INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ionotropic GABA receptors (iGABARs) are ligand-gated anion channels that mediate most inhibitory synaptic transmission. They are found throughout the CNS of vertebrates and invertebrates, and are perhaps the most widely studied inhibitory receptors.

In the course of studies of excitotoxicity in chick embryo retinal (CER) neurons, we gathered indirect evidence for the presence there of an iGABA receptor that is insensitive to picrotoxin, yet sensitive to bicuculline (Chen et al. 1999). In this tissue, excitotoxic cell death requires the entry of pathological quantities of Cl from the extracellular space (Chen et al. 1998). The routes for this Cl entry are the ligand-gated chloride channels, since cell death in the CER induced by 320 µM kainic acid (KA) can be completely prevented by a combination of GABA and glycine antagonists. Interestingly, full protection required the presence of bicuculline in addition to picrotoxin. This finding led us to predict that there is an iGABA receptor that is bicuculline sensitive and picrotoxin insensitive, which we believe is without precedent.

There is a great diversity of iGABA receptors, which differ in kinetic properties, affinity for agonists and antagonists, and sensitivity to modulators (Chebib and Johnston 1999; Mehta and Ticku 1999; Whiting et al. 1995; Wisden and Seeburg 1992). The broadest classification of iGABA receptors is into the GABA-A receptors, which can be blocked by the competitive antagonist bicuculline, and the GABA-C receptors, which are bicuculline insensitive. The basis for iGABA receptor diversity is fairly well understood. On the molecular level, iGABARs are thought to be pentameric complexes of protein subunits. There are several families of iGABA receptor subunits, known as the , , , , , , and  families, and most of these families have several members. GABA-A receptors are thought to contain  subunits, and GABA-C receptors  subunits. Structural diversity within these families is due to both multiple genes (e.g., there are 6 different  subunit genes) and to differential gene splicing [there are splice variants of the  (Bateson et al. 1991; Korpi et al. 1994) and  (Poulsen et al. 2000; Whiting et al. 1990) subunit genes expressed]. Functional iGABA receptors are most often heteromers, composed of one or more  subunits in combination with , , and others. The exceptions are the  subunits, some of which can form functional homomers, but may assemble as combinations of 's (Zhang et al. 1995) or with 2 (Qian and Ripps 1999). Differential subunit assembly provides for the functional diversity of GABA receptors. For example, different  subunits are responsible for distinct GABA affinities, and sensitivity to benzodiazepines is conferred by the presence of  subunits (Chebib and Johnston 1999).

Picrotoxin is a noncompetitive antagonist of virtually all classes of iGABARs (Whiting et al. 1995). It binds to a site within the channel and prevents ion flow. The only wild-type, endogenous iGABA receptor that we are aware of that is insensitive to picrotoxin is the GABA-C receptor of rats (Zhang et al. 1995). No native, bicuculline-sensitive GABA-A receptors are known that are not blocked by picrotoxin.

The purpose of the present study was to directly test the hypothesis that a bicuculline-sensitive, picrotoxin-insensitive iGABA receptor is present on chick embryo retinal neurons. We present here morphological and electrophysiological evidence for the presence of this novel receptor.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pathomorphology

The solutions and procedures were exactly as described in detail previously (Romano et al. 1998). Retinas from 15-day-old chick embryos were removed and cut into thirds on ice, then incubated at 25°C for 30 min in the presence of the indicated agents. The tissue was then fixed in mixed aldehydes, dehydrated, embedded in plastic, sectioned at 1 µM, stained, and examined microscopically.

Retinal slice preparation

The procedures for preparation and recording from the retinal slice were described in detail previously (Lukasiewicz et al. 1994), including modifications appropriate for recording in chick (Chen et al. 1998; Wong et al. 1998). Fertilized eggs and 8-wk-old chickens were purchased from SPAFAS (Roanoke, IL). The eggs were kept at 38°C. Chickens were anesthetized with pentobarbital and killed by decapitation. Eye cups were isolated and immersed in cold oxygenated Chick Ringer solution (in mM: 124 NaCl, 5 KCl, 2 CaCl₂, 2 MgCl₂, 1.25 KH₂PO₄, 20 glucose, and 20 Hepes, pH 7.4). Small squares were cut from isolated retina, placed ganglion cell side down on a slightly larger filter square (HA; Millipore, Bedford, MA) The tissue with filter attached was sliced at 100 µM (adult retina) or 150-200 µM (embryonic) intervals. The slices were transferred to the recording chamber and immobilized by embedding the ends of the filter into two rails of vacuum grease that had been laid down in the chamber. A Nikon ×40 long working-distance water immersion objective was used for visualization of cells on the surface of the slices.

Recording

Whole cell patch-clamp recordings were made as described (Chen et al. 1998). The pipette solution contained (in mM): 137 CsCl, 2 MgCl₂, 0.1 CaCl₂, 1.1 EGTA, and 10 Hepes, pH 7.4, with CsOH. Ganglion cells or amacrine cells were routinely held at 75 mV. The recording chamber was superfused continually with Chick Ringer solution containing: strychnine at 3 µm, D-APV at 40 µM, CNQX at 10 µM, and tetrodotoxin at 0.5 µM to block glycine receptors, NMDA receptors, non-NMDA receptors, and voltage-gated sodium channels, respectively. Picrotoxin at 500 µM was present in this bathing fluid where indicated, as were other inhibitors and modulators. Intracellular electrodes were filled with pipette solution and 0.1% Lucifer yellow; pipette resistances were 10-20 M. GABA, 500 µM in Chick Ringer's, was puffed onto the dendrites of ganglion cells with a Picrospritzer (General Valve, Fairfield, NJ). Recordings were obtained with an Axopatch 200B patch-clamp amplifer (Axon Instruments, Foster City, CA). The data were digitized and stored with a 33 MHz 486PC using a Labmaster DMA data acquisition board (Scientific Solutions, Solon, OH). Responses were filtered at 1 kHz. Patchit software (White Perch Software, Sommerville, MA) was used to generate voltage command outputs, acquire data, gate the drug perfusion valves, and trigger the Picospritzer. Data were analyzed using Tack software (White Perch Software, Sommerville, MA) and expressed as means ± SE. Peak responses were defined as the mean current in a 5-ms window around the highest current. Charge transfer was integrated over 4.5 s from the time of the GABA puff.

Agents used and sources were the following: CNQX (6-cyano7-nitroquinoxaline-2,3-dione), NO-711 (1-(2-[([diphenylmethylene] imino)oxy]ethyl)-1,2,5,6-tetrahydro-3-pyridinecarboxylic acid hydrochloride), SR-95531 (2-(3-carboxypropyl)-3-amino-6-(4-methoxyphenyl)pyridazinium bromide), and bicuculline methbromide were obtained from Research Biochemicals (Natick, MA), and D-APV (D-2-amino-5-phosphonopentanoic acid) was obtained from Precision Biochemicals (Vancouver, BC, Canada). Other reagents were obtained from Sigma Chemicals (St. Louis, MO).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Biochemical evidence suggested that bicuculline or picrotoxin alone would each provide only partial protection from KA-induced toxicity in chick embryo retina. GABA-A receptors are blocked by both antagonists, while GABA-C receptors are insensitive to bicuculline but blocked by picrotoxin. We therefore postulated that a bicuculline-sensitive yet picrotoxin-insensitive GABA receptor may exist in the retina.

Rather than blindly searching for GABA currents of the requisite pharmacology, we tried to obtain information concerning the possible cellular location of these receptors. Kainate toxicity is rapidly and dramatically manifest morphologically. Within moments of KA application, neuronal somata and processes swell, and nuclear changes are apparent, in susceptible cells. If GABA receptors with distinct inhibitor sensitivities are differentially expressed in different retinal neuronal classes, then by applying KA in the presence of bicuculline and picrotoxin individually and in combination, the cellular location of the receptor classes may be apparent.

Figure 1 illustrates the pathomorphological appearance of the retina after exposure to KA in the presence of the blockers individually or in combination. KA causes obvious and massive swelling in all layers of the retina, somal and synaptic. Many nuclear changes are evident in inner retinal neurons. In the presence of bicuculline alone, the pathology remains widespread, but appears somewhat less severe in the ganglion cell layer. In the presence of picrotoxin alone, there is pronounced preservation of the appearance of most neurons in the outer half of the INL (presumed bipolar cells). This result indicates that most if not all the GABA receptors on bipolar cells are sensitive to blockade by picrotoxin. However, marked abnormalities are still present in the ganglion cell layer, and in the inner half of the INL (presumed amacrine cells). Interestingly, the addition of bicuculline as well as picrotoxin causes most of the ganglion and amacrine somal and nuclear changes to diminish, although pathology in the IPL is still evident, presumably due to swelling of postsynaptic processes containing receptors sensitive to KA. This decreased pathology suggests that the hypothesized bicuculline-sensitive, picrotoxin-insensitive receptor contributes toward the toxicity observed in ganglion and amacrine cells of the chick embryo retina. The simplest explanation for this would be that they are present on these cells.

View larger version (67K): [in this window] [in a new window]   Fig. 1. Protection from KA-induced pathomorphology by bicuculline, picrotoxin, and the combination. Retinas were incubated for 30 min in the presence of KA (320 µM), in the presence of bicuculline (BIC, 200 µM), picrotoxin (PIC, 500 µM), or both. Bar, 20 µM.

To directly investigate the possible existence of a picrotoxin-insensitive, bicuculline-sensitive iGABA receptor, whole cell patch clamp recordings were performed using transverse slices of chick embryo retinas. Puffs of GABA elicited large inward currents in all ganglion and amacrine cells recorded from. Interestingly, the blockade afforded by a high concentration of bath-applied picrotoxin (500 µM) was quite variable from cell to cell. The majority of the current was blocked by picrotoxin in every cell tested. However, a picrotoxin-insensitive component of the current was present and varied from almost none to nearly 20%. Examples of individual ganglion and amacrine cells that had exclusively picrotoxin-sensitive currents are shown in Fig. 2, A and C, while examples of cells exhibiting a substantial percent of picrotoxin-insensitive GABA currents are shown in Fig. 2, B and D. The distribution of all cells according to amount of picrotoxin-insensitive GABA current is presented in Fig. 2, E and F. We do not believe that this difference is due to pharmacokinetic differences, as accessibility of the neurons to the bathing fluid in this transverse slice preparation is uniform and complete, and the antagonists are present for minutes. Instead, we take this as strong evidence that multiple iGABA receptors, including picrotoxin-insensitive species, are present on ganglion and amacrine cells. A summary of data from all cells tested concerning the antagonist sensitivity of the picrotoxin-insensitive component is presented in Table 1. (The table also has summary data on the modulator sensitivity of this current.)

View larger version (29K): [in this window] [in a new window]   Fig. 2. Many ganglion and amacrine cells exhibit picrotoxin-insensitive GABA currents. GABA (500 µM; duration indicated by black bar above traces) was puffed onto ganglion (A, B, E) and amacrine (C, D, F) cells in the absence and presence of 500 µM picrotoxin. Some cells had GABA currents completely eliminated by the antagonist (A, C), while others exhibited substantial picrotoxin-insensitive currents (B, D). Distribution of sensitivities shown in the histograms [ganglion cells (E); amacrine cells (F); numbers on bars represent the number of cells in each group].

Neurotransmitters may generate currents independently of interactions with traditional receptors. For example, the glutamate transporters have associated Na⁺ (Erecinska 1987; Kanner and Schuldiner 1987) and Cl currents (Fairman et al. 1995; Wadiche et al. 1995). Currents associated with GABA transporters have also been described (Kaila et al. 1992; Kamermans and Werblin 1992). Experiments were performed to determine whether the picrotoxin-insensitive GABA current might originate from a GABA transporter rather than a GABA receptor. First, the I-V relationship was characterized. A transport current would be expected to be nonlinear and inwardly rectifying, whereas a standard ligand-gated chloride channel would be expected to have a linear I-V curve, reversing at 0 mV (at this experimentally set chloride gradient). Figure 3, A and B, illustrates that the I-V curve of this GABA current had the properties expected of a ligand-gated Cl channel.

View larger version (27K): [in this window] [in a new window]   Fig. 3. Picrotoxin-insensitive GABA currents are not transporter currents. Representative traces (A) and IV curve (B) of the GABA current. The current is typical of a ligand-gated ion channel. C: representative trace illustrating that the GABA transporter blocker NO-711 (8 µM) had no effect on the GABA current. Identical results obtained in 6 ganglion and 2 amacrine cells. D: values of normalized current peak (I) and charge transfer (Q) in response to GABA puffs before, during, and after NO-711 addition to the bath (duration shown by bar above graph). E: representative trace illustrating that the GABA transporter blockade by lithium substitution for sodium had no effect on the GABA current. F: values of normalized current peak (I) and charge transfer (Q) in response to GABA puffs before, during, and after generalized transport block (duration shown by bar above graph). Identical results obtained in 2 ganglion and 2 amacrine cells.

If the current were due to the activity of the GABA transporter, it should be prevented by NO-711, a nontransported competitive blocker of GABA transport. However, NO-711 did not block the current induced by GABA (Fig. 3, C and D; Table 1), suggesting that the current does not originate in a GABA transporter.

Finally, to eliminate the possibility that the current originated from activity of an undescribed GABA transporter that is not blocked by NO-711, we replaced the Na⁺ in the medium with Li⁺ (Fig. 3, E and F; Table 1). All plasma membrane neurotransmitter transporters so far described are Na⁺ dependent. This substitution did not block the picrotoxin-insensitive GABA current. All this evidence suggests that the GABA current is not a transport-associated current, and instead is due to an iGABAR.

Bicuculline (200 µM) was able to block this picrotoxin-insensitive GABA current (Fig. 4, A and B; Table 1). Another competitive blocker of GABA-A receptors, SR-95311 (50 µM), also inhibited the picrotoxin-insensitive GABA current (Fig. 4, C and D; Table 1). These data provide electrophysiological evidence corroborating the toxicological prediction of a picrotoxin-insensitive, bicuculline-sensitive iGABA receptor.

View larger version (29K): [in this window] [in a new window]   Fig. 4. The picrotoxin-insensitive GABA current is blocked by bicuculline and SR-95311. Representative traces (A, C) and time course of blockade and washout (B, D) of inhibition of GABA current by bicuculline (A, B) and SR-95311 (C, D).

The bicuculline sensitivity of this GABA receptor may indicate that it is similar to a GABA-A receptor in its agonist binding site. To explore the properties of modulatory sites on this receptor, we examined the ability of a barbiturate, pentobarbital, and a benzodiazepine, lorazepam, to potentiate the picrotoxin-insensitive GABA current. Pentobarbital greatly potentiated the GABA current observed in the presence of picrotoxin (Fig. 5, A and B; Table 1). Lorazepam more modestly, but consistently, potentiated the GABA current (Fig. 5, C and D; Table 1).

View larger version (29K): [in this window] [in a new window]   Fig. 5. The picrotoxin-insensitive GABA currents are enhanced by both barbiturates (50 µM pentobarbital; A: representative traces; B: time course of onset and washout) and benzodiazepines (50 µM lorazepam; C: representative traces; D: time course of onset and washout).

Figures 3, 4, and 5 illustrated representative traces from embryonic ganglion cells. Similar results were found in amacrine cells, and summary data are presented in Table 1.

All the data presented so far were obtained from embryonic retinal neurons. To determine whether the picrotoxin-insensitive current was also present in mature retinal neurons, we performed similar whole cell patch clamping experiments on ganglion cells in slices prepared from 8-wk-old chickens. The results were quite similar to those observed in the embryo (Fig. 6). Recordings were made from 17 mature chicken ganglion cells, and 7 had picrotoxin-insensitive components of peak height >5%. Figure 6A illustrates a cell with a picrotoxin-insensitive component equal to 5% of the total GABA current. All the remaining panels in Fig. 6 were obtained from this cell. Figure 6B demonstrates that this current was completely blocked by bicuculline. Figure 6C demonstrates that the current was not blocked, but instead was enhanced, when Li⁺ was present to block transport. Similar results were obtained using NO-711 (not shown). This enhancement is what would be expected if GABA clearance contributes to the time course of the response and the agents effectively blocked the transporters. Pentobarbital provided a very large enhancement of this current (Fig. 6D), as it did in the embryonic neurons. These effects of antagonists and modulators were all obtained in one or two additional cells. Therefore the picrotoxin-insensitive GABA receptor is a component of adult as well as embryonic chicken retinal neurons.

View larger version (26K): [in this window] [in a new window]   Fig. 6. A picrotoxin-insensitive GABA current is present on retinal ganglion cells from 8-wk-old chickens. All the traces in this figure were from the same cell. A: picrotoxin blocks most but not all the current. B: the picrotoxin-insensitive current is nearly completely blocked by bicuculline (C, control trace; W, open circles, washout). C: lithium substitution for sodium does not block the GABA current, instead enhancing it, probably due to decreased uptake. D: pentobarbital enhances the picrotoxin-insensitive current.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have confirmed the prediction, made on the basis of toxicological experiments, of the presence of a picrotoxin-insensitive, bicuculline-sensitive GABA receptor on chick embryo retinal ganglion and amacrine cells. The receptor appears to be a GABA-A subtype. It also is present on ganglion cells in 8-wk-old chickens, indicating that it is not transiently expressed in the embryo. It will be interesting to examine the complete retinal and CNS distribution, and pattern of developmental expression, of this receptor, but that will be more efficiently and convincingly done once molecular probes are available.

As mentioned in the introduction, the only wild-type, picrotoxin-insensitive iGABA receptor so far described is the rodent GABA-C receptor (Feigenspan et al. 1993; Zhang et al. 1995). The picrotoxin-insensitivity of this receptor is determined by a single amino acid, a methionine in the second transmembrane domain of the wt 2 subunit. Mutation at this position to threonine and coexpression with the 1 subunit yields a receptor once again sensitive to picrotoxin (1 homomers are picrotoxin sensitive). Of course, this is a bicuculline-insensitive receptor, and so differs from the chicken retina receptor described here.

An invertebrate GABA receptor, which emerged as a pesticide resistant mutant, also exhibits picrotoxin insensitivity (Ffrench-Constant et al. 1993). In this case, the point mutation responsible (alanine to serine) was also localized to the second membrane spanning region.

In addition to these "naturally" occurring receptors, there are several examples of picrotoxin-insensitive GABA-A receptors that have been created by directed mutagenesis (Gurley et al. 1995). Mutations of one or two amino acids in the second transmembrane domain of either the 1, 2, or 2 subunit of the rat GABA-A receptor yielded receptors that were picrotoxin insensitive when the three subunits were coexpressed. These results indicate it is impossible at this point to predict which family (, , , etc.) the picrotoxin-insensitive subunit(s) will belong to. However, all these data lead us to predict that the key amino acid differences will be in the second transmembrane domain.

Studies are in progress to obtain the molecular identity of the picrotoxin-insensitive receptor. It will be of interest to determine whether this receptor subunit(s) confers other unique properties onto GABA-A receptors in addition to picrotoxin insensitivity, such as an altered affinity for GABA, desensitization kinetics, or single-channel parameters.

Theoretically, there is an enormous number of GABA receptor subunit permutations that could result in functional receptors, although evidence suggests many fewer are abundantly expressed (Rabow et al. 1995; Stephenson 1995). It is not clear what physiological purpose this variety serves, but certainly circuit properties are dependent on the kinetic and pharmacological properties of neurotransmitter receptors. Characterizing the molecular basis of the GABA receptor subtypes, including the picrotoxin-insensitive GABA-A receptor from chick embryo retina, is important for the understanding of physiological inhibitory transmission.

The novel GABA receptor defined here participates in excitotoxic neurodegeneration of the retina. By defining the characteristic properties, as well as the cellular and developmental expression pattern of this receptor, we may gain insights into the mechanisms of GABA-dependent excitotoxicity. The diverse properties and localization of iGABA receptor subunits suggest that subunit-combination-specific drugs may be developed and have profoundly selective actions. Indeed, some subunit-selective GABAergic agents have been described (Meadows et al. 1998). This picrotoxin-insensitive receptor therefore may be a target for drug development.