Dissimilar Aluminum and Gallium Permeation of the Blood-Brain Barrier Demonstrated by In Vivo Microdialysis

Aluminum (Al) and gallium (Ga) permeations of the blood-brain barrier (BBB) were assessed in rats. Unbound extracellular Al and Ga concentrations were ascertained at the two potential sites of BBB permeation, cerebral capillaries and choroid plexuses, by implantation of microdialysis probes in the frontal cortex and lateral ventricle, respectively. A microdialysis probe implanted in the jugular vein revealed unbound blood Al or Ga concentrations. Al or 67Ga citrate was administered via the femoral vein. Peak Al and Ga concentrations were seen within the first 10 min at all three sites. Area under the curve (concentration vs. time to final sample) values were calculated using RSTRIP. Within-rat overall frontal cortical/blood and lateral ventricular/blood ratios [brain/blood ratios (oBBRs)] were calculated from area under the curve values. Aluminum frontal cortical oBBRs were significantly higher than those for the lateral ventricle. Ga oBBRs were not significantly different between the two sites. Al and Ga oBBRs were significantly different in the lateral ventricle. These results suggest that the primary site of A1 permeation across the BBB is at cerebral capillaries, whereas Ga permeation across the BBB does not significantly differ between cerebral capillaries and choroid plexuses. The use of Ga as a model to study Al pharmacokinetics may not be appropriate in the elucidation of the site or mechanism of Al entry into the brain.     

Effects of a Single Therapeutic Dose of Glycerol on Cerebral Metabolism in the Brains of Young Mice: Possible Increase in Brain Glucose Transport and Glucose Utilization

Abstract: This is a study of the effects of a single “therapeutic” dose of glycerol [2 g(22 mmol)/kg i.p.] on brain carbohydrate and energy metabolism in normal nursing weanling mice. Findings were correlated with brain water and electrolyte content and with metabolite changes in plasma, red blood cells, and liver. Plasma glycerol levels peaked at 21 mM 7.5 min after injection and returned to the control value, 0.16 mM, by 2 h. Plasma Na⁺ concentration decreased and plasma protein increased for as long as 2 h after injection. Although red blood cells were freely permeable to glycerol, there was no evidence for glycerol metabolism in these cells. Glycerol levels in liver paralleled those in plasma. Glycerol injection increased liver glucose concentration 23% and doubled hepatic glycerol-1-phosphate levels. Liver ATP levels were reduced 24% after glycerol injection. Brain water concentration was significantly reduced from 7.5 min to 30 min after glycerol injection; brain Na⁺ and K⁺ levels were unchanged. There was no evidence for glycerol entry into brain (the amount detected in brain tissue could be explained by the glycerol content in the blood of the brain). While plasma glucose increased 33%, brain glucose increased 87%. Concomitantly there were statistically significant increases in fructose-1,6-diphosphate, lactate, α-ketoglutarate, and malate levels. The disproportionately high brain glucose value suggests increased transport of glucose from the blood to the brain. Increases in fructose-1,6-diphosphate, lactate, α-ketoglutarate, and malate are compatible with an increased metabolic flux in the glycolytic pathway and Krebs citric acid cycle. As has been previously shown for urea and/or mannitol, these changes may result from the effects of the hyperosmolar glycerol solution on the blood-brain barrier and on cerebral glucose utilization. The sustained lowering of plasma Na⁺ concentration after a single “therapeutic” glycerol injection suggests a need for monitoring plasma Na⁺ levels in the clinical situation. Possible lowering of hepatic ATP levels by the use of glycerol in humans is another concern.     

胶质细胞谷氨酸转运体-1反义寡核苷酸对脑缺血预处理神经保护效应的抑制作用

本研究应用胶质细胞谷氨酸转运体-1(glial glutamate transporter-1,GLT-1)的反义寡核苷酸(antisense oligo-deoxynucleotides,AS-ODNs)抑制Wistar大鼠GLT-1蛋白的表达,观察其对脑缺血预处理(cerebral ischemic preconditioning.CIP)增强脑缺血耐受作用的影响,探讨GLT-1在CIP诱导的脑缺血耐受中的作用.将凝闭双侧椎动脉的Wistar大鼠随机分为7组:(1)Sham组:只暴露双侧颈总动脉,不阻断血流;(2)CIP组:夹闭双侧颈总动脉3 min;(3)脑缺血打击组:夹闭双侧颈总动脉8 min;(4)CIP 脑缺血打击组:夹闭双侧颈总动脉3 min作为CIP,再灌注2 d后,夹闭双侧颈总动脉8min;(5)双蒸水组:于分离暴露双侧颈总动脉(但不夹闭)前12 h、后12 h及后36 h右侧脑室注射双蒸水,每次5 μL,其它同sham组;(6)AS-ODNs组:于分离暴露双侧颈总动脉(但不夹闭)前12 h、后12 h及后36 h右侧脑室注射GLT-1 AS-ODNs溶液,每次5 μL,其它同sham组,再根据AS-ODNs的剂量进一步分为9 nmol和18 nmol 2个亚组;(7)AS-ODNs CIP 脑缺血打击组:于CIP前12 h、后12 h及后36 h右侧脑室注射GLT-1 AS-ODNs溶液,每次5 μL,其它同CIP 脑缺血打击组,根据AS-ODNs的剂量进一步分为9 nmol和18 nmol 2个亚组.Western blot分析法观察GLT-1蛋白的表达,硫堇染色观察海马CA1区锥体神经元迟发性死亡(delayed neuronal death,DND)情况.Western blot分析显示,侧脑室注射GLT-1 AS-ODNs可剂量依赖性地抑制大鼠海马CA1区GLT-1蛋白表达.硫堇染色显示,sham组和CIP组海马CA1区未见明显的DND;脑缺血打击组海马CA1区有明显的DND:预先给予CIP可显著对抗脑缺血打击引起的DND,表明CIP可以诱导海马CA1区神经元产生缺血性耐受,对抗脑缺血打击引起的DND;而在GLT-1 AS-ODNs CIP 脑缺血打击组,侧脑室注射GLT-1 AS-ODNs后,大鼠海马CA1区出现了明显的DND,表明GLT-1 AS-ODNs通过抑制大鼠GLT-1蛋白表达从而减弱CIP对抗脑缺血打击的神经保护作用.以上结果进一步证实了GLT-1参与CIP诱导的脑缺血耐受.     

Glutathione and gamma-glutamyl transpeptidase in the adult female rat brain after intraventricular injection of LHRH and somatostatin

Glutathione content and glutamyl transpeptidase activity in different regions of adult female rat brain were determined at 10 and 30 min following intraventricular injection of LHRH and somatostatin. Hypothalamic glutathione levels were significantly elevated at 10 and 30 min after a single injection of a 0.1 micrograms dose of LHRH. On the contrary, glutathione levels significantly decreased in the hypothalamus, cerebral cortex and cerebellum at 10 and 30 min after 0.5 or 1 microgram dose. However, significant decrease in brain stem glutathione was evident at 30 min after 0.5 microgram and 10 min after the 1 microgram dose. Somatostatin at doses of 0.5 microgram and 1 microgram significantly decreased glutathione levels in all four brain regions both at 10 and 30 min following injection into the 3rd ventricle. Gamma-glutamyl transpeptidase activity in the hypothalamus and cerebral cortex was significantly elevated after intraventricular injection of LHRH. However, a significant increase in gamma-glutamyl transpeptidase activity in cerebellum and brain stem was seen only with 0.5 and 1 micrograms doses of LHRH. Somatostatin also significantly increased gamma-glutamyl transpeptidase activity in hypothalamus, cerebral cortex, brain stem and cerebellum. The decrease in glutathione levels with corresponding increase in gamma-glutamyl transpeptidase activity after intraventricular administration of LHRH and somatostatin suggests a possible interaction between glutathione and hypothalamic peptides.     

Clonidine and morphine increase atrial natriuretic peptide secretion in anesthetized rats

In order to determine whether the activity of central alpha 2-adrenergic and opioid receptors influence plasma atrial natriuretic peptide (ANP) levels, clonidine and morphine were infused into the lateral cerebral ventricle for 45 min in anesthetized Sprague-Dawley rats. The central administration of a low dose of clonidine (10 ng/min) caused a significant increase in plasma ANP without changing arterial blood pressure or central venous pressure. Pretreatment with yohimbine (5 micrograms/min) completely blocked the effect of clonidine. Central infusion of morphine (100 ng/min) also elevated plasma ANP levels and naloxone (5 micrograms/min) blunted this effect. Intravenous infusion of the same dose of clonidine or morphine did not affect plasma ANP levels. Moreover, the effect of clonidine on plasma ANP was partially blocked by pretreatment with naloxone (5 micrograms/min). These results suggest that central alpha 2-adrenergic and opioid receptors may be involved in ANP secretion.     

Method of labelling of individual ependymal areas according to periventricular structures of the rat lateral brain ventricles

Ependymal areas were studied in the lateral brain ventricles of the rat central nervous system and were labelled with a code. The presented suggestion using the coding for individual ependymal areas in rat ventricle may solve the significant problem in experimental studies, i.e. how to secure the mutual comparison of the same type of ependymal areas or ependymal cells. The periventricular structures represent a basic and stable part of brain nerve tissue and they are localized most closely to the studied part of the ventricle wall. For this quality they were chosen as reference nerve tissue for the labelling of the ependymal areas and they were used for the creation of the code. The code is composed from letters “Lv” (lateral ventricle) and “E” (ependymal area) followed by the abbreviation of the latin name of the periventricular structure, e.g., the corpus callosum abbreviation is “cc”. The code of the ependymal area over the corpus callosum is thus “LvE-cc”. The proposed labelling of the ependymal areas may offer several advantages, such as: (i) better characterization of ependymal areas in the future; (ii) preventing the interchange of different types of ependymal areas or ependymal cells; and (iii) avoiding a false interpretation in experiments.     

Vestibular efferent neurons forming projections to the saccule in the guinea pig

A comparative analysis was made of the distribution of vestibular efferent neurons projecting to the saccule and efferent cells sending out axons to the auditory nerve ("cochlear efferent neurons") in the guinea pig, using retrograde horseradish peroxidase axonal transport techniques. Saccular efferent neurons were discovered bilaterally in the subependymal granular layer at the base of the fourth cerebral ventricle and laterally to the facial nerve genu ispsilaterally in the parvocellular reticular nucleus, as well as nuclei of the superior olivary complex: the lateral olivary nucleus and lateral nucleus of the trapezoid body. Cochlear efferent neurons are located ipsilaterally in the pontine reticular caudal nucleus, in the anteroventral cochlear nucleus, and in the lateral and medial olivary nuclei. Neurons were found contralaterally in the medial nucleus of the trapezoid body. It thus emerged that location zones of vestibular saccular efferent neurons and those of cochlear efferent units partially overlapped. The possible involvement of saccular vestibular efferent neurons in the mechanisms of auditory perception is discussed.A. A. Bogomolets Institute of Physiology, Academy of Sciences of the Ukrainian SSR, Kiev. Translated from Neirofiziologiya, Vol. 22, No. 5, pp. 657–665, September–October, 1990.     

Prostacyclin (PGI2) induces rapid depletion of cyclic AMP levels in brain nuclei of rats

Intravenous injection of prostacyclin (100 micrograms/kg) in rats resulted in a decrease of systolic blood pressure within 2 minutes. Concentrations of cAMP in 15 brain regions and nuclei were determined by radioimmunoassay. In lower brain stem nuclei, such as the nucleus of the solitary tract and the lateral reticular nucleus (A1 and C1 catecholaminergic cell groups) cAMP levels were depleted significantly, while in others, including the locus coeruleus and the periaqueductal central gray, cAMP levels did not show any alterations. Levels of cAMP were also depleted in some of the hypothalamic nuclei (periventricular, anterior hypothalamic, ventromedial), and in cerebral cortical areas. Lowered cAMP levels in brain areas might indicate lower cellular activity in cells participating in baroreceptor control mechanisms.     

A serotonergic system in the brain of the Antarctic fish, Trematomus bernacchii

The immunohistochemical distribution of serotonin-containing nerve fibres and cells has been described in the brain of the Antarctic fish, Trematomus bernacchii. The largest serotonergic system was associated with the diencephalic and rhombencephalic ventricles. In particular, serotonin-positive cells have been found in the lateral recess and neuropile zone of the diencephalic ventricle, where we have identified the serotonergic portion of the paraventricular organ. Numerous serotonin cells were localized in the dorsal nucleus of the raphe, the dorsal tegmental nucleus and the central gray. Two large cell groups, arranged in a pair of well-defined columns and connecting the central gray with the dorsal reticular formation, were immunostained in the region of the trigeminal nuclei. In addition, few positive cells have been found in the preoptic area and the cerebellar valvula, and few serotonergic nerve fibres, probably belonging to the lateral lemniscus, have been identified. The distribution of serotonin elements in the brain of T. bernacchii has been compared with that described in other fish, where it showed some modifications in the immunoreactive pattern. Finally, the lack of a serotonergic system at the level of the reticular superior formation has been reported; however, it was not possible to rule out a phylogenetic or environmental explanation.     

Feeding in satiated sheep elicited by intraventricular injections of CSF from fasted sheep

Cerebrospinal fluid (CSF) was collected from the lateral ventricles of either 24-hr fasted or “satiated” sheep (donors) and injected into the lateral or third ventricle of either 24-hr fasted or “satiated” sheep (recipients). Recipient “satiated” sheep ate more following injections of CSF from fasted donors, as compared to when the same animals were injected with CSF from equally “satiated” donors (e.g., 73±10 vs 16±6g respectively, 15 min post-injection, P <0.01). Feed intake of fasted recipient sheep was slightly depressed following intraventricular injections of CSF from “satiated” donors. Apparently, the composition of cerebrospinal fluid of the donor sheep was affected by the surfeit-deficit state of their energy stores.     

The role of cholecystokinin octapeptide (CCK-8) in regulating thyrotropin secretion in rats. In vivo studies

The effect of cholecystokinin octapeptide (CCK-8) on basal and TRH-stimulated secretion of TSH was investigated in 67 male Sprague-Dawley rats. Blood for TSH determinations was sampled by cannulation of right heart auricle in urethane narcosis before and four times during 60 minutes following CCK-8 administration. It was found that CCK-8 administered to the lateral brain ventricle at a dose of 0.5 microgram per animal caused a decrease in blood serum TSH concentration but did not change the response of TSH to TRH. Intravenous administration of CCK-8 at doses of 2 and 20 micrograms/kg had no effect on blood serum concentration of TSH.     

Calcitonin gene-related peptide: Brain and spinal action on intestinal motility

The effects of intracerebroventricular (ICV) and intrathecal (IT) administration of calcitonin gene-related peptide (CGRP) on intestinal motility were examined in conscious rats chronically fitted with intraparietal electrodes in the duodeno-jejunum and a cannula in a cerebral lateral ventricle or catheter in the subarachnoid space. ICV administration of CGRP (0.5–10 μg) restores the fasted pattern of intestinal motility in fed rats in a dose-related manner. Intrathecal administration of CGRP or calcitonin also induces fasted pattern but after a 30 min delay. These effects persisted after transection of the spinal cord and no change in intestinal motility appeared after intravenous administration of CGRP at a dose effective when given IT. This study suggests that CGRP, as calcitonin, has a neuromodulatory role in the control of intestinal motility at both brain and spinal cord levels.     

Biodistribution of the radioprotective drug 35S-labeled 3-amino-2-hydroxypropyl phosphorothioate (WR77913)

3-Amino-2-hydroxypropyl phosphorothioate (WR77913), a less toxic phosphorothioate radioprotector than WR2721, has been labeled with 35S. The biodistribution of a radioprotective dose of 800 mg/kg was determined in C3H mice bearing RIF-1 tumors as a function of time after intraperitoneal injection and was expressed as percentage injected dose/gram (% ID/g). Levels of 35S in the blood peaked 10 min after injection, and radioactivity in most tissues was highest at 15 min. Label in most tissues declined markedly between 15 and 60 min, but in gut, salivary glands, tumor, and brain, the levels of radioactivity remained quite stable over 1 hr. At 30 min after injection the highest levels of labeled drug were found in submandibular salivary glands, gut, and kidney, with the lowest level in brain. Tumors had approximately the same amount of label as blood, muscle, skin, and esophagus. Two principal differences between the distribution of label from WR77913 and WR2721 were defined. Although blood levels of 35S-WR2721 also peaked 10 min after injection, the 10-min blood levels achieved for WR77913 were more than fourfold greater than those attained by WR2721. Maximum levels of WR2721 occurred in most tissues 30 to 60 min after administration of the drug, compared to 15 min for WR77913. The basis for these differences remains to be determined, but these results suggest that the optimum interval between administration of WR77913 and irradiation may be shorter than for WR2721.     

Taurine Levels in Discrete Brain Nuclei of Rats

Concentrations of taurine have been measured in 44 microdissected rat brain nuclei or areas. Taurine is ubiquitously present and distributed unevenly in the rat brain: the ratio of the highest (pyriform cortex) to lowest (midbrain reticular formation) concentrations is 4.7:1. High taurine levels were found in cerebral cortical areas, caudate-putamen, cerebellum, median eminence, and supraoptic nucleus. Acute pain stress reduced taurine levels in the hypothalamus and the lower brainstem nuclei but not in cortical areas. Increased locomotor and behavioral activities following a high dose of amphetamine elevated taurine concentrations significantly in the substantia nigra and locus ceruleus.     

Naloxone blocks the excitatory effect of ethanol and chlordiazepoxide on lateral hypothalamic self-stimulation behavior

Ethanol (ETOH, 0.5 g/kg) and chlordiazepoxide (CDP, 4.0 mg/kg) enhance operant responding by rats for lateral hypothalamic (LH) self-stimulation (SS). Naloxone (NOX, 5.0 mg/kg) does not affect LH SS or blood alcohol level, but prevents the increased responding for LH SS produced by ETOH and CDP. Thus, ETOH and CDP, but not brain stimulation reward itself, may release an endogenous opioid whose action at opiate receptors results in an excitatory behavioral (euphorigenic?) effect which can be blocked by NOX.     

Permeability of the blood-brain barrier to fructose and the anaerobic use of fructose in the brains of young mice

—Fructose levels were determined in plasma and brain of 8- to 12-day-old mice at intervals after the injection of 30 mmol/kg intraperitoneally; controls received NaCl, 15 mmol/kg. In normal animals brain fructose increased very slowly despite a rapid rise in plasma levels (120 times the control value in 5 min). At 40 min the cerebral level was 1.54 ± 0.23 mmol/kg; the corresponding plasma level was 47.1 ± 4.8 mM. The data suggest that fructose can serve as a source of energy to the brain in times of critical need: during insulin hypoglycemia brain fructose increased to only 0.88 ± 0.05 mmol/kg during the same interval (40 min) despite plasma fructose values equal to those in control animals; also 30 s after cerebral ischemia (decapitation) brain fructose fell from a zero time value of 1.19 ± 0.09 mmol/kg (20 min after fructose injection) to 0.76 ± 0.06 mmol/kg (P= 0.005). Under both circumstances (hypoglycemia and ischemie anoxia) an apparent threshold concentration of fructose for utilization was observed—0.6–0.7 mmol/kg. The most likely explanation for this finding appears to be that this level of fructose was in the extracellular space of the brain. Hexokinase activity in brain homogenates of 8- to 12-day-old mice with fructose and ATP at concentrations found in vivo and during ischemie anoxia did not appear to be rate-limiting. We concluded that the major handicap to the use of fructose by the brain was the limited penetration of fructose from the blood to the brain.     

Effect of acute surgical vagotomy on the mucosal content of 6-keto-PGF1 alpha, PGE2 and glutathione after intragastric 96% ethanol treatment in rats.

It has been observed earlier that gastric cytoprotection produced by PGI2, beta-carotene, small doses of atropine or cimetidine has failed in surgically vagotomized rats. This phenomenon may be in connection with endogenous prostaglandins (PGs) and glutathione (GSH) level of the gastric mucosa. The aims of the study were to evaluate the effect of vagus nerve on the gastric mucosal 6-keto-PGF1 alpha, PGE2 and glutathione after intragastric 96% ethanol (ETOH) treatment. The observations were carried out on CFY rats. The gastric mucosal damage was produced by intragastric administration of 1 ml 96% ETOH. Acute bilateral surgical vagotomy (ASV) was carried out 30 min prior to ETOH application. The animals were sacrificed 1, 5, 15 or 60 min after ETOH installation. The number and the severity of gastric mucosal lesions were noted and 6-keto-PGF1 alpha, PGE2 an GSH contents of gastric mucosa were measured. It has been found that: 1. the number and the severity of gastric mucosal lesions were increased after ASV compared to those with intact vagal nerve, 2. 96% ETOH treatment increased both the gastric mucosal PGs and GSH levels, 3. 6-keto-PGF1 alpha peaked at 5 min PGE2 and GSH peaked at 15 min after ETOH treatment, 4. ASV decreased the gastric mucosal PGs content and delayed the peaks of PGE2 and GSH. It has been concluded that the decreased content of PGs and the delayed GSH increase may play a pathological role in the failure of gastric cytoprotection of rats after ASV.     

Comparative Effects of Ethanol and Malnutrition on the Development of Catecholamine Neurons: Changes in Neurotransmitter Levels

Abstract: The comparative effects of exposure to ethanol and malnutrition on the concentrations of tyrosine and catecholamines in whole brain and selected regions of brain have been studied in the developing rat. These animals were the offspring of optimally nourished rats (control pups), of rats fed a diet with 35% of the calories supplied by ethanol (ETOH pups), or of animals fed a diet calorically equivalent to the latter but lacking ethanol (iso-caloric, 1C pups). These diets were administered to dams either during the last week of gestation (prenatal) or during lactation (postnatal). Tyrosine levels were elevated prior to birth in the prenatal ETOH or IC pups or at 1 and 2 weeks of age in postnatal ETOH or 1C pups as compared with values found in the control offspring. Dopamine concentration in whole brain was significantly lower in prenatal ETOH pups than in prenatal IC pups at 3 weeks of age. Levels in the brains of postnatal ETOH pups were lower than control values, but not relative to animals exposed to 1C diet. Investigation of corpus striatum showed a significant decrease in dopamine concentration compared with control or IC pup values as a result of postnatal exposure to ethanol. Norepinephrine levels in the whole brain of prenatal ETOH pups were consistently 30–40% lower than either control or matched 1C pups during development. At 3 weeks of age, the norepinephrine levels in the hypothalamus of animals exposed to ethanol pre or postnatally were 30–60% lower than values in the corresponding region in either control or 1C pups. In the rat model described, ethanol caused a decrease in catecholamine levels, perhaps solely by affecting the norepinephrine neurons.     

Modification by physostigmine of the cardiovascular responses to intracerebroventricular administration of 5-hydroxytryptamine in rats

In rats the effect of inhibition of the brain cholinesterase activity on the pressor and heart rate responses to 5-hydroxytryptamine (5-HT), administered into the lateral cerebral ventricle (l.c.v.) was examined. After administration of physostigmine (twice in a small dose of 2.5 micrograms l.c.v., 20 and 15 min before the second injection of 5-HT), the pressor effect of 5-HT (5 micrograms) was strongly reduced or almost abolished, its pure tachycardia was reduced or reversed into a bradycardia and its pure bradycardia was diminished or reversed into a tachycardia. The type of the cardiovascular response to ACh (5 micrograms l.c.v., 20 min after the second administration of 5-HT) indicates that the modification of the cardiovascular response to 5-HT was accompanied by inhibition of the brain cholinesterase activity. Thus, it seems that a functionally competent cholinesterase in the brain is necessary for the generation of the 5-HT-induced pressor response. The present experiments provide further evidence that there is a cholinergic link in the pathway by which serotonergic mechanisms in the preoptic-anterior hypothalamic area rise blood pressure and support the idea that the same link exists in the pathway(s) mediating the heart rate responses to intracerebroventricular administration of 5-HT.     

DISTRIBUTION OF FOUR POTENTIAL TRANSMITTER AMINO ACIDS IN MONKEY RETINA

Discrete layers from frozen dried sections of Rhesus monkey retina were analyzed for each of four amino acids. Peak levels of glycine were found near the border of the inner nuclear and inner reticular layers, and were high throughout these two layers. The levels were less than 50% of the peak in the adjacent ganglion cells and outer reticular layers and fell to very low levels elsewhere. GABA was much more sharply restricted to the inner reticular layer and fell off on both sides to levels of 10% or less of the peak in the fiber and photoreceptor cell layers. Glutamate and aspartate were highest in the ganglion cell layer. On a lipid-free dry weight basis the peak aspartate level was about twice that of brain. Moderately high levels of both aspartate and glutamate were found in the inner reticular and fiber layers. Elsewhere the levels ranged from 20 to 50% of the peak, and both amino acids were relatively low in optic nerve. The amino acid distributions are compatible with a transmitter function for GABA in amacrine cells and for glycine in horizontal and amacrine cells. Glutamate and aspartate may be especially high in Müller fibers, ganglion cells or both.