Divergent Regulation of Muscarinic Binding Sites and Acetylcholinesterase in Discrete Regions of the Developing Human Fetal Brain

Summary The expression of muscarinic acetylcholine binding sites and of cholinesterases was studied in extracts prepared from discrete regions of the human fetal brain, between the gestational ages of 14 and 24 weeks. The specific binding of [³H]N-methyl-4-piperidyl benzilate ([⁴H]-4NMPB) to muscarinic binding sites ranged between 0.05 and 1.30 pmol/mg protein in the different brain regions, withK _d values of 1.2 ± 0.2 nM. Binding of the cholinergic agonist oxotremorine fitted, in most of the brain regions examined, with a two-site model for the muscarinic binding sites. The density of muscarinic binding sites increased with development in most regions, with different rates and onset times. It was higher by about sixfold in some areas destined to become cholinergic, such as the cortex and midbrain, than in noncholinergic areas such as the cerebellum. In other areas destined to become cholinergic, such as the hippocampus and the caudate putamen, the receptor density remained low. Average density values increased from 0.1 ± 0.1 at 14 weeks up to 0.7 ± 0.4 pmol/mg protein at 24 weeks.The variability in the specific activities of cholinesterase was relatively low, and extracts from different brain regions hydrolyzed from 5 to 30 nmol of [³H]acetylcholine/min/mg protein. These were mostly true acetylcholinesterase (EC 3.1.1.7) activities, inhibited by 10^–5 M BW284C51, with minor pseudocholinesterase (EC 3.1.1.8) activities, inhibited by 10^–5 M iso-OMPA. The enzyme from different brain regions and developmental stages displayed similarK _m values toward [³H]acetylcholine (ca. 4 × 10^–4 M ^–1). The ontogenetic changes in cholinesterase specific activities had no unifying pattern and/or relationship to the cholinergic nature of the various brain areas. In most of the brain regions, the arbitrary ratio between the specific activity of cholinesterase and the density of muscarinic binding sites decreased with development, with average values and variability ranges of 83 ± 50 and 19 ± 19 at 14 and 24 weeks, respectively. Our findings suggest divergent regulation for cholinergic binding sites and cholinesterase in the fetal human brain and imply that the expression of muscarinic receptors is related to the development of cholinergic transmission, while acetylcholinesterase is also involved in other functions in the fetal human brain.I.B. took part in this work as partial fulfillment of the requirements of the Sackler Faculty of Medicine for an M.D. degree.     

A quantitative autoradiographic study of muscarinic cholinergic receptor subtypes in the brains of pyrithiamine-treated rats

Previous studies describe decreased acetylcholine synthesis in brain as well as neurobehavioural evidence for a central muscarinic cholinergic deficit in pyrithiamine-induced thiamine-deficient rats. In order to further evaluate this possibility, quantitative autoradiographic procedures using [³H]quinuclidinyl benzilate (for total muscarinic binding sites), [³H]pirenzepine (for muscarinic M₁ sites) and [³H]AF-DX 384 (for muscarinic M₂ sites) were performed at early (presymptomatic) and late (symptomatic) stages of thiamine deficiency induced in rats by administration of the central thiamine antagonist, pyrithiamine. No significant alterations in densities of M₁, M₂ or total muscarinic binding sites were observed in any brain structure evaluated at either early or late stages of thiamine deficiency. These findings do not support a major role for modifications of muscarinic cholinergic function in the pathogenesis of the neurological symptoms of thiamine deficiency.     

Anomalous binding of [3 H] N-methyl-quinuclidinyl benzilate methyl chloride to human lymphocyte muscarinic receptors

Using the muscarinic cholinergic ligand [³ H] N-methyl quinuclidinyl benzilate methyl chloride ([³ H] NM-QNB), we demonstrated that intact, viable human lymphocytes posses specific muscarinic binding sites. Equilibrium binding studies show that muscarinic acethylcholine receptor are devided into two subtype; high affinity (Ms) and low affinity types (Mw) for the ligand.     

STEREOSPECIFIC IN VlTRO BINDING OF [3H]DEXETIMIDE TO BRAIN MUSCARINIC RECEPTORS

A binding assay for muscarinic cholinergic receptors has been developed using labelled dexetimide as ligand and a filtration technique. The main features of this assay are its stereospecific nature, the very high affinity of the ligand for the specific receptors sitcs and its very low affinity for non-specific binding sites. The latter point was further investigated using labelled levitimide, the inactive enantiomer. The binding was found to be neither stereospecific nor saturable and displacement by both enantiomers revealed a particular curve with a very flattened course. Kinetic experiments with [³H]dexetimide suggest the occurrence of a heterogenous population of muscarinic receptors in the rat striatum. A study of the regional distribution of muscarinic receptors in rat brain showed a high concentration in the dopaminergic areas, the cortex and the hippocampus, but practically none in the cerebellum. The subcellular distribution pattern revealed a marked enrichment of [³H]dexetimide stereospecific binding sites in the microsomal fraction of rat striatum and hippocampus. Such a distribution was not found with [³H]levitimide. All the characteristics of this binding assay make dexetimide a very appropriate ligand for labelling muscarinic receptors in vitro as well as in vivo.     

Cholinergic and adrenergic receptors on mouse cardiocytes in vitro

The effects of adrenergic and cholinergic receptor agonists and antagonists on single and clustered mouse cardiocytes in culture have been studied. Cardiocytes were obtained from mice, ranging in ages from 9 days in utero to 1 day postpartum, and were grown in culture for 2–14 days. Single isolated cells of every age tested possessed the ability to respond both via a muscarinic cholinergic receptor to the cholinergic agonist, carbamylcholine, and via α- and β-adrenergic receptors to norepinephrine and epinephrine. Thus, cholinergic and adrenergic receptors are simultaneously present on the same cell. Cardiocyte clusters had considerably higher sensitivity to both autonomic agents, but, because of the extensive functional specializations between cells, the localization of functional receptors to specific cells could not be made. [³H]Alprenolol, a potent β-adrenergic receptor antagonist, and [³H]quinuclidinyl benzilate ([³H]QNB), a potent muscarinic cholinergic receptor antagonist, were used to localize β-adrenergic and muscarinic cholinergic receptors by autoradiography. Quantitation of the muscarinic ACh receptor gave ～800 sites/μm², a value comparable to that for the nicotinic ACh receptor on primary skeletal muscle in culture. Electrophysiological and fine-structural studies confirmed the myocardial nature of these cells.     

Dual muscarinic and nicotinic action on a motor program in Drosophila

The effect of cholinergic agonists and antagonists on the central pattern generator of the pharyngeal muscles has been studied in third instar larvae of Drosophila. The pharyngeal muscles are a group of rhythmically active fibers involved in feeding. Bath application of the cholinergic agonists carbachol, musarine, pilocarpine, and acetylcholine (ACh) to a semiintact preparation including the pharyngeal muscles and the central nervous system (CNS), initiated long-lasting endogenous-like bursting activity in the muscles. The muscarinic antagonists, atropine and scopolamine, blocked these responses as well as endogenous activity. Perfusion with nicotine elicited a short, tonic response that was marginally blocked by mecamylamine but not by curare, α-bungarotoxin, hexamethonium, or the muscarinic antagonists. This is the first time that a response to cholinergic drugs has been examined in Drosophila. The pharyngeal muscle preparation may prove to be a valuable system for studying mutations of cholinergic metabolism, receptors, and second messengers.     

Muscarinic cholinergic binding sites on intact human lymphocytes

Using the muscarinic chalinergic ligand [³H]-quinuclidinyl benzilate, we have demonstrated that intact, viable human lymphocytes possess specific muscarinic binding sites. The binding is saturable, proportional to cell number, and is displaceable by atropine, benztropine, trihexyphenidyl and scopolamine. The apparent kd is 67 nM and the number of binding sites per cell is on the order of 5 × 10⁴. Not only do these findings provide a pharmacological basis for the observed effects of muscarinic agents on lymphocyte function, they also demonstrate the utility of human peripheral blood lymphocytes for investigation of abnormalities of the muscarinic cholinergic system.     

Muscarinic cholinergic binding in human erythrocyte membranes

Human erythrocyte ghosts contain a small population of muscarinic cholinergic receptors, as evidenced by their high affinity binding of radiolabeled quinuclinidinyl benzilate ([³H]QNB). The apparent K_D is 1.3 × 10^?9 M and the receptor sites are saturated at a QNB concentration of 5 nM. The number of sites is 23 fmoles/mg membrane protein. The pharmacological profile of the specific binding is similar to that of neural membranes. The binding is not stereoselective for the d and 1 isomers of QNB, a situation which prevails in the muscarinic receptors of another peripheral cholinergic system, the rat iris, but not in the central nervous system.     

RELEASE OF NOREPINEPHRINE FROM NEURONS IN DISSOCIATED CELL CULTURES OF CHICK SYMPATHETIC GANGLIA VIA STIMULATION OF NICOTINIC AND MUSCARINIC ACETYLCHOLINE RECEPTORS

Abstract— Dissociated cell cultures of chick embryo sympathetic ganglia were incubated with [³H]nor-epinephrine ([³H]NE) which was taken up and stored in reserpine-sensitive sites. Exposure of the cultures to cholinergic agonists for 5 min intervals resulted in the releaseof a significant proportion (2–20%) of the intracellular stores of [³H]NE. Studies with specific cholinergic agonists and antagonists indicated that release of [³H]NE could be evoked by stimulation of either nicotinic or muscarinic receptors. Release evoked by both nicotinic and muscarinic agonists was totally blocked in the presence of 3 μM-tetrodotoxin. thus indicating that release was mediated via active electrical responses. Release by both types of agonists was also blocked in the presence of elevated Mg²⁺ or when free Ca²⁺ was removed from the extracellular medium. These findings are consistent with the presence of a stimulus-secretion coupling mechanism. Release evoked by nicotine was optimal in the presence of 1.2 mM-Ca²⁺, whereas release evoked by the muscarinic agonist methacholine increased by about 2-fold when the Ca²⁺ concentration was decreased from 1.2 to 0.3 mM. The latter observation may be due to a lowered threshold for evocation of active responses at low concentrations of Ca²⁺. Finally, no evidence was observed for interaction between the two types of receptors. These findings (a)indicate that cultured chick sympathetic neurons possess functional nicotinic and muscarinic cholinergic receptors as well as the ability to release NE via a stimulus-secretion coupling mechanism; (b) suggest that such cultures may be particularly useful for studying the molecular events which link stimulation of cholinergic receptors to neurotransmitter release; and (c) provide further evidence that muscarinic receptors may play aphysiological role in ganglionic transmission.     

Isolation by affinity chromatography of a cholinergic proteolipid from Nucleus caudatus.

A cholinergic proteolipid fraction (i.e. a hydrophobic lipoprotein) was separated from the $\text{n. caudatus}$ of the cow, using affinity chromatography with the lipophilic gel Sephadex LH-20 and p-phenyltrimethylamonium as the active group. High affinity binding studies showed that only the specific fraction, desorbed after an acetylcholine (or acid) pulse, and corresponding to 0,72% of the proteolipids, is the one that binds the cholinergic ligands. The binding of (³H)atropine and (¹⁴C)d-tubocurarine demonstrated that there are 814 picomoles/g fresh tissue of muscarinic sites and only 76 picomoles/g of nicotinic sites. The specific radioactivity for (³H)atropine is 10,000 nmoles/g protein, suggesting a high degree of purification of the specific cholinergic proteolipid.     

Regulation of putative muscarinic cholinergic receptor subtypes in rat brain

Transection of the fimbria/fornix, producing a 75% reduction in the activity of the cholinergic marker choline-o-acetyltransferase (CAT EC. 2.3.1.6) in rat hippocampus, did not change the binding characteristics of the non-subtype selective, muscarinic cholinergic receptor antagonist ligand [³H](−)quinuclidinyl benzilate {[³H](−)QNB}. Pirenzepine competition for [³H](−)QNB binding in the hippocampus was best described by a computer derived model assuming two binding sites of high affinity (putative M₁ receptors) and low affinity (putative M₂ receptors). There was no change in the proportion of high and low affinity pirenzepine binding sites in the hippocampus following cholinergic deafferentation. Thus, these data provide no evidence for a discrete localization of either putative subtype of muscarinic receptor discriminated by pirenzepine restricted to the terminals of CAT containing neurons innervating the rat hippocampus.Chronic scopolamine treatment produced a 48% increase in the B_max of [³H](−)QNB binding in the hippocampus, but again there was no change in the proportions of the sites discriminated by pirenzepine demonstrating that both putative subtypes were regulated identically. Similarly, carbachol competition for [³H](−)QNB was unaltered following cholinergic deafferentation or chronic scopolamine treatment. Furthermore, similar guanylyl-5′-imidodiphosphate [Gpp(NH)p] modulation of the proportions of high and low affinity carbachol binding sites was found in the hippocampus following transection of the fimbria/fornix or chronic scopolamine treatment. Thus there is no adaptation of receptor-effector coupling following these treatments that is reflected by changes in receptor recognition site characteristics.Carbachol competition for [³H]pirenzepine binding to putative M₁ receptors in the hippocampus was biphasic and was also modulated by Gpp(NH)p. In the brainstem, there was a homogeneous population of putative M₂ [³H](−)QNB binding sites having low affinity for pirenzepine. Carbachol competition for these binding sites was also biphasic and modulated by guanine nucleotides. Thus, both putative M₁ and M₂ muscarinic receptors, as defined by high or low affinity for pirenzepine respectively, may mediate their effects in rat brain via a guanine nucleotide regulatory subunit.     

Cholinoceptor-mediated control of catecholamine release from chromaffin cells in the American eel,Anguilla rostrata

The cholinergic agonist-induced secretion of catecholamines from chromaffin cells in the American eel, Anguilla rostrata, was assessed using a salineperfused posterior cardinal vein preparation. Direct membrane depolarization with 60 mmol·l^-1 K⁺ caused a significant release of catecholamines (adrenaline + noradrenaline) into the perfusate which was unaffected by pre-treatment with the ganglion blocker, hexamethonium (final concentration = 10^-3 mol · l^-1). The nicotinic receptor agonist, 1,1-dimethyl-4-phenylpiperazinium iodide, evoked catecholamine release in response to several doses exceeding 10^-7 mol; at 10^-5 mol the response was abolished by pre-treatment with the ganglion blocker, hexamethonium (final concentration = 10^-3 mol · l^-1). The muscarinic receptor agonist, pilocarpine, did not elicit catecholamine release in response to any of the doses administered (10^-8–10^-4 mol). A single injection of the mixed nicotinic/muscarinic cholinoceptor agonist, carbachol (10^-5 mol), caused the release of catecholamines which was abolished by pre-treatment with hexamethonium but which was unaffected by pre-treatment with the muscarinic receptor antagonist atropine (final concentration = 10^-5 mol · l^-1). The results of this study indicate that the process of cholinergic agonist-induced catecholamine secretion from the chromaffin cells in the American eel is mediated exclusively by activation of nicotinic receptors with no involvement of the muscarinic receptor.Abbreviations DMPP 1,1-dimethyl-4-phenylpiperazinium iodide - MS222 ethylaminobenzoate - HPLC high-performance liquid chromatography - PCV posterior cardinal vein - SEM standard error of the mean     

Specific binding of the muscarinic antagonist [3H]quinuclidinyl benzilate is not associated with preganglionic motor neurons in the dorsal motor nucleus of the vagus

The present study evaluates the binding of [³H]quinuclidinyl benzilate, [³H]QNB, as a measure of cholinergic muscarinic binding in six areas of the rat medulla oblongata associated with the cranial nerves. In an experimental group, the right vagus nerve was severed in the neck in order to determine whether the specific muscarinic binding sites might be located on cells that contribute efferent fibers to the vagus nerve. The level of activity of choline acetyltransferase (ChAT) also was determined in the same six areas. Additional experiments utilizing the retrograde transport of toxic ricin, a 60,000 dalton agglutinin that acts as a potent ribosomal toxin, was carried out to further evaluate localization of specific muscarinic binding in the DMN after destruction of the preganglionic efferent cells. These results support the conclusion that specific binding of the muscarinic antagonist [³H]QNB observed in the DMN of the vagus of the rat is not associated with the large cells that contribute efferent fibers into the vagus nerve. We suggest that the specific cholinergic muscarinic binding is located on interneuronal cell surfaces, on afferent terminals of local circuit neurons, or on afferent terminals of long projection axons which arise from neurons in the brainstem, hypothalamus, or forebrain.This issue is dedicated to Donald B. Tower.     

Isolated rat gastric parietal cells: Cholinergic response and pharmacology

Isolated, partially purified or enriched rat gastric muscosal parietal cells were shown to respond to carbamycholine (EC₅₀ = 2 μM) and other muscarinic cholinergic agonists as measured by an increased accumulation of ¹⁴C-aminopyrine, an indirect measure of acid secretion. The secretory response to carbamylcholine was shown to be inhibited stereoselectively and reversibly by nanomolar concentrations of muscarinic cholinergic antagonists. Non-muscarinic antagonists, including cimetidine, were either ineffective or very weak inhibitors. The affinity constants calculated for cholinergic antagonist inhibition of ¹⁴C-aminopyrine accumulation induced by carbamylcholine were similar to those previously calculated from direct binding studies on purified parietal cell particulate fractions using ³H-QNB (1). These studies support the existence of specific parietal cell muscarinic cholinergic receptors with which the natural secretagogue acetylcholine interacts to regulate gastric acid secretion.     

Muscarinic receptors in pineal

The presence of muscarinic receptors in sheep and rat pineals was detected by binding of [³H]quinuclidinyl benzilate ([³H]QNB), a potent and specific muscarinic antagonist. [³H]QNB binding to sheep pineal membrane resuspensions was saturable and reversible, with a rate constant for association at 37°C of 6×10⁸M^?1min^?1 and a rate constant for dissociation of 1×10^?2min^?1. Kinetic and saturation experiments yielded an equilibrium dissociation constant of 13–18 pM and a concentration of binding sites equivalent to 1.1 pmol/g of original wet weight. This is only about 5% of the level of β-adrenergic receptors. Competition by a variety of cholinergic drugs confirmed the muscarinic nature of the binding sites. Experiments in rats failed to detect a significant decrease in pineal [³H]QNB binding following bilateral superior cervical ganglionectomy, suggesting that the binding sites are not localized exclusively on sympathetic terminals.     

Altered Muscarinic and Nicotinic Receptor Densities in Cortical and Subcortical Brain Regions in Parkinson's Disease

Abstract: Muscarinic and nicotinic cholinergic receptors and choline acetyltransferase activity were studied in postmortem brain tissue from patients with histopathologically confirmed Parkinson's disease and matched control subjects. Using washed membrane homogenates from the frontal cortex, hippocampus, caudate nucleus, and putamen, saturation analysis of specific receptor binding was performed for the total number of muscarinic receptors with [³H]quinuclidinyl benzilate, for muscarinic M₁ receptors with [³H]pirenzepine, for muscarinic M₂ receptors with [³H]oxotremorine-M, and for nicotinic receptors with (–)-[³H]nicotine. In comparison with control tissues, choline acetyltransferase activity was reduced in the frontal cortex and hippocampus and unchanged in the caudate nucleus and putamen of parkinsonian patients. In Parkinson's disease the maximal binding site density for [³H]quinuclidinyl benzilate was increased in the frontal cortex and unaltered in the hippocampus, caudate nucleus, and putamen. Specific [³H]pirenzepine binding was increased in the frontal cortex, unaltered in the hippocampus, and decreased in the caudate nucleus and putamen. In parkinsonian patients B_max values for specific [³H]oxotremorine-M binding were reduced in the cortex and unchanged in the hippocampus and striatum compared with controls. Maximal (–)-[³H]nicotine binding was reduced in both the cortex and hippocampus and unaltered in both the caudate nucleus and putamen. Alterations of the equilibrium dissociation constant were not observed for any ligand in any of the brain areas examined. The present results suggest that both the innominatocortical and the septohippocampal cholinergic systems degenerate in Parkinson's disease. The reduction of cortical [³H]oxotremorine-M and (–)-[³H]nicotine binding is compatible with the concept that significant numbers of the binding sites labelled by these ligands are located on presynaptic cholinergic nerve terminals, whereas the increased [³H]pirenzepine binding in the cortex may reflect postsynaptic denervation supersensitivity.     

Effects of alkylating antagonists on the stimulated turnover of phosphatidylinositol produced by a variety of calcium-mobilising receptor systems

We have investigated the effects of two halogenoalkylamine drugs, dibenamine and phenoxybenzamine, on the stimulated phosphatidylinositol turnover that is produced by neurotransmitters and hormones which interact with receptors to bring about an increase in cell surface Ca²⁺ permeability. The phosphatidylinositol responses we have investigated were those evoked by muscarinic cholinergic stimuli (parotid gland and pancreas), by α-adrenergic stimuli (parotid gland, vas deferens smooth muscle), by pancreozymin or caerulein (pancreas), by phytohaemagglutinin (lymphocytes) and by either 5-hydroxytryptamine or elevation of the extracellular K⁺ concentration (ileum smooth muscle). Phenoxybenzamine inhibited the muscarinic cholinergic, α-adrenergic, 5-hydroxytryptamine and high K⁺ responses, but not the responses to phytohaemagglutinin and to pancreozymin (or caerulein). Dibenamine was less effective than phenoxybenzamine in inhibiting the α-adrenergic response and the high K⁺ response, and it did not inhibit the responses to muscarinic cholinergic stimuli, to 5-hydroxytryptamine or to the polypeptides. N,N-dimethyl-2-bromo-2-phenylethylamine (DMPEA) inhibited the α-adrenergic response, but not the response to muscarinic cholinergic stimulation. The specificity of DMPEA for the α-adrenergic response agrees with its postulated site of action at the noradrenaline-binding site of this receptor system, whereas dibenamine and phenoxybenzamine are less specific drugs which inhibit a variety of the ‘physiological’ responses of cells, including those to muscarinic cholinergic, H₁-histaminergic, α-adrenergic and 5-hydroxytryptamine stimuli. Previously, we suggested that dibenamine and phenoxybenzamine might show a constant pattern of effects on the phosphatidylinositol responses evoked through different receptors, phenoxybenzamine being inhibitory and dibenamine without effect [Jafferji & Michell (1976) Biochem. J. $\text{160}$, 163–169]. However, this pattern has not been sustained throughout the present study of a larger range of Ca²⁺-mobilising stimuli.     

A cholinergic receptor site on murine lymphocytes with novel binding characteristics

To further analyze functionally important cholinergic receptors on lymphocytes, we studied the binding of the muscarinic antagonist Quinuclidinyl benzilate (QNB) to murine splenic lymphocytes. Studies of displacement of [³H]QNB by unlabelled QNB on lymphocytes revealed at least two binding sites. Scatchard analysis of equilibrium binding isotherms also distinguished two sites with apparent Kds of 480 nM and 16 μM. There was greater specific QNB binding to B cell-enriched lymphocyte fractions than to T cell fractions. Lymphocyte binding demonstrated temperature-dependent dissociability, and specific binding occurred on isolated lymphocyte membranes as well. Both muscarinic and nicotinic ligands competed for QNB binding to lymphocytes with low and nearly equal affinity. Therefore, QNB binding sites on lymphocytes appear to be of low affinity and of mixed muscarinic and nicotinic character.     

Morphine impairs Acetylcholine Release but facilitates Acetylcholine Action at a Skeletal Neuromuscular Junction

THERE is considerable evidence that morphine impairs the release of acetylcholine (ACh) at cholinergic synapses in the brain^1–5, although there are considerable problems in determining the exact site and mechanism of this action. A simple synaptic model would be useful for pursuing this problem and the question arises whether this action of morphine is universal for cholinergic synapses or is restricted to particular sites. Morphine impairs the release of ACh at peripheral muscarinic sites^6–8 but there are no reports about the effects of morphine on ACh release at nicotinic neuromuscular sites. We have reported that both morphine and nalorphine block neuromuscular transmission in amphibian and mammalian skeletal neuromuscular preparations^9,10, apparently as a result of impairment of ACh release. We have now determined by direct measurement that morphine impairs ACh release at a skeletal neuromuscular junction.     

Ganglion cells of chicken retina possess nicotinic rather than muscarinic acetylcholine receptors

Chicken retinas were exposed to intravitreal kainic acid to destroy amacrine and bipolar cells at low concentrations, and horizontal cells at high concentrations in addition. Ganglion cells were destroyed by intravitreal injections of colchicine. Low doses of kainic acid reduced the number of binding sites for both [³H]quinuclidinyl benzilate (muscarinic acetylcholine receptors) and N-[propionyl ³H]-bungarotoxin (nicotinic acetylcholine receptors), with little additional loss at higher doses. In contrast, colchicine reduced the number of binding sites for N-[propionyl-³H]-bungarotoxin, but had little or no effect on the number of binding sites for [³H]quinuclidinyl benzilate. These results are consistent with the idea that, in chicken retina, cholinergic amacrine cells make contact with ganglion cell dendrites at sites which possess mainly nicotinic acetylcholine receptors, while both types of receptor are involved in interactions between amacrine cells and perhaps bipolar cells.