- Top of page
- Materials and methods
Thrombin levels increase in brain during ischemia and hemorrhagic episodes, and may contribute to excitotoxic neural damage. This study examined the effect of thrombin on glutamate efflux from rat cortical cultured astrocytes using 3H-d-aspartate as radiotracer. The glutamate efflux was initiated by addition of 100 mM K+ plus 1 mM ouabain (K/O) to replicate extracellular and intracellular ionic changes that occur during cerebral ischemia. Upon exposure to K/O, astrocytes swelled slowly and progressively with no evidence of volume regulation. The K/O-induced swelling was inhibited by 65% with bumetanide and 25% with BaCl2, suggesting contribution of Na+/K+/Cl− co-transporter and Kir channels. K/O-elicited 3H-d-aspartate that consisted of two phases. The first transient component of the release corresponded to 13.5% of total 3H-d-aspartate loaded. It was markedly reduced (61%) by the glutamate transporter blocker DL-threo-b-Benzyloxyaspartic acid and weakly inhibited (21%) by the volume-sensitive anion channel blocker 4-[(2-Butyl-6,7dichloro-2-cyclopentyl-2,3-dihidro-1oxo-1H-inden-5-yl)oxy] butanoic acid (DCPIB). During the second sustained phase of release, cells lost 45% of loaded of 3H-d-aspartate via a mechanism that was insensitive to DL-threo-b-Benzyloxyaspartic acid but nearly completely suppressed by DCPIB. Thrombin (5 U/mL) had only marginal effects on the first phase but strongly potentiated (more than two-fold) 3H-d-aspartate efflux in the second phase. The effect of thrombin effect was proportional to cell swelling and completely suppressed by DCPIB. Overall our data showed that under K/O swelling conditions, thrombin potently enhance glutamate release via volume-sensitive anion channel. Similar mechanisms may contribute to brain damage in neural pathologies which are associated with cell swelling, glutamate efflux and increased thrombin levels.
Swelling of brain cells, predominantly astrocytes, occurs either by a decrease in external osmolarity, or under isosmotic conditions by redistribution of ions and organic osmolytes which accumulate into the cells, generating the driving force for water influx. Isosmotic swelling occurs in brain associated with pathologies such as epilepsies, ischemia, hepatic encephalopathy and cranial trauma (Mongin and Kimelberg 2004; Pasantes-Morales and Franco 2005). Astrocytes are the brain cells which predominantly swell under these conditions, as consequence of their crucial role of clearance from the extracellular space, of potential injuring molecules such as K+, ammonium, or lactate, thus maintaining an optimal environment for neuronal function (Leis et al. 2005; Norenberg et al. 2005; Syková and Nicholson 2008). Mechanisms of uptake and/or metabolism operate specifically in astrocytes to accomplish this homeostatic function (Chen and Swanson 2003). However, during the progress of pathologies, the clearance capacities of astrocytes may be exceeded or forced to operate at maximal rate, a situation in which astrocytes not only fail to restore homeostasis, but may trigger responses that exacerbate and spread the original damage (Mongin and Kimelberg 2004; Pasantes-Morales and Franco 2005). Swelling is an early expression of this exceeded buffering capacity of astrocytes. Astrocyte swelling occurs in ischemia due to K+ and Cl− accumulation followed by osmotically-driven water. The enhanced extracellular K+ levels, which may reach concentrations of up to 80 mM, generate an ionic imbalance harmful for neuronal excitability (Walz 2000; Rossi et al. 2007; Doyle et al. 2008). Also involved in astrocytic K+ clearance is the Na+/K+ ATPase (Leis et al. 2005). If, as in ischemia, the ATPase activity is reduced or impaired, the dissipation of Na+ and K+ transmembrane gradients may further contribute to swelling and to disturb in addition, the normal operation of transporters which use the driving force of these gradients for the uptake of a variety of molecules, including the highly neurotoxic excitatory amino acid glutamate (Camacho and Massieu 2006; Doyle et al. 2008; Malarkey and Parpura 2008). This situation contributes to neuronal death by excitotoxicity, particularly at the perifocal areas of global ischemia (Won et al. 2002; Rossi et al. 2007). Under these conditions, any additional factor enhancing glutamate efflux from brain cells will aggravate the excitotoxic damage. Thrombin may be one of such factors.
Besides the role of thrombin in blood coagulation, this molecule exerts a variety of effects on brain cells, which depending on thrombin concentration may be either cytoprotective or cytotoxic (Wang and Reiser 2003). Thrombin is present in brain in low concentrations, which dramatically increase in ischemia as well as in other hemorrhagic or traumatic episodes (Xi et al. 2003; De Castro Ribeiro et al. 2006; Hua et al. 2007). Thrombin effects occur through PAR-1, PAR-3 and PAR-4 receptors, activated by a proteolytic cleavage mechanism via G protein-coupled signaling pathways (Coughlin 2000). The PAR receptors are present in astrocytes (Junge et al. 2004). The link of thrombin with glutamate efflux here investigated is based on recent reports showing that ligand activation of G protein-coupled receptors, including PAR receptors, potentiates the swelling-evoked efflux of amino acids such as taurine and glutamate, which in a variety of cells are acting as osmolytes and in brain may have the dual role of osmolytes and neurotransmitters (Fisher et al. 2008; Vázquez-Juárez et al. 2008).
In a previous study, we showed a marked effect of thrombin increasing hyposmotic-swelling induced glutamate efflux from cultured astrocytes (Ramos-Mandujano et al. 2007). The purpose of the present study is to investigate whether thrombin potentiates glutamate efflux evoked by isosmotic swelling under conditions disturbing the astrocytic capacity for K+ clearance, i.e. high extracellular K+ levels and ATPase blockade by ouabain (Leis et al. 2005). If this occurs, thrombin may exacerbate neurotoxicity and brain damage in pathologies concurrent with a disturbed K+ homeostasis. Due to the time required for the experiments, the non-metabolizable analogue of glutamate, d-aspartate, was used in this study as tracer for glutamate.
- Top of page
- Materials and methods
The present results showed swelling in cultured astrocytes under isosmotic conditions, evoked by treatment with high extracellular K+ concentrations and ouabain (K/O). The swelling time-course observed contrasts notably with that induced by 30% reduction in osmolarity in the same preparation (cultured astrocytes) (Pasantes-Morales et al. 1994; Olson et al. 1995; Cardin et al. 2003 and present results). Whereas maximal volume under the hyposmotic condition was attained almost immediately after the stimulus, it required about 30 min to be reached in K/O-treated cells. Another remarkable difference is that while hyposmotic swelling is followed by an active process of volume recovery, in the K/O-treated cells there is no evidence of volume regulation, though a plateau is reached at a certain time.
Astrocyte swelling elicited by K/O treatment is the consequence of K+ and Cl− accumulation, followed by osmotically obligated water (Ransom et al. 1996; Walz 2000). A small proportion may come also from intracellular Na+ raised by Na+/K+ ATPase blockade, which is though, counteracted by the suppressed K+ accumulation via the ATPase. Under this condition, net K+ uptake is accomplished primarily by activation of the electroneutral co-transporter NKCC. NKCC1 is the isoform expressed in cultured astrocytes and there is evidence in support of the substantial contribution of this transporter to the uptake phase of K+ clearance by astrocytes (Walz 1987; Su et al. 2002a, b; Mongin 2007). The present results showing a marked reduction in K/O-induced swelling when NKCC1 is blocked by bumetanide are in line with these previous observations. The glial inwardly rectifying K+ channels of the Kir family channels are proposed as an additional pathway for K+ accumulation and K+ buffering. The Kir4.1 isoform is expressed in cultured and in situ astrocytes and constitute the major part of the astrocytic Kir conductance (Olsen and Sontheimer 2008; Benesova et al. 2009). We found that blockade of Kir channels with barium led to a mild decrease of astrocyte swelling, suggesting a modest contribution of this mechanism to K+ uptake in cultured cortical astrocytes.
The time-course of K/O-induced swelling in cortical cultured astrocytes showed no evidence of an efficient volume regulation, which contrasts with the fast volume recovery observed after hyposmotic swelling. This is a predictable result since the typical regulatory volume decrease observed under hyposmotic conditions is accomplished to a large extent, by K+ and Cl− extrusion (Wehner 1998; Stutzin and Hoffmann 2006) which cannot occur in high extracellular K+ concentrations. The pool of organic osmolytes, including glutamate, taurine and myo-inositol (Rutledge and Kimelberg 1996; Cardin et al. 1999; Isaacks et al. 1999), is mobilized, attenuating the magnitude of swelling, and is presumably responsible for the observed swelling plateau, but appears insufficient to accomplish cell volume recovery when the K/O condition persists.
Taurine and d-aspartate efflux elicited by K/O was comparatively examined in the present study, and marked differences were found in the release pattern between the two amino acids. In contrast to the fast and large release of d-aspartate observed immediately after K/O exposure, only a marginal increase in taurine efflux was observed. Differences in the carrier properties may contribute to the difference observed, since while glutamate transporter is Na+ and K+-dependent, taurine carrier is only Na+-dependent and consequently is less influenced by changes in external K+. Results showing that prevention of taurine efflux by DCPIB, a specific blocker of the volume-sensitive anion channel (Decher et al. 2001), points to this pathway as the main route for taurine translocation.
The release of d-aspartate from astrocytes was also increased by K/O treatment, as previously reported (Rutledge and Kimelberg 1996). The efflux time-course shows two different phases: an initial phase, of fast activation and inactivation, and a second phase, of delayed and progressive efflux, detectable as long as the K/O condition persists. The pharmacological profile of these two phases revealed two different mechanisms for release. The initial phase, markedly reduced (60%) by the carrier blocker TBOA, is then likely occurring via the transport reversal, a condition favored by the dissipation of ionic gradients and depolarization. Interestingly, a fraction of 21% of d-aspartate release in this first phase was reduced by DCPIB, the volume-sensitive pathway blocker, suggesting that even small changes in cell volume as occurring in the first minutes after treatment with K/O, enhance glutamate efflux via this pathway. The simultaneous presence of TBOA and DCPIB reduced 95% the d-aspartate efflux from this first fraction, excluding mechanisms other than swelling and carrier-mediated efflux as contributors to d-aspartate release. The second phase of d-aspartate efflux showed a markedly different time-course as compared with the initial phase, and a different pattern of sensitivity to TBOA and DCPIB. While the carrier blocker had no effect, d-aspartate efflux was abolished by DCPIB, NPPB or DIDS, a result that points to swelling as the main stimulus for this release. The swelling-dependent phase of d-aspartate efflux is also evident by the effect preventing this efflux when cell swelling is reduced by treatment with bumetanide and barium. A previous study has shown a strong inhibitory effect of bumetanide on d-aspartate release elicited by high K+ concentrations (Su et al. 2002a, b), a result confirmed in the present results. All these observations clearly establish that d-aspartate efflux is elicited by both, depolarization/dissipation of the ionic gradients and cell swelling, and proceeds via two different routes, as has been previously demonstrated by Rutledge and Kimelberg (1996). The same conclusion has been reached after substantial evidence regarding glutamate efflux in a variety of experimental models of ischemia, in vitro and in vivo (Nelson et al. 2003; Phillis and O’Regan 2003; Mongin and Kimelberg 2004; Swanson et al. 2004; Kimelberg 2005). This similarity is expected since the experimental paradigm of the present study replicates intracellular and extracellular ionic changes that occur during cerebral ischemia, and has been often considered as an ischemic-like model (Rutledge and Kimelberg 1996). In contrast to glutamate, taurine efflux, which is also reported to be released in ischemia models (Phillis and O’Regan 2003; Mongin and Kimelberg 2004), seems to respond largely to swelling.
The main interest of the present study was to investigate whether thrombin potentiates glutamate efflux under ischemic-like conditions, thus potentially aggravating the risk of excitotoxicity. It should be noticed that in all experiments, d-aspartate was used as tracer for glutamate. Glutamate participates in multiple reactions related to brain energetic demands, and excitability (Dienel and Hertz 2005; Rossi et al. 2007) and in astrocytes particularly, glutamate is actively metabolized via glutamine synthetase (Isaacks et al. 1999). Therefore, the amount of glutamate released by K/O and K/O plus thrombin may be lower than that of d-aspartate. If this is too low to promote excitotoxicity remains to be demonstrated.
We showed in a previous report a marked effect of thrombin increasing glutamate efflux from cultured astrocytes swollen by hyposmolarity (Ramos-Mandujano et al. 2007) and the present study demonstrates a similar effect of thrombin in a model of isosmotic swelling, in this case elicited by intracellular K+, Na+ and Cl− intracellular accumulation. As above mentioned, in contrast to the immediate and fast increase in cell volume after a hyposmotic stimulus, swelling under the K/O condition has a temporal pattern allowing us to demonstrate that thrombin potentiation of glutamate efflux, d-aspartate in this case, occurs with a magnitude proportional to the degree of swelling. A small but significant effect of thrombin increasing d-aspartate release was observed within the first minutes after the stimulus, when only minute changes in cell volume occur. Later, the potentiation by thrombin is much higher, with a magnitude related to the extent of swelling. In full accordance with this conclusion, when swelling is prevented by bumetanide and barium, the thrombin-potentiated D-aspartate efflux was essentially suppressed.
The effect of thrombin found in the present study, in agreement with that observed on hyposmotic glutamate efflux, involved a protease-activated receptor (PAR), mainly the PAR-1 isoform, which is present in astrocytes (Wang et al. 2002; Wang and Reiser 2003; Junge et al. 2004). Thrombin activation of PAR receptors elicits a signaling pathway resulting in [Ca2+]i increase in astrocytes confirming its effect in numerous cell types. Thrombin increased [Ca2+]i in astrocytes from two main sources, extracellular Ca2+ and Ca2+ from the endoplasmic reticulum stores (Ramos-Mandujano et al. 2007). As above mentioned, thrombin potentiation of d-aspartate and taurine efflux was higher when thrombin was applied after longer times after the K/O treatment. This pattern is not due to any difference in the extent of thrombin-elicited [Ca2+]i elevation, which was found to be the same all along the experiment. It seems, in contrast, related to the degree of cell swelling which is progressively increasing. Altogether, these results show that glutamate (d-aspartate) efflux can be enhanced in swollen cells under isosmotic conditions including those replicating ischemia, provided that a threshold swelling is attained.
The thrombin-elicited increase in glutamate efflux from astrocytes might contribute to ischemic-induced neuronal death by excitotoxicity (Feustel et al. 2004; Mongin 2007) particularly since brain thrombin levels notably increase in ischemia. Other observations relate thrombin with excitotoxicity, such as the ischemia-induced up-regulation of PAR receptors (Xi et al. 2003) for which thrombin is the main substrate, or reported thrombin action increasing the efficiency of the glutamate NMDA-type receptor, which may exacerbate glutamate potential damage (Gingrich et al. 2000; Lee et al. 2007; Sharp et al. 2008). Altogether, these observations point to a possible effect of thrombin aggravating the excitotoxic damage known to occur in ischemia. In support to this possibility is the resistance to ischemic damage observed in transgenic mice defective in PAR-1, and the increased neuronal survival by treatment with PAR-1 blockers argatroban and hirudin (Kawai et al. 1996; Striggow et al. 2000, 2001; Karabiyikoglu et al. 2004).
The effects of thrombin increasing d-aspartate efflux were found abolished by preventing swelling with bumetanide and barium. Also taurine and d-aspartate release were suppressed by DCPIB. Altogether, these results point to the swelling-activated permeability pathway as the site of thrombin influence. DCPIB is a specific and potent blocker of the volume-sensitive glutamate efflux from astrocytes as shown by previous results from us and others (Abdullaev et al. 2006; Ramos-Mandujano et al. 2007). DCPIB and other Cl− channel blockers also inhibit the swelling-induced efflux of organic osmolytes (Abdullaev et al. 2006; Shennan 2008). There is still controversy on whether the swelling-sensitive Cl− channel itself is the permeability pathway for the organic osmolytes, including glutamate, a controversy raised by consistent observation of an inhibitory effect of essentially all the volume-sensitive Cl− channel blockers on the volume-sensitive efflux of organic osmolytes. If this means that the same pathways carries both Cl− and organic osmolytes or that they are so closely interconnected that blockade of one, blocks also the other one, is still uncertain. In any event, there is evidence of a strong effect of DCPIB reducing swelling-induced glutamate efflux in astrocytes and more recently DCPIB was also shown to prevent glutamate efflux evoked by middle cerebral artery occlusion-induced ischemia in adult rat. In support to the critical role played by cell swelling as a route for glutamate efflux leading to excitotoxic damage in ischemia, DCPIB showed a significant reduction of the infarct volume in this in vivo ischemia model (Zhang et al. 2008).