INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Strychnine-sensitive glycine-gated chloride channels (GlyRs) are the major provider of neuronal inhibition in spinal cord and brain stem (Aprison 1990). However, in forebrain regions of mature CNS, such as hippocampus, a role for GlyRs in modulating neuronal excitability has been largely ignored. This is due in part to numerous reports from many brain regions demonstrating that fast synaptic inhibition is abolished by GABA_A receptor (GABA_AR) antagonists (Mody et al. 1994), implicating GABA_ARs as the only ligand-gated ion channel responsible for mediating fast neuronal inhibition in brain, while in brain stem and spinal cord fast inhibition is accomplished by both GlyRs and GABA_ARs working in conjunction (Chery and de Koninck 1999; Jonas et al. 1998; O'Brien and Berger 1999). Furthermore, a current perception exists that functional GlyR expression ceases following early postnatal stages of development that likely stems from too few studies demonstrating their presence. In fact it has been reported that functional GlyRs in hippocampus are not expressed beyond the second postnatal week (Ito and Cherubini 1991; but see Ye et al. 1999). However, in several animal models of epilepsy, exogenous application of glycine can depress seizure activity in hippocampus of adult rats (Cherubini et al. 1981; Seiler and Sarhan 1984), suggesting the interesting possibility that GlyRs may mediate this inhibitory influence of glycine and implying that GlyRs are expressed in mature hippocampus, challenging current perceptions.

The relative lack of attention given to GlyRs in hippocampus by investigators interested in neuronal inhibition is surprising in light of evidence that the machinery necessary for GlyR-mediated inhibition is documented to exist in hippocampus. For instance, synaptoneurosomes obtained from adult rat hippocampus contain glycine in concentrations similar to GABA (36 nmol GABA/mg protein vs. 42 nmol glycine/mg protein) and release of both inhibitory amino acids using Ca²⁺-dependent and -independent mechanisms, suggesting the possibility of both vesicular and transporter-mediated release (Burger et al. 1991; Engblom et al. 1996). Additionally, taurine, a known GlyR agonist in other systems (Flint et al. 1998; Horikoshi et al. 1988; Hussy et al. 1997) that is present in high concentrations in hippocampus (del Rio et al. 1987; Saransaari and Oja 1994, 1997), can protect against excitotoxic cell death and depress pyramidal cell excitability (French et al. 1986; Taber et al. 1986). Presumably these effects of taurine in hippocampus are mediated via activation of GlyRs; however, this mechanism has never been tested. In situ hybridization and immunohistochemical studies show that expression of GlyR subunits in hippocampus is developmentally regulated with expression continuing throughout adulthood, although at decreased levels compared with early postnatal stages (Becker et al. 1993; Malosio et al. 1991; Sato et al. 1992). Powerful glycine transporters believed to be responsible for terminating glycinergic synaptic transmission in other regions of the CNS are present in hippocampus, with the GlyT1 subtype expressed by astrocytes and GlyT2 selectively expressed by neurons (Jursky and Nelson 1995; Zafra et al. 1995a,b, 1997). Glycine transporters and GlyRs are colocalized in hippocampus (Jursky and Nelson 1995), suggesting the possibility of GlyR-mediated synaptic transmission. In fact, strychnine-sensitive inhibitory postsynaptic currents (IPSCs) have been recorded from immature hippocampal cultures that are believed to be mediated by GlyRs (Fatima-Shad and Barry 1995). In addition, glycine transporters may also function to release glycine onto nearby postsynaptic GlyRs. Transporter-mediated glycine release has been shown recently from cholinergic presynaptic terminals in the chick ciliary ganglion after depolarization (Tsen et al. 2000). Electrophysiological studies in dissociated hippocampal cultures obtained from newborn rats demonstrate the presence strychnine-sensitive glycine-mediated currents at immature developmental stages (Fatima-Shad and Barry 1992, 1995; Krishtal et al. 1988; Shirasaki et al. 1991; Ye et al. 1999; Yoon et al. 1998). Whether or not functional GlyRs are expressed by hippocampal neurons during later stages of development is not entirely clear because a study in hippocampal slices reports a loss of glycine responses after postnatal day 13 (Ito and Cherubini 1991) while another study using dissociated CA1 pyramidal cells from postnatal day 24-30 rats observes strychnine-sensitive glycine currents (Ye et al. 1999). Although the functional studies performed thus far are informative and provide evidence that GlyRs are expressed in hippocampus, albeit during immature developmental stages, characterization of these receptors lags behind that of other ligand-gated ion channels. Thus little is known regarding their pharmacologically properties, the extent to which these receptors are expressed in mature hippocampus, or whether both major neurons types, the pyramidal cells and GABAergic interneurons, express GlyRs. Expression of GlyRs by both cell types will imply that these receptors have a fundamental role in modulating hippocampal excitability.

This study was undertaken to confirm that GlyRs are functionally expressed in mature hippocampus by both pyramidal cells and stratum radiatum interneurons located in the CA1 region and to characterize the pharmacological properties of these receptors using hippocampal slices obtained from 3- to 4-wk-old rats. We found that glycine and taurine activate GlyRs and that both CA1 pyramidal cells and interneurons express more than one GlyR subtype based on the sensitivity of glycine-induced currents (I_gly) to picrotoxin. In addition, we found that hippocampal GlyRs are bidirectionally modulated by zinc. Interestingly, low micromolar zinc concentrations that potentiate I_gly depress N-methyl-D-aspartate (NMDA) receptor-mediated currents (I_NMDA), indicating that low zinc concentrations facilitate neuronal inhibition. Our findings suggest that not only GABA_ARs but also GlyRs are important providers of neuronal inhibition in hippocampus.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

GlyR-like immunohistochemistry

Three- to 4-wk-old Sprague-Dawley rats (n = 4) were deeply anesthetized with pentobarbital and then perfused with 0.9% NaCl followed by 4% paraformaldehyde. Brains were removed and further postfixed with 4% paraformaldehyde at room temperature. Brains were cryoprotected in a 30% sucrose solution at 4°C overnight. Frozen sections (40 µM) were cut on a freezing microtome (Zeiss), rinsed with PBS, and preincubated with 10% normal goat serum (NGS). Sections were incubated at 4°C for 48-72 h with the primary monoclonal anti-GlyR antibody, mAb 4a (1:25, from Connex, Martinsried, Germany), which recognizes the 1, 2, and  subunits of the GlyR (Pfeiffer et al. 1984). Following the incubation period, the sections were washed and incubated with an Alexa 488 conjugated secondary antibody at room temperature. Omission of the primary antibody served as a negative control to ensure the specificity of labeling. The specificity of this antibody has been thoroughly characterized to confirm its selectively for GlyR subunits (Pfeiffer et al. 1984) and has been extensively used in anatomical localization studies (e.g., Dumoulin et al. 2001; Tsen et al. 2000). Stained cells were observed with a confocal laser-scanning microscope equipped with appropriate filters.

Electrophysiology

Hippocampal slices were prepared from 3- to 4-wk-old Sprague-Dawley rats as previously described (McMahon and Kauer 1997b) and maintained in a submersion holding chamber at room temperature. Artificial cerebrospinal fluid (ACSF) was used for slice preparation and recording and contained (in mM) 119 NaCl, 2.5 KCl, 2.5 CaCl₂, 1.3 MgSO₄, 1 NaH₂PO₄, 26 NaHCO₃, 10 glucose, and 1 kynurenic acid, and was saturated with 95% O₂-5% CO₂ (pH: 7.4). For experiments, slices were placed in a submersion recording chamber and continually perfused with ACSF at 2-3 ml/min at 28-30°C. Whole cell recordings of visually identified CA1 pyramidal cells and GABAergic interneurons in s. radiatum were obtained using an Olympus BX50WI fixed-stage microscope with IR-DIC optics. Patch electrodes had resistances between 4 and 6 M and were filled with (in mM) 100 CsCl or 100 Cs-gluconate, 0.6 EGTA, 5 MgCl₂, 2 ATP-Na₂, 0.3 GTP-Na, and 40 HEPES, pH: 7.2, 260-270 mOsm. In some recordings, 5 mM QX 314 was added to the internal solution to block voltage-dependent Na⁺ channels and to enhance space-clamp. Cells were held at 70 mV, except where noted.

An Axoclamp 2A amplifier was used to amplify current signals, and the output was continuously monitored on an oscilloscope and Gould chart recorder. Data were filtered at 3 kHz, stored on tape using a Vetter PCM data recorder, and analyzed off-line. Cell input resistance and series resistance (18 M) were continually monitored throughout the recording using software written in Labview and kindly provided by Dr. Richard Mooney (Duke University). Experiments were terminated when these values increased by 20%. Recorded neurons were identified as pyramidal cells or interneurons by their electrical properties (e.g., input resistance, firing pattern). In addition, neurons were filled with 0.4% biocytin and processed to allow for post hoc neuronal identification (McMahon and Kauer 1997a,b; McMahon et al. 1998). Biocytin filled cells were viewed at the light level to confirm cell identity as pyramidal cells versus interneurons and images were captured with digital microscopy. Representative images of a recorded pyramidal cell and interneuron are shown in Fig. 1B.

View larger version (72K): [in this window] [in a new window]   Fig. 1. Strychnine-sensitive glycine-coated chloride channels (GlyRs) are expressed by CA1 pyramidal cells and stratum radiatum interneurons in rat hippocampal slices. A, left: confocal microscopy images of GlyR immunoreactivity of CA1 pyramidal cells and interneurons from a 24-day-old rat. Omission of primary antibody (control) shows an absence of staining. A, right: GlyR immunoreactivity intensely labels somas and dendrites of pyramidal cells and interneurons located in s. oriens, pyramidale, and radiatum. B: Igly induced by a 3-s pressure application of 300 µM glycine () is completely unaffected by 10 µM bicuculline but reversibly blocked with 1 µM strychnine. Cells were held at 70 mV and recorded with a CsCl pipette solution. C: digital images of biocytin-filled typically recorded pyramidal cell and s. radiatum interneuron.

Drug delivery

All chemicals used in this study were purchased from Sigma-Aldrich (St. Louis, MO). Agonists and antagonists were prepared as stock solutions and diluted to appropriate concentrations in the recording solution. Glycine (or taurine) was most often applied to recorded cells via a picrospritzer with the drug-containing pipette placed within 100 µM of the recorded cell. Agonists were also applied by bath perfusion of a fixed concentration (2-3 ml/min) or by a drug pipette (100 µM from the cell) connected to a valve system which permits switching between drug solutions (4-5 ml/min) for obtaining dose-response measurements. This latter method permits faster time to peak and recovery of responses than the standard bath perfusion method (bath exchange within 30 s rather than 1-2 min so peak response occurs within 20-30 s), although the faster bath exchange did not alter the amount of desensitization as measured by the peak versus steady-state current amplitude. The glycine concentration (300 µM) used in this study was chosen because it elicits robust responses facilitating characterization of the currents and is near the glycine EC₅₀ concentration for these receptors as shown in Fig. 2A. Note that in recordings in brain slices, the exact concentration of the agonist that the receptors actually "see" is not known for sure because there are powerful re-uptake mechanisms in both neurons and glial cells that cannot be accounted for. Thus the EC₅₀ concentration we observe in slices is higher than that reported for GlyRs recorded in dissociated hippocampal cultures (40-72 vs. 270 µM) (Shirasaki et al. 1991; Ye et al. 1999; Yoon et al. 1998). This discrepancy in agonist EC₅₀ concentration between culture and slice preparations have been previously observed (Kaneda et al. 1995). Additionally, the modest speed in which agonists can be applied to recorded neurons in slices and the varying distances these cells are from the slice surface means that receptor desensitization will likely take place during the rising phase of the response that will affect the amplitude of the current; however, the extent of the desensitization is unknown (Kaneda et al. 1995).

View larger version (25K): [in this window] [in a new window]   Fig. 2. Glycine dose-response relationship and dose-inhibition relationship mediated by increasing strychnine concentrations. A, left: representative traces of Igly recorded from an interneuron at increasing glycine concentrations. Glycine was applied via a drug pipette (see METHODS) for a duration of 1 min and at increasing concentrations as noted. Glycine was reapplied at 1-min intervals. A, right: plot compares glycine dose-response curve recorded from pyramidal cells (n = 6) and interneurons (n = 6). Currents were normalized to the peak amplitude of the response at 1 mM glycine. Each point is the mean ± SE. Dose-response data were fit with the Hill equation (see METHODS). B, left: representative traces of Igly (300 µM glycine, 3 s via a picrospritzer) recorded from an interneuron in the presence of increasing concentrations of strychnine as noted. Currents partially recover following a 45-min washout of strychnine. B, right: summary plot showing the effect of increasing strychnine concentrations on Igly elicited with a 3-s application of 300 µM glycine. Responses were normalized to the peak of the Igly generated in the absence of strychnine. Each point represents the mean ± SE (pyramidal cells, n = 3; and interneurons, n = 5). The error bars for some data points are within the size of the symbol. Data were fit to the inhibition equation (see METHODS). Cells were held at 70 mV and recorded with a CsCl pipette solution.

Data analysis

Data were normalized to the maximum current for each cell. Reported mean and standard error were determined from normalized data. Glycine and taurine dose-response relationships were fit to the Hill equation I = I_max/1 + (EC₅₀/[agonist]ⁿ), where I is the peak current for a given agonist concentration, I_max is the current at the maximal agonist concentration, EC₅₀ is the concentration of agonist (glycine or taurine) required for a half-maximal response, and n is the Hill coefficient (slope factor). Strychnine or picrotoxin inhibition plots were fit using the equation I = I_max/1 + ([strychnine] or [picrotoxin]/IC₅₀)ⁿ, where I is the peak I_gly for a given strychnine or picrotoxin concentration, I_max is the current amplitude in the absence of strychnine or picrotoxin, IC₅₀ is the concentration of strychnine or picrotoxin required to block I_gly by 50% and n is the Hill coefficient (slope factor). In these experiments, one application of glycine was sufficient to obtain a steady-state block for each antagonist concentration after a 10-min application of the drug. However, to be certain, we routinely obtained three glycine responses at each concentration before increasing to the next concentration. Analysis of dose-response relationships and strychnine or picrotoxin inhibition were performed using Origin 5.0 or IGOR Pro 6.0 software. Data were expressed as means ± SE, and one-way ANOVA and unpaired T-tests were used to assess significant differences between groups. Differences were considered significant at P  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

GlyRs are functionally expressed in the CA1 region of rat hippocampus

To demonstrate that GlyRs are expressed by rat hippocampal neurons beyond the early stages of postnatal development, we obtained whole cell recordings of CA1 pyramidal cells and s. radiatum interneurons in hippocampal slices prepared from 3- to 4-wk-old animals and used GlyR immunohistochemistry to anatomically localize these receptors. In immunohistochemical staining experiments using the monoclonal anti-GlyR antibody (mAb 4a), which recognizes 1, 2, and  subunits of GlyRs (Pfeiffer et al. 1984), we consistently observed specific labeling of cell bodies and dendrites of both pyramidal cells and GABAergic interneurons located throughout all CA1 cell layers (Fig. 1A). An absence of staining was observed when the primary antibody was omitted from the reaction (Fig. 1A). These findings indicate that GlyRs are strongly and widely expressed by both pyramidal cells and interneurons in hippocampus. In Fig. 1B, we show that short pulses (3 s) of glycine (300 µM) pressure-applied to a recorded pyramidal cell and interneuron elicited inward currents (I_gly; E_Cl = 0 mV) that developed over a few seconds and decayed to baseline within 10 s following cessation of glycine application. Responses were observed in 100% of recorded pyramidal cells (n = 30) and interneurons (n = 40) and ranged in amplitude from 120 to 420 pA (240 ± 46 pA) in pyramidal cells and 160 to 1,600 pA (782 ± 142 pA) in interneurons. Bath application of the GABA_A antagonist bicuculline (10 µM; n = 8) had no effect on I_gly. However, strychnine (1 µM) reversibly depressed or abolished I_gly (n = 10), reaching maximal block within 10 min. A partial recovery was obtained following a 30- to 45-min washout of the antagonist. These data indicate that functional strychnine-sensitive GlyRs are indeed expressed by pyramidal cells and interneurons beyond the second postnatal week of development, challenging a previous report (Ito and Cherubini 1991).

To further characterize the sensitivity of GlyRs to glycine and strychnine, we recorded dose-response curves from both pyramidal cells and interneurons (Fig. 2). Normalized dose-response curves demonstrate that GlyRs expressed by pyramidal cells and interneurons respond similarly to glycine (EC₅₀ = 270 µM; Hill coefficient n = 1.4-1.9; Fig. 2A, right), although the mean current amplitude recorded at any given glycine concentration is 30-40% larger (38 ± 4% at 300 µM, approximate EC₅₀) in interneurons compared with pyramidal cells (data not shown). This finding together with a larger GlyR-induced decrease in cell input conductance in interneurons (on average a decrease of 14% in pyramidal cells vs. 30% in interneurons) suggests a higher density of somatic GlyRs expressed by these cells. Furthermore, we observed that I_gly recorded from both pyramidal cells and interneurons only partially desensitize during the 1-min exposure to glycine (concentrations 1 mM, Fig. 2A, left) and reach a steady-state level of conductance, 59 ± 10% of peak for pyramidal cells (n = 4) and 47 ± 19% of peak for interneurons (n = 4; values not significantly different; calculated at 300 µM glycine, approximate EC₅₀). Even with continuous glycine exposure 10 min, I_gly did not completely desensitize back to baseline, but continued at the steady-state conductance level reached at 1 min of exposure (data not shown). Ligand-gated ion channels are greatly desensitized by prolonged agonist exposure (Jones and Westbrook 1996); however, these findings suggest that during prolonged glycine exposure, GlyRs will still conduct current. This is an important issue because in the inherited metabolic disease, nonketotic hyperglycinemia, glycine levels in CSF are persistently elevated 8- to 28-fold (2-10 to 83-280 µMol/l) over normal levels (Steiner et al. 1996). Additionally, the CSF glycine concentration has also been reported to be elevated following seizures and hypoxic episodes (Andine et al. 1991; Castillo et al. 1996; Sherwin 1999). Our data suggest therefore that under these pathological conditions, GlyR-mediated inhibition will remain functional, albeit at a reduced level.

We next examined the sensitivity of hippocampal GlyRs to strychnine blockade (Fig. 2B). In dose-inhibition experiments, we found that GlyRs expressed by both cell types are highly sensitive to strychnine blockade with an IC₅₀ of 0.02 µM for GlyRs expressed by pyramidal cells and 0.04 µM for GlyRs expressed by interneurons (values not significantly different). The Hill coefficient for both cell types is 0.77. The high sensitivity of the glycine-mediated currents to strychnine provides strong evidence that these currents are mediated by GlyRs.

Current-voltage relationship of glycine-evoked currents

Next, we sought to confirm that I_gly is carried by chloride ions (Fig. 3). The reversal potential for I_gly recorded from pyramidal cells (n = 4) and interneurons (n = 11) was 2.1 ± 0.7 and 0.7 ± 0.4 mV, respectively, when cells were recorded with 110 mM internal chloride (calculated Nernst potential for 110 mM [Cl]_i is 3.5 mV). In cells recorded with 10 mM [Cl]_i, the I_gly reversal potential was 45.1 ± 1.7 mV for pyramidal cells (n = 4) and 41.1 ± 3.5 mV for interneurons (n = 7; calculated Nernst potential for 10 mM internal chloride is 63.9 mV). The ~40-45 mV shift observed in the I_gly reversal potential with a 100 mM increase in the internal chloride concentration is consistent with I_gly being carried by chloride ions and is near the predicted change in the reversal potential. The inexact shift between the observed and predicted values is likely do either to incomplete dialysis of the pipette solution with the cytosol of the recorded neuron or a result of varying activity of efficient chloride co-transporters having differing activities depending on internal chloride concentration (DeFazio et al. 2000).

View larger version (18K): [in this window] [in a new window]   Fig. 3. Glycine currents are carried by chloride ions. A: Igly (300 µM glycine, 3 s) recorded from an interneuron reversed at 0 mV when recorded with a CsCl-based internal solution (110 mM Cl) and at 40 mV in another interneuron when recorded with a Cs-gluconate-based internal solution (10 mM Cl). B: current-voltage relationship of Igly recorded from pyramidal cells (n = 8) and interneurons (n = 18). Currents were normalized to the peak amplitude of the response elicited at 70 mV. Each point represents the mean ± SE. The error bars for some data points are within the size of the symbol.

Taurine is an agonist at GlyRs in hippocampus

The inhibitory amino acid, taurine, is an agonist at glycine receptors in other systems (Flint et al. 1998; Hussy et al. 1997; Schmieden et al. 1992). Taurine has been shown to depress pyramidal cell excitability (Taber et al. 1986) and have an antiepileptic activity (Cherubini et al. 1981; Seiler and Sarhan 1984), although the mechanism underlying these effects is unknown. Thus we are interested to determine whether taurine also activates native GlyRs expressed in hippocampal neurons because this could be the mechanism by which taurine mediates an inhibitory effect. We obtained whole cell patch-clamp recordings from interneurons (n = 5) and observed large inward currents (I_taurine; 400-600 pA) following application of taurine (1 mM, 3 s). One millimolar taurine was used in these experiments because this concentration was close to the taurine EC₅₀ (3.5 mM) as seen in Fig. 4, C and D. To ensure that I_taurine is due to activation of GlyRs and not to activation of GABA_ARs, taurine was applied in the presence of 10 µM bicuculline. Bicuculline had no effect on I_taurine while strychnine (1 µM) reversibly blocked this current (Fig. 4A).

View larger version (23K): [in this window] [in a new window]   Fig. 4. Taurine is an agonist at GlyRs. A: taurine-mediated currents (Itaurine) induced by a 3-s pressure application of 1 mM taurine () is completely unaffected by 10 µM bicuculline but reversibly blocked with 1 µM strychnine (n = 4). B: comparison of Igly (300 µM glycine bath applied, 1 min) and Itaurine (300 µM taurine bath applied, 1 min) recorded from the same s. radiatum interneuron (n = 4). C: representative traces of Itaurine recorded from an interneuron with increasing taurine concentrations. Taurine was bath applied for 1 min at the concentrations as noted and at an interval of 1 min. D: plot compares the dose-response relationship of glycine (n = 6) and taurine-mediated currents (n = 6). The taurine dose-response curve is shifted to the right of the glycine curve indicating that taurine is a less potent agonist at GlyRs in hippocampus. Currents are normalized to the peak amplitude of the response at the highest concentration of glycine (1 mM) and taurine (30 mM). Each point is the mean ± SE. In some cases, the error bars are with the symbol. Dose-response data were fit with the Hill equation (see METHODS). In all experiments, cells were held at Vh = 70 mV. In A, C, and D, cells were recorded with a CsCl pipette solution; in B, cells were recorded with a Cs-gluconate pipette solution.

To compare the activation of GlyRs by glycine and taurine, equimolar concentrations of glycine and taurine (300 µM) were bath applied to the same recorded cell (n = 6; Fig. 4B). We found that glycine induced larger inward currents (800-2,000 pA) than taurine (60-110 pA) at the same concentration. Additionally, in dose-response experiments, the taurine dose-response curve was shifted to the right of that for glycine with an EC₅₀ of taurine (3.54 mM) 10-fold higher than the EC₅₀ for glycine (0.27 mM) (Fig. 4D). These data demonstrate that taurine is an agonist at GlyRs but has a lower affinity for these receptors than glycine.

Hippocampal GlyRs are blocked by picrotoxin

In situ hybridization studies indicate that mRNA encoding 2, 3, and  subunits is present in hippocampus (Malosio et al. 1991), suggesting the possibility that hippocampal neurons may express more than one GlyR subtype. Furthermore, the possibility exists that hippocampal neurons may express both  homomeric receptors, the immature extrasynaptic GlyR subtype and heteromeric GlyRs, the mature form of the receptor that is synaptically located (Becker et al. 1988; Kungel and Friauf 1997; Malosio et al. 1991; Takahashi et al. 1992). To pursue whether pyramidal cells and interneurons in acute slices at this stage of maturity express  homomeric and/or heteromeric GlyRs, we tested whether the GABA_AR antagonist, picrotoxin (1-1,000 µM) could suppress I_gly, taking advantage of the knowledge that the presence of the  subunit in functional GlyRs renders these receptors insensitive to blockade by low micromolar concentrations of the antagonist (Pribilla et al. 1992). This antagonist was shown in a previous study of heterologously expressed GlyRs (Pribilla et al. 1992) to inhibit  homomeric receptors but not heteromeric GlyRs at 10 µM picrotoxinin (equivalent to 20 µM picrotoxin). Additionally, this strategy has been used to demonstrate receptor heterogeneity in Mauthner cells (Legendre 1997) and cultured neonatal hippocampal neurons (Yoon et al. 1998) where the authors found that low concentrations of picrotoxin (50 µM or its equivalent) selectively blocked homomeric GlyRs but not heteromeric receptors. The presumed heteromeric receptors were not completely blocked at a concentration of 1 mM. As shown in Fig. 5, we observed that at picrotoxin concentrations 100 µM, I_gly was partially depressed (50% of peak) in both cell types (Fig. 5A), consistent with the presence of  homomeric GlyRs. However, even at 1 mM picrotoxin, complete suppression of I_gly was not achieved in pyramidal cells (n = 10) or interneurons (n = 11), with in some cells, >20% of the peak amplitude still remaining. This finding is consistent with the expression of heteromeric receptors by these cells as well. Figure 5B shows a plot of the relationship between the peak of I_gly (normalized to the maximum) and the picrotoxin concentration. Inhibition curves were fit to these data and yielded a picrotoxin IC₅₀ of 37.1 µM for pyramidal cells and 83.6 µM for interneurons (concentrations not significantly different, P  0.05). In these experiments, we observed a large variability in the amount of block at any given picrotoxin concentration between neurons in both cells groups, as indicated by the error bars in the dose-inhibition plot (Fig. 5B). Additionally, the blockade of I_gly elicited by picrotoxin was not use dependent and a steady-state block could be observed following only a 10-min exposure of the slice to picrotoxin. Thus the partial blockade of I_gly at low concentrations of picrotoxin and the lack of a complete block at 1 mM picrotoxin is implies that both  homomeric and heteromeric GlyRs are expressed by single neurons in hippocampus.

View larger version (23K): [in this window] [in a new window]   Fig. 5. Picrotoxin partially blocks Igly. A: Igly (300 µM glycine, 3 s) recorded from a pyramidal cell and interneuron in the presence of increasing concentrations of picrotoxin as noted. Currents partially recover following a 30-min washout of the antagonist. B: summary plot showing the effect of increasing picrotoxin concentrations on Igly elicited with a 3-s application of 300 µM glycine. Responses were normalized to the peak of the Igly generated in the absence of picrotoxin. Each point represents the mean ± SE (pyramidal cells, n = 10; and interneurons, n = 11). The error bars for some data points are within the size of the symbol. Data were fit to the inhibition equation (see METHODS). Cells were held at 70 mV and recorded with a CsCl pipette solution.

Bidirectional modulation of hippocampal GlyRs by zinc

The divalent cation zinc modulates many ligand-gated ion channels including GlyRs. Several studies show that GlyRs in forebrain regions are modulated by zinc. For example, zinc depresses I_gly in cerebellar neurons (Virginio and Cherubini 1997) and septal cholinergic neurons (Kumamoto and Murata 1996), but bidirectionally modulates (facilitates as well as depresses) I_gly, depending on the zinc concentration, in spinal cord neurons (Bloomenthal et al. 1994; Laube et al. 1995) and retinal ganglion cells (Han and Wu 1999). We sought to determine whether native GlyRs in hippocampus are modulated by zinc and to investigate whether the modulation of I_gly by zinc is similar or different to the well-characterized zinc modulation of I_NMDA. Short pulses (3 s) of 300 µM glycine (n = 20) or 300 µM NMDA (n = 6) were applied to recorded interneurons in the absence (control) and presence of increasing concentrations of zinc (1 µM to 1 mM) included in the extracellular solution. As shown in Fig. 6, the peak amplitude of I_gly was increased 14 ± 2% (n = 15) and 27 ± 2% (n = 10) in the presence of 1 and 10 µM zinc, respectively. However, at 100 µM and 1 mM zinc, I_gly was depressed 17 ± 4% (n = 11) and 52 ± 8% (n = 4), respectively, compared with control. In contrast, I_NMDA was depressed by all zinc concentrations applied (1-100 µM). Moreover, the effects of zinc on I_gly and I_NMDA at all zinc concentrations were significantly different (P < 0.01) as shown in Fig. 6B. Our findings show that zinc modulation of I_gly is bidirectional, with low concentrations of zinc potentiating and high concentrations depressing I_gly. Additionally, we show that zinc concentrations that potentiate I_gly depress I_NMDA.

View larger version (18K): [in this window] [in a new window]   Fig. 6. Zinc differentially modulates Igly and INMDA. A: Igly (300 µM glycine, 3 s, Vh = 70 mV) and INMDA (300 µM NMDA, 3 s, Vh = 40 mV) were recorded from s. radiatum interneurons in the absence and presence of increasing zinc concentrations. Each set of traces displayed at the various zinc concentrations were obtained from different neurons. B: bar chart shows the relationship between percentage change of Igly (n = 14) or INMDA (n = 6) vs. zinc concentration. Data are plotted as means ± SE. Cells were recorded with a CsCl pipette solution.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data convincingly show that both major neuronal cell types in hippocampus, the excitatory pyramidal cells and inhibitory GABAergic interneurons, express GlyRs. Moreover, we find that GlyRs do not completely desensitize during prolonged glycine application. This finding indicates that when glycine levels increase in CSF for extended periods of time, as occurs following seizures and ischemia, activated GlyRs will continue to conduct current (Andine et al. 1991; Castillo et al. 1996; Sherwin 1999). The data presented in this study suggest the interesting hypothesis that activation of GlyRs, in conjunction with GABA_ARs, may provide an underappreciated, fundamental inhibitory mechanism in hippocampus that will modulate neuronal excitability. In fact, activation of these receptors may be the mechanism underlying the ability of glycine and taurine to depress seizure activity in some animal models of epilepsy (Cherubini et al. 1981; Seiler and Sarhan 1984).

Is taurine a ligand acting at hippocampal GlyRs?

Taurine is an abundant amino acid in hippocampus (del Rio et al. 1987; Saransaari and Oja 1994, 1997) and interestingly, taurine is released in high levels during conditions that elicit excitotoxic cell death (Saransaari and Oja 1994, 1997; Schurr et al. 1987). Taurine has been reported to counteract excitotoxicity (Saransaari and Oja 1998a,b) and has been shown to hyperpolarize hippocampal neurons (Taber et al. 1986). However, the mechanisms underlying these actions of taurine are presently unknown. Several studies have shown that taurine activates heterologously expressed GlyRs and native GlyRs expressed in hypothalamus and cerebral cortex (Flint et al. 1998; Hussy et al. 1997; Schmieden et al. 1992). In this study, we have shown that taurine can activate hippocampal GlyRs. The action of taurine at inhibitory GlyRs could be the mechanism responsible for taurine's ability to protect against excitotoxic cell death because GlyR activation depresses neuronal excitability (Chattipakorn and McMahon 2000; Taber et al. 1986). This mechanism would counteract the pathological excitation elicited by glutamate at NMDA receptors.

Studies of GlyRs in heterologous expression systems has demonstrated that 1-containing GlyRs are more efficiently gated by taurine than 2-containing GlyRs (Schmieden et al. 1992). Our data show that taurine is less potent than glycine at activating GlyRs in hippocampal neurons, suggesting that GlyRs in mature hippocampus possibly contain 2 rather than 1 subunits. This interpretation is supported by an in situ hybridization study suggesting that the 2-to-1 subunit switch during development does not occur in hippocampus as it does in spinal cord and brain stem (Malosio et al. 1991).

Picrotoxin antagonism suggests multiple GlyR subtypes

GlyRs in mammalian CNS are formed by a combination of five membrane spanning protein subunits, and as with other ligand-gated ion channels, the physiological and pharmacological properties of GlyRs is dependent on the subunit combination. Native GlyRs in spinal cord and brain stem are either  homomeric GlyRs (immature, extrasynaptic subtype) or heteromeric GlyRs (mature, synaptic subtype) (Becker et al. 1988; Kungel and Friauf 1997; Malosio et al. 1991; Takahashi et al. 1992). Fortunately, the GABA_AR antagonist picrotoxin is a useful tool in differentiating between homomeric and heteromeric GlyRs because the presence of the  subunit in heteromeric receptors confers picrotoxin resistance (Pribilla et al. 1992). In situ hybridization studies indicate that hippocampal neurons primarily contain mRNA encoding 2, some 3 (and not 1) and  subunits (Malosio et al. 1991). Therefore we used picrotoxin to test the possibility that some fraction of the receptors expressed by hippocampal neurons is 2 homomers and that both pyramidal cells and interneurons express these receptors. Low concentrations of picrotoxin (<100 µM) partially blocked the I_gly recorded from pyramidal cells and interneurons, suggesting that a subpopulation of the GlyRs expressed by both cell types are likely to be 2 homomeric receptors. The unblocked current at the lower picrotoxin concentrations may be the result of activation of another GlyR subtype, likely heteromeric GlyRs, which have a much lower sensitivity to picrotoxin (Pribilla et al. 1992). This interpretation is consistent with a recently published report in dissociated hippocampal cultures (Yoon et al. 1998). The finding of possible heteromeric GlyRs in hippocampus raises the question that some GlyRs may be synaptically located because of evidence showing that the  subunit of GlyRs is required for receptor clustering (Kirsch et al. 1993; Meyer et al. 1995).

Zinc modulation of hippocampal GlyRs

Zinc is abundant in forebrain regions, including hippocampus, where it is especially concentrated in mossy fibers (Assaf and Chung 1984; Howell et al. 1984). Zinc is contained in vesicles with glutamate, is co-released on membrane depolarization (Assaf and Chung 1984; Howell et al. 1984), and is estimated to reach 300 µM at the synaptic cleft during strong stimulation (Frederickson et al. 1983). At physiological concentrations, zinc has been shown to modulate ligand-gated ion channels. For example, zinc significantly inhibits NMDA receptor function, slightly potentiates the function of non-NMDA receptors (Westbrook and Mayer 1987), and blocks certain subtypes of GABA_A receptor channels (GABA_ARs) (Westbrook and Mayer 1987). Although GlyRs and GABA_ARs are highly homologous, zinc concentrations <100 µM elicit opposing effects on hippocampal GlyRs and GABA_ARs, potentiating GlyRs (present report) but depressing GABA_ARs (Westbrook and Mayer 1987).

Zinc modulation of GlyRs is variable and the specific zinc effects on GlyR function depend on the brain region in which the receptors are expressed. For example, zinc depresses I_gly in rat cerebellar granular cells (Virginio and Cherubini 1997) and rat septal cholinergic neurons (Kumamoto and Murata 1996). However, zinc modulation of I_gly in rat embryonic spinal cord, human GlyR 1 and GlyR 2 expressed in HEK 293 cells and Xenopus oocytes (Bloomenthal et al. 1994; Laube et al. 1995) and isolated retinal cells (Han and Wu 1999) is bidirectional, eliciting facilitation and depression depending on the zinc concentration.

The data presented in this paper demonstrate that zinc modulates native GlyRs expressed by hippocampal neurons and that this modulation is bidirectional. At low concentrations, zinc potentiates I_gly and at high concentrations suppresses these currents. The zinc concentrations eliciting potentiation and suppression of GlyRs in our study are similar to previous studies in other brain regions. Interestingly, we show that zinc concentrations (<100 µM) that potentiate I_gly suppress I_NMDA recorded from hippocampal neurons. These results suggest that a low concentration of zinc should increase inhibition of hippocampal circuits via its differential effects on I_gly and I_NMDA, which may contribute to the control of excitability in hippocampus.

Are GlyRs involved in fast inhibitory transmission in hippocampus?

The answer to this question is presently unknown. In spinal cord and brain stem, glycine and GABA are co-released from inhibitory interneurons and simultaneously activate GlyRs and GABA_ARs located in the postsynaptic density (Jonas et al. 1998; O'Brien and Berger 1999). Elegant electrophysiological recordings from motoneurons in spinal cord and brain stem slices show that miniature IPSCs are mediated by the activation of both receptors, demonstrating cotransmission of these two inhibitory neurotransmitters (Jonas et al. 1998; O'Brien and Berger 1999). A recent study in cerebellar slices has also demonstrated GABA and glycine co-release at synapses onto Golgi cells (Dumoulin et al. 2001). However, in recordings from identified lamina I neurons in spinal cord slices, it appears that although glycine and GABA are coreleased, GABA_ARs seem to be located extrasynaptically because the GABA_AR-mediated component of the synaptic current is observed only following large but not minimally evoked monosynaptic IPSCs (Chery and de Koninck 1999).

These studies raise important questions as to whether glycine and GABA could be coreleased from GABAergic interneurons in hippocampus and as to the synaptic versus extrasynaptic location of the GlyRs expressed by pyramidal cells and interneurons. In hippocampus, GABA_AR antagonists are reported to block IPSCs, indicating that GABA_ARs, but not GlyRs, are located synaptically (Mody et al. 1994). However, GlyRs could be located just outside the synapse, similarly to the distribution of GABA_ARs in spinal cord as discussed in the preceding text. Although we have not yet investigated the presence of GlyR-mediated synaptic currents, we speculate that GlyRs in hippocampus are located at both synaptic and extrasynaptic sites. Our picrotoxin data support this idea because it suggests that heteromeric receptors make up some portion of the GlyRs expressed by hippocampal neurons and the presence of the  subunit in functional GlyRs is required for synaptic clustering by the intracellular protein gephyrin (Kirsch et al. 1993; Meyer et al. 1995).

Finally, glycine transporters, GlyT1 and GlyT2, known to be responsible for terminating glycinergic transmission in spinal cord and brain stem, are present in hippocampus and are colocalized with GlyRs, indicating that they will modulate the activity at GlyRs by regulating the local glycine concentration (Jursky and Nelson 1995; Klancnik et al. 1992). The source of glycine in hippocampus is unknown however the fact that glycine release is partially calcium and action potential dependent strongly indicates that some glycine is of neural origin (Engblom et al. 1996; Klancnik et al. 1992).

In summary, we propose that GlyRs participate in an inhibitory mechanism in hippocampus, modulating neuronal activity. We speculate that specifically enhancing GlyR activity, similarly to increasing GABA_AR activity, could be beneficial in depressing hyperexcitability that ensues in epilepsy, encouraging future investigations into the precise location and function of these understudied inhibitory receptors.