In Intact Islets Interstitial GABA Activates GABAA Receptors That Generate Tonic Currents in α-Cells

In the rat islets γ-aminobutyric acid (GABA) is produced by the β-cells and, at least, the α-cells express the GABAA receptors (GABAA channels). In this study, we examined in intact islets if the interstitial GABA activated the GABAA receptors. We used the patch-clamp technique to record whole-cell and single-channel currents and single-cell RT-PCR to identify the cell-type we recorded from, in the intact rat islets. We further identified which GABAA receptor subunits were expressed. We determined the cell-type of 43 cells we recorded from and of these 49%, 28% and 7% were α, β and δ-cells, respectively. In the remaining 16% of the cells, mRNA transcripts of more than one hormone gene were detected. The results show that in rat islets interstitial GABA activates tonic current in the α-cells but not in the β-cells. Seventeen different GABAA receptor subunits are expressed with high expression of α1, α2, α4, α6, β3, γ1, δ, ρ1, ρ2 and ρ3 subunits whereas no expression was detected for α5 or ε subunits. The abundance of the GABAA receptor subunits detected suggests that a number of GABAA receptor subtypes are formed in the islets. The single-channel and tonic currents were enhanced by pentobarbital and inhibited by the GABAA receptor antagonist SR-95531. The single-channel conductance ranged from 24 to 105 pS. Whether the single-channel conductance is related to subtypes of the GABAA receptor or variable interstitial GABA concentrations remains to be determined. Our results reveal that GABA is an extracellular signaling molecule in rat pancreatic islets and reaches concentration levels that activate GABAA receptors on the glucagon-releasing α-cells.


Introduction
The pancreatic islets consist of four major cell types: the glucagon-secreting a-cells, the insulin-secreting b-cells, the somatostatin-secreting d-cells and the polypeptide-producing PPcells. In addition to the hormones, the cells release small molecules that may act in an auto or paracrine manner [1,2]. Gammaaminobutyric acid (GABA) is an extracellular signal molecule in the islets [3,4,5,6]. GABA is formed by the enzyme glutamate decarboxylase (GAD) which catalyses the formation of GABA from glutamate and is located both in the cytoplasm and in synaptic-like vesicles [7,8,9,10,11]. Once released, GABA is thought to act in an auto and paracrine manner on the islet cells to modulate hormone secretion [6,12,13,14,15,16,17].
GABA activates ionotropic GABA A and metabotropic GABA B receptors in the plasma membrane of the islet cells [6,14,18]. In the rat islet, only the a-cells express the GABA A receptors (GABA A channels) [19], whereas in human pancreatic islets, the a, b and dcells all have GABA A receptors [12,17]. There are many subtypes of GABA A receptors whereas only one GABA B receptor has been described so far [20]. The GABA A receptors are pentameric. The subunits are grouped into eight families (a1-6, b1-3, c1-3, d, e, h, p, r1-3) and the receptors commonly contain at least 3 different types of subunits: 2 as, 2 bs and a third subunit-type. The physiological and pharmacological properties of the receptors are determined by the subunit-types that form the GABA A receptors [21].
When GABA binds to the GABA A receptor, a chloridepermeable ion channel is opened. The activation of GABA A channel is best studied in the central nervous system where the receptors evoke phasic (transient) and tonic (long-lasting) inhibition. Phasic activation is mediated by synaptic GABA A receptors and is triggered by the transient, high concentration of GABA (,mM) released from the presynaptic terminal whereas tonic activation of the extrasynaptic receptors is evoked by the ambient GABA concentration present around the neuron [22]. In the rat bcells, the vesicular release of GABA coincides with the release of the insulin containing granules when the cell is exposed to high glucose stimulation [11] whereas the non-vesicular release of GABA appears to take place both in high and low glucose concentration [23]. This raises the question of the mode of activation of the GABA A receptors in the pancreatic islet. So far, most of the electrophysiological studies of GABA A receptors in pancreatic islet cells have been conducted on dispersed cells [6] or transfected cells overexpressing GABA A receptors [12,19]. These studies have, therefore, not resolved the mode of GABA A receptors activation in intact islets. One reason why physiological experiments have predominantly used dispersed cells is related to the difficulty of identifying the cell-types in intact islets. Here we have used the method of single-cell RT-PCR to distinguish the type of cell we recorded from. Our results show, in intact rat pancreatic islets, that interstitial GABA generates tonic currents in the a-cells when the islets are exposed to 20 mM glucose. The tonic current can be enhanced by pentobarbital and inhibited by SR-95531, both drugs specific for GABA A receptors.

Preparation of Pancreatic Islets
The experiments were carried out on intact rat pancreatic islets isolated from 50-52 days old Wistar rats. Care and use of animals were in accordance with local ethical guidelines and approved by the Uppsala Djurförsöksetiska Nä mnd, Sweden (Uppsala's animal ethics committee). Isolation of pancreatic islets followed a collagenase digestion procedure [24]. Islets were picked by hand with a regular pipette. Isolated islets were cultured for no more than five days in RPMI 1640 medium (Sigma, St. Louis, MO) containing 10 mM glucose, supplemented with 10% (vol/vol) fetal calf serum, benzylpenicillin (100 U/ml), and streptomycin (0.1 mg/ml). The medium was replaced every second day.

Whole-cell Recording and Current Analysis
To record from rat islets, the intact islet was held by a holding pipette and approached by the recording pipette from the other side [17,25]. Establishing a GV seal on cells was mainly attempted on the surface of the rat islets where the most a-cells are located. After establishing a whole-cell patch-clamp configuration, currents were recorded.
All patch-clamp recordings were performed at room temperature (20-22uC). All drugs used were purchased from Sigma-Aldrich (Steinheim, Germany) or Ascent Scientific (Bristol, UK). Pentobarbital and SR-95531 were dissolved in the extracellular solution. For the voltage-clamped, gap-free recordings, pancreatic islets were perfused in the standard extracellular solution consisting of (mM): 138 NaCl, 5.6 KCl, 2.6 CaCl 2 , 1.2 MgCl 2 , 5 HEPES (pH 7.4 using NaOH) and 20 mM glucose. For the recording of the current-voltage (IV) relationship, 20 mM tetraethyl-ammonium (TEA) chloride replaced 20 mM NaCl in the extracellular solution. The pipette solution consisted of (mM): 125 CsCl, 30 CsOH, 1 MgCl 2 , 10 EGTA, 5 HEPES (pH 7.15 with HCl), and 3 Mg-ATP. The recording pipettes were made from borosilicate glass and had the resistance of 5-8 MV when filled with the pipette solution. Recordings were made at 270 or 290 mV. The access resistance was monitored and if it changed by more than 25%, the recording was rejected. Current-voltage (IV) recordings were done in the presence and absence of SR-95531, a GABA A receptor antagonist. The net current activated by the interstitial GABA is calculated by subtractions of current before and after application of SR-95531 (I -I SR-95531 = I GABA ). To record the IV relationship, the cell was held at a holding potential of 0 mV for 200 ms and then stepped to a test potential between -80 and +80 mV in 20 mV increments for 800 ms and Table 1. Primers used to amplify cDNAs of rat insulin (Ins), glucagon (Gcg-1, Gcg-2), somatostatin (Sst-1, Sst-2), GABA A receptor subunits and the reference gene b-actin (ACTB).  then returned to 0 mV for 200 ms before the next test potential was applied. Patch-clamp recordings were done using an Axopatch 200B amplifier, filtered at 2 kHz, digitized on-line at 10 kHz using an analogue-to-digital converter and analyzed with pClamp (Molecular Devices) and Sigmaplot programs. The amplitude of the tonic currents was measured as the difference in the mean holding current before and after drug application as described by Glykys and Mody [26]. The single-channel recordings were performed in the whole-cell patch-clamp configuration. The amplitude of the currents was measured either from all-point amplitude histograms or from direct measurements of the amplitude of individual currents.

Harvesting the Cytoplasm
At the end of patch-clamp recordings, the cell's cytosome was aspirated into the recording pipette by applying a negative pressure to the pipette. Harvesting was interrupted before or as soon as the seal was lost, as described by Lambolez B et al. [27]. The content in the pipette (,5 ml) was then expelled into a PCR tube and frozen immediately on the dry ice. The recording pipettes and the pipette solution were autoclaved and the recording electrode was cleaned with ethanol and RNAse away (Molecular Bioproduct, San Diego, CA, USA) before the experiment was started.

Single-cell RT-PCR
We examined two protocols for RT-PCR identification of gene product: the standard RT-PCR and multiplex RT-PCR. Both protocols worked for islet cell-type discrimination.
For the standard RT-PCR, in the first step, the reverse transcription was performed with Verso TM cDNA synthesis kit (Thermo Scientific, Bremen, Germany) in a final volume of 20 ml. The reverse transcription reaction was incubated at 42uC for 30 min and then at 95uC for 2 min. In the second step, each 3 ml cDNA was individually amplified by specific primer pairs for insulin, glucagon, and somatostatin (Ins, Gcg-2 and Sst-2 in Table 1), by a 40 cycles PCR protocol (95uC, 15 s; 59uC, 30 s; 72uC, 60 s). 4 ml of each individual PCR reaction was examined by electrophoresis on a 1.5% agarose gel stained with SYBR Gold (Invitrogen, Carlsbad, CA, USA).  The two steps of multiplex PCR were performed essentially as described by Ruano D et al. In the first step, cytosome, RT/ Platinum Taq Mix (0.4 ml, Invitrogen) and 0.67 picomole of each of the primer pairs (Ins, Gcg-1 and Sst-1 in Table 1) were added to reaction mix supplied by the manufacturer (final volume 20 ml). The set of primers were verified in jPCR program [28] to avoid dimmers or possible interactions among different primers. The reverse transcription reaction was conducted at 55uC for 30 min followed directly by 40 PCR cycles (95uC, 15 s; 59uC, 30 s; 72uC, 60 s). In the second step, 0.4 ml of the first PCR product was then used as template. Each cDNA was individually amplified using gene specific primer pair by performing 40 PCR cycles (same as described above). 4 ml of each individual PCR reaction was then electrophoresised on a 1.5% agarose gel stained with SYBR Green (Invitrogen, Carlsbad, CA).

Quantitative PCR (RT-qPCR)
Gene expression profiling of the GABA A receptor subunits in rat islets was done on total RNA from isolated rat pancreatic islets by RT-qPCR as previously described [29]. All primers are listed in Table 1.

Statistics
Data are presented as mean 6 standard error of mean (SEM). Differences in expression levels were analyzed by Student's t-test or non-parametric Mann-Whitney test. In all tests, p,0.05 was considered statistically significant. All statistical tests were

Identification of Cell-types in Intact Rat Islets by Singlecell RT-PCR
We examined the gene expression of insulin (Ins), glucagon (Gcg) or somatostatin (Sst) in islet cells using single-cell RT-PCR analysis. We did not examine the expression of the pancreatic polypepetide produced by PP cells in this study. The PP-cells constitute less than 5% of the total islet cells [30,31]. At the end of whole-cell recordings from cells in intact islets the cytosome containing the mRNAs was collected and used for analysis. Fig. 1A shows representative results from 6 different cells. The cytosome of each cell was incubated with three different primer pairs aimed at amplifying the different transcripts encoding the three hormones. The rat islet total RNA was used as positive control and water or the pipette solution as negative controls. Based on amplification of glucagon, insulin or somatostatin gene products, the cells in Fig. 1A were identified as a-cells (n = 4), one as a b-cell and another was a d-cell.
We collected cytosomes from 95 cells and detected expression of a hormone in 43 of those cells. The majority of the cells were acells expressing glucagon (21 cells, 49%, Fig. 1B). 12 cells were bcells (28%, Fig. 1B) expressing insulin and another 3 cells were d-  Fig. 1B) expressing somatostatin. The remaining 7 cells (16%) expressed somatostatin together with either insulin or glucagon (Fig. 1B). The cell surface area, which is proportional to the cell membrane capacitance, has been used to aid in determining the cell-type. Electrophysiological studies on single islet cells have shown that the b-cells have, on the average, significantly larger cell capacitance than the a-cells [25,32,33]. We, therefore, examined the relationship between the cell capacitance and the cell-type we had determined by using RT-PCR. In agreement with previous report Fig. 1C shows that the average cell capacitance was significantly larger for the b-cells (5.660.6 pF, n = 12) as compared to the a-cells (4.260.3 pF, n = 18). However, since there is a significant overlap between the capacitance of a-cells and b-cells, single-cell RT-PCR further aids in distinguishing the cell-type and may be particularly important in intact islets where cells may be electrically connected.

mRNA Expression of GABA A Receptor Subunits in Rat Islets
In order to determine which subtypes of GABA A receptors might be expressed in the rat islets, we examined expression of the 19 GABA A receptor subunit isoforms by RT-qPCR and the results are shown in Fig. 2. Seventeen subunits were detected in the islets (a1, a2, a3, a4, a6, b1-3, c1, c2, c3, d, h, p and r1-3) with the b3 subunit most prominently expressed and relatively high expression of the a1, a2, a4, a6, c1, d and r subunits. Interestingly, the a5 and the e subunits were not expressed.

Interstitial GABA Activates GABA A Receptors and Generates Tonic Currents in Intact Rat Islets
To-date the interstitial GABA concentration in rat islets has not been determined but we know that GABA is released by both vesicular and nonvesicular release processes [11,23] and in high concentrations of glucose, at least the vesicular release is stimulated [11]. We, therefore, used 20 mM glucose to evoke GABA release from the b-cells and thereby maximize the interstitial GABA concentration in the islets during our experiments. The highest concentration is expected to be around the bcells release sites from where it then diffuses to the rest of the islet. We examined if the interstitial GABA activated GABA A receptors in the islet cells and the results are shown in Fig. 3. Although no GABA was added experimentally, whole-cell currents were reduced by SR-95531, a GABA A receptor specific antagonist. SR-95531 reduced the baseline current but increased again when SR-95531 was withdrawn. The difference in the whole-cell current level in the presence and absence of SR-95531 revealed the level of the GABA-activated tonic current in the cell (Fig. 3A, B)   GABA-activated current was outwardly rectifying (Fig. 3C) demonstrating that more current passes through the GABA A receptors at depolarized as compared to hyperpolarized potentials in rat pancreatic islets. In another four cells the tonic current was significantly larger (Fig. 3D and E, 21 pA, Vp = 290 mV; Fig 3F). Whether the larger current level is due to more GABA A receptors in the plasma membrane or different subpopulations of GABA A receptors expressed in the cells is not known. However, insulin promotes insertion of GABA A receptors into the plasma membrane [34]. As insulin is released in high glucose this might have happened in some of the cells and we captured it in the four cells with the larger currents before the receptor number decreased again in the plasma membrane.
Pentobarbital modulates GABA A receptors resulting in enhanced current response to GABA [21]. We examined if pentobarbital potentiated the tonic GABA-activated currents in cells in the intact islet and the results are shown in Fig. 4A and B.
The application of 100 mM pentobarbital enhanced the tonic current by 12.4 pA (171 pS, Vp = -70 mV) and significantly increased the current baseline noise. Similar results were obtained in another 3 cells where pentobarbital increased the GABA A receptors conductance (111.8621.6 pS, n = 4). The cytosome of the cell from which the current was recorded (Fig. 4A) was harvested and the cell shown to be an a-cell (Fig. 4C). GABAactivated currents were recorded in another 7 a-cells. For the twelve identified b-cells, successful whole-cell current recordings were obtained in seven of these cells but in none of them was GABA-evoked current detected (n = 7, Fig. 4D, E, F).

Single-channel Recordings of GABA A Receptor Currents in Intact Rat Pancreatic Islets
Under our experimental conditions, the membrane resistance of the cells in the whole-cell patch-clamp configuration was often in the GV range indicating that very few channels were open in the cell membrane. In this situation, the whole-cell can be regarded as an unusually large outside-out membrane patch and ideally suited for single-channel recordings. In Fig. 5, four representative current recordings are shown where the interstitial-GABA activated singlechannel currents in cells in intact islets. The most prominent single-channel conductances were 65 pS (Fig. 5A, Vp = 290 mV), 24 pS (Fig. 5B, Vp = 270 mV), 40 pS (Fig. 5C, Vp = 270 mV) and 105 pS (Fig. 5D, Vp = 270 mV). When the islets were

Pentobarbital Enhances the GABA-activated Singlechannel Currents
An example of the effect of 100 mM pentobarbital on GABAactivated single-channels in the intact islets is shown in Fig. 7. Before the application of pentobarbital, the receptors had an average current amplitude of 1.6 pA (23 pS, Vp = 270 mV, Fig. 7A) and a low open probability as shown by the all-points histogram (15 s) to the right of the current trace. Following exposure to pentobarbital (Fig. 7B), both the single-channel current and the open probability increased as manifested in channel current of 2.7 pA (38 pS, Vp = 270 mV) and the skewed all-points histogram. Similar results were obtained in another four experiments where the pentobarbital-enhanced average conductance was 3866 pS. 100 mM SR-95531 inhibited the GABAactivated, pentobarbital-enhanced currents (Fig 7C).

Discussion
Our results demonstrate that GABA acts as a paracrine signaling molecule in the intact rat pancreatic islets. The results show that the interstitial GABA concentrations activate GABA A receptors generating tonic currents that are only present in the glucagon-releasing a-cells in rat islets. The results are consistent with the observation that antagonists of GABA A receptors inhibit glucagon secretion in both rat and human islets [17,19]. Identification of the types of cells the recordings were made from in the intact islets was done by the single-cell RT-PCR. The singlechannel conductance of GABA A receptors ranged from 24 to 105 pS. Both the whole-cell and the single-channel currents were enhanced by pentobarbital and inhibited by GABA A receptors antagonist SR-95531.
The success rate of identifying the cells we recorded from, in the intact islets, was about 45% using the single-cell RT-PCR technique. About 50% of the cells were a-cells although they represent only 20% of all the cells in a pancreatic islet. This result is related to the way the experiments were executed as we most commonly record from cells on the surface of the islets. Most of the surface cells in rat islets are non-b cells [25,30]. We, nevertheless, were also able to record from the b-cells, about 30% of the identified cells, which make up the core of the rat islet and represent 60-70% of the islet cells [25,30]. The b-cells were, on the average, significantly larger than the a-cells as determined by the cell-capacitance and the results are consistent with what has been reported previously in mouse islets [25,33]. We did, nevertheless, observe significant scatter in size for both the aand b-cells highlighting the need for an alternative method, the single-cell RT-PCR, to determine the cell-type in islets.
Although the presence of the GABA A receptors in pancreatic islets has been known since the late 1980s [6], only recently was it shown that the interstitial GABA in the islets activates the native GABA A receptors [17]. GABA A receptors can be supersensitive to GABA and respond when the GABA concentration is as low as pM to nM [35,36]. In this study, no external GABA was added but tonic, GABA-generated currents were recorded. The GABA must, therefore, have originated from cells within the islets. Furthermore, since the GABA-activated current was enhanced by pentobarbital, the interstitial GABA concentration must have been sub-saturating [17,35,36].
GABA A receptors normally contain two as, two bs plus a third type of subunit in the receptor complex [21]. In the rat islets, we detected 17 of the total 19 possible GABA A receptor subunit genes. The highest expression level was detected for the b3 subunit and a very low expression level for b2, c2, c3 and p. The a5 and e subunits were not expressed. Clearly a number of different GABA A receptor subtypes can be expressed in the rat islets and is in accord with earlier reports that have examined the expression of the subunits in rodents [19,34]. Interestingly, in human islets, both the a5 and p subunits are prominently expressed [12,17]. As the subunit composition determines both the physiological and the pharmacological properties of the receptors, it can be expected that although GABA signaling takes place both in human and rodent islets, the pharmacological profile of the response will vary somewhat between species [21,37].
The type of a subunits influences the GABA affinity of the receptors [21]. A number of different GABA A receptor a subunits are expressed in the islets potentially forming receptors that differ in their sensitivities to GABA and other drugs [21,35,36,37]. The whole-cell IV curves showed outward rectification. GABA A receptors have been shown to have either linear or outwardly rectifying IV, both at the single-channel level and at the whole-cell current level [36,38,39,40]. At the single channel level, the outward rectification is the property of the ion channel molecule whereas at the whole-cell current level it can be either due to intrinsic properties of the ion channels or that the open probability of the channels is lower at hyperpolarized potentials as compared to depolarized potentials. The tonic conductance varied significantly in magnitude, which might be related to different density of the GABA A receptors expressed in the cell plasma membrane, the concentration of interstitial GABA or the subtypes of GABA A receptors in the plasma membrane. At the single-channel level, GABA A receptors were gradually blocked by SR-95531 and potentiated by pentobarbital. Whether the single-channel conductances we determined in this study are related to different GABA A receptor subtypes or activation of a specific GABA A receptor-type by different GABA concentrations remains to be examined.
In the central nervous system, GABA-activated tonic current has an important physiological function in regulating overall excitability of neural networks [22]. Similarly, it is possible that tonic activation of GABA A receptors in pancreatic islets have a role not only in suppressing the membrane potential but even synchronizing the population of the a-cells, in addition, to coordinate their activities with that of the b-cells. GABA is known to exert inhibitory effect on glucagon release [6,17,34]. Clearly, extracellular GABA signaling in pancreatic islets is of physiological and potentially pharmacological significance for glucose homeostasis. Our results demonstrate that GABA signaling can be studied in identified cells in intact islets by combining the singlecell RT-PCR method with patch-clamp recordings.