GABAA receptor subtypes immunopurified from rat brain with alpha subunit-specific antibodies have unique pharmacological properties.

The unique cytoplasmic loop regions of the alpha 1, alpha 2, alpha 3, and alpha 5 subunits of the GABAA receptor were expressed in bacterial and used to produce subunit-specific polyclonal antisera. Antibodies immobilized on protein A-Sepharose were used to isolate naturally occurring alpha-specific populations of GABAA receptors from rat brain that retained the ability to bind [3H]muscimol, [3H]flunitrazepam, [3H]Ro15-1788, and [125I]iodo-clonazepam with high affinity. Pharmacological characterization of these subtypes revealed marked differences between the isolated receptor populations and was generally in agreement with the reported pharmacological profiles of GABAA receptors in cells transiently transfected with alpha 1 beta 1 gamma 2, alpha 2 beta 1 gamma 2, alpha 3 beta 1 gamma 2, and alpha 5 beta 1 gamma 2 combinations of subunits. Additional subtypes were also identified that bind [3H]muscimol but do not bind benzodiazepines with high affinity. The majority of GABAA receptor oligomers contains only a single type of alpha subunit, and we conclude that alpha 1, alpha 2, alpha 3, and alpha 5 subunits exist in vivo in combination with the beta subunit and gamma 2 subunit.

A novel allosteric modulatory site on the GABAA receptor beta subunit.

Cloning of cDNAs that code for GABAA receptor subunits has revealed multiple receptor populations constructed from different subunit combinations. On native rat and cloned human GABAA receptors, the anticonvulsant compound loreclezole strongly potentiated GABA-mediated chloride currents. Using different combinations of human GABAA receptor subunits expressed in Xenopus oocytes and transfected 293 cells, loreclezole was highly selective for receptors containing the beta 2 or beta 3 subunit over those containing the beta 1 subunit. Loreclezole was demonstrated to act at a site distinct from the benzodiazepine, barbiturate, and steroid sites with a unique subunit dependence. These results describe a previously unidentified modulatory site on the GABAA receptor beta subunit that allows pharmacological discrimination of different GABAA receptor subpopulations in the brain and provides a new target for putative anticonvulsant/anxiolytic drugs.

Sedative but not anxiolytic properties of benzodiazepines are mediated by the GABA(A) receptor alpha1 subtype.

Inhibitory neurotransmission in the brain is largely mediated by GABA(A) receptors. Potentiation of GABA receptor activation through an allosteric benzodiazepine (BZ) site produces the sedative, anxiolytic, muscle relaxant, anticonvulsant and cognition-impairing effects of clinically used BZs such as diazepam. We created genetically modified mice (alpha1 H101R) with a diazepam-insensitive alpha1 subtype and a selective BZ site ligand, L-838,417, to explore GABA(A) receptor subtypes mediating specific physiological effects. These two complimentary approaches revealed that the alpha1 subtype mediated the sedative, but not the anxiolytic effects of benzodiazepines. This finding suggests ways to improve anxiolytics and to develop drugs for other neurological disorders based on their specificity for GABA(A) receptor subtypes in distinct neuronal circuits.

Characterization of [35S]t-butylbicyclophosphorothionate ([35S]TBPS) binding to GABAA receptors in postmortem human brain.

BACKGROUND AND PURPOSE: The aim of the present study was to determine whether binding of [(35)S]t-butylbicyclophosphorothionate ([(35)S]TBPS) to the convulsant binding site of GABA(A) receptors in human postmortem brain samples can be used as an in vitro index of the functional activation of these receptors. EXPERIMENTAL APPROACH: Postmortem stability of [(35)S]TBPS binding was assessed in rat brain samples harvested at various times after death and the binding properties of [(35)S]TBPS binding (K(D) and B(max)) were determined in human postmortem brain using radioligand binding studies. In addition, the ability of human brain [(35)S]TBPS binding to be allosterically modulated by compounds that bind at recognition sites distinct from the convulsant binding site was measured. KEY RESULTS: Whereas binding of [(3)H]Ro 15-1788 to the benzodiazepine binding site and [(3)H]muscimol to the agonist (GABA) binding site were retained over a 20 h postmortem interval, there was a significant decrease in the affinity and number of [(35)S]TBPS binding sites. Nevertheless, [(35)S]TBPS binding in human brain could be inhibited by TBPS, picrotoxin, loreclezole and pentobarbital and modulated by GABA with potencies comparable to those observed in rats. In addition, the GABA-induced reduction in human brain [(35)S]TBPS binding could be modulated by benzodiazepine site ligands in a manner that reflected their intrinsic efficacies. CONCLUSIONS AND IMPLICATIONS: These results suggest that allosteric coupling between the [(35)S]TBPS, GABA and benzodiazepine binding sites is preserved in postmortem human brain and that [(35)S]TBPS binding in this tissue may be used to study functional characteristics of native human GABA(A) receptors.

Which GABAA-receptor subtypes really occur in the brain?

GABAA receptors are a heterogeneous family of ligand-gated ion channels responsible for mediating inhibitory neurotransmission in the CNS. Since the identification of mammalian cDNAs encoding 13 GABAA-receptor subunits, the composition of native receptor molecules and their localization in the brain has been an area of intense study. We conclude that the number of major subtypes is probably less than ten but their physiological roles have yet to be clearly defined and this represents the next step in GABAA-receptor research.

Loss of the major GABA(A) receptor subtype in the brain is not lethal in mice.

The alpha1beta2gamma2 is the most abundant subtype of the GABA(A) receptor and is localized in many regions of the brain. To gain more insight into the role of this receptor subtype in the modulation of inhibitory neurotransmission, we generated mice lacking either the alpha1 or beta2 subunit. In agreement with the reported abundance of this subtype, >50% of total GABA(A) receptors are lost in both alpha1-/- and beta2-/- mice. Surprisingly, homozygotes of both mouse lines are viable, fertile, and show no spontaneous seizures. Initially half of the alpha1-/- mice died prenatally or perinatally, but they exhibited a lower mortality rate in subsequent generations, suggesting some phenotypic drift and adaptive changes. Both adult alpha1-/- and beta2-/- mice demonstrate normal performances on the rotarod, but beta2-/- mice displayed increased locomotor activity. Purkinje cells of the cerebellum primarily express alpha1beta2gamma2 receptors, and in electrophysiological recordings from alpha1-/- mice GABA currents in these neurons are dramatically reduced, and residual currents have a benzodiazepine pharmacology characteristic of alpha2- or alpha3-containing receptors. In contrast, the cerebellar Purkinje neurons from beta2-/- mice have only a relatively small reduction of GABA currents. In beta2-/- mice expression levels of all six alpha subunits are reduced by approximately 50%, suggesting that the beta2 subunit can coassemble with alpha subunits other than just alpha1. Our data confirm that alpha1beta2gamma2 is the major GABA(A) receptor subtype in the murine brain and demonstrate that, surprisingly, the loss of this receptor subtype is not lethal.

Differential expression of GABAA receptor alpha-subunits in rat brain during development.

Unique cytoplasmic loop regions of the alpha 1, alpha 2, alpha 3, and alpha 5 subunits of the GABAA receptor have been expressed in E. coli and used to generate polyclonal antisera specific for these subunits. The antibodies identify proteins by SDS-polyacrylamide gel electrophoresis and western blotting of molecular size 51 kDa, 53 kDa, 59 kDa and 55 kDa, respectively, which show differential patterns of expression during development. Whereas the alpha 2 and alpha 3 subunits are present at early stages, the expression of alpha 1 and alpha 3 subunits is low at birth and increases with age. This differential expression could be correlated with previous studies examining the developmental expression of BZ1 and BZ2 benzodiazepine binding sites.

Regional differences in the inhibition of mouse in vivo [3H]Ro 15-1788 binding reflect selectivity for alpha 1 versus alpha 2 and alpha 3 subunit-containing GABAA receptors.

The benzodiazepines flunitrazepam, diazepam, and Ro 15-1788 and the beta-carboline DMCM bind with equivalent affinity to the benzodiazepine binding site of GABAA receptors containing different alpha subunits (i.e., alpha 1, alpha 2, alpha 3, or alpha 5); whereas, the triazolopyridazine CL 218,872 and imidazopyridine zolpidem have higher affinity for alpha 1 subunit-containing GABAA receptors. In the present study, the in vivo binding of [3H]Ro 15-1788 in mouse cerebellum and spinal cord was used to establish the occupancy of the benzodiazepine binding site of GABAA receptors containing primarily alpha 1 and alpha 2/alpha 3 subunits, respectively. Thus, the nonselective compounds flunitrazepam, diazepam, and DMCM all produced a similar inhibition of binding in cerebellum and spinal cord (respective ID50 values of 0.2 to 0.3 mg/kg, 2 mg/kg, and 10 mg/kg i.p.); whereas, the alpha 1 selective compounds CL 218,872 and zolpidem were more potent at inhibiting [3H]Ro 15-1788 binding in the cerebellum (ID50 values 4.5 mg/kg and 10 mg/kg i.p.) compared to the spinal cord (ID50 values 12 mg/kg and > 30 mg/kg i.p.). Thus, the reduction of in vivo f[3H]Ro 15-1788 binding in tissues containing alpha 1 and alpha 2/alpha 3 receptor populations reflects the in vitro affinities of subtype selective compounds and should help to interpret the behavioral profile of such compounds.

Phenoxybenzamine partially inhibits alpha 2-adrenoceptors without affecting their presynaptic function.

The turnover of alpha-adrenoceptors was assessed by administering phenoxybenzamine (PBZ) intraperitoneally to rats in order to block the receptors irreversibly. The reappearance of the binding of [3H]prazosin, [3H]clonidine and [3H]rauwolscine in membranes from cerebral cortices was then measured. Maximum inhibition of binding occurred 3 hr after administration of phenoxybenzamine. The binding of [3H]prazosin was inhibited by 95% after administration of phenoxybenzamine (2 X 4 mg/kg, i.p.), and the half life (t1/2) for the alpha 1-adrenoceptor was 1.87 days. The "turnover" of binding for the alpha 2-adrenoceptor ligands ([3H]clonidine and [3H]rauwolscine) was similar: with doses of phenoxybenzamine up to 15 mg/kg (i.p.), the binding of both ligands was inhibited to a maximum of 30%. Maximum recovery occurred 3 days after treatment with phenoxybenzamine and the alpha 2-adrenoceptor has an apparent half life for recovery of 12 hr. Since only partial blockade of alpha 2-adrenoceptors was possible with phenoxybenzamine the possibility that these blocked sites included functional presynaptic autoreceptors was investigated. Clonidine (1 microM) attenuated K+-induced release of preloaded [3H]noradrenaline from cortical synaptosomes prepared from control rats by some 35%. Clonidine inhibited this release of [3H]noradrenaline to the same extent in synaptosomes prepared from rats treated with phenoxybenzamine 3 hr prior to sacrifice. This indicates that the alpha 2-adrenoceptors which are blocked by phenoxybenzamine are not part of the functional receptor population.

Evidence that porcine native 5-HT3 receptors do not contain nicotinic acetylcholine receptor subunits.

The present investigation utilised monoclonal antibodies directed against subunits of the nicotinic acetylcholine receptor in immunoblot and immunoprecipitation studies, which failed to demonstrate that the native 5-hydroxytryptamine3 (5-HT3) receptor complex purified from porcine brain contains the alpha1, alpha3, alpha4, alpha5, alpha7 or beta2 subunits of the nicotinic acetylcholine receptor.

Effect of desipramine, phenoxybenzamine and yohimbine on beta-adrenoceptors and cyclic AMP production in the rat brain.

In rat brain, the number of beta-adrenoceptors and activity of noradrenaline-dependent adenylate cyclase were examined after treatment with desipramine (7.5 mg kg-1 day-1) for three days alone or in combination with the alpha 2-adrenoceptor antagonist, yohimbine (2 mg kg-1 12 hr-1), or with phenoxybenzamine (7.5 mg kg-1 day-1), which is a more potent inhibitor of alpha 1 than alpha 2-adrenoceptors. The only treatment which significantly decreased the specific binding of the beta-adrenoceptor antagonist, [3H]dihydroalprenolol was the combination of desipramine with yohimbine. Desipramine alone and desipramine with yohimbine also significantly reduced the formation of cyclic AMP in response to incubation with noradrenaline, the response to the drug combination being accounted for by addition of the individual effects of the drugs. The results showed that decreases in the activity of noradrenaline-dependent adenylate cyclase could become apparent before decreases in beta-adrenoceptor numbers. Whether these rapid changes in noradrenaline-dependent adenylate cyclase or in numbers of beta-adrenoceptors which are produced by combination of desipramine with an alpha 2-adrenoceptor antagonist, are of therapeutic value remains to be elucidated.

[3H]L-655,708, a novel ligand selective for the benzodiazepine site of GABAA receptors which contain the alpha 5 subunit.

A compound (L-655,708) has been identified which has at least 50-fold selectivity for the benzodiazepine site on GABAA receptors containing an alpha 5 subunit over those containing an alpha 1, alpha 2, alpha 3 or alpha 6 subunit in combination with beta 3 and gamma 2. The compound was radiolabelled with tritium and investigated as a novel radioligand which recognizes the benzodiazepine site of GABAA receptors which contain the alpha 5 subunit. [3H]L-655,708 labels one saturable and specific population of binding sites in rat hippocampus with a Kd of 2.4 +/- 0.7 nM and a Bmax of 256 +/- 42 fmol/mg protein. The pharmacology of the binding site labelled was consistent with that of receptors present in cells transfected with alpha 5, beta 2 and gamma 2 and with receptors immunoprecipitated from rat brain with an alpha 5-selective antiserum. It is concluded that [3H]L-655,708 is the first radioligand to date which is selective for any BZ2 subtype of the GABAA receptor and should provide a valuable tool for elucidating the structure and function of the alpha 5-containing GABAA receptor subtype.

Differential subcellular distribution of the alpha 6 subunit versus the alpha 1 and beta 2/3 subunits of the GABAA/benzodiazepine receptor complex in granule cells of the cerebellar cortex.

The distribution of the alpha 6 subunit of the GABAA receptor has been established in rat cerebellum and compared to the distribution of the alpha 1 (cat) and the beta 2/3 (rat, cat) subunits, using immunocytochemistry. The synapses established by Golgi cell terminals on the dendrites of granule cells were immunoreactive for the alpha 6, alpha 1 and beta 2/3 subunits in virtually all glomeruli, indicating that two variants (alpha 1 and alpha 6) of the same subunit are co-localized at the same synapses. The somatic membranes of the granule cells, which receive no synapses, were immunopositive for the alpha 1 and beta 2/3 subunits, but not for the alpha 6 subunit. Thus, the alpha 1 and the beta 2/3 subunits are located at both synaptic and extrasynaptic sites, but the alpha 6 subunit is detectable only at synaptic sites.

Density and pharmacology of alpha5 subunit-containing GABA(A) receptors are preserved in hippocampus of Alzheimer's disease patients.

The anatomical localization and pharmacology of alpha5 subunit-containing GABA type-A receptors in the human hippocampal formation of Alzheimer's disease patients were studied with an alpha5 receptor selective ligand, [3H]L-655,708 and compared to age-matched human controls. Autoradiographic analyses revealed a heterogeneous distribution of [3H]L-655,708 binding sites in CA1-CA3 areas with high levels in stratum oriens, stratum pyramidale and stratum radiatum contrasting with low levels in stratum lacunosum. The highest quantity of alpha5 receptors was found in the molecular layer of the dentate gyrus. This pattern of expression was identical in both hippocampus from control and Alzheimer's disease subjects. Quantitative studies demonstrated that the number of [3H]L-655,708 binding sites is well preserved in Alzheimer's disease with only a moderate reduction (25-30%) in the CA1 subfield and entorhinal cortex. Furthermore, saturation and competition experiments with [3H]L-655,708 and representative benzodiazepine site ligands revealed that alpha5 receptors in Alzheimer's hippocampus have an alpha5beta2/3gamma2 pharmacology and structure as in control human brain.Overall, the data reported here provide evidence for a specific expression and relative sparing of alpha5 subunit-containing gamma-aminobutyric acid type-A receptors in the hippocampus of Alzheimer's patients.

Age-related changes in GABA(A) receptor subunit composition and function in rat auditory system.

A decline in the ability to discriminate speech from noise due to age-related hearing loss (presbycusis) may reflect impaired auditory information processing within the central nervous system. Presbycusis may result, in part, from functional loss of the inhibitory neurotransmitter GABA. The present study assessed age-related changes of the GABA(A) receptor in the inferior colliculus of young-adult, middle-aged, and aged rats related to: (i) receptor subunit composition and (ii) receptor function. Western blotting was used to measure protein levels of selected GABA(A) receptor subunits in preparations obtained from the inferior colliculus of Fischer 344 and Fischer 344/Brown-Norway F1 hybrid rats. In both strains, the aged group exhibited significant increases in gamma1 subunit protein and a decrease in alpha1 subunit protein. To examine the functional consequence of this putative age-related subunit change, we measured the ability of exogenous GABA to flux/translocate chloride ions into microsac preparations derived from Fischer 344 inferior colliculus. GABA-mediated chloride influx was significantly increased in samples prepared from the inferior colliculus of aged animals. Together with previous studies, these results strongly suggest an age-related change in GABA(A) receptor composition. These changes may reflect a compensatory up-regulation of inhibitory function in the face of significant loss of presynaptic GABA release. These findings provide one example of plastic neurotransmitter receptor changes which can occur during the ageing process.

Proliferative effects of cholecystokinin in GH3 pituitary cells mediated by CCK2 receptors and potentiated by insulin.

1. Proliferative effects of CCK peptides have been examined in rat anterior pituitary GH3 cells, which express CCK2 receptors. 2. CCK-8s, gastrin(1-17) and its glycine-extended precursor G(1-17)-Gly, previously reported to cause proliferation via putative novel sites on AR4-2J and Swiss 3T3 cells, elicited significant dose dependent increases of similar magnitude in [3H]thymidine incorporation over 3 days in serum-free medium of 39 +/- 10% (P < 0.01, n = 20), 37 +/- 8% (P < 0.01, n = 27) and 41 +/- 6% (P < 0.01, n = 36) respectively. 3. CCK-8s and gastrin potentially stimulated mitogenesis (EC50 values 0.12 nM and 3.0 nM respectively), whilst G-Gly displayed similar efficacy but markedly lower potency. L-365,260 consistently blocked each peptide. The CCK2 receptor affinity of G-Gly in GH3 cells was 1.09 microM (1.01;1.17, n = 6) and 5.53 microM (3.71;5.99, n = 4) in guinea-pig cortex. 4. 1 microM G-Gly weakly stimulated Ca2+ increase, eliciting a 104 +/- 21% increase over basal Ca2+ levels, and was blocked by 1 microM L-365,260 whilst CCK-8s (100 nM) produced a much larger Ca2+ response (331 +/- 14%). 5. Insulin dose dependently enhanced proliferative effects of CCK-8s with a maximal leftwards shift of the CCK-8s curve at 100 ng ml(-1) (17 nM) (EC50 decreased 500 fold, from 0.1 nM to 0.2 pM; P < 0.0001). 10 microg ml(-1) insulin was supramaximal reducing the EC50 to 5 pM (P = 0.027) whilst 1 ng ml(-1) insulin was ineffective. Insulin weakly displaced [125I]BHCCK binding to GH3 CCK2 receptors (IC50 3.6 microM). 6. Results are consistent with mediation of G-Gly effects via CCK2 receptors in GH3 cells and reinforce the role of CCK2 receptors in control of cell growth. Effects of insulin in enhancing CCK proliferative potency may suggest that CCK2 and insulin receptors converge on common intracellular targets and indicates that mitogenic stimuli are influenced by the combination of extracellular factors present.

Heterogeneity between soluble human and rabbit splenic alpha 2-adrenoceptors.

The pharmacological and biochemical characteristics of soluble alpha 2-adrenoceptors were investigated to determine whether differences observed in membranes were maintained in solution and to probe the nature of any such differences. alpha 2-Adrenoceptors were solubilized from purified plasma membrane preparations of human and rabbit spleen using digitonin. [3H]yohimbine bound to one population of alpha 2-adrenoceptors in the preparations with dissociation constants of 2.4 nM and 7.8 nM respectively. The pharmacological profile of the alpha 2-adrenoceptors has been examined. Upon solubilization the affinity of the alpha 2-adrenoceptors for yohimbine was unchanged. In contrast, the potency of idazoxan and RX 811066 were increased, whereas the potency for prazosin (human only), phentolamine and WY 26392 was decreased 2-3-fold. The potency of the agonists oxymetazoline, UK 14304 and adrenaline were all reduced upon solubilization of alpha 2-adrenoceptors. The selectivity of yohimbine, idazoxan, RX 811066 and WY 26392 for human rather than rabbit alpha 2-adrenoceptors was maintained in solution. Possible sources of heterogeneity between human and rabbit alpha 2-adrenoceptors were investigated. The protein structure was probed by comparing the susceptibility of the receptors to inactivation by sulphydryl modifying agents. No differences were observed in the potency of N-ethylmaleimide or p-chloromercuribenzoate to inactivate the receptor. The carbohydrate component of the receptors was investigated using agarose-linked lectins. Rabbit splenic alpha 2-adrenoceptors had a lower affinity for the lectins wheatgerm agglutinin (Triticum vulgaris) and soybean (Glycine max) which bind the sugars N-acetyl d-glucosamine and N-acetyl d-galactosamine respectively. These findings suggest that heterogeneity of the alpha 2-adrenoceptor derives from its structural characteristics rather than its environment in the membrane.

Tyrosine kinase phosphorylation of GABAA receptors.

Phosphorylation of purified bovine brain GABAA receptors by the tyrosine kinase, pp60v-src was examined. pp60v-src phosphorylated two bands of 54-62 kDa and 48-51 kDa that migrated to approximately the same position as bands recognized by antisera against the beta 2 and gamma 2 GABAA receptor subunits, respectively. Bacterially expressed proteins containing the putative large cytoplasmic loops of the beta 1 and gamma 2L subunits were phosphorylated by pp60v-src, indicating that the phosphorylation sites are located in these subunit domains. The tyrosine kinase inhibitors, genistein and the tyrphostins B-42 and B-44, inhibited muscimol-stimulated 36Cl- uptake in mouse brain membrane vesicles (microsacs). magnitude of the tyrphostin B-44-induced inhibition of muscimol-stimulated 36Cl- uptake was significantly reduced in microsacs that were lysed and resealed under conditions that inhibit phosphorylation. GABA-gated Cl- currents were also inhibited by genistein and tyrphostin B-44 in Xenopus oocytes expressing alpha 1 beta 1 and alpha 1 beta 1 gamma 2L subunits. Consequently, protein tyrosine kinase-dependent phosphorylation appears to be another mechanism of regulating the function of GABAA receptors.

Molecular and functional diversity of the expanding GABA-A receptor gene family.

Fast inhibitory neurotransmission in the mammalian CNS is mediated primarily by the neurotransmitter gamma-aminobutyric acid (GABA), which, upon binding to its receptor, leads to opening of the intrinsic ion channel, allowing chloride to enter the cell. Over the past 10 years it has become clear that a family of GABA-A receptor subtypes exists, generated through the coassembly of polypeptides selected from alpha 1-alpha 6, beta 1-beta 3, gamma 1-gamma 3, delta, epsilon, and pie to form what is most likely a pentomeric macromolecule. The gene transcripts, and indeed the polypeptides, show distinct patterns of temporal and spatial expression, such that the GABA-A receptor subtypes have a defined localization that presumably reflects their physiological role. A picture is beginning to emerge of the properties conferred to receptor subtypes by the different subunits; these include different functional properties, differential modulation by protein kinases, and the targeting to different membrane compartments. These properties presumably underlie the different physiological roles of the various receptor subtypes. Recently we have identified a further member of the GABA-A receptor gene family, which we have termed theta, which appears to be most closely related to the beta subunits. The structure, function, and distribution of theta-containing receptors, and receptors containing the recently reported epsilon subunit, are described.

Localization of 5-HT3 receptors in the rat spinal cord: immunohistochemistry and in situ hybridization.

Specific antibodies raised against a fusion protein containing the amino acid sequence of the putative second intracellular loop of the cloned 5-HT3-A receptor subunit were used for the immunohistochemical visualization of 5-HT3 receptors in the rat spinal cord. A dense 5-HT3-like immunoreactivity was found in the superficial layers of the dorsal horn, which closely matched the labelling of 5-HT3 binding sites by [125I]iodo-zacopride. This immunostaining was markedly decreased following unilateral rhizotomy, consistently with a preferential location of 5-HT3 receptors on terminals of primary afferent fibres, and with the presence of 5-HT3 mRNA in dorsal root ganglia. However, a significant proportion of 5-HT3 receptors persisted after rhizotomy, and the corresponding mRNA was found in the dorsal horn of the spinal cord. 5-HT3 receptors are therefore also located on intrinsic neurones of the spinal cord.