Introduction

Brain edema is a lethal complication of acute liver failure (ALF) which leads to increased intracranial pressure and brain herniation, frequently resulting in coma and death (Jones and Weissenborn 1997). The edema in ALF is believed to be largely cytotoxic due to swelling of astrocytes, and ammonia has been strongly implicated in its formation (Traber et al. 1987). While ammonia can directly induce astrocyte swelling in ALF, it is possible that other neural cells may also be affected by ammonia and may thus indirectly contribute to the astrocyte swelling in ALF.

Microglia are bone-marrow derived cells and represent the intrinsic immune/inflammatory cells of the CNS (Garden and Moller 2006). Microglia, when activated in response to a variety of stimuli, potentiates the nuclear translocation (activation) of various transcriptional factors, including nuclear factor-kappaB (NF-κB), which are capable of inducing genes that encode for various inflammatory factors, including NADPH-oxidase (NOX), iNOS, cytokines, among others (Garden and Moller 2006). The products of these inflammatory factors include free radicals and nitric oxide which are known to cause astrocyte swelling (Norenberg et al. 2007).

Microglia may also be the target of ammonia neurotoxicity since there has also been a growing appreciation for a role of sepsis and inflammation in the development of brain edema in ALF (Wilkinson et al. 1974; Wright et al. 2007). Consistent with this view, recent reports have documented the activation of microglia in experimental models of ALF and hyperammonemia (Jiang et al. 2009; Rodrigo et al. 2010). However, the involvement of microglia in the mechanism of astrocyte swelling/cytotoxic brain edema in ALF has not been examined. This study therefore examined the role of microglia in ammonia-induced astrocyte swelling, and investigated potential mechanisms by which microglia contribute to such swelling.

Materials and methods

Astrocyte cultures

Cultures were prepared from cortices of 1–2 day old rat pups as described previously (Ducis et al. 1990). All cultures used in our experiments were 22–30 days old.

Microglial cultures

Primary cultures of rat microglia were grown on a monolayer of astrocyte cultures prepared from brains of 1 day old pups following the method of Flanary and Streit (2004) with minor modifications. The cerebral cortices were dissected; stripped of meninges and were minced in Hank’s balanced salt solution (0.137 M NaCl, 0.2 M NaH2PO4, 0.2 M KH2PO4, 5.4 mM KCl, 5 mM glucose, 58.4 mM sucrose, 0.25 μg/ml Fungizone, and 1 × 106 U penicillin/streptomycin), to which. 0.25 % trypsin was added and incubated for 30 min at 37 °C. Trypsin reaction was stopped by adding 5 ml of DMEM containing 10 % fetal bovine serum (FBS) and 1 % penicillin/streptomycin and the suspension was triturated several times. The mixed brain cell suspension was then passed through sterile filters (130 μm, 40 μm). The final cell suspension was seeded into T75 flasks and allowed to grow for 4 days; medium changed, and incubation was continued for an additional 6 days. Microglia were harvested from cortical cultures by shaking the flasks on an orbital shaker (100 rpm) for 1 h at 37 °C and the media containing the microglia was collected. centrifuged and replated at a density of 1 × 104 in 35 mm2 plates. These cultures were at least 98 % microglia as determined by ED1 immunoreactivity.

Ammonia treatment of microglia

Microglial cultures were treated with 5 mM NH4Cl for 3, 6, 12 and 24 h, as brain levels as high as 5 mM were found in experimental models of ALF (Mans et al. 1979, 1994; Swain et al. 1992; Rama Rao et al. 2010; Jayakumar et al. 2011; see also Discussion). Conditioned media (CM) from ammonia-treated microglia was added to primary cultures of astrocytes and cell volume determined 24 h later. We previously established that the amount of ammonia retained in the treatment media after 3 h was < 1 % of exogenously added ammonia. Therefore, it was not necessary to wash-out the treatment media before transferring the CM to astrocytes.

Cell volume determination

Cell volume (intracellular water space) in cultured astrocytes was determined using 3-O-methyl-[3H]-glucose (OMG, Sigma, St Louis, MO) by the method of Kletzien et al. (1975) as modified by us (Norenberg et al. 1991).

Measurement of reactive oxygen species (ROS)

ROS production in CM from ammonia-treated microglia was measured using the fluorescent probe Rhodamine123 Grzelak et al. (2001) as modified by us (Jayakumar et al. 2012).

Measurement of nitric oxide (NO)

Levels of nitrite in the CM (a measure of NO production) from ammonia-treated microglia was measured by the Griess reaction (Grisham et al. 1996) as modified by us (Jayakumar et al. 2012).

Statistical analysis

The data were expressed as means ± SEM and analyzed by ANOVA followed by Neuman-Keuls multiple-range test. A p < 0.05 was considered statistically significant.

Results

Effect of CM from ammonia-treated microglia on astrocyte cell volume

CM from microglia treated with ammonia for different time periods (3–24 h) was added to cultured astrocytes and cell volume determined (24 h). Addition of CM from ammonia-treated microglia for 12 and 24 h to astrocytes caused significant cell swelling (34 and 41 %, respectively, p < 0.05), whereas CM from microglia treated with ammonia for 3 and 6 h had no effect (Fig. 1).

Fig. 1
figure 1

Time-course of astrocyte swelling following the addition of conditioned media (CM) from ammonia-treated microglia. The addition of CM from ammonia-treated (5 mM NH4Cl) microglia to cultured astrocytes results in cell swelling at 12 and 24 h. * vs. Cont CM (control conditioned media); NH 4 ammonium chloride

Effect of simultaneous treatment of astrocytes with ammonia and CM from ammonia-treated microglia on astrocyte cell volume

Previous studies have consistently shown that the addition of ammonia to cultured astrocytes results in cell swelling (Norenberg et al. 2007). We therefore examined any potential additive/synergistic effects on astrocyte swelling by CM from microglia on ammonia-treated astrocytes. Accordingly, astrocytes were simultaneously treated with 5 mM ammonia along with CM from ammonia-exposed microglia (24 h) and cell volume was determined. Such treatment resulted in a marked synergistic effect on astrocyte swelling (Fig. 2).

Fig. 2
figure 2

Addition of CM from ammonia-treated microglia, or ammonia alone, cause astrocyte swelling. A synergistic effect on cell swelling was observed when these two treatments were combined. * vs. Con CM (control conditioned media); NH 4 ammonium chloride

Effect of ammonia on the release of reactive oxygen and nitrogen species (RONS) in microglia

A time course on the effect the ammonia on the release of RONS in microglia showed a 100–150 % increase in RONS release at 24 h (Fig. 3), whereas at 12 h such release was only 45–55 %. To establish whether RONS produced by ammonia-treated microglia contribute to astrocyte swelling, we added Tempol (25 μM) and uric acid (500 μM), scavengers of superoxide and peroxynitrite, respectively, to the CM from ammonia-treated microglia and astrocyte cell volume determined (24 h). Addition of CM from ammonia-treated microglia containing Tempol or uric acid to astrocytes resulted in a marked reduction in cell swelling (75–80 %, p < 0.001), suggesting the involvement of RONS in the mechanism of astrocyte swelling caused by CM from ammonia-treated microglia.

Fig. 3
figure 3

Effect of ammonia (24 h) on the release of RONS (reactive oxygen/nitrogen species) from microglia. * vs. Con (control); NO nitric oxide; NH 4 ammonium chloride

Discussion

This study demonstrates that CM from ammonia-treated microglia results in astrocyte swelling suggesting that microglia may contribute to the astrocyte swelling in the setting of ALF. Further, the marked synergism in cell swelling caused by the simultaneous treatment of astrocyte cultures with ammonia and CM from microglia exposed to ammonia indicates that microglia potentiates the ammonia-induced astrocyte swelling. Additionally, ammonia was shown to cause a marked increase in the release of RONS by microglia and that the addition of scavengers of RONS to the microglial CM resulted in a significantly lesser degree of cell swelling when the CM was then added to astrocytes, strongly suggesting that RONS represent one means by which microglia contribute to the ammonia-induced astrocyte swelling. Our results on the production of RONS in microglia in response of ammonia are in agreement with a recent report by Zemtsova et al. (2011).

Recent studies have documented that microglia are activated in ALF as shown by increased ED1 immunoperoxidase staining (Jiang et al. 2009), and increased expression of MHC-II molecules in a rat model of hyperammonemia (Rodrigo et al. 2010). Additionally, it was shown that minocycline, an inhibitor of microglial activation, significantly attenuated (60 %) the amount of brain edema in ALF (Jiang et al. 2009). However, factors responsible for microglial activation and the precise contribution of such activation to the pathogenesis of ALF are not clear. Additionally, the pathogenetic role of microglia in the development of cytotoxic brain edema/astrocyte swelling in ALF is not known.

Once activated, microglia are known to release inflammatory cytokines and other inflammatory mediators, including free radicals, prostaglandins and arachidonic acid (Garden and Moller 2006), all of which have been shown to cause astrocyte swelling in other conditions (Norenberg et al. 2007). Consistent with these views, our results show that microglia produce RONS following treatment with ammonia and that scavenging RONS in microglia resulted in a marked reduction in astrocyte swelling caused by CM from ammonia-treated microglia.

While the means by which microglia can produce RONS is not completely known, activated microglia are known to increase the nuclear translocation (activation) of the transcriptional factor NF-κB, and such activation is capable of inducing genes that encode for various inflammatory factors including NADPH-oxidase (NOX), iNOS and cytokines, among others (Garden and Moller 2006). We indeed found a marked increase in the protein expression of NOX and iNOS in microglia following treatment with pathophysiological concentration (5 mM NH4Cl, unpublished observations). These data, in conjunction with our observation that Tempol and uric acid significantly mitigated the astrocyte swelling caused by CM from ammonia-treated microglia strongly support a role of oxidative and nitrosative stress in the mechanism by which microglia contribute to astrocyte swelling.

The current study employed 5 mM ammonia (NH4Cl) to examine the role microglia in the mechanism of astrocyte swelling. The selection of 5 mM ammonia was based on the results of studies in experimental models of ALF which reported brain ammonia concentrations as high as 5 mM (Mans et al. 1979, 1994; Swain et al. 1992; Rama Rao et al. 2010; Jayakumar et al. 2011). Such high ammonia levels in brains of animals with ALF were observed employing reliable methods: in situ freezing technique (Pontén et al. 1973), as well as by allowing the head to drop directly into isopentane cooled with liquid nitrogen, thus avoiding any artifactual rise in brain ammonia due to postmortem changes in brain.

It should be emphasized that the use of 5 mM NH4Cl in astrocyte cultures significantly lessens the experimental period required to achieve a similar degree of astrocyte swelling as that observed after the use of lower ammonia concentrations (0.5–2 mM) for 3–5 days. Thus, treatment of astrocyte cultures with 0.5 mM NH4Cl for 5 days caused a 40 % increase in cell volume (control 4.3 μl/mg protein vs. NH +4 6.16 μl/mg protein), while a similar increase in cell volume was achieved by higher concentrations of ammonia for lesser time points (control 4.3 ml/mg protein; 2 mM treatment for 3 days resulted in an increase in cell volume to 6.32 μl/mg protein, while 5 mM treatment for 1 day resulted in a cell volume of 6.45 μl/mg protein). A similar outcome was also observed when evaluating morphological changes in ammonia-treated cultured astrocytes (Gregorios et al. 1985).

In summary, our study demonstrates that conditioned media (CM) from microglia treated with ammonia when added to cultured astrocytes result in their swelling, suggesting that microglia contribute to astrocyte swelling/brain edema in ALF. We also found a marked potentiation of cell swelling in astrocytes which were co-treated with ammonia and CM from ammonia-treated microglia. These findings are in concert with recent reports indicating that microglia are activated in experimental models of ALF and hyperammonemia. Our findings showing that ammonia increases the release of RONS in microglia and that antioxidants mitigate astrocyte swelling caused by CM from microglia, strongly suggest that ONS represents one means by which microglia contribute to astrocyte swelling. Targeting microglia represents a promising therapeutic strategy for mitigating the brain edema associated with ALF.