- Email: [email protected]
Confocal imaging of intracellular chloride in living brain slices: measurement of GABAA receptor activity
Journal of Neuroscience Methods 75 (1997) 127 – 135
Confocal imaging of intracellular chloride in living brain slices: measurement of GABAA receptor activity Jon R. Inglefield *, Rochelle D. Schwartz-Bloom Department of Pharmacology, Box 3813, Duke Uni6ersity Medical Center, Durham, NC 27710, USA Received 12 December 1996; accepted 2 January 1997
Abstract We have developed a method using UV laser-scanning confocal microscopy and the fluorescent chloride ion indicator, 6-methoxy-N-ethylquinolinium chloride (MEQ), to image GABA-mediated changes in intracellular chloride (Cl− i ) in individual neurons of the rat acute brain slice. After bath-loading slices with the cell-permeant form (reduced) of MEQ, there was intense fluorescence within neurons of diverse morphologies in the hippocampus, neocortex and cerebellum. MEQ fluorescence localized to the cytosolic compartment of both the somata and proximal dendrites. MEQ fluorescence was calibrated using the ionophores nigericin and tributyltin in the presence of varying extracellular Cl − concentrations. Neuronal MEQ fluorescence was inversely related to intracellular Cl − , with a Stern-Volmer constant of 16 M − 1 (50% quench by 61 mM Cl − ). Application of GABA in the perfusate produced a concentration-dependent decrease in MEQ fluorescence (EC50 = 40 mM) that was blocked in the presence of the Cl − channel antagonist, picrotoxin. Bath perfusion of hippocampal slices with modulators of the GABAA receptor, pentobarbital and diazepam, potentiated the GABA-mediated response by 85 and 44%, respectively. A regional comparison identified larger GABA responses for both cerebellar Purkinje and granule cells relative to pyramidal neurons of the hippocampus and neocortex and to hippocampal interneurons. Pressure ejection of the GABAA agonist, muscimol (40 mM), from a micropipet onto individual hippocampal neurons allowed the measurement of rapid responses (1 – 5 s), compared to those obtained with bath application. Thus, optical imaging of [Cl − ]i using MEQ and UV-laser-scanning confocal microscopy provides investigators with a new method to study GABAA pharmacology in neighboring neurons and perhaps even in the soma versus dendrites, simultaneously, within living brain slices. © 1997 Elsevier Science Ireland Ltd. Keywords: UV laser-scanning confocal microscopy; Hippocampus; Neocortex; Fluorescence; Cerebellum; Chloride sensitive dye; Inhibition
1. Introduction The responses of individual cells to various stimuli have been examined using fluorescent tracers for bioactive molecules (Tsien, 1988; Biwersi and Verkman, 1991; Hayashi and Miyata, 1994). For example, in cultured neurons and in brain slices, fluorescent indicators have been used to measure increases in intracellular calcium ([Ca2 + ]i), both in the somata (Yuste and Katz, 1991; Dailey and Smith, 1994; Porter and McCarthy, * Corresponding author. Tel.: +1 919 6845181; fax: + 1 919 6818609; e-mail: [email protected]
1995) and in dendrites (Regehr et al., 1989; Guthrie et al., 1991; Llano et al., 1991; Frenguelli et al., 1993; Jaffe et al., 1994) during or after neuronal excitation. In contrast, changes in intracellular chloride ([Cl − ]i) following GABAA receptor activation have been studied traditionally with electrophysiological measures (Krnjevic and Schwartz, 1967; Alger and Nicoll, 1979) and 36 Cl − flux techniques (Schwartz et al., 1985; Harris and Allan, 1985). With the advent of Cl − -sensitive fluorescent dyes such as 6-methoxy-N-[3-sulfopropyl] quinolinium (SPQ) and 6-methoxy-N-ethyl-quinolinium chloride (MEQ), changes in [Cl − ]i have been measured in non-neuronal cells (Biwersi and Verkman, 1991;
0165-0270/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 1 6 5 - 0 2 7 0 ( 9 7 ) 0 0 0 5 4 - X
128
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
Verkman, 1990 for review) and engineered cell lines (Wong et al., 1992) with fluorescence photometry. In addition, Cl − -sensitive fluorescent indicators and fluorescence photometry have been applied to the study of GABA-induced Cl − transport within both synaptoneurosomes (Engblom and Akerman, 1991) and cultured neurons (Engblom et al., 1991; Hara et al., 1992). Recently, we developed an optical imaging technique to measure GABA-mediated changes in [Cl − ]i within living brain slices using conventional epifluorescence microscopy and MEQ (Schwartz and Yu, 1995). Optical imaging in living cell preparations has become a widely used technique with the aid of a confocal microscope. Imaging with laser-scanning confocal microscopy greatly improves the resolution (both temporal and spatial) of dynamic changes in intracellular processes. It is especially useful in brain slice preparations, where cells can be imaged deep in the slice, away from cut surfaces and potentially injured cells. Thus, we took advantage of better spatial and temporal resolution using ultraviolet (UV)-laser-scanning confocal microscopy to image optically the responses to GABAA receptor activation in various types of individual neurons, simultaneously, within their native environment. The purposes of this study were to: (1) characterize the measurement of GABAA receptor responses using confocal microscopy and the Cl − -sensitive dye, MEQ; (2) calibrate the change in MEQ fluorescence intracellularly in a brain slice preparation; and (3) demonstrate rapid changes in [Cl − ]i following GABAA activation. These results have been presented in preliminary form (Inglefield and Schwartz-Bloom, 1996).
2. Materials and methods
2.1. Synthesis of dihydro-MEQ The reduction of MEQ (Molecular Probes, Eugene, OR) to a cell- permeable form, 6-methoxy-N-ethyl-1,2dihydroquinoline (dihydro-MEQ) using sodium borohydride was described previously (Biwersi and Verkman, 1991), with modifications by our laboratory (Schwartz and Yu, 1995). Prior to each experiment, the dried dihydro-MEQ extract (under N2 at -20°C) was re-suspended in 15 ml of ethyl acetate and used immediately. For optimal use with confocal imaging, the dried dihydro-MEQ extract should be used within two days.
2.2. Slice preparation and bath-loading of dihydro-MEQ Transverse hippocampal (containing somatosensory cortex) and sagittal cerebellar slices were prepared from 12 – 21 day old Sprague-Dawley rats (Charles River). Slices (300 mm for hippocampus; 200 mm for cerebel-
lum) were cut with a vibratome in oxygenated (95% O2/5% CO2 mixture) Ringer’s physiological buffer (0– 4°C), containing (in mM) 119 NaCl, 2.5 KCl, 1.0 NaH2PO4, 1.3 MgSO4, 2.5 CaCl2, 26 NaHCO3, and 11 glucose, pH 7.4. Slices were warmed to 28°C in oxygenated Ringer’s buffer and then incubated with the re-suspended dihydro-MEQ (final concentration of 300 mM) at 28°C for 0.5 h in Ringer’s buffer. Once loaded into cells, dihydro-MEQ is oxidized to MEQ, which remains trapped in the cytosolic compartment (Biwersi and Verkman, 1991). Following incubation, slices were washed once in fresh oxygenated Ringer’s buffer to remove extracellular dihydro-MEQ, and then kept in fresh buffer for 0.5–3 h before transfer to the imaging chamber.
2.3. UV laser-scanning confocal microscopy For imaging, slices were submerged in a chamber that was constantly perfused with oxygenated Ringer’s buffer (1.5–2.0 ml/min) at room temperature. The imaging chamber was positioned on the stage of an upright Nikon Optiphot microscope. The laser scanning confocal microscope (Noran Odyssey™, Noran Instruments) was equipped with an argon UV laser (80 mW output, Enterprise 653, Coherent), and a digital imaging system. A 486 personal computer along with Image1™ software (Universal Imaging) was used for image/data acquisition and Odyssey™ software was used to control the confocal microscope. MEQ fluorescence was excited with the UV laser (20% intensity) and dichroic mirror to exclude all but the 364 nm line. Fluorescent light was transmitted through a UV water immersible objective lens (40 × , NA 0.7, Olympus). Emission (Emmax = 440 nm) was imaged using a 400 nm barrier filter and photomultipliers received the signal through a confocal slit aperture of 25 mm. (Any remaining dihydro-MEQ within the slice has an emission spectrum different from MEQ and does not interfere with MEQ fluorescence detection). The video frame rates (32 images per s) of the Noran Odyssey confocal microscope permitted rapid, full image (512×480 pixels) acquisition.
2.4. Intracellular calibration of MEQ fluorescence The fluorescence of MEQ is quenched collisionally by Cl − according to the Stern-Volmer equation, Fo/FCl = 1+ Kq[Cl], where F0 is the total quenchable signal, FCl is the fluorescence in the presence of a given Cl − concentration, and Kq is the Stern-Volmer quenching constant (in M − 1) (Verkman, 1990). Intracellular MEQ fluorescence was calibrated to give estimates for intracellular Cl − according to Krapf et al. (1988), with modifications for the brain slice. MEQ-loaded slices were equilibrated for 30 min in an oxygenated, low Cl −
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
(2.5 mM) buffer (Na-gluconate replaced NaCl, and K-gluconate replaced KCl in an equimolar manner) (pH 7.4). The maximal fluorescence (when [Cl − ]i approaches 0) was obtained by adding the K + /H + ionophore nigericin (200 mM) and the OH − /Cl − antiporter tributyltin (100 mM) (maximally effective concentrations of the agents were added to ensure penetration into the slice; lower concentrations may also work). Nigericin clamps pHi during changes of [Cl − ]o and tributyltin equilibrates the [Cl − ]o and [Cl − ]i (Krapf et al., 1988). Solutions containing three different Cl − concentrations (20, 40 and 125 mM) (pH 7.4) were then sequentially added to the slices along with nigericin and tributyltin, and the fluorescence was recorded. Finally, the fluorescence was completely quenched by adding potassium thiocyanate (150 mM) and the K + ionophore, valinomycin (25 mM), to the slice preparation. The SCN − ion has a higher Kq for MEQ than Cl − , and should completely quench the MEQ fluorescence (Biwersi and Verkman, 1991). The total quenchable signal (F0) was calculated by subtracting the fluorescence in the presence of KSCN from the fluorescence in the low Cl − buffer. The ratio of F0 to the fluorescence in the presence of increasing [Cl − ] (FCl) is given as F0/FCl. A plot of F0/FCl versus [Cl − ] provides a line from which the Stern-Volmer constant (the slope, Kq in M − 1) can be determined. The 1/Kq equals the concentration (M) of Cl − which quenches fluorescence by 50%.
2.5. Data acquisition and analysis In pharmacologic experiments, fluorescence imaging was restricted to three periods of data acquisition. The initial fluorescence brightness intensity (optical density; relative units between 0 and 256) was measured over an area (typically 100 pixels2) covering each soma for several cells prior to the bath application of GABA and/or GABAA receptor modulators (basal fluorescence). The buffer was rapidly exchanged for buffer containing the drug(s) of interest. In concentration-response studies, separate slices were used for each GABA concentration (0.01 – 10 mM; average of six cells/concentration). Subsequently, brightness intensity was measured at 10 and 20 min after drug addition (only the 20 min data are presented here) to ensure that the drug had reached cells deep within the slice. At each time point, fluorescence data were collected in three 0.5 s intervals (16 frames at 32/s). This routine kept photobleaching to a minimum (20 s total illumination per experiment; see Section 3). Background signals (from a nearby portion of the slice which did not contain fluorescing cells), which ranged from 7 to 10% of cellular signals, were subtracted from the brightness intensity of each cell at every time point. The change in fluorescence in response to drug application was de-
129
F × 100, where Fb F was the fluorescence intensity following drug exposure, Fb was basal fluorescence, and DF represents the % change in response. This method permitted each cell to serve as its own control.
termined by the equation, DF= 1−
3. Results
3.1. Biophysics of confocal imaging of MEQ-loaded cells MEQ fluorescence has been characterized previously in a variety of biological preparations (cultured cells and slices) using fluorescence photometry and epifluorescence microscopy (Biwersi and Verkman, 1991; Schwartz and Yu, 1995). This dye has high Cl − sensitivity, relatively low toxicity, slow leakage rate (B10%/ h), and resists photobleaching when exposed to incoherent light (Biwersi and Verkman, 1991). To determine the extent of photochemical bleaching (photobleaching) by the UV laser (coherent light) under our conditions, MEQ-loaded hippocampal neurons were illuminated continuously for at least 1200 s and brightness intensity was measured every second. The rate of photobleaching was first-order and the rate constant for decay (k) was 0.0041890.00068 s-1 (n=3) with a halflife (t1/2) of 173.89 24.6 s. The photobleached signal reached a stable level of 36.49 5.4% of the initial brightness intensity at 12179 173 s of illumination. Factors such as laser intensity, duration of laser exposure, dye concentration, and tissue type are involved in the rate of photobleaching. To avoid significant photobleaching resulting from long periods of laser illumination, fluorescence imaging and data acquisition were restricted to several intervals (lasting 6 s each) spaced approximately 10 min apart. Total illumination during a pharmacologic experiment never exceeded 20 s, and this corresponded to a decrease of 6.29 2.7% of the initial brightness intensity, as estimated from the firstorder rate equation1. Hence, MEQ exhibits a slow photobleach rate with reasonable intensity and duration of laser illumination. Autofluorescence is a potential problem, especially when using UV illumination (Mason, 1993). When slices were illuminated (364 nm) in the absence of dye there was no autofluorescence (Fig. 1E). This was also the case for illumination of the tissue at UV wavelengths with a mercury lamp (Schwartz and Yu, 1995). 1 A =A0 e − kt where A and A0 are absorbance (brightness) at the initial time (t=0) and at a defined length of total illumination, respectively; k is the decay rate constant, and t is duration of illumination.
130
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
In addition, when slices were incubated with the cellimpermeant form of dihydro-MEQ, MEQ (parent compound), there was no fluorescence at the laser intensity used in these studies. The absence of autofluorescence may reflect the excellent optics of the UV-transmissible objective (Olympus, 40× , UV). In hippocampal slices loaded with dihydro-MEQ (cell-permeant), there was a small fluorescent signal in the neuropil (where cells were absent). This slight fluorescence in the neuropil of loaded slices probably emanated from pyramidal cell dendrites or cerebral vessels that were loaded with dye and within the plane of focus. Interestingly, in discernible capillaries, we observed small GABA-mediated increases in intracellular Cl − , confirming previous findings with other techniques (Krause et al., 1980; Imai et al., 1991).
3.2. MEQ fluorescence in discrete neurons of the hippocampus, cortex, and cerebellum
Fig. 1. Video images of single optical sections of MEQ-loaded neurons located within dorsal hippocampus using UV-laser-scanning confocal microscopy. (A) area CA3 pyramidal cell layer; (B) area CA1 pyramidal cell layer; (C) and (D) interneurons in stratum radiatum of CA1 sector. Note the heterogeneous levels of MEQ fluorescence among neighboring cells within the same plane of focus. (E) Absence of autofluorescence in an unloaded slice; image was obtained from the pyramidal cell layer of area CA1. (Bar= 20 mm for A, B, and E; 15 mm for C and D).
Using confocal microscopy, numerous brightly fluorescing neurons with diverse morphologies could be imaged in slices loaded with dihydro-MEQ (Figs. 1–3). Despite a healthy appearance of neurons within a slice (determined in transmission mode using bright field microscopy), the intensity of MEQ fluorescence was not completely uniform among the cells. This may have been due to: (1) non-uniform penetration of dye into deeper levels of the slice; (2) cells being partly below or above a single focal plane; or (3) dye leakage from certain cells in the process of dying. Occasionally, there were some cells with uneven staining within the cell bodies. These cells were not used for data collection. Individual cells that exhibited MEQ fluorescence were visualized with a narrow focal plane ( 2 mm thickness) within dense neuronal populations such as the pyramidal cell layer in areas CA3 and CA1 of hippocampus (Fig. 1A and B). In addition, individual scattered interneurons were visualized within the hippocampal dendritic fields (neuropil; Fig. 1C and D). In transverse slices of neocortex, pyramidal neurons within layer V were easily identified with their long apical dendrites projecting toward more superficial layers (Fig. 2A). In sagittal cerebellar slices, individual brightly fluorescing granule cells were distinguished within the dense cell population (Fig. 2B). Moreover, Fig. 2C shows the clear demarcation of the granule cell, Purkinje cell, and molecular layers, detected at a slightly deeper focal plane. With greater magnification (680× ; electronic zoom factor of 1.7) of a Purkinje cell, the fluorescence of thick bifurcating dendrites as they extend into the molecular layer was easily observed (Fig. 2D). In fact, in all the regions assessed, fluorescent dendrites could be visualized as they emanated from the somata. The fluorescence emitted from these proximal dendrites was consistently less intense than that within their somata (Figs. 1–3).
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
131
Fig. 2. Video images of single optical sections of MEQ-loaded neurons within neocortex and cerebellum. (A) Layer V of somatosensory cortex; (B) granule cells in cerebellum; (C) Purkinje (P) and granule cells (g) in cerebellum clearly distinguished from the molecular layer (m). (D) Increased magnification of a Purkinje cell (680 ×). In this neuron, note the distinct dendritic fluorescence (arrows) as well as the heterogeneous levels of fluorescence between the somata and dendrites. (A– C) are at the same magnification. (Bars=20 mm).
3.3. MEQ fluorescence calibration
3.4. GABA-mediated changes in MEQ fluorescence
Calibration of the changes in MEQ fluorescence with increases in Cl − was made using the Stern-Volmer equation (Section 2). The Stern-Volmer plot was linear (r =0.9994) with a Kq of 16 91 M − 1 (n = 12), quite similar to values determined for SPQ and MEQ in other whole cell preparations, e.g. 12 M − 1 in renal tubules (Krapf et al., 1988), 17 M − 1 in salivary acinar cells (Foskett, 1990), 19 M − 1 in cultured fibroblasts (Biwersi and Verkman, 1991), and 14 M − 1 in epithelial cells (MacVinish et al., 1993). A Stern-Volmer constant of 16 M − 1 for pyramidal cells in our preparation corresponds to 50% quenching of MEQ by 61 mM Cl − . This calibration is useful for estimating the change in intracellular Cl − levels in relation to the change in MEQ fluorescence. It is also useful for comparing the sensitivity of Cl − sensitive dyes within different preparations.
Pharmacologic experiments were carried out in area CA1 hippocampal pyramidal cells. Stable responses to continuous bath-application of GABA were achieved after 10 min of superfusion. The GABA-induced changes in MEQ fluorescence within discrete hippocampal CA1 pyramidal cells (Fig. 3) followed a concentration-response relationship with an EC50 of 40 mM. According to the Stern-Volmer relationship (see Section 2), this corresponds to a change in [Cl − ]i of 16.8 mM. The GABA-induced change in MEQ fluorescence was prevented in the presence of the Cl − channel antagonist, picrotoxin (Table 1). Interestingly, one of six cells imaged was insensitive to the presence of picrotoxin. Bath perfusion of the slices with other drugs that enhance GABAA receptor function in the hippocampus significantly potentiated GABA-mediated changes in MEQ fluorescence. Maximally effective concentrations of pentobarbital and diazepam each enhanced the re-
132
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
Fig. 3. Pseudo-color video images of MEQ-loaded pyramidal neurons within hippocampal area CA1 before and 20 min following GABA (50 mM) application. MEQ fluorescence is quenched as GABA induces the increase in intracellular Cl − . Values for optical density are indicated by the color bar. (Bar=15 mm).
sponse to GABA in hippocampal pyramidal cells 85 and 44%, respectively, compared to GABA alone (Table 1). In contrast, low concentrations of ethanol, which do not reproducibly affect GABAergic neurotransmission in the hippocampus (White et al., 1990; Proctor et al., 1992; Criswell et al., 1993), failed to enhance GABA-mediated changes in MEQ fluorescence (Table 1). Because of the slow nature of bath application, the GABA responses measured were desensitized responses. The length of time needed to reach a steady level (at least 10 min of perfusion) of a GABA-mediated change in MEQ fluorescence may have depended on: (1) achieving significant drug levels deep in the slice; and (2) the presence of active GABA reuptake pumps. To achieve more rapid responses, we used a pressure ejection micropipet to apply the GABAA agonist, muscimol (not a substrate for reuptake), directly to the neuronal somata. With this method, a ‘puff’ of muscimol (40 mM in the pipet) caused a change in the MEQ fluorescence within 1 s of application and the maximal DF (53%) was reached approximately 3 s after the beginning of muscimol application (Fig. 4). The effect of muscimol
was blocked by prior application of the GABAA antagonist, bicuculline (not shown). Since clear images of individual MEQ-loaded neurons could be obtained in several brain regions using UV-laser-scanning confocal microscopy, GABA-mediated responses were compared among various cell types from the hippocampus, neocortex and cerebellum. After bath perfusion of the slices with the EC50 concentration of GABA (40 mM), there were similar responses among pyramidal cells and interneurons within the neocortex and hippocampus (Table 2). However, the same GABA concentration induced significantly larger responses in both cerebellar granule and Purkinje cells, relative to the three types of cerebral cortical neurons. In addition, Purkinje cells exhibited a significantly greater GABA-mediated change in MEQ fluorescence compared to cerebellar granule cells (Table 2).
Table 1 Effect of GABAA receptor ligands on MEQ fluorescence in individual CA1 pyramidal cells Compound
DF (% decrease)
GABA (40 mM) GABA+picrotoxin (1 mM) GABA+pentobarbital (1 mM) GABA+diazepam (50 mM) GABA+ethanol (50 mM)
24.09 2.9 4.09 6.5* 44.49 3.2* 34.792.2* 21.991.8
Data are mean9 S.E.M. from 6–14 cells. Drugs were bath-applied for 20 min prior to optical measurements. *PB0.05 vs. GABA; ANOVA and Dunnett’s multiple comparison’s test.
Fig. 4. MEQ fluorescence response of a hippocampal interneuron (area CA1 stratum radiatum) to the direct application via a pressure ejection micropipet of muscimol (40 mM; bar indicates duration drug was applied). Video images (8 frame average at 32/s) were recorded to an optical disk recorder at 0.8 s intervals for the first 10 s and at t =20 and 35 s.
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135 Table 2 Regional variability of GABAA receptor responses to 40 mM GABA in morphologically identified cells n
DF (% decrease)
Hippocampus (Area CA1) Pyramidal cell Interneuron
14 4
24.092.9 19.592.1
Neocortex (Layer V) Pyramidal cell
12
20.89 3.6
Cerebellar cortex Purkinje cell Granule cell
9 15
70.69 4.7* 38.59 5.7**
Region
Cell type
Data are mean9S.E.M. GABA was bath applied for 20 min prior to optical measurements. *PB0.05 compared to all other neurons. **PB0.05 compared to neocortical pyramidal cells. ANOVA and Tukey’s multiple comparisons test.
4. Discussion We have developed a confocal imaging method to measure Cl − - sensitive MEQ fluorescence within individual neurons in living brain slices. With confocal imaging of bath-loaded slices, we could measure, simultaneously, the GABAergic responses of several neurons within the field of view. Direct, independent evidence that changes in MEQ fluorescence were a measure of changes in [Cl − ]i was provided by the calibration procedure. The Kq of 16 M − 1 is in agreement with values calculated from other cell types (Krapf et al., 1988; Foskett, 1990; Biwersi and Verkman, 1991; MacVinish et al., 1993). Considerable structural detail was obtained from the optically-sectioned, MEQ fluorescent cells and dendrites. Even greater morphologic detail can be obtained with 3-D reconstruction of several subjacent optical slices (data not shown). Before the advent of optical imaging of [Cl − ]i in living brain slices, such combinations of pharmacological and structural detail could not be achieved with either electrophysiology or 36 Cl − flux techniques alone. The pharmacologic characterization of GABAA responses demonstrated by changes in MEQ fluorescence support previous findings from electrophysiological and biochemical studies. In both ion flux and electrophysiologic studies using the cerebral cortex (Alger and Nicoll, 1979; Collingridge et al., 1984; Schwartz et al., 1985; Allan and Harris, 1986), pentobarbital and diazepam enhance the response to GABA. As expected, the GABA response was blocked by the GABAA receptor antagonists, picrotoxin and bicuculline methiodide. In addition, ethanol concentrations up to 50 mM did not affect the GABA-mediated changes in MEQ fluorescence in the hippocampus. Similarly, low concentrations of ethanol have been shown to have little or no effect on hippocampal GABAergic neurotransmission
133
using electrophysiologic methods (White et al., 1990; Proctor et al., 1992; Criswell et al., 1993). With the appropriate pharmacology established using this imaging technique, greater temporal resolution in the MEQ response was achieved by applying muscimol directly to the cell (i.e. via a puffer pipet; Fig. 4). Certain regional differences in GABAA responses were also detected by laser-scanning confocal microscopy that both support and contrast with findings obtained with other methods. For example, cerebellar neurons exhibited a greater response to GABA compared to hippocampal and neocortical neurons. Within the cerebellum, a greater relative change in [Cl − ]i occurred within Purkinje cells compared to the granule cells. Cerebellar Purkinje cells, which are under major inhibitory control, also exhibit large, spontaneous GABAA receptor-mediated currents detected with electrophysiologic techniques (Konnerth et al., 1990). In addition, Puia et al. (1994) have demonstrated that GABA-activated Cl − currents (measured in milliseconds) have different kinetics within Purkinje versus granule cells. However, our findings differ from the relative responses to GABA measured in synaptoneurosomes from the hippocampus, neocortex, and cerebellum (Allan and Harris, 1986; Luu et al., 1987). Thus, responses obtained in isolated membrane vesicles may not always be equivalent to those in living brain slices where circuitry is generally intact. The regional heterogeneity of GABAA receptor responses could arise for a variety of reasons, including: (a) diversity in the GABAA receptor subunit composition among different brain regions and neuronal types (Richards et al., 1987; Gutie´rrez et al., 1994; Inglefield et al., 1994; Fritschy and Mohler, 1995); (b) regional differences in the degree of phosphorylation, which modulates the GABAA receptor function (Stelzer et al., 1988; Heuschneider and Schwartz, 1989); or (c) regional differences in the anion/cation transporter or Cl − /HCO3-exchanger which would contribute to different Cl − gradients among neurons. Confocal imaging of multiple cell types within the living brain slice provides an opportunity to investigate these potential mechanisms underlying the regional heterogeneity of GABAA receptor responses. An advantage of both confocal and non-confocal imaging of MEQ fluorescence (compared to electrophysiology, for example) is that one can compare the GABAA responses in different parts of the neuron at the same time, without the need for multiple recording pipettes. In all neuronal types examined with UV-laserscanning confocal microscopy from post-natal day 12– 21, there was lower MEQ fluorescence in the dendrite relative to the somata. This could be due to: (1) the geometry of the thin dendrites relative to the soma; or (2) a Cl − gradient from relatively low levels within the soma to higher levels in the dendrite. In fact, Hara et al. (1992) have provided evidence for differences in Cl −
134
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
extrusion mechanisms between the somata and apical dendrite of cultured hippocampal neurons to account for a spatial unevenness of Cl − -sensitive fluorescence. In another study, local application of GABA to hippocampal pyramidal cell dendrites increased extracellular Cl − , presumably due to a pre-existing outward Cl − gradient in the dendrites (Muller et al., 1989). Whether an altered Cl − gradient is the basis for depolarizing responses to GABA application in dendrites of CA1 pyramidal cells (Alger and Nicoll, 1979) is still debated (Staley et al., 1995). While fluorescence photometry has been used successfully to study GABAA receptor activity in synaptoneurosomes, cultured cells, and engineered cell lines (Engblom et al., 1991; Engblom and Akerman, 1991; Wong et al., 1992; Hara et al., 1992), this is the first study to assess the optical characteristics and pharmacology of GABAA receptor activity in living brain slices with confocal imaging. This technique has excellent potential for several applications. First, the spatial resolution of a confocal microscope combined with the temporal resolution of electrophysiology can permit the study of signaling actions at synapses on dendrites and soma simultaneously. Whereas the bath-loading technique limits the spatial resolution somewhat to the level of the somata (as well as proximal dendrites), direct application of MEQ (non-reduced form) into the cell and its processes via a ‘patch pipet’ would provide even better spatial resolution. This would be a particular advantage for optical imaging in adult slices where bath-loading has limited success. Second, in this study, desensitized responses to GABA were measured minutes after continuous bath application. With the use of local drug application to a single cell (or dendrite) via a pressure ejection micropipet, the time between application of drug or neurotransmitter and detection of the response can be as low as B1 s. Finally, the high spatial and temporal resolution ( 0.001 s) of photoactivation (‘uncaging’) may be ultimately the best method for mapping GABA responses in situ (Wang and Augustine, 1995). In conclusion, whatever method is used for drug application, UV-confocal microscopy and Cl − sensitive dyes may have an important impact on future studies of GABA neurotransmission in both developing and adult neurons.
Acknowledgements This work was supported by a Fellowship from the American Heart Association, North Carolina Affiliate, Inc. to J.R.I., and NIH grant NS 28791 to R.D.S. We wish to thank Drs George Augustine, Victor Nadler, and Elizabeth Finch for helpful discussions relating to the development of the technique.
References Alger BE, Nicoll RA. GABA-mediated biphasic inhibitory responses in hippocampus. Nature 1979;281:315 – 7. Allan AM, Harris RA. g-Aminobutyric acid agonists and antagonists alter chloride flux across brain membranes. Mol Pharmacol 1986;29:497 – 505. Biwersi J, Verkman AS. Cell-permeable fluorescent indicator for cytosolic chloride. Biochemistry 1991;30:7879 – 83. Collingridge GL, Gage PW, Robertson B. Inhibitory post-synaptic currents in rat hippocampal CA1 neurones. J Physiol 1984;356:551 – 64. Criswell HE, Simson PE, Duncan GE, McCown TJ, Herbert JS, Morrow AL, Breese GR. Molecular basis for regionally specific action of ethanol on g-aminobutyric acidA receptors: Generalization to other ligand-gated ion channels. J Pharmacol Exp Ther 1993;267:522 – 37. Dailey ME, Smith SJ. Spontaneous Ca2 + transients in developing hippocampal pyramidal cells. J Neurobiol 1994;25:243 –51. Engblom AC, Akerman KEO. Effect of ethanol on g-aminobutryic acid and glycine receptor-coupled Cl − fluxes in rat brain synaptoneurosomes. J Neurochem 1991;57:384 – 90. Engblom AC, Holopainen I, Akerman KEO. Ethanol-induced Cl − flux in rat cerebellar granule cells as measured by a fluorescent probe. Brain Res 1991;568:55 – 60. Foskett JK. [Ca2 + ]i modulation of Cl − content controls cell volume in single salivary acinar cells during fluid secretion. Am J Physiol 1990;259:C998 – C1004. Frenguelli BG, Potier B, Slater NT, Alford S, Collingridge GL. Metabotropic glutamate receptors and calcium signalling in dendrites of hippocampal CA1 neurones. Neuropharm 1993;32:1229– 37. Fritschy J-M, Mohler H. GABAA-receptor heterogeneity in the adult rat brain: Differential regional and cellular distribution of seven major subunits. J Comp Neurol 1995;359:154 – 94. Guthrie PB, Segal M, Kater SB. Independent regulation of calcium revealed by imaging dendritic spines. Nature 1991;354:76–80. Gutie´rrez A, Khan Z, de Blas AL. Immunocytochemical localization of g2 short and g2 long subunits of the GABAA receptor in the rat brain. J Neurosci 1994;14:7168 – 79. Hara M, Inoue M, Yasukura T, Ohnishi S, Mikami Y, Inagaki C. Uneven distribution of intracellular Cl − in rat hippocampal neurons. Neurosci Lett 1992;143:135 – 8. Harris RA, Allan AM. Functional coupling of g-aminobutyric acid receptors to chloride channels in brain membranes. Science 1985;228:1108 – 9. Hayashi H, Miyata H. Fluorescence imaging of intracellular Ca2 + . J Pharmacol Toxicol Methods 1994;31:1 – 10. Heuschneider G, Schwartz RD. cAMP and forskolin decrease gaminobutryic acid-gated chloride flux in rat brain synaptoneurosomes. Proc Natl Acad Sci 1989;86:2938 – 42. Imai H, Okuno T. Wu, J.Y. and Lee, T.J.-F. (1991) GABAergic innervation in cerebral blood vessels: An immunohistochemical demonstration of L-glutamic acid decarboxylase and GABA transaminase. J Cereb Blood Flow Metab 1991;11:129 –34. Inglefield JR, Schwartz-Bloom. Confocal imaging of intracellular chloride ([Cl − ]i) in acutely prepared slices. Measurement of GABAA receptor activity. Soc Neurosci Abstracts 1996;22:324.13. Inglefield JR, Sieghart W, Kellogg CK. Immunohistochemical and neurochemical evidence for GABAA receptor heterogeneity between the hypothalamus and cortex. J Chem Neuroanat 1994;7:243 – 52. Jaffe DB, Fisher SA, Brown TH. Confocal laser scanning microscopy reveals voltage-gated calcium signals within hippocampal dendritic spines. J Neurobiol 1994;25:220 – 33.
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135 Konnerth A, Llano I, Armstrong CM. Synaptic currents in cerebellar Purkinje cells. Proc Natl Acad Sci 1990;87:2662–5. Krapf R, Berry CA, Verkman AS. Estimation of intracellular chloride activity in isolated perfused rabbit proximal convoluted tubules using a fluorescent indicator. Biophys J 1988;53:955 – 62. Krause DN, Wong E, Degener P, Roberts E. GABA receptors in bovine cerebral blood vessels: binding studies with tritiated muscimol. Brain Res 1980;185:51–7. Krnjevic K, Schwartz S. The action of g-aminobutyric acid on cortical neurones. Exp Brain Res 1967;3:320–36. Llano I, Dreessen J, Kano M, Konnerth A. Intradendritic release of calcium induced by glutamate in cerebellar purkinje cells. Neuron 1991;7:577 – 83. Luu MD, Morrow AL, Paul SM, Schwartz RD. Characterization of GABAA receptor-mediated 36chloride uptake in rat brain synaptoneurosomes. Life Sci 1987;41:1277–87. MacVinish LJ, Reancharoen T, Cuthbert AW. Kinin-induced chloride permeability changes in colony 29 epithelia estimated from 125I − efflux and MEQ fluorescence. Br J Pharmacol 1993;108:469 – 78. Mason WT. Fluorescent and Luminescent Probes for Biological Activity. New York: Academic Press, 1993:433. Muller W, Misgeld U, Lux HD. g-Aminobutryic acid-induced ion movements in the guinea pig hippocampal slice. Brain Res 1989;484:184 – 91. Porter JT, McCarthy KD. GFAP-positive hippocampal astrocytes in situ respond to glutamatergic neuroligands with increases in [Ca2 + ]i. Glia 1995;13:101–12. Proctor WR, Soldo BL, Allan AM, Dunwiddie TV. Ethanol enhances synaptically evoked GABAA receptor-mediated responses in cerebral cortical neurons in rat brain slices. Brain Res 1992;595:220 – 7. Puia G, Costa E, Vicini S. Functional diversity of GABA-activated Cl − currents in Purkinje versus granule neurons in rat cerebellar slices. Neuron 1994;12:117–26.
.
.
135
Regehr WG, Connor JA, Tank DW. Optical imaging of calcium accumulation in hippocampal pyramidal cells during synaptic activation. Nature 1989;341:533 – 6. Richards JG, Schoch P, Ha¨ring P, Takacs B, Mo¨hler H. Resolving GABAA/benzodiazepine receptors: Cellular and subcellular localization in the CNS with monoclonal antibodies. J Neurosci 1987;7:1866 – 86. Schwartz RD, Jackson JA, Wiegert D, Skolnick P, Paul SM. Characterization of barbiturate-stimulated chloride efflux from rat brain synaptoneurosomes. J Neurosci 1985;5:2963 – 70. Schwartz RD, Yu X. Optical imaging of intracellular chloride in living brain slices. J Neurosci Methods 1995;62:185 – 92. Staley KJ, Soldo BL, Proctor WR. Ionic mechanisms of neuronal excitation by inhibitory GABAA receptors. Science 1995;269:977– 81. Stelzer A, Kay AR, Wong RKS. GABAA-receptor function in hippocampal cells is maintained by phosphorylation factors. Science 1988;241:339 – 41. Tsien RY. Fluorescence measurement and photochemical manipulation of cytosolic free calcium. Trends Neurosci 1988;11:419–24. Verkman AS. Development and biological applications of chloridesensitive fluorescent indicators. Am J Physiol 1990;259:C375–88. Wang SS-H, Augustine GJ. Confocal imaging and local photolysis of caged compounds: Dual probes of synaptic function. Neuron 1995;15:755 – 60. White G, Lovinger D, Weight F. Ethanol inhibits NMDA-activated current but does not alter GABA-activated current in an isolated adult mammalian neuron. Brain Res 1990;507:332 – 6. Wong G, Sei Y, Skolnick P. Stable expression of type I g-aminobutyric acidA/benzodiazepine receptors in a transfected cell line. Mol Pharmacol 1992;42:996 – 1003. Yuste R, Katz LC. Control of postsynaptic Ca2 + influx in developing neocortex by excitatory and inhibitory neurotransmitters. Neuron 1996;6:333 – 44.
Confocal imaging of intracellular chloride in living brain slices: measurement of GABAA receptor activity Jon R. Inglefield *, Rochelle D. Schwartz-Bloom Department of Pharmacology, Box 3813, Duke Uni6ersity Medical Center, Durham, NC 27710, USA Received 12 December 1996; accepted 2 January 1997
Abstract We have developed a method using UV laser-scanning confocal microscopy and the fluorescent chloride ion indicator, 6-methoxy-N-ethylquinolinium chloride (MEQ), to image GABA-mediated changes in intracellular chloride (Cl− i ) in individual neurons of the rat acute brain slice. After bath-loading slices with the cell-permeant form (reduced) of MEQ, there was intense fluorescence within neurons of diverse morphologies in the hippocampus, neocortex and cerebellum. MEQ fluorescence localized to the cytosolic compartment of both the somata and proximal dendrites. MEQ fluorescence was calibrated using the ionophores nigericin and tributyltin in the presence of varying extracellular Cl − concentrations. Neuronal MEQ fluorescence was inversely related to intracellular Cl − , with a Stern-Volmer constant of 16 M − 1 (50% quench by 61 mM Cl − ). Application of GABA in the perfusate produced a concentration-dependent decrease in MEQ fluorescence (EC50 = 40 mM) that was blocked in the presence of the Cl − channel antagonist, picrotoxin. Bath perfusion of hippocampal slices with modulators of the GABAA receptor, pentobarbital and diazepam, potentiated the GABA-mediated response by 85 and 44%, respectively. A regional comparison identified larger GABA responses for both cerebellar Purkinje and granule cells relative to pyramidal neurons of the hippocampus and neocortex and to hippocampal interneurons. Pressure ejection of the GABAA agonist, muscimol (40 mM), from a micropipet onto individual hippocampal neurons allowed the measurement of rapid responses (1 – 5 s), compared to those obtained with bath application. Thus, optical imaging of [Cl − ]i using MEQ and UV-laser-scanning confocal microscopy provides investigators with a new method to study GABAA pharmacology in neighboring neurons and perhaps even in the soma versus dendrites, simultaneously, within living brain slices. © 1997 Elsevier Science Ireland Ltd. Keywords: UV laser-scanning confocal microscopy; Hippocampus; Neocortex; Fluorescence; Cerebellum; Chloride sensitive dye; Inhibition
1. Introduction The responses of individual cells to various stimuli have been examined using fluorescent tracers for bioactive molecules (Tsien, 1988; Biwersi and Verkman, 1991; Hayashi and Miyata, 1994). For example, in cultured neurons and in brain slices, fluorescent indicators have been used to measure increases in intracellular calcium ([Ca2 + ]i), both in the somata (Yuste and Katz, 1991; Dailey and Smith, 1994; Porter and McCarthy, * Corresponding author. Tel.: +1 919 6845181; fax: + 1 919 6818609; e-mail: [email protected]
1995) and in dendrites (Regehr et al., 1989; Guthrie et al., 1991; Llano et al., 1991; Frenguelli et al., 1993; Jaffe et al., 1994) during or after neuronal excitation. In contrast, changes in intracellular chloride ([Cl − ]i) following GABAA receptor activation have been studied traditionally with electrophysiological measures (Krnjevic and Schwartz, 1967; Alger and Nicoll, 1979) and 36 Cl − flux techniques (Schwartz et al., 1985; Harris and Allan, 1985). With the advent of Cl − -sensitive fluorescent dyes such as 6-methoxy-N-[3-sulfopropyl] quinolinium (SPQ) and 6-methoxy-N-ethyl-quinolinium chloride (MEQ), changes in [Cl − ]i have been measured in non-neuronal cells (Biwersi and Verkman, 1991;
0165-0270/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 1 6 5 - 0 2 7 0 ( 9 7 ) 0 0 0 5 4 - X
128
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
Verkman, 1990 for review) and engineered cell lines (Wong et al., 1992) with fluorescence photometry. In addition, Cl − -sensitive fluorescent indicators and fluorescence photometry have been applied to the study of GABA-induced Cl − transport within both synaptoneurosomes (Engblom and Akerman, 1991) and cultured neurons (Engblom et al., 1991; Hara et al., 1992). Recently, we developed an optical imaging technique to measure GABA-mediated changes in [Cl − ]i within living brain slices using conventional epifluorescence microscopy and MEQ (Schwartz and Yu, 1995). Optical imaging in living cell preparations has become a widely used technique with the aid of a confocal microscope. Imaging with laser-scanning confocal microscopy greatly improves the resolution (both temporal and spatial) of dynamic changes in intracellular processes. It is especially useful in brain slice preparations, where cells can be imaged deep in the slice, away from cut surfaces and potentially injured cells. Thus, we took advantage of better spatial and temporal resolution using ultraviolet (UV)-laser-scanning confocal microscopy to image optically the responses to GABAA receptor activation in various types of individual neurons, simultaneously, within their native environment. The purposes of this study were to: (1) characterize the measurement of GABAA receptor responses using confocal microscopy and the Cl − -sensitive dye, MEQ; (2) calibrate the change in MEQ fluorescence intracellularly in a brain slice preparation; and (3) demonstrate rapid changes in [Cl − ]i following GABAA activation. These results have been presented in preliminary form (Inglefield and Schwartz-Bloom, 1996).
2. Materials and methods
2.1. Synthesis of dihydro-MEQ The reduction of MEQ (Molecular Probes, Eugene, OR) to a cell- permeable form, 6-methoxy-N-ethyl-1,2dihydroquinoline (dihydro-MEQ) using sodium borohydride was described previously (Biwersi and Verkman, 1991), with modifications by our laboratory (Schwartz and Yu, 1995). Prior to each experiment, the dried dihydro-MEQ extract (under N2 at -20°C) was re-suspended in 15 ml of ethyl acetate and used immediately. For optimal use with confocal imaging, the dried dihydro-MEQ extract should be used within two days.
2.2. Slice preparation and bath-loading of dihydro-MEQ Transverse hippocampal (containing somatosensory cortex) and sagittal cerebellar slices were prepared from 12 – 21 day old Sprague-Dawley rats (Charles River). Slices (300 mm for hippocampus; 200 mm for cerebel-
lum) were cut with a vibratome in oxygenated (95% O2/5% CO2 mixture) Ringer’s physiological buffer (0– 4°C), containing (in mM) 119 NaCl, 2.5 KCl, 1.0 NaH2PO4, 1.3 MgSO4, 2.5 CaCl2, 26 NaHCO3, and 11 glucose, pH 7.4. Slices were warmed to 28°C in oxygenated Ringer’s buffer and then incubated with the re-suspended dihydro-MEQ (final concentration of 300 mM) at 28°C for 0.5 h in Ringer’s buffer. Once loaded into cells, dihydro-MEQ is oxidized to MEQ, which remains trapped in the cytosolic compartment (Biwersi and Verkman, 1991). Following incubation, slices were washed once in fresh oxygenated Ringer’s buffer to remove extracellular dihydro-MEQ, and then kept in fresh buffer for 0.5–3 h before transfer to the imaging chamber.
2.3. UV laser-scanning confocal microscopy For imaging, slices were submerged in a chamber that was constantly perfused with oxygenated Ringer’s buffer (1.5–2.0 ml/min) at room temperature. The imaging chamber was positioned on the stage of an upright Nikon Optiphot microscope. The laser scanning confocal microscope (Noran Odyssey™, Noran Instruments) was equipped with an argon UV laser (80 mW output, Enterprise 653, Coherent), and a digital imaging system. A 486 personal computer along with Image1™ software (Universal Imaging) was used for image/data acquisition and Odyssey™ software was used to control the confocal microscope. MEQ fluorescence was excited with the UV laser (20% intensity) and dichroic mirror to exclude all but the 364 nm line. Fluorescent light was transmitted through a UV water immersible objective lens (40 × , NA 0.7, Olympus). Emission (Emmax = 440 nm) was imaged using a 400 nm barrier filter and photomultipliers received the signal through a confocal slit aperture of 25 mm. (Any remaining dihydro-MEQ within the slice has an emission spectrum different from MEQ and does not interfere with MEQ fluorescence detection). The video frame rates (32 images per s) of the Noran Odyssey confocal microscope permitted rapid, full image (512×480 pixels) acquisition.
2.4. Intracellular calibration of MEQ fluorescence The fluorescence of MEQ is quenched collisionally by Cl − according to the Stern-Volmer equation, Fo/FCl = 1+ Kq[Cl], where F0 is the total quenchable signal, FCl is the fluorescence in the presence of a given Cl − concentration, and Kq is the Stern-Volmer quenching constant (in M − 1) (Verkman, 1990). Intracellular MEQ fluorescence was calibrated to give estimates for intracellular Cl − according to Krapf et al. (1988), with modifications for the brain slice. MEQ-loaded slices were equilibrated for 30 min in an oxygenated, low Cl −
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
(2.5 mM) buffer (Na-gluconate replaced NaCl, and K-gluconate replaced KCl in an equimolar manner) (pH 7.4). The maximal fluorescence (when [Cl − ]i approaches 0) was obtained by adding the K + /H + ionophore nigericin (200 mM) and the OH − /Cl − antiporter tributyltin (100 mM) (maximally effective concentrations of the agents were added to ensure penetration into the slice; lower concentrations may also work). Nigericin clamps pHi during changes of [Cl − ]o and tributyltin equilibrates the [Cl − ]o and [Cl − ]i (Krapf et al., 1988). Solutions containing three different Cl − concentrations (20, 40 and 125 mM) (pH 7.4) were then sequentially added to the slices along with nigericin and tributyltin, and the fluorescence was recorded. Finally, the fluorescence was completely quenched by adding potassium thiocyanate (150 mM) and the K + ionophore, valinomycin (25 mM), to the slice preparation. The SCN − ion has a higher Kq for MEQ than Cl − , and should completely quench the MEQ fluorescence (Biwersi and Verkman, 1991). The total quenchable signal (F0) was calculated by subtracting the fluorescence in the presence of KSCN from the fluorescence in the low Cl − buffer. The ratio of F0 to the fluorescence in the presence of increasing [Cl − ] (FCl) is given as F0/FCl. A plot of F0/FCl versus [Cl − ] provides a line from which the Stern-Volmer constant (the slope, Kq in M − 1) can be determined. The 1/Kq equals the concentration (M) of Cl − which quenches fluorescence by 50%.
2.5. Data acquisition and analysis In pharmacologic experiments, fluorescence imaging was restricted to three periods of data acquisition. The initial fluorescence brightness intensity (optical density; relative units between 0 and 256) was measured over an area (typically 100 pixels2) covering each soma for several cells prior to the bath application of GABA and/or GABAA receptor modulators (basal fluorescence). The buffer was rapidly exchanged for buffer containing the drug(s) of interest. In concentration-response studies, separate slices were used for each GABA concentration (0.01 – 10 mM; average of six cells/concentration). Subsequently, brightness intensity was measured at 10 and 20 min after drug addition (only the 20 min data are presented here) to ensure that the drug had reached cells deep within the slice. At each time point, fluorescence data were collected in three 0.5 s intervals (16 frames at 32/s). This routine kept photobleaching to a minimum (20 s total illumination per experiment; see Section 3). Background signals (from a nearby portion of the slice which did not contain fluorescing cells), which ranged from 7 to 10% of cellular signals, were subtracted from the brightness intensity of each cell at every time point. The change in fluorescence in response to drug application was de-
129
F × 100, where Fb F was the fluorescence intensity following drug exposure, Fb was basal fluorescence, and DF represents the % change in response. This method permitted each cell to serve as its own control.
termined by the equation, DF= 1−
3. Results
3.1. Biophysics of confocal imaging of MEQ-loaded cells MEQ fluorescence has been characterized previously in a variety of biological preparations (cultured cells and slices) using fluorescence photometry and epifluorescence microscopy (Biwersi and Verkman, 1991; Schwartz and Yu, 1995). This dye has high Cl − sensitivity, relatively low toxicity, slow leakage rate (B10%/ h), and resists photobleaching when exposed to incoherent light (Biwersi and Verkman, 1991). To determine the extent of photochemical bleaching (photobleaching) by the UV laser (coherent light) under our conditions, MEQ-loaded hippocampal neurons were illuminated continuously for at least 1200 s and brightness intensity was measured every second. The rate of photobleaching was first-order and the rate constant for decay (k) was 0.0041890.00068 s-1 (n=3) with a halflife (t1/2) of 173.89 24.6 s. The photobleached signal reached a stable level of 36.49 5.4% of the initial brightness intensity at 12179 173 s of illumination. Factors such as laser intensity, duration of laser exposure, dye concentration, and tissue type are involved in the rate of photobleaching. To avoid significant photobleaching resulting from long periods of laser illumination, fluorescence imaging and data acquisition were restricted to several intervals (lasting 6 s each) spaced approximately 10 min apart. Total illumination during a pharmacologic experiment never exceeded 20 s, and this corresponded to a decrease of 6.29 2.7% of the initial brightness intensity, as estimated from the firstorder rate equation1. Hence, MEQ exhibits a slow photobleach rate with reasonable intensity and duration of laser illumination. Autofluorescence is a potential problem, especially when using UV illumination (Mason, 1993). When slices were illuminated (364 nm) in the absence of dye there was no autofluorescence (Fig. 1E). This was also the case for illumination of the tissue at UV wavelengths with a mercury lamp (Schwartz and Yu, 1995). 1 A =A0 e − kt where A and A0 are absorbance (brightness) at the initial time (t=0) and at a defined length of total illumination, respectively; k is the decay rate constant, and t is duration of illumination.
130
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
In addition, when slices were incubated with the cellimpermeant form of dihydro-MEQ, MEQ (parent compound), there was no fluorescence at the laser intensity used in these studies. The absence of autofluorescence may reflect the excellent optics of the UV-transmissible objective (Olympus, 40× , UV). In hippocampal slices loaded with dihydro-MEQ (cell-permeant), there was a small fluorescent signal in the neuropil (where cells were absent). This slight fluorescence in the neuropil of loaded slices probably emanated from pyramidal cell dendrites or cerebral vessels that were loaded with dye and within the plane of focus. Interestingly, in discernible capillaries, we observed small GABA-mediated increases in intracellular Cl − , confirming previous findings with other techniques (Krause et al., 1980; Imai et al., 1991).
3.2. MEQ fluorescence in discrete neurons of the hippocampus, cortex, and cerebellum
Fig. 1. Video images of single optical sections of MEQ-loaded neurons located within dorsal hippocampus using UV-laser-scanning confocal microscopy. (A) area CA3 pyramidal cell layer; (B) area CA1 pyramidal cell layer; (C) and (D) interneurons in stratum radiatum of CA1 sector. Note the heterogeneous levels of MEQ fluorescence among neighboring cells within the same plane of focus. (E) Absence of autofluorescence in an unloaded slice; image was obtained from the pyramidal cell layer of area CA1. (Bar= 20 mm for A, B, and E; 15 mm for C and D).
Using confocal microscopy, numerous brightly fluorescing neurons with diverse morphologies could be imaged in slices loaded with dihydro-MEQ (Figs. 1–3). Despite a healthy appearance of neurons within a slice (determined in transmission mode using bright field microscopy), the intensity of MEQ fluorescence was not completely uniform among the cells. This may have been due to: (1) non-uniform penetration of dye into deeper levels of the slice; (2) cells being partly below or above a single focal plane; or (3) dye leakage from certain cells in the process of dying. Occasionally, there were some cells with uneven staining within the cell bodies. These cells were not used for data collection. Individual cells that exhibited MEQ fluorescence were visualized with a narrow focal plane ( 2 mm thickness) within dense neuronal populations such as the pyramidal cell layer in areas CA3 and CA1 of hippocampus (Fig. 1A and B). In addition, individual scattered interneurons were visualized within the hippocampal dendritic fields (neuropil; Fig. 1C and D). In transverse slices of neocortex, pyramidal neurons within layer V were easily identified with their long apical dendrites projecting toward more superficial layers (Fig. 2A). In sagittal cerebellar slices, individual brightly fluorescing granule cells were distinguished within the dense cell population (Fig. 2B). Moreover, Fig. 2C shows the clear demarcation of the granule cell, Purkinje cell, and molecular layers, detected at a slightly deeper focal plane. With greater magnification (680× ; electronic zoom factor of 1.7) of a Purkinje cell, the fluorescence of thick bifurcating dendrites as they extend into the molecular layer was easily observed (Fig. 2D). In fact, in all the regions assessed, fluorescent dendrites could be visualized as they emanated from the somata. The fluorescence emitted from these proximal dendrites was consistently less intense than that within their somata (Figs. 1–3).
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
131
Fig. 2. Video images of single optical sections of MEQ-loaded neurons within neocortex and cerebellum. (A) Layer V of somatosensory cortex; (B) granule cells in cerebellum; (C) Purkinje (P) and granule cells (g) in cerebellum clearly distinguished from the molecular layer (m). (D) Increased magnification of a Purkinje cell (680 ×). In this neuron, note the distinct dendritic fluorescence (arrows) as well as the heterogeneous levels of fluorescence between the somata and dendrites. (A– C) are at the same magnification. (Bars=20 mm).
3.3. MEQ fluorescence calibration
3.4. GABA-mediated changes in MEQ fluorescence
Calibration of the changes in MEQ fluorescence with increases in Cl − was made using the Stern-Volmer equation (Section 2). The Stern-Volmer plot was linear (r =0.9994) with a Kq of 16 91 M − 1 (n = 12), quite similar to values determined for SPQ and MEQ in other whole cell preparations, e.g. 12 M − 1 in renal tubules (Krapf et al., 1988), 17 M − 1 in salivary acinar cells (Foskett, 1990), 19 M − 1 in cultured fibroblasts (Biwersi and Verkman, 1991), and 14 M − 1 in epithelial cells (MacVinish et al., 1993). A Stern-Volmer constant of 16 M − 1 for pyramidal cells in our preparation corresponds to 50% quenching of MEQ by 61 mM Cl − . This calibration is useful for estimating the change in intracellular Cl − levels in relation to the change in MEQ fluorescence. It is also useful for comparing the sensitivity of Cl − sensitive dyes within different preparations.
Pharmacologic experiments were carried out in area CA1 hippocampal pyramidal cells. Stable responses to continuous bath-application of GABA were achieved after 10 min of superfusion. The GABA-induced changes in MEQ fluorescence within discrete hippocampal CA1 pyramidal cells (Fig. 3) followed a concentration-response relationship with an EC50 of 40 mM. According to the Stern-Volmer relationship (see Section 2), this corresponds to a change in [Cl − ]i of 16.8 mM. The GABA-induced change in MEQ fluorescence was prevented in the presence of the Cl − channel antagonist, picrotoxin (Table 1). Interestingly, one of six cells imaged was insensitive to the presence of picrotoxin. Bath perfusion of the slices with other drugs that enhance GABAA receptor function in the hippocampus significantly potentiated GABA-mediated changes in MEQ fluorescence. Maximally effective concentrations of pentobarbital and diazepam each enhanced the re-
132
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
Fig. 3. Pseudo-color video images of MEQ-loaded pyramidal neurons within hippocampal area CA1 before and 20 min following GABA (50 mM) application. MEQ fluorescence is quenched as GABA induces the increase in intracellular Cl − . Values for optical density are indicated by the color bar. (Bar=15 mm).
sponse to GABA in hippocampal pyramidal cells 85 and 44%, respectively, compared to GABA alone (Table 1). In contrast, low concentrations of ethanol, which do not reproducibly affect GABAergic neurotransmission in the hippocampus (White et al., 1990; Proctor et al., 1992; Criswell et al., 1993), failed to enhance GABA-mediated changes in MEQ fluorescence (Table 1). Because of the slow nature of bath application, the GABA responses measured were desensitized responses. The length of time needed to reach a steady level (at least 10 min of perfusion) of a GABA-mediated change in MEQ fluorescence may have depended on: (1) achieving significant drug levels deep in the slice; and (2) the presence of active GABA reuptake pumps. To achieve more rapid responses, we used a pressure ejection micropipet to apply the GABAA agonist, muscimol (not a substrate for reuptake), directly to the neuronal somata. With this method, a ‘puff’ of muscimol (40 mM in the pipet) caused a change in the MEQ fluorescence within 1 s of application and the maximal DF (53%) was reached approximately 3 s after the beginning of muscimol application (Fig. 4). The effect of muscimol
was blocked by prior application of the GABAA antagonist, bicuculline (not shown). Since clear images of individual MEQ-loaded neurons could be obtained in several brain regions using UV-laser-scanning confocal microscopy, GABA-mediated responses were compared among various cell types from the hippocampus, neocortex and cerebellum. After bath perfusion of the slices with the EC50 concentration of GABA (40 mM), there were similar responses among pyramidal cells and interneurons within the neocortex and hippocampus (Table 2). However, the same GABA concentration induced significantly larger responses in both cerebellar granule and Purkinje cells, relative to the three types of cerebral cortical neurons. In addition, Purkinje cells exhibited a significantly greater GABA-mediated change in MEQ fluorescence compared to cerebellar granule cells (Table 2).
Table 1 Effect of GABAA receptor ligands on MEQ fluorescence in individual CA1 pyramidal cells Compound
DF (% decrease)
GABA (40 mM) GABA+picrotoxin (1 mM) GABA+pentobarbital (1 mM) GABA+diazepam (50 mM) GABA+ethanol (50 mM)
24.09 2.9 4.09 6.5* 44.49 3.2* 34.792.2* 21.991.8
Data are mean9 S.E.M. from 6–14 cells. Drugs were bath-applied for 20 min prior to optical measurements. *PB0.05 vs. GABA; ANOVA and Dunnett’s multiple comparison’s test.
Fig. 4. MEQ fluorescence response of a hippocampal interneuron (area CA1 stratum radiatum) to the direct application via a pressure ejection micropipet of muscimol (40 mM; bar indicates duration drug was applied). Video images (8 frame average at 32/s) were recorded to an optical disk recorder at 0.8 s intervals for the first 10 s and at t =20 and 35 s.
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135 Table 2 Regional variability of GABAA receptor responses to 40 mM GABA in morphologically identified cells n
DF (% decrease)
Hippocampus (Area CA1) Pyramidal cell Interneuron
14 4
24.092.9 19.592.1
Neocortex (Layer V) Pyramidal cell
12
20.89 3.6
Cerebellar cortex Purkinje cell Granule cell
9 15
70.69 4.7* 38.59 5.7**
Region
Cell type
Data are mean9S.E.M. GABA was bath applied for 20 min prior to optical measurements. *PB0.05 compared to all other neurons. **PB0.05 compared to neocortical pyramidal cells. ANOVA and Tukey’s multiple comparisons test.
4. Discussion We have developed a confocal imaging method to measure Cl − - sensitive MEQ fluorescence within individual neurons in living brain slices. With confocal imaging of bath-loaded slices, we could measure, simultaneously, the GABAergic responses of several neurons within the field of view. Direct, independent evidence that changes in MEQ fluorescence were a measure of changes in [Cl − ]i was provided by the calibration procedure. The Kq of 16 M − 1 is in agreement with values calculated from other cell types (Krapf et al., 1988; Foskett, 1990; Biwersi and Verkman, 1991; MacVinish et al., 1993). Considerable structural detail was obtained from the optically-sectioned, MEQ fluorescent cells and dendrites. Even greater morphologic detail can be obtained with 3-D reconstruction of several subjacent optical slices (data not shown). Before the advent of optical imaging of [Cl − ]i in living brain slices, such combinations of pharmacological and structural detail could not be achieved with either electrophysiology or 36 Cl − flux techniques alone. The pharmacologic characterization of GABAA responses demonstrated by changes in MEQ fluorescence support previous findings from electrophysiological and biochemical studies. In both ion flux and electrophysiologic studies using the cerebral cortex (Alger and Nicoll, 1979; Collingridge et al., 1984; Schwartz et al., 1985; Allan and Harris, 1986), pentobarbital and diazepam enhance the response to GABA. As expected, the GABA response was blocked by the GABAA receptor antagonists, picrotoxin and bicuculline methiodide. In addition, ethanol concentrations up to 50 mM did not affect the GABA-mediated changes in MEQ fluorescence in the hippocampus. Similarly, low concentrations of ethanol have been shown to have little or no effect on hippocampal GABAergic neurotransmission
133
using electrophysiologic methods (White et al., 1990; Proctor et al., 1992; Criswell et al., 1993). With the appropriate pharmacology established using this imaging technique, greater temporal resolution in the MEQ response was achieved by applying muscimol directly to the cell (i.e. via a puffer pipet; Fig. 4). Certain regional differences in GABAA responses were also detected by laser-scanning confocal microscopy that both support and contrast with findings obtained with other methods. For example, cerebellar neurons exhibited a greater response to GABA compared to hippocampal and neocortical neurons. Within the cerebellum, a greater relative change in [Cl − ]i occurred within Purkinje cells compared to the granule cells. Cerebellar Purkinje cells, which are under major inhibitory control, also exhibit large, spontaneous GABAA receptor-mediated currents detected with electrophysiologic techniques (Konnerth et al., 1990). In addition, Puia et al. (1994) have demonstrated that GABA-activated Cl − currents (measured in milliseconds) have different kinetics within Purkinje versus granule cells. However, our findings differ from the relative responses to GABA measured in synaptoneurosomes from the hippocampus, neocortex, and cerebellum (Allan and Harris, 1986; Luu et al., 1987). Thus, responses obtained in isolated membrane vesicles may not always be equivalent to those in living brain slices where circuitry is generally intact. The regional heterogeneity of GABAA receptor responses could arise for a variety of reasons, including: (a) diversity in the GABAA receptor subunit composition among different brain regions and neuronal types (Richards et al., 1987; Gutie´rrez et al., 1994; Inglefield et al., 1994; Fritschy and Mohler, 1995); (b) regional differences in the degree of phosphorylation, which modulates the GABAA receptor function (Stelzer et al., 1988; Heuschneider and Schwartz, 1989); or (c) regional differences in the anion/cation transporter or Cl − /HCO3-exchanger which would contribute to different Cl − gradients among neurons. Confocal imaging of multiple cell types within the living brain slice provides an opportunity to investigate these potential mechanisms underlying the regional heterogeneity of GABAA receptor responses. An advantage of both confocal and non-confocal imaging of MEQ fluorescence (compared to electrophysiology, for example) is that one can compare the GABAA responses in different parts of the neuron at the same time, without the need for multiple recording pipettes. In all neuronal types examined with UV-laserscanning confocal microscopy from post-natal day 12– 21, there was lower MEQ fluorescence in the dendrite relative to the somata. This could be due to: (1) the geometry of the thin dendrites relative to the soma; or (2) a Cl − gradient from relatively low levels within the soma to higher levels in the dendrite. In fact, Hara et al. (1992) have provided evidence for differences in Cl −
134
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135
extrusion mechanisms between the somata and apical dendrite of cultured hippocampal neurons to account for a spatial unevenness of Cl − -sensitive fluorescence. In another study, local application of GABA to hippocampal pyramidal cell dendrites increased extracellular Cl − , presumably due to a pre-existing outward Cl − gradient in the dendrites (Muller et al., 1989). Whether an altered Cl − gradient is the basis for depolarizing responses to GABA application in dendrites of CA1 pyramidal cells (Alger and Nicoll, 1979) is still debated (Staley et al., 1995). While fluorescence photometry has been used successfully to study GABAA receptor activity in synaptoneurosomes, cultured cells, and engineered cell lines (Engblom et al., 1991; Engblom and Akerman, 1991; Wong et al., 1992; Hara et al., 1992), this is the first study to assess the optical characteristics and pharmacology of GABAA receptor activity in living brain slices with confocal imaging. This technique has excellent potential for several applications. First, the spatial resolution of a confocal microscope combined with the temporal resolution of electrophysiology can permit the study of signaling actions at synapses on dendrites and soma simultaneously. Whereas the bath-loading technique limits the spatial resolution somewhat to the level of the somata (as well as proximal dendrites), direct application of MEQ (non-reduced form) into the cell and its processes via a ‘patch pipet’ would provide even better spatial resolution. This would be a particular advantage for optical imaging in adult slices where bath-loading has limited success. Second, in this study, desensitized responses to GABA were measured minutes after continuous bath application. With the use of local drug application to a single cell (or dendrite) via a pressure ejection micropipet, the time between application of drug or neurotransmitter and detection of the response can be as low as B1 s. Finally, the high spatial and temporal resolution ( 0.001 s) of photoactivation (‘uncaging’) may be ultimately the best method for mapping GABA responses in situ (Wang and Augustine, 1995). In conclusion, whatever method is used for drug application, UV-confocal microscopy and Cl − sensitive dyes may have an important impact on future studies of GABA neurotransmission in both developing and adult neurons.
Acknowledgements This work was supported by a Fellowship from the American Heart Association, North Carolina Affiliate, Inc. to J.R.I., and NIH grant NS 28791 to R.D.S. We wish to thank Drs George Augustine, Victor Nadler, and Elizabeth Finch for helpful discussions relating to the development of the technique.
References Alger BE, Nicoll RA. GABA-mediated biphasic inhibitory responses in hippocampus. Nature 1979;281:315 – 7. Allan AM, Harris RA. g-Aminobutyric acid agonists and antagonists alter chloride flux across brain membranes. Mol Pharmacol 1986;29:497 – 505. Biwersi J, Verkman AS. Cell-permeable fluorescent indicator for cytosolic chloride. Biochemistry 1991;30:7879 – 83. Collingridge GL, Gage PW, Robertson B. Inhibitory post-synaptic currents in rat hippocampal CA1 neurones. J Physiol 1984;356:551 – 64. Criswell HE, Simson PE, Duncan GE, McCown TJ, Herbert JS, Morrow AL, Breese GR. Molecular basis for regionally specific action of ethanol on g-aminobutyric acidA receptors: Generalization to other ligand-gated ion channels. J Pharmacol Exp Ther 1993;267:522 – 37. Dailey ME, Smith SJ. Spontaneous Ca2 + transients in developing hippocampal pyramidal cells. J Neurobiol 1994;25:243 –51. Engblom AC, Akerman KEO. Effect of ethanol on g-aminobutryic acid and glycine receptor-coupled Cl − fluxes in rat brain synaptoneurosomes. J Neurochem 1991;57:384 – 90. Engblom AC, Holopainen I, Akerman KEO. Ethanol-induced Cl − flux in rat cerebellar granule cells as measured by a fluorescent probe. Brain Res 1991;568:55 – 60. Foskett JK. [Ca2 + ]i modulation of Cl − content controls cell volume in single salivary acinar cells during fluid secretion. Am J Physiol 1990;259:C998 – C1004. Frenguelli BG, Potier B, Slater NT, Alford S, Collingridge GL. Metabotropic glutamate receptors and calcium signalling in dendrites of hippocampal CA1 neurones. Neuropharm 1993;32:1229– 37. Fritschy J-M, Mohler H. GABAA-receptor heterogeneity in the adult rat brain: Differential regional and cellular distribution of seven major subunits. J Comp Neurol 1995;359:154 – 94. Guthrie PB, Segal M, Kater SB. Independent regulation of calcium revealed by imaging dendritic spines. Nature 1991;354:76–80. Gutie´rrez A, Khan Z, de Blas AL. Immunocytochemical localization of g2 short and g2 long subunits of the GABAA receptor in the rat brain. J Neurosci 1994;14:7168 – 79. Hara M, Inoue M, Yasukura T, Ohnishi S, Mikami Y, Inagaki C. Uneven distribution of intracellular Cl − in rat hippocampal neurons. Neurosci Lett 1992;143:135 – 8. Harris RA, Allan AM. Functional coupling of g-aminobutyric acid receptors to chloride channels in brain membranes. Science 1985;228:1108 – 9. Hayashi H, Miyata H. Fluorescence imaging of intracellular Ca2 + . J Pharmacol Toxicol Methods 1994;31:1 – 10. Heuschneider G, Schwartz RD. cAMP and forskolin decrease gaminobutryic acid-gated chloride flux in rat brain synaptoneurosomes. Proc Natl Acad Sci 1989;86:2938 – 42. Imai H, Okuno T. Wu, J.Y. and Lee, T.J.-F. (1991) GABAergic innervation in cerebral blood vessels: An immunohistochemical demonstration of L-glutamic acid decarboxylase and GABA transaminase. J Cereb Blood Flow Metab 1991;11:129 –34. Inglefield JR, Schwartz-Bloom. Confocal imaging of intracellular chloride ([Cl − ]i) in acutely prepared slices. Measurement of GABAA receptor activity. Soc Neurosci Abstracts 1996;22:324.13. Inglefield JR, Sieghart W, Kellogg CK. Immunohistochemical and neurochemical evidence for GABAA receptor heterogeneity between the hypothalamus and cortex. J Chem Neuroanat 1994;7:243 – 52. Jaffe DB, Fisher SA, Brown TH. Confocal laser scanning microscopy reveals voltage-gated calcium signals within hippocampal dendritic spines. J Neurobiol 1994;25:220 – 33.
J.R. Inglefield, R.D. Schwartz-Bloom / Journal of Neuroscience Methods 75 (1997) 127–135 Konnerth A, Llano I, Armstrong CM. Synaptic currents in cerebellar Purkinje cells. Proc Natl Acad Sci 1990;87:2662–5. Krapf R, Berry CA, Verkman AS. Estimation of intracellular chloride activity in isolated perfused rabbit proximal convoluted tubules using a fluorescent indicator. Biophys J 1988;53:955 – 62. Krause DN, Wong E, Degener P, Roberts E. GABA receptors in bovine cerebral blood vessels: binding studies with tritiated muscimol. Brain Res 1980;185:51–7. Krnjevic K, Schwartz S. The action of g-aminobutyric acid on cortical neurones. Exp Brain Res 1967;3:320–36. Llano I, Dreessen J, Kano M, Konnerth A. Intradendritic release of calcium induced by glutamate in cerebellar purkinje cells. Neuron 1991;7:577 – 83. Luu MD, Morrow AL, Paul SM, Schwartz RD. Characterization of GABAA receptor-mediated 36chloride uptake in rat brain synaptoneurosomes. Life Sci 1987;41:1277–87. MacVinish LJ, Reancharoen T, Cuthbert AW. Kinin-induced chloride permeability changes in colony 29 epithelia estimated from 125I − efflux and MEQ fluorescence. Br J Pharmacol 1993;108:469 – 78. Mason WT. Fluorescent and Luminescent Probes for Biological Activity. New York: Academic Press, 1993:433. Muller W, Misgeld U, Lux HD. g-Aminobutryic acid-induced ion movements in the guinea pig hippocampal slice. Brain Res 1989;484:184 – 91. Porter JT, McCarthy KD. GFAP-positive hippocampal astrocytes in situ respond to glutamatergic neuroligands with increases in [Ca2 + ]i. Glia 1995;13:101–12. Proctor WR, Soldo BL, Allan AM, Dunwiddie TV. Ethanol enhances synaptically evoked GABAA receptor-mediated responses in cerebral cortical neurons in rat brain slices. Brain Res 1992;595:220 – 7. Puia G, Costa E, Vicini S. Functional diversity of GABA-activated Cl − currents in Purkinje versus granule neurons in rat cerebellar slices. Neuron 1994;12:117–26.
.
.
135
Regehr WG, Connor JA, Tank DW. Optical imaging of calcium accumulation in hippocampal pyramidal cells during synaptic activation. Nature 1989;341:533 – 6. Richards JG, Schoch P, Ha¨ring P, Takacs B, Mo¨hler H. Resolving GABAA/benzodiazepine receptors: Cellular and subcellular localization in the CNS with monoclonal antibodies. J Neurosci 1987;7:1866 – 86. Schwartz RD, Jackson JA, Wiegert D, Skolnick P, Paul SM. Characterization of barbiturate-stimulated chloride efflux from rat brain synaptoneurosomes. J Neurosci 1985;5:2963 – 70. Schwartz RD, Yu X. Optical imaging of intracellular chloride in living brain slices. J Neurosci Methods 1995;62:185 – 92. Staley KJ, Soldo BL, Proctor WR. Ionic mechanisms of neuronal excitation by inhibitory GABAA receptors. Science 1995;269:977– 81. Stelzer A, Kay AR, Wong RKS. GABAA-receptor function in hippocampal cells is maintained by phosphorylation factors. Science 1988;241:339 – 41. Tsien RY. Fluorescence measurement and photochemical manipulation of cytosolic free calcium. Trends Neurosci 1988;11:419–24. Verkman AS. Development and biological applications of chloridesensitive fluorescent indicators. Am J Physiol 1990;259:C375–88. Wang SS-H, Augustine GJ. Confocal imaging and local photolysis of caged compounds: Dual probes of synaptic function. Neuron 1995;15:755 – 60. White G, Lovinger D, Weight F. Ethanol inhibits NMDA-activated current but does not alter GABA-activated current in an isolated adult mammalian neuron. Brain Res 1990;507:332 – 6. Wong G, Sei Y, Skolnick P. Stable expression of type I g-aminobutyric acidA/benzodiazepine receptors in a transfected cell line. Mol Pharmacol 1992;42:996 – 1003. Yuste R, Katz LC. Control of postsynaptic Ca2 + influx in developing neocortex by excitatory and inhibitory neurotransmitters. Neuron 1996;6:333 – 44.