Walter J. Freeman Journal Article e-Reprint

How does diazepam work? Do its antihistaminic properties contribute to its actions? Cell biology gives answers, but neural networks offer different and more penetrating answers. The purpose of this review is to point out what new experiments should be done on the basis of modeling.

These mathematical models of the functions of neurons in networks and in brain systems cannot prove experimental results. They can bring out new relationships among disparate data and suggest new hypotheses for experimental testing. A case in point is the complex pharmacology at the intersection of research on GABA (gamma aminobutyric acid), benzodiazepine receptors, brain amine systems, and anxiolytics.

Few myths are more firmly entrenched in modern neurobiology than the view that GABAergic neurons are inhibitory -- period. A counter-example is derived from the mammalian olfactory bulb, which has two main populations of GABAergic interneurons (Ribak et al 1977). The internal granule cells have dendrodendritic reciprocal synapses with mitral cell basal dendrites, and the external granule (periglomerular) cell have axons synapsing onto each other and onto mitral cell apical dendrites in the glomeruli.

A neural network model of periglomerular function predicted nearly twenty years ago that, whereas the internal granule cells inhibited the mitral cells, the external cells excited each other and mitral cells as well (Freeman 1975). The model indicated that the periglomerular cells generated an excitatory bias that was essential to "turn on" (arouse) the olfactory system. This proposal was vigorously disputed (Getchell and Shepherd 1975), in part owing to a confusion between presynaptic (Voronkov and Gusel'nikova 1969) and postsynaptic -(Shepherd 1974). inhibition, but experimental evidence has since accumulated (Nowycky et al 1981; Martinez and Freeman 1984; Rhoades and Freeman 1990), largely driven by the logic of mathematical models (Freeman 1987), that the main action of periglomerular cells is excitatory.

The study of GABAergic neurons is made extremely difficult by their small size, dense packing, and multiplicity of synaptic connections. The difficulty is compounded in the bulb by the close contiguity of two populations with opposing actions. The most compelling histochemical evidence thus far is the recent finding, based on a technique pioneered 40 years ago by van Harreveld and Schade (1960), that some periglomerular dendrites and probably the mitral cell apical dendrites as well, accumulate intracellular chloride (Siklos et al., submitted). The finding implies that activated GABA-A receptors depolarize these cells by an outflow of chloride ions from their dendrites.

These observations were made in response to experiments in which the application of chloride-free solutions (Cook et al 1991), and likewise the application of histamine (Rhoades and Freeman 1988), tended to enhance bulbar activity. In the past two years there have been reports of excitatory actions of GABA in several other systems, including the hippocampus (Michelson and Wong 1991), hypothalamus (Charles et al 1992), retina (Kamermans and Werblin 1992), lateral geniculate nucleus (McCormick and Williamson 1991) and superior colliculus (Arakawa and Okada 1988), adding to the extensive literature from invertebrates on the variety of actions of GABA, and showing that, like acetylcholine and norepinephrine, GABA is not a "single-action" neurotransmitter.

The significance of the findings extends beyond experimental pharmacology, because they offer a new perspective on possible roles of histamine in brain function. The histaminergic system suffers unjustified neglect, as demonstrated by a recent review article in IBRO News (Robbins et al 1992) on arousal systems of the brain, covering four amines - norepinephrine, dopamine, serotonin and acetylcholine - but failing to mention histamine. There is substantial histochemical evidence for the wide distribution in the forebrain of histaminergic axons arising from nuclei in the posterior hypothalamus (Panula 1990; Wada et al 1991; Watanabe and Wada 1991), and there is equally compelling evidence for a substantial role of histamine in arousal reactions (Lin et al 1990; McCormick and Williamson 1991).

The targets of these histaminergic projections include the olfactory bulb. Our model generates the hypothesis that histamine increases the chloride uptake of periglomerular cells, which increases the depolarizing action of GABA through A receptors and thereby the sustained excitatory bias that activates the olfactory neural networks, leading to arousal. There are major unanswered cellular questions. Do the periglomerular cells have an inwardly directed chloride pump? If there is a pump, is it activated by histamine? Is there any similarity between chloride transport in cortex and in the stomach? Are the EPSPs and the excitatory bias of the periglomerular cells augmented by histamine? Is the concentration gradient of chloride sufficient to explain the physiological effects? Are there other GABAergic populations in the forebrain that provide excitatory bias under histamine control for arousal?

Several major pharmacological questions arise. Do peripherally administered antihistamines decrease or reverse the putative chloride transport? Is the rate of inward chloride transport in periglomerular cells augmented during anxiety states, and is it reduced by anxiolytics and sedatives? An obvious inference is that centrally acting antihistamines, phenothiazines and benzodiazepines may have a common thread of action through chloride channels and GABA and histamine receptors. Conversely, is the rate of chloride transport abnormally low in clinical depression, as a part or even a main aspect of diminished arousal? The bulbectomized rat has been described as the best available animal model for clinical depression, (Jesberger and Richardson 1985; van Riezen and Leonard 1990). Might a search among agents that promote chloride transport reveal new ways of treating clinical depression?

The extant literature on the pharmacological actions of behaviorally neuroactive substances relating to chloride transport and GABA (e.g. Morrow et al 1988; van Riezen and Leonard 1990; Zorumski and Isenberg,1991; File, 1991; Amejdki-Chab et al 1992) reveals lack of full understanding of the actions of these agents on the complex synaptic systems involving GABA and the modulator amines. Excitatory actions of GABA have been explained at the cellular molecular levels by synaptic changes (Kamermans and Werblin, 1992; Alkon et al 1992) without reference to histamine, and with no impact as yet on the belief that diazepam acts to enhance an inhibitory action of GABA. There has been little change in this belief for two decades, owing in part to a widespread agnosia about the multifaceted actions of GABA, but even more to an experimental difficulty. Intracellular chloride concentrations are extremely labile.

As observed by van Harreveld and Schade (1960), massive shifts of chloride ions from neurons into glia may occur with experimental trauma to brain tissue. This tendency should greatly concern researchers, because a normal excitatory action of GABA might be reversed to an inhibitory action in standard laboratory assays, owing to a substantial change in the direction of transmembrane chloride concentration gradients for which the GABA-A channels are opened. The questions should be asked, What is the direction of the chloride gradient normally, and what is it as the pharmacological measurements are made? If these controls are not now in hand, then some preceding research may have to be re-done.

In conclusion, a new breakthrough from neural networks may be the insight that GABA can be excitatory, possibly under regulation by histamine. A further point of this story concerns the proliferation of receptor subtypes (Nicoll et al 1990), feedback regulatory mechanisms (Kamermans and Werblin 1992), ionic dependencies (Alkon et al 1992), and multitransmitter interactions (Zorurnski and Isenberg 1991) of neurochemicals. Evolutionary pressures for selection act on behavior, not on single genes or molecular receptors, and the effects of shaping will be seen in the performance of densely interactive masses of neurons. A molecular approach one receptor at a time may further confuse rather than enlighten investigators. Mathematical models that encompass interactive wholes (Mandell and Selz 1992) can give broad pictures that may prove to be necessary for informed development of a more rational psychoneuropharmacology.

Acknowledgements

This work was supported by grant MH06686 from the National Institute of Mental Health.

References

Alkon DL, Sanchez-Andres J-V, Ito E, Oka K, Yoshioka T and Collin C (1992) Long-term transformation of an inhibitory into an excitatory GABAergic synaptic response. Proceedings National Academy of Science USA 89: 11862-11866.

Amejdki-Chab N, Costentin J, Bonnet J-J (1992) Kinetic analysis of the chloride dependence of the neuronal uptake of dopamine and effect of anions on the ability of substrates to compete with the binding of the dopamine uptake inhibitor GBR 12783. Journal of Neurochemistry 58: 793-800.

Arakawa T, Okada Y (1988) Excitatory and inhibitory action of GABA on synaptic transmission in slices of guinea pig superior colliculus. European Journal of Pharmacology 158: 217-224.

Charles A, Sanderson M, Hales T (1992) GABA has excitatory effects on hypothalamic neurons. Transactions, American Society of Neurochemistry 23: 276.

Cook PB, Rhoades BK, Freeman WJ (1991) GABAergic modulation of averaged evoked potentials in rat olfactory bulb. Chapter in: Eeckman FA (ed) Analysis and Modeling of Neural Systems, Boston MA, Kluwer Academic. pp. 311-318.

File SE (ed) (1991) Psychopharmacology of anxiolytics and antidepressants. New York, Pergamon Press.

Getchell TV, Shepherd GM (1975) Short-axon cells in the olfactory bulb: dendrodendritic synaptic interaction. Journal of Physiology (London) 251: 523-548.

Jesberger JA, Richardson JS (1985) Animal models of depression: parallels and correlates to severe depression in humans. Biological Psychiatry 20: 764-784.

Kamermans M, Werblin F (1992) GABA-mediated positive autofeedback loop controls horizontal cell kinetics in tiger salamander retina. Journal of Neuroscience 12: 2451-2463.

Lin JS, Sakai K, Vannimercier G, Arrang JM (1990) Involvement of histarminergic neurons in arousal mechanisms demonstrated with H-3-receptor ligands in the cat. Brain Research 23: 325-330.

Mandell AJ, Selz KA (1992) Dynamical systems in psychiatry: Now what? Biological Psychiatry 32: 299-301.

Martinez DM, Freeman WJ (1984) Periglomerular cell action on mitral cell in olfactory bulb shown by current source density analysis. Brain Research 308: 223-233.

McCormick DA, Williamson A (1991) Modulation of neuronal firing mode in cat and guinea pig LGND by histamine - possible cellular mechanisms of histaminergic control of arousal. Journal of Neuroscience 11: 3188-3199.

Michelson HB, Wong RKS (1991) Excitatory synaptic responses mediated by GABA(A) receptors in the hippocampus. Science 253: 1420-1423.

Morrow AL, Suzdak PD, Paul SM (1988) Benzodiazepine, barbiturate, ethanol and hypnotic steroid hormone modulation of GABA-mediated chloride ion transport in rat brain synaptoneurosomes. In: Chloride Channels and their Modulation by Neurotransmitters and drugs. Biggio G, Costa E. (eds) New York, Raven Press.

Nicoll RA, Malenka RC, Kauer JA (1990) Functional comparison of neurotransmitter receptor subtypes in mammalian central nervous system. Physiological Reviews 70: 513-565.

Nowycky MC, Mori K, Shepherd GM (1981) Blockade of synaptic inhibition reveals long-lasting synaptic excitation in isolated turtle olfactory bulb. Journal of Neurophysiology 46: 649-658.

Panula P, Airaksinen MS, Pirvola U, Auvinen S (1990) A histamine-containing neuronal system in human brain. Neuroscience 34: 127-132.

Rhoades BK, Cook PB, Freeman WJ (1988) Histamine effects on EEG and evoked potentials in the rat olfactory bulb. Neuroscience Society Abstracts 14: 474.2, 1187.

Rhoades BK, Freeman WJ (1990) Excitatory actions of GABA in the olfactory bulb. Neuroscience Society Abstracts 16: 170.2, 403.

Ribak C, Vaughn JE, Saito K, Barber R, Roberts E (1977) Glutamate decarboxylase localization in neurons of the olfactory bulb. Brain Research 126: 1-18.

Robbins TW, Everitt BJ, Muir JL, Harrison A (1992) Understanding the behavioural functions of neurochernically defined arousal syptems. IBRO News 20: 7.

Shepherd GM (1974) Synaptic organization of the Brain. New York, Oxford.

Siklos L, Rickmann M, Joo' F, Freeman WJ, Wolff JR (1995) Chloride is preferentially accumulated in certain intraglomerular dendrites and periglomerular cells of the rat olfactory bulb. Neuroscience 64: 165-172.

Voronkov GS, Gusel'nikova KG (1969) Presynaptic inhibition in the frog olfactory bulb. Neuroscience Translations 7: 775-777.

van Harreveld A, Schade JP (1960) On the distribution and movements of water and electrolytes in the cerebral cortex. In: Tower DB, Schade JP (eds.) Structure and Function of the Cerebral Cortex. Amsterdam, Elsevier, pp 239-256.

van Riezen H, Leonard BE (1990) Effects of psychotropic drugs on the behavior and neurochemistry of olfactory bulbectomized rats. Pharmacology and Therapeutics 47: 21-34.

Wada H, Inagaki N, Yamatodani A, Watanabe T (1991) Is the histaminergic neuron system a regulatory center for whole-brain activity? Trends in NeuroScience 14: 415-418.

Watanabe T, Wada H (1991) Histaminergic Neurons: Morphology and Function. Boca Raton FL, CRC Press.

Zorumski CF, Isenberg KE (1991) Insights into the structure and function of GABAbenzodiazepine receptors: ion channels and psychiatry. American Journal of Psychiatry 148: 162-173.