The Journal of Mind and Behavior

Volume 8, Number 4 ~ Special Issue, Inhibition in the Brain

Part I: Molecular Biology and Localization of GABA and Glycine

Biochemistry of Glycinergic Neurons
Edward C. Daly, Roudebush VA Medical Center and Indiana University Medical Center
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 477 [1]-490 [14], ISSN 0271-0137, ISBN 0-930195-04-3
As in other organs, the ß-carbon of serine, generated in the conversion of serine to glycine by serine hydroxymethyltransferase (SHMT), is probably the major source of de novo one-carbon units within the central nervous sytem (CNS). SHMT also catalyzes the final reaction in the major pathway for the synthesis of glycine from glucose. The synthesis of glycine via SHMT is the major source of glycine within the CNS and the in vitro activity of the SHMT correlates fairly well with the tissue content of glycine within regions of the rat CNS. As glycine functions in the additional capacity of a neurotransmitter within at least certain regions of the CNS, these two neurochemical demands, one-carbon generation and neurotransmitter glycine synthesis, would require careful, and perhaps unique, regulation of the interconversion of serine and glycine within the CNS. Within the liver of the rat, SHMT exists as separate mitochondrial and cytoplasmic isoenzymes, but the subcelluar distribution of SHMT within the CNS reveals almost exclusively a mitochondrial localization. An indefinite portion of the mitochondrial SHMT in liver is associated with the glycine cleavage system (GCS), the major pathway of glycine catabolism in the liver. However, very low levels of the GCS are found in homogenates of the medulla/pons and spinal cord, two regions in which glycine is thought to function as a neurotransmitter. These two aspects of the interconversion of serine and glycine within the CNS (predominance of mitochondrial SHMT and the marked regional distribution of the GCS) may be manifestations of unique neurochemical mechanisms to regulate the supply of glycine and one-carbon units. Current models for the interrelationship of serine, glycine, and one-carbon generation in the periphery do not appear to be applicable to the known CNS biochemistry of SHMT and the GCS. A model under investigation in our laboratory is outlined and provides for: (a) the synthesis of glycine with minimal perturbation of the one-carbon pool; (b) the degradation of glycine with minimal perturbation of the one-carbon pool; and (c) the generation of the one-carbon units with a minimum perturbation of intracellular glycine levels.
Requests for reprints should be sent to Edward C. Daly, M.D., Ph.D., Departments of Neurology and Neurobiology (Psychiatry), Indiana University Medical Center, VAMC, 1481 W. Tenth Street, Indianapolis, Indiana 46202.

Immunocytochemical Characterization of Glycine and Glycine Receptors
R.J. Wenthold, National Institutes of Health and R.A. Altschuler, University of Michigan
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 491 [15]-502 [26], ISSN 0271-0137, ISBN 0-930195-04-3
Antibodies against the glycine postsynaptic receptor and against glycine conjugated to BSA with glutaraldehyde were used to immunocytochemically localize and characterize glycinergic neurons in auditory nuclei of the brainstem. Intense glycine receptor immunoreactivity is found throughout the cochlear nucleus, appearing as puncta often present around cell bodies. In the superior olivary complex, the most intense labeling is present in the lateral superior olive. Electron microscopic studies show that immunoreactivity is restricted to the synaptic region. With mAb GlyR 2 antibody, which recognizes the ligand binding subunit of the glycine receptor, reaction product is present in the synaptic cleft; while mAb GlyR 7 antibody, which recognizes an associated protein, shows reaction product on the cytoplasmic side of the postsynaptic membrane. Labeling is restricted to specific populations of synapses, and in the cochlear nucleus it is often associated with terminals containing flattened synaptic vesicles. Glycine immunoreactivity is seen in cell bodies and in puncta, which resemble presynaptic terminals. Most intense cell body labeling is in the medial nucleus of the trapezoid body. These cells project to the lateral superior olive where heavy glycine receptor labeling was seen. In the cochlear nucleus several populations of glycine immunoreactive cell bodies are found, with most being in the dorsal cochlear nucleus. A population of glycine immunoreactive cells in the dorsal cochlear nucleus appears similar to a population of GABA immunoreactive cells identified in an earlier study. To determine if the same cells are labeled with both antibodies, a double label study was done using rabbit anti-glycine and guinea pig anti-GABA. This confirmed that some cells in the dorsal cochlear nucleus are immunoreactive for both glycine and GABA. Similar results were seen in the cerebellum where Golgi cells were glycine immunoreactive, whereas other GABA containing cells were not labeled.
Requests for reprints should be sent to R.J. Wenthold, PhD., National Institutes of Health, Laboratory of Neuro-otolaryngology, Building 36, Room 5D08, Bethesda, Maryland 20892.

Distribution of Inhibitory Amino Acid Neurons in the Cerebellum With Some Observations on the Spinal Cord: An Immunocytochemical Study With Antisera Against Fixed GABA, Glycine, Taurine, and ß-Alanine
Ole P. Otterson and Jon Storm-Mathisen, University of Oslo
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 503 [27]-518 [42], ISSN 0271-0137, ISBN 0-930195-04-3
We have raised antisera against three inhibitory amino acid transmitter candidates: GABA, glycine and taurine. The immunogens were amino acid-glutaraldehyde-carrier conjugates. All sera were purified by immunosorption with the glutaraldehyde-treated carriers, and with one or more amino acid-glutaraldehyde-bovine serum albumin conjugates made from amino acids different from those used for immunization. All purified antisera yielded specific staining that could be abolished by preincubation with glutaraldehyde complexes of the amino acid against which the serum was raised, but not by complexes of other amino acids. The three antisera produced very different staining patterns in the normal cerebellum. Immunoreactivity for GABA occurred in cell bodies and processes of stellate, basket and Golgi cells, and in the axons of Purkinje cells. However, it was low in Purkinje cell bodies and dendrites. Immunoreactivity for glycine was selectively localized in Golgi cells, and was found to co-exist with that for GABA in a subpopulation of Golgi cells. Immunoreactivity for taurine was found in all Purkinje cells, but only in a very small proportion of the neurons in the molecular layer. These results show that the different inhibitory amino acids have selective distributions in the cerebellum.
Requests for reprints should be sent to Ole P. Otterson, M.D., Anatomical Institute, University of Oslo, Karl Johansgt. 47, 0162 Oslo 1, Norway.

GABA-Peptide Neurons of the Primate Cerebral Cortex
Edward G. Jones, University of California, Irvine
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 519 [43]-536 [60], ISSN 0271-0137, ISBN 0-930195-04-3
The neuropeptide containing neurons of the neocortex do not seem to be a highly heterogeneous group as commonly supposed. Virtually all of the known peptides appear to be contained in a limited type, perhaps a single class, of GABAergic intrinsic neuron. VIP may be more commonly found in a cholinergic neuron that shares the morphological features of the GABA-peptide class. The majority of cortical neurons, including all the pyramidal neurons and spiny intrinsic neurons, amounting to about 75% of the total neuronal population, and most varieties of the 25% GABAergic intrinsic neurons, are not immunoreactive for known peptides. The roles of those peptides, that are apparently co-released with GABA in cortical funtion, are still unknown.
Requests for reprints should be sent to Edward G. Jones, M.D., Ph.D., Department of Anatomy and Neurobiology, University of California, Irvine, California College of Medicine, Irvine, California 92717.

Part II: Functional Role of GABA and Glycine
GABAergic Inhibition in the Neocortex
K. Krnjevic, McGill University
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 537 [61]-548 [72], ISSN 0271-0137, ISBN 0-930195-04-3
Though long suspected, inhibition in the neocortex has become fully accepted as a major mechanism of control only in the last 25 years. The principal mechanism is a large increase in chloride permeability, mediated by the synaptic release of GABA from a variety of interneurons. There is still no evidence of any significant glycinergic inhibition in cortex. The functional role of cortical inhibition has been demonstrated in several ways: by global inactivation (with GABA depletion or antagonists) which causes paroxysmal activity; or by very localized release of GABA antagonists, which reveals the important role of GABAergic inhibition in determining both qualitative and quantitative aspects of receptive-field characteristics in visual and somatosensory cortex. Prolonged synaptic inhibitions are probably generated by an increase in potassium permeability. They probably account for some bicuculline resistant inhibitory actions and may be mediated either by a non-bicuculline sensitive GABA action (via GABA B receptors) or by non-GABAergic pathways.
Requests for reprints should be sent to K. Krnjevic, M.D., Anaesthesia Research Department, McGill University, McIntyre Medical Sciences Building, 3655 Drummond Street, Room 1208, Montreal, Quebec, Canada H3G 1Y6.

Physiology of GABA Inhibitions in ths Hippocampus
R.C. Malenka, R. Andrade and R.A. Nicoll, University of California, San Francisco
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 549 [73]-558 [82], ISSN 0271-0137, ISBN 0-930195-04-3
Application of GABA to hippocampal pyramidal cells causes three types of responses. Low doses applied at the soma cause a chloride-and bicuculline-sensitive huperpolarization mediated by GABA_A receptors. Low doses applied in the dendrites cause primarily a bicuculline-sensitive depolarization which may also be mediated by chloride-dependent GABA_A receptors. Higher doses of GABA elicit a bicuculline-resistant hyperpolarization which is due to the opening of K+ channels by GABA_B receptors. The coupling of GABA_B receptors to K+ channels requires a GTP binding protein which may directly activate the K+ channel. Stimulation of afferent fibers in the hippocampus elicits a fast and slow IPSP. The fast IPSP is due to the activation of GABA_A receptors and it is proposed that the slow IPSP is due to the activation of GABA_A receptors. Under conditions of enhanced release of GABA or enhanced responsiveness of GABA_A receptors with barbiturates a very slow (100's of msecs to secs) bicuculline-sensitive depolarizing IPSP can be recorded. The different properties of these putative GABAergic IPSPs make it likely that they subserve distinct functions in the hippocampus.
Requests for reprints should be sent to R.A. Nicoll, M.D., Department of Pharmacology, School of Medicine, University of California, San Francisco, California 94143.

Inhibitory Processes in the Thalamus
M. Steriade and M. Deschenes, Universite Laval
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 559 [83]-572 [96], ISSN 0271-0137, ISBN 0-930195-04-3
Three major thalamic inhibitory processes are involved in the genesis of synchronized EEG spindle oscillations during sleep, in spindle disruption upon arousal, and in local synaptic operations within glomeruli that are probably related to discrimination processes during wakefulness. Spindle oscillations are characterized intracellularly by rhythmic (7 to 14 Hz) depolarizations and spike burst in GABAergic reticular (RE) thalamic neurons, with the consequence of rhythmic huperpolarization-rebound sequences in thalamocortical neurons. The rebound component triggers postsynaptic events in cortical neurons, within the frequency range of spindles. The absence of spindle rhythms in thalamic nuclei deprived of inputs from the RE nucleus and the preservation of spindle rhythmicity in RE neurons disconnected from the thalamus and cerebral cortex demonstrate that the RE nucleus is the pacemaker of spindle oscillations. The disruption of spindle oscillations upon arousal from sleep is realized by upper brainstem reticular afferents to thalamic and basal forebrain structures. Stimulation of peribrachial (pedunculopontine) nucleus elicits, in addition to an early excitation, a long-lasting hyperpolarization associated with a marked conductance increase in RE neurons. This effectively blocks ongoing spindle sequences in the thalamic pacemaker.
Requests for reprints should be sent to M. Steriade, M.D., Ph.D., Laboratoire de Neurophysiologie, Faculte de Medecine, Universite Laval, Quebec, Canada G1K 7P4.

Neurotransmitter Modulation of Thalamic Neuronal Firing Pattern
David A. McCormick and David A. Prince, Stanford University Schoolof Medicine
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 573 [97]-590 [114], ISSN 0271-0137, ISBN 0-930195-04-3
Thalamic neurons can generate two basic patterns of neuronal activity (burst firing and single spike activity) depending upon the membrane potential of the cell. Here we review the mechanisms of action of acetylcholine (ACh), norepinephrine (NE) and _T-amino-butyric acid (GABA) in the thalamus to show how these agents modulate the firing mode of thalamic neurons. ACh can cause three different responses in the thalamocortical relay cells: a fast nicotinic depolarization; an increase in potassium conductance (muscarinic); and a decrease in membrane potassium conductance (muscarinic). Norepinephrine causes a decrease in membrane potassium conductance. GABA causes an increase in membrane chloride conductance followed by an increase in potassium conductance. The transmitter induced increases in potassium conductance result in the inhibition of single spike activity and the promotion of burst firing, while the decreases in potassium conductance have the opposite effect. GABA-induced increases in chloride conductance inhibited both burst firing and single spike activity. The alterations in ionic currents produced by activation of ascending brainstem cholinergic and noradrenergic systems and local GABAergic circuits can thus have a variety of consequences for regulation of activity in thalamocortical circuits.
Requests for reprints should be sent to David A. McCormick, Ph.D., Section of Neuroanatomy, Yale University School of Medicine, 333 Cedar Street, New Haven, Connecticut 06510.

What Do GABA Neurons Really Do? They Make Possible Variability Generation in Relation to Demand
Eugene Roberts, Beckman Research Institute of the City of Hope
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 591 [115]-604 [128], ISSN 0271-0137, ISBN 0-930195-04-3
It is proposed that GABA neurons play key roles in maintaining meaningful communications within and among neural units by making possible variability generation in relation to demand. Activities of GABAergic inhibitory projection neurons from command centers and local circuit GABAergic inhibitory interneurons allow adaptive nervous system function to take place in a manner characterized by freedom without license. Through their multiple activities and connections these neurons make possible smooth transitions between modes of nervous system function over a range of increasing demands (neural pressures), enabling organisms to explore full ranges of their options.
Requests for reprints should be sent to Eugene Roberts, Ph.D., Department of Neurobiochemistry, Beckman Research Institute of the City of Hope, Duarte, California 91010.

Part III: Functional Aspects of Inhibition Related to Neurological Diseases
GABAergic Abnormalities Occur in Experimental Models of Focal and Genetic Epilepsy
Charles E. Ribak, University of California, Irvine
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 605 [129]-618 [142], ISSN 0271-0137, ISBN 0-930195-04-3
Two types of experimental models of epilepsy were studied with morphological methods. The first type resembles post-traumatic focal epilepsy and can be generated by alumina gel implants into the sensorimotor cortex of monkeys. The epileptic focus in these monkeys displays a preferential loss of cortical GABAergic neurons and axon terminals. Recent studies indicate that this loss occurs following the alumina gel treatments and prior to the onset of clinical seizures. These findings add further support to the hypothesis that a loss of GABAergic inhibition plays a causal role in focal epilepsy. Two genetic models of epilepsy were analyzed with biochemical and immunocytochemical methods for GABAergic neurons and synapses. The seizure-sensitive gerbil and the genetically epilepsy-prone rat display an increase in the number of GABAergic neurons and terminals in specific brain regions that also show significant increases in the biochemical level of GABA as compared to non-epileptic animals of the same species. These brain regions are essntial for seizure activity because lesions in these areas block seizures. Furthermore, it appears that these brain regions are involved in the analysis of the stimuli that generate seizures. One possible hypothesis to explain seizures in these genetic models is increased disinhibition of excitatory projection neurons in the affected brain regions. These findings suggest that GABAergic abnormalities occur in experimental models of both focal and genetic epilepsy and provide further support for the use of antiepileptic drugs that act at the GABA receptor.
Requests for reprints should be sent to Charles E. Ribak, Ph.D., Department of Anatomy and Neurobiology, University of California, Irvine, California 92717.

Inhibition, Local Excitatory Interactions and Synchronization of Epileptiform Activity in Hippocampal Slices
F. Edward Dudek, Tulane University School of Medicine and Edward P. Christian, University of Maryland School of Medicine
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 619 [143]-634 [158], ISSN 0271-0137, ISBN 0-930195-04-3
Inhibition counteracts local excitatory influences between hippocampal neurons in the normal brain, thus maintaining independent channels for information processing. Removal of synaptic inhibition has provided the most widely employed experimental model for epileptogenesis. Inhibitory mechanisms reduce the efficacy of excitatory chemical synapses, thereby increasing the probability of transmission failure through local excitatory pathways. After removal of inhibition, local excitatory chemical synapses recruit increasing numbers of pyramidal cells in a positive feedback manner. Differences in the epileptogenicity of various areas of the hippocampus (e.g., CA3 versus CA1), when synaptic inhibition is blocked, may be due largely to the density and strength of recurrent excitation. Evidence is also available for electrotonic coupling and electrical field effects among hippocampal neurons. Electrical mechanisms can synchronize hippocampal neurons when chemical synapses are blocked with low-[Ca²⁺] solutions. Removal of inhibition would also be expected to strengthen electrical interactions, although less is known about them. Electrical interactions probably combine with local chemical excitation to synchronize neurons when inhibition is compromised.
Requests for reprints should be sent to F. Edward Dudek, Ph.D., Mental Retardation Research Center, UCLA School of Medicine, 760 Westwood Plaza, Los Angeles, California 90024.

Inhibition in Huntington's Disease
M. Flint Beal, David W. Ellison and Joseph B. Martin, Massachusetts General Hospital
The Journal of Mind and Behavior , Autumn 1987, Vol. 8, No. 4, Pages 635 [159]-642 [166], ISSN 0271-0137, ISBN 0-930195-04-3
Huntington's disease (HD) is an autosomal dominant disease in which severe atrophy of the basal ganglia is accompanied by progressive dementia and chorea. An initial biochemical observation was that there was a marked deficiency of GABA and its biosynthetic enzyme glutamate decarboxylase (GAD) in HD basal ganglia. This was true of both the striatum and its sites of projection, the globus pallidus and substantia nigra. In our own studies we have confirmed the GABA deficiency and have shown that it correlates with pathologic grade. There is a gradient of GABA loss with the caudate being most severely affected followed by the putamen and nucleus accumbens. There were no significant changes in cerebral cortex. Studies of GABA receptors have shown reductions in the striatum with increased numbers of receptors in the pallidum, consistent with denervation hypersensitivity. Numerous trials of GABA replacement therapy using various agents have been unsuccessful despite evidence that these agents increase CSF concentrations of GABA. Therefore, the GABA deficiency alone is unlikely to be crucial for the clinical manifestations of HD. GABA deficiency appears to be a marker for loss of striatal spiny neurons in HD and knowledge of its deficiency has led to improved animal models of the disease.
Requests for reprints should be sent to M. Flint Beal, M.D., Neurology Research 4, Massachusetts General Hospital, Boston, Massachusetts 02114.