Inhibition of hydrogen peroxide-induced apoptosis but not arachidonic acid release in GH3 cell by EGF

Reactive oxygen species (ROS) and arachidonic acid (AA) can both function as extra- and intra-cellular messengers to regulate various cell functions including cell death. The effect of ROS on phospholipase A₂ (PLA₂) activity and/or AA release has not been extensively studied in neuronal cells. In this study, we investigated the effects of H₂O₂ on AA release and apoptosis in GH3 cells, a clonal strain from rat anterior pituitary. Incubation with H₂O₂ for 1 h stimulated [³H]AA release in a concentration-dependent manner from prelabeled GH3 cells. [³H]AA release was inhibited by arachidonyl trifluoromethyl ketone, a specific inhibitor of cytosolic PLA₂, and cytosolic PLA₂ protein with a molecular mass of 100 kDa was detected by immunoblotting. Culture with 0.2 mM H₂O₂ and 30 μM AA for 24 h induced lactate dehydrogenase (LDH) leakage, DNA laddering and DNA fragmentation in GH3 cells. In GH3 cells pretreated with EGF (50 ng/ml) for 24 h, LDH leakage and DNA fragmentation by H₂O₂ and AA were inhibited, although H₂O₂-induced [³H]AA release was not modified. Mastoparan, a wasp venom peptide, induced [³H]AA release and cell death in GH3 cells. Neither effect of mastoparan was inhibited by EGF treatment. These findings suggest that (1) H₂O₂ stimulates AA release via activation of cytosolic PLA₂, (2) H₂O₂ and AA induce apoptotic death of GH3 cells and (3) treatment with EGF protects H₂O₂- and AA-, but not mastoparan-, induced GH3 cell death.     

Dopamine metabolism in characterised neurones ofPlanorbis corneus

A sensitive chromatographic procedure was used to study the metabolism of [¹⁴C]tyrosine, [³H]DOPA and [³H]dopamine in 3 defined cell-types situated in the nervous system ofPlanorbis corneus. One of the cell-types contains dopamine (GDC), the other serotonin (GSC) and the other neither amine (GC). The GDCs metabolise [¹⁴C]tyrosine to form DOPA and dopamine while the other two cells lack this ability. In contrast, the GDCs and the GSC, but not the GCs, metabolise [³H]DOPA to form dopamine. In addition the GDCs incorporate radioactivity from [³H]DOPA into DOPAC, homovanillic acid and methoxytyramine. After incubation of cells in [³H]dopamine, only the GDCs metabolise it to form DOPAC, homovanillic acid and methoxytyramine. In no instance did the GDCs form significant amounts of noradrenaline from the incorporated radioactive substances.

These results, together with data on the amine histochemistry of the individual cell-types following pretreatment of animals with drugs known to affect specific enzymes in the synthesis of amine transmitter substances, clearly demonstrate that the GDCs alone have the enzymes requisite for the biosynthesis and catabolism of dopamine, but not noradrenaline.     

Evidence for defective incorporation of proteins in myelin of the quaking mutant mouse

The defect in myelinogenesis present in the Quaking mutant mouse was investigated using a double radioisotope technique for comparing the incorporation of amino acid into myelin proteins of normal and mutant mice. Quaking mice and littermate controls recieved intracranial injections of 150 μCi [³H]glycine and 25 μCi of [¹⁴C]glycine respectively. After 2 h their brains were combined and jointly processed to obtain subcellular fractions. The ³H/¹⁴C ratio for the myelin subfraction was 1.88 as compared to a ³H/¹⁴C ratio of 3.0 for the other subfractions, indicating a 40% decrease in glycine incorporation into myelin of Quaking mice. Myelin proteins were separated by discontinuous gel electrophoresis in the presence of sodium dodecyl sulfate (SDS) and the ³H/¹⁴C ratios determined in each gel slice. In contrast to the microsomal subfractions which gave a ³H/¹⁴C ratio of 2.6 across the gel, the ³H/¹⁴C ratio of myelin showed large variations with values ranging from 0.54 for proteolipid protein to 2.0 for some of the high molecular weight proteins. During development, the Quaking mutant exhibited a preferential depression in glycine incorporation into proteolipid protein in 18-day-old mice, while in older animals (32–54 days) the fast migrating basic protein, as well as the proteolipid protein, was labeled to a significantly lesser extent.     

Transport, uptake, and metabolism of blood-borne vasopressin by the blood-brain barrier

Transport, binding, and metabolism of [phenylalanyl-3,4,5-³H(N)]arginine vasopressin (AVP) by the blood-brain barrier (BBB) was studied in adult guinea-pigs by means of a novel vascular brain perfusion (VBP)/capillary depletion technique and HPLC. A time-dependent, progressive brain uptake of ³H-radioactivity was measured over the 10 min period of VBP both in brain homogenates and in brain tissue depleted of cerebral microvessels. The unidirectional blood-to-brain transport constant, K_IN, estimated by multiple-time tissue uptake analysis of the homogenate and postcapillary supernatant, indicated that the BBB transfer rat ffor [³H]AVP (K_IN = 2.37±0.25 μl min⁻¹ per gram brain homogenate) was almost 10 times higher than for simultaneously perfused [¹⁴C]sucrose, a cerebrosvascular space marker. In contrast to homogenate and postcapillary supernatant, the [³H]radioactivity determined in the vascular pellet after dextran density centrifugation of the brain homogenate was very low and only somewhat higher than for [¹⁴C]sucrose. HPLC analysis of the perfused brain tissue revealed time-dependent degradation of the blood-borne neuropeptide. The percentage of intact [³H]AVP as determined in the postcapillary supernatant progressively declined during brain perfusion, from 49% at 1 min to 11.9% at 10 min. The major detectable labeled metabolite was [³H]phenylalanine, the labeled amino acid residue of [³H]AVP. The aminopeptidase inhibitor bestatin (0.5 mM), perfused simultaneously with [³H]AVP by the VBP technique, did not alter tissue uptake of [³H]AVP, indicating that there was no significant hydrolysis of peptide by the luminal BBB surface. The results suggest that rapid in vivo metabolism of AVP occurs after BBB transport in the brain parenchyma with no evidence of significant capillary sequestration, or degradation of AVP by the BBB.     

Pirenzepine-insensitive muscarinic autoreceptors regulate acetylcholine release in human neocortex

The release of [³H]acetylcholine ([³H]ACh) and its modulation mediated by autoreceptors were investigated in synaptosomes prepared from fresh human cerebral cortex prelabelled with [³H]choline ([³H]Ch) and depolarized in superfusion with 15 mM KCl. The K⁺-evoked release of tritium was almost totally accounted for by unmetabolized [³H]ACh and was largely calcium-dependent. Exogenous ACh decreased the depolarization-evoked release of [³H]ACh in a concentration-dependent manner (EC₅₀ = 1.5 μM). The inhibitory effect of ACh on [³H]ACh release was counteracted by the non-selective muscarinic antagonist atropine. In contrast, the selective M₁ receptor antagonist pirenzepine was ineffective. It is concluded that muscarinic autoreceptors regulating the release of ACh are present on cholinergic nerve terminals of human cerebral cortex and appear to belong to a pirenzepine-insensitive subtype.     

Antithrombin III modulates the effect of thrombin on the metabolism of glycosaminoglycans in cultured endothelial cells

We previously reported that a treatment of cultures of endothelial cells from bovine aorta with thrombin resulted in a less accumulation of glycosaminoglycans (GAG) in the cell layer. In the present study, we found that thrombin-induced decrease in the accumulation of [³⁵s]sulfate-labeled GAG (³⁵S-GAG) such as heparan sulfate was prevented by antithrombin III (AT III) but not by heparin cofactor II (HC II). However, AT III did not show a significant effect on the ³⁵S-GAG accumulation individually. Pretreatment of the cell layer with neither AT III not HC II showed any preventive effect. When GAG in the cell layer was labeled with both [³⁵s]sulfate and [³⁵h] glucosamine, neither thrombin nor a combination of thrombin with AT III changed the ratio of the radioactivity of ³⁵S to that of ³H. Although thrombin stimulated the release of ³⁵S-GAG from the cell layer, AT III completely prevented the stimulatory effect. In conclusion, it was suggested that AT III may inhibit the thrombin action on GAG metabolism of endothelial cells to prevent thrombosis in vivo.     

Alterations in neurotransmitter receptors and glucose use after unilateral orbital enucleation

Serotonin (5-hydroxytryptamine; 5-HT), acetylcholine and γ-aminobutyric acid (GABA) are neurotransmitters in the rat visual system. Using quantitative autoradiography, the effect of unilateral orbitral enucleationon [^3H]5-HT, [³H]ketanserin, [³H]quinuclidinyl benilate (QNB) and [³H]muscimol binding to 5-HT₁, 5-HT₂, muscarinic and GABA_A receptors has been examined within anatomical components of the visual pathway at 4 time points up to 20 days after the lesion. The functional deficit was assessed in the same animals using quantitative [¹⁴C]2-deoxyglucose autoradiography. At 1 day after unilateral orbital enucleation, there were no significant alterations in ligand binding although local cerebral glucose use was reduced in primary visual structures in the visually deprived hemisphere. At 5 days post-enucleation, however, [³H]5-HT binding was significantly reduced in both the visually deprived superior colliculus (by 17%) and dorsal lateral geniculate body (DLG) (by 33%). There were similar alterations in the binding of this ligand in these primary retinal projections areas at 10 and 20 days after orbital enucleation, but there were no changes in secondary areas (e.g. visual cortex) at any time point. [³H]Muscimol binding was significantly reduced in the visually deprived DLG (30%) and visual cortex (21%) only at 20 days post-lesion, whilst [³H]ketanserin and [³H]QNB were not altered in any region in the visually deprived hemisphere at any time point post-enucleation. At 10 and 20 days post-enucleation, the degree of [³H]5-HT, and [³H]muscimol binding deficits in visually deprived structures correlated significantly with the level of reduced metabolic activity in these areas (r = 0.700andr = 0.543respectively). The specificity and regional and temporal heterogeneity of neurotransmitter receptor binding alterations provides evidence of selective adjustments within visual system component6s in response to orbital enucleation.     

Binding of [125I]insulin to specific receptors and stimulation of nucleotide incorporation in cells cultured from rat brain

The occurrence of insuling receptors and biological responses to insulin has been investigated in trypsin-dissociated fetal rat brain cells maintained in culture for 8 days. Binding of [¹²⁵]insulin to brain cells in culture was time- and pH-dependent and 85–90% specific. Porcine insulin competed for [¹²⁵]insulin binding in a dose-dependent manner. Unrelated polypeptides, including angiotensin II, glucagon, bovine growth hormone, and bovine prolactin did not compete for [¹²⁵]insulin binding. The half-life of [¹²⁵]insulin dissociation from receptors at 24°C was 15 min and a plot of ln[B/Bo] vs time suggested two dissociation rate constants of2.7 × 10⁻⁴ sec⁻¹ and5.0 × 10⁻⁵ sec⁻¹. Scatchard analysis of the binding data gave a curvelinear plot which may indicate negative cooperativity or the occurrence of both high affinity(K_a = 2 × 10¹¹M⁻¹) and low affinity(K_a = 4 × 10¹⁰M⁻¹) sites. Of the estimated total of 4.9 × 10⁴ binding sites per cell, 28–30% appear to be high affinity sites.

Incubation of cultures with insuling caused a time- and dose-dependent stimulation of [³H]thymidine and [³H]uridine incorporation into TCA-precipitable material. Maximum stimulation of thymidine incorporation (2–5-fold) occured 11 h after incubation with 167 nM insulin. The same concentration of insulin caused a 2.2-fold increase in [³H]uridine incorporation in 2 h. These results indicate that cells cultured from rat brain contain specific insulin receptors capable of mediating effects of insulin on macromolecular synthesis in the central nervous system.     

Regulation of peripheral-type benzodiazepine receptors in cultured astrocytes by monoamine and amino acid neurotransmitters

Exposure of primary cultured astrocytes for 3 days to 1 μM of either dopamine, serotonin or norepinephrine resulted in upregulation (25–34% increase in B_max) of the peripheral-type benzodiazepine receptors (PBRs) labeled with [³H]Ro5-4864. A similar treatment with γ-aminobutyric acid [GABA] caused a 2-fold increase in the affinity (K_d) of [³H]Ro5-4864. The monoamines tested and GABA had no effect on the binding parameters of [³H]PK 11195, another selective PBR ligand. The present study indicates that Ro5-4864 binding sites are susceptible to regulation by specific neurotransmitters and provides further evidence for the distinction between Ro5-4864 and PK 11195 binding sites of the PBRs in cultured astrocytes.     

The differentiation of synthesis from incorporation of basic protein in quaking mutant mouse myelin

The incorporation of radioactive label into the myelin basic protein isolated from whole brain and from purified myelin of Quaking mice and normal littermates was compared. Four Quaking mice (32 days) and 4 littermate controls were injected intracranially with 150 μCi [2-²H]glycine and 25 μCi of [2-¹⁴C]glycine, respectively. One hour later, the 8 mice were sacrificed and their brains combined for common homogenization. The ³H/¹⁴C ratios of the small and large basic proteins in whole brain were 3.44 and 2.48 respectively, while the ³H/¹⁴C ratios for these proteins in myelin were 0.79 and 1.00, respectively. In the same experiment, the microsomal fraction had a ³H/¹⁴C ratio of 2.98 which is the expected ratio for normal incorporation. The results indicate that the synthesis of basic protein in whole brain of Quaking mouse proceeds at a normal rate, but specifically, the incorporation of basic protein into myelin is depressed suggesting a defect at the step of assembly of myelin components into a final membrane product.     

Functional damage of dopamine nerve terminals following intrastriatal kainic acid injection

The release of [³H] dopamine ([³H]DA) previously taken up into rat striatal slices was studied one week after a monolateral intrastriatal injectuion of kainic acid (KA). Different releasing stimuli (electrical pulses, veratrine, high-K⁺) were applied. The electrically evoked release in the KA-lesioned striata was drastically reduced with respect to the unlesioned contralateral striata. In contrast, KA had no effect on the release of [³H]DA evoked by veratrine or high-K⁺. In unlesioned striatal slices, depolarized with 15 mM KCl, apomorphine reduced and (−)sulpiride increased the release of [³H]DA. The effect of apomorphine was antagonized by (−)sulpiride indicating the presence of an autoreceptor system similar to that seen in unlesioned striata stimulated electrically. However, the effects of apomorphine and of (−)sulpiride were dramatically reduced in K⁺-depolarized slices prepared from KA-lesioned striata. The results suggest that the axon terminals in KA-treated areas remain intact in several of their properties but may be damaged in some critical processes.     

Amino acid transport in cerebral microvessels during plasmodium yoelii infection in mice

Plasmodium yoelii infected cerebral microvessels of mice had an enhanced time-dependent, temperature-sensitive, and saturable uptake of [¹⁴C]-amino acid. viz. leucine, valine and glycine. Metabolic inhibitors caused a noticeable inhibition of amino acid uptake in normal microvessels as compared to infected cerebral microvessels indicating that the uptake of [¹⁴C]-L-leucine, [¹⁴C]-L-valine and [¹⁴C]-glycine is an energy dependent process.     

Comparison of two different benzodiazepine binding proteins by peptide mapping after limited proteolysis

Proteins P₅₁ and P₅₅ were photolabeled by [³H]flunitrazepam, [³H]clonazepam or ³H-Ro 15–4513 and then compared by peptide mapping after limited digestion with various proteases. Results indicated that [³H]flunitrazepam or [³H]clonazepam irreversibly bind to the same and [³H]Ro 15–4513 to a different part of these proteins. The different radiolabeled peptide patterns obtained from P₅₁ or P₅₅ irrespective of the photolabel used, suggest a difference in the molecular structure of these proteins.     

Estimation of EDRF and nitric oxide release using [3H]GTP-labeled human platelets

We have developed a new bioassay for endothelium-derived relaxing factor (EDRF) or nitric oxide (NO) using human [³H]guanosine triphosphate (GTP)-labeled platelets. The labeled platelets were preincubated with isobutyl-methylxanthine and co-cultured with endothelial cells and the [³H]cyclic guanosine monophosphate (cGMP) formed was isolated by ion-exchange chromatography. Endothelial cells, either in monolayer cells or in suspension, increased platelet cGMP accumulation dose-dependently, a significant increase being detected with 5,000 endothelial cells or more/assay when suspended cells were used. Co-culturing with the same number of skin fibroblasts failed to elevate platelet cGMP. Preincubation of endothelial cells with bradykinin and superoxide dismutase (SOD) synergistically potentiated the increase in platelet cGMP, but was attenuated by N^ω-nitro-L-arginine, with partial restoration by L-arginine but not by D-arginine. These compounds, however, did not affect cGMP accumulation by sodium nitroprusside. Moreover, preincubation of the labeled platelets with the NO synthase inhibitor prior to EDRF assay had no effect. We conclude that [³H]GTP-labeled platelets could provide a simple, sensitive and specific bioassay for estimating EDRF or NO release.     

Inositol 1,4,5-trisphosphate metabolism in the cerebella of Lurcher mutant mice and patients with olivopontocerebellar atrophy

We have investigated inositol 1,4,5-trisphosphate (InsP₃) metabolism in cerebellar membranes of normal humans and patients with dominant ataxia (‘C’ kindred), and also in cerebellar microsomes of Lurcher mutant mouse (a suggested model for cerebellar ataxia). Various [³H]InsP₃ metabolites formed separated by HPLC using 3 successive convex gradients of 1.7 M ammonium formate, pH 3.7. [³H]InsP₃ metabolism was rapid and in 15- and 45-day-old control mice cerebella about 50% of [³H]InsP₃ was metabolized within 20 s. In 15-day-old Lurcher mice the rate of [³H]InsP₃ metabolism was significantly low (40% of normal). [³H]InsP₃ metabolism was almost absent in 45-day-old Lurcher mice cerebellar microsomes. The decreased [³H]InsP₃ metabolism was consistent with decreased recovery of the various inositol polyphosphates formed. Similarly, in cerebellar membranes of human patients with olivopontocerebellar (OPCA) a significant decrease in [³H]InsP₃ metabolism was observed when compared with normal controls. These data suggest that altered phosphoinositide turnover may be associated with the onset of neuronal degeneration in human OPCA.     

Differences in uptake kinetics of cis-3-aminocyclohexane carboxylic acid into neurons and astrocytes in primary cultures

The uptake of [³H]ACHC and [³H]GABA into cultured neurons and astrocytes was studied. [³H]ACHC uptake was less efficient than that of GABA in both cell types and K_m values for ACHC uptake into neurons and astrocytes were 40.3 μM and 210.8 μM, respectively. The corresponding V_max values were 0.321 and 0.405 nmol·min⁻¹·mg⁻¹ cell protein, respectively. Kinetic studies of the effects of GABA on ACHC uptake and vice versa showed that GABA is a linear competitive inhibitor of ACHC uptake in both cell types with a K_i value of 15 μM. On the other hand, ACHC turned out to be a complex inhibitor of astrocytic GABA uptake being competitive at lower concentrations and non-competitive at higher concentrations. ACHC inhibited GABA uptake into neurons competitively with a K_i of 69 μM. It is concluded that ACHC acts primarily on neuronal GABA uptake sites but its uptake is much more complicated than hitherto anticipated.     

[H]Nipecotic acid binding to GABA uptake sites in human brain

The binding of [³H]nipecotic acid to frozen post-mortem human brain tissue has been characterized. Competition experiments with γ-aminobutyric acid (GABA), GABA uptake inhibitors, ligands active at post-synaptic GABA receptors and receptors for other neurotransmitter systems, suggest that [³H]nipecotic acid binds to the neuronal (but not glial) GABA uptake site. Competition and kinetic experiments suggest that 85% of the binding is to high affinity site. The dissaciation constants (K_d) measured in kinetic and equilibrium experiments were in the same range (0.5–0.6 μM). The regional distribution was studied in 19 brain regions and the binding was relatively homogenous. It is concluded that [³H]nipecotic acid binding can be used as a marker for neuronal GABA uptake sites in post-mortem human brain tissue.     

Selective alterations in glutamate receptor subtypes after unilateral orbital enucleation

Glutamate is the major excitatory neurotransmitter in the rat visual system. Using quantitative autoradiography the effect of unilateral orbital enucleation on [³H]kainate, [³H]-amino-3-hydroxy-5-methyl-isoxazole-4-propionic acid ([³H]AMPA) and [³H]glutamate binding to kainate, quisqualate and NMDA receptors respectively has been examined within anatomical components of the visual pathway at 4 time points up to 20 days post-lesion. The time course for the degeneration of retinal projection fibres was assessed in a separate group of animals by quantifying [³H]cyclohexyladenosine ([³H]CHA) binding to presynaptic adenosine A₁ receptors. Over the first 5 days after orbital enucleation, there were no significant alterations in glutamate or adenosine A₁ receptor binding in visual structures of the visually deprived hemisphere. However, at 10 days post-lesion [³H]AMPA binding was significantly reduced (30%) in the visually deprived superior colliculus but unaltered in other visual structures. At this time point there was also a significant reduction (50%) in [³H]CHA binding in the visually deprived superior colliculus but not in other retino-recipient nuclei. There were similar changes in [³H]AMPA and [³H]CHA binding at 20 days post-enucleation. [³H]Kainate binding was significantly increased in the visually deprived superior colliculus only at 20 days post-enucleation. Saturation analysis of [³H]kainate and [³H]AMPA binding at this time point indicated a selective increase in the b_max value for the high affinity [³H]kainate binding site and a concomitant decrease in the b_max value for the high affinity [³H]AMPA binding site in the visually deprived superior colliculus. There were, however, no significant alterations in [³H]AMPA or [³H]kainate binding in other primary projection areas or in secondary visual areas (e.g. visual cortex) at any time point. NMDA sensitive [³H]glutamate binding was unaltered in the visually deprived hemisphere up to 20 days post-enucleation. These results suggest an upregulation of kainate receptors in the visually deprived superior colliculus after orbital enucleation and a loss of presynaptic quisqualate receptors on degenerating retinal fibres. The plastic alterations in kainate receptors in the superior colliculus are supportive of electrophysiological data suggesting a physiological role for these sites in mediating excitatory postsynaptic potentials in tectal neurons.     

Glucose metabolism in cholinoceptive cortical ratbrain regions after basal forebrain cholinergic lesion

To address the question whether the changes in cortical glucose metabolism observedin patients with Alzheimers disease are interrelated with, or consequences of, basal forebraincholinergic cell loss, an experimental approach was employed to produce cortical cholinergicdysfunction in rat brain by administration of the cholinergic immunotoxin 192IgG-saporin. [¹⁴C] -glucose utilization in brain homogenates, -glucose-displaceable [³H]cytochalasin B binding to glucose transporters (GLUT),Northern and Western analyses, as well as in vivo [¹⁴]2-deoxyglucoseautoradiography were used to quantify the regional glucose metabolism.

Basal forebrain cholinergic lesion resulted in transient increases in glucose transporterbinding in cortical regions displaying reduced acetylcholinesterase activity, already detectableseven days after lesion with peak values around 30 days post lesion. Western analysis revealedthat the changes in total glucose transporter binding are mainly due to changes in the GLUT3subtype only, while the levels of GLUT1 and GLUT3 mRNA (Northern analysis) were notaffected by cholinergic lesion. Both immunocytochemistry and in situ hybridizationdemonstrated preferential localizations of GLUT1 on brain capillaries and GLUT3 on neurons,respectively. A lesion-induced transient decrease in [¹⁴C] -glucoseutilization seven days post lesion was detected in the lesion site, whereas cholinoceptive corticalregions were not affected. In vivo [¹⁴C]deoxyglucose uptake was transientlyincreased in cholinoceptive cortical regions and in the lesion site being highest between three toseven days after lesion.

The cholinergic lesion-induced transient up-regulation of cortical glucose transporters anddeoxyglucose uptake reflects an increased glucose demand in regions depleted by acetylcholinesuggesting functional links between cortical cholinergic activity and glucose metabolism incholinoceptive target regions.     

Tracing specific transmitter pathways in the rat CNS: combination of [H]serotonin retrograde labelling with immunocytochemical detection of endogenous transmitters

Selective retrograde labelling with [³H]serotonin ([³H]5-HT) can be used to identify serotonergic cell bodies after specific [³H]5-HT uptake by the corresponding nerve terminals. In the present study, we demonstrate that autoradiography of this [³H]5-HT radiolabelling can be combined with immunocytochemical detection of endogenous serotonin, GABA or substance P on the same tissue section. The midbrain raphe serotonergic projections to the olfactory bulb and the spinal projections of medullary serotonergic nuclei were investigated. The specificity of retrograde labelling with [³H]5-HT was confirmed by immunoreactivity of the radiolabelled cells for serotonin, using an antiserum specific for formaldehyde-fixed serotonin. After spinal injections of [³H]5-HT, many retrogradely labelled cells in the medullary raphe were immunopositive for substance P, and a few for GABA. These results are in agreement with the available information on the co-existence of putative transmitters in the spinal projections of caudal raphe neurons. Therefore, autoradiography of [³H]5-HT retrograde labelling combined with immunocytochemistry offers a possibility to test the specificity of transmitter-selective retrograde labelling, to identify transmitter-defined neuronal interactions and to investigate the projection fields of multitransmitter containing neurons.