Acute astrocyte loss following CNS insult

Astrocyte swelling is a prominent feature of both cerebral ischemia (Jenkins et al., 1982) and TBI (Castejon, 1998) and is considered to be an "exaggerated extension of normal astrocyte function'' (Kimelberg, 1992). Astrocyte swelling is likely a prominent component of raised ICP after both ischemia and TBI and resulting in part from alterations in ion-transport mechanisms that leads to osmotically obligated water entry (Kimelberg et al., 1990). Cortical biopsies of human TBI complicated with subdural hematoma have shown severe astrocytic edema, dissociation of the perivascular astrocyte endfeet from the capillary basement membrane, and disruption of inter-astrocyte gap junctions (Castejon, 1998). Electron microscopy of brain tissue from human cerebral contusion patients has revealed evidence of astrocytic swelling as early as 3 h after injury, persisting for as long as 3 days (Bullock et al., 1991). It is important to consider that human tissue is essentially never available for analysis during the first hours after insult. Thus, there is a distinct possibility that human astrocyte swelling may occur very rapidly after injury as in animal models of ischemia and TBI. In addition to cell swelling and raised intracranial pressure, the alterations in astrocyte ionic homeo-stasis can likely initiate cell damage and death.

There has been a dogma that astrocytes are highly resistant to hypoxic/ischemic insults compared to neurons (Lukaszevicz et al., 2002). This dogma has been challenged by early suggestions and by later experiments demonstrating that as-trocytes are indeed vulnerable to ischemic conditions. Over 20 years ago, Plum (Plum, 1983) suggested that the degree of infarction from focal ischemia might depend upon the viability of as-trocytes. This proved to be an insightful prediction and subsequent studies indicate that this suggestion applies to TBI as well as cerebral ischemia. A number of in vitro studies have demonstrated that astrocytes are indeed vulnerable to ischemia. As-trocytes exposed to hypoxia alone can require up to 24 h exposure to produce widespread death (Yu et al., 1989; Sochocka et al., 1994) even in the absence of glucose (Kelleher et al., 1993). However, combining hypoxia with acidosis, a more realistic model, accelerates astrocyte death to just a few hours (Swanson et al., 1997). Recent in vitro studies have examined this issue utilizing a novel experimental condition in which cultured as-trocytes were exposed to hypoxic, acidic, ion-shifted Ringers (HAIR) solution that mimics the environmental characteristics of ischemia and TBI (Bondarenko and Chesler, 2001a). Their microenvironment included glucose and thus mimicked incomplete ischemia as well as severe TBI. They found that as little as 15-20 min exposure to their HAIR solution resulted in death of 60% of the astrocytes (Bondarenko and Chesler, 2001a). The same group subsequently reported that the NHE-1 inhibitor amiloride protected astrocytes exposed to the HAIR protocol. Furthermore, blocking the Na + /Ca2+ exchanger with benzamil or KB-R7943, which preferentially inhibits reverse Na + / Ca2+ exchange, protected astrocytes from the HAIR protocol (Bondarenko and Chesler, 2001b; Bondarenko et al., 2005). These studies provide evidence that astrocytic intracellular aci-dosis activates NHE-1 leading to Na+ entry and the reversal of the Na + /Ca2+ exchanger, triggering astrocyte cell loss after ischemia.

Recent in vivo studies provide further compelling evidence of early astrocyte loss following cerebral ischemia. Liu et al. (1999) reported a 50% decrease in GFAP mRNA and protein in the ischemic core within 5 h after permanent middle cerebral artery occlusion, and almost complete absence of GFAP mRNA by 12 h. The loss of GFAP mRNA preceded the loss of the mRNA for GLUT3, a neuron-specific glucose transporter. Immunohisto-chemical experiments using other astrocyte markers including S100beta, GSTYb1, and vimentin produced similar results. Interestingly, mRNA began to increase in the peri-infarct area by 3 h and peaked at 48 h after ischemia, suggestive of reactive astrocyte proliferation in the ischemic penumbra. Using a rat subdural hematoma-induced ischemia model, Jiang et al. (2000) reported complete loss of GFAP immunostaining at 4 h after the insult in the ischemic core. Protoplasmic astrocytes appear to be more vulnerable to ischemia than fibrous as-trocytes. For example, following permanent middle cerebral artery occlusion in mice, Lukaszevicz et al. (2002) found that protoplasmic astrocytes (type 1, accounting for most of the astrocytes in gray matter) lost GFAP immunolabeling within minutes in the ischemic core. In contrast, fibrous astrocytes (type 2, found predominantly in white matter) displayed a transient hypertrophy with no conspicuous cell death (Lukaszevicz et al., 2002). These studies provide compelling evidence that ischemia is associated with an early loss of astrocytes.

Glutamate excitotoxicity can lead to both ne-crotic cell death and apoptotic programed cell death following application of glutamate agonists (Du et al., 1997; Nicotera et al., 1997) ischemia (Namura et al., 1998; Graham and Chen, 2001)

and TBI (Rink et al., 1995; Newcomb et al., 1999). Most of these studies have focused on apoptotic cell death of neurons. Astrocytes are also vulnerable to apoptotic cell death that may contribute to the pathogenesis of many acute and chronic neurodegenerative disorders (Takuma et al., 2004). Hypoxia/ischemia induces mitochondrial depolarization in astrocytes (Smith et al., 2003), that triggers caspase-dependent cell death pathways and by poly(ADP-ribose) polymerase-1 cell death pathway (Giffard and Swanson, 2005). Astrocyte apoptosis can be detected by caspase activation by a variety of in vitro insults including staur-osporin (Keane et al., 1997), cycloheximide (Tsuc-hida et al., 2002), and ischemic conditions (Yu et al., 2001; Gabryel et al., 2002). Activation of apoptotic cascades in astrocytes has also been reported after in vivo TBI. For example, accumulation of active caspase-3 was observed for 5 days after controlled cortical impact TBI in rats in both neurons and astrocytes but, interestingly with a higher proportion of activated caspase-3 in astrocytes than in neurons (Johnson et al., 2005).

TBI and ischemia share a number of patho-physiological similarities including excessive glutamate release and selective neuronal cell loss. The role of astrocytes in uptake of glutamate may have particular importance to the traumatically injured brain where large fluxes of extracellular glutamate (Faden et al., 1989; Katayama et al., 1990) likely contribute to acute excitotoxic processes (Olney and Ho, 1970; Hayes et al., 1988). Specifically, early impairment of astrocyte function after TBI may compromise maintenance of homeostasis in the extracellular microenvironment and disrupt critical neuronal-glia interactions. Zhao et al. (2003) demonstrated selective loss of GFAP and glutamine synthetase immunoreactivity in rat brains as early as 1 h after lateral fluid percussion TBI. The loss of immunoreactivity was limited to regions directly adjacent to areas of selective neuronal vulnerability, the hippocampus CA3. The loss of immunoreactivity for astrocytes markers preceded fluoro-jade detection of neuronal degeneration by several hours suggesting that loss of supporting astrocytes may contribute to subsequent neuronal cell loss.

Unraveling the mechanisms of acute astrocyte damage with in vitro mechanical injury

It is clear from the studies described above that in vivo experimental TBI or ischemia produces rapid and profound damage to astrocytes. Moreover, several in vitro models of ischemia also show acute astrocyte pathology. Additionally, several of the astrocytic intracellular events occurring acutely after TBI have been characterized in a series of in vitro studies using a mechanical injury model developed by Ellis et al. (1995). In this model, cells are grown on a membrane that is subjected to a rapid mechanical deformation that approximates the tensile strain or stretch that is a key component in the acceleration/deceleration type of closed human TBI (Meaney et al., 1995) and in vivo brain injury (Margulies et al., 1990). This model has been validated by demonstrating that in vitro strain injury produces many of the post-traumatic responses observed with in vivo injury models including intracellular lesions to the mitochondria, Golgi, and cytoskeletal elements in astrocytes and neurons (Dietrich et al., 1994; Ellis et al., 1995; McKinney et al., 1996), increases in astrocyte and neuronal permeability (Ellis et al., 1995; Weber et al., 1999; Willoughby et al., 2004), activates phospholipases (Lamb et al., 1997; Floyd et al., 2001), and induces free radical and isoprostane formation (McKinney et al., 1996; Lamb et al., 1997; Hoffman et al., 2000) as well as release of the injury markers S-100p and neuron-specific enolase (NSE; Slemmer et al., 2002, 2004). Strain-injury of cultured neurons also alters the Mg2+ block of the NMDA receptor (Zhang et al., 1996) and produces a novel stretch-induced delayed neuronal depolarization, which may be related to transient neuronal dysfunction observed in vivo (Hamm et al., 1994; Tavalin et al., 1995). Strain-injury in neurons also increases activation of the AMPA receptor by decreasing AMPA receptor desensitization (Goforth et al., 1999). Thus, this well characterized mechanical injury model recapitulates many aspects of in vivo TBI and is a useful tool to examine injury-induced changes in acute astrocyte pathology.

A critical element in astrocytic calcium signaling is regulation of intracellular free Ca2 + .

Mechanical strain injury to astrocytes produces a rapid rise in [Ca2 + ]; that returns to near-basal levels within 15 min. However, application of either glutamate or the metabotropic glutamate group I (mGluRI) agonist trans-1-aminocyclopentyl-1-3-dicarboxylic acid (tACPD) after injury results in a significantly blunted rise in [Ca2 + ]; for up to 24 h post-injury, suggesting that the calcium signaling machinery is disturbed by injury (Rzigalinski et al., 1998). Alterations in intracellular calcium dynamics after mechanical strain injury may be related to alterations in inositol trisphosphate (IP3) signaling as strain injury to astrocytes causes a significant increase in IP3 at 5, 15, and 30min and at 24 and 48 h post-injury (Floyd et al., 2001). Interestingly, the injury-induced increases in IP3 at 15 and 30 min post-injury correspond to time points when [Ca2+]; had returned to near normal levels; however, high IP3 should produce increased [Ca2 + ]; if Ca2+ signaling were functioning normally. Instead the opposite was observed — elevated IP3 at time points when intracellular calcium has returned to basal levels which indicates that IP3 may be uncoupled from its target, the intracellular Ca2+ store. Additionally, antagonism of mGluRIs and inhibition of PLC attenuated injury-induced uncoupling of IP3-mediated Ca2+ signaling, reduced astrocyte damage (Floyd et al., 2001, 2004), and reduced injury-induced depletion of intracellular calcium stores (Chen et al., 2004). Thus, mechanical strain injury to astrocytes causes acute elevations in intracellular calcium and disruption of IP3-mediated intracellular calcium signaling.

Mechanical strain injury to astrocytes also significantly increases intracellular sodium (Floyd et al., 2005). Mild and moderate strain injury produced a rapid rise in intracellular sodium that returned to near-basal levels by 20-25 min postinjury while severe injury produced increased in-tracellular sodium for at least 50 min (the duration of the experiment). These elevations in intracellu-lar sodium were similar to those induced by exogenous glutamate application and were reduced by pre-injury application of the sodium-dependent glutamate uptake inhibitor TBOA, indicating that sodium which accompanies glutamate uptake contributes to injury-induced increases in intra-cellular sodium. Comparison of the time course of injury-induced elevations in intracellular sodium and calcium indicated that there is significant overlap in the first 30 min post-injury, which raises the possibility that sodium/calcium exchange may be involved in the acute astrocyte pathology after injury. The magnitude and direction of calcium flux of the sodium/calcium exchanger is dependent, in part, on the Na+ electrochemical gradient, and under normal conditions the sodium/calcium exchanger operates in the "forward" or calcium-efflux mode to extrude calcium. Importantly, manipulations in uninjured astrocytes that raise intracellular sodium have been shown to increase intracellular calcium by "reversal" of the sodium/ calcium exchanger or calcium-influx operation (Goldman et al., 1994). Thus, it was predicted that the elevated intracellular sodium observed after mechanical strain injury may be sufficient to reverse the sodium/calcium exchanger and cause an influx of calcium. It was found that pre-injury blockade of the calcium-influx (reversed) mode with the compound KB-R7943 significantly reduced intracellular calcium in the moderate and high levels of injury (but not at the low levels of injury), the magnitudes of injury with the largest and more sustained elevations in intracellular sodium (Floyd et al., 2005). Additionally, this severity of injury-dependent reversal of the sodium/ calcium exchanger was predicted by mathematical modeling of the reversal potential of the sodium/ calcium exchanger using experimental values of intercellular sodium and calcium concentrations after mechanical strain injury. Taken together, these data suggest that increased intracellular sodium can cause a reversal of the sodium/calcium exchanger to cause influx of calcium into the cells, and that this only occurs with moderate or severe injury.

Mechanical strain injury to astrocytes causes release of ATP (Ahmed et al., 2000; Neary et al., 2005). ATP release following mechanical injury not only has consequences for cellular energetics (Bambrick et al., 2004), but also may alter intercellular calcium signaling via P2 purinergic receptors. Cortical astrocytes express both the ionotropic P2X receptors and the G-protein coupled P2Y receptors (Lenz et al., 2000; Weisman et al., 2005); however, P2Y expression is much more prevalent in astrocytes than P2X (Illes and Ribeiro, 2004). Nucleotides released from injured cells can activate either the P2X or P2Y and cause an increase in intracellular calcium (Illes and Ribeiro, 2004) as well as activation of mitogen-activated protein kinase (MAPK) pathways (Gendron et al., 2003). Importantly, activation of P2X receptors on astrocytes is linked to neurode-generative events and astrogliosis (for review, see Neary et al., 1996); yet, activation of P2Y receptors on astrocytes is protective. For example, activation of P2Y2 increases Bcl-2 family genes (Chorna et al., 2004), and activates extracellular signal-regulated protein kinase (ERK) which has been linked to the protective activation of CREB (Neary et al., 1996; Chorna et al., 2004). In an eloquent series of studies, Neary and colleagues have detailed the effects of mechanical strain injury to astrocytes on P2 receptor activation and found that injury induces a rapid release of ATP which, via P2 receptors, activates ERK (Neary et al., 2003) and Akt (Neary et al., 2005); increases expression and release of a protein that induces synapse formation, thrombospondin-1 (Tran and Neary, 2006); and also phosphorylates GSK3b at the Ser-9 location, thereby inhibiting activity (Neary and Kang, 2006). Thus, release of ATP from injured astrocytes can act not only as a stimulus for gliosis, but also as a promoter of cell survival and synaptic plasticity. Further work is needed in this exciting area to better elucidate the complex relationship between P2 activation, gliosis, and cell survival.

Mechanical strain injury to astrocytes disrupts many of the critical features of intra- and intercellular calcium signaling in astrocytes. The known changes in astrocyte signaling after mechanical injury are summarized schematically in Fig. 1. Strain injury to astrocytes causes a rapid increase in in-tracellular calcium and sodium that is dependent on the magnitude of injury (Floyd et al., 2005). There are several mechanisms by which calcium is elevated following mechanical injury. Excessive glutamate following injury agonizes mGluRs to cause IP3-mediated calcium release (Floyd et al., 2004). Theoretically, elevated glutamate following injury could also activate the AMPA ionotropic glutamate channel as has been shown in neurons (Goforth et al., 2004). Although the effects of

Fig. 1. Pathological events known to occur in astrocytes after mechanical strain injury. Mechanical strain injury to astrocytes results in a rapid rise in intracellular calcium and sodium. Sources of injury-induced increases in intracellular calcium include activation of the mGluRs, release of calcium from IP3-mediated intracellular calcium stores, reversal of the sodium/calcium exchanger, and possibly AMPA receptor activation. Increases in intracellular sodium after mechanical injury are largely due to sodium-dependent uptake of excessive glutamate. Injured astrocytes also release ATP. ATP could activate P2YRs and further increase intracellular calcium but also could initiate cell survival pathways such as Bcl-2. Astrocytes are coupled by gap junctions, and small molecules such as IP3 or Ca2 + could travel between cells after injury. Propidium iodide, a marker for dead or damaged cells, could enter the astrocyte through disruptions in the cell membrane and intercalate into the DNA. (See text for further details.) Abbreviations: protein kinase B/Akt (Akt): (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)propionate (AMPA), extracellular signal-regulated protein kinase (ERK), glutamate transporter (GLT-1), inositol trisphosphate (IP3), metabotropic glutamate receptor (mGluR), sodium/calcium exchanger (NCX), purinergic 2Y receptor (P2YR), phospholipase C (PLC).

Fig. 1. Pathological events known to occur in astrocytes after mechanical strain injury. Mechanical strain injury to astrocytes results in a rapid rise in intracellular calcium and sodium. Sources of injury-induced increases in intracellular calcium include activation of the mGluRs, release of calcium from IP3-mediated intracellular calcium stores, reversal of the sodium/calcium exchanger, and possibly AMPA receptor activation. Increases in intracellular sodium after mechanical injury are largely due to sodium-dependent uptake of excessive glutamate. Injured astrocytes also release ATP. ATP could activate P2YRs and further increase intracellular calcium but also could initiate cell survival pathways such as Bcl-2. Astrocytes are coupled by gap junctions, and small molecules such as IP3 or Ca2 + could travel between cells after injury. Propidium iodide, a marker for dead or damaged cells, could enter the astrocyte through disruptions in the cell membrane and intercalate into the DNA. (See text for further details.) Abbreviations: protein kinase B/Akt (Akt): (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)propionate (AMPA), extracellular signal-regulated protein kinase (ERK), glutamate transporter (GLT-1), inositol trisphosphate (IP3), metabotropic glutamate receptor (mGluR), sodium/calcium exchanger (NCX), purinergic 2Y receptor (P2YR), phospholipase C (PLC).

mechanical injury on the subunit composition in astrocytes has not been examined, yet, if the GluR2 subunit is present, injury induced activation would likely contribute to increase intracellular Na + . However, if the GluR2 subunit were absent, the channel would also be permeable to Ca2+ (Frandsen and Schousboe, 2003). Activation of either configuration of the AMPA receptor would likely contribute to the pathophysiology of astrocyte injury, as both increased intracellular calcium and sodium are intimately involved. Excessive glutamate from mechanical injury also affects the astrocyte by activation of sodium-dependent glutamate uptake that significantly elevates intracellular sodium. At moderate and severe levels of injury, the elevation in intracellular sodium is sufficient to cause reversal of the sodium/calcium exchanger and produce calcium

Fig. 2. Effect of mechanical injury on functional coupling in astrocytes. Confluent primary astrocyte cultures we mechanically injured and then functional coupling was analyzed using fluorescence recovery after photobleach (FRAP). Panel (A) shows representative micrographs of uninjured (left column) or moderately injured (right column) astrocytes before, immediately after, and 15 s after photobleach. Regions of interest (ROIs) were drawn around four cells per field. The ROIs outlined in red and green were bleached but the ROIs in yellow and blue were not. A representative trace of a mildly injured astrocyte at 24 h post-injury is shown in panel (B). Panel (C) shows quantification of the maximal recovery at various post-injury times. Mild injury increased and moderate injury decreased functional coupling in astrocytes (n = 3 separate experiments).

Uninjured

Moderate Strain

Before Bleach

Bleach

After Bleach

Uninjured

Moderate Strain

Before Bleach

Bleach

After Bleach

Was this article helpful?

0 0

Post a comment