Glial and neuronal metabolism is intimately connected. In contrast to neurons, astrocytes can have net synthesis of glutamate as they are able to convert pyruvate to oxaloacetate via the brain’s main anaplerotic enzyme, pyruvate carboxylase (Shank et al. 1985). Net synthesis of neuronal tricarboxylic acid cycle metabolites and compounds like glutamate and GABA require the entry of a four-carbon unit. Pyruvate carboxylase in astrocytes therefore transforms pyruvate to oxaloacetate resulting after condensation with acetyl CoA in the formation of the TCA cycle intermediate citrate, which can be further converted to α-ketoglutarate. From the latter, glutamate can be formed via glutamate dehydrogenase or different transaminases, but more importantly, in neurons glutamate can emerge from glutamine after hydrolysis by phosphate-activated glutaminase. The latter pathway is part of the so-called ‘glutamine-glutamate cycle’ (Hertz 1979). Shortly, astrocytes release glutamine into the extracellular space, from where it is taken up by neurons, converted to glutamate and further, in GABAergic cells, to GABA. After release upon depolarization, glutamate is cleared from the synaptic cleft by astrocytes and converted to glutamine, which closes the cycle. Note that GABA is predominately taken up into neurons (Schousboe 2003). The significance of astrocytic clearing of glutamate as part of the ‘glutamine-glutamine-cycle’ is illustrated by the fact that glutamate/aspartate transporter and glutamate transporter which account for most of the glutamate transport, are primarily found on astroglial cells (Danbolt 2001).
As astrocytes are so intimately connected to glutamate metabolism, it seems necessary to assume that they play a major role in epilepsy and depression. Indeed, titles of recent reviews such as ‘Astrocytes have a key role in epilepsy’ (Siva 2005), ‘Astrocytes get in the act in epilepsy’ (Rogawski 2005), and ‘Gliogenesis and glial pathology in depression’ (Rajkowska and Miguel-Hidalgo 2007) corroborate this notion. Astrocytic activation and gliosis, together with neuronal loss, are the most significant histological features of hippocampal sclerosis seen in mesial TLE (Binder and Steinhauser 2006). Moreover, as astrocytic modulation of synaptic transmission between neurons is now well-recognized (Verkhratsky and Toescu 2006), an increase in glial cell number or volume may contribute to hyperexcitability of hippocampal neurons in epilepsy and a decrease to hypoexcitability of the frontal cortex in depression. Indeed, decreased amounts of glutamate, glutamine and GABA have been reported in the frontal lobe of patients with major depression (Hasler et al. 2007) in addition to a decrease in number of astrocytes (Ongur et al. 1998). This implies disturbed glial-neuronal interactions in the frontal lobe during depression, which also have been reported for this and other areas in experimental TLE (Binder and Steinhauser 2006; Melo et al. 2006). Changes in astrocyte membrane channels, receptors and transporters have all been associated with epileptogenesis and seizures (for a review, see Binder and Steinhauser 2006). As both extracellular K+ concentration and osmolarity have great impact on neural excitability, it is likely that alterations of astrocytic K+ and aquaporine water channels, detected in TLE specimens (Eid et al. 2005), contribute to epileptic hyperexcitability. Simple transitory opening of the blood-brain-barrier, which are covered by astrocytic end-feet, can under certain circumstances be sufficient for focal epileptogenesis (Seiffert et al. 2004). Newly generated glial cells can migrate into the hippocampus and contribute to enhanced seizure susceptibility (Parent et al. 2006). In some models of epilepsy, blockade of neuronal death in the hippocampus may prevent limbic brain damage, but not epileptogenesis (Halonen et al. 2001; Brandt et al. 2003). This implies that neurodegeneration alone may not lead to epilepsy. In a study by Kang et al. on pilocarpine-induced status epilepticus, microgliosis and astroglial death occurred first and preceded neuronal damage, abnormal neurotransmission of glutamate and GABA, and mossy fiber sprouting in the dentate gyrus. In addition, expressions of glutamine synthetase, glutamate dehydrogenase, and GABA transporters were down-regulated in newly generated astrocytes (Kang et al. 2006). Thus, glial reactions to status epilepticus probably add to epileptogenesis and hyperexcitability of temporal lobe structures. As astrocytes take up the great bulk of synaptic glutamate (Danbolt 2001), it is reasonable to assume that impaired glial glutamate metabolism and astrocytic-neuronal interactions play the greatest part in neurotransmitter disturbances in epilepsy and depression. In mesial temporal sclerosis down-regulation of glutamine synthetase causes a diminuation of the glutamate-glutamine cycling and accumulation of the transmitter in astrocytes and in the extracellular space (Eid et al. 2004). This has been confirmed by 1H and 13C MRS in resected human epileptic hippocampus after injection of [2-13C]glucose (Petroff et al. 2002a,b). Petroff et al. concluded that hippocampal sclerosis seems to be characterized by slow rates of glutamate-glutamine cycling, decreased glutamine content, and a relative increase of glutamate levels. The authors suggested that the low rate of glutamate-glutamine cycling may result from a failure of glial glutamate detoxification because of slow clearance from synapses and continuing excitotoxicity (Petroff et al. 2002a,b). However, in animals 13C MRS may be used to study metabolism specific to neurons and astrocytes (Sonnewald and Kondziella 2003). When [1-13C]glucose and [1,2-13C]acetate are given simultaneously, glial-neuronal interactions can be studied in the same animal because of the fact that [1,2-13C]acetate is exclusively taken-up by astrocytes, whereas most of the acetyl-CoA derived from [1-13C]glucose is metabolized in neurons (Sonnewald and Kondziella 2003). In animal studies of TLE using 13C MRS increased astrocytic activity in rats 1 day after status epilepticus (Qu et al. 2003) resulted after 2 weeks of epileptogenesis in an increased amino acid turnover in neurons (Muller et al. 2000). Chronic TLE 2 months after status epilepticus then again lead to decreased neuronal metabolism in the rat hippocampal formation with lowered glutamate levels (Melo et al. 2006; Alvestad et al. 2007). Epileptic kindling with pentylenetetrazole, a convulsant decreasing GABA activity, alters mainly metabolism of astrocytes in young and of glutamatergic neurons in old mice (Kondziella et al. 2002, 2003).
Of particular importance is the novel observation that astrocytes show Ca2+-induced release of glutamate, which directly excites surrounding neurons (Volterra and Meldolesi 2005). Thus, not only is glutamate-uptake of astrocytes in epilepsy reduced, but astrocytes are also capable of releasing glutamate through a Ca2+-dependent process, which might be involved in seizure generation (Kang et al. 2005; D’Ambrosio 2006). In a study of acute epilepsy models, Tian et al. (2005) reported that astrocytes can initiate synchronized epileptiform activity because of paroxysmal depolarizing shifts induced by glutamate release. This may indeed be an ‘astrocytic basis for epilepsy’ (Tian et al. 2005). Should further studies confirm the finding, this pathway may present a promising novel therapeutic target.