Loss of cerebellar glutamate transporters EAAT4 and GLAST differentially affects the spontaneous firing pattern and survival of Purkinje cells

Abstract Loss of excitatory amino acid transporters (EAATs) has been implicated in a number of human diseases including spinocerebellar ataxias, Alzhiemer’s disease and motor neuron disease. EAAT4 and GLAST/EAAT1 are the two predominant EAATs responsible for maintaining low extracellular glutamate levels and preventing neurotoxicity in the cerebellum, the brain region essential for motor control. Here using genetically modified mice we identify new critical roles for EAAT4 and GLAST/EAAT1 as modulators of Purkinje cell (PC) spontaneous firing patterns. We show high EAAT4 levels, by limiting mGluR1 signalling, are essential in constraining inherently heterogeneous firing of zebrin-positive PCs. Moreover mGluR1 antagonists were found to restore regular spontaneous PC activity and motor behaviour in EAAT4 knockout mice. In contrast, GLAST/EAAT1 expression is required to sustain normal spontaneous simple spike activity in low EAAT4 expressing (zebrin-negative) PCs by restricting NMDA receptor activation. Blockade of NMDA receptor activity restores spontaneous activity in zebrin-negative PCs of GLAST knockout mice and furthermore alleviates motor deficits. In addition both transporters have differential effects on PC survival, with zebrin-negative PCs more vulnerable to loss of GLAST/EAAT1 and zebrin-positive PCs more vulnerable to loss of EAAT4. These findings reveal that glutamate transporter dysfunction through elevated extracellular glutamate and the aberrant activation of extrasynaptic receptors can disrupt cerebellar output by altering spontaneous PC firing. This expands our understanding of disease mechanisms in cerebellar ataxias and establishes EAATs as targets for restoring homeostasis in a variety of neurological diseases where altered cerebellar output is now thought to play a key role in pathogenesis.

EAATs are responsible for the maintenance of low extracellular glutamate levels which prevents neurotoxicity whilst ensuring accurate synaptic communication (1). However, the full detail of the pathogenic effects of transporter loss in disease states is not yet clear and evidence is lacking for a primary role of EAAT4 loss in cerebellar ataxia.
EAAT4 is expressed in Purkinje cells (PCs) (24), the principal neurons and sole output of the cerebellar cortex, whereas GLAST/EAAT1 is the predominant glial transporter, being expressed in Bergmann glia 6-fold more abundantly than GLT-1/EAAT2 (25). GLAST/EAAT1 is believed to be present uniformly throughout the cerebellum in functional excess (26). In contrast, EAAT4 displays a differential pattern of expression within parasagittal bands, mapping onto that of aldolase C (zebrin II), with zebrin-positive (Zþ) PCs expressing EAAT4 to a much higher level than zebrin-negative (Z-) PCs (27,28). This variable EAAT4 density has been shown to have synaptic physiological consequences with larger glial AMPAR-mediated current amplitudes observed in regions where PCs have lower endogenous levels of EAAT4 (29) and the synaptic activation of mGluR1a, which exhibits the same peri-synaptic distribution as EAAT4 (28,30), found to be dampened in regions of high EAAT4 expression despite no differential expression of mGluR1a (31). This physiological effect of EAAT density on synaptic activity is due to the slow glutamate translocation time of EAATs with multiple cycles of transport not effective in the rapid removal of glutamate from the synaptic cleft (32). Instead it is the density of EAATs that governs the rapidity of binding and sequestration of glutamate after synaptic release (33).
PCs, in addition to receiving excitatory and inhibitory synaptic input, also exhibit spontaneous high frequency repetitive firing (34,35), the frequency of which was recently reported to vary between Zþ and Z-bands of the cerebellar cortex in rats (36) and mice (37). To investigate whether a pathological loss of EAAT4 and GLAST/EAAT1 protein has an effect on the spontaneous activity of PCs, and whether such a defect plays a key role in the pathogenesis of cerebellar ataxia, we performed wholecell patch clamp recordings of PCs in vitro in EAAT4 (ET4 -/-) and GLAST (GLAST -/-) knockout mice. Here, we report that loss of EAAT4 enhances the spontaneous firing frequency and decreases the regularity of action potential firing of Zþ PCs by a mechanism mediated through mGluR1 signalling. In contrast, the intrinsic activity of Z-PCs is reduced in mice lacking GLAST and this is mediated by NMDA receptor signalling. Motor deficits arising from loss of EAAT4 and GLAST function are alleviated by mGluR1 and NMDA receptor blockade, which restore normal spontaneous simple spike activity in ET4 -/and GLAST -/knockout mice respectively. Furthermore, loss of EAAT4 and GLAST results in loss of Zþ and Z-PCs, respectively. Thus, this study demonstrates that within the cerebellum, plasma membrane glutamate transporters have differential roles in regulating the intrinsic firing frequency and survival of PCs, both key elements in normal cerebellar function by ensuring accurate and efficient processing of sensory information. These data refine our understanding of how glutamate transporter dysfunction can disrupt the encoding of information within discrete functional units and contribute to disease pathogenesis in cerebellar disorders.

Differential effect of cerebellar transporter loss on the spontaneous firing frequency of PCs
To elucidate what effect reduced levels of the two predominant cerebellar glutamate transporters have on the intrinsic properties and spontaneous firing pattern of PCs, we carried out in vitro whole-cell patch clamp recordings on PCs from ET4 -/and GLAST -/mice. All recordings were carried out in the presence of AMPA and GABA A receptor antagonists to block fast synaptic inputs. This revealed heterogeneity within the pattern of spontaneous simple spike activity ( Fig. 1A and B). All wild-type PCs possessed a regular (tonic) rate of action potential firing, but 30% of PCs in ET4 -/mice were observed to fire irregularly with periods of high frequency activity (bursting), although regular tonic firing could be restored in all bursting cells by delivery of a negative current. The regularity of simple spike activity was quantified by calculating the coefficient of variation of interspike interval (CV ISI) with a value >1 indicating high variance/ irregularity (Fig. 1C) and used to classify whether a PC was burst firing. In contrast, 34% of PCs in GLAST -/mice were found to be silent with no spontaneous action potential firing recorded within a minimum of 2 min. However, action potentials could be elicited in all silent cells by a depolarizing current step. The remaining PCs in ET4 -/and GLAST -/mice (70 and 66%, respectively) exhibited tonic firing with a CV ISI < 1.
EAAT4 modulates the firing pattern of zebrin II positive (Zþ) PCs whereas GLAST modulates the firing pattern of zebrin II negative (Z-) PCs Given EAAT4 expression levels differ through the cerebellum (27,28), we next determined whether the heterogeneity in firing observed in ET4 -/and GLAST -/knockout mice correlated with EAAT4 expression profile. This was achieved by creating ET4 -/--eGFP and GLAST -/--eGFP mice by crossing ET4 -/and GLAST -/mice with a fluorescent reporter BAC mouse (eGFP) in which eGFP cDNA is under the control of the EAAT4 promoter (38).
To confirm the fluorescent intensity of GFP accurately reflects EAAT4 promoter activity, we first visualized mid-sagittal cerebellar sections from eGFP reporter mice. This revealed a higher GFP signal in posterior lobules VI-X compared with anterior lobules I-V ( Fig. 2A) consistent with the fact EAAT4 expression is higher in posterior lobules. Next, coronal cerebellar sections from eGFP reporter mice were immunostained using antibodies against either EAAT4 or aldolase C (zebrin II), a protein known to exhibit the same para-sagittal banding pattern as EAAT4 (28). In both instances, the immunoreactivity was identical to that of the eGFP banding pattern ( Fig. 2A), confirming that the eGFP fluorescence signal does accurately reflect EAAT4 expression profile and can be used to visually distinguish Zþ and Z-PC subtypes. Of note no difference in EAAT4, GLAST or GLT1 levels were observed between WT and WT-eGFP mice and neither full-length nor any truncated form of EAAT4 protein was detected in the cerebellum of ET4 -/--eGFP mice (Fig. 2B).
Finally, although not all high GFP expressing (Zþ) ET4 -/cells were found to be bursting (CV ISI > 1) they exhibited a significantly higher tonic firing frequency (64.5 6 9.4 Hz; P ¼ 0.047) to that of Zþ WT cells (39.8 6 4.3 Hz). Conversely, all the low GFP expressing (Z-) GLAST -/cells that were not silent were significantly less active, exhibiting a much lower firing frequency (9.6 6 3.8 Hz; P ¼ 0.001) to that of Z-WT cells (41.6 6 3.9 Hz; Fig. 2E). Of note in 6-week-old WT animals a slightly higher firing frequency was detected in Z-WT cells compared with Zþ WT cells but a significant increase in firing rate was only observed in mice older than 3 months (Zþ 34.5 6 2.9 Hz, n ¼ 12; Z-53.4 6 7.5 Hz, n ¼ 9; P ¼ 0.04), an age similar to that used by Zhou and colleagues (37).

mGluR1 activation downstream of EAAT4 loss mediates irregular firing of Zþ PCs
To test whether an acute loss of transporter activity can alter spontaneous PC activity similar to a constitutive loss, WT cerebellar slices were pre-incubated with DL-TBOA (100 lM), a nonselective glutamate transporter antagonist prior to performing in vitro whole cell patch recordings. A concentration of 100 lM TBOA was chosen as it blocks a significant proportion of, but not all, transporter molecules and results in a physiologically relevant increase in ambient glutamate (39). Following preincubation for 1 h, the majority of PCs were found to exhibit an irregular pattern of firing (CV ISI 3.75 6 1.67; Fig. 3), similar to that of high GFP expressing (Zþ) ET4 -/cells (2.9 6 0.8; Fig. 2C and D). If incubation with DL-TBOA was extended to 2 h, further elevating ambient glutamate levels, then the majority of PCs were found to fall silent ( Fig. 3A and C), with those not silent possessing a regular but lower frequency of firing (18.8 6 1.04 Hz; CV ISI 0.06 6 0.005; Fig. 3B). The pre-incubation with DL-TBOA was not pathological as normal tonic firing could be restored in all silent cells with a depolarizing current step.
Next we investigated the effect a group 1 mGluR agonist, dihydroxyphenylglycine (DHPG), had on the intrinsic firing properties of Zþ WT PCs. This pharmacological manipulation was carried out because regional differences in synaptic mGluR1a activity have been shown to correlate with the expression pattern of EAAT4 (31) and mGluR1 agonists were shown to induce burst activity in cultured Purkinje neurons (40). Excessive mGluR activation induced by the application of DHPG resulted in the irregular spontaneous firing of Zþ WT PCs (CV ISI 8.8 6 1.5) but which was reversible following drug wash-out ( Fig. 4A and B). Moreover, the TBOA-induced bursting phenotype in WT PCs was blocked in all but one cell by the co-incubation of TBOA for 1 h with the mGluR1 antagonist CPCCOEt (10 lM). The restoration of normal tonic firing by CPCCOEt (26.8 6 5.5 Hz; CV ISI 0.14 6 0.07) indicates that aberrant mGluR1 signalling was responsible for the irregular PC firing following elevated ambient glutamate levels ( Fig. 4B and C).

NMDA receptor activation downstream of GLAST loss abolishes Z-PC spontaneous activity
Depolarization-induced potentiation of inhibition has been reported in the cerebellum whereby activation of pre-synaptic NMDA (ionotropic glutamate) receptors on molecular layer  interneuron terminals results in a sustained increase in inhibitory efficacy onto PCs (41). Furthermore, glial glutamate transporters have been shown to influence synaptic inhibition of PCs by limiting glutamate spillover and hence the activation of presynaptic NMDA receptors (42). At 6 weeks of age PCs are not believed to express functional NMDA receptors (43). We therefore determined whether excessive presynaptic NMDA activation could be responsible for the silencing of Z-PCs in GLAST -/mice due to glutamate spillover in regions of the cerebellar cortex with low EAAT4 levels. This was achieved by pre-incubating acute cerebellar slices from GLAST -/--eGFP mice with AP5 (50 lM), an NMDA receptor antagonist, for 30 min prior to performing whole cell patch recordings. Following NMDA receptor blockade all Z-PCs in GLAST -/mice were no longer silent, but were found to be spontaneously active (21.2 6 3.9 Hz; CV ISI 0.15 6 0.04; N ¼ 3, n ¼ 8; Fig. 5A and B), in contrast to control slices where all silent PCs remained silent (N ¼ 3, n ¼ 13). We then investigated whether NMDA receptor blockade could rescue motor deficits in GLAST deficient mice by administering AP5 (2.5 mg/kg). The ability of animals to stay on the rotarod at 10 rpm following drug administration was compared with their pre-treatment performance (Fig. 5C). This revealed a significant improvement in motor performance following administration of AP5 (51.7 6 10.5 s) compared with vehicle (19.4 6 4.3 ; P ¼ 0.0096 paired t-test; N ¼ 5). There was no adverse effect of AP5 treatment on the motor behaviour of wild-type animals (vehicle, 120 s; AP5, 118.5 6 1.5 s; A similar effect on Z-PC spontaneous activity was observed with age. In 1-year-old WT mice 45.5% of PCs from the predominantly Z-anterior lobules (I-V) were found to be silent whereas only 8.3% of cells were silent in the predominantly Zþ posterior lobules (VIII-X) (N ¼ 3, n ¼ 22, lobules I-V; N ¼ 5, n ¼ 36, lobules VIII-X; P ¼ 0.0023; Fig. 5C). This is in contrast to all Zþ and Z-PCs in young mice exhibiting regular spontaneous activity (Fig. 2C). A significant reduction in GLAST protein levels was observed in 1-year-old compared with 1-month-old WT mice (65.3 6 1.9% of 1-month-old levels, N ¼ 4, P ¼ 0.0004; Fig. 5D) indicating the loss of GLAST protein could be a factor of ageassociated silencing of Z-PCs and in turn age-related motor decline. These findings are also consistent with our previous study in a disease model of SCA type 5 where one third of in vivo PC recordings from lobule V in 8-month-old b-III -/mice exhibited no simple spike firing (5). At this age b-III -/mice have a significant reduction in GLAST protein (5) supporting the case that (C) Ratio of WT PCs that are tonically firing or bursting following pharmacological treatment to manipulate mGluR1 signalling. loss of GLAST function results in a decrease in PC activity and reduced output from the anterior cerebellar cortex.

EAAT4 knockout mice display progressive ataxia
Because this study revealed Zþ PCs in ET4 -/mice exhibit irregular spontaneous intrinsic activity and with age there is loss of Zþ PCs, combined with previous studies that reveal EAAT4 loss affects synaptic events (7,44,45), we carried out a longitudinal behaviour analysis of ET4 -/mice to determine whether they exhibit any motor deficits. This involved carrying out gait analysis, rotarod and an elevated beam task on 6-week, 6-month and 10month-old WT and ET4 -/animals.
Next we investigated whether mGluR1 blockade could rescue the motor deficits observed in 6-week-old ET4 -/mice by administering JNJ16259685 (0.15 mg/kg) (Fig. 8C and D). The ability of animals to stay on the rotarod at 10 rpm, the only motor task found to be impaired in young ET4 -/mice, following drug administration was compared with their pre-treatment performance (Fig. 8D). This revealed a significant improvement in motor performance following administration of JNJ16259685 (93.7 6 8.5 s) compared with vehicle (65.7 6 6.5 s; P ¼ 0.007 paired t-test; N ¼ 15). No adverse effect of JNJ16259685 treatment was observed on rotarod performance of wild-type animals (118.3 6 1.1 s; P ¼ 0.724) but mice lacking GLAST were further impaired on the rotarod following administration of JNJ16259685 (5.9 6 2.3 s; P ¼ 0.011). Together the data indicate mGluR1 blockade can alleviate motor deficits resulting from loss of EAAT4, but not that of GLAST.

Discussion
The pathogenic consequences of cerebellar glutamate transporter dysfunction have, to date, been largely related to excitotoxicity. Here we demonstrate for the first time that the two predominant cerebellar glutamate transporters, EAAT4 and GLAST/EAAT1, have critical but differential roles in the modulation of spontaneous PC firing patterns. This study expands our understanding of how glutamate transporter dysfunction through elevated extracellular glutamate and the dysregulation of peri-and extra-synaptic receptors, as well as receptors on nearby cells, contributes to disease pathogenesis in cerebellar disorders. Of note the study also highlights that in disease states the effect on cerebellar cortical output from specific PC populations and hence inhibition onto neurons in the cerebellar nuclei will differ depending on the EAAT dysfunction. High levels of EAAT4 protein are required to constrain the inherent excitability and plasticity of Zþ PCs by limiting mGluR1 signalling. In contrast GLAST/EAAT1 appears to play no role in modulating the firing pattern of the Zþ subpopulation of PCs but it does have a fundamental role in sustaining normal spontaneous simple spike activity in low EAAT4 expressing (Z-) PCs by preventing NMDA receptor activation. Loss of GLAST from these regions of the cerebellum results in a substantial decrease in spontaneous activity of Z-PCs, with the majority falling silent and the other Z-cells exhibiting a very low intrinsic firing frequency. In addition, both these cerebellar transporters were found to have differential effects on PC survival, with EAAT4 expression important in preventing Zþ PC death in the posterior cerebellum, whereas GLAST expression is required for the longterm survival of PCs expressing low levels of EAAT4. Together the data provide strong evidence that the consequence of GLAST/EAAT1 or EAAT4 loss varies between cerebellar regions, but in both cases it has a profound effect on intrinsic properties of PCs, cell survival and motor coordination. This new knowledge that ambient elevations in extracellular glutamate due to glutamate transporter dysfunction can alter the spontaneous firing pattern of PCs contributes to a better understanding of disease mechanisms.

Loss of EAAT4 results in ataxia and selective PC death
Here we show for the first time that loss of EAAT4 results in an ataxic phenotype and that the motor deficits are progressive. Although it has previously been reported that EAAT4 knockout mice do not have a cerebellar phenotype, this observation was simply based on a visual inspection of cage behaviour (46). Without direct assessment of motor task function motor deficits are likely to be overlooked; indeed given the young age of ET4 -/mice studied it is also possible that the visual observation of cage behaviour was not carried out on sufficiently aged ET4 -/animals (46). This study is also the first to implicate loss of EAAT4 in PC death and specifically of Zþ cells in the posterior cerebellum. Previous studies using ET4 -/mice either did not look (44,45) or analysis was not carried out on aged animals (47). Furthermore in the latter study using 8-to 12-week-old mice PC density was only quantified within the anterior cerebellum (lobule III) and not posterior lobules. The authors report PC loss following brain ischemia in lobule III of GLAST knockout mice but not EAAT4 knockout mice (47). This is in agreement with our findings but we also find that low EAAT4 expressing PCs in the posterior cerebellum, especially out with the central vermis, are vulnerable to cell death in the absence of GLAST protein. Therefore, in studies of ataxia where a loss of EAAT4 is observed this can now be directly implicated in a motor phenotype, so contributing to a better understanding of mechanisms that underpin cerebellar ataxia. The defects in cerebellar cortical output would likely arise from a combination of effects following EAAT4 loss; notably the findings reported here that Zþ PCs in the posterior cerebellum exhibit irregular firing patterns due to aberrant mGluR1 signalling and altered intrinsic properties and that they are more vulnerable to cell death following EAAT4 loss but also that EAAT4 loss affects synaptic events, altering the size and/or duration of excitatory postsynaptic currents at parallel fibre-PC synapses (7,44,45) and impairing the clearance of released glutamate at climbing fibre-PC synapses (46,48,49). Of note the latter defect is confounded by the fact that greater glutamate release was shown to occur from climbing fibre terminals in zebrin-positive regions (50).

Intrinsic Zþ PC membrane firing pattern is regulated by EAAT4 and mediated by mGluR1 signalling
Our data suggest that the irregular firing of Zþ PCs and motor incoordination resulting from EAAT4 loss are mediated by mGluR1 signalling. This is consistent with other recent studies that have implicated dysregulated mGluR1 signalling in altered PC excitability and ataxia. For example, prolonged mGluR1 activity at cerebellar parallel fibre-PC synapses was observed in moderately ataxic SCA1 mice (51) and mGluR1 activation was also shown to increase intrinsic PC firing in cerebellar slices from SCA2 mice (52). Moreover, consistent with our findings in regard to mGluR1 antagonist improvement in motor performance in ET4 -/mice, studies show that modulating mGluR1 activity and glutamatergic stimulation can restore motor performance in other mouse models of ataxia (51,53). Removing excessive mGluR1 activity by treating with JNJ16259685 improved motor performance in young SCA1 (51) but not old SCA1 mice (54). This difference in response to mGluR1 modulation may be due to the subsequent loss of GLAST in later stages of disease in SCA1 mice (4) as we find that mGluR1 blockade further impairs motor performance in mice lacking GLAST. In SCA28 mice reducing mGluR1 expression or increasing EAAT2 expression does alleviate ataxia (53). Of note, intrinsic excitability of wild-type PCs was found to be increased following induction of long-term depression in the parallel fibre-PC synapses (55), an event dependent on mGluR1 activity (56).
Currently it is unclear what downstream PC dendritic ionic conductance(s) are affected and responsible for the changes in neuronal firing observed in this study. However, deciphering the complex interplay of metabotropically regulated currents will be essential in elucidating the downstream cellular mechanisms that underpin alterations to PC intrinsic properties and which contribute to aberrant cerebellar output in disease states. Possibilities include the activation of low-threshold T-type calcium channels (57), changes to the surface expression of HCN channels, regulated by PLCb/IP 3 R and protein kinase C activities (58), or differences in potassium channels some of which have been implicated in ataxia (59)(60)(61)(62). Whether the differential expression of isoforms of mGluR1 (63), PLC beta (64) or other downstream signalling molecules such as aCaMKII are important in the regional effects mediated by the expression profile of cerebellar EAATs remains to be elucidated. Given mGluR1b does not interact with homer proteins (65) and is complementary to zebrin II (63), hence EAAT4, it could be that differences in the postsynaptic effectors of homer-dependent and homerindependent mGluR1 isoforms are relevant in the differential regulation of PC intrinsic properties within parasagittal compartments. For instance mGluR1a possesses higher affinity and coupling efficacy to G proteins and phospholipase C than mGluR1b (66). This would result in differences in Ca 2þ release from intracellular stores and may account for Zþ PCs reacting differently to that of Z-PCs following loss of EAAT function. The differences in glutamate release (50) and calcium signalling downstream of mGluR1a activation may also account for the fact that the proximal dendrites, sites of climbing fibre innervation, are the first to degenerate in PCs from the posterior cerebellum of b-III -/mice that lack functional EAAT4 (5,7).
In the future identifying all the proteins that underlie the different physiological profiles of PC subpopulations will be key to understanding the functional diversity of microzones within the cerebellar cortex and how distinct changes to firing patterns occur in different disease states.

Modulation of Z-PC intrinsic properties by GLAST/ EAAT1 activity as a potential factor in neurological disorders
Physiological defects have been reported to occur from changes to GLAST/EAAT1 activity with GLAST knockout mice exhibiting progressive motor incoordination (7,67) and mutations in SLC1A3, the gene encoding EAAT1, being associated with episodic ataxia type 6 (EA6) (9)(10)(11). In the case of one EA6 point mutation, substitution of proline 290 by an arginine, there is both an abnormal chloride conductance and a decrease in glutamate transport rates (9,68). A recent study in Drosophila proposes it is excessive chloride extrusion by the mutant transporter that is relevant to disease pathogenesis, as increased chloride uptake rescues locomotor performance (69). It is proposed that changes to intracellular chloride concentrations affect the osmotic balance, reduce cell volume and therefore ensheathment of synapses by astrocytes (69). A recent study also reports a retraction of Bergmann glial processes from PC dendrites in GLAST knockout mice (70). Such a reduced glial coverage of synapses superimposed on loss of GLAST/EAAT1 would further facilitate glutamate diffusion, extra-synaptic crosstalk and altered ionic conductances within dendrites. Together these studies improve our understanding of how PC dysfunction can arise following loss of the glial glutamate transporter GLAST in regions where EAAT4 levels are low. Given NMDA receptors composed of GluN1/GluN2C and GluN1/GluN2D subunits are less sensitive to Mg 2þ block (71) it is likely that these are the molecular layer interneuron receptor subtypes being activated by glutamate spillover. Furthermore as all recordings were performed in the presence of picrotoxin, a GABA A receptor antagonist it is proposed that aberrant dendritic GABA B receptor activation is involved, providing insights into the cellular mechanisms downstream of loss of GLAST function that underpin cerebellar dysfunction. This is consistent with the finding that baclofen, a GABA B agonist inhibits PC firing and that this inhibition persists in the presence of a GABA A antagonist (72).
Conversely increased EAAT1 activity has been suggested to be a factor in some neurodevelopmental disorders. A functional SLC1A3 variant has been reported to be more prevalent in individuals with Tourette syndrome than controls and found to result in increased EAAT1 membrane insertion and glutamate uptake (73). A gene duplication of SLC1A3 has also been reported as a possible risk factor for autism and attention deficit/hyperactivity disorder-like behaviour (74). The current study therefore points towards altered EAAT1 glutamate transport activity being implicated in neurodevelopmental disorders by modulating dendritic ionic conductances and hence spontaneous PC activity. In fact similar regional cerebellar differences to that observed in GLAST -/and ET4 -/mice have been reported in Shank2-deficient mice that exhibit autism-like behaviour (75). A significant decrease in simple spike frequency was observed in anterior lobules of global Shank2 -/mice, but not in their posterior lobules whereas a consistent irregularity of simple spikes was observed in PCs from the posterior but not anterior lobe. Depolarizing current steps have also revealed that PCs in the posterior, but not the anterior, cerebellum of b-III -/mice exhibit increased membrane excitability compared with wild-type PCs (data not shown) even though they display reduced spontaneous simple spike activity due to loss of ankyrin R and lower densities of Na v 1.1 and Na v 1.6 (5,76). Similarly GLT1/EAAT2 dysfunction has been implicated in enhanced pyramidal neuronal excitability in fragile X syndrome (77) and pathological repetitive behaviours (78). It therefore appears that alterations to distinct physiological profiles and intrinsic activity of PCs within microzones of the cerebellar cortex downstream of aberrant EAAT activity impairs the correct processing and efficient transmission of information possibly by decoupling coordinated activation of neighbouring functional microzones (79). Future studies are needed to address how aberrant cerebellar EAAT activity modulates intrinsic PC dendritic properties, not only in disorders of the motor system, but also in autism spectrum disorders and other cognitive and neuropsychiatric disorders given the emerging role of cerebellar dysfunction in these disorders (18)(19)(20)(21)(22)(23).
To conclude, this study identifies a new and central role for plasma membrane glutamate transporters in regulating spontaneous PC activity and survival, both key determinants of cerebellar output, by restricting the activation of peri-and extrasynaptic receptors, as well as receptors on nearby cells. Consequently altering functional cerebellar EAAT levels may prove very effective in restoring dendritic ionic conductances and neuronal function in a number of neurological disorders where intrinsic neuronal excitability is altered.

Animals
All procedures involved in generation and analysis of mutant mice were carried out according to the United Kingdom Animals (Scientific Procedures) Act (1986) and other Home Office regulations under specific-pathogen-free conditions. GLAST -/and ET4 -/mice, both on a C57BL/6 genetic background, were maintained and crossed with eGFP-reporter mice to generate GLAST -/--eGFP and ET4 -/--eGFP mice. The genotypes of all animals were determined by PCR analysis on genomic DNA extracted from ear notch biopsies using ChargeSwitch gDNA tissue kit (Invitrogen, Carlsbad, CA) and both sexes were used in all experiments. For ET4 -/mice a common upstream primer (5 0 -ttcctgattgctggaaagattctgg-3 0 ) and primers specific for the wild-type allele (5 0 -agttcagggaaaggccataccttgg-3 0 ) and the PGK-neo cassette in the mutant allele (5 0 -ggatcggccattgaacaagatgg-3 0 ) were used for amplification at an annealing temperature of 57 C. The 220-bp (from wild-type allele) and 1200-bp (from targeted allele) PCR products were resolved by electrophoresis on a 1.6% (w/v) agarose gel. For GLAST -/mice the specific primer sets used for amplification of wild-type allele were (5 0aagtgcctatccagtccaacga-3 0 ; 5 0 -aagaactctctcagcgcttgcc-3 0 ) and mutant allele (5 0 -aatggaaggattggagctacgg-3 0 ; 5 0 -ttccagttgaagg ctcctgtgg-3 0 ) at an annealing temperature of 54 C. The 214-bp (from wild-type allele) and 362-bp (from targeted allele) PCR products were resolved by electrophoresis on a 1.6% (w/v) agarose gel. The primer set specific for GFP was 5 0 -aagttcatctgcaccaccg-3 0 and 5 0 -tccttgaagaagatggtgcg-3 0 and used at an annealing temperature of 55 C. The 220-bp GFP PCR product was resolved by electrophoresis on a 1.6% (w/v) agarose gel. Immunoblot analysis using polyclonal antibody against EAAT confirmed ET4 -/genotype in ET4 -/--eGFP mice (presence of EAAT4-BAC construct prevented detection of homozygosity by PCR). All knockout mice were viable although pups from GLAST -/and GLAST -/--eGFP mice were routinely fostered with CD1 mothers to ensure survival.

Immunohistochemistry
Brains were removed and immersion-fixed with 4% (w/v) paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.4 overnight at 4 C and cryoprotected by immersion in 0.1 M sodium phosphate buffer (pH 7.4) containing 30% (w/v) sucrose. 30-lmthick coronal or mid-sagittal free-floating cerebellar sections were incubated for at least 3 h with blocking solution [5% (v/v) normal goat serum with 0.4% (v/v) Triton X-100 in PBS] prior to applying rabbit anti-EAAT4, anti-aldolase C and anti-ITPR1 antibody (Millipore) [2% (v/v) normal goat serum/0.1% (v/v) Triton X-100 in PBS] overnight at 4 C. Sections were washed three times in PBS before applying either Cyanine 3 (Cy3)-conjugated goat anti-rabbit IgG (Jackson laboratories) or fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (Cappel) for 1 h at room temperature followed by three rinses in PBS and coverslipping with Vectashield (Vector Laboratories, Burlingame, CA). All quantification was carried out blind to genotype and involved counting the number of PCs, along 1 mm linear lengths in anterior (II-IV) and posterior (VIII-X) lobules from three sections/animal and the counts averaged for each animal. Images were captured with either a Zeiss inverted LSM510 or Nikon confocal laser scanning microscope in the IMPACT imaging facility, Centre for Discovery Brain sciences. Colours were applied using Image J.

Motor coordination tests
Footprint patterns were analyzed using a runway (80 cm by 10.5 cm wide) with white paper at the bottom. Hind paws of animals were dipped in non-toxic, water-soluble black ink (Indian Ink, Winsor & Newton, Harrow, UK). Three consecutive strides were measured for each animal and base width was measured as the distance between the centre of one paw print to the centre of the next print on the opposite side. The elevated beam test was performed using a narrow horizontal beam (2 cm wide, 80 cm long, held at a height of 30 cm from the table). The number of hind paw slips the animal made whilst traversing the beam were counted. In the rotarod test the time a mouse remained on a rotating (3, 5 and 10 rpm) 3cm-diameter cylinder was recorded (maximum 120 s; four trials per speed; TSE Rotarod, Bad Homburg, Germany). 6-weekold ET4 -/mice received intraperitoneal injections of vehicle (sterile saline with 10% (w/v) hydroxypropyl-betacyclodextrin) 30 min before rotarod testing. Two hours after completion of rotarod testing, the same mice were injected with 0.15 mg/kg JNJ16259685 (Tocris Bioscience) dissolved in vehicle. Administration of JNJ at higher doses resulted in greater passivity, reduced rearing, a crouching behaviour and impaired motivation. About 6-month-old GLAST deficient mice received intraperitoneal injections of vehicle (sterile water) 30 min before rotarod testing. Two hours after completion of rotarod testing, the same mice were injected with 2.5 mg/kg AP5 (Tocris Bioscience) dissolved in vehicle.

Statistics
Statistical analysis was carried out using SPSS v21 (SPSS Inc., Chicago, IL, USA). Analysis was performed using Student's t-test for two groups of data and one-way analysis of variance for comparisons of three or more data groups. Contingency table and Fisher's tests were used for categorical data evaluation. Significance was accepted at P-values < 0.05. N indicates the number of animals and n indicates the number of cells.