Complete but not partial inhibition of glutamate transporters exacerbates cortical excitability in the R6/2 mouse model of Huntington’s disease

Summary

Aim

Deficient glutamate reuptake occurs in the cerebral cortex of Huntington's disease (HD) patients and murine models. Here, we examine the effects of partial or complete blockade of glutamate transporters on excitatory postsynaptic currents (EPSCs) of cortical pyramidal neurons (CPNs).

Methods

Whole‐cell patch clamp recordings of CPNs in slices from symptomatic R6/2 mice and wild‐type (WT) littermates were used to examine the effects of selective or concurrent inhibition of glutamate reuptake transporters.

Results

Selective inhibition of the glial glutamate transporter 1 (GLT‐1) or the glutamate aspartate transporter (GLAST) produced slight decreases in decay time of evoked EPSCs in CPNs from WT and R6/2 mice with no significant differences between genotypes. In contrast, concurrent inhibition of both transporters with DL‐TBOA induced a significant increase in area and decay time and this effect was significantly greater in R6/2 CPNs. Furthermore, full blockade also reduced spontaneous EPSC frequency and exacerbated epileptiform activity in CPNs from symptomatic R6/2 mice.

Conclusions

R6/2 CPNs are more sensitive to glutamate accumulation during full inhibition of both glutamate transporters, and these neurons have homeostatic mechanisms to cope with inhibition of GLT‐1 or GLAST by a mechanism that involves upregulation of either transporter when the other is deficient.

1. INTRODUCTION

Huntington's disease (HD) is a hereditary neurodegenerative disorder caused by a mutation in the Huntingtin gene (HTT) that results in an increase in the number of glutamine‐encoding CAG repeats.1, 2 When the mutant Htt (mHtt) protein contains >39 glutamine repeats, the illness manifests itself by the onset of cognitive and psychiatric disturbances followed by the progressive development of motor symptoms such as chorea. HD histopathology presents with a significant loss of medium‐sized spiny neurons (MSNs) in the striatum and in its innervation arising from cortical pyramidal neurons (CPNs).3 The corticostriatal pathway is critically involved in movement control, and brain imaging studies in presymptomatic HD carriers have shown that cortical atrophy occurs early, develops progressively, and correlates with the expression and severity of cognitive and motor symptoms.4, 5, 6, 7 More recent studies have shown that when the expression of mHtt is switched off only in CPNs, the electrophysiological alterations in MSNs are improved, thus reducing the motor phenotype in a conditional HD mouse model.8, 9 Altogether this evidence suggests that CPNs play a crucial role in the etiology of HD (for review, see10). Therefore, understanding the synaptic changes underlying aberrant functioning of CPNs in HD might provide novel drug targets to improve HD symptoms.

Central to CPN alterations is the failure of the glutamate reuptake system during glutamatergic neurotransmission. This system functions to regulate the synaptic concentration of glutamate in order to avoid its accumulation and spillover, which in turn could initiate an excitotoxic neuronal death cascade (for review, see ref. 11). Glutamate transporter 1 (GLT‐1) and glutamate aspartate transporter (GLAST) are proteins expressed predominantly in astrocytes, although GLT‐1 is also expressed in neuronal terminals, and they take up glutamate after its release from presynaptic terminals.12, 13 In addition to preventing glutamate accumulation, proper functioning of glutamate transporters is essential to ensure efficient neuronal transmission and flow of information14, 15. Decreased GLT‐1 mRNA and protein levels have consistently been reported in the striatum and cortex of postmortem HD brains and transgenic mouse models.16, 17, 18, 19, 20, 21 While consistent alterations have been observed in GLT‐1, less is known about changes in GLAST. Although some studies have shown decreased GLAST protein expression, others describe no changes in protein content.17, 18, 20, 21 Nevertheless, the reduction in protein content of both GLT‐1 and GLAST in the cortex of HD patients and animal models could render the striatum more vulnerable to glutamate‐induced neurotoxicity.21, 22 Together, this evidence suggests that failure of glutamate reuptake might contribute to the impaired cortical processing described in HD models both in vitro and in vivo.23, 24, 25, 26, 27 However, the precise effects of faulty glutamate reuptake on excitatory inputs to CPNs remain to be explored. In the present study, we evaluated changes induced by partial or complete blockade of the glutamate reuptake transporters. Evoked and spontaneous excitatory postsynaptic currents (eEPSCs and sEPSCs) were recorded in layer 2/3 CPNs from symptomatic R6/2 transgenic mice (10‐12 weeks of age) and wild‐type (WT) littermates.

2. METHODS

2.1. Animals and genotyping

Experimental procedures were performed under the United States Public Health Service Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care Committee at the University of California, Los Angeles (UCLA). All efforts were made to minimize suffering and the number of animals used. WT and R6/2 mice were from our breeding colony at UCLA. Mice were housed under controlled temperature (21‐24ºC) and humidity conditions (30%‐60%) and maintained on a 12‐h light/dark cycle with free access to food and water. For this study, we used a total of 45 male or female mice (n = 19 WT and n = 26 R6/2), at an average age of 75±5 days for WT and 74±3 days for R6/2 (t = 0.089, df = 43, P = 0.929). Mouse genotyping was performed using DNA obtained from tail samples collected at weaning [for more details about genotyping procedures, see 28]. After completion of electrophysiological recordings, mouse genotypes were corroborated by a secondary, blind genotyping process. The mean number of CAG repeats was 151±2 as determined by Laragen Inc (Culver City, CA).

2.2. Slice preparation and electrophysiological recordings

Mice were deeply anesthetized with isoflurane and intracardially perfused with ice‐cold high‐sucrose solution that had been oxygenated with 95% O₂‐5% CO₂. After removing the brain, it was immediately placed in the same oxygenated high‐sucrose slice solution. Coronal brain slices from frontal cortex (300 μm thick) were cut with a Vibratome (Leica VT1000S; Leica Microsystems, Buffalo Grove, IL) and transferred to an incubating chamber containing artificial cerebrospinal fluid (in mmol L⁻¹): 130 NaCl, 3 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 2 MgCl₂, 2 CaCl₂, and 10 glucose oxygenated with 95% O₂‐5% CO₂ (pH 7.2‐7.4, osmolality 290‐310 mOsm, at 31±2^ºC for 20 min), followed by exposure to room temperature. Recordings began after 1 h. Frontal CPNs from layer 2/3 were visualized using a 40x water‐immersion lens on an Olympus microscope (BX50WI) equipped with infrared video microscopy (QICAM‐IR Fast 1394; QImaging, Surrey, BC, Canada). Whole‐cell patch clamp recordings from slices were obtained with a MultiClamp 700A amplifier (Molecular Devices, Sunnyvale, CA) and pClamp software (version 10.2). Patch pipettes (3‐5 MΩ) were filled with a Cs‐methanesulfonate‐based internal solution containing (in mM): 130 Cs‐methanesulfonate, 10 CsCl, 4 NaCl, 1 MgCl₂, 5 MgATP, 5 EGTA, 10 HEPES, 5 GTP, 10 phosphocreatine, and 0.1 leupeptin (pH 7.2 with CsOH, 270 mOsM). For stimulation experiments, QX‐314 (4 mmol L⁻¹) was added to the internal solution to block Na⁺channel activation.

A depolarizing step voltage command (10 mV) and the membrane test function integrated in pClamp were used to estimate passive cell membrane properties: membrane capacitance (Cm; in pF), input resistance (Rm; in MΩ), and decay time constant (tau in msec). Cells were held at −70 mV and sEPSCs were recorded first in the presence of bicuculline (BIC, 5 μmol L⁻¹) to block gamma‐aminobutyric acid type A (GABA_A) receptor–mediated currents. After 7 min of BIC incubation, glutamate reuptake blockers were applied in the perfusate for 15‐18 min. We used either the nonspecific glutamate reuptake inhibitor, DL‐threo‐beta‐benzyloxyaspartate (DL‐TBOA; 30 μmol L⁻¹; Tocris Bioscience, Minneapolis, MN), the GLT‐1‐specific inhibitor, dihydrokainic acid (DHK; 100 μmol L⁻¹; Tocris Bioscience), or the inhibitor for GLAST, 2‐amino‐5,6,7,8‐tetrahydro‐4‐(4‐methoxyphenyl)‐7‐(naphthalen‐1‐yl)‐5‐oxo‐4H‐chromene‐3‐carbonitrile (UCPH‐101; 10 μmol L⁻¹; Hello Bio, Princeton, NJ). The concentrations of the glutamate reuptake inhibitors used in the present study were chosen based on their ability to block glutamate reuptake and induce significant changes in the decay time of evoked responses.29, 30, 31, 32, 33 To evaluate the effect of glutamate reuptake inhibition during cortical electrical stimulation, we used a monopolar glass electrode (impedance 1 MΩ) which was placed 200‐300 μm lateral from the recorded CPN. Glutamate receptor–mediated currents were evoked by an electrical pulse (0.1 mA, 20 s interpulse interval, 3 times). For group comparisons between genotypes, replicates from all cells were averaged to increase power and reduce variability. Similarly, percent changes before and after drug application were calculated for each replicate and then averaged. We examined the neuronal response before and 18 min after perfusion with the glutamate reuptake inhibitors.

2.3. Data analysis and statistics

As no gender differences were observed, data were pooled. sEPSC data were analyzed offline, initially with an automated detection protocol using the Mini Analysis program (Justin Lee, Synaptosoft, version 6.0) and then checked manually for accuracy by an investigator blind to the experimental condition. The threshold amplitude for the sEPSC event detection was 5 pA. We quantified large‐amplitude paroxysmal activity manually using a threshold of 400 pA. For the stimulation experiments, we used ClampFit software to determine peak amplitude, area, and decay time of eEPSCs. Decay time was considered the time 90%‐10% of the decay phase, where 100% was the peak amplitude and 0% was the baseline.32 In the text and figures, values are expressed as mean±SEM. GraphPad Prism version 7 (GraphPad Software, La Jolla, California) was used for statistical analyses. Basic membrane properties, latency to paroxysmal activity, and evoked responses of WT and R6/2 mice were analyzed with an unpaired Student's t test. Changes in sEPSC frequency after the incubation with BIC alone or with the glutamate transporter inhibitors were analyzed by two‐way repeated‐measures ANOVA followed by Sidak's multiple comparison tests. Average frequency of paroxysmal discharges was analyzed by two‐way ANOVA followed by Fisher's multiple comparison tests. For the effect of glutamate reuptake inhibition on eEPSCs, we analyzed the peak amplitude, the area of the response, and decay time using two‐way repeated‐measures ANOVA followed by Sidak's multiple comparison tests. Percent changes obtained after the incubation with DHK, UCPH‐101, or DL‐TBOA and changes in the cumulative interevent interval probability were analyzed by two‐way ANOVA followed by uncorrected Fisher's multiple comparison tests. Differences were considered significant when P<0.05.

3. RESULTS

3.1. Basic membrane properties of CPNs

CPNs in frontal cortex were recorded at a holding potential (V_hold) of −70 mV. CPNs from R6/2 mice displayed a significant decrease in mean cell membrane capacitance (unpaired Student's t test t = 3.72, df = 101, P = 0.0003, n = 53 WT and n = 50 R6/2), a significant increase in mean input resistance (t = 2.05, df = 101, P = 0.043), and a significant decrease in the mean decay time constant (t = 3.79, df = 101, P = 0.0003) (Table 1 ). These results suggest functional and morphological changes occur in frontal CPNs and are consistent with those observed in CPNs from other cortical areas in R6/134 and R6/2 mice.35

Table 1.

Passive membrane properties of CPNs in WT and R6/2 mice

	Cm (pF)	Rm (MΩ)	τ (ms)
WT (n = 53)	197 ± 11.9	139 ± 15.4	4.38 ± 0.2
R6/2 (n = 50)	144 ± 7.1*	180 ± 12.8*	3.34 ± 0.1*

Cm, cell membrane capacitance; Rm, input resistance; τ, decay time constant. Data are presented as mean±SEM. Data were analyzed by unpaired Student's t test.

*P < 0.05.

3.2. Evoked responses before and during partial or complete glutamate reuptake inhibition

We examined CPN responses evoked by electrical stimulation in the proximity (200‐300 μm) of the recorded cell to induce phasic release of glutamate. First, we obtained a basal response to an electrical stimulus (0.1 mA intensity, three sweeps applied every 20 s). As shown in Figure 1A–D, R6/2 CPNs (n = 20 cells, 60 replicates) displayed significant increases in peak amplitude, area, and decay time of eEPSCs as compared with WT CPNs (n = 33 cells, 99 replicates) [peak amplitude: t = 3.83, df = 157, P = 0.0002; area: t = 7.648, df = 157, P < 0.0001; decay time: t = 6.606, df = 157, P < 0.0001]. These results corroborate previous findings from our laboratory.35

Figure 1.

(A) Representative traces showing the glutamate receptor–mediated currents evoked by electrical stimulation (0.1 mA) in WT and R6/2 CPNs. The evoked response in the R6/2 CPNs is characterized by a significant increase in peak amplitude (B), area (C), and decay time (D). Data were analyzed by unpaired Student's t test. * Statistically different relative to the effect seen in WT CPNs (n = 20 WT and n = 33 R6/2 CPNs)

We next examined the effect of GLT‐1 inhibition on eEPSCs. As shown in Figure 2 A, in both WT and R6/2 CPNs, the incubation with DHK led to a slight decrease (~20%) in peak amplitude as compared with their own basal response (n = 12 WT, n = 7 R6/2), but the difference between genotypes was not statistically significant. Similarly, the area of evoked responses was reduced after incubation with the GLT‐1 blocker, but again there was no significant difference between genotypes (Figure 2B). The reduction in evoked response area was due to a combination of reduced amplitude and, unexpectedly, a decrease in eEPSC decay time (Figure 2C). This could suggest that when one glutamate transporter is blocked, the other can compensate and even overcompensate for the pharmacologically induced loss of function. In contrast to what we observed during the specific blockade of GLT‐1, inhibition of GLAST with UCPH‐101 induced very small increases in the peak amplitude in both WT and R6/2 CPNs (n = 11 WT and n = 7 R6/2), but the difference between genotypes was not statistically significant. However, similar to DHK, the GLAST inhibitor produced small decreases in decay times, which led to negligible changes in response area. Incubation with DL‐TBOA induced a slight decrease in eEPSC amplitude in both WT and R6/2 CPNs (n = 10 WT and n = 6 R6/2). However, the difference was not statistically significant between genotypes. In contrast, R6/2 neurons showed significantly larger areas and decay times relative to WTs (Figure 2B,C), as a consequence of the inhibition of both transporters, and this effect was significantly greater in CPNs from R6/2 mice (area: t = 3.049, df = 153, P < 0.002; decay time: t = 2.487, df = 153, P < 0.01). Figure 2D shows some example traces illustrating the significant increase in decay time in R6/2 CPNs after application of DL‐TBOA.

Figure 2.

Percent change of the evoked glutamate receptor–mediated responses in CPNs from WT and R6/2 mice induced by the incubation with DHK, UCPH‐101, or DL‐TBOA. While peak amplitude shows no change between the groups (A), CPNs from the R6/2 showed a significant increase in area (B) and decay time (C) after DL‐TBOA. ^*Statistically different relative to the effect seen in WT CPNs. Data were analyzed by two‐way ANOVA followed by Fisher's multiple comparison tests. DHK n = 12 WT and n = 7 R6/2; UCPH‐101 n = 11 WT and n = 7 R6/2; DL‐TBOA n = 10 WT and n = 6 R6/2 CPNs. (D). Representative eEPSCs recorded in CPNs from WT and R6/2 mice before and after incubation with DL‐TBOA

3.3. sEPSC frequency before and during glutamate reuptake inhibition

First, sEPSC frequency of CPNs was determined in the presence of BIC (5 μmol L⁻¹) alone. sEPSC frequency was similar between WT and R6/2 CPNs (Figure 3A, B; n = 21 WT and n = 35 R6/2, P = 0.87). Next, we evaluated the effect of DHK, UCPH‐101, and DL‐TBOA on the frequency of CPN sEPSCs. We observed that incubation with DHK tended to decrease the frequency of sEPSCs in R6/2 CPNs, but the resulting difference did not prove to be statistically significant (n = 9 WT and n = 9 R6/2; Figure 3A,C). A similar effect occurred after the selective inhibition of GLAST with UCPH‐101 (n = 5 WT and n = 16 R6/2; Figure 3A,C). Interestingly, the inhibition of both transporters with DL‐TBOA significantly decreased the sEPSC frequency in the R6/2 CPNs as compared to the response observed in WT CPNs (n = 7 WT and n = 10 R6/2, P = 0.036; Figure 3A,C). This effect is illustrated by the cumulative interevent interval histogram in Figure 3E.

Figure 3.

(A) Representative traces showing the effects of bicuculline (BIC; 5 μmol L⁻¹), BIC+DHK, BIC+UCPH‐101, and BIC+DL‐TBOA on spontaneous excitatory postsynaptic currents (sEPSCs) in CPNs from 10‐week‐old WT and R6/2 mice. (B) sEPSC frequency in WT and R6/2 CPNs in the presence of BIC. sEPSC frequency in CPNs was practically identical between WT and R6/2 mice. (C). DHK and UCPH‐101 failed to induce significant changes in sEPSC frequency. In contrast, DL‐TBOA induced a significant decrease in sEPSC frequency in CPNs from R6/2 mice. ^*Difference statistically significant relative to the effect seen in WT CPNs. (D). Variable percent changes in frequency occurred in WT and R6/2 CPNs after incubation with DHK or UCPH‐101. In contrast, DL‐TBOA consistently reduced sEPSC frequency and the reduction was greater in cells from R6/2 mice. Data were analyzed by two‐way repeated‐measures ANOVA followed by Sidak's multiple comparison tests. (E). Cumulative interevent interval probability distribution in R6/2 CPNs before and after incubation with DL‐TBOA. The inhibition of both transporters with DL‐TBOA induces a rightward shift, indicating reductions in sEPSC frequency. Data were analyzed by two‐way ANOVA followed by Fisher's multiple comparison tests (n = 10 CPNs)

We analyzed the changes in sEPSC amplitude and kinetics after the inhibition of glutamate transporters (Table 2). DHK incubation only reduced the amplitude (t = 3.13, df = 16, P = 0.006, n = 9 WT and n = 9 R6/2), while no changes in the rise time, decay time, and half‐amplitude duration occurred. Incubation with UCPH‐101 and DL‐TBOA failed to modify the amplitude or any of the kinetic parameters.

Table 2.

Amplitude and kinetics of sEPSCs after glutamate transporter blockade

	Ampl. (pA)	Rise (ms)	Decay (ms)	½‐Ampl. Duration (ms)
DHK
WT (n = 9)	14.1 ± 0.6	1.3 ± 0.1	5.0 ± 0.4	6.6 ± 0.5
R6/2 (n = 9)	11.9 ± 0.2*	1.2 ± 0.1	4.3 ± 0.2	6.1 ± 0.3
UCPH‐101
WT (n = 5)	12.4 ± 1.3	1.5 ± 0.1	5.1 ± 0.4	7.4 ± 0.6
R6/2 (n = 16)	13.5 ± 0.5	1.3 ± 0.1	4.5 ± 0.2	6.2 ± 0.3
DL‐TBOA
WT (n = 7)	13.2 ± 0.7	1.1 ± 0.12	4.3 ± 0.3	6.1 ± 0.4
R6/2 (n = 10)	14.6 ± 0.9	1.0 ± 0.06	4.0 ± 0.2	5.3 ± 0.3

Data are presented as mean ± SEM. Data were analyzed by unpaired Student's t test.

*P < 0.05.

We also observed that some CPNs (WT = 33% and R6/2 = 45%) developed paroxysmal, epileptiform activity as a consequence of the application of glutamate reuptake blockers in the presence of BIC. We quantified the frequency of paroxysmal events (>400 pA) induced by DHK, UCPH‐101, and DL‐TBOA. CPNs from R6/2 mice were more prone to develop paroxysmal‐like activity during inhibition of either of the glutamate transporters (Figure 4 ), but this increase was statistically significant only during full inhibition of the glutamate reuptake system by DL‐TBOA (P = 0.048, n = 6 WT and n = 10 R6/2; Figure 4A and B). In addition, the latency to induce paroxysmal activity in the R6/2 CPNs was significantly shorter relative to WT CPNs (Figure 4C; t = 4.05, df = 14, P = 0.0014).

Figure 4.

Inhibition of glutamate transporters and paroxysmal activity. (A) Traces show an example of paroxysmal activity in WT and R6/2 CPNs during the incubation with DL‐TBOA. More discharges were seen in R6/2 CPNs. (B) Frequency of paroxysmal events obtained with DHK, UCPH‐101, and DL‐TBOA. The frequency of paroxysmal discharges was significantly increased in R6/2 CPNs. Data analyzed by two‐way ANOVAs followed by Fisher's multiple comparison tests. DHK n = 9 WT and n = 9 R6/2; UCPH‐101 n = 5 WT and n = 16 R6/2; DL‐TBOA n = 6 WT and n = 10 R6/2. (C) The latency to induce paroxysmal activity in R6/2 CPNs was significantly lower relative to WT. Data analyzed by unpaired t test, P = 0.0073). ^*Statistically different relative to the effect seen in WT CPNs

4. DISCUSSION

In the present study, we show that CPNs from R6/2 mice display an overall increase in evoked responses (amplitude, area, and decay time) compared to WTs indicating that R6/2 CPNs are more prone to develop overactivation of glutamate receptors than their WT counterparts. Perhaps this might explain the exacerbation of paroxysmal activity during glutamate reuptake block with DL‐TBOA. Further, our data demonstrate that while partial blockade of each glutamate transporter produces variable changes in peak amplitude, area, and decay time, full blockade with DL‐TBOA consistently increases evoked response area (relative charge) and decay time, and this increase is significantly greater in CPNs from R6/2 mice. In addition, we demonstrate that partial inhibition of the glutamate reuptake system in R6/2 CPNs with either DHK or UCPH‐101 fails to modify the frequency of sEPSCs. However, simultaneous inhibition of both GLT‐1 and GLAST with DL‐TBOA reduces the frequency of sEPSCs. Altogether our results suggest that CPNs in the R6/2 mice are more sensitive to compromised glutamate reuptake and that failure in this system might lead to overactivation of glutamate receptors in the frontal cortex.

Our results demonstrate for the first time that the frontal cortex also is susceptible to glutamate reuptake failure, which was reported previously in the striatum of R6/2 mice.21 Moreover, as frontal CPNs in the R6/2 are more prone to paroxysmal activity, this brain area might be a trigger for the development of epileptic seizures observed in R6/2 mice.35, 36 This is consistent with the finding that with disease progression, there is a progressive increase in sEPSC frequency in CPNs which is associated with reduced inhibition.35 Interestingly, application of glutamate reuptake blockers reduced the frequency of sEPSCs in R6/2 CPNs and this effect was statistically significant when both transporters were blocked with DL‐TBOA. Reduced sEPSC frequency could represent a particular form of homeostatic synaptic downscaling37 as a result of failure of glutamate reuptake from the synaptic cleft and designed to prevent excessive activation of postsynaptic glutamate receptors. An alternative explanation could involve activation of metabotropic receptors located on the presynaptic terminals. Indeed, studies have shown that inhibition of glutamate reuptake causes a decrease in synaptic release of glutamate due to overactivation of these receptors.12, 38 Why with disease progression these homeostatic mechanisms fail to reduce synaptic activity remains unknown. However, it was recently demonstrated that RE1‐silencing transcription factor (REST) plays a critical role for the downscaling of neuronal hyperexcitability,39, 40 and in HD, the function of REST is greatly disrupted.41

In control conditions, before any drug application, evoked responses of R6/2 CPNs showed an increase in amplitude, area, and decay time compared with WTs, suggesting that glutamate transport is impaired in R6/2 CPNs. It is interesting that when both glutamate transporters are blocked by DL‐TBOA, we only observed an increase in area and decay time, but not in the amplitude of the response. One explanation for this outcome could be that a ceiling effect has occurred in R6/2 CPNs. In striatal neurons from HD mice, we have shown that acute blockade consistently increases decay time, whereas changes in amplitude of the response are more variable.32 In addition, after acute application of DL‐TBOA, rapid adaptive changes in the number of receptors located at the postsynaptic terminal may occur. A redistribution and concomitant decrease in the number of postsynaptic glutamate receptors has been suggested as a compensatory mechanism induced by increased neuronal activity or ambient glutamate.42, 43 For example, addition of glutamate to the culture medium induced a pronounced decrease in the frequency of miniature EPSCs in hippocampal neurons.42 Interestingly, relevant to our present findings, recent evidence has demonstrated that the mechanisms of spontaneous and evoked glutamate release differ in many respects.44, 45 For example, single neuron responses to either evoked or spontaneous activation showed a differential localization and dynamics of postsynaptic responses.46 Therefore, glutamate reuptake inhibition could affect spontaneous and evoked EPSCs differentially.

Impaired function of the glutamate reuptake system has been described in HD, with GLT‐1 as the main focus [for review, see47]. Reduced functioning and protein expression of this transporter in the striatum and cerebral cortex have been described in both human HD patients and transgenic models.16, 17, 18, 19, 20, 48 In the R6/2 model, a decrease in mRNA and protein levels of GLT‐1 has been observed in the striatum and cortex at about 10 weeks of age,17, 18 although reduced function can occur as early as 8 weeks of age.49 However, despite the fact that several studies indicate early changes in the function of GLT‐1, the consequences of such alterations during glutamate neurotransmission remain unclear. Alterations of GLT‐1 seem to play a critical role in the neuropathology of HD since the sole expression of mHtt in astrocytes is enough to induce an HD‐like behavioral phenotype and decreased functioning of GLT‐1.50 Aberrant GLT‐1 functioning might compromise the dynamics of synaptic transmission, as recently described.15 The authors showed that the membrane diffusion of GLT‐1 is a critical factor in shaping synaptic transmission. Based on these findings, it is possible that deficient function of this transporter may underlie some of the altered neuronal processing described in R6/2 mice and in other HD mouse models.9, 23, 24, 25, 48 As previously reported, CPNs in the R6/2 model present changes in GLT‐1 as early as eight weeks of age.24 In this study, we examined the effects of inhibiting GLT‐1 with the selective inhibitor DHK and evaluated the changes in the frequency of sEPSCs. Interestingly, no significant changes in the frequency of sEPSCs or the presence of paroxysmal activity occurred in R6/2 mice. Considering that recording sEPSCs may be a passive way to measure the CPN response to further inhibition of GLT‐1, we decided to challenge the neurons by applying electrical stimuli that would lead to the phasic release of glutamate and potential spillover from the synaptic cleft. Surprisingly, instead of increasing the amplitude, area, and decay time of the evoked response, which typically occurs when a high concentration of glutamate is present in the synaptic cleft, we observed the opposite; DHK inhibition of GLT‐1 induced a small reduction in peak amplitude, area, and decay time in both WT and R6/2 CPNs. This effect suggests that a compensatory mechanism might have taken place and that GLAST could compensate for the compromised functioning of GLT‐1. Interestingly, changes in protein content of GLAST have been less consistent. While some studies reported no change,17, 18 others have found reduced levels in R6/2 mice at 14 weeks of age but not earlier.21 In line with this observation, the sole inhibition of GLAST with UCPH‐101 failed to reduce the glutamate response peak amplitude, as occurred during the selective blockade of GLT‐1. The possibility that upregulation of GLAST function compensates for GLT‐1 deficiency might serve to clarify some controversies regarding GLT‐1 function in HD. For example, Petr and colleagues described that reduced expression of GLT‐1 in the R6/2 model does not worsen the progression of the phenotype.51 Moreover, an in vitro experiment evaluating glutamate clearance in striatum showed normal and even accelerated striatal reuptake in the R6/2 model.52 More recently, it was demonstrated that GLT‐1 inhibition by DHK slows clearance rates to only a small fraction (~20%) of that produced by DL‐TBOA, suggesting that non‐GLT‐1 transporters make a significant contribution to glutamate clearance, in particular when GLT‐1 is dysfunctional, as occurs in HD.53 These observations are compatible with our finding that, in spite of selective inhibition of GLT‐1 in HD frontal cortex, the remaining transporter appears to be working more efficiently. It also suggests that slower eEPSC decay time in R6/2 mice is not due exclusively to deficient glutamate transport but also possibly due to postsynaptic changes in glutamate receptor subunits and localization.54

These results emphasize that more experiments are needed to understand further the role of GLAST in HD neuropathology. If GLAST can cope and compensate for the failure of GLT‐1, it might be a possible target for future therapeutic approaches. Interestingly, studies in primary astrocyte cultures from cortex demonstrated that expression of mHtt in astrocytes significantly decreased both mRNA and protein levels of GLT‐1 but not those of GLAST.55

5. CONCLUSIONS

In summary, our results support the emerging understanding of the role of astrocytes in the neuropathology of HD.56 Changes in the expression and functioning of both glutamate transporters might play a critical role during the onset and progression of HD. However, further studies are necessary to fully understand the effect of GLT‐1 and GLAST alterations on neuronal functioning in vivo. Finally, the study of GLAST as it pertains to HD might bring forth a broader understanding of the roles of the glutamate reuptake system and astrocytes during HD pathogenesis.

DISCLOSURE

The authors declare no competing financial interests.

AUTHOR CONTRIBUTIONS

AME‐S, MSL, and CC conceived and designed the experiments. AME‐S, DC, KP‐O, KJ, and MN‐H performed the experiments. AME‐S, DC, KP‐O, KJ, and MN‐H analyzed the data. AME‐S, DC, KP‐O, KJ, and MN‐H prepared the figures and the first draft of the manuscript. AME‐S, CC, and MSL developed the final version of the manuscript with input from DC, KP‐O, KJ, and MN‐H. All authors approved the final version of this manuscript.

Estrada‐Sánchez AM, Castro D, Portillo‐Ortiz K, et al. Complete but not partial inhibition of glutamate transporters exacerbates cortical excitability in the R6/2 mouse model of Huntington’s disease. CNS Neurosci Ther. 2019;25:509–518. 10.1111/cns.13070