Volume Versus Wiring Transmission in the Brain: A New Theoretical Frame for Neuropsychopharmacology

Luigi F. Agnati Department of Human Physiology, University of Modena, Modena, Italy

Borje Bjelke and Kjell Fuxe Department of Neuroscience, Karolinska [nstitutet, Stockholm, Sweden

I . Historical Aspects, Basic Concepts, and Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 11. Chemical Messages for Volume Transmission . .

A. Neuropeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 B. Classical Transmitters . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

39 40 40

111. Electrical (Ionic) Messages for Volu IV. Carbon Dioxide as a Regulator of WT and VT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Nitric Oxide as a VT Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

VI. The Replacement Therapy in the Frame of the WT and VT Concept ......................... A. WT Potentiation Therapy . . . . . . . . . . . . . .

C. Pharmacological Interventions of WT and VT .......................................... 42 marks ............................ .... 43 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

B. VT Potentiation Therapy ...............

I. HISTORICAL ASPECTS, BASIC CONCEPTS, AND DEFINITIONS

From 1985 to 1986 we suggested that two types of intercellular communication could be distinguished in the Central Nervous System (CNS): the Volume Transmission (VT) and the Wiring Transmission (WT).1,2 VT is characterized by the signal conduction in a 3-dimensional mode within the extracellular fluid (ECF). Thus the interneuronal commu- nication channel is not confined to the classical substrate proposed by the “neuron doctrine” (see below): presynaptic knob/synaptic cleft/postsynaptic membrane, typical for the WT. Hence, the essence of VT is the conduction of electrical and chemical messages along multiple, largely unpredictable, channels, while the essence of WT is the presence of channels physically constrained to the neuronal network ~ i r i n g . 3 , ~ Previ- ously Guillemin discussed paracrine secretion in brain and Nicholson described the dif- fusion of electrical signals within the interconnected microenvironments of the extracel- lular fluid .5,6 Furthermore Vizi (1984)7 described the so-called nonsynaptic transmission representing a diffusion of chemical signals over a very short distance (Fm) to regulate the presynaptic release of transmitters and Schmitts also in 1984 described the parasyn- aptic system in the brain working in parallel with the synapses involving hormones, transmitters, and growth factors in the control of the electrical and metabolic activity of the networks.

The terms WT and VT derive from very early definitions of the conduction phenome- non. Thus, already Volta (Alessandro Volta, 1745-1827), the inventor of the voltaic pile in 1793, pointed to the existence of two types of conductors for the electrical current: the

Medicinal Research Reviews, Vol. 15 No. 1, 33-45 (1995) 0 1995 John Wiley & Sons, Inc. CCC 0198-6325/95/010033-13

34 AGNATI ET AL.

“metal conductors” (I class conductors, according to his original definition) made by metal wires, and the ”wet conductors” (I1 class conductors, according to his original definition) made by solutions.9 In addition, the term WT is also supported by the mean- ing of the word ”wire,” that may also indicate ”the telegram system i.e. send a message by wire”l0 and this is, in our opinion, a good analogy of the classical synaptic transmis- sion in the CNS. The term VT refers to the theory of volume conduction,” that is based on the ECF capability to work as a I1 class conductor. In fact, this theory describes the flow of ionic currents, generated by nerve cells, through ECF under various conditions of cellular activity.

The existence of such a dichotomy in the interneuronal communication in the CNS,*J was already present at the origin of the modern neuroscience, since Golgi and Cajal had opposite views on the organization of the neurons.’2-*4 Golgi supported the reticular theory of neuronal continuity, originally proposed by Kolliker,l5,16 while Cajal favored the neuron theory of nerve cell contiguity, originally proposed by Waldeyer.17 Actually, Golgi was very cautious about the controversy as it appears from a Golgi’s letter reported by Luciani.18 As a matter of fact, he pointed out that the techniques, at that time available, did not allow to prove or to disprove any one of the two theories, but from a functional standpoint he was inclined to surmise that, in executing several tasks, the CNS works as a holistic system. From the anatomical point of view the controversy has been definitely settled in favor of the neuron theory by electron microscopic studies. Golgi’s view from the functional standpoint has today even more solid grounds than before.19.20

The present concept of VT may explain some aspects of the holistic way of the opera- tion of the CNS that are not easily interpreted on the basis of a wired organization of the brain. Furthermore, the VT concept can also offer the rationale for some otherwise mysterious features of the CNS as, e.g., the existence of several neurotransmitters, when only two (one inhibitory, one excitatory) would be sufficient and necessary for the synaptic transmission. The strangeness of this phenomenon can be fully appreciated on the basis of Occam’s razor (William of Occam, 1290-1350, English philosopher) who stated: “frustra fit per plura quod potest fieri per pauciora.” This famous sentence can be translated: “It is vain to do with more what can be done with fewer”). According to the VT concept the necessity of multiple coexisting transmitters derives from the fact that while axon and axon terminals assure a ”private” channel to convey signals from the source neuron to the target neurons, different chemical signals and high affinity recep- tors, selective for each of these signals, fulfill this aim for VT. Thus, the existence of the VT mode of communication in the brain may be one of the reasons for the multiple coexistences in the vast majority of neurons.

Luigi F. Agnati is Professor of Human Physiology a t the University of Modena, Modena, Italy. He obtained a Doctorate Degree in Biostatistics in 1969 in Padova and became Associate Professor in Human Physiology at the University of Bologna in 1973. He obtained his Doctorate Degree in medicine in 2977. He became Professor at the Department of Human Physiology in Naples in 1980 and obtained his present position in 2981. He then became Director of Post-Graduate School in Nutrition at the University of Modena in 2984 and Honorary Doctor at the Karolinska lnstitutet in 1993.

Borje Bjelke received his M.D. from the Uppsala-Linkoping Medical School in 1984 and started work on volume transmission a few years later in the Department of Histology and Neurobiology at the Karolinska lnstitutet. His research on brain damage and recovery will result in the first Ph.D. based on volume transmission.

Kjell Fuxe presently holds a position as Professor of Histology at the Karolinska Znstitutet, Stockholm. He became a Doctor of Medicine in 2965, Associate Professor of Histology at Karolinska lnstitutet in 2968, and obtained his present position in 2979. He became a member of the Swedish Royal Academy of Sciences in 1980 and a member of the Bologna Academy in Italy in 1984. In 1989 he became a member of the Academia Europaea and in 1992 Honorary Doctor at the Claude Bernard University in Lyon.

NEUROPSYCHOPHARMACOLOGY 35

Figure 1. Schematic representation of the different types of intercellular communications in Volume Transmis- sion in the Central Nervous System. Abbreviation: ac = autocrine; nel = neuroendocrine-like; nba = nerve bundle-assisted; pc = paracrine; CSF = cerebrospinal fluid.

Diffusion of chemicals in the ECF may follow facilitated pathways. The necessity of such pathways derives from the fact that the extracellular space is filled not only with ECF, but also with extracellular matrix, which represents an obstacle for the free diffu- sion.21 Recently, it has been discovered that the space around nerve bundles seems to be one of these facilitated pathways (Bjelke ef al. , in preparation). Furthermore, not only diffusion, but also convective transport of substances can occur in the CNS, the vector being the cerebrospinal fluid (CSF) and the ECF.

Thus, different types (see Fig. 1) of VT can be surmised, each of these are particularly suited to some tasks.4

11. CHEMICAL MESSAGES FOR VOLUME TRANSMISSION

A. Neuropeptides

The major neurotransmitters for VT are the neuropeptides which are usually stored in large granular vescicles far away from the synaptic cleft.22~23 These large vescicles can undergo parasynaptic exocytotic release and thus their content may represent a major source of VT signals. Of substantial interest is that several peptides are often costored in these large vescicles24 so that multiple signals can be coreleased for VT.4

The coexistence of classical transmitters and neuropeptides in the same axon terminal should also be mentioned, beeing present in different vesicles. In this case, the small synaptic vesicles contain transmitters such as glutamate or GABA and are predomi- nantly located close to the synaptic cleft,= and involved in synaptic transmission. Thus, one and the same axon terminal may be the source of signals for both VT and WT.

The neuropeptides are very well suited for VT in that they are not taken up again via reuptake mechanisms into the neuron but rather split into fragments by proteases and neuropeptidases.25.26 It should be noted that these fragments may be inactive or active. In the latter case they can interact with their unique subtypes of peptide receptors and trigger a set of interrelated responses, i.e., a “syndromic response.” Thus, the syndromic response is not simply the elicitation of positive or negative feedback actions with respect to the parent peptide response, but are instead responses which constitute the follow-up of the parent peptide response in order to achieve the proper elicitation of a f~nct ion.2~-3~

36 AGNATI ET AL.

It should also be emphasized that there exists not only an extracellular processing of neuropeptides, but also an intracellular processing.33 In fact, within the large granular vescicles the neuropeptides may be produced from protein precursors by the action of peptidases: although the precursor peptides are synthetized within the cell body area, the final processing takes place within the large granular vescicles of the preterminal axons and/or of the nerve terminals themselves. There exists evidence that the forma- tion of the active peptides in the terminal regions becomes accelerated by increases in the firing rates of the neurons.3 It should, therefore, be considered that an important trigger for VT may be represented by the increase in nervous impulse flow, which may lead to an increase of the preferential exocytosis of large granular vescicles. This process may be described as a switch of neuronal signaling from the WT towards the VT mode.

Evidence for long-distance diffusion of neuropeptides have been obtained by the use of prolactin secreting transplants in n e o s t r i a t ~ m , ~ ~ and also by the use of antibody microprobes.36 The latter experiments have demonstrated that, following noxious cuta- neous stimulation, immunoreactive neurokinin A is present far away from the neuro- kinin A immunoreactive nerve terminals in the dorsal horns of the spinal cord.

B. Classical Transmitters

All the classical transmitters, such as glutamate, GABA, acetylcholine, and mono- amines also operate via high affinity, slow receptors, coupled to G-proteins. It seems, therefore, likely that they may represent chemical signals not only for WT, but also for VT. However, since the vast majority of classical transmitter receptor seem to be fast ion channel linked receptors of the low affinity subtype, it is probable that classical transmit- ters are mainly involved in the WT.

When classical transmitters mediate VT they may reach the G-protein coupled recep- tors after being released either by the action potential at the level of boutons en passage, or via the reversal of the reuptake carriers for the respective transmitters during depo- larisation.37

The reuptake mechanism may also play an important role in controlling the diffusion of classical transmitters in the ECF. Thus, a local control of the reuptake mechanism in the different innervation areas can effectively control the extent of the VT. In the case of the monoamine and cholinergic neurons, with highly divergent projections to several distant brain regions, local control of the reuptake mechanism may effectively lead to a differential regulation of VT versus WT in discrete parts of the CNS.

Experimental evidence for dopamine (DA) involvement in the VT mode of interneuro- nal communication has been obtained using adenohypophyseal transplants in neo- striaturn35 as well as in a model of hemiparkinson’s disease.38

Adenohypophyseal transplants survive within the neostriatum and release prolactin into the extracellular space. By confocal laser microscopy indications were obtained that prolactin diffuses within ECF pathways. This process could be modulated by treatment with a DA agonist (bromocriptin) and a DA antagonist (haloperidol). Furthermore, fol- lowing disappearance of striatal DA terminals surrounding the transplant by means of an intranigral 6-OH-DA injection, the diffusion volume of prolactin was substantially increased. These results suggest that there exists a DA control of the transplanted prolac- tin secreting cells, and that the DA extracellular concentration is capable of reducing the prolactin secretion in an innervated neostriatum.

Recently evidence has been obtained for a very long-distance diffusion of DA via CSF in a model of hemiparkinsonism induced by unilateral intranigral injection of 6-OH-

NEUROPSYCHOPHARMACOLOGY 37

DA.38 Thus, in spite of an almost complete DA denervation of the neostriatum the catecholamine releasing drug, d-amphetamine, still induced c-fos-like immunoreactivity in a large number of striatal nerve cells on the denervated side and reduced the in- creased firing rate of the striatal neurons on the denervated side, as evaluated by electro- physiological recordings. Thus, d-amphetamine mimicked the DA action in the dener- vated striatum. The d-amphetamine action was abolished by reserpine pretreatment in combination with tyrosine hydroxylase inhibition. These results may be explained on the basis of a d-amphetamine releasing action on the unlesioned side leading to the penetration of DA into the CSF followed by its diffusion into the denervated neo- striatum, in a mediolateral direction, from the lateral ventricle.

111. ELECTRICAL (IONIC) MESSAGES FOR VOLUME TRANSMISSION

Already Golgi put forward the hypothesis that ionic current could take place between nerve cells without any direct contact between them, due to the interposition of a I1 class conductor, and that these currents could represent intercellular communications (Golgi, 1903). More recently, this field has been exploited by several research workers, in partic- ular by Nicholson.39

Thus, VT signals may also be ions such as K+, Ca++, H+ and these ions could represent not only interneuronal signals, but also neuron-glial and glial-glial intercellu- lar signals.40 It should also be observed that there exists evidence in favor of mutual interactions among these ion signals (as well as with other ones such as C1-, Na+, HC0,-). Thus, it has been suggested that glial cells increase their intracellular K+ con- centrations in response to the release of this ion by neighboring active neurons. This increase may control glycogen metabolism in glial cells, which in turn may influence neuronal metabolism and function.41.42 There are links between Ca++ and K+, as well as between K+ and H+. The extracellular increase in K+ concentration induces depolariza- tion of glial cells, resulting in the opening of voltage dependent Ca++ channels, with consequent changes in Ca+ + dependent K+ conductances. Furthermore, intracellular calcium ions deeply affect the glial cell intracellular biochemical machinery. The relation- ships between K+ and H+ are established, even if the underlying mechanism is still unknown. Thus, it has been observed that the intracellular pH of astrocytes (resting value around 7.1) turns more alkaline if exposed to elevated external K+.43

As far as the regulation of H+ concentration in the central nervous system is con- cerned, one should distinguish not only neurons from glial cells (involving in fact differ- ent glial cell populations), but also the brain regions and even the various parts of the neuron under study.4 Furthermore, one should have evaluations of the extracellular versus the intracellular pH, and finally the bidirectional link between pH and neuronal activity should be kept in mind. From a general standpoint it has been shown that neuronal activity causes an early alkaline shift followed by a more marked and long- lasting acid shift in the interstitial space.4 It seems that these pH changes in the intersti- tial fluid are due to the rapid acid efflux from glial cells, which is opposed by the alkalizing effect due to neuronal activity. Accordingly only small pH shifts can be ob- served. On the contrary, the late phase is due to a synergistic glial and neuronal acidify- ing action and, therefore, a marked and prolonged pH shift towards acidity can be detected. These local pH changes may be large enough to affect ion channels and enzyme activities. Thus, they could play a role in modulating neuronal function.4 In particular, NMDA-evoked currents are significantly enhanced by extracellular alkaline shifts of a few tenths of a pH unit.45

38 AGNATI ET AL.

Figure 2. Fluorescence microphotograph of a coronal section through the globus pallidus in a 17-day-old male rat using a two-color immunofluorescence procedure. The carbonic acid anhydrase I1 immunoreactivityhl (dilution 1:W) is visualized by FITC fluorescence and the myelin-associated glycoprotein (MAG) immunoreactivity62 (dilution 1:500) by Texas Red fluorescence. Arrows indicate oligodendrocytes containing both markers. Bar indicates 5 0 ~ m .

Since the interstitial fluid of the brain contains negligible concentrations of proteins and organic acids,46 it has to rely upon the COJHCO, buffer. The efficacy of the C02/HC0, buffer is dependent on the presence of carbonic anhydrase, which is capable of increasing C 0 2 hydration speed by around 400 times. Thus, this enzyme is of funda- mental importance to allow the C02/HC03 buffer to operate at a rate compatible with the homeostatic needs of the CNS.

NEUROPSYCHOPHARMACOLOGY 39

Several years ago histochemical evidence for the existence of carbonic anhydrase (CA) in the nervous system was provided.47 Recently, an immunocytochemical mapping of the distribution of CA I1 in the CNS of the rat has been carried out demonstrating its predominant presence in interfascicular and perineuronal oligodendroglia (see Fig. 2).

In the present article, we put forward the hypothesis that CO, also has other effects on neuronal function beside being part of the CO,/HCO, buffer. As a matter of fact we will present indirect evidence supporting the view that this gas may also work as a multifacet VT signal.

IV. CARBON DIOXIDE AS A REGULATOR OF WT AND VT As outlined above, the relationships between pH and CO, may be highly relevant in

determining neuronal function.48 However, another aspect of CO, presence in the inter- stitial fluid has to be considered, namely the carbamate forrnati0n.~~,5~ Some L-ami- noacids can be transformed in the CNS into an a-aminocarbamate endowed with NMDA-like excitatory actions.51 We would like to put forward the hypothesis that this mechanism before being excitotoxic can subserve a physiological excitatory role. Actu- ally, according to the scheme of Figure 3, carbonic anhydrase has a central role in:

protecting the neuron from excessive aminocarbamate formation, since it can favor

allowing a prompt switching of the microenvironment composition towards either the disappearance of CO, and formation of H+;

CO, or H+.

The latter aspect may be basic to a CO, mediated negative feedback control of neuro- nal activity. Thus, it may be surmised that CO,, in view of its high diffusibility (CO, has a diffusibility, through organic fluids and membranes, around 20 times higher than that of O,52) tends to uniformly permeate the entire CNS, to give a powerful contribution to the brain acid-base balance as well as to provide the CNS with a homeostatic control of

Amino Carbarnates

+ -

+ Carbonic

Anhydrase

Neuronal*K+- Glial Activity + + Depolarisation

+ Figure 3. Schematic representation of the possible CO, indirect action on NMDA receptors (R,,,,), via the CO,/HCO,- buffer and the formation of amino carbamates. See text for further details.

40 AGNATI ET AL.

neuronal activity superimposed on the local control. In fact, in a region with low neuronal activity there are low interstitial K+ levels and low glial activity which, in turn, favors alkaline pH in the microenvironment associated with a high efficacy of the NMDA decoding system. The CO, derived carbamate will therefore be very efficient, but the alkaline pH favors the formation of HC03- by the CA. In a region with high neuronal activity the high interstitial K+ levels and glial activity will lead to an acidic microenviron- ment and thus to a reduced efficacy of the NMDA decoding system. The CO, derived carbamate will therefore have a low activity. However, the acidic pH favors the formation of CO, by the CA, which will also reach distant areas due to a high diffusibility. In conclusion, a negative feedback controlling neuronal activity is therefore operating and its efficacy is directly dependent on the presence of carbonic anhydrase and pH.

This regulatory role of CO, on neuronal function may explain the widespread distribu- tion of carbonic anhydrase in the CNS. In fact, our mapping (Fig. 2) has demonstrated carbonic anhydrase I1 immunoreactivity in all brain regions, mainly located in oligo- dendroglial cells (perineuronal and interfascicular).

V. NITRIC OXIDE AS A VT SIGNAL

Several unrelated research lines demonstrate a role of NO in the regulation of a variety of cell functions and intercellular communication processes. 53

Nitric oxide is synthesized from the semi-essential amino acid L-arginine. Two types of NO synthases have been purified. Both enzymes are flavoproteins: the one present in the brain is a constitutive Ca2+-calmodulin-dependent enzyme, while the other one is an inducible Ca,+-independent enzyme.

Nitric oxide is released in response to increases of intracellular Ca2+ as seen following the activation of NMDA receptors.54 It may be surmised that a cross-talk between CO, and NO may take place. In fact, according to our scheme (see Fig. 3) CO, by regulating NMDA receptor sensitivity can modulate NO release.

Nitric oxide diffuses out of the producing cells and, via inter ulia the ECF, reaches the target cells where it activates the soluble guanylate cyclase. It seems that this gas may diffuse up to 100 pm in 5 s.55 However, it has a fast inactivation with a 50% decay of its activity in 4 s. Thus, its action is particularly suited for a paracrine fast type of VT (see Fig. 1).

VI. THE REPLACEMENT THERAPY IN THE FRAME OF THE WT AND VT CONCEPT

In the present review article different chemical signals have been mentioned as VT signals, in particular, classical transmitters, neuropeptides, NO, and ions.

The discovery of VT signals and of their interactions may open up new vistas on physiological and pathological aspects of brain function. Already now, however, on the basis of the available knowledge it is possible to explain some features of complex brain functions such as sleep/wakefulness, mood, pain, and central autonomic control, as well as to work with a new conceptual frame which allows us to discuss in a completely different way the action of neuropsychoactive drugs, especially of those used to treat neurodegenerative diseases.

Let us examine this last issue. Neuropathological studies have demonstrated that a neurodegenerative process sometimes hits preferentially, even if not exclusively, neu- rons which synthetize a certain transmitter. Then, on the basis of the neurotrophic hypothesis and the just mentioned neuropathological evidence, two types of therapy have been suggested:

NEUROPSY CHOPHARMACOLOGY 41

The “replacement therapy,” which has as a goal the compensation for the deficit of the transmitter. In order to obtain such a goal, administration of drugs capable of mimicking the transmitter or to enhance its synthesishelease or to block its catabol- ism have been developed. The “trophic support therapy,” which has as a goal the development of drugs capable to increase the trophic support selective for the dying neurons.

In the present article the former therapeutical approach will be discussed in the frame of the WT and VT modes of intercellular communication that exist in the CNS.

Chemical interneuronal communication rests on the release of a first messenger from the ”source neuron” and its action on the ”target neuron,” i.e., on the cell which possesses the appropriate molecular recognition-decoding system. However, the case must be distinguished in which the first messenger diffuses in the synaptic cleft (WT) from the case in which the diffusion takes place in the extracellular fluid, outside the synaptic cleft, sometimes also involving the cerebrospinal fluid to rapidly reach far located target neurons (VT).

Let us again consider some basic features of the two modes of interneuronal commu- nication.

WT: the first messenger is released by the source neuron at synaptic level according to an electro-temporal code. This coding of the transmitter release is fundamental for the information transfer from the source to the target neuron. WT is phasic.

VT: the first messenger is extrasynaptically released and diffuses to reach the target neuron. In this case, the information carried by the first messenger is substantially freed from the electro-temporal code that has caused the transmitter release. VT is tonic.

It is, therefore, possible to develop two types of therapeutical approaches-one which potentiates WT, and another one which potentiates VT.

A. WT Potentiation Therapy

Since its goal is the enhancement of synaptic transmission, the drug intervention must respect the electro-chemical code. Hence, it is not possible to use postsynaptic agonists. On the contrary, suitable drugs may be transmitter precursors and blockers of transmit- ter inactivation processes. In fact, transmitter precursors may increase the amount of transmitter released per impulse; blockers of inactivation processes tend to enhance the transmitter synaptic levels in the interimpulse period. Thus, the impulse released quan- ta of transmitter add to the amount already present in the synaptic cleft allowing the activation of the low affinity postsynaptic receptors. This type of treatment will obvi- ously also increase VT.

B. VT Potentiation Therapy

Its goal is to enhance the extracellular fluid levels of the first messenger, with little demand for an electro-temporal code. Thus, the source neuron can be ignored and also postsynaptic agonists and blockers of the inactivation processes can be considered as suitable drugs.

These considerations may help to differentiate between the neuropathological sub- strate of the various neurodegenerative diseases. In fact, it may be surmised that if a postsynaptically acting drug does not work, then one or several of the following phe- nomena exists:

a single first messenger is deficient, but WT is the mode of transmission;

42 AGNATI ET AL.

a single first messenger is deficient, VT is the mode of transmission, but the target

several first messengers are deficient; neurons are degenerated;

If a presynaptically acting drug does not work, then one or several of the following phenomena exists:

a single first messenger is deficient, but VT is the mode of transmission and the drug is not capable of enhancing the extracellular levels of transmitter in such a way that it can reach and activate the target neurons; a single first messenger is deficient, WT is the mode of transmission, but the target neurons are degenerated; several first messengers are deficient.

On the contrary, it may be surmised that if a postsynaptic drug is active, then very likely:

VT is the mode through which the affected communication operates; furthermore, mainly one single first messenger is deficient and, therefore, most of the symp- tomatology is caused by this single deficit.

If a presynaptic drug is active, then WT may be the mode through which the affected communication operates together with VT; furthermore, mainly only one single first messenger is deficient and, therefore, most of the symptomatology is caused by this single deficit.

C. Pharmacological Interventions of WT and VT

order to better understand failures in drug actions: Some other features of the interneuronal communication should be considered in

(a) Interneuronal communication (above all synaptic transmission) is controlled by very effective negative feedbacks, which tend to maintain the constancy of the features of the communication channels in spite of possible disturbances. This may produce tolerance and drug dependency.

(b) Several transmission lines are present at the synaptic level (this holds true also for VT). It seems that a pharmacological intervention on one of them can cause compensatory changes in other ones. This phenomenon may reduce treatment efficacy.

(c) Several networks cooperate to execute a certain complex function. It is possible that the same first messenger is employed in some of these networks. Thus, it may happen that if this first messenger simultaneously activates all its target neurons the appropriate integration among networks cannot take place.

All these problems may be present for both WT and VT, but they may be less relevant for VT. In fact:

(a) VT provides the target neurons with a tonic activation, so down-regulation problems should not be of a great importance. Furthermore, VT is not orga- nized in a fast operating feedback loop as the WT circuits, which usually show neuronal negative feedback loops.

NEUROPSYCHOPHARMACOLOGY 43

(b) VT target neurons may integrate the different VT signals impinging on them in several separate membrane domains, while WT is integrating them only at the postsynaptic membrane level.

(c) Due to the free diffusion of VT signals to reach distant target neurons, it is likely that a certain drug, mimicking the VT first messenger, triggers consistent ef- fects in the various networks which cooperate in performing a certain function.

VII. CONCLUDING REMARKS

The experimental evidence for the existence of two intercellular communication modes in the CNS is rapidly growing.3,35,38,56-58

This new conceptual view has both a theoretical and practical impact. As a matter of fact, several aspects of the CNS morphofunctional organization becomes more easily understandable on the basis of this dual mode of intercellular communication4 (see also above, e.g., the problem of multiple transmitters). It is also tempting to suggest that the slow global regulation by VT of entire neuronal networks may be relevant for some CNS functions, such as sleep/wakefulness, circadian rhythms, mood, nociceptive- antinociceptive mechanisms, hunger, sexual behavior, and acclimatation processes. In many instances, these functions are characterized by a slow onset and have a time-scale ranging from minutes to days, and they are accompanied by a reset of the entire network activity. Thus, as Golgi pointed out 90 years ago some sort of widespread signaling process may be the basis for their operation.

The practical impact of this new conceptual view has also been elaborated upon in the present review article by showing that the WT and VT concept may be the appropriate framework to discuss not only the complex brain functions mentioned above, but also the action of neuropsychoactive drugs.

As a last point, one should consider that, very likely, other tissues, besides the ner- vous tissue, may similarly use these two types of intercellular communication, in fact, “Natura non facit saltus.” This sentence, in the present case, means that also in other tissues is it possible to distinguish an intercellular communication based on structural contacts (e.g., gap-junctions59 which work as WT contacts also in the brain) from another one based on diffusion of chemical signals in the ECF. This phenomenon has been, for example, recently suggested in the bone tissue, for the regulation of osteoblast activity.60

VIII. SUMMARY

A volume transmission mode of communication in brain was implicit already in the early work of Golgi, who postulated the existence of electrical signals in the extracellular fluid (ECF) based on Volta’s ”wet conductor” made by solutions. The term volume trans- mission is taken from the term volume conduction describing the flow of ionic currents in the ECF as a basis for the electrocorticogram. The slow VT mode includes also chemical signals and is opposed to the fast synaptic (wiring) transmission. Every neuron may function in a dual mode, the synaptic and the volume transmission mode, when consid- ering the autocrine and synaptic classes of communication. The paracrine- and neuroendocrine-like classes only involve the VT mode in the latter case including the CSF as a route. The chemical signals for VT are the neuropeptides, but also the classical transmitters, the monoamines, acetylcholine, GABA, and glutamate can participate, when they operate via slow, high affinity G protein coupled receptors. Ions such as K+, Ca++, and H+ also function as VT signals. The hypothesis is also introduced that C02

44 AGNATI ET AL.

can act as a multifacit long-distance VT and WT regulator besides being part of the CO,/HCO, buffer. C02 via regulating NMDA receptor sensitivity can also regulate N O formation, which represents a paracrine and fast VT signal. The therapy of CNS disor- ders is also discussed in the frame of the wiring and VT concept. Two therapeutical approaches can therefore be developed, one based on increasing WT and one based on increasing VT. In contrast to the WT therapy, which must preserve the electrotemporal code, the VT therapy can operate also with postsynaptic agonists. Therefore, a therapeu- tic effect with such a drug indicates that the deficiency in the communication process operates via VT. In view of the lack of very effective negative feedbacks in VT vs. WT, VT therapy may produce less tolerance and drug dependency.

REFERENCES

1. L. F. Agnati, K. Fuxe, and M. Zoli, Crescita, 14, 34-38 (1985). 2. L. F. Agnati, K. Fuxe, M. Zoli, I. Zini, G. Toffano, F. Ferraguti, Acta Physiol. Scand., 128, 201-207 (1986). 3. K. Fuxe and L. F. Agnati, in Advances in Neuroscience, Vol. 1 , Volume Transmission in the Brain. Novel

Mechanismsfor Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 1-9. 4. L. F. Agnati, B. Bjelke, and K. Fuxe A m . Sci., 80, 362-373 (1992). 5. R. Guillemin, Science, N.Y. 202, 390-402 (1978). 6. C . Nicholson, J. M. Phillips, and A. R. Gardner-Medwin, Brain Res., 169, 580-584 (1979). 7. E. S. Vizi, Non-Synaptic Interactions Between Neurons: Modulation of Neurochemical Transmissions.

8. F. 0. Schmitt, Neuroscience, 13, 991-1001 (1984). 9. A. Volta, Scritti scelti (a cura d i Gliozzi M.) , UTET, Torino, 1967.

Wiley, New York, 1984.

10. Webster's New Collegiate Dictionary, G. & C. Merriam Co., Springfield, MA 1961. 11. J. H. Martin, in Principles of Neural Science, E. R. Kandel and J. H. Schwartz, Eds., Elsevier, Amsterdam,

12. C. Golgi, in Opera Omnia, Ulrico Hoepli, Milano, 1903, pp. 579-605. 13. S. R. Cajal, in Nobel Lectures: Physiology or Medicine, 2901-1921, Elsevier, Amsterdam, 1967, pp. 220-253. 14. G. M. Shepherd, Foundations of the Neuron Doctrine, Oxford University Press, New York, 1991. 15. A. Kolliker, Handbuch der Gewebelehre des Menschen, 5th Ed., Engelmann, Leipzig, 1867. 16. J. von Gerlach, Zentralbl. med. Wiss., 10, 273-275 (1872). 17. W. Waldeyer, Deutsche Med. Woch., 1352-1356 (1891). 18. L. Luciani, Fisiologia dell'Uomo, Societa Editrice Libraria, Milano, 1912. 19. Changeux, in Advances in Neuroscience, Vol. 1 , Volume Transmission in the Bruin. Novel Mechanisms for Neural

20. F. Jacob, La logique du vivant. Une historie de l'hiriditi, Editions Gallimard, Paris, 1970. 21. W. D. Comper, R. P. Williams, and 0. Zamparo, Conn. Tissue Res., 25, 89-102 (1990). 22. Klein A. Thureson, and R. L. Klein, Znt. Rev. Cytol., 121, 67-126 (1990). 23. M. A. Matteoli, T. Reetz, and P. DeCamilli, in Advances in Neuroscience, Vol. 1 , Volume Transmission in the

Brain. Novel Mechanisms for Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York,

24. T. Hokfelt, J. F. Rehfeld, L. Skirboll, 8 . Ivemark, M. Goldstein, and K. Markey, Nature, 285, 476-478 (1980). 25. J-C. Schwartz, M-L. Bouthenet, 8 . Giros, C. Gros, C. Llorens-Cortes, and H. Pollard, in Advances in

Neuroscience, Vol. 1 , Volume Transmission in the Brain. Novel Mechanisms for Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 381-394.

26. L. Terenius and F. Nyberg, in Advances in Neuroscience, Vol. 1 , Volume Transmission in the Brain. Novel Mechanismsfor Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991,415-424.

27. L. F. Agnati, K. Anderson, K. Fuxe, V. Locatelli, V. Mutt, T. Hokfelt, Neurosci. Lett. S u p p l . , 5, S204 (1980). 28. L. F. Agnati, K. Fuxe, and C . Blake, et al., INSERM, 110, 87-102 (1983). 29. D. DeWied, and J. Jolles, Physiol. Rev., 62, 976-1059 (1982). 30. K. Fuxe, K. Anderson, V. Locatelli, L. F. Agnati, T. Hokfelt, L. Skirboll, and V. Mutt, Eur. 1. Pharmacol., 67,

329-331 (1980). 31. K. Fuxe, K. Anderson, V. Locatelli, V. Mutt, J. Lundberg, T. Hokfelt, L. F. Agnati, P. Eneroth, and

P. Bolme, in Neural Peptides and Neuronal Communication, E. Costa and M. Trabucchi, Eds., New York, Raven Press, 1980, pp. 37-50.

1985, pp. 636-647.

Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 569-585.

1991, pp. 181-193.

NEUROPSYCHOPHARMACOLOGY 45

32. K. Fuxe, L. F. Agnati, and A. Harfstrand et al., Progress in Brain Research, Elsevier, Amsterdam, 1986, pp. 341-368.

33. J. del Rio-Garcia and D. G. Smyth, in Advances in Neuroscience, Val. 1 , Volume Transmission in the Brain. Novel Mechanisms for Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 407- 414.

34. J. M. Lundberg, in Advances in Neuroscience, Vol. I , Volume Transmission in the Brain. Novel Mechanisms for Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 425-432.

35. B. Bjelke, L. F. Agnati, and K. Fuxe, in Volume Transmission in the Brain: Novel Mechanisms for Neural Transmission. Advances in Neuroscience, Vol. 1 , K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 463-478.

36. A. W. Duggan, P. J. Hope, B. Jarrott, H.-G. Schaible, and S. M. Fleetwood-Walker, Neuroscience, 35, 195- 202 (1990).

37. D. K. Meyer, in Advances in Neuroscience, Vol. 1, Volume Transmission in the Brain. Novel Mechanisms for Neural Transmission, K. Fuxe and L. F. Agnati, Eds., Raven Press, New York, 1991, pp. 433-440.

38. B. Bjelke, 1. Stromberg, W. T. O'Connor, B. Anbjer, L. F. Agnati, and K. Fuxe, Brain Research, 1994, in press. 39. C. Nicholson, in The Neurosciences: Fourth Study Program, F. D. Schmitt and F. G. Worden, Eds., The MIT

40. W. Walz, Progress in Neurobiology, 33, 309-333 (1989). 41. M. Wozniak, 8. Rydzewski, S. P. Baker, and M. K. Raizada, Neurochem. Int., 22, 1-10 (1993). 42. R. Dringen, R. Gebhardt, and B. Hamprecht, Brain Res., 623, 208-214 (1993). 43. M. Cheder and R. P. Kraig, Am. 1. Physiol., 253, R666-R670 (1987). 44. M. Cheder, Progress in Neurobiology, 34, 401-427 (1990). 45. M. Morad, M. Ditcher, and C-M. Thang, Soc. Neurosci. Abstr., 14, 791 (1988). 46. V. Fencl, in Handbook of Physiology. The Respiratory System. Control of Breathing, Sect. 3, Vol. 11, Am. Physiol.

SOC., Bethesda, 1986, pp. 115-140. 47. E. J. Giacobini, Neurochem., 9, 169-177 (1962). 48. J. H. Coote, TIPS, 12, 450-455 (1991). 49. B. Max, TIPS, 12, 408-411 (1991). 50. E. A. Mroz, Science, 243, 1615 (1989). 51. P. B. Nunn, A. J. Davis, and P. O'Brien, Science, 251, 1619 (1991). 52. M. P. Hlastala, in Textbook of Physiology, H. D. Patton, et al., Eds., Vol. 11, W. B. Saunders Company, 1989,

53. S. Moncada, R. M. Palmer, E. A. Higgs, Biochem. Pharm., 38, 1709-1715 (1989). 54. T. Mccall and P. Vallance, TIPS, 13, 1-5 (1992). 55. J. A. Gally, P. R. Montague, G. N. Reeke, Jr., and G. M. Edelman, Proc. Natl. Acad. Sci. U S A , 87, 3547-3551

56. B. Bjelke, K. Fuxe, and L. F. Agnati, Neurosci. Lett., 93, 139-145 (1988). 57. B. Bjelke, K. Fuxe, L. F. Agnati, Neurosci. Lett., 107, 33-38 (1989). 58. B. Bjelke and K. Fuxe, Neuroreport, 5, 265-268 (1993). 59. P. 8. Detwiler and W. E. Crill, in Textbook of Physiology, H. D. Patton et al., Eds., W. 8. Saunders, Phila-

60. G. Marotti, S. Palazzini, and C. Palumbo, Calcified Tissue, 53, 440 (1993). 61. K. T. Sundquist, M. Leppilampi, K. Jarvelin, T. Kumpulainen, and H. K. Vaananen, Bone, 8, 33-38 (1987). 62. D. E. H. Afgar, J. L. Salzer, J. Roder, P. E. Braun, and J. C. Bell, 1. Neurochem., 55, 1418-1426 (1990).

Press, Cambridge, MA: 1979, pp. 457-476.

pp. 1012-1025.

(1990).

delphia, 1989, pp. 231-284.