Synaptic plasticity and learning. II: Do different kinds of plasticity underlie different kinds of learning?

This paper discusses certain issues connected with the question of whether synaptic plasticity is involved in information storage by the brain. We begin by contrasting two well documented types of synaptic plasticity--activity-dependent modulation of presynaptic facilitation ("Kandel synapses") and NMDA receptor triggered alterations in excitatory amino acid transmission ("Hebb synapses"). We then propose that, embedded-into appropriate circuitry, these different forms of plasticity might underlie different kinds of learning. In partial support of this idea, we show that, in freely moving rats, intrahippocampal microinfusion of the NMDA-receptor antagonist D,L-AP5 causes a dose-dependent impairment of a type of spatial learning known to be sensitive to disruption by hippocampal lesions (water-maze place-navigation). In a second experiment, the same drug infusion protocol is shown to cause a blockade of hippocampal long-term potentiation in vivo across a comparable dose range. Finally, third, autoradiographic experiments indicate that diffusion of the drug was largely restricted to the hippocampus. Together, these results (1) suggest that blockade of hippocampal NMDA receptors, under conditions which leave baseline synaptic transmission relatively unaffected, blocks a type of learning with which the vertebrate hippocampus has been implicated on the basis of neuropsychological work; and (2) supports Abrams and Kandel's Trends Neurosci. 11 (1988) recent proposal that there may be several logically distinct forms of synaptic plasticity.

Human embryonal carcinoma stem cells expressing green fluorescent protein form functioning neurons in vitro: a research tool for co-culture studies.

Neural differentiation is controlled by complex molecular mechanisms that determine cell fate and diversity within the nervous system. Interactions between developing tissues play an important role in regulating this process. In vitro co-culture experiments offer a method to study cell differentiation and function under controlled conditions, with the additional benefit of investigating how interactions between populations of cells influence cell growth and behavior. However, it can often be difficult to distinguish between populations of co-cultured cells. Here we report the development of a human embryonal carcinoma (EC) stem cell line (named TERA2.cl.SP12-GFP) that expresses the genetic marker, green fluorescent protein (GFP). Here, we demonstrate that TERA2.cl.SP12-GFP stem cells stably express GFP and that this remains detectable during retinoic acid-induced differentiation. Regulated expression of neural markers during cell development correlated with the formation of morphologically identifiable neurons. Populations of post-mitotic GFP-positive neurons were readily purified and electrophysiological characterization confirmed that such neurons were functionally active. Thus, cultured TERA2.cl.SP12-GFP cells can be readily distinguished from alternative cell types in vitro and provide an amenable system for live cell imaging to study the development and function of human neurons in isolation, and in co-culture with other tissue types.

A patch clamp study of the effects of ciprofloxacin and biphenyl acetic acid on rat hippocampal neurone GABAA and ionotropic glutamate receptors.

The neurotoxic effects of 4-quinolones alone and in combination with certain non-steroidal anti-inflammatory drugs (NSAIDs) may be related to an interaction at GABAA and/or ionotropic glutamate receptors. In the present study, the effects of the fluoroquinolone, ciprofloxacin, alone and in combination with the NSAID, biphenyl acetic acid (BPAA), were examined on GABAA-, NMDA-, AMPA-, and kainate-evoked current responses recorded from cultured rat hippocampal neurones, using the whole cell patch clamp technique. GABA-evoked currents were reversibly inhibited by bicuculline (3 microM) and ciprofloxacin (100 microM) to 11 +/- 5 and 38 +/- 7% of control, respectively. BPAA (100 microM) had little affect on the GABA current (the response was 82 +/- 4% of control) but enhanced the inhibitory potency of ciprofloxacin by approx. 3000-fold. The antagonist effects of ciprofloxacin (30 microM) and ciprofloxacin (0.03 microM) together with BPAA (100 microM) on the GABA-evoked current were not voltage-dependent. Whole cell currents evoked by NMDA, AMPA or kainate were little influenced by ciprofloxacin (100 microM), BPAA (100 microM), or ciprofloxacin plus BPAA (both at 100 microM); the responses being > or = 90% of control in all cases. These data suggest that the proconvulsant effects of quinolones when combined with BPAA may be related to antagonism of central GABAA receptors but not to an interaction at ionotropic glutamate receptors.

Characterization of the interaction between a novel convulsant agent, norbiphen, and GABA(A) and other ligand-gated ion channels.

A hybrid molecule composed of the antimicrobial, norfloxacin, linked to the non-steroidal anti-inflammatory drug (NSAID), biphenylacetic acid, which we have termed norbiphen, is a lethal convulsant in vivo and an antagonist of rodent GABA(A) receptors in vitro. In the present study, the selectivity, molecular site(s) and mechanism of action of this novel convulsant were investigated using electrophysiological techniques. Sub-maximal GABA-evoked currents recorded from rodent hippocampal neurons were reversibly inhibited by norbiphen (1 microM) to 5+/-2% of control whereas glutamate, NMDA and glycine activated responses were little or unaffected. Sub-maximal GABA-evoked currents recorded from oocytes expressing recombinant human alpha1beta2gamma2s or alpha1beta2 GABA(A) receptors were also reversibly inhibited by norbiphen (1-1000 nM) with an IC(50) (+/-s.e.m.) of 5.7+/-1 and 8.8+/-1 nM, respectively. Similarly, GABA currents recorded from alpha1beta1gamma2s, alpha1beta1 and beta2gamma2s receptors were inhibited with IC(50)s of 16.1+/-1, 18.8+/-1 and 4.2+/-1 nM, respectively. In contrast, norbiphen (100 nM) had little or no effect at rho1 GABA(C) homomers. At alpha1beta2gamma2s receptors, norbiphen had no affect on the GABA reversal potential, and inhibition was not voltage-dependent, suggesting that this compound does not act at the ion channel. The GABA concentration response curve was shifted in a competitive-like fashion by norbiphen (10-300 nM) and a Schild analysis of these data yielded a slope of 0.94+/-0.1 and a pA(2) of 7.77. Our data reveal a novel, selective and highly potent antagonist of GABA(A) receptors. Norbiphen should be a valuable agent in future studies of this receptor complex.

Selective antagonism of the GABA(A) receptor by ciprofloxacin and biphenylacetic acid.

1. Previous studies have shown that ciprofloxacin and biphenylacetic acid (BPAA) synergistically inhibit y-aminobutyric acid (GABA)A receptors. In the present study, we have investigated the actions of these two drugs on other neuronal ligand-gated ion channels. 2. Agonist-evoked depolarizations were recorded from rat vagus and optic nerves in vitro by use of an extracellular recording technique. 3. GABA (50 microM)-evoked responses, in the vagus nerve in vitro, were inhibited by bicuculline (0.3-10 microM) and picrotoxin (0.3-10 microM), with IC50 values and 95% confidence intervals (CI) of 1.2 microM (1.1-1.4) and 3.6 microM (3.0-4.3), respectively, and were potentiated by sodium pentobarbitone (30 microM) and diazepam (1 microM) to (mean+/-s.e.mean) 168+/-18% and 117+/-4% of control, respectively. 5-Hydroxytryptamine (5-HT; 0.5 microM)-evoked responses were inhibited by MDL 72222 (1 microM) to 10+/-4% of control; DMPP (10 microM)-evoked responses were inhibited by hexamethonium (100 microM) to 12+/-5% of control, and alphabetaMeATP (30 microM)-evoked responses were inhibited by PPADS (10 microM) to 21+/-5% of control. Together, these data are consistent with activation of GABA(A), 5-HT3, nicotinic ACh and P2X receptors, respectively. 4 Ciprofloxacin (10-3000 microM) inhibited GABA(A)-mediated responses in the vagus nerve with an IC50 (and 95% CI) of 202 microM (148-275). BPAA (1-1000 microM) had little or no effect on the GABA(A)-mediated response but concentration-dependently potentiated the effects of ciprofloxacin by up to 33,000 times. 5. Responses mediated by 5-HT3, nicotinic ACh and P2X receptors in the vagus nerve and strychnine-sensitive glycine receptors in the optic nerve were little or unaffected by ciprofloxacin (100 microM), BPAA (100 microM) or the combination of these drugs (both at 100 microM). 6. GABA (1 mM)-evoked responses in the optic nerve were inhibited by bicuculline with an IC50 of 3.6 microM (2.8-4.5), a value not significantly different from that determined in the vagus nerve. Ciprofloxacin also inhibited the GABA-evoked response with an IC50 of 334 microM (256-437) and BPAA (100 microM) potentiated these antagonist effects. However, the magnitude of the synergy was 48 times less than that seen in the vagus nerve. 7. These data indicate that ciprofloxacin and BPAA are selective antagonists of GABA(A) receptors, an action that may contribute to their excitatory effects in vivo. Additionally, our data suggest that the molecular properties of GABA(A) receptors in different regions of the CNS influence the extent to which these drugs synergistically inhibit the GABA(A) receptor.

The mechanism of action and pharmacological specificity of the anticonvulsant NMDA antagonist MK-801: a voltage clamp study on neuronal cells in culture.

1. Some possible molecular mechanisms of action of the anxiolytic, anticonvulsant and neuroprotective agent MK-801 have been examined in 'whole-cell' voltage clamp recordings performed on rat hippocampal and cortical neurones, bovine adrenomedullary chromaffin cells and N1E-115 neuroblastoma cells maintained in cell culture. 2. Transmembrane currents recorded from rat hippocampal and cortical neurones in response to locally applied N-methyl-D-aspartate (NMDA) were antagonized by MK-801 (0.1-3.0 microM). Blockade was use-dependent, and little influenced by transmembrane potential. MK-801 (3 microM) had no effect on currents evoked by kainate (100 microM). 3. The antagonism of NMDA-induced currents by MK-801 was only slowly and incompletely reversed when the cell membrane potential was clamped at -60 mV during washout. Prolonged applications of NMDA at +40, but not -60 mV during washout, markedly accelerated recovery from block. 4. In contrast to MK-801, ketamine (10 microM) blocked NMDA-induced currents in a voltage-dependent manner. Blockade increased with membrane hyperpolarization and was completely reversible upon washout. 5. MK-801 (1-10 microM) produced a voltage- and concentration-dependent block of membrane currents elicited by ionophoretically applied acetylcholine (ACh) recorded from bovine chromaffin cells. The block was readily reversible upon washout. 6. gamma-Aminobutyric acidA (GABAA) receptor-mediated chloride currents of chromaffin cells were unaffected by MK-801 (1-100 microM). In contrast, such currents were potentiated by diazepam (1 microM). MK-801 (100 microM) had no effect on currents evoked by GABA on hippocampal neurones. 7. MK-801 (10 microM) had little effect on membrane currents recorded from N1E-115 neuroblastoma cells in response to ionophoretically applied 5-hydroxytryptamine (5-HT). Such currents were antagonized by the 5-HT3 receptor antagonist GR 38032F (1 nM) and also by MK-801 at high concentration (100 microM). 8. Voltage-activated, tetrodotoxin-sensitive, sodium currents of chromaffin cells were unaffected by 10 microM MK-801. However, at a relatively high concentration (100 microM), MK-801 reduced the amplitude of such currents to approximately 77% of control. 9. The relevance of the present results to the central actions of MK-801 is discussed.

A further study of the neuromuscular effects of vesamicol (AH5183) and of its enantiomer specificity.

1. The effects of vesamicol (2-(4-phenylpiperidino) cyclohexanol), an inhibitor of acetylcholine storage, and its two optical isomers have been studied on neuromuscular transmission in rat and frog muscle, and on nerve conduction in frog nerve. 2. Racemic vesamicol produced a pre-block augmentation of twitch tension that also occurred in directly-stimulated muscle. This effect is thus at least partially due to an increase in muscle contractility. 3. (-)-Vesamicol was approximately 20 times more potent than (+)-vesamicol in blocking twitches elicited at 1 Hz. This degree of stereoselectivity is similar to that measured for inhibition of acetylcholine uptake by isolated synaptic vesicles. Both enantiomers were equally weak in reducing nerve action potential amplitude in frog nerve. 4. Further studies with the active isomer, (-)-vesamicol, showed that, like that produced by racemic vesamicol, the neuromuscular block was highly frequency-dependent. The block was not reversed by choline or neostigmine, but was partially reversed by 4- or 3,4-aminopyridine. 5. Preliminary electrophysiological studies showed that vesamicol reduced miniature endplate potential amplitude in rapidly-stimulated frog nerve-muscle preparations. Addition of lanthanum ions increased the frequency of miniature endplate potentials and led to the appearance of apparently normal-sized potentials amongst those of reduced amplitude. 6. The results show the close agreement between pharmacological and biochemical observations indicating the suitability of the rat diaphragm as a test model for substances of this nature. The degree of reversibility of the vesamicol-induced neuromuscular block by aminopyridines was unexpected, and it is suggested that in the presence of a drug which greatly increases release, a pool of acetylcholine is capable of being released which is not normally releasable after block of storage by vesamicol. It is also considered possible that the results from the intracellular recording studies may be explained in these terms.

Purification and amino acid sequence of a highly insecticidal toxin from the venom of the brazilian spider Phoneutria nigriventer which inhibits NMDA-evoked currents in rat hippocampal neurones.

A new insecticidal toxin Tx4(5-5) was isolated from the fraction PhTx4 of the venom of the spider Phoneutria nigriventer by reverse phase high performance liquid chromatography (HPLC) and anion exchange HPLC. The complete amino acid sequence determined by automated Edman degradation showed that Tx4(5-5) is a single chain polypeptide composed of 47 amino acid residues, including 10 cysteines, with a calculated molecular mass of 5175 Da. Tx4(5-5) shows 64% of sequence identity with Tx4(6-1), another insecticidal toxin from the same venom. Tx4(5-5) was highly toxic to house fly (Musca domestica), cockroach (Periplaneta americana) and cricket (Acheta domesticus ), producing neurotoxic effects (knock-down, trembling with uncoordinated movements) at doses as low as 50 ng/g (house fly), 250 ng/g (cockroach) and 150 ng/g (cricket). In contrast, intracerebroventricular injections (30 microg) into mice induced no behavioural effects. Preliminary electrophysiological studies carried out on whole-cell voltage-clamped rat hippocampal neurones indicated that Tx4(5-5) (at 1 microM) reversibly inhibited the N-methyl-D-aspartate-subtype of ionotropic glutamate receptor, while having little or no effect on kainate-, alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid- or gamma-aminobutyric acid-activated currents.

The effects of quinolones and NSAIDs upon GABA-evoked currents recorded from rat dorsal root ganglion neurones.

Recent animal studies have demonstrated a proconvulsant effect of certain quinolone and non-steroidal anti-inflammatory drug combinations. Radioligand binding experiments have indicated that these actions may be mediated by antagonism of the GABAA receptor. The present study has further investigated this hypothesis in a functional assay by examining the effects of the quinolones ciprofloxacin and ofloxacin alone and in combination with either fenbufen or biphenyl acetic acid (BPAA) upon GABA-evoked currents recorded from voltage-clamped rat dorsal root ganglion neurones (DRG) maintained in cell culture. GABA-evoked whole cell currents were weakly but dose-dependently (30 microM-1 mM) reduced in the presence of ciprofloxacin and ofloxacin. The IC50 for ciprofloxacin was 100 microM but greater than 1 mM for ofloxacin. Application of either fenbufen (100 microM) or BPAA (100 microM) alone produced little effect on the GABA-evoked currents. However, the inhibitory action of ciprofloxacin was enhanced in the presence of 100 microM fenbufen by approximately five-fold whereas the antagonism of GABA responses by ofloxacin was unaffected. In contrast, BPAA (100 microM) had a dramatic effect on the inhibitory actions of both antibiotics such that the IC50 for ciprofloxacin and ofloxacin was reduced to 0.03 and 0.3 microM respectively. The present results support earlier binding studies and extend them by demonstrating electrophysiologically a potent quinolone/NSAID drug interaction at the GABAA receptor. The mechanism(s) of this novel interaction remains to be determined. These results are commensurate with clinical observations of an increased risk of fits in patients prescribed certain quinolones together with certain NSAIDs.

An electrophysiological study of the effects of propofol on native neuronal ligand-gated ion channels.

1. Pharmacological evidence suggests that some of the clinical actions of propofol may be mediated, at least in part, by positive modulation of the GABA(A) receptor chloride channel. The effect of propofol at other native neuronal ligand-gated ion channels is unclear. 2. To gain some insight into the effects of propofol at a range of native neuronal receptors, the present study has used an extracellular recording technique and determined its effects at GABA(A), 5-HT3, P2X and nicotinic acetylcholine (nACh) receptors of the rat isolated vagus nerve and the GABA(A) and strychnine-sensitive glycine receptor of the rat isolated optic nerve. In addition, we have used patch-clamp recording techniques to further investigate the effects of propofol at the GABA(A) and strychnine-sensitive glycine receptors in rat cultured hippocampal neurons. 3. Propofol (0.3-100 micromol/L) concentration-dependently potentiated submaximal GABA-evoked responses in the vagus nerve and shifted the GABA concentration-response curve to the left. In contrast, propofol at concentrations ranging from 1 to 10 micromol/L had little or no effect on 5-HT3, P2X or nACh receptor-mediated responses in the vagus nerve but, at 100 micromol/L, propofol inhibited these responses to approximately 50% of control. In the optic nerve, EC20 GABA-evoked responses were also potentiated by propofol (10 micromol/L), while EC20 glycine-evoked responses were minimally enhanced. 4. Further investigations using cultured hippocampal neurons showed that submaximal (10 micromol/L) GABA-evoked currents were potentiated by propofol (1-10 micromol/L), in a non-voltage-dependent manner, whereas submaximal (100 micromol/L) glycine-evoked currents were unaffected. 5. These data suggest that propofol, at therapeutic concentrations, exerts its principle pharmacological actions at GABA(A) receptors with relatively little effect at other neuronal ligand-gated ion channels.

Subunit-selective modulation of GABAA receptors by the non-steroidal anti-inflammatory agent, mefenamic acid.

Mefenamic acid (MFA) has anti-convulsant and pro-convulsant effects in vivo, and has been shown to potentiate and inhibit GABAA (gamma-aminobutyric acid) receptors in vitro. In this study, whole-cell currents were recorded from Xenopus oocytes and human embryonic kidney (HEK) cells expressing human recombinant GABAA receptors to resolve the molecular mechanisms by which MFA modulates GABAA receptor function. We demonstrate that MFA potentiated GABA-activated currents for alpha1beta2 gamma2S (EC50 = 3.2 +/- 0.5 microM), but not for alpha1beta1 gamma2S receptors. MFA also enhanced GABA-activated responses and directly activated alpha1beta2/beta3 GABAA receptors, but inhibited responses to GABA on alpha1beta1 constructs (IC50 = 40 +/- 7.2 microM). A comparison of beta1, beta2 and beta3 subunits suggested that the positive modulatory action of MFA involved asparagine (N) 290 in the second transmembrane domain (TM2) of the beta2 and beta3 subunits. Mutation of N290 to serine (S) markedly reduced modulation by MFA in alpha1beta2(N290S)gamma2S receptors, whereas alpha1beta1(S290N)gamma2S constructs revealed potentiated responses to GABA (EC50 = 7.8 +/- 1.7 microM) and direct activation by MFA. The potentiation by MFA displayed voltage sensitivity. The direct activation, potentiation and inhibitory aspects of MFA action were predominantly conferred by the beta subunits as the spontaneously active homomeric beta1 and beta3 receptors were susceptible to modulation by MFA. Molecular comparisons of MFA, loreclezole and etomidate, agents which exhibit similar selectivity for GABAA receptors, revealed their ability to adopt similar structural conformations. This study indicates that N290 in TM2 of beta2 and beta3 subunits is important for the regulation of GABAA receptor function by MFA. Our data provide a potential molecular mechanism for the complex central effects of MFA in vivo.