GABA_A increases calcium in subventricular zone astrocyte-like cells through L- and T-type voltage-gated calcium channels

Neurogenesis persists in two regions of the adult brain, the SVZ along the lateral ventricle and the subgranular zone in the hippocampal dentate gyrus (Lledo et al., 2006 ; Zhao et al., 2008 ). The SVZ contains a mosaic of cell types including neuroblasts ensheathed by cells with astrocytic features such as glial fibrillary acidic protein (GFAP) expression. A subpopulation of these GFAP-expressing cells (also called astrocyte-like cells or SVZ astrocytes) generates intermediate progenitors called transit amplifying cells. The latter generate neuroblasts that differentiate into interneurons in the olfactory bulb. Previous studies have shown that the distinct steps of neurogenesis (i.e. migration, proliferation and differentiation) are influenced by local molecules. These molecules differentially affect intracellular Ca²⁺ dynamics and regulate Ca²⁺-dependent intracellular pathways (Bordey, 2006 ; Pathania et al., 2010 ). One such local molecule is the amino acid γ-aminobutyric acid (GABA) that is synthesized and released by SVZ neuroblasts as shown using immunohistochemistry for glutamic acid decarboxylase (GAD) (the enzyme that catalyzes the decarboxylation of glutamate to GABA) and GABA, and electrophysiology to show functional release from neuroblasts (Stewart et al., 2002 ; Nguyen et al., 2003 ; Bolteus and Bordey, 2004 ; De Marchis et al., 2004 ; Liu et al., 2005 ; Gascon et al., 2006 ; Platel et al., 2008 ).

GABA acts through specific receptors, GABA_A receptors, which are expressed in both SVZ neuroblasts and astrocytes (Stewart et al., 2002 ; Nguyen et al., 2003 ; Wang et al., 2003b ; Bolteus and Bordey, 2004 ; Liu et al., 2005 ). In developing systems, the GABA_A receptor is thought to regulate the behavior of immature cells through depolarization leading to the canonical activation of VGCCs and intracellular Ca²⁺ increases (Owens and Kriegstein, 2002 ). Such a mechanism has been reported in SVZ neuroblasts and involves nifedipine-sensitive L-type VGCC (Nguyen et al., 2003 ; Wang et al., 2003b ; Gascon et al., 2006 ). It has been speculated that this canonical Ca²⁺ increase may not operate in SVZ astrocytes due to their biophysical properties (low input resistance and hyperpolarized resting potential) (Liu et al., 2006 ; Bordey, 2007 ). We thus set out to investigate whether and, if so, how GABA_A increases Ca²⁺ in SVZ astrocytes.

One limitation to address this issue has been the inability to distinguish Ca²⁺ indicator-loaded astrocyte-like cells from neuroblasts in acute SVZ slices. To study Ca²⁺ activity selectively in SVZ astrocytes, we used two lines of transgenic mice where astrocyte-like cells express intracellular DsRed or membrane-associated GFP. Using these mice revealed that traditional bath loading of Ca²⁺ indicators preferentially loaded neuroblasts at the slice surface while astrocyte-like cells resided deeper inside the tissue. Using pressure loading of a Ca²⁺ indicator dye inside the tissue, we preferentially loaded astrocyte-like cells. We found that a GABA_A receptor agonist increased Ca²⁺ in a subset of astrocyte-like cells (∼50%) through L- and T-type VGCCs. In addition, ambient GABA tonically regulated the frequency of Ca²⁺ activity in ∼80% of SVZ astrocytes. GABA increased or decreased the frequency, subdividing the SVZ astrocytes into two subpopulations. For the first time this finding illustrates a functional difference among astrocyte-like cells of the SVZ. Such a GABA_A-regulation in selective astrocyte-like cells may impact Ca²⁺-dependent mechanisms, including proliferation and the release of diffusible molecules (e.g. ATP; Striedinger et al., 2007 ).

Experiments were performed in several lines of transgenic mice: (1) Mice expressing DsRed under the human GFAP promoter (hGFAP-DsRed mice) were produced by the co-authors, N.A. Jensen and J.V. Nielsen (Noraberg et al., 2007 ). To generate the hGFAP-DsRed transgene, the pDsRed2-1 plasmid (Clontech) was initially modified to introduce a PacI site downstream of the SV40 polyadenylation signal, by filling in an AflII site. Subsequently, the hGFAP promoter (Brenner et al., 1994 ) was cloned into the BglII and SalI sites, before a rabbit β-globin intron was inserted into the BamHI sites between the hGFAP promoter and the DsRed coding sequence. For microinjection, the transgene was excised from the plasmid backbone by digestion with BglII and PacI, gel-purified and microinjected, at a concentration of 6 ng/μl, into the pronucleus of fertilized B6D2F1 mouse eggs as previously described (Nielsen et al., 2007 ). Transgenic hGFAP-DsRed mice were identified by PCR with primers: 5′-TCTGGGCACAGTGACCTCAGTG and 5′-GGGACATCTTCCCATTCTAAAC. (2) hGFAP-DsRed mice were crossed with homozygote mice carrying GFP under the doublecortin promoter (DCX-GFP mice, FVB/N strain, a gift from Dr. R. Miller, University of Chicago, originally from Gensat) to generate hGFAP-DsRed/DCX-GFP mice. These mice were used in 25% of the experiments instead of hGFAP-DsRed mice. (3) hGFAP-tTA/TetO-MrgA1:GFP mice (called hGFAP-MrgA1:GFP mice) were a gift from Dr. Ken McCarthy (University of North Carolina at Chapel Hill). In the absence of doxycycline, astrocytes express MrgA1 receptors fused to GFP (Fiacco et al., 2007 ). All experimental protocols were approved by the Institutional Animal Care and Use Committees of Yale School of Medicine and by the Danish National Animal Care and Use Committee. Mice were used between postnatal day (P) 20 and P42.

Slice preparation, immunostaining, and image acquisition and analysis were as previously described (Platel et al., 2009 ). Primary antibodies included: anti-DCX (goat or rabbit, 1:100, Santa Cruz, SC8066 and SC28939), anti-GFAP (rabbit, 1:1000, Dako, Z0334), anti-EGFR (sheep, 1:100, Millipore, 06-129), and anti-GFP (chicken, 1:500, Abcam). Each staining was replicated at least in four to five slices from three different mice. The appropriate secondary antibodies were Alexa fluor series (1:1000, Invitrogen, USA) or Cyanine series (1:500, Jackson Labs). Z-section images (spaced by 1–2 μm over 10–20 μm) were acquired on a laser-scanning confocal microscope (Olympus FluoView 1000) with a 20× dry objective (N.A. 0.75) or a 60× oil objective (N.A. 1.42). Images were analyzed using Imaris 4.0 (Bitplane AG, Switzerland) and reconstructed in ImageJ 1.39t (Freeware, Wayne Rasband, NIH, USA) and Photoshop CS3 (Adobe, USA).

Acute coronal or sagittal brain slices (250–300 μm-thick) containing the SVZ were prepared as we previously described (Bolteus and Bordey, 2004 ; Bolteus et al., 2005 ). Slices were placed in a flow-through chamber and continuously superfused with oxygenated artificial cerebrospinal fluid (aCSF) containing (in mM): NaCl 125; KCl 2.5; CaCl₂ 1.8; MgCl₂ 1; NaHCO₃ 25; glucose 10. Experiments were performed on an upright Olympus BX61WI microscope equipped with an Olympus FluoView 1000 confocal microscope and a water-immersion Nomarski phase-contrast and fluorescence 60× objective (N.A. 0.9).

Whole-cell patch clamp recordings were obtained as previously described (Wang et al., 2003a ,b ; Bolteus and Bordey, 2004 ; Liu et al., 2006 ). Pipettes had resistances of 6–8 MΩ when filled with an intracellular solution containing the following: 110 mM KCl, 1.0 mM CaCl₂, 10 mM EGTA, 10 mM HEPES, 50 μM Alexa Fluor 488 dye and an ATP-regenerating solution that included 4 mM K₂ATP, 20 mM K₂-phosphocreatine, 50 U/ml creatine phosphokinase, and 6 mM MgCl₂. The pH and the osmolarity were adjusted to 7.2 and 290 mOsm, respectively. The liquid junction potential (∼4 mV) was not corrected. Whole-cell recordings were performed using an Axopatch 200B amplifier, and current signals were low-pass filtered at 2–5 kHz and digitized on-line at 5–20 kHz using a Digidata 1320 digitizing board (Axon Instruments, Foster City, CA, USA). Recorded cells were held at −60 mV. Voltage steps were applied from −100 to +100 mV by 20 mV increment. Capacitive and leak currents were not subtracted.

SVZ cells were loaded by pressure application of Fluo-4 AM (100 μM in aCSF, 0.4% Pluronic acid F-127, Invitrogen). We observed no differences in data obtained with either hGFAP-DsRed or hGFAP-MrgA1:GFP mice regarding the pharmacology of muscimol responses; we therefore pooled the data. Images were acquired every 2 s with FluoView acquisition software. In acute slices from hGFAP-MrgA1:GFP mice, Fluo-4 AM-astrocyte-like cells were distinguished from other SVZ cells due to their enhanced green fluorescence on their cell membrane. In addition, application of a selective peptide agonist of MrgA1 receptors (that has no endogenous receptors) was routinely used to induce Ca²⁺ responses selectively in SVZ astrocytes to identify them at the end of the experiments. ROIs were placed on SVZ cells that responded to the peptide during a 10–20 s application period.

For spontaneous movies, images were acquired every 2 s (0.5 Hz) for 10 min in each condition with 5 min of wash-in or wash-out between each movie. F₀ (i.e. baseline) and F are the mean fluorescence intensities measured throughout all of the regions of interest (ROIs) and in each ROI, respectively. A change in fluorescence was considered to be a Ca²⁺ increase if it was >15% F/F₀ increase. Intracellular Ca²⁺ changes were calculated using Calsignal (Platel et al., 2007 ) and Clampfit 10. F/F₀ was detected with Calsignal and traces were exported into Clampfit for peak analysis using the threshold detection function. For peak analysis, the baseline for each ROI trace was manually adjusted to zero. In addition, traces from control and drug-treated movies were concatenated and the same threshold for peak detection was used. ROIs were designated as “responding” if the cell was responsive at least 50% of muscimol applications in order to perform subsequent pharmacology. “Non-responsive” cells were cells that displayed no increase in F/F₀.

Data are expressed as mean ± standard error. Statistical analysis used a two-tailed t-test except where noted.

Ca²⁺ imaging experiments were performed in the presence of the GABA_B blocker CGP 52432 (1 μM) and blockers of glutamate receptors including D-APV (50 μM) for NMDA receptors, NBQX (20 μM) for AMPA/kainate receptors, MPEP (10 μM) and JNJ 16259685 (100 nM) for group I metabotropic glutamate receptors. We also used for selective experiments: nickel (100 μM) to block N- and R-type VGCCs; nifedipine (10 μM, Sigma-Aldrich) and mibefradil dihydrochloride (10 μM) for blocking L-type and T-type VGCCs, respectively; BayK 8644 (10 μM) to enhance L-type VGCC activity; 2-aminoethoxydiphenyl borate 2-APB (100 μM) for blockade (although non-selective) of IP₃-related signaling. Drugs were from Tocris Biosciences (MO, USA), except where noted.

We first verified that DsRed is expressed in astrocyte-like cells of the SVZ. We immunostained for GFAP (blue) in sections from hGFAP-DsRed/DCX-GFP mice (Figures 1 A,B). We observed bright DsRed-positive cells, which were GFAP-positive (arrows in Figure 1 B), and some faint DsRed cells, which were GFAP-negative and DCX-GFP-positive (arrowhead in Figure 1 B). This result was confirmed by staining for both GFAP and DCX in sections from hGFAP-DsRed mice (Figure S1 in Supplementary Material). DsRed-positive cells occasionally stained positive for the transit amplifying cell marker epidermal growth factor receptor (EGFR, Figure S1B in Supplementary Material, arrowhead). The known half-life of DsRed (∼4 days) allows it to persist at a lower level in daughter cells. These findings are thus in agreement with GFAP-positive cells (i.e. SVZ astrocytes) being neural progenitors that generate EGFR-cells and neuroblasts (Doetsch et al., 2002 ; Platel et al., 2009 ).

We next performed patch clamp recordings of DsRed-fluorescent cells in acute slices where cells with different fluorescence intensity could be observed (Figures 1 C,D). Every bright DsRed-fluorescent cell recorded had a hyperpolarized resting potential (mean of −80.8 ± 1.0 mV), low input resistance (57.4 ± 9.1 MΩ, n = 7), and a linear current–voltage relationship following depolarizing steps (from −160 to +100 mV, Figures 1 E,F) from a holding potential of −80 mV. Thus, bright DsRed-fluorescent have characteristics of astrocytes of the SVZ (Liu et al., 2006 ). By contrast, every recorded, faint DsRed-fluorescent cell displayed characteristics of neuroblasts (Wang et al., 2003a ; Bolteus and Bordey, 2004 ). These characteristics include high input resistance (mean of 4.4 ± 1.3 GΩ), depolarized zero-current resting potentials (−42.7 ± 8.7 mV, n = 5), and the presence of voltage-dependent outward currents (Figures 1 E,F). These data suggest that bright and faint DsRed-fluorescent cells can be unambiguously identified as astrocyte-like cells and neuroblasts, respectively, in acute slices.

One important limitation of the hGFAP-DsRed mice for performing Ca²⁺ imaging is the fact that not every astrocyte-like cell is DsRed-fluorescent (Figures 1 A,B). This is more apparent in sections from hGFAP-DsRed mice crossed with hGFAP-GFP mice. While some DsRed-fluorescent cells are GFP-positive (arrows), not all GFP-positive cells are DsRed-fluorescent (arrowheads, Figures 2 A–C). We thus acquired transgenic mice that express an exogenous Gq-protein coupled receptor (called Mas-related gene A1, MrgA1) in GFAP-expressing cells (i.e. astrocytes). MrgA1 has no endogenous ligand in the brain (Fiacco et al., 2007 ). The MrgA1 receptor fused to GFP was targeted to astrocytes using the inducible tet-off system. Mice expressing the tetracycline transactivator (tTA) under the human GFAP promoter were crossed to mice in which the MrgA1:GFP receptor was transcribed using the tet-off (tetO) minimal promoter. In the SVZ of hGFAP-tTA/tetO-MrgA1:GFP mice (referred henceforth as hGFAP-MrgA1:GFP mice), GFP displays a membrane expression selectively in all astrocyte-like cells (GFAP+ cells, red) but not in neuroblasts (DCX+ cells, blue, Figures 2 D,E). In addition, SVZ astrocytes loaded with the Ca²⁺ indicator dye Fluo-4 AM can be further identified with a peptide agonist for MrgA1 receptors that does not bind endogenous receptors in the brain. Pressure application of phe-leu-arg-phe amide peptide (FLRFa, 50 μM, 10–20 s) induced Ca²⁺ increases in GFP-decorated cells, i.e. astrocyte-like cells (Figures 2 F,G).

We previously reported that astrocyte-like cells express functional GABA_A receptors using patch clamp recordings (Liu et al., 2005 ). However, it remained unknown whether these receptors led to Ca²⁺ increases in these cells. We pressure loaded slices with Fluo-4 AM in slices from hGFAP-DsRed and hGFAP-MrgA1:GFP mice (Figures 3 A,E, respectively). Application of the GABA_A receptor agonist muscimol (50 μM, 5 s) increased Ca²⁺ in 38.1 ± 4.8% and 51.3 ± 3.5% of SVZ astrocytes from hGFAP-DsRed and hGFAP-MrgA1:GFP mice (106 cells, n = 17 slices and 396 cells, 16 slices, respectively, Figures 3 B,D,F). Experiments were performed in the presence of glutamate and GABA_B receptor blockers (see Materials and Methods). GABA_A-induced Ca²⁺ responses were abolished by either bicuculline or picrotoxin (50 μM), two GABA_A receptor antagonists, in 94 ± 4.67% of the muscimol-responding cells (27 cells, n = 5 slices total, Figures 3 C,G and 4 E,F).

GABA_A receptor activation is known to depolarize parenchymal astrocytes (Barakat and Bordey, 2002 ; Bekar and Walz, 2002 ). Following GABA_A receptor-induced depolarization, intracellular Ca²⁺ increases in immature cells are also thought to result from the canonical activation of VGCCs (Owens and Kriegstein, 2002 ; Meier et al., 2008 ). We thus tested the effects of three VGCC blockers on the percentage of SVZ astrocytes displaying muscimol-induced Ca²⁺ increases in slices from both hGFAP-DsRed and hGFAP-MrgA1:GFP mice. VGCCs are grouped into three families: (1) the high-voltage activated dihydropyridine (DHP)-sensitive channels (L-type, Ca_V1.x), (2) the high-voltage activated DHP-insensitive channels (P-, N- and R-type, Ca_V2.x), and (3) the low-voltage activated channels (T-type, Ca_V3.x). We tested the following three blockers: Nickel (Ni²⁺, 100 μM) for R-type Ca_V2.3 and T-type Ca_V3.2 channels (high sensitivity), and for other T-type channels (low-sensitivity); the DHP antagonist nifedipine (10 μM) for L-type channels, and mibefradil (10 μM) for T-type channels. Ni²⁺ significantly decreased the percentage of muscimol-responding astrocytes but by only 30 ± 9.4% (n = 50 cells, six slices, Figures 4 E,F). Nifedipine and mibefradil significantly decreased the percentage of muscimol-responding astrocytes by 61 ± 10% Figures 4 A,B and 48 ± 4%, respectively (n = 217 cells, six slices, and 78 cells, four slices, respectively, Figures 4 E,F). When used together, these three VGCC blockers abolished muscimol responses in 90 ± 4% of the responding SVZ astrocytes (Figures 4 E,F). Of the cells that continue to respond, none of these drugs had any effect on the amplitude or area of the Ca²⁺ responses (Figure 4 G). These data suggest that GABA_A-induced Ca²⁺ responses in astrocyte-like cells of the SVZ are predominately mediated by Ca²⁺ influx through L- and T-type VGCCs.

As illustrated in Figures 3 D and 4 E, 40–50% of SVZ astrocytes display Ca²⁺ increases in response to GABA_A receptor activation. Considering that nearly all SVZ astrocytes tested have been reported to display functional GABA_A currents (Liu et al., 2005 ), the lack of Ca²⁺ responses could be due to either the absence of functional VGCCs or the fact that GABA_A depolarization does not reach VGCC threshold. We thus tested the L-type VGCC activator BayK 8644. In the presence of BayK 8644, there was a significant increase (to 190% of control) in the % of cells responding to muscimol (from ∼35% to 75%, Figures 4 C,D,E,F). These data suggest that most astrocyte-like cells express DHP-sensitive VGCCs. BayK 8644 also increased the area, but not the amplitude, of muscimol-induced Ca²⁺ increases (Figure 4 G), suggesting that BayK 3644 prolonged Ca²⁺ responses due to GABA_A receptor activation.

Considering that increases in intracellular Ca²⁺ can further trigger Ca²⁺ increases from intracellular stores, we tested 2-APB, a blocker of inositol-1,4,5-trisphosphate receptors (IP3) and IP3-sensitive intracellular stores, among other channels at 100 μM (Bootman et al., 2002 ; Peppiatt et al., 2003 ). Bath application of 2-APB resulted in a 30 ± 9% decrease in the number of muscimol-responding astrocytes, but this decrease was not significant (Figures 4 E,F, t-test, p > 0.05). Nevertheless, when analyzing cells that continue to respond after drug application, 2-APB significantly reduced the amplitude and area of muscimol-induced Ca²⁺ transients, while nifedipine and other VGCC blockers did not (Figure 4 G). These data suggest that Ca²⁺ release from intracellular stores contributes to the increase in intracellular Ca²⁺ concentration following GABA_A depolarization-induced Ca²⁺ influx through VGCCs.

We previously reported that 60% of SVZ astrocytes recorded with the patch clamp technique displayed a tonic current due to GABA_A receptor activation (Liu et al., 2005 ). We thus examined whether ambient GABA exerted a tonic regulation of Ca²⁺ dynamics in SVZ astrocytes. Recordings were performed in the presence of blockers of GABA_B and glutamate receptors, as described in the Methods. We found that 77.8% of astrocyte-like cells displayed baseline Ca²⁺ activity in the form of Ca²⁺ transients at an average frequency of 0.252 Hz (10 min of recordings, n = 196 cells, four slices, see traces in control conditions in Figure 5 D). To test the effect of drugs on Ca²⁺ activity, a 10-min recording was following by a 5-min drug wash-in and an additional 10-min recording in the presence of the drug. With no drug application (control movies), spontaneous Ca²⁺ activity was stable over time. Following analysis of Ca²⁺ frequency in the first and second 10-min periods, we found that 26.3 ± 4.8% and 19.0 ± 2.5% of the cells displayed a non-significant increase and decrease in the frequency of Ca²⁺ transient to 117.0 ± 6.3% and 81.7 ± 6.8% of control, respectively (n = 47 cells, three slices). Spontaneous Ca²⁺ activity was eliminated following 15–20 min of 2-APB application (data not shown), as expected, since regenerative Ca²⁺ transients have been shown to involve Ca²⁺ from internal stores that rely on the activation of IP3 receptors in various cell types (D’Andrea et al., 1993 ; Ciapa et al., 1994 ; Liu et al., 2001 ; Bellamy, 2006 ). When recorded in the presence of bicuculline (Bic) under baseline conditions, wash-out of bicuculline had two distinct effects on SVZ astrocytes. In 79 ± 3% of the astrocyte-like cells bicuculline wash-out resulted in a significant increase in the frequency of Ca²⁺ transients to 216 ± 39% of control (from 0.154 to 0.36 Hz, n = 90 cells, Figures 5 A,B, p < 0.01). In addition, in 18 ± 6% of the SVZ astrocytes bicuculline wash-out resulted in a significant decrease in the frequency of Ca²⁺ transients to 72.9 ± 4.6% of control (from 0.45 to 0.34 Hz, n = 15/90 cells, Figure 5 B). Wash-out of SR-95531 (gabazine, 100 nM), a non-competitive GABA_A receptor antagonist, had similar effects (data not shown). Gabazine wash-out increased and decreased the frequency of Ca²⁺ transients in 41% and 27% of the SVZ astrocytes, respectively (n = 3 slices, data not shown). Bath application of bicuculline had the opposite effects to that of bicuculline wash-out, as expected. Bicuculline wash-in significantly decreased the frequency of Ca²⁺ transients to 60 ± 6% of control in 46 ± 8% of astrocyte-like cells (from 0.252 to 0.175 Hz, n = 185 cells) and increased the frequency to 223 ± 32% of control in 31 ± 3% of the cells in the same slices (from 0.10 to 0.25 Hz, n = 185 cells, four slices, Figures 5 C,D, p < 0.05).

In this study we first report the use of two transgenic mouse lines to perform Ca²⁺ imaging in astrocyte-like cells of the SVZ. Without these mice, identifying Fluo-4 AM-loaded astrocytes among other SVZ cells was not feasible. hGFAP-DsRed mice are the most practical mouse line for performing Ca²⁺ imaging with the green fluorescent dye Fluo-4 AM since DsRed fluorescence is readily distinguishable from Fluo-4 fluorescence. However, in these mice not every SVZ astrocyte fluoresces red, and it is unclear whether a sub-population is targeted. In addition, the long half-life of DsRed could be an issue for positive identification. Newly born neuroblasts are DsRed fluorescent but are dimmer than astrocytes. The hGFAP-MrgA1:GFP mouse line appears to be the best model to study Ca²⁺ activity in SVZ astrocytes. GFP being fused to the MrgA1 receptor highlights the cell membrane and appears brighter than cytoplasmic Fluo-4 AM-loading. In addition, and perhaps more importantly, SVZ astrocytes can be selectively identified following increases in intracellular Ca²⁺ using the FLRFa peptide. In our system bath loading of the Ca²⁺ indicator Fluo-4 AM resulted in surface labeling of neuroblasts (data not shown). In the SVZ, cells are densely packed, thus limiting dye diffusion inside the tissue. In addition, neuroblasts tend to “bulge out” of the slice surface while astrocytes remain anchored inside the tissue as previously reported (Wang et al., 2003a ). Instead, we used pressure loading inside the tissue resulting in preferential loading of SVZ astrocytes. With these experiments, we recommend using hGFAP-MrgA1:GFP mice to study Ca²⁺ activity in astrocyte-like cells of the SVZ in future studies.

Next, our pharmacological analyses suggest that the GABA_A-mediated Ca²⁺ increases arise from Ca²⁺ influx through L- and T-type VGCCs that are then amplified by Ca²⁺-induced Ca²⁺ release from internal stores possibly involving IP3 receptors. This pathway is well-described in other neuronal and non-neuronal cell populations, and implies that GABA_A receptor activation results in membrane depolarization that, in turn, activates VGCCs. Indeed, astrocytes are known to be depolarized by GABA_A receptor activation resulting in GABA_A-induced Ca²⁺ increases (Bekar and Walz, 2002 ; Meier et al., 2008 ). GABA_A depolarization is likely due to high intracellular Cl⁻ in SVZ astrocytes, although expression of the Cl⁻ importing transporter Na⁺/K⁺/2Cl⁻ (NKCC1) and importantly the absence of the exporting Cl⁻ transporter K⁺/2Cl⁻ (KCC2) have not been examined in SVZ cells.

Considering the low-activation threshold of T-type VGCCs (as low as −70 mV) (Nilius et al., 2006 ), it is easily conceivable that these VGCCs can be opened by GABA_A-induced depolarization based on the published biophysical properties of SVZ astrocytes. Indeed, GABA_A currents (100 μM GABA) range from −20 to −400 pA in gap-junction coupled SVZ astrocytes held at −84 mV (Liu et al., 2005 ). The current has a mean of −60 pA in SVZ astrocytes recorded in the presence of a gap junction channel blocker (unpublished data). These cells were recorded with an intracellular solution containing near-physiological chloride concentration. SVZ astrocytes’ mean input resistance is 250 MΩ (range from 50 to 500 MΩ) and their mean resting potential is −85 mV ranging from −73 to −95 mV (Liu et al., 2006 ). Thus, a 60-pA GABA_A-current in a 250-MΩ astrocyte would result in a 15-mV depolarization, which is sufficient to reach the threshold for T-type VGCC activation.

L-type VGCCs are high-voltage activated ranging between −50 and −40 mV (Lacinova, 2005 ). Around ∼40–50% of total astrocytes are muscimol-responsive in control conditions (from Figure 3 E). Of those, ∼40% are nifedipine-sensitive (see Figure 4 F). Therefore, the population of astrocytes that reach the threshold for activation by L-type VGCCs comes to ∼16–20% of all SVZ astrocytes. Similarly, SVZ astrocytes with higher input resistances (above 250 MΩ) correspond to ∼40% of the total SVZ astrocytes (see Figure 6 in Liu et al., 2006 ). Of those with higher input resistance, 20% also have resting potentials more depolarized than −80 mV, both of which would be required to reach L-type VGCC activation threshold. In this population, a current of 120 pA or less will be sufficient to reach the activation threshold of L-type VGCCs.

In the non-responding SVZ astrocytes, it is possible that either these cells do not express functional VGCCs or that GABA_A-depolarization did not reach the activation threshold for VGCCs. The fact that in the presence of BayK 8644, three-quarters of the SVZ astrocytes responded to muscimol suggests that the majority of, and perhaps all, astrocyte-like cells expresses DHP-sensitive VGCCs. These data suggest that GABA_A-depolarization in non-responding astrocyte-like cells does not reach the threshold to open VGCCs.

A question associated with the above data is whether there is enough GABA in the ambient milieu to activate GABA_A receptors and modulate Ca²⁺ activity. GABA is synthesized and released by neuroblasts (Stewart et al., 2002 ; Nguyen et al., 2003 ; Bolteus and Bordey, 2004 ; De Marchis et al., 2004 ; Liu et al., 2005 ; Gascon et al., 2006 ) that may provide a tonic versus a phasic release of GABA (for review and discussion see Bordey, 2007 ). There are no known phasic or synaptic sources of GABA in the SVZ. It was suggested that ambient GABA levels may be in the μM range in the SVZ (Bolteus and Bordey, 2004 ; Bolteus et al., 2005 ). GABA levels are expected to fluctuate over time because neuroblasts migrate, thus changing the microenvironment of astrocyte-like cells. GABA_A receptors in astrocyte-like cells have a reported EC₅₀ for GABA of 15 μM and are thus expected to be tonically activated by low μM ambient GABA (Liu et al., 2005 ). Indeed, based on published data, GABA_A receptors are tonically activated in 60% of the astrocyte-like cells (Liu et al., 2005 ). It was shown that bicuculline or picrotoxin applications induced a small shift in the current baseline that would correspond to a small depolarization of astrocyte-like cells.

Consistent with the predictions discussed above, we report that astrocyte-like cells of the SVZ display spontaneous Ca²⁺ transients, the activity of which is tonically regulated by ambient GABA acting on GABA_A receptors. In particular, GABA_A receptor inhibition revealed that tonic GABA_A activation had two effects on SVZ astrocyte-like cells, dividing SVZ astrocytes into two functionally distinct populations. In one subpopulation of these cells, tonic GABA_A receptor activation increased the frequency of Ca²⁺ transients while it decreased it in the other subpopulation. This latter GABA_A action may result from a shunting effect of GABA_A conductance in astrocyte-like cells. In future work, it will be important to understand how Ca²⁺ transients are generated in SVZ astrocytes.

In conclusion, tonic regulation of Ca²⁺ activity by ambient GABA could have major implications for the behavior of the Ca²⁺-dependent release of diffusible molecules from astrocytes such as ATP (Striedinger et al., 2007 ). Considering that GABA_A receptor activation has been shown to limit SVZ astrocyte proliferation (Liu et al., 2005 ), it remains to be determined whether GABA_A’s effect on proliferation involves L- and/or T-type VGCCs.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

This work was supported by grants from the National Institute of Health (NS048256 and DC007681, A.B.), Yale Brown-Coxe fellowship (J.-C.P.), and NRSA 1F31NS063758-01A1 (S.Z.Y.). We thank Tiffany Lin for assistance and discussion on the experiments and analysis.

Thank you to M. Morgan Taylor for comments and editing.

The Supplementary Material for this article can be found online at http://www.frontiersin.org/cellularneuroscience/paper/10.3389/fncel.2010.00008/