Monoamine oxidase activity in brain microvessels determined using natural and artificial substrates: relevance to the blood-brain barrier.

The possible contribution of cerebrovascular monoamine oxidase (MAO) to the blood-brain barrier to catecholamines was studied in isolated porcine and rat microvessels by determining its activity with various substrates. Michaelis-Menten kinetic constants, Km and Vmax, were determined using noradrenaline (NA) as substrate in a Tris medium. Km values were 0.25 +/- 0.05 mM in control and 0.16 +/- 0.09 mM in ultrasonically disintegrated (USD) preparations (difference not significant); Vmax in USD preparations (1.83 +/- 0.20 n.atoms O2 min-1 mg protein-1) was slightly higher (p less than 0.05) than in control preparations (1.35 +/- 0.11 n.atoms O2 min-1 mg protein-1), suggesting a certain restriction by the plasma membrane of substrate access to the enzyme. This phenomenon was confirmed in a more physiological, ionic medium; the activity was then approximately doubled for 1 mM NA, whereas that for 1 mM beta-phenylethylamine (beta-PEA), a lipid-soluble substrate, tended to decrease with USD treatment. These results show that this highly active form of MAO is unlikely to be saturated by physiological concentrations of catecholamine. It can be estimated that, for a plasma concentration of NA of 1 microM, a facilitated diffusion accelerating the entry of the catecholamine into the cells by at least 15-fold would be necessary in order to exceed the catabolic capacity of MAO. It is concluded that circulating catecholamines are not likely to cross the endothelial barrier of cerebral microvessels intact, and that the small quantities of radioactivity detected in the parenchyma in measurements of the brain uptake index essentially represent metabolites due to MAO activity.

Monoamine oxidase activity in the cerebral vasculature: comparison between fresh microvessels from different structures and cell cultures derived from microvessels.

Monoamine oxidase (MAO) activity was studied in various preparations of porcine brain microvessels to explore further the role of this enzyme in the blood-brain barrier to catecholamines. No difference was noted (Vm and Km) between microvessels isolated from three structures (caudate nucleus, thalamus, and cerebral cortex) in which the responses to circulating catecholamines in vivo are markedly different. Large and small microvessels from the caudate nucleus and the thalamus presented the same specific activity. Cell cultures obtained from small microvessels were rich in endothelial cells as identified by the presence of Factor VIII-related antigen. These preparations displayed an MAO activity about ninefold less than freshly isolated microvessels, although their prostaglandin synthetase activity appeared normal. These results suggest that MAO activity is not the main factor determining the regional differences in the cerebrovascular reactions to catecholamines, that MAO is not specifically localized in the endothelium but must be also present in the smooth muscle, and that the MAO activity is greatly decreased during cell culture.

Microvessels isolated from brain: localization of muscarinic sites by radioligand binding and immunofluorescent techniques.

The present investigation was carried out to determine the extent to which muscarinic acetylcholine receptors (mAChRs) in vascular and perivascular structures were colocalized with glial fibrillary acidic protein (GFAP)-positive structures. To this aim, an immunocytochemical approach on free-floating cryosections and isolated microvessels obtained from rat brain was performed to study the possible colocalization of immunostaining with the anti-mAChR protein antibody (M35) and an anti-GFAP antibody. Double-labeling experiments were carried out by fluorescent techniques. Confocal microscopic observations of GFAP and M35 immunoreactivities on free-floating sections showed a high degree of colocalization on astrocyte processes associated with large vessels or capillaries. This pattern suggests that muscarinic receptors are associated with astrocytic endfeet. Confocal microscopic observations of immunoreactivity from isolated cerebral microvessels strengthen this conclusion since double-labeling of M35 and GFAP showed that perivascular astrocytic structures remained attached to the isolated microvessels and were present on vascular segments showing M35 immunoreactivity. In another set of experiments, the specific binding of [3H]quinuclidinylbenzylate ([3H]QNB) to isolated microvessel membrane preparations from cerebral cortex, caudate nucleus, thalamus, and cerebellum showed that a constant binding yield (20% in bovine and 40% in rat) was observed for microvessels compared with the corresponding brain region. According to our immunocytochemical results, the astrocytic membrane remaining attached to microvessels may account for the majority of the muscarinic binding to isolated microvessels. [3H]QNB binding values found in isolated microvessels cannot therefore be considered as artifacts without any link with vascular function. Taken together, the present study strengthens the idea that the muscarinic receptors may be implicated in the functional relationship between glial and vascular structures.

Regional study of the co-localization of neuronal nitric oxide synthase with muscarinic receptors in the rat cerebral cortex.

There is increasing evidence that nitric oxide is an important molecular messenger involved in a wide variety of biological processes including the regulation of the cerebral circulation. For instance, it has been implicated in the vascular response to nucleus basalis magnocellularis stimulation, a structure which is widely recognized as the predominant source of cholinergic fibres projecting to the neocortex. The present investigation was carried out to determine if muscarinic receptors are present on cortical neurons expressing neuronal nitric oxide synthase (nitric oxide-producing enzyme). To this aim, double labelling of both neuronal nitric oxide synthase/vessels and neuronal nitric oxide synthase/muscarinic receptors was performed on free-floating cryosections obtained from rat brain. The observations were made by confocal laser scanning microscopy. The double labelling of neuronal nitric oxide synthase with the arterioles demonstrated the presence of nitroxidergic fibres in the wall of intraparenchymal vessels. A rich network of nitroxidergic fibres independent of the vessels was also seen in the parenchyma. Since the maximal surface of a square of tissue without any nitroxidergic fibres corresponded to 1400 +/- 105 microns2, the distance separating any cortical point from its closest neuronal nitric oxide synthase-positive fibre was never higher than 25 microns (half diagonal of square). According to models of the diffusional spread of nitric oxide, it is likely that nitric oxide can reach the whole cortical volume. Our results on the regional study of neuronal nitric oxide synthase/muscarinic receptors showed a high density of neuronal nitric oxide synthase-positive neurons principally in the frontal and perirhinal cortices and a low density in the occipital cortex. These data fit well with the known pattern of cortical projections from the nucleus basalis magnocellularis as revealed by anterogradely transported markers. The double labelling showed that about 10% of neuronal nitric oxide synthase-positive neurons were co-localized with muscarinic receptors in the frontoparietal cortex. In agreement with previous papers, the vascular innervation by nitroxidergic neuronal processes was often found to lie near the branching points of arterioles. Such localization allows neuronal nitric oxide synthase-positive neurons an extensive control of the vascular tree without requiring a large number of neuronal commands. Therefore, despite the low level of neuronal nitric oxide synthase/muscarinic receptor co-localization, this neuronal subpopulation could represent a possible relay implicated in the vascular effects of the nucleus basalis magnocellularis.

Hypervascularization in the magnocellular nuclei of the rat hypothalamus: relationship with the distribution of aquaporin-4 and markers of energy metabolism.

In the magnocellular nuclei of the hypothalamus, there is a rich vascular network for which the function remains to be established. In the supraoptic nucleus, the high vascular density may be one element, which together with the water channel aquaporin-4 expressed in the astrocytes, is related to a role in osmoreception. We tested the osmoreception hypothesis by studying the correlation between vascular and cellular densities in the paraventricular nucleus and the supraoptic nucleus. Whether aquaporin-4 is likely to contribute to osmoreception was tested by studying the distribution in the magnocellular nuclei of the hypothalamus. The high vascular density may also reflect a high metabolic activity due to the synthesis of vasopressin and oxytocin. This metabolic hypothesis was tested by studying the regional cytochrome oxidase histochemistry, the local cerebral blood flow, and the density of glucose transporter type-1 in the supraoptic and paraventricular nuclei. All the magnocellular nuclei were characterized by an extended and intense aquaporin-4 labelling and a weak cytochrome oxidase histochemistry. The highest vascular density was found in the supraoptic nucleus and the magnocellular regions of the paraventricular nucleus. The local cerebral blood flow rates were surprisingly low in the paraventricular nucleus and the supraoptic nucleus in comparison to the cerebral cortex. Furthermore in these nuclei, the antibody for glucose transporter type-1 revealed two populations of vessels differing by their labelling intensity. The similarities observed between the different nuclei suggest that, in the hypothalamus, all magnocellular regions sense the plasma osmolarity. The low local cerebral blood flow, and the patterns of glucose transporter type-1 labelling and cytochrome oxidase histochemistry suggest that the high vascularization of these hypothalamic nuclei is not related to a high metabolic capacity in basal conditions.

Effects of a glutamate uptake inhibitor on glutamate release induced by veratridine and ischemia.

It has been postulated that a reversal of glutamate reuptake ("uptake reverse") may contribute to glutamate release during cerebral ischemia. We tested this hypothesis by studying the effect of threo-3-hydroxy-DL-aspartic acid (THA), a glutamate uptake inhibitor, on extracellular glutamate accumulation measured by microdialysis during 4-vessel ischemia (20 min). The inhibitory effect of THA on sodium-dependent glutamate uptake was measured in vitro on rat hippocampal slices (Ki = 45 +/- 11 microM). We examined in vivo the effect of THA (400 microM in the dialysis solution) on the extracellular glutamate release from the rat hippocampus, during veratridine depolarization and ischemia. THA decreased the amount of glutamate appearing in the extracellular space during veratridine depolarization (61%). In contrast, the glutamate release induced by ischemia was not affected by THA. We conclude that a reversal of the sodium-dependent uptake contributes to an increase in extracellular glutamate during veratridine depolarization. In contrast, glutamate release occurring during ischemia is not mediated by uptake reverse.

Effect of monoamine oxidase inhibition on the regional cerebral blood flow response to circulating noradrenaline.

The effect of an acute i.v. infusion of noradrenaline (NA) on regional cerebral blood flow (rCBF) was investigated in the awake rat using [14C]iodoantipyrine as diffusible tracer. The contribution of vascular monoamine oxidase (MAO) to the efficiency of the enzymatic blood-brain barrier (BBB) to catecholamines was assessed by measuring the multiregional cerebrovascular response to circulating NA given alone or after i.v. administration of the monoamine oxidase inhibitor, clorgyline. Since i.v. infusion of NA elevates blood pressure, the influence of NA on the cerebrovascular bed was first studied by determining the relationship between rCBF and the mean arterial pressure (MAP). When the MAP was only slightly increased (to approximately 130 mm Hg), a trend to flow decrease under NA infusion was observed. Secondly, we compared the effects of NA on rCBF in animals treated or not treated with clorgyline. This was performed under moderate hypertension (within the 'autoregulated' range of MAP) to avoid any risk of mechanical damage to the BBB. Clorgyline administration alone did not significantly modify rCBF, but the subsequent i.v. infusion of NA induced an increase in rCBF (weighted mean 14%) in all structures investigated. The differences being statistically significant (P less than 0.05) in 5 out of 13 structures by up to 20%. Compared to studies involving disruption of the morphological BBB in which plasma NA elicits a widespread important increase in blood flow, the weak cerebrovascular effects we observed provide indirect evidence for the efficiency of the BBB to catecholamines in the conscious rat within the autoregulated range of arterial pressure.(ABSTRACT TRUNCATED AT 250 WORDS)

Adenosine modulation of amino acid release in rat hippocampus during ischemia and veratridine depolarization.

This study was undertaken to determine whether endogenous adenosine modulates 'in vivo' neurotransmitter amino acid release via its presynaptic receptors. Two conditions were compared: neuronal depolarization by local infusion of veratridine (600 microM), and transient global ischemia by four-vessel occlusion. Both stimuli were applied for 20 min. Extracellular amino acid (glutamate, taurine/GABA, glycine) variations in concentration were determined in the rat hippocampus by microdialysis and HPLC. Modulation of adenosine receptor activity was objectified by continuous local infusion of an adenosine agonist (R-phenylisopropyladenosine R-PIA) or an antagonist (theophylline), starting one hour before stimulation of amino acid release. R-PIA (100 microM) significantly decreased the glutamate release (50%) evoked by veratridine, whereas it did not significantly modify the ischemia-induced glutamate release. In contrast, theophylline did not significantly affect veratridine-induced glutamate release, but it significantly potentiated glutamate efflux (400%) under ischemic conditions. Neither treatment altered the release of the other amino acids. These data suggest that endogenous adenosine appearing in the extracellular space during veratridine-induced depolarization cannot control glutamate release. In contrast, ischemia-induced glutamate release was strongly inhibited by the concomitant increase in extracellular adenosine.

Cerebrovascular changes elicited by electrical stimulation of the centromedian-parafascicular complex in rat.

Electrical stimulation of the centromedian-parafascicular complex (CM-Pf) in anesthetized (chloralose) and paralyzed (tubocurarine) rats elicits a widespread cerebrovascular dilatation. Regional cerebral blood flow (rCBF) was measured in dissected tissue samples of 10 brain regions (medulla, pons, cerebellum, inferior colliculus, superior colliculus, frontal parietal and occipital cortices, caudate-putamen and corpus callosum) by [14C]iodoantipyrine method. In unstimulated and sham-operated rats rCBF ranged from 40 +/- 3 (ml/100 g/min) in corpus callosum to 86 +/- 6 (ml/100 g/min) in inferior colliculus. During CM-Pf stimulation, rCBF increased significantly (P less than 0.05, analysis of variance and Scheffe's test) in all cerebral regions bilaterally ranging from +118% in parietal cortex to +38% in cerebellum. Although cerebral vasodilation elicited by CM-Pf stimulation persisted after unilateral transection of the cervical sympathetic trunk, the cortical CBF was significantly reduced (P less than 0.05) on the denervated side. Acute adrenalectomy significantly (P less than 0.05) decreased elevated rCBF during CM-Pf stimulation in all cortical regions (frontal-36%, parietal -34%, and occipital -27%) and in caudate nucleus (-37%). Thus, excitation of neurons originating in, or fibers passing through the CM-Pf can elicit a powerful cerebral vasodilation. The cerebral vasodilation is modulated by cervical sympathectomy and circulating adrenal hormones. We conclude that CM-Pf elicited vasodilation is at least partly mediated by intrinsic neural pathways.

Distribution of muscarinic receptors on the endothelium of cortical vessels in the rat brain.

Functional and pharmacological studies have suggested that there are muscarinic receptors (mAChRs) on the endothelial cells of major cerebral arteries, while recent immunological studies indicate that there are no mAChRs on the endothelium of brain capillaries. This difference may be because the distribution of mAChR on the endothelium varies with the type of vessel. This paper examines the distribution of mAChR on the vascular endothelium along intraparenchymal blood vessels in the rat brain by immunolabelling and laser confocal microscopy. Sections were immunostained by combinations of an anti-mAChR antibody (M35) with antibodies to endothelial (anti-GLUT1), or to smooth muscle markers (anti-actin). Antibody labellings were detected with fluorescent second antibodies. Most of the penetrating vessels bore mAChR immunolabelling which coincided over almost all the vessel surface with endothelial labelling. The mAChR immunolabelling was less widespread over the endothelium on the medium sized vessels (diameter < 50 microm) and only 50% of these vessels had mAChR staining on the endothelium. There was no mAChR immunostaining on the endothelium of the capillaries. In contrast with the basilar artery, there was no mAChR immunolabelling on the smooth muscle layer of the intracortical vessels. These data indicate that the intensity of mAChR immunolabelling decreases along the vascular tree from large conducting vessels to capillaries.

Effects of an A1 adenosine receptor agonist on the neurochemical, behavioral and histological consequences of ischemia.

Untreated rats and rats given the A1 receptor adenosine agonist, R-phenylisopropyladenosine (R-PIA), were subjected to four vessel ischemia. The effect of R-PIA on hippocampal amino acid release, hippocampal neuronal damage, exploratory behavior, learning capacity and global neurological score were evaluated. R-PIA decreased by half the glutamate released during ischemia and improved the global neurological scores 3, 24, 48, 78 h and 7 days after ischemia. But R-PIA had no effect on taurine/GABA release (during ischemia), hippocampal neuronal damage (7 days post-ischemia), exploratory behavior (48 h post-ischemia) or learning capacity (7 days post-ischemia). Thus, a decrease in glutamate release by R-PIA is not systematically correlated with an improvement of histological damage or learning capacity. Reduced glutamate release is not therefore a sufficient criterion on which to evaluate the neuroprotective capacity of a drug.

7-Nitroindazole, a selective inhibitor of nNOS, increases hippocampal extracellular glutamate concentration in status epilepticus induced by kainic acid in rats.

The glutamate extracellular concentration is controlled by metabolic and neuronal pathways via release and uptake mechanisms. Stimulation of glutamate receptors induces neuronal nitric oxide (NO) release, which in turn modulates glutamate transmission. In this study, the influence of neuronally derived NO on hippocampal glutamate extracellular concentration was investigated in conditions of intense metabolic activation, i.e., during status epilepticus induced by systemic kainic acid (KA). Glutamate, arginine and citrulline concentrations were measured by microdialysis coupled to HPLC. Experiments were performed in conscious rats implanted with a microdialysis probe within the hippocampal CA3 area. Three groups were used: (1) rats treated with KA i.p. (12 mg/kg) and vehicle locally, via the microdialysis probe (n = 9); (2) rats given KA i.p. and a selective inhibitor of neuronal NO synthase, 7-nitroindazole (7-NI, 1.25 mM) locally (n = 13); (3) rats treated with saline i.p. and 7-NI locally (n = 7). Infusion of 7-NI or vehicle was performed throughout the second hour of status epilepticus. In groups 1 and 3, no significant modifications of extracellular glutamate, arginine and citrulline concentrations were measured. In group 2, the local application of 7-NI in the hippocampus during status epilepticus significantly increased extracellular glutamate and arginine concentrations, whereas citrulline concentration remained constant. The concomitant increases of extracellular glutamate and arginine concentrations under local 7-NI perfusion in seizure conditions, suggest that glutamate and arginine are linked in a common metabolic pathway and/or that glutamate is involved in the cross-talk between glia and neurons. A cerebrovascular effect of 7-NI which triggers glutamate release may also occur.

Effect of two different routes of administration of R-PIA on glutamate release during ischemia.

Considering that adenosine decreases glutamate release from brain slices by stimulating presynaptic A1 receptors, we have attempted to modulate glutamate release in vivo during global ischemia with an agonist (R-phenylisopropyladenosine, R-PIA) of A1 receptors. Extracellular hippocampal glutamate was sampled by microdialysis and measured by HPLC. Conscious rats were submitted to transient global ischemia for 20 min. Ischemia induced a significant increase (10 fold) in extracellular glutamate. R-PIA (20 micrograms/kg) administered i.p. 30 min before ischemia significantly reduced (-64%) glutamate release. Conversely, R-PIA (100 microM) continuously infused through the hippocampal dialysis probe did not significantly modify glutamate efflux. The efficiency of infused R-PIA was evidenced by the decrease (-47%) of glutamate release induced by veratridine depolarization. These results indicate that the depressive action of R-PIA during ischemia results from various effects which are not restricted to a local action on the hippocampus.

Presence of aquaporin-4 and muscarinic receptors in astrocytes and ependymal cells in rat brain: a clue to a common function?

Using combined double immunofluorescence and laser confocal microscopy, we studied the common cellular localization of cholinergic muscarinic receptors (mAChRs) and aquaporin-4 water channels (AQP4) in the cortex, the corpus callosum and in ependymal cells of the rat brain. In the cortex, AQP4 staining was restricted to the perivascular end-feet of astrocytes. It was more widely distributed on the astrocytes of the corpus callosum. On astrocytes, mAChRs were often present in regions immunoreactive to AQP4. Ependymal cells bordering the third ventricle were also stained by both antibodies. The double staining of mAChRs with AQP4 on two different cell-types might indicate that further interactions exist which may be important in the regulation of water and electrolyte movements in the brain.

Vascular monoamine oxidase activity in the rat brain: variation with the substrate and the vascular segment.

Brain vascular monoamine oxidase (MAO) was assayed in order to determine (a) whether microvessel MAO is more or less specific for certain substrates and (b) if the extraparenchymal, pial arteries possess an MAO activity as high as that in the microvessels. Rat brain microvessels were prepared by gentle homogenisation of grey matter, followed by filtration and differential centrifugation of the matter retained. Pial arteries were carefully freed of the meninges and cut into small segments. For comparison, rat mesenteric arteries were also dissected out and cut up. MAO was assayed by measuring the rate of oxygen consumption in a small cell with a Clark electrode. Although a high microvessel MAO activity (2.2 +/- 0.3 nmol min-1 mg prot.-1) was found using noradrenaline as substrate, significantly higher rates were found with tyramine, serotonin and beta-phenyl-ethylamine. By contrast, both pial and mesenteric arteries showed a 6-7 fold lower activity (substrate tyramine). These results indicate first, that a certain specialisation of the microvessel MAO activity exists which is apparently independent of the classical A or B-form category of the substrates, and second, that the extraparenchymal vessels (pial arteries) appear to possess significantly lower MAO activity, in accordance with the concept that blood-brain properties are induced by the cerebral parenchyma.

Adrenergic receptors on cerebral microvessels in control and parkinsonian subjects.

The binding of adrenergic ligands (3H-prazosin, 3H-clonidine, 3H-dihydroalprenolol) was studied on a preparation of cerebral microvessels in the prefrontal cortex and putamen of control and Parkinsonian subjects. The adrenergic receptor density in microvessels of control patients was less than 0.5% and 3.3% respectively of the total binding. A significant decrease in the number of alpha-1 binding sites was observed on microvessels in the putamen of patients with Parkinson's disease.

UT-B1 urea transporter is expressed along the urinary and gastrointestinal tracts of the mouse.

Selective transporters account for rapid urea transport across plasma membranes of several cell types. UT-B1 urea transporter is widely distributed in rat and human tissues. Because mice exhibit high urea turnover and are the preferred species for gene engineering, we have delineated UT-B1 tissue expression in murine tissues. A cDNA was cloned from BALB/c mouse kidney, encoding a polypeptide that differed from C57BL/6 mouse UT-B1 by one residue (Val-8-Ala). UT-B1 mRNA was detected by RT-PCR in brain, kidney, bladder, testis, lung, spleen, and digestive tract (liver, stomach, jejunum, colon). Northern blotting revealed seven UT-B1 transcripts in mouse tissues. Immunoblots identified a nonglycosylated UT-B1 protein of 29 kDa in most tissues and of 36 and 32 kDa in testis and liver, respectively. UT-B1 protein of gastrointestinal tract did not undergo N-glycosylation. Immunohistochemistry and in situ hybridization localized UT-B1 in urinary tract urothelium (papillary surface, ureter, bladder, and urethra), prominently on plasma membranes and restricted to the basolateral area in umbrella cells. UT-B1 was found in endothelial cells of descending vasa recta in kidney medulla and in astrocyte processes in brain. Dehydration induced by water deprivation for 2 days caused a tissue-specific decrease in UT-B1 abundance in the urinary bladder and the ureter.

Capacity for energy metabolism in microvessels isolated from rat brain.

Numerous methods used for the isolation of brain microvessels involve procedures which disturb the structural integrity of the cells and their organelles. In the present study, analysis of the adenylate energy charge and content as well as the incorporation of adenosine derivatives in isolated rat brain microvessels indicated a lesion of the mechanisms of energy production. The results show that experiments on isolated microvessels prepared by a mechanical homogenization exerting shear forces should be interpreted with caution when the rate of energy metabolism is a significant factor in the study.

Local cerebral blood flow in gently restrained rats: effects of propranolol and diazepam.

The cerebrovascular consequences of the gentle restraint commonly used for the measurement of local cerebral blood flow (LCBF) in conscious rats has been tested by using two drugs, propranolol and diazepam. Propranolol induced small LCBF decreases in 7 structures suggesting that the cerebral circulation was partially controlled by the activation of intra or extracerebral aminergic pathways in this protocol. Sedative doses of diazepam reduced LCBF in most of the structures but anxiolytic doses increased it in 4 structures. This effect may be due to a regionally differentiated modification of activity in the brain due to the selective inhibition of limbic structures by diazepam. Since propranolol and diazepam induced only small changes, the LCBF of gently restrained rats appeared to be minimally affected by the stressful situation imposed by the protocol.

Muscarinic binding of pial vessels and arachnoid membrane.

The muscarinic sites in arachnoid and pial vessels were compared by analysis of the binding of quinuclidinyl benzilate (QNB) to membrane preparations. Saturation analysis indicated that the process was saturable, high affinity, and related to protein concentration in both structures. Although the affinities in the two structures [KD = 0.039 (arachnoid) and 0.097 nM (pial vessels)] were similar, the arachnoid had approximately 10-fold more binding sites (Bmax = 2,100 fmol/mg of protein) than the pial vessels (Bmax = 250 fmol/mg of protein). This difference was found in both bovine and porcine fractions. Pharmacological analysis of [3H]QNB displacement by muscarinic and nonmuscarinic ligands gave the typical pattern of muscarinic receptors in both structures. Inhibition of binding to pial vessels by the M1 antagonist pirenzepine revealed only one low-affinity site (Ki = 7.8 x 10(-7) M), whereas, the arachnoid had a small proportion (21%) of high-affinity sites (Ki = 2.2 x 10(-9) M) associated with low-affinity sites (Ki = 5.50 x 10(-7) M). It is concluded that muscarinic-mediated effects that do not involve the M1 subtype are induced in bovine pial vessels by a relatively low concentration of binding sites. The high content of muscarinic binding sites and their diversity in the arachnoid suggest a functional role for muscarinic cholinergic receptors in this structure.