Isovaline Does Not Activate GABAB Receptor-Coupled Potassium Currents in GABAB Expressing AtT-20 Cells and Cultured Rat Hippocampal Neurons

Isovaline is a non-proteinogenic amino acid that has analgesic properties. R-isovaline is a proposed agonist of the γ-aminobutyric acid type B (GABAB) receptor in the thalamus and peripheral tissue. Interestingly, the responses to R-isovaline differ from those of the canonical GABAB receptor agonist R-baclofen, warranting further investigation. Using whole cell recording techniques we explored isovaline actions on GABAB receptors coupled to rectifying K+ channels in cells of recombinant and native receptor preparations. In AtT-20 cells transfected with GABAB receptor subunits, bath application of the GABAB receptor agonists, GABA (1 μM) and R-baclofen (5 μM) produced inwardly rectifying currents that reversed approximately at the calculated reversal potential for K+ R- isovaline (50 μM to 1 mM) and S-isovaline (500 μM) did not evoke a current. R-isovaline applied either extracellularly (250 μM) or intracellularly (10 μM) did not alter responses to GABA at 1 μM. Co-administration of R-isovaline (250 μM) with a low concentration (10 nM) of GABA did not result in a response. In cultured rat hippocampal neurons that natively express GABAB receptors, R-baclofen (5 μM) induced GABAB receptor-dependent inward currents. Under the same conditions R-isovaline (1 or 50 μM) did not evoke a current or significantly alter R-baclofen-induced effects. Therefore, R-isovaline does not interact with recombinant or native GABAB receptors to open K+ channels in these preparations.


Introduction
Receptors for γ-aminobutyric acid (GABA), especially of the GABA B type, are a promising target in analgesia. However, the use of the prototypical GABA B agonist, baclofen, is limited due to the development of severe side effects [1]. Isovaline is an unusual non-proteinogenic amino acid that is anti-nociceptive without side effects typical of GABA B agonists [2,3]. A component of the analgesia produced by isovaline is attributable to activation of GABA B receptors [2]. GABA B receptors are obligate heterodimers, made up of a GABA B1 and a GABA B2 subunit [4][5][6]. Surprisingly, there are no pharmacologically distinct GABA B receptor subtypes [7,8]. Critical residues for orthosteric agonist and antagonist binding are located on the GABA B1 subunit within a region known as the Venus flytrap domain [9][10][11], while allosteric modulators of the GABA B receptor act at sites on the GABA B2 subunit [12]. GABA B receptors are G-protein coupled receptors that associate with the pertussis toxin sensitive G αi/o family of G-proteins [13]. Conventional cellular effects of GABA B receptor agonism include inhibition of adenylate cyclase, G βγ -mediated activation of G-protein coupled inwardly rectifying K + (GIRK) channels and G βγ -mediated inhibition of voltage gated Ca 2+ channels [8,14,15].
Isovaline's action at GABA B receptors is atypical as demonstrated in thalamic neurons of brain slices. Compared to R-baclofen, R-isovaline-evoked K + currents are slow in onset and long lasting [16,17]. Also in contrast to R-baclofen, the response to R-isovaline is blocked by pre-application of a GABA B receptor antagonist, but not by application of the GABA B antagonist subsequent to the initiation of the K + current [16]. Furthermore, some neurons that respond to R-baclofen do not respond to R-isovaline [16]. The ability of R-isovaline to activate currents in sensitive neurons does not appear to result from GABA release and subsequent postsynaptic activation of GABA B -mediated K + currents, since inhibition of other GABA-mediated currents does not alter R-isovaline responses [16]. Here we test the effects of R-isovaline on GABA B receptors in a neuronal expression system as well as in isolated hippocampal neurons natively expressing GABA B receptors.

AtT-20 cell culturing, transfection and electrophysiology
Mouse pituitary AtT-20 cells [18,19] obtained from Dr. C. Chavkin's laboratory at the University of Washington (March, 2010) were grown in Gibco high glucose Dulbecco's modified Eagle medium (Invitrogen) with 1% fetal bovine Serum, 10% horse serum, penicillin-streptomycin and 0.2 mM L-glutamine and kept in an incubator at 37°C with 5% CO 2 . Cells were passaged every 2-3 days. For transient transfection cells were plated in a 35mm culture dish at 90% confluency and left for 24 hours. Cells were then co-transfected overnight in serum containing media using Lipofectamine 2000 (Invitrogen, Life technologies, Burlington, ON) following the manufacturer's protocols. Cells were co-transfected with cDNA for the GABA B1a and GABA B2 subunits as well as green fluorescent protein (GFP) to aid identification of transfected cells (1:2:5 ratio of GFP:GABA B1a :GABA B2 ). The following day the transfection media were removed and cells were re-plated onto poly-D-lysine coated glass cover slips which were then incubated for a further 24 hours before use.
To confirm membrane expression of the GABA B receptor in cells that were suitable for patching (isolated from one another), we performed immunohistochemistry for the GABA B1 subunit which is unable to traffic to the membrane without associating with a GABA B2 subunit [4][5][6]. Cells were fixed with 4% formaldehyde for 30 minutes then blocked with 0.5% bovine serum albumin for 1 hour. Cells were incubated with or without mouse anti-GABA B1 (1:1000 dilution, ab55051, Abcam, Toronto) overnight at 4°C. Goat anti-mouse conjugated to Alexa Fluor 546 (Life technologies) was applied at a 1:200 dilution for 1 hour at room temperature. DAPI (4',6-diamidino-2-phenylindole) was applied at 1:10,000 dilution for 5 minutes to stain the nuclei. Coverslips were washed with phosphate buffered saline 3 times between each step. The glass coverslips were mounted onto glass slides and left to dry for at least 72 hours. Confocal microscopy was performed using an Olympus FluoView FV1000 confocal microscope with a 60x objective (Tokyo, Japan).
For electrophysiological recordings, cells were removed from the incubator and the medium was replaced at room temperature with high K + extracellular solution containing (in mM): NaCl (130), KCL (20), CaCl 2 (2), MgCl 2 (1), HEPES (10), glucose (35), and 0.1-1 μM tetrodotoxin. The solution had a pH = 7.4 (adjusted with NaOH) and an osmolality of 330 ± 5 mOsm. The glass coverslip was cut into sections and placed into a 50 μL fast exchange diamond bath (Warner Instruments, Hamden, CT) constantly perfused (*2 ml/min) with extracellular solution. Recording electrodes were made out of thin-walled borosilicate glass (World Precision Instruments, Sarasota, FL) and had a resistance between 3-5 MO when filled with an intracellular pipette solution containing (in mM): K-gluconate (130), NaCl (20), MgCl 2 (1), EGTA (10), glucose (10), HEPES (10), MgATP (5) and Na 3 GTP (0.1) with pH = 7.2 (NaOH/ HCl) and an osmolality of 310 ± 3 mOsm. The calculated Nernst equilibrium potential for K + was approximately -47 mV. Cells were visualized with an Axiovert 25 inverted fluorescent microscope (Zeiss, Germany). Whole cell recordings from fluorescing cells were performed using a List EPC 7 amplifier (HEKA, Germany). Recordings were filtered at 3 kHz, digitized at 10 kHz and analysed using pClamp 8.2 software (Molecular Devices, Sunnyvale, CA). Cells were voltage-clamped at -50 mV. The current-voltage (IV) relationship was measured each minute by applying a family of 400 ms voltage steps, from -100 mV to +20 mV, in 10 mV increments. A junction potential of -12.5 mV was accounted for offline. The current was measured as the average current in the last 100 ms of each voltage step. In all experiments except one, drugs were applied via the bath for 2-5 min, and then washed off. In one experiment R-isovaline was added into the intracellular pipette solution. Cells were held in voltage clamp for an extra 5 min before further drug application to allow for intracellular dialysis. In several cells GABA B action was confirmed by blockade with a GABA B antagonist.

Hippocampal cell culturing and electrophysiology
Cultured hippocampal neurons were prepared from 18 day old Sprague-Dawley rat embryos (male and female) following methods described in detail by Xie et al. [20]. Rat housing and culture preparation methods were in accordance with the Canadian Council on Animal Care and approved by the University of British Columbia Animal Care Committee. Pregnant rats were sacrificed with 5% CO 2 and care was taken to ensure an absence of nociceptive response prior to decapitation and removal of embryos. Neurons were plated on poly-L-lysine coated glass coverslips at a density of 130 cells/mm 2 and were incubated with 5% CO 2 at 37°C for 14 to 28 days until used for electrophysiological analysis in vitro.
On the day of recording cultured hippocampal neurons were removed from the incubator and the glass coverslip was placed into a culture dish filled with room temperature (*22°C) extracellular solution containing (in mM): NaCl (140), KCl (5.4), CaCl 2 (2), MgCl 2 (1), HEPES (20) and glucose (20) at pH = 7.4 and 325 ± 5 mOsm. In some experiments, hippocampal neurons were incubated with R-isovaline (50 μM) in extracellular solution for at least an hour before recordings. Recording electrodes (4-5.5 MO) were filled with a solution containing (in mM): K-gluconate (120), KCl (20), NaCl (10), HEPES (10), EGTA (5) MgATP (3) and Na 2 GTP (0.2) at pH 7.2 and 305 ± 5 mOsm. Cells were voltage clamped at -80 mV and the membrane current was constantly recorded. To enhance inward K + currents, a high [K + ] extracellular solution, whereby 20 mM NaCl was replaced with equimolar KCl, was used to examine drug effects. Drugs were mixed with the high [K + ] extracellular solution and applied via the bath perfusate for a minimum of 30 s and a maximum of 5 min. The average current during 20 s of the peak drug effect was measured and compared with the current in the last 20 s before the drug was washed in (baseline). In the case of no obvious drug effect, the average current within the last 20 s before drug wash-off was measured. Recordings were filtered at 3 kHz, digitized at 10 kHz and analysed using pClamp 8.2 software.

Drugs
Extracellular solution was prepared on the day of the experiment. Intracellular solution was prepared and frozen in aliquots and used within a week. Stock solutions of R-baclofen HCl (25 mM, Sigma-Aldrich, St. Louis, MO), GABA (50 mM, Sigma-Aldrich), CGP 52432 (10 mM, Tocris, UK) and R-isovaline HCl or S-isovaline HCl (100 mM, BioFine International, Vancouver, BC), were made using double distilled H 2 0 and kept at 4°C. Final drug concentrations were prepared on the day of experiment.

Analysis and Statistics
For data from AtT-20 cells, agonist action at the GABA B receptor was defined as the ability to induce a current that was inwardly rectifying and reversed at approximately E K . The net current was determined by subtracting the baseline current from the current during drug application at each voltage step. The IV relationships were then plotted to determine whether the net current was inwardly rectifying. To determine the reversal potential, the linear section of the baseline subtracted IV curve (-112.5 mV to -52.5 mV) for each recording was fitted using linear regression, the point at which Y = 0 for each curve was determined and mean ± SEM and 95% confidence interval calculated. For clarity, example currents from AtT-20 cells depict only the current recorded at the maximum hyperpolarising step (-112.5 mV). Bar graphs display the untransformed current at the maximum hyperpolarising step (-112.5 mV) for AtT-20 cells or at the holding potential (-80 mV) for hippocampal cells. For analysis of the effect of intracellular R-isovaline, the baseline current was measured by stepping to -112.5 mV from a holding potential of -62.5 mV at t = 1 min. This baseline value was then subtracted from all subsequent measurements. Repeated measures 1 or 2-way ANOVAs with Bonferroni's post-hoc test were used as appropriate to test for drug induced changes in currents. For analysis of changes in the R-baclofen-evoked current in hippocampal cells, the baclofen-evoked current (net current) was determined by subtracting the baseline current from the current during application of Rbaclofen. A 1-way ANOVA was used to compare the magnitude of R-baclofen-evoked currents. Data are expressed as mean ± SEM. "n" refers to the number of cells. Results were considered significant if p < 0.05.
We then transiently transfected AtT-20 cells with the GABA B1a and GABA B2 subunits and GFP. To confirm membrane expression of GABA B receptors in AtT-20 cells, we used a mouse monoclonal antibody to the GABA B1 subunit, which is unable to traffic to the cell surface without GABA B2 [4][5][6]. Fig. 2A,B illustrates cell surface expression of GABA B1 and colocalisation with GFP alongside the negative control. Next, we assessed if AtT-20 cells transiently transfected with both the GABA B1a and GABA B2 subunits responded to GABA, R-baclofen or R-and S-isovaline. GABA (1 μM) applied for 2 min produced an inwardly rectifying current that was antagonised by co-application with 1 μM CGP 52432 (Fig. 3A). GABA evoked a maximum current of -41.0 ± 9.7 pA. Co-application of GABA and a GABA B antagonist, CGP 52432, reduced the current to -8.0 ± 4.1 pA (paired t-test, t = 4.316, df = 6, p = 0.005, n = 7). The reversal potential of the GABA evoked current was -52 ± 3 mV (95% CI: -59 mV to -45 mV, n = 7).
To test if R-isovaline could modulate GABA action, we applied GABA (1 μM) in the presence or absence of R-isovaline (250 μM). As shown in Fig. 4A,B there was no statistically significant difference in the currents with application of GABA alone (-82.9 ± 16.3 pA) compared to GABA + R-isovaline (-81.5 ± 18.1 pA, repeated measures 1-way ANOVA with Bonferroni's  To determine if R-isovaline is a positive modulator, GABA was applied at 1/100 of an effective concentration in the presence of R-isovaline after confirming a response to a high concentration of GABA. As shown in Fig. 4A,C, co-application of R-isovaline with a low concentration of GABA did not significantly change the measured current as determined by repeated measures 1-way ANOVA with Bonferroni's post hoc To investigate the possibility that R-isovaline could only induce a current by an intracellular action, we applied R-isovaline via the intracellular pipette solution to AtT-20 cells while measuring the GABA-evoked inward current. We assumed that the concentration of R-isovaline that would transport across the cell membrane would be lower than that applied extracellularly. Therefore, we used a lower concentration for these experiments (10 μM). There was no difference in the currents recorded with electrodes containing R-isovaline or control solution as determined by a 2-way ANOVA (main effect of R-isovaline F (1, 105) = 1.77, p = 0.2, n = 6 for both electrode groups, Fig. 5A,B). The difference in the GABA-evoked current recorded with electrodes containing R-isovaline or control solution was not statistically significantly different (control -59.5 ± 33.6 pA vs. R-isovaline -58.0 ± 17.7 pA, Bonferroni's post hoc test, p>0.05).

Discussion
Our studies show that R-isovaline did not induce an inward current in AtT-20 cells heterologously expressing GABA B receptors or in cultured hippocampal neurons natively expressing GABA B receptors. While GABA and baclofen activated GABA B receptors, co-application with R-isovaline did not occlude or modulate GABA or baclofen responses. The data are in contrast to reports that show R-isovaline-mediated activation of GABA B receptors [2,16]. However, others have demonstrated that isovaline does not alter the postsynaptic electrical properties of hippocampal pyramidal neurons [22].
Mouse pituitary AtT-20 cells contain endogenous GIRK channels [23], a common effector channel of GABA B receptors [15]. In the present experiments, AtT-20 cells transiently transfected with GABA B receptor subunits responded to GABA or R-baclofen with an inwardly rectifying current that reversed near the calculated equilibrium potential for K + . GABA B receptors mediated the inward currents as confirmed by antagonism with CGP 52432. R-baclofen also induced a GABA B receptor-dependent inward current in cultured hippocampal neurons, attributable to activation of GIRK channels [21]. In contrast to the GABA-or R-baclofen-induced activation of endogenously expressed GIRK channels, R-isovaline did not produce a response in either cell type. Previous work has demonstrated that R-isovaline induces a large increase in GABA B -mediated K + conductance in thalamocortical neurons of brain slices, albeit with slower response kinetics than with R-baclofen [16,17]. One explanation for the difference may be that R-isovaline only acts at a specific isoform of the GABA B receptor. GABA B receptors are obligate heterodimers consisting of a , R-baclofen + CGP 52432 (n = 7) and R-isovaline at low (1 μM; n = 3) and high (50 μM; n = 8) concentrations (shaded bars). b) Example shows R-baclofen-mediated current c) Example shows lack of effect of low concentration of R-isovaline in a cell that responded to R-baclofen both before and after R-isovaline application. d) Example shows the effect of R-baclofen after 5 min pretreatment with 50 μM R-isovaline. e) Example shows the effects of two applications of R-baclofen after prolonged incubation (> 1 hour) with R-isovaline. f) Graph shows the magnitude of the current evoked by 5 μM R-baclofen in control extracellular solution (n = 8), after 5 minutes pretreatment with R-isovaline (n = 4) or after > 1 hour pretreatment with R-isovaline (n = 8). Data represent mean ± SEM. ** = P < 0.01. doi:10.1371/journal.pone.0118497.g006 GABA B1 and GABA B2 subunit [4][5][6]24]. While there is only one isoform of the GABA B2 subunit, two isoforms of the GABA B1 subunit are functionally expressed in the mammalian central nervous system [7]. If R-isovaline interacted selectively with GABA B receptors containing the GABA B1b subunit, we would not observe an effect because AtT-20 cells were transfected with cDNA coding only for the GABA B1a subunit. However, GABA B1b subunits differ only on the N terminus, a region that has been shown not to affect ligand binding or receptor function [9,25,26]. There are no known pharmacological or functional differences between GABA B1aor GABA B1b -subunit containing GABA B receptors upon heterologous expression, thus, a selective action at a specific isoform of the GABA B1 subunit is unprecedented [7,26].
Another explanation for the failure of isovaline to activate GABA B receptors in our studies is that its action could depend on an alternative protein that associates with the GABA B receptor or signalling components. GABA B receptors can display pharmacological and functional differences depending on cell type and subcellular location. This is due to variation in cell specific proteins that regulate GABA B receptors or their responses [27][28][29][30][31][32][33][34]. Cooke et al. (2012) [16] suggested a cell-specific mechanism of action in thalamic slices because significantly fewer neurons responded to R-isovaline than to R-baclofen; in addition, a subset of neurons responded to R-baclofen but not R-isovaline [16]. Candidate cell specific proteins include potassium channel tetramerization domain (KCTD) proteins, which directly associate with GABA B receptors to alter ligand affinity and response kinetics such as desensitization [27]. However, differential expression of KCTD proteins may not explain a lack of effect of R-isovaline on K + currents in hippocampal pyramidal neurons because the hippocampus has been reported to express all 3 KCTD proteins that act as GABA B auxiliary subunits [27]. Alternatively, specific regulators of G-protein signalling proteins that alter GABA B receptor efficacy [28,31] may be required for R-isovaline to have sufficient efficacy to evoke GABA B -mediated GIRK currents. It also is possible that isovaline acts as a biased agonist [35] at GABA B receptors in such a way that it does not influence the membrane delimited coupling of the GABA B receptor to GIRK channels. Instead, isovaline may initiate GABA B -mediated signalling cascades that indirectly influence membrane conductances, for example activation of Src-kinases [36,37].
Isovaline's effects, particularly in the central nervous system, may depend on cellular location. For example, isovaline has anti-epileptic properties which are not likely GABA B -mediated, but instead are postulated to result from a selective enhancement of hippocampal interneuronal activity by non-synaptically increasing inhibitory input and/or eliciting a shunting phenomenon onto pyramidal neurons [22,38]. Our experiments demonstrated no effect of isovaline on hippocampal pyramidal neurons, but do not exclude isovaline actions on interneurons.
In summary, these studies demonstrate that R-isovaline does not activate GABA B -mediated GIRK currents in AtT-20 cells or isolated hippocampal pyramidal neurons. Furthermore, intracellular or extracellular application of R-isovaline does not occlude or modulate the actions of other GABA B receptor agonists in these isolated cell systems. Future studies should be aimed at determining if cell specific modulators of GABA B receptor signalling are required for isovaline-induced K + currents in thalamic and other responsive neurons.