Glial cells play a major role in the central nervous system (CNS). Comprising different cell types, glial cells regulate many physiological processes: Astrocytes, participating in the tripartite synapse, remove excess levels of glutamate, thereby enabling proper glutamatergic neurotransmission and preventing extrasynaptic N-methyl-D-aspartate (NMDA) receptor-induced excitotoxicity [1, 2]. Microglia are CNS resident immune cells that regulate the immune response to threatening stimuli [3]. Oligodendrocytes ensheath axons to protect and support them and to enable rapid information transfer inside neuronal networks [4].

The excitatory amino acid glutamate plays an important role in the homeostasis of the brain [5], thus its regulation by glial transporters is particularly relevant in diseases. In the CNS, neuronal damage can occur through increased extracellular glutamate, which then triggers cell death-inducing cascade by the activation of extrasynaptic NMDA receptors, giving rise to the so-called mechanism of excitotoxicity [6, 7]. This has strong implications for various diseases of the CNS such as Alzheimer’s disease or Huntington disease [8]. To regulate glutamate homeostasis, astroglial cell express glutamate exporters and glutamate importers. Through this interplay the level of extracellular glutamate can be regulated. The major glutamate exporter in the CNS is the system x${}_c$${}^{-}$. It is mainly expressed in the astroglia [9], and during the development of the rat, the expressions of the system x${}_c$${}^{-}$ subunits ‘xCT’ or ‘SLC7A11’ (light chain) and ‘4F2hc’ (heavy chain) increase until at least 3 months of age [10]. With aging, system x${}_c$${}^{-}$ is overexpressed in rats [11]. These findings suggest that with the growing nervous systems, the need to balance oxidative stress and glutamate homeostastis remains an important task that is fulfilled by system x${}_c$${}^{-}$.

Astroglial glutamate importers such as the excitatory amino acid transporters 1 and 2 (EAAT1/2) counteract the extrusion of glutamate. EAATs are sodium-dependent glutamate transporters that are responsible for the removal of extracellular glutamate by importing glutamate into the astrocytes’ cytosol [12]. The examination of both glutamate-exporting and -importing transporters provides important insights into the overall astroglial contribution to glutamate homeostasis and its pathophysiological relevance. The detailed molecular aspects and the physiology of glutamate transporters are beyond the scope of this plasticity-focused review. We refer the reader to excellent reviews that comprehend this knowledge [13, 14, 15, 16].

Oxidative stress can strongly interfere with physiologic function. System x${}_c$${}^{-}$ imports cystine to fuel the antioxidant response of the cells [17] in exchange for glutamate, which is exported to the extracellular fluid. System x${}_c$${}^{-}$ activity in astrocytes is not only regulated by glutamate but also by its second substrate, cystine, since the intracellular glutathione (GSH) levels can regulate the expression of system x${}_c$${}^{-}$ [18]. However, while system x${}_c$${}^{-}$ contributes to the antioxidative response [17], its deficiency in system x${}_c$${}^{-}$ does not induce oxidative stress by itself, but may disrupt glutamate homeostasis [19].

Neurons in the CNS are susceptible to oxidative stress due to their high oxygen consumption during metabolism and the high abundance of lipids, which represent targets for oxygen radicals [20]. As some neurons are especially vulnerable to reactive oxygen species, it is important for the brain to maintain an antioxidative response [21]. The glutamate-cystine antiporter system x${}_c$${}^{-}$ is involved in the antioxidative response, which has been also strongly investigated in the context of cancer biology [22]. On a molecular level, the import of cystine through system x${}_c$${}^{-}$ fuels the formation of antioxidant glutathione. Inhibiting cystine import through system x${}_c$${}^{-}$ inhibitors such as erastin promotes ferroptosis, an iron-dependent form of cell death triggered by missing antioxidants [23]. In the field of neurooncology, system x${}_c$${}^{-}$ thus represents a potential therapeutic target, and tumor cell death during chemotherapy could potentially be boosted by add-on therapy with system x${}_c$${}^{-}$ inhibitors [24, 25].

In this review, we will focus on the issue of whether alterations in glutamate transporter activity are a consequence of the disease pathology or are part of the pathogenesis in glioma, tumor-associated epilepsy, Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis, or multiple sclerosis. We further address the benefits of boosting/inhibiting these proteins and we are going to approach the question of how to selectively target glial cells, because of their various responsabilities in the nervous system.

There is no final consensus yet on the expression levels of the cystine-glutamate antiporter system x${}_c$${}^{-}$ in different CNS cell types. Soria et al. [26] reported co-expression of system x${}_c$${}^{-}$ with markers for neurons, as it was shown in other studies [10, 27], and for oligodendrocytes. Recent findings questioned these data by claiming that system x${}_c$${}^{-}$ is expressed mainly in astrocytes, where it is coexpressed with the sodium-dependent glutamate transporter EAAT1 [9]. System x${}_c$${}^{-}$ expression was not found in neurons, oligodendrocytes, or microglia. In line with this, Mesci et al. [28], could not detect system x${}_c$${}^{-}$ in motor neurons. In a systematic approach to evaluate antibody specificity and technical specifications, it became evident that the antibody production process and the staining protocols can heavily impact if system x${}_c$${}^{-}$ can be detected [29]. Further studies are required to confirm in which cell system x${}_c$${}^{-}$ may be expressed. While we wanted readers to bear in mind this uncertainty when evaluating study results, system x${}_c$${}^{-}$ modulation affects every cell type that is in contact with system x${}_c$${}^{-}$ modulated extracellular glutamate level regardless of their individual expression.

Although system x${}_c$${}^{-}$ expression in neurons is uncertain, its effect on neuronal activity is undebated. Around the time when Sato et al. [30] used molecular-biological techniques to clone human system x${}_c$${}^{-}$ and to describe system x${}_c$${}^{-}$ expression across the brain with in situ hybridization [31], the first electrophysiological evidence for system x${}_c$${}^{-}$′s impact on vesicular release was obtained: in acute brain slices containing the medial prefrontal cortex, bath application of cystine reduced the frequency of miniature excitatory postsynaptic currents (mEPSCs) that are the electrophysiological expression of the readily releasable pool (RRP) of synaptic vesicles [32]. The effect was suppressed by system x${}_c$${}^{-}$ inhibitor (S)-4-carboxyphenylglycine [33]. (S)-4-carboxyphenylglycine is a rather unspecific blocker for system x${}_c$${}^{-}$ and also acts as an antagonist for the group 1 metabotropic glutamate receptors [34]. This weak receptor specificity of (S)-4-carboxyphenylglycine could interfere with the inhibition of system x${}_c$${}^{-}$. There is no indication that the impact of (S)-4-carboxyphenylglycine on system x${}_c$${}^{-}$ than its modulation of metabotropic glutamate receptors, since data reveal no direct connection between the RRP size and group I mGluR activity, at least in the functional context of synaptic long-term depression [35].

In electrophysiological recordings, mEPSCs can only be detected when action potential-driven release is blocked, e.g., with the voltage-gated sodium channel blocker tetrodotoxin (TTX). Spontaneous EPSCs (measured in the absence of TTX) are not decreased by cystine [33]. Cystine’s dampening effect on mEPSCs is mediated through group II metabotropic glutamate receptors that balance glutamate release by suppressing synaptic firing upon above-threshold extracellular glutamate detection [36]. The increase in extracellular cystine could elevate extracellular glutamate concentrations, which triggers metabotropic glutamate receptors, which in turn suppress neuronal vesicle release [37, 38]. When evoked EPSCs (eEPSC) from cortical layer 2/3 neurons were recorded, system x${}_c$${}^{-}$ inhibitor sulfasalazine could decrease these eEPSCs [39]. In mixed primary hippocampus cultures with both astrocytes and neurons, treatment with system x${}_c$${}^{-}$ inhibitor erastin led to a reduction of the RRP size without affecting the recycling pool size [40]. In these cultures, treatment with system x${}_c$${}^{-}$ inhibitors erastin [23] and sorafenib [41] resulted in increased extracellular glutamate levels. In addition to acting on system x${}_c$${}^{-}$, sorafenib inhibits several targets, such as the RAF pathway and the vascular endothelial growth factor receptor family (VEGFR), which is of particular importance since VEGF has direct synaptic effects [42]. The broad range of targets on which sorafenib acts makes it difficult to determine which target is particularly relevant in this context. Thus, there is a need for more specific inhibitors to narrow down the targets involved when ‘dirty drugs’ are used.

The action of system x${}_c$${}^{-}$ is not restricted to the presynaptic site. A study in Drosophila revealed that a reduction in system x${}_c$${}^{-}$ reduced extracellular glutamate levels [43]. Specifically, in Drosophila mutants with genetic elimination of the so-called xCT gene “genderblind”, extracellular glutamate falls off by 50% compared to wild-type-level [43]. In response to this reduced glutamate level, the postsynaptic receptors become more expressed [43]. This elevated postsynaptic receptor expression could be reversed by exposing postsynaptic glutamate receptors to normal levels of glutamate [43]. System x${}_c$${}^{-}$′s regulation of extracellular glutamate hence has a direct effect on synaptic transmission. This finding was further substantiated by recordings from the Schaffer collaterals, the CA3-to-CA1 synapse in the hippocampus, showing that a knock-out of system x${}_c$${}^{-}$ leads to more $\alpha$-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors being expressed at the membrane, and to a strengthened synaptic transmission [44]. Furthermore, in system x${}_c$${}^{-}$ deficient sut mice, long-term potentiation and long-term memory were decreased, while the basic synaptic transmission was unaltered [45]. In line with that normal synaptic transmission, the assessment of mouse behavior in terms of spontaneous alternation, rotarod, and open field behavior, revealed that a genetic knockout of system x${}_c$${}^{-}$ does not influence these basic kinds of behavior [46].

Extracellular glutamate may bind to a variety of receptors, amongst them AMPA receptors, N-methyl-D-aspartate (NMDA) receptors, or metabotropic glutamate receptors (mGluRs), and also to glutamate transporters such as EAAT1/2. While mEPSCs abate upon an excess extracellular glutamate, mGluR1-mediated EPSCs swell [47]. This indicates competition for the extracellular glutamate between neuronal receptors and glial transporter. Extracellular glutamate levels cannot only be increased by system x${}_c$${}^{-}$ activation. Also the inhibition of EAAT1 (GLAST) and of EAAT2 (GLT1) has similar effects [47]. The interplay between all different glutamate transporters is complex and essential in determining how the synaptic transmission is affected by the manipulation of one of them. This is illustrated for example by the finding that system x${}_c$${}^{-}$ deletion led to increased expression of EAAT2 [48], which does not represent a compensatory, but rather a reinforcing mechanism of decreasing extracellular glutamate. The extent to which glial glutamate transporter-mediated glutamate uptake (via EAAT1/2) affects neurotransmission is even brain region-specific: compared with the neonatal rat hippocampus, the expression of EAAT1 and EAAT2 is lower, and the entry of glutamate into neonatal rat cortical astrocytes is slower [49].

When pharmacological action on the glutamate transporters is planned, it is important to assess how this broad inhibition would impact the glutamate transporters on the diverse glial cells. One promising approach to tackle this question is the investigation of cell-type specific knock-outs, e.g., of EAAT2. When EAAT-2 was genetically knocked out using neuron- or astrocyte-specific Cre-loxP systems, it turned out that the astrocytic EAAT2-deletion had more severe effects than the neuronal deletion of EAAT2 [50]. When astrocytes had lost their glutamate-importing capacities, mortality and seizure susceptibility increased [50]. Interestingly, during aging an astrocyte-specific EAAT2 knock-out differs from a neuron-specific knock-out on behavioral and transcriptional levels [51], indicating that the different contributions of neuron- and astroglial-mediated glutamate imports are directly linked to behavioral states. These findings stress the fact that it is important to precisely investigate each transporter and its role in physiological and pathophysiological settings.

Altogether, glial glutamate transporters control important aspects of the physiologic neuronal activity that gives rise to behavior. In the following paragraph, we will discuss how these transporters shape astrocytes’ pathophysiological role in the context of different CNS diseases.

The knowledge that astrocytes are associated with human disease is growing rapidly [52]. One of the most devastating diseases of the human CNS represents malignant glioma. Resections—if even possible at all—are incomplete at best. The tumors are often treatment resistant and can progress to higher CNS WHO grades even after early detection, up to glioma of WHO grade 4 (mostly glioblastoma) [53, 54]. The expression levels of system x${}_c$${}^{-}$ in such highly malignant human glioblastoma cells are positively correlated with tumor invasion, and negatively correlated with patient survival [55, 56]. In contrast to the high expression of system x${}_c$${}^{-}$ in tumor cells [57, 58], the glutamate importers EAAT1 and EAAT2 are abnormally low expressed in glioma cell lines [59], animal models [60] and human glioblastoma [61]. It has even been shown that increased EAAT2 expression inhibits tumor growth [62], further underscoring the idea of glutamate as a proponent of malignant growth [63].

In addition to the harms caused by the tumor itself, healthy CNS function is often affected in the peritumoral area as well. In a very recent study, the transcriptome of human astrocytes that stem from tissue surrounding the tumor site was analyzed [64]. The authors found that peritumoral astrocytes downregulated genes related to reacting to their microenvironment and synaptic function. Interestingly, amongst those genes were the EAAT2-encoding gene SLC1A2 and the EAAT1-encoding gene SLC1A3 [64]. In the peritumoral area of human tumor samples, system x${}_c$${}^{-}$ expression was not elevated [39], suggesting that solely glutamate importers were downregulated and thereby could contribute to increase extracellular glutamate levels. In a tumor-transplant model performed in mice, it was found that primary CNS tumors release glutamate via system x${}_c$${}^{-}$, which evokes epileptic activity in the peritumoral area [65]. High expression of system x${}_c$${}^{-}$ and low expression of EAAT are biomarkers for glioma-associated seizures [66, 67]. While the primary tumor enhances its glutamate-releasing capabilities (high system x${}_c$${}^{-}$ expression), astrocytes in the peritumoral area decrease their glutamate-importing capacities (low EAAT1/2 expression). This leads to a higher glutamatergic tone in the peritumoral area. As a result, glioma often coincides with epilepsy [68].

Glutamate transporters are involved in the pathomechanisms of epileptic seizures also without glioma as the underlying cause. In biopsies from patients with pharmacoresistant temporal lobe epilepsy, qPCR analysis revealed an upregulation of xCT mRNA [69]. In humans, EAAT1/2 is also involved in epilepsy. Astroglial glutamate transporters EAAT1/2 are investigated in the context of seizures since almost thirty years ago, in mice, the EAAT2 gene was located in a chromosomal region known to modulate neuroexcitability and seizure frequencies [70]. In patients suffering from pharmacoresistant temporal lobe epilepsy but without hippocampal sclerosis (i.e., without significant cell loss), EAAT2 was upregulated in hippocampal subregions, as confirmed by immunohistochemical stainings and in situ hybridization from post-mortem tissue [71]. In human epileptic foci in the neocortex, EAAT2 protein expression was decreased [72], similar to other cohorts of patients with intractable epilepsy in which EAAT1 and EAAT2 were decreased [73, 74]. Taken together, it appears that in human patients suffering from epilepsy, the glutamate homeostasis is shifted towards an increase in the extracellular glutamate level. However, human studies on measurable levels of glutamate in epilepsy have yielded mixed results and further studies using standardized experimental paradigms are needed to elucidate this topic [75].

The hypothesis that glutamate excess is contributing to epileptic seizures [76] was tested in several animal models with blocked system x${}_c$${}^{-}$ function: at first, it was described that xCT${}^{-/-}$ mice have an elevated threshold for seizures induced by pilocarpine or kainic acid [19]. Thus, fewer seizures occur in the absence of system x${}_c$${}^{-}$. In these experiments, pilocarpine applications induce seizures in the temporal lobe via its action on the M1 receptor and represent a widely used experimental model to study epileptic seizures [77]. Seizures are induced by kainic acid because of its analogism with glutamate and the agonistic action on ionotropic glutamate receptors [78]. N-acetylcysteine as an activator of system x${}_c$${}^{-}$ could not exert its usually proconvulsive effects in xCT${}^{-/-}$ mice – ascribing a seizure-promoting role to system x${}_c$${}^{-}$ through its glutamate-extruding activity [19]. In hippocampal rat brain slices, convulsive agents such as kainic acid, pilocarpine, and veratridine reduced the EAATs activity and could therefore be rescued by the application of a mGluR III agonist [79]. In this study, the injection of an EAAT inhibitor into rats exacerbated kainic acid-induced seizures. Thus, more seizures occured in the absence of EAATs because blockage of glutamate transporters increased their extracellular concentrations. Involving mGluRs, EAAT inhibition in hippocampal pyramidal cells activated mGluR group I and II receptors and led to epileptic-like activity [80]. Even in zebrafish, EAAT2 deficiency led to increased glutamate levels as well as light-induced seizures in neurons and glial cells [81].

In a model of self-sustained status epilepticus—where amygdala-implanted electrodes were used for stimulation that results in ongoing seizure events—genetic system x${}_c$${}^{-}$ inactivation had anticonvulsive and antiepileptogenic effects, weighing also towards the seizure-promoting role of system x${}_c$${}^{-}$ [82]. In their model, pharmacological system x${}_c$${}^{-}$ inhibition (by sulfasalazine treatment) also prevented seizures. In line with this, system x${}_c$${}^{-}$ null mice (xCT${}^{\text{sut/sut}}$) displayed a reduced occurrence of seizures in a pentylenetetrazole kindling model, a model in which inhibition of GABA${}_{\text{A}}$ receptors increases synaptic stimulation and excitation [83, 84]. Although the same authors, however, later reported lower seizure thresholds in system x${}_c$${}^{-}$ null mice after kainic acid or pentylenetetrazole and more severe seizures [85]—ascribing a seizure-preventing role to system x${}_c$${}^{-}$—it generally appears that targeting glial system x${}_c$${}^{-}$ may support therapy of seizures. In a very recent study, the system x${}_c$${}^{-}$ inhibitor sulfasalazine could ameliorate seizures that were elicited by astrogliosis [86], ruling out pharmacological manipulations that potentially may occur during seizure induction.

In summary, increased expression of system x${}_c$${}^{-}$ and reduced expression of EAATs promote epileptic activity and seizures in peritumoral astrocytes and in temporal lobe epilepsy. From zebrafish to humans, a decrease in astroglial EAAT activity promotes seizures. This raises the hypothesis that a system x${}_c$${}^{-}$-driven increase in glutamate promotes seizures, while the EAAT-driven decrease in extracellular glutamate prevents seizures. In conclusion, epilepsy treatment might benefit from selectively targeting astrocytic glutamate transporters in a manner that promotes the activity of EAATs but decreases system x${}_c$${}^{-}$ activity.

Neurodegenerative diseases like Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis, is heavily impacting human society. Despite years of research efforts, our understanding of the pathophysiology of neurodegeneration is still ill-defined [87]. Moreover, up to now there is no cure for diseased patients. Amongst many cellular and molecular suspects, astrocytes have been implicated in neurodegeneration and in neuroprotection [88, 89]. In the following, we specifically discuss the roles of system x${}_c$${}^{-}$ and EAAT1/2 in Alzheimer’s disease, amyotrophic lateral sclerosis, and Parkinsonian disease.

In this review we brigded the experimental data with clinical data on glial glutamate transporters system x${}_c$${}^{-}$ and EAAT1/2. We present evidence for the described plasticity as a consequence of the disease pathology as well as part of the pathogenic process. We took a closer look at (tumor-associated) epilepsy, Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis, and multiple sclerosis.

Under physiological conditions, the astroglial glutamate transporters system x${}_c$${}^{-}$ and EAAT1/2 control neuronal function through pre- and postsynaptic modulation of synaptic transmission.

For epileptic seizures induced by adjacent tumor tissue, an aberrant glutamate transporter activity appears as a causative. In contrast, the resulting glutamate levels in epileptic patients with non-tumor related epileptic disorders are yet unclear. It would be interesting to take a further look at the many different pharmacological models that are used to induce epileptic seizures with many underlying signaling cascades.

The evidence presented in Alzheimer’s disease showed that glial glutamate transporter plasticity can be directly induced via Amyloid-$\beta$-induced changes.

In ALS, studies showed how restoring glutamate transporter works, strongly indicates their pathogenetic role since those interventions proved beneficial to ameliorate symptoms and prolong survival. Similarly, the role of glutamate transporters in Parkinson’s disease may be causative for the disease since treatment approaches directly to these transporters were beneficial for the disease outcome. Since multiple sclerosis is the interplay of many cell types and molecular events that are not fully understood yet, it is more difficult to categorize changes in glutamate transporters as causative or consequential.

The role of glial glutamate transporter-mediated plasticity is still ambiguous. The relevance of these proteins in some brain diseases is increasing and worth for further studies. Approaches addressing the selectivity of inducers or inhibitors to glutamate transporters could further deepen our knowledge of the pathophysiological processes of these brain disorders in order to improve treatment management.

A$\beta$PP23, amyloid-protein precursor 23 gene; ALS, amyotrophic lateral sclerosis; CA, cornu ammonis; CNPase, 2′,3′-cyclic nucleotide 3′-phosphodiesterase; CNS, central nervous system; EAAT1/2, excitatory amino acid transporters 1 and 2; EAE, experimental autoimmune encephalomyelitis; EPSC, excitatory postsynaptic currents; FGF-2, fibroblast growth factor 2; GLAST, glutamate aspartate transporter 1; GLT-1, glutamate transporter-1; GSH, glutathione; Iba-1, ionized calcium binding adaptor molecule 1; IL 1$\beta$, Interleukin-1 beta; L-DOPA, l-3,4-dihydroxyphenylalanine; mGluR1, metabotropic glutamate receptor 1; MPTP, 1 methyl 4 phenyl 1,2,3,6-tetrahydropyridine; mRNA, messenger ribonucleic acid; MS, multiple sclerosis; NMDA receptor, N-methyl-D-aspartate receptor; 6-OHDA, 6-hydroxydopamine; PET, positron emission tomography; qPCR, quantitative polymerase chain reaction; RRP, readily releasable pool; SLC1A2, solute carrier family 1 member 2; SLC1A3, solute carrier family 1 member 3; SLC7A11, solute carrier family 7 member 11; SOD-1, superoxide dismutase 1; TMEV, Theiler’s murine encephalomyelitis virus; TTX, tetrodotoxin; xCT, cystine/glutamate antiporter; WHO, world health organization.

EY, NS and MD designed the study. EY, JKD and MD collected and analyzed the literatures. EY, HHS, JKD, MD and NS wrote, read and approved the final manuscript.

Not applicable.

We acknowledge financial support by Deutsche Forschungsgemeinschaft and Friedrich-Alexander-Universität Erlangen-Nürnberg within the funding programme.

Deutsche Forschungsgemeinschaft and Friedrich-AlexanderUniversität Erlangen-Nürnberg funding Programme of “Open Access Publication Funding”.

The authors declare no conflict of interest.