cells derived from different batches of astrocyte and neuronal cultures. Five to six individual culture plates were used in each experimental group for TSP-1, and four to six for synaptophysin measurements. Six animals from each group were used for in vivo studies. Data of all experiments were subjected to analysis of variance followed by Tukey's post hoc comparisons. A value of p < 0.05 was considered significant. Error bars, mean ± SE.

Results

Intra- and extracellular TSP-1 in cultured astrocytes after treatment with ammonia

Primary cultures of rat cortical astrocytes treated with ammonia were used in this study. The use of such cultures as a model of HE is highly appropriate since substantial evidence invokes a crucial role of ammonia in the pathogenesis of HE, and astrocytes are the principal cells affected in this condition (Norenberg et al. 2009). Moreover, many of the findings occurring in HE in vivo are also observed in ammonia-treated astrocyte cultures, including characteristic morphologic changes, cell swelling, defects in glutamate transport, up-regulation of the 18-kDa translocator protein, reduction in levels of glial fibrillary acidic protein (Sobel et al. 1981; Kretzschmar et al. 1985; Kimura and Budka 1986) and myo-inositol (Norenberg et al. 2009), disturbance in energy metabolism, and evidence of oxidative/nitrative stress (ONS) (Lange et al. 2012).

Pathophysiological concentrations of ammonia (0.5, 1.0, and 2.5 mM NH4Cl) (Dejong et al. 1993; Singh and Trigun 2010; Carbonero-Aguilar et al. 2011) (also see Table 1) were added to astrocyte cultures for different time periods (1, 5, and 10 days with regular media changes, once in 2 days for both 5- and 10-day treatment). At the end of treatment (24 h after the addition of last ammonia exposure), TSP-1 protein levels in the cell extracts and culture media were measured by western blots. Fresh addition of inhibitors and antioxidants was performed with each change of culture medium during the entire period of incubation.

Table 1. Blood and brain ammonia levels of rats treated with thioacetamide (TAA)
 ControlTAA
  1. Control = 5; TAA = 5. Blood and brain ammonia levels were determined 24 h after the last injection of TAA (100 mg/kg bw).

  2. a

    p < 0.05 versus control.

Blood ammonia (μM)191.6 ± 11.7611.2 ± 32.1a
Brain ammonia (mM)0.34 ± 0.070.96 ± 0.16a

Exposure of astrocytes to 0.5, 1.0, and 2.5 mM NH4Cl for 10 days caused a reduction in extracellular TSP-1 levels (58.6, 43.9, and 67.5%, respectively) (Fig. 1). Intracellular TSP-1 levels were also reduced (16.2, 57.4, and 40.7%, respectively) (Fig. 2). We additionally found a significant decline in TSP-1 mRNA level when astrocytes were exposed to 0.5, 1.0, and 2.5 mM ammonia for 10 days (Fig. 2c).

Figure 1.

Extracellular thrombospondin-1 (TSP-1) level after a 10-day treatment of cultured astrocytes with 0.5–2.5 mM ammonia (NH4Cl). (a) Representative western blots from cell culture media of ammonia-treated astrocytes show a significant decrease in TSP-1 levels. (b) Quantification of ammonia-induced changes in TSP-1 protein levels. *p < 0.05 versus control. C, control; NH4+, ammonia.

Figure 2.

Intracellular thrombospondin-1 (TSP-1) level after a 10-day treatment of cultured astrocytes with 0.5–2.5 mM ammonia. (a) Representative western blots from ammonia-treated astrocytes (1 and 2.5 mM) show a significant decrease in intracellular TSP-1 levels. (b) Quantification of NH4Cl-induced changes in TSP-1 protein. TSP-1 levels were normalized against α-tubulin. (c) TSP-1 mRNA expression after ammonia treatment. *p < 0.05 versus control. C, control; N, NH4Cl.

Exposure of astrocytes to 0.5 and 1.0 mM ammonia for 5 days had no effect on extra- (data not shown) or intracellular TSP-1 levels (Figure S2a). However, exposure to 2.5 mM ammonia significantly reduced TSP-1 levels by 51.6% (Figure S2a and b). Although TSP-1 protein levels were not altered in 0.5 and 1.0 mM ammonia-treated astrocytes, a significant decline in TSP-1 mRNA level was observed at these ammonia concentrations (Figure S2c).

Hevin is another matricellular protein that is known to influence neuronal synaptic integrity in other conditions (Kucukdereli et al. 2011). Its level also decreased in ammonia-treated cultured astrocytes (intra- and extracellular levels by 35.7% and 31.1%, respectively, n = 5, p < 0.05 vs. control). As decreased transforming growth factor beta (TGF-β1) was shown to influence TSP-1 levels (Okamoto et al. 2002; Mimura et al. 2005; McGillicuddy et al. 2006), we examined whether ammonia also influenced TGF-β1 levels. A reduction (57.8%) in intracellular levels of TGF-β1 was detected when astrocytes were treated with ammonia (Figure S3). The reduction in the level of hevin, however, was of a lesser magnitude than that observed with TSP-1.

Effect of conditioned media from ammonia-treated astrocytes on neuronal synaptic proteins

We examined whether an ammonia-induced decline in extracellular TSP-1 concentration contributes to a reduction in neuronal synaptophysin. Accordingly, cultured astrocytes were treated with ammonia (0.5, 1.0, and 2.5 mM, NH4Cl) for 10 days, and at the end of treatment, 1 mL of CM from ammonia-treated astrocytes was added to neuronal cultures. Synaptophysin level was determined 24 h later by immunofluorescence and western blots. Cultured neurons exposed to CM from 0.5–2.5 mM NH4Cl-treated astrocytes showed a significant reduction in synaptophysin levels in a dose-dependent manner, as measured both by immunofluorescence (Figure S4a–e) and western blots (Fig. 3a), which corresponded well with levels of TSP-1 reduction. Quantification of the immunocytochemistry data also corresponded well with findings in western blots (Figure S4e and Fig. 3a). While CM from ammonia-treated astrocytes (0.5–2.5 mM) reduced synaptophysin levels, the extent of reduction among the three ammonia-treated groups was not statistically significantly different from each other. Exposed neuron cultures to CM from control astrocytes had no effect on synaptophysin protein levels (data not shown).

Figure 3.

Synaptophysin, PSD95 and synaptotagmin protein levels in cultured neurons. (a) Representative western blots from astrocytes treated with ammonia (0.5–2.5 mM) for 10 days and the conditioned media (CM) added to neurons for 24 h. Such treatment led to a significant reduction in synaptophysin in a dose-dependent manner. (b) CM from ammonia-treated astrocytes also decreased PSD95, and synaptotagmin levels (c). *p < 0.05 versus control. C, control.

It should be emphasized that the direct exposure of cultured cortical neurons to 0.5–2.5 mM ammonia for 24 h had no significant effect on synaptophysin level. Additionally, we found no residual ammonia in the ammonia-treated astrocytic CM (at 24 h) that was subsequently applied to cultured neurons. We also found no detectable levels of ammonia 2 h after exposure of cultured astrocytes to ammonia (Jayakumar et al. 2006). This absence of ammonia is likely caused by a rapid conversion of ammonia into glutamine via glutamine synthetase activity (Cooper et al. 1979). These findings indicate that the ammonia-mediated reduction in neuronal synaptophysin levels is caused by an effect of ammonia on astrocytes, and not to a direct effect of ammonia on neurons.

We also examined whether PSD95, a post-synaptic density protein, and synaptotagmin-1, a Ca2+ sensor in the membrane of the pre-synaptic axon terminal, which has also been implicated in the maintenance of synaptic integrity, were similarly affected by CM from ammonia-treated astrocytes. We indeed found that when cultured neurons were exposed to CM from ammonia-treated astrocytes (1.0 mM, 10 days), decreased levels of PSD95 (32.7%) and synaptotagmin-1 (26.8%) were observed (Fig. 3b and c). However, the extent of reduction in PSD95 and synaptotagmin-1 levels was less than that observed in synaptophysin. The direct exposure of cultured neurons to ammonia (1 mM, 24 h) also had no significant effect on PSD95 protein levels (data not shown).

CM from TSP-1 over-expressing cultured astrocytes that were treated with ammonia, when added to cultured neurons, reversed the decrease in synaptic proteins

We then examined whether the effect of CM from ammonia-treated cultured astrocytes on the decline in levels of neuronal synaptic proteins is indeed caused by diminished astrocytic TSP-1 levels. Accordingly TSP-1 was over-expressed (with 100 and 250 ng/mL TSP-1 cDNA) in cultured astrocytes. TSP-1 over-expressed cells that were treated with ammonia (1 mM, 10 days) showed a significant increase in extracellular levels of TSP-1 (1.2- to 1.5-fold more than the effect of CM from ammonia-treated astrocytes that were exposed to an ‘empty’ vector or with the transfection reagent alone). Exposure of cultured cortical neurons to CM from ammonia-treated astrocytes showed a reduction in synaptic proteins. On the other hand, exposure of neurons to CM from ammonia-treated astrocyte cultures that had been transfected with 250 ng TSP-1 cDNA led to a lesser reduction in synaptophysin, PSD95, and synaptotagmin content by 73.5%, 65.9%, and 78.2%, respectively (n = 4, p < 0.05 vs. respective controls).

We also examined whether the addition of recombinant TSP-1 (rTSP-1) also prevented the effect of CM from ammonia-treated (1.0 mM) cultured astrocytes on the synaptophysin loss in neurons. For this purpose, cultured cortical neurons were exposed to CM from ammonia-treated (1.0 mM) cultured astrocytes, along with recombinant TSP-1 (rTSP-1; 50, 100, 200 ng/mL) for 24 h, and levels of synaptophysin were measured by western blots. Neurons exposed to CM from ammonia-treated astrocytes showed a 43.6% reduction in synaptophysin content, and such effect was significantly reversed by 50, 100, and 200 ng/mL rTSP-1 (39.5%, 68.7%, and 35.6%, respectively) (Fig. 4).

Figure 4.

Effect of recombinant thrombospondin-1 (rTSP-1) on synaptophysin level. Representative western blots from cultured astrocytes treated with ammonia (1.0 mM) for 10 days and the CM was then added to cultured neurons for 24 h, along with rTSP-1. Such treatment caused a reversal of the synaptophysin loss. *p < 0.05 versus control. †p < 0.05 versus CM from ammonia-treated (1.0 mM) astrocytes. C, control; CM, conditioned medium.

In addition, we investigated whether depletion of TSP-1 in the astrocytic CM affects neuronal synaptophyin content. We found that TSP-1 levels in the CM of astrocytes were depleted by immunoprecipitation. The resulting CM when added to cultured neurons (for 24 h) exhibited a significant loss of synaptophysin content (43.6%), supporting the concept that TSP-1 is critical for the maintenance of synaptophysin levels in CHE.

Effect of c-Myc on TSP-1 level after ammonia treatment to cultured astrocytes

c-Myc is a transcription factor expressed in astrocytes in vitro and in vivo (Saljo et al. 2002; Liu et al. 2006), and its over-expression was shown to cause a decrease in TSP-1 protein by exerting a repressor effect on TSP-1 mRNA (Watnick et al. 2003). We therefore examined whether c-Myc contributes to the ammonia-induced reduction in astrocytic TSP-1. Exposure of cultured astrocytes to ammonia (1 mM, 10 days) increased levels of c-Myc protein by 86.2% (Fig. 5a).

Figure 5.

Effect of c-Myc on thrombospondin-1 (TSP-1) expression. (a) Representative western blot from ammonia-treated astrocytes shows a significant increase in c-Myc protein. (b) Treatment with c-Myc siRNA significantly reduced c-Myc concentration. (c and d) Inhibition of c-Myc by siRNA significantly increased intra- and extracellular TSP-1 protein level. *p < 0.05 versus control; †p < 0.05 versus scrambled control. C, control; NH4+, ammonia; Scr, scrambled control (control siRNA).

We then examined whether silencing c-Myc with siRNA prevents or diminishes the reduction in TSP-1 levels by ammonia. c-Myc was silenced in cultured astrocytes with siRNA as described previously (Lu and Hong 2009), and found that transfection with 40, 60, and 80 nM of c-Myc siRNA significantly reduced the concentration of c-Myc (by 47.8%, 69.4%, and 76.5%, respectively) (Fig. 5b). Exposure of c-Myc-silenced cells (80 nM siRNA) to 1 mM ammonia (10 days) caused a 64% and 60% recovery in intra- and extracellular levels of TSP-1, respectively (Fig. 5c and d) (n = 5, p < 0.05 vs. respective controls) as compared to the effect of ammonia on scrambled siRNA-treated cells, which showed a 40–45% reduction in TSP-1. Exposure of cultured neurons (24 h) to CM from ammonia-treated astrocytes (10 days), in which the c-Myc-gene was silenced, resulted in a lesser reduction in the level of synaptophysin (by 61%) (Fig. 6a).

Figure 6.

Effect of astrocytic c-Myc inhibition on neuronal synaptophysin levels. (a–c) Astrocytes in which the c-Myc-gene was silenced, or exposed to a c-Myc inhibitor, 10058-F4, along with ammonia and the CM then added to cultured neurons (for 24 h) resulted in a lesser reduction in synaptophysin level. *p < 0.05 versus control; †p < 0.05 versus scrambled control. C, control; NH4+, ammonia; Scr, scrambled control (control siRNA); CM, conditioned media.

We also examined whether a pharmacological inhibition of c-Myc prevents or diminishes the reduction in TSP-1 levels by ammonia. We found that treatment of astrocytes with ammonia plus 10058-F4 (50 μM), an inhibitor of c-Myc, for 10 days enhanced intra- and extracellular TSP-1 levels by 64.2% and 61.2%, respectively (Fig. 6b). Higher doses of 10058-F4 (100 and 150 μM) showed no additional effect (data not shown). Furthermore, exposure of cultured neurons (24 h) to CM from ammonia-treated astrocytes (10 days), in which c-Myc was inhibited by 10058-F4, caused a reversal of the synaptophysin loss (59.4%), as compared to the effect of CM only from ammonia-treated astrocytes (Fig. 6c).

Attenuation of the ammonia-induced inhibition of TSP-1 by antioxidants and an inhibitor of nuclear factor kappa B

A considerable body of evidence suggests that the formation of reactive oxygen/nitrogen species and the resulting ONS plays a major role in HE (Jayakumar and Norenberg 2012). In addition, activation of the transcription factor nuclear factor kappa B (NF-кB) was identified in ammonia-treated cultured astrocytes (Schliess et al. 2002; Sinke et al. 2008). Since ONS and NF-кB have been strongly implicated in the mechanism of TSP-1 down-regulation in other conditions (De Stefano et al. 2009; Tan et al. 2009; Chen et al. 2011), we examined whether ONS and NF-кB are likewise involved in the mechanism of the ammonia-induced reduction in TSP-1 levels in cultured astrocytes. Astrocyte cultures were treated with the antioxidants Mn(III) tetrakis (4-benzoic acid) porphyrin (MnTBAP, 10 μM), dimethylthiourea (DMTU, 100 μM), and the nitric oxide synthase inhibitor N-nitro-l-arginine methyl ester (l-NAME) (250 μM), as well as SN50 (0.5–1.0 μM), an inhibitor of NF-кB, along with ammonia (1.0 mM) for 10 consecutive days (with regular media changes), and the level of TSP-1 in the culture media was determined by western blots. MnTBAP, DMTU, and l-NAME significantly reduced the inhibition of TSP-1 level in the culture media after the exposure to 1.0 mM ammonia (Figure S5a). Similarly, SN50 significantly reduced the inhibition of extracellular TSP-1 level after the exposure to 1.0 mM ammonia (Figure S5b). These findings indicate that ONS and the activation of NF-кB indeed contribute to the ammonia-induced inhibition of TSP-1 release.

The effect of antioxidants on the ammonia-induced reduction in TSP-1 mRNA expression was also examined in cultured astrocytes. Treatment of astrocytes with MnTBAP (10 μM), DMTU (100 μM), and l-NAME (250 μM) significantly diminished the ammonia-induced reduction in TSP-1 mRNA (Figure S5c), suggesting that the reduction in TSP-1 levels in ammonia-treated astrocytes is caused by a defective transcriptional regulation of TSP-1 by ONS.

We further examined whether an agent known to enhance TSP-1 synthesis and release is capable of reversing the ammonia-induced reduction in TSP-1. For this purpose, we investigated the effect of metformin, a commonly used anti-diabetic agent that is known to enhance the release of TSP-1 (Tan et al. 2009). Metformin was also shown to exert a protective effect in diabetic patients who also had HE (Ampuero et al. 2012). We found that metformin (25 μM) diminished the ammonia-induced reduction in intra- and extracellular levels of TSP-1 in astrocytes (by 77.3% and 68.9%, respectively, p < 0.05 vs. ammonia group). Furthermore, exposure of cultured neurons (24 h) to CM from ammonia-treated astrocytes (10 days), in which TSP-1 levels were enhanced by metformin, caused an elevation in synaptophysin levels by 65.8% (p < 0.05), as compared to the effect of CM astrocytes treated only with ammonia.

TSP-1 and synaptophysin levels in cerebral cortex of rats with CHE

To examine whether comparable alterations in TSP-1 and synaptophysin protein level also occur in an in vivo model of CHE, rats were treated with the liver toxin thioacetamide (TAA, 100 mg/kg bw) for 10 days, and the extent of TSP-1 and synaptophysin protein expression in cortical sections were examined. To identify changes in TSP-1 in astrocytes, sections were co-immunostained with GFAP (astrocyte marker) and DAPI (nuclear marker), and the degree of immunofluorescence examined with a confocal laser scanning microscope. A significant reduction (52.4%) in TSP-1 fluorescence was detected in astrocytes from TAA-treated rats as compared to control rats (Fig. 7a–c). We also found a comparable reduction in synaptophysin protein in cortical sections of rats treated with TAA (Fig. 8). A slight reduction in GFAP fluorescence was identified in TAA-treated rats (Fig. 7). Such reduction was previously reported in humans with HE, as well as in ammonia-treated cultured astrocytes (Sobel et al. 1981; Kretzschmar et al. 1985; Kimura and Budka 1986). However, we did not observe a significant change in the total number of astrocytes in this in vivo model of CHE as measured by counting the astrocyte nuclei (stained with DAPI) that were co-stained with GFAP. In addition, it should be noted that a decrease in the number of astrocytes is not a feature of humans with chronic HE (Norenberg et al. 1992). We also measured TSP-1 level in cerebral cortex of TAA-treated rats by western blots. Cortical tissues from rats treated with TAA for 3 days showed a significant decline in TSP-1 content (42.3% decrease, as compared to control; n = 5) (Fig. 7d and e), which corresponded well with data obtained by immunofluorescence.

Figure 7.

Thrombospondin-1 (TSP-1) protein levels in astrocytes from cerebral cortex of rats with chronic hepatic encephalopathy (CHE) following the administration of the hepatotoxin thioacetamide (TAA, 100 mg/kg) for 10 days. (a) Control brain: glial fibrillary acidic protein (GFAP, red, astrocytes), TSP-1 (green), and DAPI (blue, nuclei). The co-localization (merged) image of TSP-1 and GFAP illustrates the enrichment of TSP-1 in astrocytes. (b) TAA-treated rat brain showed a reduction in astrocytic TSP-1 content. (c) Relative quantification of TSP-1 immunofluorescence staining. Scale bar = 20 μm. (d). Cortical TSP-1 level after a 10-day treatment of rats with TAA. Representative western blots from TAA-treated rats show a significant decrease in intracellular TSP-1 level. (e) Quantification of TSP-1 immunoblots.*p < 0.05 versus control. C, control.

Figure 8.

Synaptophysin level in cerebral cortex of rats with chronic hepatic encephalopathy (CHE) following the administration of the hepatotoxin thioacetamide (TAA) for 10 days. (a) Synaptophysin level in control (normal) brain. (b) A significant reduction in synaptophysin level was found in TAA-treated rats as compared to control animals. (c) Relative quantification of synaptophysin immunofluorescence staining. *p < 0.05 versus control. Cont, control; ΤΑΑ, thioacetamide. Scale bar = 20 μm.

Discussion

Our results demonstrate that chronic ammonia toxicity in cultured astrocytes results in decreased extracellular levels of TSP-1. In addition, synaptophysin levels declined in cultured neurons after exposure to CM derived from ammonia-treated astrocytes, and such effect was diminished when recombinant TSP-1 (rTSP-1) was added to the CM of ammonia-treated astrocytes. We also found decreased levels of other synaptic proteins (PSD95 and synaptotagmin-1), when neurons were exposed to CM derived from ammonia-treated astrocytes. Increased c-Myc protein content was detected in ammonia-treated astrocytes, while silencing c-Myc or pharmacological inhibition of c-Myc significantly enhanced intra- and extracellular levels of TSP-1 after ammonia treatment in cultured astrocytes. Furthermore, exposure of cultured neurons (24 h) to CM from ammonia-treated astrocytes, in which c-Myc was silenced or inhibited by 10058-F4, caused a reversal of the synaptophysin loss. A reduction in TSP-1 and neuronal synaptophysin was also detected in an in vivo rat model of CHE. Altogether, these findings strongly suggest that an ammonia-induced reduction in TSP-1, and possibly other factors in astrocytes (see below), results in a decline in synaptic protein levels, which likely contributes to the pathogenesis of CHE.

TSP-1, also known as THBS1 or THP1 protein, is a member of the thrombospondin family that in humans is encoded by the THBS1 gene. This protein can bind to fibrinogen, fibronectin, laminin, type V collagen, and integrins alpha-V/beta-1, and thereby initiate cell–cell and cell–matrix interactions (Li et al. 2002). TSP-1 expression was identified in post-natal and young adult animal brains (Lu and Kipnis 2010; Yonezawa et al. 2010), as well as in normal human cortical astrocytes (Asch et al. 1986). Furthermore, cultured astrocytes are known to synthesize and secrete TSP-1 (Christopherson et al. 2005; Tran and Neary 2006; Yonezawa et al. 2010). Under normal conditions, TSP-1 causes an increase in the total number of synapses (Christopherson et al. 2005; Eroglu et al., 2009), as well as accelerates synaptogenesis (Xu et al., 2010). Astrocyte-derived TSP-1 was also shown to mediate the development of pre-synaptic plasticity in vitro (Crawford et al. 2012). Conversely, defective astrocytic TSP-1 release is known to be associated with neuronal dysfunction (Garcia et al. 2010).

Yu et al. (2008) demonstrated that retinal ganglion cell survival and neurite outgrowth were improved when co-cultured with bone marrow stromal cells (which are known to release TSP-1), a process mediated by the up-regulation of synaptophysin. On the other hand, decreasing TSP-1 expression by siRNA silencing led to a reduction in neurite outgrowth and a decrease in the expression of synaptophysin in retinal ganglion cells (Yu et al. 2008). In comparable studies, Rama Rao et al. (2013) showed that exposure of cultured astrocytes to β-amyloid peptide significantly reduced extracellular TSP-1, and that CM from these β-amyloid peptide-treated astrocytes when added to cultured neurons led to a decrease in synaptophysin protein levels.

In the present study, we observed a significant decrease in both TSP-1 synthesis and release when cultured astrocytes were exposed to a pathophysiologically relevant concentration of ammonia (1 mM), and that inhibition of c-Myc reversed the ammonia-induced reduction in TSP-1 (see below), suggesting the involvement of a transcriptional regulatory mechanism in the effect of ammonia on TSP-1. However, as noted just above, β-amyloid peptide reduced only the extracellular, but not intracellular TSP-1 levels, suggesting that the mode of action of ammonia on astrocytic TSP-1 content is different from the effect of β-amyloid peptide. We also found that exposure of cultured neurons to CM from ammonia-treated astrocytes led to a decrease in synaptophysin protein levels, while the addition of recombinant TSP-1 diminished this effect. These findings indicate that a decrease in both intra- and extracellular TSP-1 contributes to the reduction of neuronal synaptophysin observed in CHE. It is, however, possible that a diminution in astrocytic factors other than TSP-1 (e.g., secreted protein acidic and rich in cysteine, glypicans, nerve growth factor, basic fibroblast growth factor, as well as cholesterol) may also have contributed to the reduction in synaptophysin levels. In this regard, it is noteworthy that we found a reduction in hevin in ammonia-treated astrocytes, which may also have contributed to the reduction in synaptophysin.

As noted above and shown in Figure 2, TSP-1 protein levels were not changed at day 10 after 0.5 mM ammonia treatment, although a significant decline in TSP-1 mRNA level was detected at that time. Similar findings were also observed with 0.5 and 1 mM ammonia at day 5 after treatment (Figure S2b and c). The reason for the delayed decrease in TSP-1 protein compared to mRNA may result from a slow turnover rate of TSP-1 protein. Further, exposure of astrocytes to 1 mM ammonia for 10 days showed a reduction in intra- and extracellular levels of TSP-1 as well as TSP-1 mRNA (Fig. 2), and that inhibitors of ONS, c-Myc or NF-кB prevented such effects (Figs 5 and 6 and Figure S5a, b and c). These findings suggest that the reduction in TSP-1 levels in 1 mM ammonia-treated astrocytes may be due to a defective transcriptional regulation of TSP-1. On the other hand, exposure of astrocytes to 0.5 mM ammonia for 10 days showed a reduction in extra-cellular, but not intracellular, TSP-1 levels (Figs 1b and 2b), suggesting an impairment in the release of TSP-1 or that enhanced extra-cellular degradation mechanisms are likely involved.

While the mechanism by which ammonia reduces the synthesis and secretion of TSP-1 in CHE is not known, it is likely that ONS and the activation of NF-кB are involved (De Stefano et al. 2009; Tan et al. 2009; Chen et al. 2011). Markers of ONS have been identified in animal models of CHE, and in humans with CHE (Jayakumar and Norenberg 2009). In addition, activation of NF-кB was identified in ammonia-treated cultured astrocytes (Schliess et al. 2002; Sinke et al. 2008; Jayakumar et al. 2011). In this study, the decrease in intra- and extracellular TSP-1 levels when astrocytes were treated with ammonia was inhibited by antioxidants, as well as by an NF-кB inhibitor, implicating ONS and NF-кB in the reduction of astrocytic TSP-1 by ammonia. In agreement with these findings, it was shown that exposure of astrocyte cultures to cobalt chloride (an inducer of reactive oxygen species) inhibited TSP-1 mRNA expression (Chen et al. 2011), whereas the NF-кB inhibitor SN50 reduced the TSP-1 down-regulation in cultured human retinal glial cells following viral infection (Cinatl et al. 2000).

c-Myc is a transcription factor that is expressed in astrocytes in vitro and in vivo (Saljo et al. 2002; Liu et al. 2006). Over-expression of c-Myc protein was shown to cause a decrease in TSP-1 protein level by exerting a repressor effect on TSP-1 mRNA (Watnick et al. 2003). Our study also identified an increase in c-Myc protein level in ammonia-treated astrocytes (1.0 mM for 10 days), while the pharmacological inhibition of c-Myc, or silencing c-Myc by siRNA prevented the reduction in TSP-1 levels by ammonia (Figs 5 and 6). In addition to c-Myc inhibition, antioxidants and NF-кB inhibitors also prevented the reduction in TSP-1 levels by ammonia (see above), suggesting that ONS and NF-кB also contribute to the regulation of TSP-1 in CHE.

The activation of ONS, NF-кB, and c-Myc were identified in astrocytes at similar time points (8–10 days) after ammonia treatment. Since all of these factors appear to have contributed to the reduction in TSP-1 synthesis and release, it is possible that these factors acted independently to reduce TSP-1 synthesis and release. However, interaction among these factors may also have occurred. For example, ONS is known to activate NF-кB (Bowie and O'Neill 2000), NF-кB is known to stimulate c-Myc activation (Ji et al. 1994; La Rosa et al. 1994), and ONS is known to increase c-Myc expression (Höcker et al. 1998; Joseph et al. 2001). Nevertheless, the precise mechanisms by which these factors ultimately lead to a reduction in TSP-1 levels in ammonia-treated astrocytes remains to be determined.

It should also be noted that ONS and NF-кB can inhibit TGF-β (Bitzer et al., 2000; Choi et al. 2013), which is known to stimulate the production of TSP-1 in astrocytes (Cambier et al. 2005; Yonezawa et al. 2010). Furthermore, TGF-β has been shown to reduce c-Myc expression (Pietenpol et al. 1990; Warner et al. 1999). Since we found decreased levels of TGF-β in ammonia-treated astrocytes, it is possible that the ONS- and NF-кB-mediated decrease in TGF-β may have enhanced c-Myc expression resulting in reduced levels of TSP-1.

In addition to ONS, NF-кB, and c-Myc, a number of other signaling factors, including p53, specificity protein 1, activating transcription factor-1, activator protein 1, Forkhead box O 1, E2F transcription factor 1, Runx2/3, as well as TGF-β, have been implicated in the mechanism of TSP-1 regulation in other conditions (Salnikow et al. 1997; Janz et al. 2000; Li and Rossman 2001; Ji et al. 2010; Roudier et al. 2013; Shi et al. 2013). Whether these factors are also involved in the ammonia-induced reduction of TSP-1 in astrocytes, and the associated loss of synaptophysin, is not known.

Synaptophysin is an abundant integral membrane protein of pre-synaptic vesicles involved in the regulation of neurotransmitter release and synaptic plasticity (Alder et al. 1995; Janz et al. 1999). It also participates in the biogenesis and recycling of synaptic vesicles, as mice lacking synaptophysin exhibit behavioral alterations and learning deficits (Schmitt et al. 2009). In addition, the loss of synaptophysin in the hippocampus was shown to correlate with a cognitive decline in patients with Alzheimer's disease (Sze 1997), in keeping with the involvement of synaptophysin in the maintenance of synaptic integrity. Our study likewise identified a decrease in synaptophysin level when cultured neurons were exposed to CM derived from ammonia-treated astrocytes. We similarly found decreased synaptophysin levels in brains from rats with experimental CHE. Furthermore, PSD95, a post-synaptic density protein and synaptotagmin-1, a Ca2+ sensor in the plasma membrane of pre-synaptic axon terminals which is involved in the maintenance of synaptic integrity, were similarly negatively affected by ammonia.

It should be emphasized that a reduction in synaptic proteins in CHE was not caused by neuronal loss, as such loss is not a feature of CHE (Norenberg 1981). Moreover, no neuronal loss or reactive glial changes were observed in these rats. Instead, the presence of Alzheimer type II astrocytes was the major neuropathological abnormality observed in this condition (Figure S1), consistent with the concept that defective astrocytes are a major factor in the loss of synaptic integrity in CHE.

The means by which a reduction in TSP-1 decreases synaptic protein levels in CHE is not known. Studies have shown that TSP-1 binds and activates integrin alpha-V/beta-1, leading to the stability of synaptophysin (DeFreitas et al. 1995), and that a defect in this process may lead to a reduction in synaptophysin levels, with or without synaptic loss. Whether integrins alpha-V/beta-1 or other integrin receptors are involved in the reduction in synaptic proteins in CHE remains to be established.

In conclusion, a significant reduction in TSP-1 protein level was observed in ammonia-treated astrocyte cultures. Such reduction was associated with a decrease in neuronal synaptic proteins when conditioned media from ammonia-treated astrocytes were added to cultured neurons. We further observed a significant reduction in astrocytic TSP-1 and in neuronal synaptophysin protein in a rat model of CHE. Our findings suggest that dysfunctional astrocytes (a glyopathy) resulting from ammonia treatment negatively impacts neuronal synaptic integrity, which may contribute to the neurological abnormalities known to occur in patients with CHE. Targeting astrocytic TSP-1 may provide a novel therapeutic strategy for the neurological abnormalities associated with CHE.

Acknowledgments and conflict of interest disclosure

This work was supported by a Merit Review from the Department of Veterans Affairs and by the National Institutes of Health grant DK063311. KMC was supported by an NIH-NIAMS post-doctoral fellowship (7F32AR062990-02). The authors thank Alina Fernandez-Revuelta for the preparation of cell cultures. The authors declare no competing financial interests.

All experiments were conducted in compliance with the ARRIVE guidelines.

Ancillary