History of the Synapse

Citation preview

History of the Synapse

Dedicated to Bernard Katz and to the memory of John Newport Langley

History of the Synapse Max R.Bennett University of Sydney, Australia

harwood academic publishers Australia • Canada • France • Germany • India • Japan • Luxemburg Malaysia • The Netherlands • Russia • Singapore • Switzerland

This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”

Copyright © 2001 OPA (Overseas Publishers Association) Amsterdam N.V. Published by license under the Harwood Academic Publishers imprint, part of the Gordon and Breach Publishing Group. All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, without permission in writing from the publisher. Printed in Singapore. Amsteldijk 166 1st Floor 1079 LH Amsterdam The Netherlands British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 0-203-30254-0 Master e-book ISBN

ISBN 0-203-34506-1 (Adobe eReader Format) ISBN 90-5823-233-6 (Print Edition) Cover: Hippocampal neuron stained for MAP2 together with synaptotagmin. Courtesy M. Matteoli and © Cell Press. Cover design by Lee McLachlan.

Contents

Introduction

ix

The early history of the synapse: from Plato to Sherrington

1

1.1

Plato and Aristotle: ‘vital pneuma’ is necessary to initiate organ function

1

1.2

Galen: pneuma is conducted and transmitted from nerve to muscle

3

1.3

Descartes: the replacement of pneuma by mechanical corpuscles

4

1.4

Borelli: a corpuscular description of conduction and transmission

6

1.5

Fontana: nerves are composed of many cylinders along each of which conduction occurs

8

1.6

Galvani: electricity is conducted and transmitted not corpuscles

9

1.7

Matteucci and du Bois-Reymond: transient electrical changes are conducted (the action potential)

12

1.8

Helmholtz: the action potential has a finite velocity

15

1.9

Kuhne and Auerbach: identifying the structure of nerve endings on muscle and neurons

16

1.10

Cajal: nerve endings are not continuous with the cells on which they impinge

18

1.11

Sherrington: the adoption of the word ‘synapse’

22

Emergence of the concept of transmitter release at peripheral and central synapses

26

2.1

Research on the synapse in the laboratories of Sherrington and Langley before the Great War

26

2.2

Sherrington’s concept of the inhibitory and excitatory states of central synapses

28

2.3

Lucas, Adrian, and the electrical concept of the inhibitory state of central synapses

29

2.4

Loewi, Dale and Eccles examine the inhibitory state at autonomie neuromuscular junctions

32

2.5

Eccles develops the electrical concept of the excitatory state at autonomie synapses

32

2.6

Katz, Kuffler and Eccles establish the motor endplate as the paradigm synapse for electrophysiology

35

2.7

Eccles elucidates the electrical signs of the inhibitory and excitatory states of central synapses

36

2.8

Katz’s concept of quantal transmitter release at the motor endplate and the vesicle hypothesis

39

2.9

Conclusion: the establishment of Sherrington’s concept of the synapse in the central nervous system and central synaptic transmission

41

The discovery of acetylcholine and the concept of receptors at synapses

43

3.1

Introduction

43

3.2

Claude Bernard and curarization: the notion of an intermediate zone between nerve and muscle

44

3.3

Paul Ehrlich and the idea of the ‘receptive side chains’ of cells

46

3.4

John Langley and T.R.Elliott: the emergence of the concept of chemical transmission between sympathetic 46 nerves and smooth muscle

3.5

The action of curare and John Langley’s development of the idea of transmitter receptors

51

3.6

The Langley-Ehrlich receptor theory

54

Chapter 1

Chapter 2

Chapter 3

vi

3.7

The discovery of acetylcholine and its physiological action at autonomic neuroeffector junctions

54

3.8

The physiological action of acetylcholine in autonomic ganglia

59

3.9

The identity of acetylcholine as the transmitter substance at somatic neuromuscular junctions

60

3.10

The discovery of the physiological action of single acetylcholine receptors

61

3.11

Conclusion

64

The discovery of adrenaline and the concept of autoreceptors at synapses

65

4.1

Introduction: the discovery of noradrenaline as a transmitter

65

4.2

Early observations leading to the idea of autoreceptors

65

4.3

Direct experimental evidence for autoreceptors

68

4.4

Identification of presynaptic adrenergic autoreceptors different from postsynaptic adrenergic receptors

71

4.5

Presynaptic adrenergic autoreceptors in the central nervous system

72

4.6

Evidence that endogenous autoreceptor mechanisms exist

72

4.7

The ionic basis of the action of alpha 2 adrenoceptors

73

4.8

Conclusion

76

The discovery of amino acid transmission at synapses in the central nervous system

77

5.1

Introduction

77

5.2

Identification of excitant and depressant amino acids

77

5.3

Glycine accepted as an inhibitory transmitter in the spinal cord

78

5.4

The emergence of GAB A as an inhibitory transmitter in the brain

81

5.5

L-Glutamate as a neurotransmitter: synaptic excitation, ion fluxes and neurotransmitter transporters

84

5.6

NMDA receptors: the first amino acid receptor identified at central excitatory synapses

86

5.7

Non-NMDA receptors at excitatory synapses

87

5.8

GAB A receptors

87

5.9

Conclusion

87

Monoaminergic synapses and schizophrenia: the discovery of neuroleptics

90

6.1

Introduction

90

6.2

Chlorpromazine

90

6.3

Haloperidol

91

6.4

The dopamine hypothesis for neuroleptics

91

6.5

Dopaminergic projections in the brain

92

6.6

Identification of the D1-like and D2-like dopamine receptors

92

6.7

Determination of different classes of dopamine receptors

95

6.8

Clozapine

95

6.9

Distribution of D1 and D2 receptors in the striatum of schizophrenics

98

Chapter 4

Chapter 5

Chapter 6

6.10

Mixed aminergic actions of the neuroleptics: serotonin and dopamine receptor blockade

102

6.11

Cellular and molecular mechanisms of action of dopamine receptors

102

6.12

The time course of action of neuroleptics on dopamine receptors and the emergence of antipsychotic effects 103

vii

6.13

Conclusion

104

The discovery of transmitters other than noradrenaline and acetylcholine at synapses in the peripheral nervous system

105

7.1

Introduction: J.N.Langley, H.H.Dale and non-adrenergic, non-cholinergic (NANC) transmission

105

7.2

Parasympathetic neuromuscular junctions in the gastrointestinal tract: mechanical studies

109

7.3

Parasympathetic neuromuscular junctions in the gastrointestinal tract: electrophysiological studies

116

7.4

NANC transmission: the new autonomic paradigm

121

7.5

Contemporary views on the identity of NANC inhibitory transmitters

127

7.6

Ionic mechanisms involved in generating the IJP

132

7.7

The secretion of NANC transmitters responsible for the IJP

133

7.8

NANC excitatory transmission in the gastrointestinal tract

134

7.9

Conclusion

135

Development of the concept of a calcium sensor in transmitter release at synapses

136

8.1

Introduction

136

8.2

Calcium is necessary for the release of transmitter

136

8.3

Electrophysiological evidence that calcium is necessary for the release of transmitter: the concept of a calcium sensor for secretion

137

8.4

The calcium action potential

140

8.5

Are calcium movements across the nerve terminal necessary for evoked secretion?

143

8.6

Direct evidence for calcium entry across the nerve terminal membrane during an impulse

147

8.7

Calcium channels in the nerve terminal

149

8.8

Identification of the calcium sensor molecule

158

8.9

Conclusion

162

The discovery of quantal secretion and the statistics of transmitter release at synapses

164

9.1

Introduction

164

9.2

Evoked quantal release as a binomial or Poisson variate

165

9.3

Kinetics of release of a quantum

176

9.4

Maximum likelihood estimation of parameters in statistical models of quantal release

186

9.5

Autocorrelation function used to detect quantal release

190

9.6

Model discrimination: statistical methods for discrimination between different statistical models of transmitter release

193

9.7

Appendix

195

The discovery of long-term potentiation of transmission at synapses

216

10.1

Introduction: the hippocampus, memory and long-term potentiation (LTP)

216

10.2

LTP at synapses in the brain

219

10.3

The induction of associative LTP in the brain

220

10.4

The maintenance of associative LTP

226

10.5

Biochemical pathways implicated in the maintenance phase of associative LTP

229

Chapter 7

Chapter 8

Chapter 9

Chapter 10

viii

10.6

Evidence that associative LTP is involved in memory

230

10.7

The induction of non-associative LTP

233

10.8

Summary and Conclusion

233

Emergence of the concept of synapse formation molecules

238

11.1

Introduction

238

11.2

Synapse formation in muscle

238

11.3

Synapse formation molecules in muscle and the elimination of polyneuronal innervation

243

11.4

Elimination of polyneuronal innervation during muscle development described by dual-constraint theory

247

11.5

Elimination of polyneuronal innervation during reinnervation of muscles described by dual-constraint theory

255

11.6

Elimination of polyneuronal innervation and establishment of topographical maps in muscle described by 262 dual-constraint theory

11.7

Identification of the synapse formation molecules

271

11.8

Synapse formation in autonomic ganglia

279

11.9

Elimination of polyneuronal innervation in autonomic ganglia

279

11.10

Elimination of polyneuronal innervation during reinnervation of ganglia

280

11.11

Identification of synapse formation molecules in autonomic ganglia

281

11.12

Conclusion

282

Epilogue

286

References

288

Illustration Acknowledgements

328

Acknowledgements

332

Index

334

Chapter 11

Introduction

This work is an attempt to provide a history of those discoveries concerning the identification and function of synapses which provide the foundations for research during this new century. It is written in the conviction that errors in the development and application of contemporary concepts to the understanding of synapses arise if there is failure to probe the origins of the scientific paradigm at present in use. The idea that blood vessels are the means by which muscles are activated lasted some five hundred years, from Aristotle in the fourth century BC to Galen in the second century AD, when it was shown that nerves are the means of communication to muscle. The concept that the ventricles of the brain, containing the psychic pneuma identified by Aristotle, are the sites at which sensations, thinking and memory are experienced lasted for a much longer period. The most posterior ventricle was taken as initiating the flow of psychic pneuma to the nerves and hence to muscles for their activation. This concept lasted for over fifteen hundred years, from shortly after Galen to Thomas Willis at the end of the seventeenth century. Is the historical development of these facts, involving both the discoveries and thoughts of men of genius, to be regarded as an oddity divorced from ‘the common-sense’ which we now bring to the solution of problems regarding the workings of synapses? The first chapter considers the wonderful story, evolving over two and a half thousand years, of how progress was made in the establishment of a conceptual structure that would allow the synapse to be identified at the beginning of the twentieth century. It was founded on the idea of conduction and transmission to muscle by Aristotle, the identification of nerves as providing the conducting medium by Galen and on Galvini’s discovery that it is electricity that is conducted, not Cartesian corpuscles. The concept of the synapse which was accepted for most of the twentieth century, was certainly not in place at the end of the nineteenth century through the work of Sherrington, as is popularly accepted. Rather, he highlighted the fact that a mechanism must be sought by which conduction could proceed from one excitable cell to another following the discovery of Cajal that neurones are separate cells. However, this difficulty had been a focus of physiological enquiry for centuries, particularly to Galen and to Descartes’ contemporary Borelli. In naming this region of apposition between excitable cells ‘the synapse’ Sherrington helped focus the physiological enquiry which was to be so brilliantly brought to fruition by his mentor at Cambridge, John Newport Langley. The genius of Langley is exemplified in the experiments he performed. These first drew attention to the similarities between the effects of sympathetic nerve stimulation on autonomic effectors and that of extracts of the adrenal glands, leading to the audacious speculation of his student Elliott that electrical conduction is transmitted chemically and that in the case of sympathetic nerves this chemical is adrenaline (Chapters 2 and 3). However, a reading of the papers of that period (the beginning of the twentieth century) points up that it was Langley’s discovery of transmitter receptors at the somatic neuromuscular junction that was the pivotal work in establishing the concept of the synapse. Langley championed the idea of the receptor in the face of fierce opposition, particularly from the founder of immunology, the chemist Paul Ehrlich. It was Langley’s work that gave us the modern concept of the synapse, namely an area of apposition between a nerve ending and another cell at which a chemical substance is released from the ending onto a ‘receptive substance’ found on the cell. All of these definitive concepts which gave ‘substance’ to the abstract term ‘synapse’ were established in a period of about eight years, principally from the one laboratory! The subsequent identification of acetylcholine as the chemical released from motor nerve terminals had to wait several more decades (Chapter 3). It was primarily the laboratory in London of Dale, a former student of Langley, together with that of Loewi in Germany that was then responsible for generalising the principle of chemical transmission at peripheral synapses. The study of central synaptic transmission did not reach a level of sophistication comparable to that at peripheral synapses for nearly half a century after the period when the idea of transmitter substances and their receptors was first conceived. The 1950s were a remarkable period for central synaptic transmission. First amino-acids were identified as likely central transmitters (Chapter 5) and then neuroleptic agents were synthesised and shown to act at central monoaminergic synapses in ways which were to have a profound impact on the alleviation of mental suffering, such as in schizophrenia (Chapter 6). It was also at this time that the mechanism of synaptic transmission was greatly illuminated due to introduction of the glass microelectrode for studying the same junction that gave us the concept of the receptor, namely the somatic neuromuscular junction. Katz and his colleagues showed that transmitter release occurs in units (Chapter 9) and that the probability of release

x

of these units is controlled by calcium ions. The search for the calcium sensor upon which these ions act has become a major research program of contemporary neuroscience (Chapter 8). The paradigm established by Langley, Elliott, Dale and Loewi, namely that adrenaline and acetylcholine are the transmitters at peripheral autonomic and somatic synapses, was shown to be incorrect in the early 1960s. At this time evidence was provided through application of glass microelectrode techniques to autonomic effectors, that at least two other transmitters must be active in synaptic transmission. The synapses using these transmitters were termed non-adrenergic noncholinergic (NANC) (Chapter 7). The subsequent discovery that these NANC transmitters include nitric oxide, neuropeptides and ATP has substantially contributed to our understanding of the process of chemical transmission. As the pace of research on synapses accelerated in the 1970s, three further discoveries were made that have now become principal foci of research. The first of these was the phenomenon of long-term potentiation of transmitter release whereby some nerve terminals greatly increase their efficacy of transmission following a high-frequency train of impulses, with this increased efficacy maintained for very long periods of time in excess of hours (Chapter 10). The second was the discovery that synapse formation molecules exist, of the kind originally envisaged by Langley at the beginning of the century, and that these are laid down on muscle cells by the motor-nerve terminal during early development (Chapter 11). This initiated a research program to identify such molecules, which is another area of intense activity amongst the contemporary synaptic neuroscience community. Finally, the 1970s saw the first analysis of transmitter receptors at the single receptor level by Sakmann and Neher (see Chapter 3). Suitably, they studied the receptors of the somatic neuromuscular junction to introduce their technique of patch-clamping, the same preparation that Langley had used seventy years earlier to give us the idea of transmitter receptors in the first place. This history provides a personal view of the process by which new concepts concerning the workings of the synapse have developed. The emphasis is naturally then on those aspects of synaptic transmission that have fascinated me over a forty-year period of research and brought me so much pleasure. I have written this book in the hope that others might share in the excitement of synaptic physiology and perhaps even help them in placing the development of their own concepts and research in an historical perspective. It will be clear to some that I have over emphasised those areas of research to which I contributed, such as the discovery of synapses that are non-adrenergic non-cholinergic (NANC) in the peripheral nervous system (Chapter 7), the discovery of the calcium action potential (Chapter 8), as well as the discovery that synapse formation molecules exist at the somatic neuromuscular junction (Chapter 11). On the other hand very important areas of research have been left out completely, such as the delivery of neurotrophic molecules to nerve terminals and the modification of proteins at synapses that determine the efficacy of transmission. I ask the readers indulgence in these distortions of perspective. If this work should seem worthwhile to some then these omissions will be rectified in the future. It will be a pleasure to receive from readers their own views of the events chronicled here which might then lead to adjustment of my perspective on this great subject. I have attempted an intellectual history of the synapse, charting the development of concepts and insights regarding synaptic structure and function that can be gleaned from reading primary sources, namely those research papers and occasionally other scholarly works in which the research was originally announced. I apologise to the many students of synaptic transmission whose work has not been quoted where they may see it as relevant. However, I have had to make difficult judgements as to the extent to which the literature should be referenced on a particular topic. This work does not provide much in the way of detail concerning the controversies surrounding the emergence of new ideas and claims of priority of discovery, issues that often involve the dramatic clash of personalities that can make the narrative of a history come ‘alive’. However, this detail has been provided in a few cases where it is necessary to understand the issue under consideration. Such cases arise in the disagreement between Langley and Ehrlich concerning the possibility of transmitter receptors at the beginning of the twentieth century (Chapter 3) as well as in the arguments during the middle of that century concerning the idea that there must be several transmitters in the peripheral nervous system other than just noradrenaline and acetylcholine (Chapter 7). It emerges again in the incisive arguments that the biophysicists such as Adrian and Lucas used in the 1920s to argue for electrical transmission in the central nervous system and that Eccles used at that time in favour of such transmission in the peripheral nervous system, all in opposition to the arguments of Sherrington on the one hand and Dale on the other (Chapter 2). Such disagreements emerge again concerning the concept that either a fluid or a set of corpuscles must pass from nerve into muscle at the site of transmission in order for a muscle to shorten, championed in their different ways by Aristotle, Galen, Descartes and Borelli. This idea was negated by the definitive experiments of Swammerdam in the eighteenth century and by the experiments of Galvani soon after showing that it is electricity that is conducted and transmitted at this junction, not fluid or corpuscles. Some will argue that there is an excessive number of illustrations in this history, which is not balanced by either an appropriate depth of analysis of the experiments that are highlighted nor of sufficient attention to many aspects of the history of the synapse that have been neglected. All I can say in reply is that I have tried to keep as close to the raw data of the experiments as possible. This has seemed best accomplished through figures explaining either the technical approach or the experimental finding. However no excuse is offered for failing to present a more comprehensive history, which I hope to produce at some time in the future, especially after receiving the criticisms of my fellow neuroscientists.

xi

The chapters in this book were previously published as essays on the history of various aspects of synaptic function in response to requests from my colleagues. This soon turned into a fascinating intellectual journey, which provided me with the foundations for a subject that I have found so personally satisfying. I hope the reader is as stimulated as I have been in surveying our efforts to understand the workings of the synapse.

1 The Early History of the Synapse: from Plato to Sherrington

1.1 Plato and Aristotle: `vital pneuma' is necessary to initiate organ function In the beginning there were four elements, fire, air, water and earth. Various proportions of these composed the blood, muscle, bone, tendons and nerves from which the body or soma was constructed. The ingredients of the blood determined intelligence so that the heart was the basis of the intellect and of mental life. This pre-Socratic idea of the fifth century BC, principally due to Empedocles, was developed further by Democritus who considered that each of the four elements was composed of a different kind of particle. He argued that the psyche or soul was composed of the lightest, fastest moving and most nearly spherical particles which are to be found throughout the body, especially concentrated in the brain. Particles of a lesser quality were to be found in the heart giving it a role in emotion whilst the most coarse particles were located in the liver, responsible for functions such as lust. Plato, in the fourth century BC, assigned specific geometrical shapes to each of the four kinds of particles. In addition he confronted the problem which these pre-Socratic ideas presented of how to relate the psyche to the body. A living thing for Plato was matter properly arranged to permit effectual intervention of the soul. Following Democritus, he claimed that there were three different kinds of psyche, namely that concerned with rational thought and behaviour which was associated with the head, that involved with passion and the emotions associated with the breast and the heart therein, and that concerned with desires which was associated with the liver. Only the rational psyche was immortal (Finger, 1994; Gross, 1998; White, 1996). The problem of what form the association between soul and body took was formulated in terms of the geometrical principles that played such a large role in Plato’s cosmology (Conford, 1956). As he considered the fundamental units of the elements themselves to be geometrical figures, such as the triangle, so the body composed of these elements must ultimately be thought of in mathematical terms. It was the appropriate organisation of these geometrical figures from which the body was ultimately composed that allowed the bonding of the soul to the body. It was only through such bonding that the soulbody complex could manifest life-as-action. Plato placed this bonding in what is now called the nervous system. In his ‘Timaeus’ he describes the soul as bonded to a substance that is found in its purest form in the cranial and spinal cavities where it appears as ‘marrow’, or what is now called brain and spinal cord. The marrow is the primary life stuff in which ‘were fastened the bonds of life by which the soul is bound to the body’. This marrow is not composed of the four elements, or rather of the elementary geometrical figures which make up the elements, but of specially well-formed examples of those triangles which are the common components of these elements (Hall, 1975). Thus Plato, following Pythagoras, developed the notion of the body as a temporary receptacle for a separate soul, which was associated with rationality, located in the marrow or nervous system, and which could pass from one body into another at death. The other kinds of soul, those associated with the emotions and desires and therefore with the heart and liver, were not capable of this transmigration. Physiological function of an organ was considered in terms of the associated psyche or soul giving life to the propensity of the organ to carry out its function. In this way organs came to be seen as possessing faculties or propensities to carry out a physiological act that was energised by the psyche (Plato, 1892). Aristotle, in the fourth century BC, developed radically different concepts concerning the functions of the body and soul that were to have a profound influence on physiological thought concerning the activation of organs and of muscle for two thousand years [see Aristotle, 1984, Fig. 1.1]. The Aristotlean concept of the soul will be considered in some detail here as Aristotle’s ideas are so different from those of Descartes, which embody a dualism like that of Plato which still dominates thinking on these issues to this day (Cottingham, 1995; Everson, 1996). For Aristotle the soul or psuche or psyche was the form of the thing under consideration. This form constituted the reason for a thing being as it is and could be considered as providing explanations for what it is made of (the material cause), what actually makes it (the efficient cause), what shape is used to identify it (the formal cause) and the ultimate reason for its existence (the final cause). Thus in the case of a muscle, the material cause is the fibres that it is made of, the efficient cause is the grouping of the fibres in relation to each other, the formal cause is that this grouping is done to a particular design in order

2

EARLY HISTORY OF THE SYNAPSE

Fig. 1.1. Theories of the nervous system before Descartes: conduction and transmission of psychic pneuma. (A) This anonymous fifteenth-century drawing illustrates pre-Cartesian brain theories, which followed the views of Aristotle. The senses of touch and taste are shown connected to the heart, while the boxes on the head are the ‘cerebral cells’ where mental faculties such as memory and fantasy are located. Anonymous, 15th Century, Bayerische Staatsbibliothek. Munich (from Posner, 1994). (B) A 15th century illustration due to Gregor Reisch, showing the four routes of communication connecting the organs of taste, smell, seeing and hearing to the anterior cerebral ventricle. This is divided into sensus communis, fantasia and imaginativa. The vermis connects it to the second ventricle, with the two faculties called cogitativa and estimativa, whereas the faculty spoken of as memorativa is in the third. The curving lines around the ventricles can be interpreted as representations of the brain’s convolutions (from Biblioteca Nazionale Centrale, Florence; see also Corsi, 1991; The Enchanted Loom). (C) In the 4th century A.D. and for several hundred years to follow, the faculties of the mind were thought to be housed in four ventricles of the brain as in (B). A 15th Century illustration, with indistinct lettering, designed to illustrate the 1494 edition of Aristotle’s de Anima. Four regions of the brain are labelled: sensus communis, virtus cogitativa, virtus imaginativa, and memoria (Courtesy of the Incunabula Collection at the National Library of Medicine, Bethesda; from Brazier, 1984).) (D) Leonardo da Vinci’s localization of the sensus communis. In the lower left of the figure his own words state: ‘Where the line a-m is intersected by the line c-b there the meeting place of all the senses (sense commune) is made’. (Courtesy of the Royal Collection © 2000, Her Majesty Queen Elizabeth II; this work later attributed to Cesare da Sesto).

to produce a muscle of a particular shape and the final cause is the fulfilment of the purpose of the muscle, which is to contract and produce, say, the movement of a limb. In this way an organ’s form or psyche is not material but is inherent in the organ and cannot exist separate from the organ. If then the constituents which make up the form are specified, so is the soul or psyche. In this way Aristotle lays stress on the activities of living things and on the distinction between ‘living’ and ‘dead’

HISTORY OF THE SYNAPSE

3

rather than in the distinction emphasised by Plato between ‘mental’ and ‘physical’. ‘Mind’ does not figure largely in Aristotle’s work, perhaps because of his emphasis on the psyche, that is on the activities of living things such as organs. If the psyche disappears from an organ it then ceases to be such a thing except in name only. The loss of psyche, of the soul of a living thing, means it ceases to exist (Hall, 1975). The concept of the psyche of an organ was not abandoned until Descartes. Meantime, the effect of Aristotle’s ideas was to lead to the search for the form of an organ so that scholars sought to identify the psyche of an organ (Clarke, 1995). This had two effects: first, it lead enquiry away from the mechanical workings of the organ; second, it placed emphasis on the final cause component of the psyche, that is the potential of an organ to carry out its function, of how the potentiality of the psyche of an organ could be realised, a problem which will now be considered. For Aristotle the heart was the central organ of perception, rather than the brain as postulated by Plato. If an animal’s perception gives rise to action, that is to the contraction of muscle leading, for example, to locomotion, it will occur as follows (Everson, 1996): ‘if the region of the origin (i.e. the heart) is altered through perception and thus changes, the adjacent parts change with it and they too are extended or contracted, and in this way the movement of the animal necessarily follows’. According to Aristotle, perception occurs in the heart with its particular psyche (Fig. 1.1 A). Perception is not an activity that involves two different substances, as later suggested by Descartes, for the affections of the psyche are common to the psyche of the body. ‘It is apparent that all the affections of the psyche are with the body…in all these the body undergoes some affection.’ The central sense organ is therefore the heart, which is connected to the individual sense organs (Fig. 1.1 A & 1.1C). When these are affected by their objects, the affections pass through the blood stream to the heart. Thus the movements around the heart bring about the movements of the limbs by acting through the blood stream. An organ ceases to be an organ if separated from the body as its psyche no longer exists. It is only as a consequence of being part of the body that its psyche is intact, which includes the final cause or ultimate reason for the organs existence. Given that the heart is the centre of perception and of the appetites and responds to these by initiating animal motion it is responsible for the activation of the muscular organs. The key question which now arises is how does the psyche of the heart conduct and transmit through the blood stream information to the psyche of the muscular organs which is responsible for their final cause, that is contraction. The emphasis of Aristotle on the natural world rather than the Platonic mathematical world led him to consider the most likely method for conduction and transmission from the heart to a muscle through the blood stream in terms of the elements. To the organs whose substance was made up of the four elements (fire, air, water and earth) he introduced a fifth element. This element was not restricted to this world only but also belonged to the stars and heavens, so that it permeated the entire universe: he named it the ‘ether’. The concept of the ether was to have a major impact on both the physical as well as the biological sciences. Aristotle considered that the ether element was taken into the body during breathing and conveyed from the lungs to the heart in which it was transformed to ‘vital pneuma’ or ‘vital heat’ (Peck, 1943). This vital pneuma was then distributed from the heart throughout the body by blood vessels, where it was able to mediate between the psyche of the heart and the psyche of the organs including muscles. It is then vital pneuma that is conducted from the heart along the blood vessels to be transmitted to the muscles and in so doing initiating their final cause, contraction (Hall, 1975). Aristotle had in one brilliant stroke introduced a means for mediating between the psyche of the heart and the psyche of muscles by introducing a fifth element that was not just of this world but seemed to possess a heavenly property associated with the stars. But the great contribution here is the introduction of the concept of a substance of a kind, be it of somewhat mysterious qualities, which had to be conducted to an organ to allow it to function, even though this function was taken to be simply the ability of the organ to release its propensities to action as dictated by the final cause of its psyche. 1.2 Galen: pneuma is conducted and transmitted from nerve to muscle Galen and his students, in the second and third centuries AD, greatly refined the concept of the conduction of pneuma to the organs of the body (see Galen, 1978; Galen, 1968; Galen, 1962). They retained Aristotle’s conceptual scheme with the four elements constituting the tissues and organs of the body, and a fifth composing the vital pneuma acting as a mediator for the psyche to give life to the organ and allowing it to release its propensities for action. Erasistratus argued that the pneuma of the inspired air became vital pneuma as it passed from the bronchioles of the lungs via the intrapulmonary veins to the pulmonary vein and into the heart. The heart on dilation sucked in the pneuma from the pulmonary vein and on contraction forced the vital pneuma to the rest of the body through the arteries. Blood is carried by veins not arteries. The vital pneuma which reaches the brain in this way is converted to ‘psychic pneuma’ there from which it travels outwards along nerves (Galen, 1821– 1833). The brain, rather than the heart, becomes once more the centre of perception in this scheme. Galen had already established that nerves arise from the brain and spinal cord, that conduction of psychic pneuma is necessary in these nerves for sensation and motor action for if they are cut or damaged there is no sensation or movement, and that there are two classes

4

EARLY HISTORY OF THE SYNAPSE

of nerves, one motor (if damaged no motor action) and the other sensory (if damaged no sensation). It was therefore estab lished that sensitive psyche possessed its own nerve supply as did the locomotor psyche. These observations and speculations of Galen and his students set the stage for the consideration of the mechanism of conduction of psychic pneuma along motor nerves and of the transmission of pneuma into muscle. Galen comments ‘All muscles require to receive a nerve from the brain or from the spinal cord and this nerve is small to behold but by no means slight in power’. Three possibilities for the conduction of the effects of psychic pneuma were entertained: one, that the psychic pneuma flows along the nerves like a liquid along a conduit; second, that the psychic pneuma in the brain pushes the pneuma resident in the nerve so that some is released at the ends of the nerves; finally, that there is only a flow of ‘potency’ through the psychic pneuma that is resident in the nerve. This last is akin to nerve conduction as we understand it today. However Galen did not speculate further on which of the three modes of psychic pneuma conduction was most likely to occur. The next problem concerns that of the transmission of the psychic pneuma into the muscle necessary for the muscle psyche, in Aristotle’s scheme, to realise its final cause and so contract. All that Galen says on this is that transmission must be such as to allow the psychic pneuma to reinforce and initiate the muscle’s intrinsic propensity to contract, that is, to achieve its final cause. However, he did entertain the possibility that this might occur by the psychic pneuma being pushed out of the end of the nerve (Hall, 1975). It was this idea that was to pave the way for a revolution in the approach to conduction and transmission, which followed 1300 years later, and is due to Descartes. 1.3 Descartes: the replacement of pneuma by mechanical corpuscles The great contribution of Descartes (1596–1650; Fig. 1.2A) was to dismantle the concept due to Aristotle two thousand years earlier that all manifestations of life, such as locomotion, nutrition, and sensation are to be attributed to the psyche; engagement of the causal entity then leading to the expression of the inherent capacity of a particular organ to be expressed. As he pointed out (Descartes, 1662): The error is that, from observing how all dead bodies are devoid of heat, and consequently of movement, it has been thought that it is the absence of the soul which has caused these movements and this heat to cease; and thereby, without reason we have come to believe that our natural heat and all the movements of the body depend on the soul. What, on the contrary, we ought to hold is that the reason why soul absents itself on death is that this heat ceases and that the organs that operate in moving the limbs disintegrate. The psyche, as elaborated by Aristotle, was abandoned. This opened up for enquiry the mechanisms by which organs move and heat is produced, cessation of which leads to death. It made transparent the fact that the idea of each organ possessing a psyche, which had prevented the development of physiology for two thousand years, was merely a means of declaring an ignorance concerning the mechanisms of how a particular organ functioned. The loss of the psyche as the causal agent meant that psychic pneuma was no more, leaving open the questions of what is conducted along nerves and transmitted into muscle and how does conduction and transmission occur. To these questions Descartes gave detailed answers, based on his new mechanistic philosophy. In this the body consists of a set of corpuscularly constituted mechanically interacting parts, so that the ultimate level of analysis concerns corpuscular motion. Each part of the body can be activated by the transfer to it of motion that is ultimately derived from heat, which itself is just the agitation of particles engaged in fermentation. Descartes thought this took place in the heart and therefore involved blood particles. This description has a modern ring about it, except of course for the placing of heat generation in the heart. In Descartes scheme large blood particles when they reached the brain were used to nourish it whereas fine blood particles were transformed into a different kind of particle that could be used by the brain for the purposes of conduction along the nerves leaving the brain and spinal cord. This different kind of fine particle to that found in the blood he referred to as animal spirits. Such a name tends to remind one of the psychic pneuma, but in Descartes case the animal spirits were fine particles and accessible to physiological enquiry. Descartes own dissections of the nervous system in his early twenties led him to describe nerves as hollow tubules with a sleeve like double outer sheath, the inner and outer membranes of the sheath being continuous with the inner and outer meninges of the brain. Each nerve tubule contained a central marrow of longitudinal fibrils, surrounded by animal spirits moving outward from the brain, the animal spirits being composed, as we have commented, of highly volatile material particles derived from the blood. In Descartes own words (Descartes, 1972; Fig. 1.3 A): Now in the same measure that spirits enter the cavities of the brain they also leave them and enter the pores or conduits in its substance, and from these conduits they proceed to the nerves. And depending on their entry or their mere

HISTORY OF THE SYNAPSE

5

tendency to enter some nerves rather than others, they are able to change the shapes of the muscles into which these nerves are inserted and in this way to move all the members. Conduction in nerves involves the passage of small particles derived from the heart. Transmission is due to these particles leaving the ends of the nerves and entering the muscle. In order to make sure that the reader is aware of the mechanism of conduction and transmission that he is proposing Descartes comments: But to make you understand all this distinct, I wish to speak to you first of the fabric of the nerves and the muscles, and to show you how from the sole fact that the spirits in the brain are ready to enter into certain of the nerves they have the ability to move certain members at that instant. This description is worth quoting at some length as the 1662):

first detailed account of conduction and transmission (Descartes,

Observe in Fig. 1.3 A [Fig. 1.3 A of this book] for example, nerve A whose external membrane is like a large tube containing several other small tubes, b, c, k, l, and so on, composed of a thinner, internal membrane; and observe that these two membranes (outer and inner) are continuous with the two, K(pia) and L(dura), that envelop the brain MNO. Observe also that in each of the little tubes there is a sort of marrow composed of several very fine fibrils which come from the actual substance of the brain N and whose two extremities end one at the internal surface of the cavities of the brain and the other at the membranes and flesh on which the tubule containing them terminates. But because this marrow is not used to move the members, it will suffice for now that you know that it does not completely fill the tubes containing it but leaves room enough for animal spirits to flow easily through them from the brain into the muscle whither these little tubes, which should be thought of as so many little nerves, make their way. Descartes goes on to say, with respect to Fig. 1.3B: And consider that although there is no evident passage through which the spirits contained in muscle D and E can leave them except to go from one to the other nevertheless because their particles are very small and indeed because they are made incessantly finer through the force of their agitation, some always escape across the membranes and flesh of the muscle while others return through the two nerve-tubes bf and cg to replace those that escape. It will be noted that Descartes retained the basic Galenic idea that the heart was the source of the material used to allow conduction by the nerves, after its transformation in the brain. In the case of Galen and his students that material passed from the heart as vital pneuma, was transformed in the brain to psychic pneuma whence it was used by the nerves that leave the brain and spinal cord for conduction. For Descartes, coarse and fine particles in the blood leave the heart and are sorted by the brain in such a way that the coarse particles are used to nourish it whereas the small particles cease to have the form of blood and become animal spirits. These are able to enter the ‘pores and conduits’ of the brain from which they are guided eventually into appropriate nerves to mediate a particular motor action. The Galenic and Cartesian schemes are very similar except that in the latter we are dealing with definite particles, sorted in a definite way by the brain, with the properties of the particles and their passage in the brain and nerves open to further physiological enquiry. It seems likely that Descartes conceived that conduction of the particles occurs by the mechanism favoured by Galen for psychic pneuma, namely that the particles are forced out of the peripheral end of the nerve in the muscle as a consequence of the entry of particles into the central end of the nerve. As for transmission, it probably required the direct entry of particles from the nerve endings into the muscle cells on which they impinge. However Descartes does not make these points explicit, commenting in relation to the nerves that animal spirits flow… easily through them from the brain without specifying whether this is to be thought of as a travelling wave in time. Indeed it was taken to be a wave of infinite velocity until the experiments of Helmholtz in the nineteenth century. As for transmission from nerve to muscle, he comments in relation to the nerve fibres in the muscle that animal spirits enter therein they cause the whole body of the muscle to inflate and shorten and so pull… while on the contrary, when they withdraw, the muscle disinflates and elongates again. This certainly seems to imply that there is direct flow of animal spirits into the muscle that causes the inflation, although that is not specified and explained in detail until the work of Borelli a few years later.

6

EARLY HISTORY OF THE SYNAPSE

Fig. 1.2. The principal contributors to our understanding of nerve conduction and transmission in the 17th and 18th centuries: from corpuscles to electricity. (A) R.Descartes (1596–1650) (from Bennett, 1997, with permission). (B) Giovanni Alfonso Borelli (1608–1679) (from Brazier, 1959, with permission). (C) Felice Fontana (1730–1805) (from Bentivoglio, 1996, with permission). (D) Luigi Galvani (1737–1798) (from Brazier, 1959, with permission).

1.4 Borelli: a corpuscular description of conduction and transmission The Cartesian hypothesis concerning the mechanism of conduction and transmission was taken up and elaborated on in great detail in the new tradition of physiological enquiry by Borelli (1608–1679; Fig. 1.2B), a young contemporary of Descartes who outlived him by some thirty years. Borelli’s (1670) description of the animal spirits used for conduction by the nerves follows closely that of Descartes (1664):

HISTORY OF THE SYNAPSE

7

Fig. 1.3. The elucidation and refutation of the corpuscular theory of transmission in the 17th and early 18th centuries. (A) The Cartesian model of the nerves cast a long shadow into the 19th century. Descartes (L’homme, 1664) conceived of motor impulses conveyed in the space between the pipes and outer sheath while sensory impulses were conveyed in the inner pipes. For a full description of this figure see the text (from Descartes, 1662). (B) On the left is the Cartesian model of how the nerves proceed to a muscle and control its shortening. In Descartes words ‘Next observe how the tube or little nerve bf proceeds to muscle D, which I assume to be one of those that move the eye, and how it there divides into several branches composed of a loose membrane which can extend, enlarge and shrink according to the quantity of animal spirits that enter or leave it, and whose branches of fibres are so arranged that when animal spirits enter therein they cause the whole body of the muscle to inflate and shorten and so pull the eye to which it is attached, while on the contrary, when they withdraw, the muscle disinflates and elongates again. Observe further that in addition to the incoming nerve-tube bf there is still another, namely ef, through which the animal spirits can enter muscle D, and another, namely dg, through which they can leave it. And quite similarly that muscle E, which I assume is used to move the eye in the contrary direction receives animal spirits from the brain through nerve-tube cg from muscle D through dg, and sends them back toward D through ef’ (from Descartes, 1662). On the right is the original sketch by Descartes illustrating ‘the canals by which the spirits of one muscle can pass into that which opposes it’. Valves in the canals, ‘i’ can open or shut as required for reciprocal innervation (from Brazier, 1984). (C) Croone’s diagram depicting the route (EFG) by which nervous fluid flows from the brain (H) to little bladders which inflate the muscle to expand from contour ABCD (solid line) to AQV (dotted line). On the right is a diagram of the direction of forces when bending the elbow (from Brazier, 1959; Croone, 1665). (D) Swammerdam’s experiments including the one by which he proved that muscles were not swollen by an influx of nervous fluid when they contracted. Fig. V is of an experiment to show the change in shape of a muscle when stimulated by pinching its nerve. Fig. VI illustrates the pulling together of the pins holding the tendons when the muscle contracts. Fig. VIII is the crucial one in which a drop of water is imprisoned in the narrow tube projecting from the vessel enclosing the muscle. Further description of this experiment is given in the text (Biblia Naturae, Amsterdam, 1738; from Brazier, 1959; Swammerdam, 1737).

In the animals, besides liquids such as blood, there is another extremely spirituous fluid substance which is the direct motive cause of the animal body. This appears from the effects of this substance. This spirituous humour is not wind or air but has a liquid consistency such as spirit of wine. It is generated from blood in the brain and diffused by the nerves.

8

EARLY HISTORY OF THE SYNAPSE

All modern authors admit this point. The exact structure and composition of the nervous juice, although unknown, can be surmised somewhat from its motions through the nerves. The mechanism of conduction of the animal spirits along the nerve is due to the compression of the spirits at the central end leading to secretion of spirit at the peripheral end (Descartes, 1664): The spongy cavities of these nervous fibres thus are conceived as being always soaked and filled up to turgescence by some juice or spirit transmitter from the brain. In a bowel full of water and closed at both ends, impulse at an extremity compressed and slightly percussed is instantly transmitted to the other extremity of the turgid bowel. The adjacent elements of the liquid are aligned in a long row. By pushing and shaking each other, they transmit the movement to the extremity of the bowel. Similarly, as a result of some slight compression, jolt or irritation at the origins of the canals of the nervous fibres which are in the brain itself, these fibres thus shaken and activated must secrete some drops of this juice which swells their internal spongy substance, into the fleshy mass of the muscles. Transmission involves the movement of the spiritious juice from the nerve endings in the muscle directly into the muscle cells (Descartes, 1670): The distal orifices of these nervous fibres are scattered everywhere in the mass of the muscle although they are open, the spongy structure itself with which the fibres are provided plays the role of valvules. Indeed droplets hanging from wet sponges do not flow out. A shaking force is required to express them. This may be the cause why the nervous juice is secreted and instilled in all the mass of the muscle by order of the will. The cause and mechanism by which nervous juice is instilled in the muscles with a convulsive force by an order of the will and produces their instantaneous swelling, are deduced from what was said above. Contraction will continue as long as the cause of the bursting is present i.e. the instillation of nervous juice. When it stops, the turgescence of the muscles disappears, as light disappears when the flame which continuously renews it is removed. In summary, Borelli conceives conduction and transmission thus (Descartes, 1670): Consequently, this slight motion of the spirits provoked by the will in the brain can shake or excite the fibres or spongy ducts of some nerves turgid with spirituous juice. As a result of this convulsive irritation which shakes all the length of the nerves, some spirituous droplets can be expressed and spilled from the orifices of their extremities into the corresponding muscle. This results in the boiling and bursting by which muscle is contracted. At the end of the seventeenth century, William Croone summarised for the Royal Society of London the revolution in understanding of conduction and transmission that had taken place that century, involving rejection of the concept of psychic pneuma for that of juices consisting of corpuscles. Fig. 1.3C shows his diagram of the mechanism of conduction along motor nerves and transmission to muscle. What is now called the nervous fluid flows from the brain along the motor nerves to inflate small bladders in the muscle which cause it to expand and shorten (Croone, 1665). 1.5 Fontana: nerves are composed of many cylinders along each of which conduction occurs The composition of the nerves along which conduction occurs was illuminated in the eighteenth century. At its beginning the Dutch microscopist Antoni van Leeuwenhoek (1632–1723) used his one lens microscope to give a description of the composition of nerves, commenting that Often and not without pleasure, I have observed the structure of the nerves to be composed of very slender vessels of an indescribable fineness, running length-wise to form the nerve (Fig. 1.5A). These vessels were taken to be hollow tubes, in agreement with the Cartesian concept that animal spirits flowed in nerves (van Leeuwenhoek, 1685; van Leeuwenhoek, 1717). The relation between these hollow tubes and the nerve was spelt out in detail through the nerve dissections of Felice Fontana (1730–1805; Fig. 1.2C). These were performed, after immersing the nerve threads in water, with very sharp needles under a magnification of X 700, and allowed Fontana to claim that (Fontana, 1760; Fig. 1.4B): The basic structure of nerves is as follows: a nerve is formed of a large number of transparent, uniform, and simple cylinders. These cylinders seem to be fashioned like a very thin, uniform wall of tunic which is filled, as far as one can see, with transparent, gelatinous fluid insoluble in water. Each of these cylinders receives a cover in the form of an outer sheath which is composed of an immense number of winding threads. A very large number of transparent

HISTORY OF THE SYNAPSE

9

cylinders together can form a nerve so small that it is barely visible but which shows the white bands on the outside. Several of these nerves together form the larger nerves seen in animals. I am fully convinced by my own observations, which I repeated many times with the same result, that the cylinders I have described are the simple and first organic elements of nerves, for I have not succeeded in dividing them further, no matter what investigations I carried out with the help of the sharpest and finest needles. I could easily tear and break them here and there; but they always remained indivisible. I could strip them off their sheaths and separate the winding cylinders of which they are formed, although they were very small. The primitive nerve cylinder then appeared transparent, homogeneous, and of equal diameter everywhere. He goes on to say that: After having dissected a very small nerve and its minimal nervous threads made of the different nervous primitive cylinders I have extensively dealt with in my work, I succeeded in stripping from the inner sheath, or rather from the tortuous threads, some nervous primitive cylinders. These were transparent, homogeneous, not empty, as I had found them in previous occasions. As to the constituents that made up the cylinders, that were ‘not empty’, Fontana in 1781 describes the microscopic features of the axoplasm extruded from the cut end of an axon as (Fontana, 1781): …glutinous, elastic, transparent material, insoluble in water, that decomposed itself into very little round grains of a diameter four or five times less than a red blood globule’. ‘I am not sure that Physiologists would be willing to consider those little grains as animal spirits, and the mechanical principle of all movements. This hypothesis would not explain the instantaneous speed of animal movements, since those little grains seem too lazy to move inside the nerve, where they form instead a viscous and inert glutine. Animal movements would be easier to explain considering that such grainy material is elastic, and continuous along all the nervous canal, as the observation in fact demonstrates. The movement could be transmitted at the moment that would follow a mechanical alteration of the nerve or any of its parts. These descriptions of the larger nerves as composed of smaller nerves which are not divisible further and which contain ‘a glutinous, elastic, transparent material’ has modern resonances. However Fontana produced this description in the year that Galvani began his most important discoveries. These were to identify electricity as the conducting material for nerves rather than the Cartesian corpuscles of fine particles derived from blood. Fontana then adheres still to the Cartesian animal spirits and so has difficulty in reconciling the size of the lazy particles in the nerve cylinders with that of the speed of animal movement. He then comes to emphasise the possibility favoured by Galen that it is the extrusion of the particles at the peripheral ends of the nerves following the entry of particles at the central ends of the nerves that provides the appropriate speed for nerve action (Fontana, 1767; Fontana, 1775). 1.6 Galvani: electricity is conducted and transmitted not corpuscles Borelli had commented in relation to the idea of the flow of a nervous fluid in nerve to muscle that All modern authors admit this point. Indeed when Boerhaave produced the first figure of the neuromuscular junction in the early part of the eighteenth century (Boerhaave, 1735; Fig. 1.4C), it emphasised continuity between the nerve ending and muscle, as expected if there was to be a direct flow of nervous fluid into the muscle required for muscle shortening. This whole conceptual scheme was dealt a major setback with the brilliant physiological experiments of Swammerdam, published in 1738. These showed that muscles were not swollen by an influx of nervous fluid during contraction. In this work, illustrated in Fig. 1.3D (VIII), he placed a muscle with its nerve supply in a narrow tube which was then filled with water in such a way that water could be expelled from the tube if the muscle swelled on contraction (Swammerdam, 1737). A wire attached to the muscle nerve in the tube was then pulled on to excite the nerve and contract the muscle. The result was unequivocal, muscle contraction did not lead to the expulsion of water from the tube, so that muscle swelling could not have taken place. Nervous fluid could not, by flowing directly into a muscle, cause contraction. The whole concept of a nervous fluid, consisting of small Cartesian corpuscles, was now thrown into doubt. What could be the nature of the substance that was conducted by nerves? The science of electricity emerged in the sixteenth and seventeenth centuries. William Gilbert (1544–1603) had constructed the first electroscope, consisting of a suspended needle that was attracted to static electricity, so that it turned on being brought near a piece of rubber amber. This apparatus allowed the amount of attraction due to the static electricity to be given in quantitative terms according to the extent of deflection of the needle. Gilbert used the electroscope to detect static electricity in a number of other rubbed objects consisting of glass, wax and sulphur. This led to the invention by Otto von

10

EARLY HISTORY OF THE SYNAPSE

Fig. 1.4. Anatomical identification in the 17th±19th centuries of nerve fibres and their junctions with muscle. (A) Leeuwenhoek’s drawing of a small nerve (ABCDEF) composed of many ‘vessels’ in which ‘the lines or strokes denote the cavities or orifices of these vessels’. This nerve is surrounded in part by five other nerves (one of which is labelled G), in which only ‘external coats’ are represented (from Bentivoglio, 1996; van Leeuwenhoek, 1685). (B) Nerve fibres drawn by Fontana. The drawing illustrates ‘a nerve torn with a needle, to determine the continuity of the primitive nervous cylinders’, a indicates the ‘two ends of the nerve’, c,n,o indicate ‘several of the primitive cylinders’ (from Bentivoglio, 1996). (C) Boerhaave’s concept of the neuromuscular junction. He believed that the nerve (EC) flowed directly into the substance of the muscle (HB) (from Boerhaave, 1735; Brazier, 1984). (D) Schematic summary view of the mammalian neuromuscular junction (from Kuhne, 1869; Shepherd, 1991).

Guericke (1602–1686) of the frictional machine which was constructed from a sulphur ball mounted on a spindle and rotated by hand to generate large quantities of static electricity. The opportunities for the discovery of animal electricity were in place with the subsequent invention of the Leyden jar for storing static electricity by Petru van Musschenbroek (1692–1761; Fig. 1.4A). Indeed speculations that electricity might compose the Cartesian animal spirits were made at this time by the mathematician Christian August Hausen (1693–1743). Luigi Galvani (1737–1798: Fig. 1.2D) discovered ‘animal electric fluid’, a phrase reminiscent of the ‘animal spirits’ used by Descartes in his mechanical description of nerve conduction. This story begins on the famous occasion during which one of Galvani’s collaborators touched with a lancet the exposed nerve of a frog muscle near a frictional machine (Fig. 1.5C),

HISTORY OF THE SYNAPSE

11

which occasionally sparked giving rise to the transfer of charge by induction to the frog’s nerve and thereby a twitch contraction. Galvani investigated this phenomenon further using the apparatus shown in Fig. 1.5C (Galvani, 1791). This consisted of a frog’s exposed spinal cord-leg preparation suspended in a sealed jar by means of a wire passed through the spinal cord and then through a seal at the top of the jar; lead shot was present in the bottom of the jar (right hand side of Fig. 1.5C). A wire was then strung across the ceiling to pick up the charge from a frictional machine and convey it to the wire from which the spinal cord was strung, as shown in Fig. 1.5C. This apparatus allowed for the unequivocal demonstration that when the machine sparked the legs twitched. From this Galvani concluded that frog nerves conduct electricity. Animal spirits had become electricity. Galvani devised a number of other experimental procedures in the years 1781 to 1791 which showed the existence of electrical conduction in nerves (Galvani, 1783; Galvani, 1791; Galvani, 1794). In one experiment he used a pair of jars in one of which there was enclosed a frog spinal cord-leg preparation suspended over lead shot as before by means of a fine iron wire; this wire then lead into another jar which in turn had a layer of lead shot, together with a coil of attached wire to collect the discharge from the frictional machine, as shown in Fig. 1.5B. This discharge was accompanied by sparking in the upper jar and twitching of the frog’s legs in the lower jar, due to what Galvani described as the passage of ‘electric fluid’ down the wire in the upper jar and down the spinal cord and nerves into the leg muscles. His conclusion from these experiments was that there must be a nervous ‘electric fluid’. That this electric fluid must flow along individual nerves and not just the spinal cord was confirmed by work in which the sciatic nerve of one leg of a frog was dissected and used in the experimental apparatus instead of the spinal cord. In this case the leg twitched on discharge of the friction machine as had been the case with the isolated spinal cord-leg preparation. An investigation of Galvani’s in 1794, which was to have far reaching repercussions in the following century in the hands of du Bois-Reymond, involved experiments that were to lead to the discovery of ‘animal electricity’, that is the existence of electricity generated by nerve and muscle itself (Galvani, 1791). In this experiment Galvani placed the severed end of a nerve, belonging to a leg-muscle preparation, on the intact portion of the nerve and obtained movement of the leg (Fig. 1.5E). This showed the existence of electrical potential in nervous tissue and that electrical flow could occur in nerves as a consequence of the nerves producing a potential. He published this work anonymously as ‘Dell’ uso e dell’ attivita dell’ Arco conduttore nelle contrazioni dei muscoli’ (On the application and activity of the Arco conduttore in the contraction of muscle). In 1797 Galvani showed that if he allowed one nerve of a nerve-leg preparation to fall from a glass rod on which it was suspended onto the cut region of another nerve from the same frog then the legs moved (Galvani, 1797). It was not then necessary that the same nerve be used to excite itself, but that any injured nerve would suffice. Indeed electricity could be led by a suitable conductor from the cut end of the spinal cord where the potential was generated to the leg directly in order to obtain a twitch (Fig. 1.5D). One of Galvani’s most famous demonstrations of the flow of electricity in nerve involved the observation that frog’s legs twitched when hung from brass hooks to an iron railing even in the absence of a thunderstorm. Galvani interpreted this as due to the generation of animal electricity rather than, as Volta was later to show, to the flow of current between dissimilar metals connected in a circuit. However Volta went further and attempted to analyse all of Galvani’s experiments as an artefact due to this phenomenon (Volta, 1918a; Volta, 1918b). It was left to Alexander von Humboldt (1769–1859) to confirm Galvani’s experiments and show that they occurred independently of any current flow due to dissimilar metals being incorporated into the experimental design, something which Galvani himself had shown (see Fig. 1.5E). Galvani met Volta’s criticism by cutting both sciatic nerves of a frog where they leave the spinal cord. He then lifted the cut end of one nerve with a glass rod so that it touched the other nerve with its cut end. When this occurred the muscle of the touched nerve contracted. Fig. 1.5 (F3) shows one of von Humboldt’s experiments in which he applied a charged tube of glass to the nerve of an isolated frog nervemuscle preparation and obtained a contraction. Fig. 1.5 (F6) shows another experiment in which he turned the cut end of the nerve against the muscle and obtained a contraction, a variation of the Galvani experiment in which the cut end of a sciatic nerve was placed on another sciatic nerve (von Humboldt, 1797). None of these experiments were open to the kinds of criticism that Volta had aimed at Galvani. As we have seen, at the end of the seventeenth century nerves were thought to conduct animal spirits rather than psychic pneuma, with the former envisaged as corpuscular in nature, derived from fine particles of blood. Galvani died at the end of the eighteenth century, by which time he had shown that nerves could conduct electricity and further that the potential for generating electricity could be found in nerve and muscle itself (Bresadola, 1998). It was generally accepted after this work and that of von Humboldt that the conduction of electricity in nerve was like the way in which metallic wire conducts voltaic electricity. Animal spirits had become electricity.

12

EARLY HISTORY OF THE SYNAPSE

Fig 1.5. The emergence of the concept in the late 18th century that electricity is conducted by nerves. (A) Two early Leyden jars in the collection from the Boerhaave Museum, Leyden. (Photograph by courtesy of the Boerhaave Museum). (B) Galvani’s sketch of his preparation of inverted flasks containing the frog’s nerve muscle preparation from an experiment dated December 10, 1781 (from Brazier, 1984). (C) This figure shows Galvani’s frictional machine, a Leyden jar, and a wire strung across the room to collect the charge (from Brazier, 1984; Galvani, 1791). (D) An artist’s depiction of Galvani’s favourite preparation published as part of the first illustration to the famous Commentary published in 1791 (from Brazier, 1984). (E) The critical experiment by Galvani on muscle contraction in the absence of all metals (from Brazier, 1984). (F) Von Humboldt’s experiments in which he demonstrated contraction of nerve-muscle preparations in the absence of any metals. His Fig. 3 depicts a frog nerve-muscle preparation to which he applied a tube of glass (x), producing a contraction. His Fig. 6 shows an experiment in which he turned back the nerve against the muscle without interposing the glass rod (from Brazier, 1984; von Humboldt, 1797).

1.7 Matteucci and du Bois-Reymond: transient electrical changes are conducted (the action potential) The triumph of nineteenth century physiology was to take Galvini’s discoveries and show that the nervous primitive cylinders of Fontana possess a potential across their membranes that could give rise to a propagating transient potential change, the action potential.

HISTORY OF THE SYNAPSE

13

As is so often the case in the history of neurophysiology, the development of concepts concerning electricity in nerve and muscle at the beginning of the nineteenth century was dependent on advances in instrumentation. Gilbert’s use of a suspended needle to detect electricity permitted only slight deviations from the meridian because of the earth’s magnetic field. In 1820 Schweigger, following Oersted’s discovery that a magnet tends to set itself at right angles to a loop of bent wire carrying an electric current (Oersted, 1820), designed the galvanometer. In this instrument many turns of wire were wound on a rectangular frame inside which a compass needle was placed that was balanced on a vertical pivot or in some cases suspended from a thread. Leopold Nobili in 1825 manufactured the first astatic galvanometer in which he wound two coils of wire on the rectangular frame of Schweigger in opposite directions, so as to cancel the effects of the earth’s magnetism. Nobili used this instrument in 1827 to detect currents passing up the body of a frog away from the legs towards a cut spinal cord and in this way made the first measurement of animal current, or as he called it the ‘intrinsic current’. However he attributed the current to a thermoelectric effect caused by the unequal cooling of nerve and muscle produced by evaporation rather than due to an intrinsic biological phenomenon (Nobili, 1824; Nobili, 1828). Carlo Matteucci (1811–1865; Fig. 1.6A) used the Nobili galvanometer to great effect on isolated nerve-muscle preparations. Although not well known he may be considered to be a founding father of electrophysiology. Matteucci showed that a twitching muscle generated current sufficient to stimulate the nerve of another muscle laid across it and so produce a twitch in the other muscle. Importantly he detected current flow between the cut end of a muscle and the intact end. These currents were correctly interpreted as generated by the muscles themselves (Matteucci, 1842a; Matteucci, 1942b). This was emphasised by his experimental technique of preparing a pile of sectioned frog’s thighs arranged in a series so that the intact surface of one thigh was in contact with the sectioned surface of the next one. The currents generated were in proportion to the number of thigh sections in the pile. However Matteucci’s most important observation was that the current between the cut end of a muscle and the intact end declined during a tetanus caused by strychnine, that is there was a negative variation in the current. Thus excitability was associated with a decrease in the potential that gives rise to the current. Although Matteucci was unable to detect with his instruments a negative variation in the nerve current, his observations laid the ground work for the emergence of the concept of the action current and of the action potential (Matteucci, 1838; Matteucci, 1844; Matteucci, 1848; see also Matteucci & Humboldt, 1843). du Bois-Reymond (1818–1896; Fig. 1.6B), confirmed Matteucci’s experiments on nerve-muscle preparations, and on muscles isolated from their nerve supply, calling current flow in the latter case ‘muscle current’. Most importantly the negative variation or ‘negative Schwankung’ of the muscle current during a tetanus was shown in 1843 to be produced by other means than strychnine, for instance by direct faradic stimulation. du Bois-Reymond, using more sensitive instrumentation than that available to Matteucci, was able to detect in 1834 the negative variation in nerves as well as muscle. The concept of the action potential with its action current showing up as a negative variation was clearly envisaged by du Bois-Reymond (1841; 1842). He hypothesised that a resting potential existed between the middle of muscle cells at positive potential and that of the tendons at negative potential: it was this potential which decreased during stimulation so that a negative variation was recorded. He went on to develop the concept of ‘electromotive particles’ or ‘electrical molecules’ (du Bois-Reymond, 1877). These possessed a positive charge in their middle and a negative charge at each of the polar regions. He postulated that these were situated along the length of the surface of muscle cells and nerve fibres and that it was these that gave rise to the polarisation of the cells (Fig. 1.6D). At rest these molecules were postulated to be arranged in an ordered longitudinal array, so that if a nerve or muscle was sectioned transversely this gave rise to the muscle or nerve current between the injured regions and the intact surface. An electrical stimulus was envisaged to perturb the ordered longitudinal array, producing an electrotonic disturbance leading to the initiation of the negative variation. In this ‘molecular hypothesis’ muscle and nerve fibres are composed of strings of so-called peripolarelectric molecules each of which possess an equator corresponding to the electropositive metal zinc and two poles corresponding to the electronegative metal copper. The current attributed to the internal potential difference thus created could be led off by placing one end of a conductor on the ‘natural longitudinal section’ of a nerve or muscle and the other end on the ‘natural cross section’; in this case the longitudinal section acted as the positive pole and the cross section as the negative pole. The term ‘natural cross section’ as applied to muscle refers to the tendon covered ends of the muscle, regarded as prisms or cylinders, while the term ‘natural longitudinal section’ refers to the lateral surface of these prisms or cylinders. The corresponding artificial cross section and longitudinal section are obtained by dividing the muscle length-wise or crosswise. In this sense the proximal cross section is the upper one, and the distal cross section the lower one. The same applies mutatis mutandis for the nerve. The negative variation involved the discharge of this electromotive force, an idea that clearly presaged the concept of the resting membrane potential and its depolarisation during the action potential (du Bois-Reymond, 1860; du Bois-Reymond, 1845; for a recent detailed account, see Piccolino, 1998). Matteucci had discovered the negative variation in muscle that accompanies activity and du Bois-Reymond that in nerve. Although it seemed very likely that this negative variation of electrical polarisation was the animal spirit of Descartes, it was still endowed with the mysterious property that it could travel at infinite velocity. This was accepted by all the leading physiologists of the first half of the nineteenth century. For example in 1846 E.Weber summarised his observations on the

14

EARLY HISTORY OF THE SYNAPSE

Fig 1.6. Identification of the action potential as the electrical means of conduction. (A) Carlo Matteucci (1811–1865) (from Brazier, 1959). (B) Emil du Bois-Reymond (1818–1896) (from Brazier, 1959). (C) H.von Helmholtz (1821–1894) shown as a young man when he made his greatest contribution to the understanding of impulse conduction in nerve (from Brazier, 1959). (D) Schemata of du Bois-Reymond’s postulated method for transmission at the motor end plate (from du Bois-Reymond, 1845). (E) Helmholtz’s apparatus for measuring the time course of muscle contraction and the propagation velocity of the nerve impulse. On the left, Figure 1 shows the entire apparatus; on the right, Figure 2 shows the arrangement when the nerve w is attached and more than one point on the nerve can be stimulated (from Cahan, 1993; Helmholtz, 1850a). (F) Helmholtz’s muscle curve (from Cahan, 1993; Helmholtz, 1850a).

conduction of the action current in muscle nerves with the comment that (Weber, 1846): When one stimulates a muscle through a motor nerve its movement occurs at the same moment that is the movement begins and ends with the stimulus. Muller’s comment was that (Muller, 1840; Muller, 1851): the time the stimulus takes to travel to the brain and back is infinitely small and unmeasurable. The existence of such a phenomenon as a travelling wave with infinite velocity left the mechanism of conduction opaque to further analysis, rather in the way that the idea of the psychic pneuma had until the time of Descartes. This impasse was broken through the experimental skill of Helmholtz (1821–1894; Fig. 1.6C).

HISTORY OF THE SYNAPSE

15

1.8 Helmholtz: the action potential has a finite velocity In 1848 Helmholtz began, in his own words (Cahan, 1993): …to study the processes occurring in the simple contraction of a muscle; by such an action I mean one that results from a stimulus of vanishingly small duration. I have now finished building my frog-tracing machine and have already carried out a few tracing experiments on mica sheets. Instead of the frog muscles, I inserted a spring. The weight hung from it, oscillated up and down, and recorded its movements. The traces are much prettier than the earlier ones, very fine and regular [see Fig. 1.6F]. The previously unknown fact that in animal muscles too, as in the case in much longer time intervals in organic muscles, the energy of the muscle does not develop completely at the moment of an instantaneous stimulus. Rather, in most cases after the stimulus has already ceased, it increases gradually, reaches a maximum, and again subsided. [Fig. 1.6F]. The force of the muscle was not strongest directly after the stimulation, but rather increases for a time and then falls. In October 1849 he set out to give a more accurate account of this apparent delay between the electrical stimulus and the muscle’s response, with its implications for a finite velocity of conduction of the nerve action potential. To this end he used a method for measuring small time intervals based on the fact that the length of the arc through which the magnetic needle of a galvanometer moves when a transitory current passes through its coil is proportional to the duration of the current. The time interval in question was then measured by ensuring that this movement was led to the deflection of a mirror. Helmholtz next arranged his apparatus so that the beginning and end of the time interval marking the currents duration coincided with the time interval that began with the application of the stimulus and ended with the onset of the muscle’s mechanical action. The latter was obtained by the mechanical action of the muscle lifting a weight that then placed a break on the electrical current. It is worthwhile analysing the apparatus that Helmholtz used for this experiment, both for the beauty of its design and for the unequivocal outcome that it led to, namely the accurate measurement of the velocity of the action potential. With reference to Fig. 1.6E, and following the description in Cahan (1993): …a muscle was hung suspended from a screw I to which was attached a series of screws and contact surfaces that would break the flow of the current when the muscle raised the weight suspended on the scale pan K. The apparatus was placed in a container with humidity enriched air in order to prevent the muscle from drying out; this set up remained in a usable state for three to four hours. At a certain point following stimulation of the nerve w (see the enlargement on the right hand side of Fig. 1.6E), where v is the current carrying wire to the nerve, the energy of the muscle would equal the load suspended from its lower end. After that point, any increase in the energy of the muscle would elevate the load a little and separate point m from the point n on the apparatus; if, however, weights were put on the scale pan K, such that the muscle was acted upon by an additional overload, then the stimulated muscle could raise the combined weight only if its energy (elastic Spannung) equalled the sum of the weights of the load and the overload. Helmholtz arranged his apparatus so that the current, whose time interval was to be measured, would break when the elastic Spannung of the muscle increased by an amount sufficient to raise the weight of the overload. With this approach Helmholtz discovered that the time interval between a stimulus applied to the nerve and the moment when the muscle produced enough force to lift the overload depended on the distance between the point of stimulation of the nerve and the muscle. A method for measuring the velocity of propagation of the action potential was now available. On 29 December 1849, Helmholtz measured the velocity of propagation as 30.8 metres per s. The curves obtained in these experiments (Fig. 1.6F) showed that the velocity was finite and the apparatus (Fig. 1.6E) that this velocity could be measured (Helmholtz, 1850a; Helmholtz, 1950b; Helmholtz, 1850c; see also Olesko & Holmes, 1993). Helmholtz had made the great discovery which transformed the nervous system from consisting of cylinders through which animal spirits flowed with infinite velocity to one which was amenable to quantitative measurement for the testing of hypotheses (Helmholtz, 1851; Helmholtz, 1852). He knew that as long as physiologists insist on reducing the nerve effect to the propagation of an imponderable or psychic principle, it will appear unbelievable that the velocity of the current should be measurable. Although the mechanism by which this propagation of the action potential occurred was not indicated by these experiments, Helmholtz proposed the first hypothesis to be tested, namely that the process was the same as that of the conduction of sound in the air and in elastic matter or the burning of a tube filled with an explosive mixture. J.Bernstein (1839–1917) began the use of quantitative measurement of nervous phenomena to investigate the mechanism of propagation of the action potential in 1868 with his measurement of the time course of the potential, its latency, rise-time and decay (Bernstein, 1868). This led him to his famous theory that the membrane of nerve and muscle is normally polarised at rest with an excess of negative ions on the inside and of positive ions on the outside. The action potential then becomes a self-

16

EARLY HISTORY OF THE SYNAPSE

propagating loss of this polarization or a depolarisation. Injury currents arise as a consequence of the loss of the polarization at some point in the membrane due to its injury. Bernstein was led to this theory by comparing the negativity of the current arising from an injured nerve and that of the action current (Bernstein, 1871). As the former was attributed to the inside of the tissue becoming continuous with the outside it was conjectured that a transient exposure of the inside to the outside occurs during the action current (Bernstein, 1912). The fact that the action potential could conduct in the orthograde or anterograde directions followed from this theory of the polarization of the nerve and muscle cell membrane. Perhaps the most original contribution of Bernstein was his theory of the origin of the polarization of the membrane of muscle and nerve, which was based on Nernst’s concept of the diffusion potential developed at about the same time (Nernst, 1888; Nernst, 1908). In this theory the potential arose as a consequence of the high permeability of the membrane to potassium compared with other ions. Given that the concentration of potassium is higher inside than outside a negative polarization of the inside of the membrane with respect to the outside is generated. 1.9 Kuhne and Auerbach: identifying the structure of nerve endings on muscle and neurons The advent of superior histological stains allowed the first descriptions of neurons to be given in 1836–1837 by C.G.Ehrenberg for single nerve cells in the leech nervous system (Fig. 1.7A) and by J.Purkinje for the large nerve cells of the mammalian cerebellum named after him (Fig. 1.7B; see also Purkinje, 1837a; Purkinje, 1837b; Purkinje, 1937c). Sixteen years later A.Kolliker (1817–1905; Fig. 1.8A) showed that the nerve fibres of Fontana originated from nerve cells with a description of the beginnings of the acoustic VIII nerve, or in his words in relation to Fig. 1.7C: Nerve-cell with the origin of a fibre (from the acoustic VIII nerve) of the Ox; a, membrane of the cell; b, contents; c, pigment; d, nucleus; e, continuation of the sheath region of the nerve-fibre; f, nerve-fibre. In the same year of 1852 he described how motor nerves originated from the anterior horn nerve cells of the spinal cord (Fig. 1.7D). The question then arose as to the relationship between these motor nerves and the muscle cells on which they impinge. R.Wagner showed in 1847 that the terminal branches of nerves going to the electric organ of the electric ray split into ever finer branches when they entered the electric organ until nothing was left of them in the fine-grained parenchyma of the organ (Wagner, 1847). He extended this by analogy to the nerve supply of muscle. W.Kuhne (1837–1900) described histological differences between the end of the motor nerve and the muscle cell on which it abuts in frogs in 1862, namely at the end-plate (Fig. 1.4D; see also Kuhne, 1862; Kuhne, 1888; Kuhne, 1869). However this did not illuminate the functional problem of how the action potential was transmitted from motor nerve to muscle any more than did the diagram of the motor endplate by Boerhaave’s some one hundred and forty years earlier (Fig. 1.4C) indicate how animal spirits left the nerve to enter the muscle. Boerhaave had simply acquiesced in the current physiological concept of Descartes, developed by Borelli, that animal spirits passed directly from the nerve into the muscle to increase the muscle volume and so shorten it. Kuhne, likewise, took the current physiological paradigm of the action potential and suggested that the action current of the nerve invaded the muscle at the endplate. This idea was developed in some detail by W.Krause in 1863, who drew attention to the similarity between the nerve endings in muscle and the electric plate in the organ of the electric catfish, which was taken to act as a Leyden jar. He argued that the nerve ending, the nerve endplate, charged like an electric plate when the nerve was excited, giving to the contractile substance of the primitive muscle fascicle an electric shock so stimulating it to contract. Once it had been established that nerve fibres originate from nerve cells the question arose as to the relationship between the nerve fibre endings in the nervous system and nerve cells. Given that the histological methods using silver staining as well as microscopical techniques current at that time were not able to give an appropriately detailed account of the relationship between nerve ending and muscle at the endplate, they were certainly not able to illuminate the question of the relationship between nerve ending and nerve cell. This area of study was ripe for speculation, without necessarily furthering understanding, and as a result controversy followed. In 1865 O.Deiter (Deiters, 1865) showed that dendrites (or as he called them protoplasmic processes; these were named dendrites by W.His in 1889, (Waldeyer, 1891)), in addition to nerve fibres (or as he called them axis cylinders; these were termed axons by Kolliker in 1896; Kolliker, 1856a, b), arose from the nerve cell (named the neuron by Waldeyer in 1891), a fact that he illustrated beautifully with his drawings of neurons dissected from the ox spinal cord (Fig. 1.9A). He noted in passing that the dendrites possessed trumpet-like expansions on their surfaces attached to very fine fibres (Fig. 1.9A), which he speculated could represent the input to the dendrites. In 1898 L.Auerbach showed, with silver stain, the ‘end bulbs’ of nerve fibres on the surface of neurons in the facial nucleus, which he unequivocally identified as such (Fig. 1.9B; Auerbach, 1898). He also subscribed to the Cartesian/Borelli idea that there was continuity in the propagation from nerve to target cell, commenting that (Auerbach, 1898):

HISTORY OF THE SYNAPSE

17

Fig. 1.7. Identification in the first half of the 19th century of the neuron as the cell body giving rise to the nerve fibre or axon. (A) Single neurons from the leech nervous system representing one of the first two illustrations of a neuron (from Ehrenberg, 1836). (B) The large corpuscles of the cerebellum, which became known as Purkinje cells after their discoverer, giving the other first illustration of a neuron. This was also the first published view of the cellular composition of the histological layers within a brain region. From below: fibres, granules, large corpuscles (Purkinje cells), molecular layer (from Purkinje, 1837a; Shepherd, 1991). (C) A nerve originates from a cell body. A nerve cell with its nerve in Kolliker’s Handbuch der Gewebelehre des Menschen (1852). In his words: ‘Nerve cell with the origin of a fibre from the acoustic VIII nerve of the Ox; a, membrane of the cell; b, contents; c, pigment; d, nuclear, e, continuation of the sheath region of the nerve fibre; f, nerve-fibre…’ (from Kolliker, 1896; Shepherd, 1991). (D) Another nerve cell from Kolliker’s (1832) textbook for comparison with Fig. C. This is a ‘large nerve cell with processes from the anterior cornua (horn) of the spinal end in man’ (from Shepherd, 1991).

As I understand that theory, the axon terminals exert their effects on the cell surface of the ganglion cell by means of close contact of the end bulbs, without intervention of any intermediate substance.

18

EARLY HISTORY OF THE SYNAPSE

Fig. 1.8. Major figures at the end of the 19th century responsible for the emerging definition of the synapse. (A) From left to right, Guilio Bizzozero (a friend of Golgi), Albrecht Kolliker (1817–1905) and Camillo Golgi (1844–1926) at Golgi’s home in Padua in 1887 (from Prof. P.P.C.Graziadei; from Shepherd, 1991). (B) Santiago Ramon y Cajal (1852–1934) (from the frontispiece in Cajal, 1995). (C) Charles Scott Sherrington (1858–1952) (from Granit, 1966).

1.10 Cajal: nerve endings are not continuous with the cells on which they impinge The use by C.Golgi (1842–1926; Fig. 1.8A) in 1886 of a silver stain in which potassium bichromate and silver nitrate are applied to produce a black impregnation of neurons did not help resolve the nature of the region of nerve fibre endings on cells. On the one hand Golgi claimed that he could detect intracellular neurofibrils in nerves with his silver technique and that these extended from the end bulbs of the nerves into the neurons on which they ended, providing continuity for the transmission of the action potential (Golgi, 1886a; Golgi, 1886b). No better illustration of the implications of this continuous

HISTORY OF THE SYNAPSE

19

reticular network of neurofibrils through nerve fibres, end bulbs and neurons is provided than that of Golgi’s drawing from his stained material of a transverse section through the hippocampus (Fig. 1.9C). Golgi emphasises in the legend to this figure that the axons ending on the granule cell neurons (upper right part of the figure in the fascia dentata; axons of the pyriform pathway) merge with the dendrites of the granule cells and that the nerves which emerge from the granule cell neurons (the mossy fibre axons, shown converging to form a single nerve in area CA3) merge with the dendrites of the CA3 pyramidal neurons. In his words (Golgi, 1886b): The diagram illustrates particularly the mode by which a fascicle of nerve fibers comes in relation to the small ganglion cells of the fascia dentata. Between the fascicle of nervous fibers still maintaining themselves as individual elements and the nervous prolongations (axons) of the small cells, exist a complicated network, occupying a semicircular area, which, especially toward the deep part, has indeterminate borders. It is on entering into this network, that, ramifying, a part of the nervous prolongations (axons) lose themselves, as well as the fibers deriving from the fascicle. The latter, issuing from the semicircle formed by the fascia dentata, traverse the zone of the grey layer of the convolution, occupied by the bodies of the cells which belong to this layer, and go to join the fibers of the Alveus and Fimbria. The conceptual framework here is that of Descarte/Borelli for the neuromuscular junction, namely with continuity between nerves and the structures on which they end. However in 1866, when Golgi stated that continuity existed between nerves, quite different conclusions were being reached by other histologists (Finger, 1994). W.His suggested, on the basis of Kühne’s description of nerve endings on muscle fibers (Kuhne, 1862; Kuhne, 1888), that ‘the motor endplates give the indisputable example of transmission of a stimulus without continuity of substance’ (His, 1886). In 1887, F.Nansen, using Golgi’s silverstaining technique on the nervous system of invertebrates, concluded that ‘a direct combination between ganglion cells, by direct anastomosis of the protoplasmic process, does not exist’ and that ‘the branches of the nervous processes do not anastomose’ (Nansen, 1887). Finally, A.Forel in 1887 used the Golgi technique to show that after a lesion ‘total atrophy is always confined to the processes of the same group of ganglion cells, and does not extend to the remoter elements in merely functional connections with them’ (Forel, 1887). None of these observations gave clear evidence one way or the other as to whether nerve endings are continuous with the cells on which they impinge, although the research of Forel on the effects of a lesion paved the way for the definitive work on this problem by a remarkable neurohistologist. S.Ramon y Cajal (1852–1934; Fig. 1.10B) learnt of Golgi’s silver staining technique in 1887 and applied it to blocks of the cerebellum in 1889. From this earliest work Cajal claimed that the terminal baskets of the stellate neurons could be seen to envelop the Purkinje neurons on which they ended without any sign of their being in continuity with the neurons. He developed from this the neuron doctrine, namely that each neuron is an independent cell that does not anastomose with surrounding cells (Cajal, 1995; Clarke, 1995). This doctrine is deservedly attributed to Cajal for two reasons: one is the trenchant way in which he defended it against the contrary claims of other neurohistologists, which he did by using their experimental techniques to show how their claims were based on artefacts; the other was his definitive degeneration studies in which he showed that the loss of a particular neuron type could leave behind the synaptic terminals that impinge on it, indicating that the latter were not in continuity with the former. Examples of his forceful style in relation to the defence of his work claiming that nerve terminals can be shown in silver stained material to abut but not anastomose with neurons are as follows (Cajal, 1995): Gerlach concluded that certain axons anastomose at their endings with the tips of dendrites, and, thus that central axonal arborizations do not end freely but instead merge with dendrites. Because Golgi thought that central axonal arborizations do not end freely but anastomose instead, his hypothesis is actually based on Gerlach’s theory. To demolish the theory, it was necessary to show by direct observation in the adult brain that axonal arborizations terminate freely, and in the final analysis to do so under conditions that no one could object to because observations were in embryonic material or because material was improperly stained. We were the first, in 1888, to demonstrate unequivocally and irrefutably that terminal arborizations end freely. One need only recall the varied and often profound alterations that occur in dendrites stained with Ehrlich’s method giving varicosities, cyanophilic masses and abnormal thickenings, which may condense or fuse with one another. When this happens the resulting images look so much like anastomoses that they may readily be mistaken for them. In the embryonic and adult spinal cord, in the cerebellum, cerebral cortex, Ammons horn, striatum, and olfactory bulb, in the autonomic nervous system, in the spinal ganglia, retina, and elsewhere, the terminal arborizations of axons and dendrites invariably end absolutely freely—a fact that can be demonstrated equally well by the Golgi and the Cox methods. He therefore concluded that (Cajal, 1995):

20

EARLY HISTORY OF THE SYNAPSE

Fig. 1.9. The search during the second half of the 19th century for nerve endings in the brain. (A) The concept that protoplasmic processes (b) and the axis cylinder (a) are different prolongations of the same cell was represented by Deiters in 1865 in a neuron dissected from the spinal cord of an ox (from Max Schultze, 1870). Deiters thought that input to the protoplasmic processes (dendrites) was via fine fibres connecting by means of a trumpet-like expansion (arrows added by another author) (from Deiters, 1865; Jacobson, 1993; Schultze, 1870). (B) The first representation of synaptic endings (Endknopfchen) in the central nervous system (facial nucleus; reduced silver preparation; paraffin section). From Leopold Auerbach (1898). He concluded that this was evidence in support of the contact theory of nerve connections (from Auerbach, 1898; Jacobson, 1993). (C) A diagram by Golgi of the nervous elements of the hippocampus and fascia dentata (from Golgi, 1886a; Shepherd, 1991). (D) A diagram by Cajal illustrating the structure and connections of the hippocampus and fascia dentata. A, retrosplenial area; B, subiculum; C, Ammon’s horn; D, dentate gyrus; E, fimbria; F, cingulum; G, angular bundle or dorsal hippocampal commissure (crossed temporoammonic path); H, corpus callosum; K, recurrent collaterals from pyramidal cells to the stratum lacunosum of Ammon’s horn (Schaffer collaterals); a, axon entering cingulum; b, cingulum fibres ending in the retrosplenial area; c, fibres of the perforant or direct temporoammonic path; d, perforant fibres of the cingulum; e, plane of dorsal perforant path fibres; g, subicular cell; h, pyramidal cells in field CA1 (regio superior) of Ammon’s horn; i, ascending (Schaffer) collaterals of large pyramidal cells; r, collaterals of alvear fibres (from Cajal, 1995).

…the only opinion that is in harmony with the facts (is) that nerve cells are independent elements which are never anastomosed, with by means of their protoplasmic expansions (dendrites) or by the branches of their prolongations of Deiters (axons), and that the propagation of nervous action is made by contacts at the level of certain apparatuses of dispositions of engagement.

HISTORY OF THE SYNAPSE

21

In 1892 Cajal gave his description of the nerve networks of the hippocampus using the Golgi silver technique, six years after Golgi himself had described the network (Fig. 1.9C). Cajal’s work could not be more different (compare Fig. 1.9D with 1.9C in which the sections through the hippocampus are oriented in the same way, with the fascia dentata uppermost). Here neurons are clearly circumscribed according to the neuron doctrine that each neuron is a separate cell, rather than in continuity with other cells through a neurofibrillar network that joins them all together at the site of the end bulbs. The neurofibrils were intracellular to the neuron according to Cajal (Fig. 1.10E), and did not extend out of the terminal boutons into the neurons on which they abutted (Fig. 1.10G). With Golgi and many other neurohistologists these neurofibrils were both intracellular and intercellular, joining the terminal boutons to the neuron on which they ended, and so giving rise to a reticulum of neurofibrils that gave continuity to the entire neural network in places such as the hippocampus (Fig. 1.9C). On technical grounds it is not easy to say unequivocally that Cajal had objective proof for his doctrine. This was particularly the case when considering the speculations which he and other neurohistologists engaged in concerning the role of the neurofibrils in the conduction of the action potential. For example Cajal comments that (Cajal, 1995): The existence of a conductive pathway in the cytoplasm was also postulated on the basis of observations made with the Nissl method… So that the reader may judge the extent to which the discovery of the neurofibrillar network justifies these assumptions, we shall reproduce a drawing [Fig. 1.10F published long ago and based on the work of Bethe]. Neurofibrils are the sole conductors of neuronal activity. They form bundles in the dendrites and axon, and course between the Nissl bodies as they cross the perikaryon on their way from one process to another without anastomosing among themselves. Long neurofibrils, most of which converge on the axon, are not the only type found in the cell. Cajal went much further in this conjecturing, for which there was not a scintilla of physiological evidence, that the arrangement of the neurofibrils within a single neuron is such that the conduction of the action potential could only occur from dendrites to soma to axon (Fig. 1.10F). This then led him to place arrows of action potential flow on so many of his drawings summarising the results of silver staining of a particular block, such as that of the hippocampus (Fig. 1.9D). The fact that these have ended up being approximately correct does not mean that Cajal should be credited for their discovery, which would be for profoundly wrong reasons, namely based on the conduction of action potentials by neurofibrils. It might be commented on in passing that Cajal’s work was not subject to review until towards the end of his life, as it was published privately. In summary then the neuron doctrine could not be considered to be definitively supported by this silver stain work. An entirely different conclusion may be reached when considering Cajal’s work on degeneration of neural centres which does give definitive support to the neuron doctrine. Again the major evidence was provided by the relationship between terminals on Purkinje cells in the cerebellum and the state of the terminals when the Purkinje cells degenerate (Cajal, 1995): As an example of a convincing and well known case let us mention the disappearance of the Purkinje cells in general paralysis with maintenance of the basket and stellate cells of the molecular layer. This persistence, revealing the independence of the baskets and the cells they surround, can also be produced experimentally by sectioning the axons of the Purkinje cells at the level of the granular layer or even below as is shown in Fig. 1.10B (compare with the normal in Fig. 1.10A). This remarkable conservation of the baskets, despite the disappearance of the cells in connection with them…’ ‘The basket cells resist pathological influences much more that do the Purkinje cells. The baskets in young traumatised animals appear nearly normal, even at the level of regions where the cells have disappeared (Fig. 1.10B(D)).’ ‘Baskets of the Purkinje cells…are perfectly formed in animals from twenty to thirty days old, while they are constantly altered in the vicinity of wounds, although they never react so actively and energetically as the axons of the Purkinje cells. The lesions most commonly found are as follows: (a) Baskets whose descending appendices have terminal balls. As can be seen in Fig. 1.10C(A), the Purkinje cells have been resorbed, and the descending branches of the baskets, notably thickened and intensely stained, end in a terminal ball or in a series of clubs. One may entirely agree with Cajal in his comment that (Cajal, 1995): The neuron doctrine is compatible with the well-documented phenomenon of secondary degeneration in neural centers. In fact, if neurons were not completely independent, it would be impossible to account for the precise localisation of degeneration following ablation of cell groups or fiber tracts. Cajal’s speculations concerning the physiological role of neurofibrils had led him to his polarisation of the neuron doctrine, namely that action potential flow was only from terminal bulb to dendrite (or sometimes soma), and then from soma to axon. However it was a physiologist, namely Sherrington, who supplied the experimental findings which showed that the region of contact between the end bulb and neuron might only allow the direction of action potential transmission in one direction through the end bulb.

22

EARLY HISTORY OF THE SYNAPSE

Fig. 1.10. Evidence at the end of the 19th century that nerve terminals are not in continuity with the cells on which they impinge. (A) The endings of cerebellar basket cells in the albino mouse viewed using the Golgi method. A, Purkinje cell stained with osmic acid; B, basket cell, a, b, pericellular axonal ramifications forming baskets; C, axon (from Cajal, 1995). (B) The endings of cerebellar basket cells in a cat (twenty-five days old) twenty-four hours after axotomy of the Purkinje cells, which destroy them. A, almost normal Purkinje cell; B, Purkinje cell undergoing atrophy and granular in appearance; D, baskets surrounding the empty spaces previously occupied by Purkinje cells that are now destroyed (from Cajal, 1959). (C) The endings of cerebellar basket cells in a rabbit (two months old) thirty hours after axotomy of the Purkinje cells. A, terminal clubs; B, molecular layer; a, transversal fibres of this zone (from Cajal, 1959). (D) Motor endplates on skeletal muscle fibres during reinnervation of the muscle by the motor nerve. Different stages of restoration are shown. A, nerve; B, fibre that gives rise to several plates; E, H, plates as yet without ramifications; C, F, plates with a well developed arborization (from Cajal, 1959). (E) The concept of a neurofibrillary network within an individual neuron, that was taken to conduct impulses by Cajal and others. Shown is a giant pyramidal cell (10-day-old dog). A, is a cell with a pericellular plexus. A, axon; B, summit of the axon hillock; F, dendritic branch with a single neurofibril (from Cajal, 1995). (F) Diagram due to Cajal showing the presumed direction of conduction of impulses along the neurofibrillary network with a cortical pyramidal cell. A, axon; B, nucleus; a, channels or pathways for the neuronal currents of the impulses; b, Nissl bifurcation cones; c, nuclear hood; d, recurrent path of currents in a process bordering the axon; e, elongated Nissl body. Arrows indicate the direction of current flow (from Cajal, 1995). (G) Terminal boutons surrounding a funicular neuron in the spinal cord (from Cajal, 1995).

1.11 Sherrington: the adoption of the word `synapse' Sherrington (1858–1952; Fig. 1.8C), as a consequence of his work on spinal reflexes in the 1890s, had reached the conclusion that transmission of the action potential across the end bulbs of sensory nerve terminals to neurons in the spinal cord involved different principles to that of the conduction of the action potential along nerve fibres (see Liddell, 1960).

HISTORY OF THE SYNAPSE

23

The nerve centre exhibits a valve like function, allowing conduction to occur through it in one direction only. How securely the circuits of the nervous system are valved against regurgitation is shown by the Bell-Magendie law of the reactions of the spinal nerve roots. It was work on spinal reflexes that had led to the idea of the ‘valve like function’ of the region of apposition between the end bulbs and the neuron, and so distinguished this region from the rest of the nerve fibre. This is reminiscent of Descartes model of the nervous mechanism of reciprocal inhibition of voluntary movement which was based on the idea of a flow of animal spirits in hollow nerve fibres; in this case valves could differentially alter the flow through anastomoses between nerves to antagonistic muscles, e.g. lateral and medial rectus muscles of the eye (see Fig. 1.3B). As Michael Foster said in 1901 (Foster, 1897): If we judge Descartes from the severe standpoint of exact anatomical knowledge, we are bound to confess that he, to a large extent, introduced a fantastic and unreal anatomy in order to give clearness and point to his exposition… If we substitute in place of the subtle fluid of the animal spirits, the molecular changes which we call a nervous impulse, if we replace his system of tubes with their valvular arrangements by the present system of concatenated neurons… Descartes’ exposition will not appear so wholly different from the one which we give today. How far then did the work on spinal reflexes in the late nineteenth century allow for a new principle to be enunciated concerning the operation of the region where end bulbs impinged on neurons, different from the speculations of the seventeenth century? Sherrington’s emphasis on some kind of discontinuity at the region of apposition between end bulb and neuron mostly rests on the results obtained from degeneration studies. He says (Sherrington to Schafer, 1897a; Sherrington to Schafer, 1897b): The evidence of Wallerian secondary degeneration is clear in showing that that process observes strictly a boundary between neurone and neurone in the reflex arc. The characteristics distinguishing reflex-arc conduction from nervetrunk conduction may therefore be largely due to inter-cellular barriers, delicate transverse membranes. He goes on to comment that: the characteristics distinguishing reflex arc conduction from nerve-trunk conduction may therefore be largely due to intercellular barriers, delicate transverse membranes… If the conductive element of the neurone be fluid and if at the nexus between neurone and neurone there does not exist any actual confluence, there must be a surface of separation. Even should a membrane visible to the microscope not appear, the mere fact of non-confluence of the one with the other implies the existence of a surface of separation. Such a surface might restrain diffusion, bank up osmotic pressure, restrict the movement of ions, accumulate electric charges, support a double electric layer, alter in shape and surface tension with changes in difference of potential, alter in difference of potential with changes in surface tension and in shape, or intervene as a membrane between dilute solutions of electrolytes of different concentration or colloidal suspensions with different sign of charge. It would be a mechanism where nervous conduction, especially if predominantly physical in nature might have grafted upon it characteristics just such as those differentiating reflex—arc conduction from nerve—trunk conduction. For instance, change from reversibility of direction of conduction to irreversibility might be referable to the membrane possessing irreciprocal permeability. In Foster’s textbook of 1897 he goes on to comment on the nervous impulse ‘sweeping along’ the axon of one neuron until it is (Foster, 1897): brought to bear through the terminal arborisation on the dendrites of another neuron where the lack of continuity between the material of the arborisation of the one cell and that of the dendrite (or body) of the other cell offers an opportunity for some change in the nature of the nervous impulse as it passes from one cell to the other. There is no doubt that the results of Wallerian degeneration pointed to the likelihood of end bulbs possessing membranes as did the rest of their parent nerve fibre. The problem then presented itself of how such a membrane might relate to the membrane of the underlying neuron membrane on which the end bulb impinged. It was to this region then that the irreversibility of nerve transmission must be ascribed and an explanation sought. It was the histological work on Wallerian degeneration together with the physiological discovery of the irreversibility of transmission that indicated the special nature of this region. It was clear that this region deserved a name that might focus the attention of experimenters and so help delineate its properties. In Foster’s textbook of 1897 Sherrington provided the name (Foster, 1897):

24

EARLY HISTORY OF THE SYNAPSE

So far as our present knowledge goes, we are led to think that the tip of a twig of the arborescence is not continuous with but merely in contact with the substance of the dendrite or cell-body on which it impinges. Such a special connection of one nerve cell with another might be called a ‘synapses’. The origins of this use of the word ‘synapsis’ or as it became ‘synapse’ can be found in letters Sherrington wrote to his colleagues Sharpy-Schafer in 1897 and Fulton in 1937, in reply to enquiries concerning the derivation of the word. To the former he wrote [see Sherrington to Schafer, 1897a; Sherrington to Schafer, 1897b; Sherrington to Fulton, 1937): As to nomenclature—its sole object is I take it clearness combined with brevity… Definition is wanting when a penny has to pass for 5 and 3 farthings as well as for 4: the one symbol is then too little… All I think we ought to be careful not to do is ‘commit barbarisms,’ e.g., impossible adjectival form, using prefixes and affixes with false signification or in impossible ways—that simply adds new terms which like other ‘monsters’ can’t live long, and may be misleading during life, does all of us harm as giving the impression of carelessness or ignorance’. ‘As to ‘junction’ I feel we are less easily reconcilable. If a Latin form capere not jungere should be the root. The mere fact that junction implies passive union is alone enough to ruin the term… I think it does not want the gift of prophecy to foretell that it [the word junction] must become more and more obviously inapplicable as research progresses. Synapse, which implies a catching on, as e.g., by one wrestler of another—is really much closer to the mark. But I am not a bit wedded to the word: if you could suggest an English word containing the notion which is not already overburdened with applications. I have been trying to find one but cannot. Conjunction is even worse than junction. Sherrington wrote to Fulton that (Sherrington to Fulton, 1937): You enquire about the introduction of the term ‘synapse’; it happened thus. Michael Foster had asked me to get on with the Nervous System part (Part III) of a new edition of his ‘Textbook of Physiology’ for him. I had begun it, and had not got far with it before I felt the need of some name to call the junction between nerve-cell and nerve-cell (because that place of junction now entered physiology as carrying functional importance). I wrote him of my difficulty, and my wish to introduce a specific name. I suggested using ‘syndesm’ ( ). He consulted his Trinity friend Verrall, the Euripidean scholar, about it, and Verrall suggested ‘synapse’ (from ), and as that yields a better adjectival form, it was adopted for the book. The concept at root of the need for a specific term was that, as was becoming clear, conduction which transmitted the impulse along the nerve fibre could not—as such— obtain at the junction, a membrane there lay across the path, and conduction per se was not competent to negotiate a cross-wise membrane. At least so it seemed to me then, perhaps A.V.Hill and Gasser and Bishop could tell us differently today. I do not know when the term ‘synapsis’ was introduced for a phase of the karyokinetic process. Neither Foster nor I knew of it in that connection. I fancy Salvin Moore, a cytologist, put it forward. He once told me he had not known the term was in use in physiology. I think that your proposed synaptic knobs would be very clear and helpful. Pace Verrall’s memory (Verrall was a delightful and charming man). ‘Synapsis’ strictly means a process of contact, that is, a proceeding or act of contact, rather than a thing which enables contact, that is, an instrument of contact. ‘Syndesm’ would not have had the defect, that is, it would have meant a ‘bond’. The credit for the word ‘synapse’ then goes to a classical scholar at Cambridge (for a detailed outline of this claim, see Tansey, 1997). Although the word ‘junction’ was abandoned by Sherrington as appropriate to describe the functional relationship between the end bulb and neuron it was preserved for that between motor nerve and muscle at the endplate. Here Wallerian degeneration had also indicated the discreteness of the nerve terminal from the muscle at the endplate, suggesting that neither Boerhaave nor Kuhn were any more correct than Golgi in ascribing continuity between the end of nerve terminals and the cells on which they impinged. The research of Tello, working in Cajal’s laboratory, was particularly persuasive on this issue, as it showed the postjunctional endplate apparatus was intact in frog muscle after denervation, and that reinnervating nerve fibres could be found at different stages of terminal formation on these regions of the muscle (Fig. 1.9D; see Cajal, 1928). Sherrington’s prescient comment that (Sherrington to Fulton, 1937): …conduction per se was not competent to negotiate a cross-wise membrane was followed by the caveat that At least so it seemed to me then, perhaps A.V. Hill and Gasser and Bishop could tell us differently today. Sherrington’s claim that conduction could not per se negotiate the synapse was soon challenged by K.Lucas and later his colleague E.D.Adrian. They produced credible biophysical explanations of how conduction per se could negotiate a cross-

HISTORY OF THE SYNAPSE

25

wise membrane. Their theory showed how even the process of inhibition at synapses could function without the necessity of evoking any new principles other than those involved in the conduction of the action potential. Cajal’s use of arrows showing the direction of action potential conduction on his drawings of silver-stained neurons seem to independently support Sherrington’s notions, with conduction only possible in one direction across the synapse. But of course Cajal’s arrows are placed according to an erroneous idea of conduction by neurofibrils. Too much importance has been placed in the history of neuroscience on Cajal’s arrows and Sherrington’s introduction of the work ‘synapse’. These researchers made great contributions to the delineation of the types of neurons to be found together with their spatial relationships on the one hand and to that of the excitatory and inhibitory processes that these neurons participate in on the other. However one must turn to the research of other investigators in order to find the observations which warrant the use of the words ‘junctions’ and ‘synapses’. Such research was supplied in the twentieth century, and is the subject of a more contemporary history (see Chapters 2 and 3).

2 Emergence of the Concept of Transmitter Release at Peripheral and Central Synapses

2.1 Research on the Synapse in the Laboratories of Sherrington and Langley before the Great War Over ninety years ago Charles Sherrington gave his Silliman Lectures at Yale University which were later published as The Integrative Action of the Nervous System (Sherrington, 1906). In that great work Sherrington laid the conceptual foundations for much that was to dominate research on the central nervous system for the rest of the century. Sherrington had begun his studies on the central nervous system at Cambridge in the physiological laboratory of John Langley and later at Liverpool. In his Silliman Lectures Sherrington pointed out that: In view, therefore, of the probable importance physiologically of this mode of nexus between neurone and neurone it is convenient to have a term for it. The term introduced has been ‘synapse.’ Sherrington had already defined the synapsis in Foster’s Textbook of Physiology some ten years earlier (Sherrington, 1897). He went on to say in the Silliman Lectures that: The neurone itself is visibly a continuum from end to end, but continuity, as said above fails to be demonstrable where neurone meets neurone—at the synapse. There a different kind of transmission may occur. The delay in the gray matter may be referable, therefore, to the transmission at the synapse. Regarding how synapses operate, he said: It would be a mechanism where nervous conduction, especially if predominantly physical in nature, might have grafted upon its characters just such as those differentiating reflex-arc conduction from nerve-trunk conduction. Sherrington had developed these ideas as a consequence of his studies on the reflex contractions of muscles following stimulation of muscle and skin afferents. His summary diagram of the place of excitation and inhibition in reflex pathways for flexor activation and extensor inhibition, shown in Fig. 2.1, is a masterpiece of fruitful speculation. This diagram not only indicates the concept of excitatory and inhibitory synapses, developed clearly by 1908, but draws the experimentalist into Sherrington’s line of inquiry as to what other nervous pathways may be delineated by this approach: in particular, how is the information transferred at the nerve terminal across the synaptic gap in the drawing, and what is the mechanism of inhibition. A research program entirely different from Sherrington’s was directed by his mentor J.N.Langley at Cambridge. In 1903 Langley, who had introduced Sherrington to neurophysiology (Langley & Sherrington, 1884), was at that time laying the foundations for our understanding of the chemical nature of transmission at synapses. In 1901 Langley published a remarkable paper (Langley, 1901) showing that stimulation of the sympathetic component of the autonomic nervous system, which Gaskell and he had already defined, resulted in changes in the effectors that in many cases could be mimicked by application of suprarenal extract (adrenaline). In Langley’s words: I have formerly divided the autonomic nervous system into sympathetic, cranial, sacral and enteric. It is a noteworthy fact that the effect of supra-renal extract in no case corresponds to that which is produced by stimulation in normal conditions of a cranial autonomic or of a sacral autonomic nerve…. It is equally noteworthy that the effects produced by supra-renal extract are almost all such as are produced by stimulation of some one or other sympathetic nerve. In many cases the effects produced by the extract and by electrical stimulation of the sympathetic nerve correspond exactly (see Fig. 2.2A)…. It is hardly possible to avoid the conclusion that in these cases the extract acts directly on the unstriated muscle, and if this is so, it is probable that in all cases the action is direct. The theory of direct action cannot, however,

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

27

Fig. 2.1. Sherrington's 1906 diagram `indicating connections and actions of two afferent spinal root-cells (dorsal root ganglia), and ' in regard to their reflex influence on the extensor and flexor muscles of the two knees. Flexor and extensor muscles of the knee joint on the right (R) and left (L) sides are shown together with the inputs to the spinal cord by a cutaneous afferent ( ) and a muscle spindle afferent ( ). The reflex pathways postulated show flexor (F) excitation (+) and extensor (E) inhibition (−) ipsilaterally and flexor (F) inhibition (−) and extensor (E) excitation (+) contralaterally. In Sherrington’s 1906 words, ‘the sign +indicates that at the synapse which it marks the afferent fibre ( (and ( ) excites the motor neurone to discharge activity, whereas the sign —indicates that at the synapse which it marks the afferent fibre (and ) inhibits the discharging activity of the motor neurones. The effect of strychnine and of tetanus toxin is to convert the minus sign into a plus sign’.

be regarded as more than provisional until it is shown experimentally that the inhibitory action of suprarenal extract on certain unstriated muscle, and its stimulating action on salivary gland cells take place in the absence of nerve-endings. These points I propose to consider in a later paper. These experiments were carried out by Langley’s student T.R.Elliott who concluded in a note to The Journal of Physiology (owned and edited by Langley) in 1904 that: Adrenalin might then be the chemical stimulant liberated on each occasion when the impulse arrives at the periphery (Elliott, 1904a). The idea of chemical transmission at the synapse, and indeed of receptors on the effector organ for receiving the chemical substance released by the nerves, was already a central part of the research program in Cambridge physiology under Langley (Langley, 1906). This research was furthered in 1906 by W.E.Dixon. Working in the Cambridge physiology laboratory on the effect of suprarenal extracts on the lung (Brodie & Dixon, 1904) Dixon decided to perform an experiment in which he took an extract of a dog’s heart that had undergone vagal stimulation and applied it to the exposed heart of a frog, obtaining an interruption of the heart beat (Fig. 2.2B). This work was similar in design to Otto Loewi’s famous 1921 experiment (see Fig. 2.3A) some 14 years later, establishing the idea of chemical transmission in the heart unequivocally (Loewi, 1921; Dale, 1934). Henry Dale had observed these experiments of Dixon’s. Dale came across acetylcholine accidentally in 1914, as a constituent of a particular sample of ergot. Here he describes his finding: I was led to make a detailed study of its action. This, I think, gave the first hint that acetylcholine might have an interest for physiology. Then I was struck by the remarkable fidelity with which it reproduced the various effects of parasympathetic nerves, inhibitor on some organs and augmentor on others—a fidelity which I compared to that with which adrenaline reproduces the effects of the other, true sympathetic, division of the autonomic system (Dale, 1914b) At the beginning of the century Sherrington had already placed both excitatory and inhibitory synapses at center stage in the integrative behaviour of the spinal cord. Furthermore, Langley’s school had shown that synapses at the autonomic neuroeffector junctions were likely to operate by the secretion of a chemical substance, which in the case of the sympathetic was related to adrenaline. Research over the next 25 years attempted to unravel the principles of operation of synapses within a conceptual framework that originated in the great research schools formed by Langley and Sherrington.

28

HISTORY OF THE SYNAPSE

Fig. 2.2. The evolution of the idea of chemical transmission at synapses. (A) Langley’s 1901 table showing the effect of suprarenal extract (adrenaline) in the cat and rabbit arranged roughly in the order of the amount of extract required per body weight to produce an obvious effect. Langley notes that: ‘I have formerly divided the autonomic nervous system into sympathetic, cranial, sacral and enteric. It is a noteworthy fact that the effect of supra-renal extract in no case corresponds to that which is produced by stimulation in normal conditions of a cranial autonomic or of a sacral autonomie nerve. It is equally noteworthy that the effects produced by supra-renal extract are almost all such as are produced by stimulation of some one or other sympathetic nerve. It is hardly possible to avoid the conclusion that in these cases the extract acts directly on the unstriated muscle, and if this is so, it is probable that in all cases the action is direct. (B) Unpublished record from a 1906 experiment by W.E.Dixon showing the beat of the exposed heart of a frog. At the first mark, extract from a dog’s heart that had been inhibited by vagal stimulation was applied; at the second mark, atropine was applied (from Dale, 1934b). This is the first known record of an attempt to determine if a nerve secretes a substance that, when placed on another organ, will mimic the effects of nerve stimulation to that organ.

2.2 Sherrington's concept of the inhibitory and excitatory states of central synapses John Eccles was born in 1903 and he entered Melbourne University Medical School at the very young age of 15, in the year that saw the end of the Great War, and later won a Rhodes Scholarship to Oxford in 1925 to work with Sherrington. Eccles entered an intellectual environment on synaptic transmission that was now dominated by Loewi’s recent experiments (Loewi, 1921) indicating that chemical transmission occurred between the vagus and the heart, and Dale’s work indicating a role for

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

29

acetylcholine in synaptic transmission (Dale, 1914; Dale, 1910; Loewi, 1935) (Fig. 2.3A). Eccles entered Oxford the year J.N.Langley died at the age 73 after completing a six-hour experiment in Cambridge. He arrived at a time when Sherrington was engaged in research with Liddell on the characteristics of the myotatic reflex (Liddell & Sherrington, 1924; Liddell & Sherrington, 1925) and with Creed on the flexion reflex (Creed & Sherrington, 1926). Sherrington had just produced a masterly summary of work on inhibition, in which he concluded that: In relation to inhibition at the synapse that it might be mediated by an agent, moreover, one whose existence lies outside the intrinsic properties of pure nerve-fibre and with a, so to say, more chemical mode of origin and function than the nerve impulse per se (Sherrington, 1925). This comment was made only four years after the experiments of Loewi. Eccles joined Creed in his first experimental work, which was on the subject destined to dominate his research life for over 40 years: the mechanism of inhibitory synaptic transmission (Creed & Eccles, 1928). Then in 1929 he joined Sherrington in a technical improvement of the torsion myograph (Eccles & Sherrington, 1930a) in preparation for a collaboration (that lasted but a few years, 1929–1931) concerned with research on the flexion reflex and inhibition (Eccles & Sherrington, 1930b; Eccles & Sherrington, 1931), These experiments were to see the last flowering of Sherrington’s scientific genius at the age of 75. The work on the ipsilateral spinal flexion reflex introduced Eccles to the technique of stimulating first with a just threshold conditioning volley, then at later intervals with a subsequent test volley in order to tease out the time course of the central excitatory state (Eccles & Sherrington, 1930a) (Fig. 2.4A). This approach, when applied to the mechanism of monosynaptic transmission in the spinal cord, gave a very precise measure of the time course of the central excitatory state or, as we now know, the excitatory postsynaptic potential (Fig. 2.4). With hindsight it might be expected that chemical transmission would seem to be the most likely mechanism for determining the central excitatory state (c.e.s.) and the central inhibitory state (c.i.s.), based on the experiments of Langley and his school, along with those of Loewi and Dale. This was certainly not the case, as is shown in the next section. 2.3 Lucas, Adrian and the electrical concept of the inhibitory state of central synapses Towards the end of Langley’s career, the Cambridge School of Physiology came to be dominated by those such as Keith Lucas and E.D.Adrian who were introducing electrophysiological techniques into the study of how impulses conduct in excitable tissue. Lucas had published a Physiology Monograph entitled The Conduction of the Nervous Impulse in which he argued that central inhibition might be brought about by the interference of high-frequency discharges in the nerves as they approach their synaptic connections on neurones (Lucas, 1917). In this way the refractory state of the axon following an impulse could operate to produce inhibition. This idea was followed up in detail in 1924 by Adrian (Fig. 2.5A), who was skeptical about the recent research of Loewi and Dale, commenting: It appears, then that the fluid coming from the stimulated organ reproduces the characteristic effects of the vagus or sympathetic on different tissues, though whether every detail of the nervous effect is copied by the fluid remains an open question. The nature of the ‘vagus substance’ is uncertain. If these results can come to be generally accepted we shall have a new and extremely interesting picture of the action of the autonomic system. The difficulty is that the effects seem to be capricious and are not easily reproduced. Some observers have failed to satisfy themselves that they occur at all outside the margins of the experimental error. In view of this uncertainty we can only wait until there is a more general agreement as to the experimental basis of the humoral theory (Adrian, 1924). In expanding on the Lucas theory of the electrical basis of spinal cord inhibition, Adrian went on to say that: If it (the humoral theory) is correct, the explanation of peripheral inhibition resolves Itself into that of (a) the secretory mechanism which produces the inhibiting substance whenever impulses pass along certain fibres to the muscle, and (b) the way in which an inhibiting substance, adrenalin for instance, exerts its effect on the muscle fibre… If we compare the present theory (the electrical theory of Figure 2.5A), or some modification of it, with the view which supposes that inhibition is due to the production of a special substance which blocks the excitatory paths to the motor neurone, it will be seen that there is actually not very much difference between them. If an inhibiting substance is produced, its production must be almost instantaneous and it must disappear very rapidly; what the present theory assumes is that the ‘substance’ is to be identified with the refractory state and that sustained inhibition is due to a series of refractory states and not to a steady production of an inhibiting substance (Adrian, 1924).

30

HISTORY OF THE SYNAPSE

Fig. 2.3. The first apparently unequivocal demonstration of chemical transmission at a synapse and characterisation of the accompanying synaptic potential. This was performed for the vagal inhibition of the heartbeat. (A) Otto Loewi’s 1921 original record in which at 1 the heartbeat is shown in normal ringer; at 2, the decline in the heartbeat is due to the addition of a ringer that had been in contact with another heart whose vagus had been stimulated for 15 minutes; at 3, the heartbeat is normally in the presence of a ringer from another heart in which the vagus had not been stimulated; finally, in 4, atropine was added in a normal ringer, increasing the heartbeat. (B) Curve showing the extent of inhibition of the heartbeat (on the vertical axis) due to a single stimulus in an experiment performed by Brown and Eccles (1934). A single stimulus is applied to the vagus nerve and the lengthening of each cardiac cycle (i.e., the amount by which it exceeds a normal cycle) is expressed as a fraction of the normal cycle (of about 305 msec) on the ordinate; the abscissa gives the interval between the vagal stimulus and the end of that particular cycle. There is a latent period of rather more than 100 msec before a volley in the postganglionic fibres produces an inhibitory effect on the pacemaker. (C) Intracellular records of the hyperpolarisation in the arrested frog’s heart due to a single vagal volley by del Castillo and Katz (1957). Note that the latency between the vagal volley and the hyperpolarisation is several hundred msec. Calibration is (small vertical bar) 1mV and indicates the moment of stimulation, which occurs 400 msec before the hyperpolarisation commences.

Adrian arrived at this conclusion as a consequence of experiments performed with Bronk on the frequency of discharges in motoneurones accompanying reflex and voluntary contractions. It was not until 1929 that Adrian felt able to abandon the idea that central inhibition could be explained along the lines suggested by Lucas (i.e., by the depressant effects produced by highfrequency impulse discharges) (Adrian & Bronk, 1929). Adrian carried great authority on matters concerned with electrical activity in nerves at the time Eccles arrived at Oxford in 1925. He had just successfully recorded for the first time the trains of

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

31

Fig. 2.4. The time courses of the central excitatory states (c.e.s.) and central inhibitory states (c.i.s.) compared with those of the excitatory and inhibitory postsynaptic potentials. A, time course of the central excitatory state (c.e.s.) determined using muscle reflexes. It shows the reflex response of the tibialis anticus muscle to two stimuli (each one of which alone is just sub-threshold) to medial gastrocnemius nerve and lateral gastrocnemius nerve at various intervals, as measured by Eccles and Sherrington (1930). The abscissa gives the time, approximately in msec, and the ordinate the tension in grams. Right of zero shows the interval by which the stimulus to the lateral gastrocnemius nerve is leading; left of zero shows the interval by which stimulus to the medial gastrocnemius nerve is leading to. Each point plotted shows the tension developed at the indicated interval between stimuli. The excitatory state, set up by the conditioning volley, lasts for about 15 msec and is called the ‘central excitatory state’ (c.e.s.). B, time course of the central excitatory state (c.e.s.) and central inhibitory state (c.i.s.) determined by measuring the compound action potential in a muscle nerve, a, the extent to which the response to a test reflex is enhanced by a prior conditioning reflex (facilitation) at the different intervals given in the abscissa, as measured by Lloyd (1946). The facilitation of the biceps reflex by afferent volleys in the semitendinosus nerve and of one head of the gastrocnemius by afferent volleys in nerves to the other head are given. Conditioning volleys of near reflex threshold strength were used. The relative facilitation, expressed in percent maximum, is plotted as a function of time and gives the c.e.s.; this is similar to that given by the method used by Eccles and Sherrington in A. b, the extent to which the response to a test reflex is inhibited by a prior conditioning reflex at the different intervals given in the abscissa, as measured by Lloyd (1946). The inhibition of the reflex of tibialis anterior by weak volleys to the gastrocnemius afferent nerve are given. The ordinate gives the degree of inhibition, in percent of maximum, to the time interval between volleys on the abscissa. This relative inhibition, expressed in percent maximum, is plotted as a function of time and gives the c.i.s. C, intracellular potentials recorded by Brock et al. (1952) from a biceps semitendinosus motoneuron due to two afferent volleys in the biceps semitendinosus nerve; note that the dorsal root spikes accompanying each volley are shown beneath the intracellular recordings. These excitatory postsynaptic potentials give the electrical signs of the central excitatory state measured by Eccles and Sherrington in 1930 and shown in A; note the similar time course. D, intracellular potentials recorded by Brock et al. (1952) from a biceps semitendinosus motoneuron due to a single volley in a quadriceps nerve, of increasing size downward; note the increasing size of the dorsal root spikes shown beneath the successive intracellular recordings. These inhibitory postsynaptic potentials give the electrical signs of the central inhibitory state shown in B, b; note the similar time course.

nerve impulses traveling in single sensory and motor nerve fibers which, according to A.L.Hodgkin, ‘marks a turning point in the history of physiology’ (Hodgkin, 1978).

32

HISTORY OF THE SYNAPSE

2.4 Loewi, Dale and Eccles examine the inhibitory state at autonomic neuromuscular junctions The chemical or electrical nature of synaptic transmission in either the peripheral or the central nervous system was an entirely unsettled issue when Eccles arrived at Oxford. It seemed likely, despite Adrian’s skepticism, that vagal inhibition of the heartbeat was chemical and that the ‘vagus substance’ or ‘Vagusstoff’ was acetylcholine. However, it must be remembered that Otto Loewi and Henry Dale did not win the Nobel Prize for their investigations until 1936 and Loewi still felt constrained to defend this idea as late as 1935. Eccles, after completing his last work with Sherrington in 1931 on spinal cord inhibition, sought to delineate the characteristics of this chemical inhibition; he used the only appropriate preparation available at the time, namely the vagus to the heart. Together with G.L.Brown he determined that the time course of the inhibitory state (analogous to the c.i.s.) following a single stimulus to the vagus nerve was of the order of a second or more and did not arise for over 100 msec after a stimulus (Brown & Eccles, 1934) (Fig. 2.3B). This determination was supported 20 years later with the introduction of intracellular recording of the inhibitory junction potential in the heart by del Castillo and Katz (del Castillo & Katz, 1957) (Fig. 2.3C). Brown and Eccles commented that: If the vagal volley is set up late in a cardiac cycle, that cardiac cycle is not inhibited, the latent period of the inhibition being usually 100 to 160 ms. Of this amount the conduction time to the region of the pacemaker probably only accounts for about 10 ms, ie. the greater part of the latent period appears to occur after the arrival of the inhibitory impulses at the nerve fibres of the pacemaker. It is probable that most of this time is occupied in the liberation of the acetylcholine substance and its diffusion to the point of its action (Brown & Eccles, 1934). This slow time course of the only reasonably well established chemical synapse seemed to set the temporal characteristics of chemical transmission. This was particularly so for other synapses at which nerve terminals were shown to secrete acetylcholine. 2.5 Eccles develops the electrical concept of the excitatory state at autonomic synapses Eccles then turned his attention to the only readily accessible synapse on neurons for which acetylcholine was known to be released on nerve stimulation, namely that in sympathetic ganglia (Feldberg & Gaddum, 1934; Eccles, 1937) (Fig. 2.6A). It was only natural that he should first approach the problem of defining the excitatory state in the ganglion by the same methods developed to study the time course and other characteristics of the c.e.s. of motoneurones in the spinal cord (Fig. 2.4A). Examination was made of the interaction of submaximial volleys to each of two preganglionic inputs to a ganglion delivered at different intervals apart in test-conditioning pairs, and the ganglionic action potential measured (Fig. 2.6B(a)). In this case the time course of compound action potentials was being determined rather than the time of reflex contractions of muscles used to determine the c.e.s. of motoneurons (Fig. 2.4A). Compound action potential waveforms had to be subtracted in the manner indicated in the legend to Fig. 2.6B(a) before the c.e.s. could be estimated; this figure shows that the test compound action potential seems to be elevated compared with the compound action potential in the absence of a conditioning impulse. This elevation occurs from the earliest times for the test-conditioning interval, then declines to zero at an interval of 4.5 msec, increases again, and reaches a peak at 17.6 msec from which it slowly declines over the next 40 msec or so to zero. The time course of these events is indicated in Figure 2.6B(b) which shows the early fast phase of the potentiation of the test compound action potential followed by the later developing slow component. The fast phase, termed the ‘detonator response’ (Eccles, 1937) was not affected by anticholinesterases as is vagal inhibition of the heart; furthermore, it is much faster than the action of acetylcholine on the heart (Fig. 2.3B). The detonator response was not then attributed to the action of acetylcholine but rather to the action currents in the preganglionic nerve terminals that produced a potential response in the ganglion cells; the later phase was identified as the excitatory state. This analysis led to the proposition that: On present evidence however, it seems that the action-current hypothesis offers a more probable explanation for direct synaptic transmission, the acetylcholine liberated in sympathetic ganglia possibly having a secondary excitatory action as already suggested (Eccles, 1936). The analysis of synaptic transmission from the postganglionic nerves to the smooth muscle of the nictitating membrane also revealed an excitatory state that had an initial fast component followed by a slower late component (Monnier & Bacq, 1935; Eccles & Magladery, 1936; Eccles & Magladery, 1937). In this case the anti-adrenaline drug 933F (piperidonethyl-3benzodioxane) blocked the slow electrical response and the associated contraction but not the fast response, suggesting that the latter might also indicate the signs of electrical transmission (Eccles, 1936; Eccles & Magladery, 1936). Similar

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

33

Fig. 2.5. Electrical theories of central inhibition. (A) Scheme due to Adrian (1924) based on an idea of Keith Lucas (1917; see his Fig. 22) in which inhibition occurs as a consequence of the refractory state left in a nerve following an impulse. In this diagram, A is the excitatory pathway and B the inhibitory pathway converging on the motoneuron M; the shaded areas conduct with a decrement. By making the decremental path from B to M longer than that from A, we can account for the fact that an impulse from B can never succeed in reaching M, whereas impulses from A can do so provided the path is given time for complete recovery between each impulse. With such an arrangement, an impulse from B, though not itself exciting M, would leave the common pathway in a refractory state, absolute or relative, which would hinder the passage of impulses from A for a time depending on the rate of recovery of the path and the extent of the decrement in it. A rapid succession of impulses from B would produce continuous inhibition by never giving time for the complete recovery of the common pathway. If the impulses from B also pass by a more direct route, dotted in the figure, to the antagonistic motoneuron M , the periods of inhibition of M would synchronize with the discharge of motor impulses from M . (B) Scheme due to Brooks and Eccles (1947) of how electrical inhibition could occur in the spinal cord. The diagram indicates current flow at a schematic synapse of a Golgi cell G on a motoneuron M according to this electrical theory of inhibition. E shows the excitatory line to M, and I the inhibitory line that subliminally excites G and so generates the current flow producing an electrotonic focus on M. According to this theory, an intracellular electrode would record a brief positively going electrical field at X.

difficulties were arising with the parasympathetic innervation of the bladder. In this case only the slow phase of contraction could be blocked by atropine and mimicked by applied acetylcholine; this left the fast phase to be accounted for in terms of an electrical component of transmission (Henderson & Roepke, 1934). In order to escape this awkward fact, Dale and Gaddum (1930) suggested that the concentration of adrenaline secreted at the sympathetic postganglionic nerve terminal was very high, giving the fast response, and was therefore insensitive to 933F; the slow reaction was attributed to the escape of adrenaline and its secondary diffusion from the synaptic cleft to act on other cells that were sensitive to 933F. As late as 1937 Dale was reiterating that:

34

HISTORY OF THE SYNAPSE

Fig. 2.6. Claims for electrical and chemical transmission at synapses. (A) Demonstration of how stimulation of the cervical sympathetic causes liberation from the ganglion of a substance pharmacologically identified by Feldberg and Gaddum (1934) as acetylcholine. The top panels show the response of the frog’s beating heart and the bottom panels that of contraction of leech muscle as a consequence of adding Ringer’s fluid collected from a ganglion during preganglionic stimulation (A); adding Ringer’s fluid with different concentrations of acetylcholine (B and C), and adding Ringer’s fluid from an unstimulated ganglion (D). Note that the fluid from the stimulated ganglion has identical effects on slowing the beating heart and contracting the leech muscle, as does acetylcholine. (B) Electrophysiological evidence involving the study of action potentials interpreted by Eccles (1937) as showing an electrical component to synaptic transmission. In (a), single submaximal stimuli were applied to two different preganglionic nerve branches at various intervals apart; the continuous lines show the ganglionic action potentials produced by a second stimulus at the indicated intervals and the broken lines show the action potential set up by the second stimulus alone. It will be noted that when the two volleys are simultaneous, spatial facilitation is maximal (i.e., it is effective in producing a discharge from the largest number of ganglion cells); this number is still large at the interval of 2.3 msec. At 4.5 msec, effective summation occurs in very few, if any, ganglion cells. At intervals longer than 4.5 msec, spatial and temporal facilitation again develop and the time course of decay of this second facilitation wave is due to the excitatory state of the ganglion cells, which would now be called the excitatory postsynaptic potential. In (b) the time course of the ‘detonator response’ is shown and the excitatory state derived from experiments such as those in (a); the former is attributed to electrical transmission at the synapse and the latter to chemical transmission using acetylcholine to give the synaptic potential (from Eccles, 1936). (C) Electrophysiological evidence involving the direct study of synaptic potentials in the absence of action potentials interpreted by Eccles (1943) as showing that only chemical transmission occurs in ganglia, (a) shows a single synaptic potential, recorded with extracellular electrodes, which lasts for about 100 msec. The faster curve is a theoretical estimate of the time course of transmitter action that gives rise to the synaptic potential, based on the ‘local potential’ theory of A.V.Hill. Ordinates for transmitter actions are in arbitrary units, (b) shows the theoretical times for the decline in the amount of transmitter remaining within a sphere of either radius 2 µm (B) or 1 µm (D), as well as a cylinder of radius 2 µm (A) or 1 µm (C) following the transmitter’s instantaneous deposition (from Ogston, 1955). The curves show that the concentration declines with a similar time course to that of the time course of transmitter action given in (a), suggesting that free diffusion of acetylcholine out of a synaptic region with the dimensions of 1 µm or 2 µm could account for the observed results. (D) First synaptic potentials recorded with an intracellular electrode from a mammalian sympathetic ganglion (from R.Eccles, 1955).

This antagonist (atropine) cannot similarly intervene, when acetylcholine is liberated front nerve endings in immediate contact with, or even inside the cell membrane (Dale, 1937).

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

35

Dale thought that the fast phase of contraction of smooth muscle on nerve stimulation was due to the high transmitter concentration reached at the nerve endings, with the slow contraction attributed to its later diffusion to other muscle cells (Dale, 1937). He had often used different renditions of this argument to escape the unpalatable fact that neither adrenaline nor acetylcholine seemed suitable candidates for transmission to some smooth muscles. For example, in 1934 Dale said that: There are some parasympathetic effects, such as the action of the vagus on the intestine, and the vaso-dilator action of parasympathetic nerves in general, which are resistant to atropine, though the otherwise similar actions of injecting or applying acetylcholine are readily abolished by it… Gaddum and I suggested that in such cases the nerve impulses liberate acetylcholine so close to the reactive structures that atropine cannot intervene, whereas it can prevent its access to them when it is artificially applied from without (Dale, 1934). Following the discovery of inhibitory junction potentials in smooth muscle in 1963 (Bennett et al., 1963) that are mediated by nonadrenergic noncholinergic (NANQ) synapses (Bennett et al., 1966) we now know of many synapses that do not conform to Dale’s paradigm. Eccles developed his ideas on electrical transmission in Oxford between 1934 and 1937. He then returned to Australia in 1937 to become Director of the Kanematsu Memorial Institute of Pathology at Sydney Hospital. His first studies there were concerned with the possibility of electrical transmission at the somatic neuromuscular junction. In 1936 G.L.Brown had shown that close interarterial injection of acetylcholine into the cat’s gastrocnemius gave repetitive impulse firing and contraction of the muscle (Fig. 2.7B). Protection of the metabolism of acetylcholine with eserine allowed it to appear in a perfusate after stimulation of motor nerves (Dale et al., 1936; Brown et al., 1936). By 1938 both Eccles and O’Connor (1938) as well as Gopfert and Schaefer (1938) had recorded the endplate potential with extracellular electrodes in curarized muscles (Fig. 2.7C). This was probably the first time a synaptic potential had been observed without distortion due to electrotonic conduction. The possibility that the endplate potential was due to the release of acetylcholine was not grasped. In 1939 Eccles and O’Connor used the method of applying conditioning-test volleys to mammalian motor nerves and recording the impulses in a muscle in order to determine the time course of the excitatory state. They concluded that a nerve impulse exerts two excitatory actions at the neuromuscular junction: (1) Newborn muscle impulses are set up by a brief excitatory action probably no more than 1 msec in duration and analogous to the detonator action described for synaptic transmission. (2) The much more prolonged end-plate potential is set up independently of the newborn impulses, but if the growth of these impulses is sufficiently delayed, it appears to aid in their growth to the fully propagated size. It is analogous to the N wave and the associated central excitatory state of synaptic transmission, and analogous responses have also been described at the neuromuscular junction of smooth muscle (Eccles & O’Connor, 1939). Thus the test-conditioning volley approach used so successfully to determine the time course of the c.e.s. for motoneurons now led to the erroneous conclusion that a very fast response, much faster than the endplate potential recorded at the neuromuscular junction in curarized preparations, was the primary means of transmission. 2.6 Katz, Kuffler and Eccles establish the motor endplate as the paradigm synapse for electrophysiology J.N.Langley and Keith Lucas initiated the studies of A.V.Hill on the biophysics of nerve and muscle at Cambridge in 1909, and Hill in turn supervised the first research of Bernard Katz on electrical excitation and conduction of the nerve impulse at University College London in 1935 (Hill et al., 1936a,b; Katz, 1978). The frog isolated gastrocnemius sciatic nerve preparation was used by Katz in his investigations of Hill’s theory of excitation in order to provide an index for the duration of maintained nerve excitation (Katz, 1936). However it was the Cambridge zoologist Carl Pantin, who had acted as a guide to A.L.Hodgkin’s studies, that was responsible for the first explicit research by Katz on neuromuscular transmission (Katz, 1936). In this work it was shown that magnesium ions could block neuromuscular transmission in crabs. By 1939 Katz was using the frog isolated sartorius-nerve preparation after curarization (Katz, 1939) to confirm the work of Gopfert and Schaefer that showed: A small non-conducted potential change is to be found in the myoneural region, which reaches a maximum 4 msec after arrival of the nerve impulse, and then falls at a slow rate, similar to the electrotonic potential (Gopfert & Schaefer, 1938).

36

HISTORY OF THE SYNAPSE

It was natural that Katz should have wanted to continue research on this ‘small non-conducted potential change’ or endplate potential, using the frog sartorius-nerve preparation, when he joined Eccles and Stephen Kuffler at the Sydney Kanematsu Institute in 1939 (Fig. 2.7A). Rather than working on the innervation zones of the cat’s soleus muscle with Eccles, Katz did the following: I ganged up with Stephen Kuffler, and I was very pleased when we succeeded in getting hold of some nice Australian tree frogs whose sartorius muscles proved to be very suitable for the experiments we wanted to do, and this kept me busy and moderately happy for two years (Katz 1986). The work of Katz, Kuffler, and Eccles in Sydney (Eccles & Kuffler, 1941; Eccles et al., 1942) marks the beginning of a new era in synaptic physiology after the one begun 50 years earlier by Langley and Sherrington. It is characterised by the use of progressively more sophisticated electrical techniques to probe the mechanism of synaptic transmission. The first experiments of Eccles, Katz, and Kuffler did not involve test-conditioning volleys to estimate (the time course of the excitatory state but rather concentrated on the properties of the extracellular signs of the endplate potential made subthreshold by a suitable dose of curare (Figure 2.7D). They showed, using an analysis provided by A.V.Hill (1933), that the time course of the underlying transmitter action lasted for only a few msec. As Eccles, Katz, and Kuffler stated: Thus it seems that most of the declining phase of the e.p.p. is a passive decay of a negative membrane charge after the depolarizing agent has ceased to act. The earlier suggestion, therefore, that the decline of the endplate potential follows the time course of a passively decaying electrotonic potential is confirmed (Eccles et al., 1941). They stated further: By making plausible assumptions it is shown that the observed curare and eserine actions are reconcilable with the hypothesis that acetylcholine is responsible for all the local potential changes set up by nerve impulses (Kuffler, 1942). Eccles then abandoned the hypothesis of an early ‘detonator’ electrical response followed by a slower endplate potential due to the secretion of acetylcholine. In the following year, Eccles used an analysis similar to that applied to the neuromuscular junction when he evaluated extracellular recordings of the synaptic potential in curarised sympathetic ganglia. This led to the conclusion that (Fig. 2.6C(a)): The results conform well with the postulate of a single depolarizing agent…it is concluded that most and possibly all of the evidence for the detonator action may now be attributed to the brief transmitter action (Eccles, 1943). The time course of transmitter action following a single impulse here could be shown to conform to free diffusion of acetylcholine from the synaptic cleft (Ogston, 1955) rather than to the hydrolysis of acetylcholine by cholinesterase (Fig. 2.6C (b)). Katz returned to A.V.Hill’s laboratory in 1946. Using the frog sartorius muscle nerve preparation once more, and as a result of the introduction of the microelectrode in 1949 by Ling and Gerard (Ling & Gerard, 1949), he was able to confirm with Fatt that the endplate potential alone initiated the muscle action potential (Fatt & Katz, 1951); (compare Fig. 2.7D with Fig. 2.7E). This idea was soon generalised for the nervous system when Rosemary Eccles (Fig. 2.6D) showed that the synaptic potential in sympathetic ganglia alone initiated the action potential (Eccles, 1955). 2.7 Eccles elucidates the electrical signs of the inhibitory and excitatory states of central synapses The introduction of the microelectrode also allowed for the first time an investigation of whether the c.e.s. and c.i.s. of a motoneuron could be described in terms of synaptic potentials, and also whether these synaptic potentials were likely to be due to a ‘detonator’ electrical effect or the secretion of a transmitter. Lloyd (Lloyd, 1946) had already utilized the testconditioning volley method described 16 years earlier by Eccles and Sherrington (1930) to determine this c.e.s. He used the stretch reflex and the more accurate method of electrical recording from the muscle nerves rather than muscle contraction. Lloyd’s results for the time course of the c.e.s. of motoneurons were similar to those of Eccles and Sherrington (compare Fig. 2.4B(a) with Fig. 2.4A); he also gave the time course of the c.i.s. using this method (Fig. 2.4B(b)). Eccles still thought it possible that electrical transmission might account for the c.i.s. of motoneurons; he provided a Golgi cell model of this as late as 1947 (Brooks & Eccles, 1947) (Fig. 2.5B), even though he had abandoned the idea of electrical transmission in the peripheral nervous system. It must be remembered that at this stage the chemical transmitters acting on motoneurons were

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

37

Fig. 2.7. The electrical and chemical analysis of synaptic transmission at the neuromuscular junction. (A) S.W.Kuffler, J.C.Eccles and B.Katz in 1941 at the time of their experiments on elucidating the nature of the endplate potential. They are shown in Martin Place, Sydney, walking from the Kanematsu Memorial Institute of Pathology at Sydney Hospital to catch a tram to the University of Sydney to give a lecture in the Physiology Department (B) First recordings of the electrical and mechanical responses of the cat’s gastrocnemius muscle to close intra-arterial injections of two different concentrations of acetylcholine, taken by G.L.Brown and communicated to J.C.Eccles (1936). (C) First endplate potentials to be recorded with extracellular electrodes in curarized (a) cat soleus muscle (Eccles & O’Connor, 1938) and (b) frog sartorius muscle (Gopfert & Schaefer, 1938). (D) Endplate potential recorded with an extracellular electrode in a curarised frog sartorius muscle by Eccles et al. (1941); the numbers on the records refer to the distance in mm from the pelvic end of the muscle. (E) Tracings of endplate potentials recorded with an intracellular electrode from the frog sartorius by Fatt & Katz (1951); the numbers refer to different distances in mm of the recording electrode from the endplate (compare with D above).

unknown, research having shown that neither acetylcholine nor adrenaline were likely to be secreted at inhibitory or excitatory synapses. The first intracellular recordings of synaptic potentials in motoneurons were awaited with considerable interest. Eccles had recently left Dunedin and taken up the foundation Chair of Physiology at the John Curtin School of Medical Research in Canberra. While there, he, Brock, and Coombs published their classic paper ‘The recording of potentials from motoneurones with an intracellular electrode’ in 1952. As in the peripheral nervous system, the c.e.s. and the c.i.s. were

38

HISTORY OF THE SYNAPSE

Fig. 2.8. The discovery of transmitter quanta and synaptic vesicles. (A) Spontaneous miniature endplate potentials in frog sartorius muscle treated with 10−6 prostigmine bromide (Fatt & Katz, 1952). (B) Distribution of amplitudes of spontaneous miniature endplate potentials from an endplate in the frog sartorius muscle treated with prostigmine; there are 800 miniature potentials in the sample (Fatt & Katz, 1952a). (C) Histogram showing distribution of amplitudes of spontaneous miniature potentials and endplate responses at a low calcium endplate in the frog sartorius muscle. The lower part shows a continuous curve that has been calculated according to the hypothesis that the responses are built up statistically of units whose mean size and amplitude distribution are identical to those of the spontaneous potentials. The expected number of failures are shown by the arrows. Abscissae: scale units mean amplitude of spontaneous potentials (0.875 mV) (del Castillo & Katz, 1954a). (D) Electron micrograph of a section of an area of complex axonal entanglement from the neuropile of the earthworm. Numerous profiles of axonal membranes can be distinguished, varying in density from place to place. Mitochondria (M) and endoplasmic reticulum are distinguishable in several places. Y denotes an area of specialized axonal contact identified as synaptic in nature. Numerous synaptic vesicles (SV) are seen in the presynaptic neuron (PRSN), whose profile is of irregular outline. The profile of the postsynaptic member (PSN) is identified as a section through a finger-like axonal projection indenting the presynaptic axon, producing puckerings or folds (FO) in the presynaptic axonal membrane. Enlarged 25 000x. (de Robertis & Bennett, 1955). (E) Electron micrograph of catecholamine-containing granules in synaptic vesicles of sympathetic nerve terminals in the rat vas deferens. The distance between nerve and muscle membranes is about 25 nm (Richardson, 1962; 79). (F) Vesicle hypothesis as first enunciated (del Castillo and Katz, 1956). A diagram is shown of a nerve-muscle junction, with several features described after Robertson’s (1956) electron micrographs of the junction. N, nerve terminal; M, muscle fiber. In the lower part, an enlarged part of the nerve terminal is shown, containing ‘ACh-carrier corpuscles,’ as then described by del Castillo and Katz, or synaptic vesicles. Release of ACh is supposed to occur as a result of critical collisions between these synaptic vesicles and the membrane. This is indicated formally by labelling certain ‘critical spots’ on the surface of both.

identified with monophasic synaptic potentials, one in the depolarising direction (Fig. 2.4C) and the other in the hyperpolarising direction (Fig. 2.4D). Furthermore, these potentials had time courses similar to those predicted for the c.e.s. and the c.i.s. in the original work of Eccles and Sherrington in 1930 (compare Fig. 2.4C with Fig. 2.4A) and with that of Lloyd in 1946 (compare Fig. 2.4D with Fig. 2.4B(b)). On observing the inhibitory postsynaptic potential, Eccles concluded that:

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

39

…the potential change observed is directly opposite to that predicted by the Golgi-cell hypothesis, which is thereby falsified… It may therefore be concluded that inhibitory synaptic action is mediated by a specific transmitter substance that is liberated from the inhibitory synaptic knobs and causes an increase in polarization of the subjacent membrane of the motoneurone (Brock et al., 1952). So concluded the long saga that established the supremacy of chemical transmission at peripheral and central synapses. It enabled Eccles finally to show that Sherrington’s c.e.s. and the c.i.s. of motoneurons were due to chemical transmission (Fig. 2.10A). In hindsight, Keith Lucas and E.D.Adrian, the founding fathers of the biophysical approach to the study of impulse conduction, had inappropriately tried to transfer knowledge gained on conduction to the theoretical analysis of the mechanisms of transmission. Langley, Loewi, and Dale had exerted a counteracting influence that was founded on a tremendous amount of ingenious experimentation on transmission. However, it was the analysis of the endplate potential offered by the biophysicists A.V.Hill and Bernard Katz that finally clarified the whole matter. 2.8 Katz's concept of quantal transmitter release at the motor endplate and the vesicle hypothesis The discovery of spontaneous miniature endplate potentials in the frog sartorius muscle by Fatt and Katz revolutionised our understanding of neurotransmitter release (Fatt & Katz, 1952a) (Fig. 2.8A). Although the endplate potential in response to a nerve impulse had been identified in 1938, it was not possible to observe spontaneous endplate potentials without intracellular microelectrodes. The amplitude-frequency distribution of the spontaneous potentials was approximately Gaussian (Fig. 2.8B), although Fatt and Katz did note that: …there is indication of several discharges of about twice the mean amplitude, and of one isolated discharge of three or four times the mean size (Fatt & Katz, 1952a), They attributed this to the coincidence of two (or three) unitary discharges that could not be resolved given that detection was only possible down to 5 msec; the calculated chances of units occurring at such small intervals apart supported their conclusion. This question concerning the composition of the unitary discharges is still a matter of great interest. It was natural to consider if the endplate potential was composed of these spontaneous unitary discharges. Del Castillo and Katz showed that this was likely to be the case. They determined that the amplitude-frequency distribution of the endplate potential under conditions of low transmitter release could be built of units whose mean size and amplitude distribution were identical to those of the spontaneous unitary discharges (Fig. 2.8C). They concluded that: statistical analysis indicates that the end-plate potential is built up of small all-or-none quanta which are identical in size and shape with the spontaneous occurring miniature potentials (del Castillo & Katz, 1954). Del Castillo and Katz noted that the statistical analysis failed under conditions of reasonably high-evoked transmitter release, which they thought may occur because some synaptic units respond more readily than others. With the application of more refined electrophysiological techniques it is now known that there is indeed nonuniformity in the probability of secretion of quantal unit at different release sites within a nerve terminal. Attempts to both measure this non-uniformity and see if the Katz statistical paradigm for quantal secretion holds for different peripheral and central synapses constitute a major research effort at this time. The use of the microelectrode to study neurotransmitter release was complemented by the development of refined biochemical techniques for determining the constituents of nerve terminals, as well as by introduction of the electronmicroscope. The major biochemical contributions came from the Karolinska Institute in Stockholm, where U.S. von Euler, who had trained with H.H.Dale in London in 1934, first showed definitively in 1946 that the catecholamine noradrenaline was a transmitter at sympathetic nerve terminals (von Euler, 1946). In a letter to Dale in 1945, von Fuler commented that: perhaps it will interest you to hear about the sympathetomimetic substance in the spleen which I have been working with lately. It appeared that ordinary alcoholic extracts of cattle spleen contain the somewhat surprising amount of some 10 mg adrenaline pressor equivalent per kg. After perfusion the active substance was found to differ somewhat from adrenaline, and, on the basis of your admirable analysis with Berger in 1910 of the action of sympathomimetic amines, it emerged that it resembled definitely more an amino-base like nor-adrenaline than adrenaline or methylated compounds (see Blaschko, 1985).

40

HISTORY OF THE SYNAPSE

Fig. 2.9. The calcium dependence of neurotransmitter release. (A) Evidence that calcium is required for the release of acetylcholine (ACh) at sympathetic preganglionic nerve terminals. Both contraction of the cat’s nictitating membrane (above) and the cat’s blood pressure (below) were used to assay acetylcholine in the venous effluent collected during perfusion of the superior cervical ganglion during the following corresponding periods. A–C, Perfusion with normal Locke’s solution containing eserine: A, no stimulation; B, maximal preganglionic stimulation, 10 per sec; C, at the arrow, injection of 2 mg KCI. D–F, perfusion with Ca-free Locke’s solution containing eserine: D, no stimulation; E, 10 min later, maximal preganglionic stimulation, 10 per sec, producing no further contraction of the nictitating membrane; F, at the arrow, injection of 2 mg KCL G, time signal, and effect of 0.005 µg of Ach (Harvey & Macintosh, 1940). (B) Relationship between calcium concentration and amplitude of endplate potential in the frog sartorius muscle. Each symbol gives the results for a different magnesium concentration (open circles, 0.5 mm; crosses, 2.0 mM; filled circles, 4.0 mM). The coordinates are logarithmic, giving straight lines with a slope of approximately 4 (Dodge & Rahamimoff, 1967). (C) Suppression of transmitter release during a large ‘positive voltage step’ of the presynaptic membrane potential of the giant synapse in the stellate ganglion of the squid, treated with tetrodotoxin to block nerve impulses. The presynaptic terminal is loaded with tetraethylammonium ions. Blocks 1 to 8 show increasing pulse intensity. In each block of records, the bottom trace shows the presynaptic voltage step, the middle trace shows postsynaptic response, and the top trace monitors the current pulse. There is a progressive suppression of ‘on’-response and replacement by ‘off’ response, as presynaptic voltage is increased from 100 to 200 mV (records 4 to 8), indicating that, if the movement of calcium ions into the terminal is blocked by depolarization to 200 mV, then transmitter release fails to occur until after the depolarization is removed (Katz & Miledi, 1967b).

In 1953, Hillarp, Lagerstedt, and Nilsson in Lund, Sweden (Hillarp et al., 1953), as well as Blaschko and Welch in Oxford, showed that catecholamines such as noradrenaline were stored in granules within the adrenal medulla. It was not long before von Euler together with Hillarp (von Euler & Hillarp, 1956) showed that noradrenaline was also stored in the particulate fraction of sympathetic nerves. Noradrenaline was therefore likely to be stored in granules within sympathetic nerve terminals.

EMERGENCE OF THE CONCEPT OF TRANSMITTER RELEASE AT PERIPHERAL AND CENTRAL SYNAPSES

41

This concept of the packaging of neurotransmitters in particles or granules within nerve terminals was greatly enhanced by the first electron microscope images of the terminals by Palade (1954) as well as de Robertis and Bennett (1954) (Fig. 2.8D). De Robertis and Bennett commented in 1955 that: A granular or vesicular component, here designated the synaptic vesicles, is encountered on the presynaptic side of the synapse and consists of numerous oval or spherical bodies 20 to 50 nm in diameter, with dense circumferences and lighter centers. Synaptic vesicles are encountered in close relationship to the synaptic membranes (de Robertis & Bennett, 1955). At sympathetic nerve terminals these synaptic vesicles were seen to actually contain the granules of catecholamines (Fig. 2.8E), as predicted by von Euler and Hillarp. By 1956 the quantal unit of transmitter release had been discovered at the motor endplate, along with the storage of catecholamines in granules within sympathetic nerves, and synaptic vesicles within both central and peripheral nerve terminals. That year del Castillo and Katz enunciated the vesicle hypothesis, attributing quantization of transmitter release to its association with synaptic vesicles (Fig. 2.8F). In their own words: Recent electron microscope studies (Robertson, 1956) have shown that the motor nerve terminals contain a fairly dense population of microsomes, granules or vesicles, of less than 0.1 µ diameter, which may well be the intracellular corpuscles to which ACh is attached. It has been known since Loewi’s investigations that most of the ACh which is present in ‘homogenised’ nerve tissue can only be extracted into an aqueous solution after chemical destruction of the cell proteins. It appears then that the discharge of ACh from a nerve terminal requires the disruption of more than one diffusion barrier: first the release from its intracellular attachment, and secondly a passage through a nerve membrane. One might suppose that when a ‘critical’ collision occurs between an infracellular ACh-carrier and the membrane of the nerve terminal, the two barriers are opened simultaneously and the ACh-contents of the carrier particle are suddenly discharged. This picture, though purely speculative, is nevertheless in accord with recent experimental findings; it takes account of the evidence discussed below that the release of ACh from nerve terminals occurs in multi-molecular units or ‘quanta’ and of the evidence, already cited, for the bound state of intracellular ACh content (del Castillo & Katz, 1956). The mechanism by which the ‘contents of the carrier particle are suddenly discharged’ is perhaps the major focus of research on neurotransmitter release at the present time. Katz next tackled the problem of determining the necessary and sufficient conditions for the ACh-contents of the carrier particle to be suddenly discharged. It had been known since the work of Locke reported in 1894 that calcium was necessary for transmission at the neuromuscular junction (Locke, 1894). The reason for this, as Harvey and Macintosh showed in 1940, was that the release of acetylcholine at nerve terminals required calcium (Harvey & Macintosh, 1940).This transmitter was only released in the perfused superior cervical ganglion of the cat upon stimulating the preganglionic nerves in the presence of calcium (Fig. 2.9A). Kuffler and Eccles had shown by 1942 that the amplitude of the endplate potential was affected by the calcium concentration in accord with the observations of Harvey and Macintosh on the calcium-dependence of acetylcholine release. However, it was not until 1967 that Katz’s laboratory produced two most important observations on how calcium might govern transmitter release. The first was due to Dodge and Rahamimoff, who showed that the endplate potential in low concentrations of calcium ions increased as the fourth power of the calcium concentration (Dodge & Rahamimoff, 1967) (Fig. 2.9B). This observation gave rise to the idea that the cooperative action of about four calcium ions is necessary for release of each quantal packet of transmitter by the nerve impulse. The second observation was due to Katz and Miledi, who determined that the site of this cooperative action of calcium ions was on the inside of the nerve terminal rather than on the outside (Katz & Miledi, 1967b) (Fig. 2.9C). The basis of this cooperative action of calcium ions on the inside of the nerve terminal membrane to release the contents of synaptic vesicles is a theme of much current research. 2.9 Conclusion: the establishment of Sherrington's concept of the synapse in the central nervous system and central synaptic transmission While Katz was researching the mechanism of neurotransmitter release, Eccles and his colleagues in Canberra explored synaptic mechanisms at successively higher levels of the central nervous system. Studies were carried out on inhibition of Purkinje cells in the cerebellum (Eccles et al., 1966) (Fig. 2.10D), on thalamic-cortical relay cells (Eccles, 1969) (Fig. 2.10B) and on the CA3 pyramidal cells in the hippocampus (Andersen et al., 1963) (Fig. 2.10C). This work was revolutionary in as much as it supplied a functional microanatomy of the synaptic connections to be found in these different nerve centers. In addition, it

42

HISTORY OF THE SYNAPSE

Fig. 2.10. Inhibitory pathways in the central nervous system elucidated by Eccles (1969) and colleagues. (A) Synaptic connection involved in the inhibitory action produced in extensor motoneurons by afferent volleys in the Group 1b afferent fibers from the Golgi tendon organs in extensor muscles (Eccles, 1969). (B) Pathway to the sensory-motor cortex for cutaneous fibres from the forelimb; inhibitory neurons are shown in black in both the cuneate nucleus and ventrobasal nucleus of the thalamus. Note that the inhibitory pathway is of the feed-forward type in the cuneate nucleus and feed-back type in the thalamus (Eccles, 1969). (C) Results of recording from a CA3 hippocampal pyramidal cell in response to commissural (Com), septal (Sept), and local (Loc) stimulation, a, responses recorded by a microelectrode penetrating CA3 following local stimulation, b, graph in which the size of the positive waves of the responses to commissural, septal, and local stimulation is plotted against depth, with the positivities measured at a time indicated by the stippled line in (a) c, a CA3 pyramidal cell, semi-diagrammatically drawn to scale to facilitate comparison with (b). The arrows indicate the extracellular flow of current generated by the inhibitory postsynaptic potential (from Andersen et al., 1963). (D) Perspective drawing of a cerebellar folium to show the synaptic connections of the inhibitory interneurons. The cerebellar cortex is seen to be divided into three layers: molecular layer (ML) Purkinie cell layer (PL), and granular layer (GL). The input to the cortex is by two types of fibre: mossy fiber (MF) and climbing fiber (CF). Single examples are shown of four types of interneurons: granule cells (GrC), Golgi cells (GoC), basket cells, and outer stellate cells (SC). Also shown are two Purkinje cells, one (PC) with its dendritic ramifications, and both axons (PA), one with two collaterals (PAC) ending on the Golgi cell and the basket cell. The mossy fibre shown with numerous branches and thickenings at the sites of its synapses on granule cell dendrites, so forming the glomeruli (Glo). Collaterals of the climbing fiber (CF) are shown making synapses on the Golgi cell and basket cell. The axons of the granule cells bifurcate to give rise to the parallel fibers (PF) in the molecular layer. Arrows show directions of normal propagation in: the mossy fiber, climbing fiber, and its collaterals; the Purkinje axons and collaterals; and the axons of the interneurons BC, SC, and GoC. (Eccles et al., 1966).

provided insights into the ionic basis of inhibition at different synapses and the role this inhibition plays in controlling the excitability of the major neuron types. In this way the research program initiated by Sherrington’s concepts of inhibition and excitation in 1906 was brought to fruition.

3 The Discovery of Acetylcholine and the Concept of Receptors at Synapses

3.1 Introduction The idea of the ‘receptive substance’ or receptors as we now call them, was developed by John Langley of Cambridge 90 years ago (Fig. 3.1 A). Between 1901 and 1905 Langley laid the foundations for the idea of chemical transmission with his student Thomas Elliott (Fig 3 1B) through their investigations on sympathetic neuroeffector transmission. In an extraordinary act of creative ability, Langley then carried out a series of

Fig. 3.1. The founding fathers of chemical transmission at synapses. (A) J.N.Langley (1852–1925), Fig 20.4 in Finger (1994). (B) T.R.Elliott (1877–1961), portrait facing p.53 in Dale (1961) (C) H.H.Dale (1875–1968), portrait facing p.77 in Feldberg (1970). (D) O.Loewi (1873–1961), Fig. 20.5 in Finger (1994).

44

HISTORY OF THE SYNAPSE

Table 1. Chronological table of significant events in the history of receptors 1844 1866 1899 1901 1901 1904 1905 1905 1906 1906 1914 1921 1926 1929 1934 1936 1936 1970 1976

Curare paralyses rabbits without affecting the heart Curare acts on an intermediate zone between nerve and muscle Supra-renal extract (adrenaline) contracts and relaxes different smooth muscles Supra-renal extract (adrenaline) contracts or relaxes different smooth muscles as does stimulation of their sympathetic nerve supply Nicotine stimulates sympathetic ganglion cells directly Adrenaline acts at the junction between nerves and smooth muscle cells not on nerve terminals Nicotine stimulates skeletal muscles directly and this is blocked by curare The concept of a ‘receptive substance’ on skeletal muscles first described The ‘receptive substance’ shown to provide the receptor for alkaloids such as nicotine and curare Acetylcholine synthesised and shown to have powerful effects on the circulation Acetylcholine has similar actions on smooth muscles and cardiac muscle as stimulating the vagus nerve ‘Vagusstoff’ is released by the vagus nerve and controls the heart beat Physostigmine potentiates the effects of applied acetylcholine on the heart Acetylcholine shown to be a natural constituent of horse and ox spleen; likely to be ‘Vagusstoff’ Acetylcholine is released in autonomic ganglia on nerve stimulation; likely to be the transmitter Acetylcholine collected in venous fluid from skeletal muscles on nerve stimulation Acetylcholine injected into the arteries of skeletal muscles initiates contraction Acetylcholine applied at the endplate gives membrane noise due to the opening of channels Acetylcholine receptor channels give electrical signal that may be recorded directly

Bernard Vulpian, 1866 Lewandowsky, 1899 Langley, 1901 Langley, 1901 Elliott, 1904 Langley, 1905 Langley, 1905 Langley, 1906 Hunt and Taveau, 1906 Dale, 1914a Loewi, 1921 Loewi and Navratil, 1926b Dale and Dudley, 1929 Feldberg and Gaddum, 1934 Dale, Feldberg and Vogt, 1936 Brown, Dale and Feldberg, 1936 Katz and Miledi, 1970b Neher and Sakmann, 1976

investigations between 1905 and 1907 on the somatic neuromuscular junction that established the idea of transmitter receptors. This historical review traces the development of Langley’s ideas over this period, especially in relation to the concept of the ‘chemoreceptor’ developed by Paul Ehrlich. The review then examines how this work was applied by a number of investigators to place the concept of transmitter substances and their receptors on a firm foundation for the modern molecular approaches to the delineation of receptor types and their function. In order to assist the reader, a chronological table of significant experiments in the history of receptors is provided (Table 1), together with a list of the major contributors to these experiments (Table 2) and the agents they used to delineate the receptor concept (Table 3). 3.2 Claude Bernard and curarization: the notion of an intermediate zone between nerve and muscle In June of 1844 Claude Bernard wrote in his experimental note book that: A poisoned arrow obtained from a friend who had connections with South American natives was thrust into the subcutaneous tissue of a rabbit at the internal part of the thigh and maintained there for 30 seconds. The animal was then observed. At first, nothing happened. But after six minutes it became totally paralysed: no reflex movements were observed on pinching the rabbit, although the heart continued to beat. The animal subsequently died and at autopsy it was not possible to find any lesion capable of explaining paralysis and death (Fessard, 1967). Table 2. Scientists who contributed significantly to the idea of Name

Location

Dates of research

C.Bernard J.Langley T.Elliott H.Dale O.Loewi W.Feldberg G.Brown

Paris Cambridge Cambridge London Austria London London

1844–1883 1874–1908 1904–1905 1914–1936 1921–1926 1934–1936 1936–1937

Mentor M.Forster J.Langley

H.Dale H.Dale

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

45

Table 3. Definition of some of the agents used to delineate the receptor concept Agent

Site of action

Muscle type

Curare Pilocarpine Supra-renal extract Adrenaline Nicotine Acetylcholine Vagusstoff Eserine Acetylcholine

Nicotinic receptors Muscarinic receptors

Skeletal Smooth and cardiac Smooth and cardiac Smooth and cardiac Skeletal Smooth and cardiac Cardiac

Nicotinic receptors Muscarinic receptors Muscarinic receptors Cholinesterase Nicotinic receptors

Skeletal muscle

Antagonist Atropine

Curare Atropine

Curare

Although this observation had been made in 1811 by Brodie, who later went on to show that curarized animals could be maintained alive on an artificial respirator, the advent of Bernard into this area of research brought a keen experimental mind to bear on the problem of the action of curare. Bernard constructed a galvanic stimulator for exciting either nerves or muscle fibres which allowed him to carry out investigations on curare that led him to report that: Electrical stimulation of a motor nerve in a curarized frog has no effect, whereas its muscles contract when directly stimulated. This pointed to curare acting on nerves rather than generally acting as some kind of anaesthetic. In order to test whether both the motor and sensory nerves were affected by curare, Bernard designed an experiment in which a ligature was passed around the waist of a frog, so that the lower limbs were isolated from the rest of the body, except for the sciatic nerve, as shown in Fig. 3.2A. Bernard then reported the following observations: Curare is introduced under the skin of the back. It poisons the anterior part of the body and prevents movement there; but sensation in this part is conserved, for stimuli applied to this paralysed portion cause energetic reflex movements in the isolated posterior half. Curare is thus a poison which not only produces physiological separation of nerves and muscles, but also separation of two major kinds of nervous manifestations. It suppresses movement but has no action on sensation; so that in a way it dissects out the neuromotor system and separates it from the muscular system, the sensory nervous system, and other tissues. He then designed an experiment that is illustrated in Fig. 3.2B: here electrical stimulation was applied to the sciatic nerve lying in a bath of curare, as shown in V, and contraction of the muscle outside the curare bath was present. On the other hand, when the muscle was placed in the curare bath as shown in V, stimulation of the nerve outside the bath did not give rise to contraction. The obvious conclusion to this experiment would seem to be that some junctional structure between the nerve and the muscle had been affected by curare. However Bernard did not reach this conclusion. Fessard (1967) has conjectured as to why Bernard did not follow the appropriate deduction from his observations. Perhaps Bernard was concerned that he was dealing with organs that were separated from the body and not subjected to the ‘milieu interieur’ and circulation of the blood? It seems that Bernard was persuaded of the idea that the action of curare should be related to the circulation of the blood, perhaps as Bernard himself suggested through an alteration of the gas exchange between the blood and the air in the lungs or the tissues of the capillaries. Bernard then turned to the experiment illustrated in Fig. 3.2C. The technique involves a kind of close arterial injection, as illustrated by the insert. Curare is injected into the artery supplying a muscle, so that it does not come into contact with the nerve trunk; furthermore there is an outlet in the vein which prevents the curare containing blood from reaching the central nervous system. This beautiful experiment seemed to Bernard to show that curare acted on the nerve terminals within the muscle. There is no mention in his books of the notion that curare might work at a junction formed between the nerve and the muscle although in his notebooks there is mention that ‘curare must act on the terminal plates of motor nerves’ and that ‘Curare does no more than interrupt something motor which puts the nerve and the muscle into electrical relationship for movement’ (Fessard, 1967). These quotes suggest that he had envisaged the notion of a neuromuscular junction, although this was never pursued in his formal statements to be found in his books concerning these experiments. The explicit claim that curare does not act on motor-nerve terminals, but rather on some intermediate zone between nerve and muscle was left to Vulpian (1866) in his Lecons sur la Physiologie Generale et Comparee du Systeme Nerveux. The nature of this intermediate zone was next investigated by histologists, seeking to find the site at which curare works. Chief amongst these at this time

46

HISTORY OF THE SYNAPSE

was Bernard’s student Kuhne, however it is now clear in retrospect that the histological approach to this problem had to await the advent of ultrastructural techniques, nearly a century later. As to the proper development of the functional approach, that might have been pursued immediately. Although there were no technical limitations to such an approach, another forty years had to pass before the problem was elucidated by Langley. 3.3 Paul Ehrlich and the idea of the `receptive side chains' of cells In 1885 Ehrlich presented his thesis to the University of Leipzig in which he described for the first time his ‘side chain theory’ of cellular action. The protoplasm of a cell was considered to be a giant molecule incorporating a central structure responsible for the specific activity of a particular cell type (such as a muscle cell or a neurone), which possessed chemical side chains. The side chains were envisaged as carrying out processes common to all cells. For example one such side chain might be involved in the process of oxidation, following which the chain had to be regenerated by the cell. Two years later, in 1897, Ehrlich elaborated this idea into his influential side chain theory of immunity. He postulated that a ‘receptive side chain’ of a particular cell, for example one involved in nutrition, has an atom group which by mere coincidence possessed specific combining properties for a particular toxin, such as tetanus toxin. The normal function of the side chain is lost once the toxin binds to the group, triggering the cell to produce a large number of such side chains. Many of these excess side chains then break off from the cell and are so released into the blood stream. Here they act as antibodies or antitoxins, combining with the toxin in the blood stream and so preventing it from combining with cells. Ehrlich in this work likened the relation between toxin and receptive side chain, which by 1900 he referred to simply as ‘receptor’, to that between a ‘lock and a key’. In his Croonian Lecture to the Royal Society of London in 1900, Ehrlich specifically excluded his receptor theory for the actions of toxins as being applicable to the action of drugs on cells. He came to the conclusion that drugs are not bound firmly to cells like toxins as most of the former are easily extracted from tissues by solvents. Thus toxins are bound to the protoplasmic molecule by chemical union whereas pharmacological drugs are not as they do not possess appropriate groups. It follows that they are not capable of eliciting the production of antibodies. If alkaloids, aromatic amines, antipyretics, or aniline dyes be introduced into the animal body it is a very easy matter, by means of water, alcohol, or acetone, according to the nature of the substance, to remove all these things quickly and easily from the tissues… We are therefore obliged to conclude that none of the foreign bodies just mentioned enter synthetically into the cell complex; but are merely contained in the cells in their free state. The combinations into which they enter with the cells, and notably with the not really living parts of them are very unstable, and usually correspond only to the conditions in solid solutions, while in other cases only a feeble salt-like formation takes place. The conclusion reached by Ehrlich then in 1900, and reiterated in 1902, was that pharmacological substances do not possess the necessary atomic groups which would allow them to combine with the appropriate groups of the cell protoplasm (Ehrlich & Morgenroth, 1900). The ‘lock and key’ concept did not then apply to the interaction of drugs with cells, so that the ‘receptor’ concept did not apply in this instance. However by 1907 Ehrlich had completely changed his mind on this issue, even introduced the word ‘chemoreceptor’ to describe the interaction of drugs with cells. What had happened in the five years between 1902 and 1907 to change his mind on this issue was largely due to the work during this period of the laboratory of Langley, which will now be described. 3.4 John Langley and T.R.Elliott: the emergence of the concept of chemical transmission between sympathetic nerves and smooth muscle In 1899 Lewandowsky observed that supra-renal extract causes in cats dilation of the pupil, withdrawal of the nictitating membrane (Fig. 3.3 A), separation of the eyelids and protrusion of the eyeball. Lewandowsky suggested that the extract acted directly on the smooth muscle and not on the nerve endings in the muscle as he obtained the same results with the extract after excision of the superior cervical ganglion and degeneration of the postganglionic nerves as in the normal animal. This was an extraordinary insightful interpretation, which formed the basis for the subsequent comprehensive study of the effects of supra-renal extract by Langley. In 1901 he inquired into the effects produced by supra-renal extract in the cat and rabbit on different organs, and arranged them in order as regards the amount of extract required per body weight to produce an obvious effect, as shown in the Table of Fig. 3.3B. This table shows that the extract in some cases contracts smooth muscles of a particular organ and in other cases relaxes the muscle. Langley had already, in 1898, defined the autonomic nervous system which he divided into sympathetic, cranial, sacral and enteric components. In his 1901 paper he makes the historic remarks:

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

47

Fig. 3.2. The experiments of Claude Bernard and John Langley establishing the concept of `receptive substances' at somatic neuromuscular junctions. (A) A frog preparation illustrating an experiment by Claude Bernard. A ligature passed around the waist of the frog isolated the lower-limbs from the rest of the body, except for the sciatic nerve trunk. The experiment is described in the text. (Reproduced from Lecons 1883 edition, Fig. 26, p. 345; from Fig. 26 in Bernard, 1883; reproduced by Fessard as Fig. 2, p.111 in Grande & Visscher, 1967). (B) Illustration of an experiment with curare carried out by Claude Bernard (reproduced from Lecons 1883 ed., Fig. 23, p. 329; from Fig. 23 in Bernard, 1883; reproduced by Fessard as Fig. 3, p.112 in Grande & Visscher, 1967). (C) A drawing made by Claude Bernard to illustrate one of his experiments on neuromuscular curarization (see text). Reproduced from the original Note-book ‘Cahier rouge’ (also recently reproduced in Cahier de Notes, 1965, p. 76). Inset: an explanatory scheme of the drawing that is described in the text. (From pg. 76 in ‘Cahier de Notes’, 1965, Bernard; reproduced by Fessard as Fig. 4, p. 113 in Grande & Visscher, 1967). (D) Frog killed by destroying the whole of the central nervous system. Contraction of the muscles of the forelimbs caused by nicotine. (From Fig. 7 in Langley, 1906). (E) Nicotine injected into the abdominal cavity of a frog, whose spinal cord and brain had been destroyed. For details of the experiment see the text. (From Fig. 9 in Langley, 1906). (F) Fowl, anaesthetised with morphia and A.C.E. mixture, balanced on its thorax in a V-shaped piece of wood. The neck and legs hang down and are flaccid, the eyes are shut. (From Fig. 1 in Langley, 1906). (G) The same fowl as in F. Two minutes after injection of 5 mg of nicotine into the jugular vein. The injection caused a gradual and fairly quick extension of the legs, retraction and twisting of the neck, and opening of the eyes. In order to show the eyes, the beak was held when taking the photograph. The fowl was unfastened throughout, and the injection caused no general movement nor any decrease of the anaesthesia. (From Fig. 2 in Langley, 1906).

It is a noteworthy fact that the effect of supra-renal extract in no case corresponds to that which is produced by stimulation in normal conditions of a cranial autonomic or of a sacral autonomic nerve. It does not produce the effect of stimulating the third nerve on the eye, nor of the vagus on the stomach or the heart, nor the effect of stimulating the

48

HISTORY OF THE SYNAPSE

pelvic nerve on the bladder, the rectum, the anus, or the generative organs. It is true that it causes a free secretion of saliva, but the secretion is not accompanied in its first stages by increased vascularity such as is caused by stimulation of the chorda tympani of Jacobson’s nerve. It is equally noteworthy that the effects produced by suprarenal extract are almost all such as are produced by stimulation of some one or other sympathetic nerve. In many cases the effects produced by the extract and by electrical stimulation of the sympathetic nerve correspond exactly. Having made these observations, and the fact that the effects of the extract persist after denervation of the organs, as Lewandowsky had first observed, Langley reached the conclusion that: …the difference in action on different autonomic tissues must depend upon their intrinsic differences. However, at this time Langley did not comment on the possiblity that the sympathetic nerves exerted their effects by the release of a substance equivalent to suprarenal extract. In 1904, Langley’s student Elliott reported experiments that showed even further the parallel effects of sympathetic nerve stimulation and of supra-renal extract (now identified as adrenaline by Takamine, 1901) on autonomic effectors. He showed that stimulation of the sympathetic nerves causes the sphincter at the junction between the small and large intestine to contract at the same time inhibiting the circular muscle in the wall of the ileum and colon adjoining the sphincter. Adrenaline produced the same effect as sympathetic nerve stimulation, thus contracting the sphincter (Fig. 3.3C) and relaxing the circular muscle of the surrounding ileum. These observations emphasised the parallel actions of adrenaline and of sympathetic nerve stimulation on the smooth muscle of different organs, and in this case within an organ. In that same year Elliott carried out an extensive study of the parallel actions of sympathetic nerve stimulation to the smooth muscles of different organs and that of the action of adrenaline on these, such as inhibition of the stomach by the splanchnic nerves (Fig. 3.3D and 3.3E), examining all apparent exceptions to this rule by previous investigators (not including Langley) and came to the conclusion, published in 1905, that: In all vertebrates the reaction of any plain muscle to adrenalin is of similar character to that following excitation of the sympathetic (thoracico-lumbar) visceral nerves supplying that muscle. The change may be either to contraction or relaxation. In default of sympathetic innervation plain muscle is indifferent to adrenalin. A positive reaction to adrenalin is a trustworthy proof of the existence and nature of sympathetic nerves in any organ. Sympathetic nerve cells with their fibres, and the contractile muscle fibres are not irritated by adrenalin. Elliott was therefore led to conclude that since some plain muscles that do not receive a sympathetic innervation are not affected by adrenaline, then the contractile apparatus cannot be the site of action of this substance. Furthermore, as Lewandowsky, Langley and Elliott himself had shown that the actions of adrenaline were not dependent on an intact sympathetic nerve supply, then it was concluded that adrenaline did not exert its effects through the nerve supply. Elliott was then led to the important conclusion that: The stimulation takes place at the junction of muscle and nerve (Fig. 3.3F). The irritable substance at the myoneural junction depends for continuance of life on the nucleoplasm of the muscle cell, not of the nerve cell. However nowhere in this classic paper of 1905 is there any mention that stimulation of the muscle by the nerve involves the release of a chemical substance, let alone that this substance in the case of sympathetic nerves is adrenaline. Yet in a proceedings note to the Physiological Society of 1904 Elliott makes his claim for chemical transmission at sympathetic nerve terminals and that this might be adrenaline. In that famous note he presents the evidence in favour of his two hypotheses as follows (Elliott, 1904a): 1. the effect of adrenalin upon plain muscle is the same as the effect of exciting the sympathetic nerves supplying that particular tissue. 2. [the] medulla and the sympathetic ganglia have a common parentage’, (see Kohn, 1903a, b). 3. …the facts suggest that the sympathetic axons cannot excite the peripheral tissues except in the presence and perhaps through the agency, of the adrenalin or its immediate precursor secreted by the sympathetic paraganglia. 4. Adrenalin does not excite sympathetic ganglia when applied to them directly, as does nicotine. Its effective action is localised to the periphery. 5. …even after such complete denervation, whether of three days’ or ten months’ duration, the plain muscle of the dilatator pupillae will respond to adrenalin.

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

Fig. 3.3. Elliott and Langley establish the concept of adrenaline as a transmitter at the autonomic neuromuscular junction. (A) Blutdruck und Membrana nictitans. Injection of 1ccm of extract of the adrenal bodies. (From Fig. 2 in Lewandowsky, 1899). (B) The effects produced by supra-renal extract in the cat and rabbit may be arranged roughly in the order shown as regards the amount of extract required per body weight to produce an obvious effect. (From the Table in Langley, 1901). (C) Cat. Vagi cut. Injection of 0.3 mgm. adrenalin into external jugular vein. A is the record of the ileo-colic sphincter under pressure of 15 cm. B gives the period of injection. The figures 12, 14, 16 indicate the blood-pressure in cms, given in the upper trace. Bottom trace gives time marker in seconds (not detectable in the original figure). (From Fig. 6 in Elliott, 1904b).

In summary then: Therefore it cannot be that adrenalin excites any structure derived from, and dependent for its persistence on, the peripheral neurone. But since adrenalin does not evoke any reaction from muscle that has at no time of its life been innervated by the sympathetic (for example the absence of action on the muscle of the bronchioles and of the pulmonary blood vessels, as shown by Brodie and Dixon, 1904), the point at which the stimulus of the chemical excitant is received, and transformed into what may cause the change of tension of the muscle fibre, is perhaps a mechanism developed out of the muscle cell in response to its union with the synapsing sympathetic fibre, the function of which is

49

50

HISTORY OF THE SYNAPSE

Fig. 3.3. (D) Inhibition of cat’s stomach by splanchnic nerves. Ether. Vagi and splanchnic cut within thorax and placed on shielded electrodes. Artificial respiration. Record of volume change of stomach under constant pressure of 7 cm. water. Stimulation of splanchnics caused rise of blood pressure and relaxation of stomach by 30 c.c. Period of stimulation of the splanchnics is given by the second trace from the bottom. Bottom trace gives time marker in seconds (barely detectable in original figure). (From Fig. 6 in Elliott, 1905). (E) From same experiment as in D. Injection of .18 mgm. adrenalin completely relaxed the stomach (second trace). Tone did not return until the vagi were again stimulated. Period of application of adrenalin given by the second trace from the bottom. The uppermost trace gives the blood pressure. The numbers on the traces are almost illegible: 210 mm Hg is indicated from above a base of 70 mm Hg. Bottom trace gives time marker in seconds (not clear in original record). (From Fig. 8 in Elliott, 1905). (F) Diagram of doubly innervated muscle-nerve system. (1) the sympathetic motor ganglion cell, (2) its axon, and (3) the nerve ending; (4) the myoneural junction; (5) the contractile muscle fibre. (1a) and (4a) the corresponding parts of the inhibitor mechanism. To simplify the diagram the motor myoneural junction (4) is represented as spatially separated from the inhibitor (4a). (From Fig. 10 in Elliott, 1905).

to receive and transform the nervous impulse. Adrenalin might then be the chemical stimulant liberated on each occasion when the impulse arrives at the periphery. We therefore have for the first time a succinct statement of the concept of chemical transmission but also identification of a transmitter substance. Although Langley undoubtedly supplied the intellectual environment in the laboratory for the development of these hypotheses, there is no sign in his papers to this time (1904) that he had joined together the set of

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

51

numbered observations indicated above to arrive at the conclusion that chemical transmission is most likely to occur at sympathetic nerve terminals and that the transmitter is adrenaline. Langley was always loathe to speculate and develop hypotheses. His great experimental career consists of generating a formidable set of facts that lead inexorably to a conclusion. On no occasion did he draw a diagram of the kind shown in Fig. 3.3F from Elliott (1905) that so brilliantly concentrates one’s interest on the neuromuscular junction. Indeed this figure may be compared for its fruitful prescience with that of Sherrington’s figure of 1906 in the Integrative Action of the Nervous System, showing the monosynaptic connections of the motor and sensory nerves in the spinal cord. Langley’s reticence meant that he missed out generating the brilliantly fruitful hypotheses of Elliott. It may also be that Elliott himself was dissuaded by Langley from continuing down the path of elaborating his ideas further, as following the note of 1904 there is no mention by Elliott in the very substantial paper of 1905 of either the chemical transmission hypothesis or the possibility that adrenalin is a transmitter. 3.5 The action of curare and John Langley's development of the idea of transmitter receptors By 1904 it was clear that adrenalin acted on those smooth muscles that received a sympathetic innervation and that this action was independent of the nerve supply to the muscles. Elliott did not elaborate further on his concept of chemical transmission in his 1905 paper that there is a: mechanism developed out of the muscle cell in response to its union with the synapsing sympathetic fibre, the function of which is to receive and transform the nervous impulse, (Fig. 3.3F). However there were undoubtedly discussions in the Cambridge physiological laboratory concerning his hypotheses. This is made to some extent explicit by Langley in his first paper on the actions of curare on striated muscle in 1905 in which he says in the introduction: Elliott brings forward further and most striking evidence that adrenalin stimulates tissues which are stimulated by sympathetic nerves and these only. This leads him to look on adrenalin as acting on some substance common to sympathetic nerves. He finds, however, that degeneration of the nerves does not diminish the action of adrenalin, and as he considers that the axon endings degenerate, the substance affected by adrenalin must be in trophic connection with the muscle. This as I have pointed out above is, I think, the same as saying that it is part of the muscle. But in view of the close relation of adrenalin to sympathetic nerves, and because he considers it improbable that the varying action of adrenalin can be due to intrinsic differences in the muscle, he concludes that when sympathetic nerves unite with unstriated muscle they cause the formation in it of a new substance, the myo-neural junction, and it is this which is acted upon by adrenalin. Now supposing that nervous connection does cause in the muscle the formation of a new substance, this does not make the new substance any the less part of the muscle. The fundamental fact of Elliott’s view is then, I think, the same as mine, viz. that adrenalin acts directly on muscle. The concept of Elliott’s ‘new substance’ therefore had a major influence on how Langley designed his experiments concerning the manner by which curare acted. These were not only based on the conceptual framework of Elliott but also on Langley’s own discovery that nicotine stimulates sympathetic nerve cells by a ‘direct action upon them’ (Langley, 1901). Furthermore, it must not be forgotten that Langley’s first experiments in 1874, while still a student at Cambridge under the guidance of Michael Foster, involved an investigation into the actions of atropine and pilocarpine (muscarine like) on the secretion of saliva by the submaxillary gland. He found that these had opposite effects and in his full paper on this antagonism, published in 1878, there is the comment that: …we may, I think, without much rashness, assume that there is a substance or substances in the nerve endings or gland cells with which both atropine and pilocarpine are capable of forming compounds. On this assumption then the atropine or pilocarpine compounds are formed according to some law of which their relative mass and chemical affinity for the substance are factors. So Langley, some thirty years or more before the experiments of Elliott or for that matter of his own on curare, was already developing the idea of pharmacological agents forming compounds with the substances in cells. The concept of the receptor is clearly present in these early formulations, which are in contrast to those of Ehrlich in 1900 mentioned above. Furthermore, the observations upon which these conjectures were developed were made well before Ehrlich began his research in 1878. In 1905 Langley showed that injection of nicotine into the vein of an anaesthetised fowl led to gradual stiffening and extension of the hindlimbs over a couple of minutes due to tonic contraction of the red muscles (Figs. 3.2F and 3.2G). This

52

HISTORY OF THE SYNAPSE

effect still occurred after section of the sciatic and crural nerves, so that it did not involve the nerve supply. In order to provide quantitative details of this effect, Langley took measurements of the gastrocnemius muscle in the fowl following injection of nicotine into the vein, without interfering with the muscle’s blood supply and after cutting the sciatic and crural nerves. The results showed that the muscle contracted for several minutes (Fig. 3.4A upper). Langley then injected curare about one minute after the beginning of the nicotine-induced contraction: the muscle then relaxed (Fig. 3.4A lower). Repeating this experiment after cutting the sciatic and crural nerves several weeks previously so as to allow their peripheral extensions to fully degenerate, did not alter the results of injecting nicotine and curari, as shown in Fig. 3.4B. Langley (1905) commented on these experiments that: I conclude then that nicotine acts upon the muscle substance, and not upon the axon-endings. It has been shown above that curari acts upon the same substance as nicotine. It follows then that curari acts upon the muscle substance and not upon the axon-endings. Since, in the normal state, both nicotine and curari abolish the effect of nerve stimulation, but do not prevent contraction from being obtained by direct stimulation of the muscle or by a further adequate injection of nicotine, it may be inferred that neither the poisons nor the nervous impulse act directly on the contractile substance of the muscle but on some accessory substance. Since this accessory substance is the recipient of stimuli which it transfers to the contractile material, we may speak of it as the receptive substance of the muscle. This is the first occasion on which the phrase ‘receptive substance’ or as we now simply call it ‘receptor’ was used. In developing this concept, Langley (1905) makes his indebtedness to Elliott quite clear when he comments that: The subsequent work mentioned in the Introduction, and especially that of Elliott on the action of adrenalin, made the issues clearer He goes on to say that: In my view, the myo-neural junction is a part of the receptive substance localized in the neighbourhood of the axonending. Although Langley (1907) went on to give detailed descriptions of how nicotine acts to block transmission at the myo-neural junction, especially at high concentrations (Fig. 3.4E), and of the fact that the effects of nicotine are only found in those parts of individual muscles where nerve endings are found, his concept of the receptive substance was fully matured by the time he gave a Croonian lecture to the Royal Society of London in 1906. In that lecture he presented most elegant tracings of how contractions of the fowl’s gastrocnemius due to nicotine are blocked by curari (Fig. 3.4C). In addition he gave graphic demonstrations of the effects of nicotine when injected into the abdominal cavity of frogs whose brain and spinal cord were destroyed. These passed from a state in which the muscles are flaccid, so that when the limbs are raised they at once fall, to a condition in which there is maximum flexion of the forelimbs, as shown in Fig. 3.2D (see also Fig. 3.4D). The same experiment when performed on toads gives rise to a cataleptic condition in both the forelimbs and the hindlimbs. The contraction of the flexors and extensors of the arm are about equal so that there is little movement or no movement. The forelimbs can then be moved about almost as if made of lead, and stay with but slight return movement in any position in which they are placed consistent with the arrangement of the joints and ligaments (Fig. 3.2E). These cataleptic conditions are completely abolished by sufficient doses of curari. The conclusion is reached that: The mutual antagonism of nicotine and curari on muscle can only satisfactorily be explained by supposing that both combine with the same radicle of the muscle, so that nicotine-muscle compounds and curari-muscle compounds are formed.’ ‘Since neither curari nor nicotine, even in large doses, prevents direct stimulation of muscle from causing contraction, it is obvious that the muscle substance which combines with nicotine or curari is not identical with the substance which contracts. It is convenient to have a term for the specially excitable constituent, and I have called it the receptive substance. Langley (1906) concludes his lecture by drawing attention to the fact that if the set of experiments on muscle with nicotine and curari are carried out on the excitability of nerve cells in sympathetic ganglia, then analogous results are obtained (Langley, 1901). Thus: ‘I conclude that the substance affected by the poisons is a special receptive substance and not the fundamental substance of the cell.’ As to transmission between nerve endings and smooth muscle as well as glands, the arguments outlined above concerning the action of adrenalin that have been developed in particular by Elliott indicate that:

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

53

Fig. 3.4. The experiments of John Langley that established the existence of `receptive substances' at the somatic neuromuscular junction. (A) Effects of curari and nicotine. Upper panel shows a contraction produced by a large dose of nicotine, applied for a time given by the second trace from the bottom. The bottom trace gives a time marker in 10s intervals. There is no calibration of the size of the contractions. Lower panel shows a similar contraction annulled by curari; the upper curve gives the blood pressure. (From Figs. 4 and 5 in Langley 1905). (B) Denervated muscle. Effect of nicotine and of curari after nicotine applied at times indicated in the trace second from the bottom. Time in sees, omitting every tenth is shown in the bottom trace, but is not clear in the original figure. There is no calibration of the size of the contraction. (From Fig. 7 in Langley 1905). (C) Abolition by curari of the contraction in the gastrocnemius muscle of the fowl caused by nicotine. The line second from the bottom gives the period of application of the drugs. The lowest line marks intervals of 10 seconds. No calibration is given of the size of the contraction. (From Fig. 4 in Langley 1906). (D) Frog. Brain and spinal cord destroyed. A thread was tied to the manus and connected with an unweighted lever, so that flexion of the arm caused a rise of the lever; 1 c.c. of a 1 per cent, nicotine injected into the abdominal cavity at the time shown by the signal in the bottom trace (the ‘1cc1pc nicotine’ is printed on the trace). Time marked in 10 seconds. No calibration is given of the size of the upper trace. (From Fig. 8 in Langley 1906). (E) Contraction of sartorius with nicotine compared with that due to direct stimulation (top trace). Effect of stimulating with make and break induction shocks before and after nicotine: 0’—Muscle in Ringer’s fluid (this is not shown in the original figure). 2’— Stimulate first with make and then with break shock, raise lever to base line. 7’—pour .005 per cent, nicotine into muscle chamber, beginning and end of pouring is marked. 11’—Stop drum, and run off nicotine. 11½’ and 12½’—Stimulate with make and break shocks. Bottom trace, time in 1s intervals. Period of application of make and then break shocks and of nicotine indicated in second bottom trace. No calibration is given of the size of the contraction in the upper trace. (From Fig.1 in Langley, 1907).

The legitimate statement from the premises is that it does not act on any muscle substance or on any nerve substance outside the limits of the myoneural junction. As regards the localisation of the receptive substance, strong evidence that this occurs to a considerable extent is afforded by the action both of adrenalin and of chrysotoxin on tissues which have a double nerve supply, but the evidence cannot be regarded as conclusive.

54

HISTORY OF THE SYNAPSE

It seems likely, despite some of Langley’s arguments to the contrary (viz. that Elliott adopted the theory that nerve and unstriated muscle are continuous), that Elliott had conceived of Langley’s receptive substance, but without giving it a name for he comments: But since adrenalin does not evoke any reaction from muscle that has at no time of its life been innervated by the sympathetic, the point at which the stimulus of the chemical excitant is received, and transformed into what may cause the change of tension of the muscle fibre, is perhaps a mechanism developed out of the muscle cell in response to its union with the synapsing sympathetic fibre, the function of which it is to receive and transform the nervous impulse. One cannot do better in this early part of the history of the receptor than to quote the conclusion of Langley’s great Croonian lecture: In the foregoing account we have seen reason to believe that in each of the three great types of connection of the peripheral end of an efferent nerve with a cell it is some constituent of the cell substance which is stimulated or paralysed by poisons ordinarily taken as stimulating or paralysing nerve endings. This theory adds to the complexity of the cell. It necessitates the presence in it of one or more substances (receptive substances) which are capable of receiving and transmitting stimuli, and capable of isolated paralysis, and also of a substance or substances concerned with the main function of the cell (contraction or secretion, or, in the case of nerve cells, of discharging nerve impulses). The Croonian lecture of 1906 on the antagonism between curare and nicotine giving rise to the concept of the transmitter receptor brings to a conclusion the research program embarked on by Langley some thirty years earlier while still a student at Cambridge working on the antagonism between atropine and pilocarpine. The antagonism of this set of drugs is seen as acting on the protoplasmic substance or substances in the muscle or effector organ, and does involve the combination of an alkaloid with protoplasm, in contradistinction to the suggestions of Ehrlich. Following these experiments of Langley, Ehrlich accepted the idea of the receptor for alkaloids such as nicotine and curare. 3.6 The Langley-Ehrlich receptor theory It is fascinating to trace the productive interaction of the ideas of Ehrlich and Langley over this period from 1878 to 1908, beginning as each did from quite different research programs. In the case of Ehrlich, his research was concerned with drug resistance as a consequence of studies on the chemotherapy of trypanosomes. The receptive side chain concept was developed in order to give a theoretical underpinning to his work on the chemotherapy of such micro-organisms, in particular in the use of substances that act in a manner that is largely irreversible, such as the arsenicals. For Langley, the starting point involved his research on the effects on muscle and nerve of alkaloids such as nicotine, curare, atropine, pilocarpine, strychnine and adrenaline, all of which produce an action that is relatively reversible compared with the actions studied by Ehrlich. Nevertheless by 1908 Langley could say: My theory of the action is in general on the lines of Ehrlich’s theory of immunity. I take it that the contractile molecule has a number of ‘receptive’ or side-chain radicles and the nicotine by combining with one of these causes contraction and by combining with another causes twitching… To which Ehrlich commented in 1914: For many reasons I had hesitated to apply these ideas about receptors to chemical substances in general, and in this connection it was, in particular, the brilliant investigations by Langley, on the effects of alkaloids, which caused my doubts to disappear and made the existence of chemo-receptors seem probable to me. Thus was conceived the tremendously fruitful concept of the receptor. 3.7 The discovery of acetylcholine and its physiological action at autonomic neuroeffector junctions In 1906 Hunt and Taveau synthesized acetylcholine and reported that:

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

55

…as regards its effect upon the circulation, it is the most powerful substance known. They went on to say: we have not determined the cause of the fall of blood pressure from acetyl-cholin, but from the fact that it can be prevented entirely by atropine, I am inclined to think that it is due to an effect upon the terminations of the vagus in the heart. In the same year, Dixon (1906) gave a description of his experiments on the vagus inhibition of the heart, and stated that he was of the opinion that: the heart contains a substance—‘pro-inhibitin’, which, as a result of vagus excitation, is converted into a chemical body —‘inhibitin’. This substance, combining with the heart muscle, results in cardiac standstill. These suggestions were extraordinarily prescient of Loewi’s experiments on the identity of acetylcholine as the inhibitory transmitter released by the vagus in the heart subsequently carried out in the early 1920s (see below). In 1914, Ewins showed that extracts of certain specimens of the fungus ergot contained an active principle, which he identified as acetylcholine. Much of this work of Ewins was done at the instigation of Dale (Fig. 3.1C), as a consequence of his knowledge of the work of Hunt and Taveau (1906) who had already shown that synthesized acetylcholine had powerful depressor activity, indeed Dale reported to Elliott: We got that thing out of our silly ergot extract. It is acetylcholine and an most interesting substance. It is much more active than muscarine, though so easily hydrolysed that its action, when it is injected into the blood-stream, is remarkably evanescent, so that it can be given over and over again with exactly similar effects, like adrenaline. Here is a good candidate for the role of a hormone related to the rest of the autonomic nervous system, I am perilously near wild theorising… I shall be surprised, however, if this principle, once identified, does not turn up in all sorts of tissueextracts (see Letter, Dale to Elliott, 11 Dec, 1913, Contemporary Medical Archives Centre, Wellcome Institute, GC/42 ‘T.R.Elliott’; quoted in Tansey 1991). It was then of considerable interest to see if the depressor effect of acetylcholine was its principal effect. In 1914, Dale determined to carry out a systematic study of the effects of acetylcholine on various organs of the autonomic and somatic nervous systems (Dale, 1914a, b). In that work he showed that following the injection of acetylcholine into an animal, there were considerable responses elicited in a number of organs: cat’s blood pressure declined (Fig. 3.5A); there was complete cessation of the heart beat in frogs (Fig. 3.5B) and the rabbit small intestine contracted vigorously (Fig. 3.5C). Dale was alert to the fact that these effects are those produced by stimulation of the vagus nerve, and in a sense constitute a complementary set of effects to those observed by Langley and Elliott with respect to adrenalin and sympathetic nerve stimulation. He comments that: The question of a possible physiological significance, in the resemblance between the action of choline esters and the effects of certain divisions of the involuntary nervous system, is one of great interest, but one for the discussion of which little evidence is available. Acetylcholine is, of all the substances examined, the one whose action is most suggestive in this direction. The fact that its action surpasses even that of adrenine both in intensity and evanescence, when considered in conjunction with the fact that each of these two bases reproduces those effects of involuntary nerves which are absent from the action of the other, so that the two actions are in many directions at once complementary and antagonistic, gives plenty of scope for speculation. The main problem at this time in making a physiological claim for acetylcholine was: On the other hand, there is no known depot of choline derivatives, corresponding to the adrenine depot in the adrenal medulla, nor, indeed, any evidence that a substance resembling acetylcholine exists in the body at all. So Dale concludes in this paper which gives the first systematic account of the actions of acetylcholine in the peripheral nervous system that: Acetylcholine occurs occasionally in ergot, but its instability renders it improbable that its occurrence has any therapeutic significance’.

56

HISTORY OF THE SYNAPSE

Fig. 3.5. Henry Dale and Otto Loewi establish acetylcholine as a transmitter substance at autonomic neuromuscular junctions. (A) Cathether. Plethysmograph records from intestine and limb with calibration in c.c. Carotid blood-pressure shown with calibration in mmHg. Injection of 0.001 mg acetylcholine indicated by the second trace from the bottom. Bottom trace is time marker in 10s intervals. Third trace from the top is not specified (From Fig. 3 in Dale, 1914a). (B) Perfused heart of frog, recorded by suspension-lever. Tracings read from right to left, the vertical line in each case indicating change from pure Ringer’s solution to similar solution containing a choline-ester (the vertical line is about 7 beats from the right in each case), a Acetylcholine 1 in 100 millions, b Acetylcholine 1 in 200 millions, c Acetylcholine 1 in 500 millions, d Acetylcholine 1 in 1000 millions, e Nitroso-choline 1 in 100 000. 1 Nitroso-choline 1 in 1 million. Bottom trace is time in seconds. There is no calibration of the size of the contraction (From Fig. 11 in Dale, 1914a) (C) Loop of Rabbit’s small Intestine in 50 cc of Tyrode’s solution. At A 0.01 mg synthetic Acetyl-choline. At B 0.01 mg of Acetylcholine from ergot, added to the bath. At R, R, fresh Tyrode’s solution. Bottom trace is time marker in 10 s intervals. No calibration is given for the size of the contraction (From Fig. 13 in Dale, 1914a).

There the matter rested until after the first World War. In 1921 Loewi (Fig. 3.1D) reported from Austria his experiments on the heart indicating that chemical transmission occurred between the vagus nerve and the heart, mediated by a substance which he called ‘Vagusstoff’. Fig. 3.5D shows his original record, in which the heart beat in 1 is in normal Ringer, whereas the subsequent decline in the heart beat in 2 is due to the addition of Ringer that has been in contact with another heart whose vagus had been stimulated for 15 minutes; at 3 the heart beat returns to normal as a consequence of it being exposed to a Ringer that had been in contact with another heart for which

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

57

Fig. 3.5. (D) Frog heart contractions. At 1. Ringer. 2. Ringer from a heart that received 15 min of vagal stimulation. 3. Ringer from a heart that did not receive stimulation. 4. Addition of 0.1 mg atropine. Bottom trace is not specified and there is no time or contraction calibration given. (From Fig. 1 in Loewi, 1921). (E) Frog heart contractions (a) The testing material is heart extract (1:20 dilution) plus acetylcholine (dilution 1:100,000). 2 is inactivated heart extract plus acetylcholine (dilution 1:1000). 3 is heart extract plus acetylcholine again, (b). The starting material is again heart extract (dilution 1:20) plus acetylcholine (dilution 1:100,000). 1 is heart extract inactivated by standing for 30 min plus acetylcholine (dilution 1: 750). 2 is heart extract inactivated by standing for 90 min plus acetylcholine (dilution 1:1000). 3 is heart extract inactivated by standing for 90 min plus acetylcholine (dilution 1:150). Bottom trace is not specified and no time or contraction calibrations are given. (From Fig, 5 in Loewi & Navratil, 1926a). (F) Frog heart contractions. 1=Ringer. 2=Vagusstoff (alcoholic extract from the heart). 3=Acetylcholine (dilution 1:100 million). Bottom trace not specified and no time or contraction calibrations given. (From Fig. 4 in Loewi & Navratil, 1926a). (G) Effect of stimulating the heart nerve on contraction of the frog’s heart in relation to the effects of applied acetylcholine and physostigmine given at the times indicated. In the left-hand panel, samples were taken from the heart at the time indicated by I and another taken at the time indicated by II (Inhaltsentnahme). In the right-hand panel the effects of applying these samples (I and II) to another heart are shown. Bottom trace is not specified and no time or contraction calibration is given. (From Fig. 3 in Loewi & Navratil, 1926b). (H) Effect of eserine on the frog’s heart. The caption reads: ‘Heart extract with Eserine without. Left standing for 2 hours.’ Bottom trace not specified and no calibrations given for the time of contraction. (From Fig. 5 in Loewi & Navratil, 1926b). (I) Frog heart, a=Acetylcholine (dilution 1 :million). I=Bovine blood plus acetylcholine. II=Heated bovine blood plus acetylcholine. 1=Inactivated bovine blood plus acetylcholine. 2=Heated bovine blood plus acetylcholine. (From Fig. 7 in Engelhart & Loewi, 1930). (J) Comparison of purified spleen extract (S.E.) and acetylcholine (A.C.) solution on the blood-pressure of a cat under ether. Calibration of the blood pressure is given in mm Hg. The amount of A.C. given is 0.01 mg. About 0.4 cc of the spleen extract gave the same result as that of 0.01 mg of A.C. (From Fig. 1 in Dale & Dudley; 1929)

the vagus was not stimulated; at 4 atropine was added to the normal Ringer and this increases the heart beat. By 1926 Loewi and Navratil had produced sufficient evidence to mount a persuasive case that Vagusstoff was acetylcholine: application of very low concentrations of acetylcholine to the heart greatly decreased the heart beat (Loewi & Navratil, 1926a; Fig. 3.5E and

58

HISTORY OF THE SYNAPSE

3.5F). These authors went on to show that inactivation of esterases in the heart for acetylcholine, using eserine (physostigmine; Fig. 3.5G, 3.5H and 3.5I; Loewi & Navratil, 1926b; Engelhart & Loewi, 1930) greatly potentiated the decrease in the heart beat brought about by the exogenous application of acetylcholine. This at last provided an explanation for the problem that Dale had referred to with regards to acetylcholine, that: when it is injected into the blood-stream, it is remarkably evanescent. Although it had been known for some time that physostigmine ‘sensitized’ the heart to stimulation of the vagus, this demonstration by Loewi and Navratil that physostigmine acted to potentiate the actions of exogenous acetylcholine led directly to the concept by these authors that the heart possesses an endogenous esterase for acetylcholine, for which they later coined the term ‘cholinesterase’. Dale continued in the late 1920s to try and find a natural source of acetylcholine in the body like that provided for adrenaline in the adrenal medulla. He commented to a friend in 1929: We are still struggling with the acetylcholine problem, which I mentioned to you when I saw you in the autumn. I am more and more convinced that the thing is there to be found, if only we can overcome the technical difficulties’ (Letter, Dale to Richards, 22 Mar 1929, Archives of the National Institute for Medical Research, File 647; quoted in Tansey, 1991). The breakthrough occurred in 1929, when Dale and a chemist Dudley discovered that acetylcholine was a natural constituent of both horse and ox spleens (Fig. 3.5J), thus giving the long sought after ‘depot of choline derivatives’ in the body. This then provided the necessary impetus for once more examining the role of acetylcholine as the mediator of the effects of transmission from parasympathetic nerves to the effectors of the autonomic nervous system as well as at other sites of transmission in the body. For as they state in their paper (Dale & Dudley, 1929): But there has been a natural and proper reluctance to assume, in default of chemical evidence, that the chemical agent concerned in these effects, or in the humoral transmission of vagus action, was a substance known, hitherto, only as a synthetic curiosity, or as an occasional constituent of certain plant extracts… It appears to us that the case for acetylcholine as a physiological agent is now materially strengthened by the fact that we have now been able to isolate it from an animal organ and thus to show that it is a natural constituent of the body. They go on to say: We feel, however, that its definite isolation from one organ has so far altered the position that, when an extract from, or a fluid in contact with the cells of, an animal organ can be shown to contain a principle having the actions, and the peculiar instability, of acetylcholine, it will be reasonable in future to assume the identification. On such lines a physiological survey of its distribution should be practicable. Similarly, when there is evidence associating some physiological event with the liberation of a substance indistinguishable from acetylcholine by its action, the presumption that it is, indeed, that ester will be strengthened by the knowledge that acetylcholine occurs in the normal body. By 1930, then, three decades of research had shown firstly that acetylcholine, either synthesized or extracted from ergot, had dramatic depressor effects (Ewins, 1914; Hunt & Taveau, 1906; Dale, 1914b); secondly, that stimulation of the vagus to the heart released a substance that seemed to mediate transmission and which had properties remarkably similar to that of acetylcholine (Loewi, 1921), and finally that acetylcholine occurred naturally in mammals (Dale & Dudley, 1929). The question that came to dominate the 1930s was whether acetylcholine could be detected in the overflow from other organs than the heart during nerve stimulation, and also whether the effects of such stimulation could be mimicked by the close arterial injection of acetylcholine into these organs. Loewi had already shown that the effects of exogenous acetylcholine on the heart or of stimulation of the vagus could be greatly enhanced if esterases for acetylcholine in this organ were first inhibited using physostigmine or eserine (Fig. 3.5G, H and I; Engelhart & Loewi, 1930). In 1930, Matthes, working in Dale’s laboratory, showed that the destruction of acetylcholine in the blood was due to the action of an esterase and that the action of this enzyme could be inhibited by eserine. Thus inhibition of acetylcholinesterases was a necessary requirement of any attempt to show that acetylcholine was released from a particular nerve ending, either by using an intravenous injection of eserine or by adding it to the organ bath. Shortly after this work of Matthes, Feldberg arrived in Dale’s laboratories as a refugee from Germany to Great Britain. He introduced the eserinized leech muscle preparation which had previously been shown to be exquisitely sensitive to applied acetylcholine by the German pharmacologist Fuhner in 1918. Feldberg was inspired to use this approach because of the experiments of Loewi and Navratil (1926b) on the actions of physostigmine in potentiating the effects of applied acetylcholine on the frog’s heart. In the hands of Dale and Feldberg (1934) the eserinized leech muscle was

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

59

Fig. 3.6. The identification of acetylcholine as a transmitter substance in autonomic ganglia. A. Effect of fluid collected from stimulated cat sympathetic ganglia on contraction of the frog’s heart and of eserinised leech muscle. Upper panel: frog’s heart (Straub’s method). Lower panel: leech muscle treated with eserine. (A), fluid collected from the ganglion during the stimulation. (D), control fluid. (B), (C), acetylcholine (15 and 30 mg per litre respectively). No time base or contraction calibration is given in the original figure. (From Fig. 6 in Feldberg & Gaddum, 1934). (B) Effect of eserine on nictitating membrane contraction due to stimulation of the cervical sympathetic. Responses to equal groups of submaximal shocks to cervical sympathetic; (A) and (B) before, (C) and (D) during perfusion of eserine 1 in 106. Time calibration is 10 ms; no contraction calibration is given. (From Fig. 6 in Feldberg & Vartiainen, 1934). (C) The evaluation of esterase activity in normal and decentralized sympathetic ganglia using rabbit’s jejunum contraction as test. Numbers give the contractions to the following amounts of acetylcholine: 1, 1.5 µg; 2, 2 µg; 7, 1.8 µg; 8, 1 µg; 10, 1.5 µg; 11, 1.7 µg; 12, 1.9 µg; 14, 2µg. Numbers 9 and 13 give the contraction in the presence of 10µg of acetylcholine plus an extract of normal sympathetic ganglia. Numbers 4 and 5 give the contraction in the presence of 3 µg and 2 µg of acetylcholine respectively in the presence of extract of denervated sympathetic ganglia. No time or contraction calibration is given. (From Fig. 1 in Brucke, 1937).

first used to show that acetylcholine appeared in the venous blood after stimulation of the vagus nerve to the stomach. This was followed by experiments that showed acetylcholine also appeared after stimulation of the splanchnic nerves to the suprarenal glands (adrenal medulla; Feldberg et al, 1934). 3.8 The physiological action of acetylcholine in autonomic ganglia In 1934 experiments were also begun to see if acetylcholine could be detected at neuronal synapses in addition to neuroeffector junctions. To this end Feldberg and Vartiainen (1934) were able to show that: …when the superior cervical ganglion of the cat is perfused with warm, oxygenated Locke’s solution containing a small proportion of eserine, acetylcholine appears in the venous effluent whenever the cervical sympathetic nerve is effectively stimulated, and only then The assay for this acetylcholine was either the frog’s heart (Fig. 6A, upper panel) or the leech muscle treated with eserine (Fig. 3.6A, lower panel). They regarded these observations as (Feldberg & Gaddum, 1934): …support for the theory that the mechanism by which each impulse normally passes the synapse consists in the liberation of a small quantity of acetylcholine This discovery of the release of acetylcholine in autonomic ganglia then raised the possibility that the ‘Vagusstoff’ of Loewi was in fact liberated from ganglia in the heart rather than at the neuroeffector junction. Whilst this would certainly be the case, Feldberg and Gaddum (1934) were persuaded of the view that most likely the Vagustoff collected by Loewi came from both the synapses in the intramural ganglia as well as the neuroeffector junction, although no very effective argument was offered in defence of this proposition. One caveat in these experiments on the ganglia was that Eccles had in the same year shown that the electrical signs of the action potential set up in the postganglionic nerve trunk of the superior cervical ganglion by volleys in the preganglionic trunk

60

HISTORY OF THE SYNAPSE

were depressed by eserine rather than potentiated as would be expected as a consequence of an enhanced transmission through the ganglion on inactivation of acetylcholinesterase with eserine (Eccles, 1934). Later in 1934, Feldberg and Vartiainen offered a vigorous defense of the idea that acetylcholine is the transmitter substance in autonomic ganglia. Using the nictitating membrane of the cat as a measure of the postganglionic volleys in response to stimulation of the preganglionic supply to the ganglion, they were able to show that 18 impulses delivered in 10s gave rise to a greatly potentiated response of the membrane after the ganglion had been perfused with eserine (Fig. 3.6B(C) and 3.6B(D)) compared with that in the absence of eserine (Fig. 3.6B(A) and (B)), leading them to conclude that: …these new items of evidence entirely support the conception of transmission at ganglionic synapses by liberation of acetylcholine. The idea that eserine has a depressant action on cholinesterase, which may be located at nerve terminals in ganglia, so that eserine potentiates the effects of endogenously released acetylcholine, was shown to be very likely when Brucke (1937) found high concentrations of cholinesterase in the superior cervical ganglion. He showed that this mostly disappeared on section and degeneration of the preganglionic nerves to the ganglion. These results are illustrated in Fig. 3.6C, where the contractile responses of the rabbit’s jejunum to different concentration (in gamma units) of acetylcholine alone (1=1.5; 2=2; 6=2; 7=1.8; 8=1; 10=1.5; 11=1.7; 12=1.9; 14 =2), or acetylcholine together with extract of normal ganglia that therefore contains cholinesterase (9 and 13 =10) or acetylcholine plus extract of denervated ganglion plus eserine (4=3; 5=2) are shown. It will be noted that much higher concentrations of acetylcholine and extract of normal ganglion had to be used in order to get responses comparable to that of acetylcholine alone, indicating the effects of cholinesterase in this case (Fig. 3.6C). The consensus opinion at the end of the 1930s was that acetylcholine acted as the transmitter of impulses in autonomic ganglia. Eccles provided the main continuing resistance to this idea with some persuasive arguments that are detailed elsewhere (Bennett, 1994). 3.9 The identity of acetylcholine as the transmitter substance at somatic neuromuscular junctions Although the concept of the transmitter receptor was developed primarily in relation to striated muscle, as detailed above, identification of acetylcholine as the transmitter substance that acts on these receptors came relatively late, well after the establishment of acetylcholine as the transmitter from the vagus nerve to the heart. In the late 1920s and early 1930s the problem of the relationship between motor nerves and muscle revolved around questions relating to their relative excitability and the actions of agents thought to exert effects at the junction between nerve and muscle, such as curare, had on this excitability. Lapicque had at the beginning of the century carried out a series of experiments on the excitability of nerve and muscle in which he had defined the chronaxie and rheobase of the strength-duration curve for setting up excitation in these tissues, as are now described in many text books (Lapicque & Lapicque, 1906; Lapicque, 1926). In these works he developed the concept that nerve and muscle possessed the same chronaxie which he defined as isochronism. This was challenged by Rushton (1930) who showed, following the work of Lucas (Lucas & Mines, 1907), that in general muscle possessed two different excitabilities, one associated with the intramuscular nerves and the other with the muscle fibres themselves (Rushton, 1930, 1932), so that nerve and muscle did not possess isochronism. The possibility that nerve terminals and the muscle or endplate could be brought into isochronism by the action of acetylcholine was then entertained as involved in the direct transmission of the nerve impulse into the muscle. This proposition was put forward as agents, such as acetylcholine, could shorten the chronaxie (Fredericq, 1924) whereas curare lengthened the chronaxie (Fig. 3.7A; Lapicque, 1934). The heated arguments used in the considerable controversy between Lapicque and Rushton outlined in their papers in the Journal of Physiology seemed to offer at this time the possibility for an esoteric interaction between the action of agents known to affect transmission at the endplate and the electrical properties of this region of the muscle. These electrical controversies concerning isochronism in relation to the mechanism of transmission between nerve and muscle declined in the second half of the 1930s. First, using the eserinized leech preparation, Dale et al. (1936) showed that considerable quantities of acetylcholine could be collected in venous fluids from the cat gastrocnemius muscle following nerve stimulation (Fig. 3.7C). This also occurred if the muscle was directly stimulated, even in the presence of curare, but not if the muscle had previously been denervated. Dale and his colleagues were careful at this time not to claim this showed acetylcholine, released at nerve terminals, transmitted the impulse from the terminals to muscle. They were still open to an interpretation: That the propagated disturbance in the nerve fibre is directly transmitted to the effector cell, but that the latter cannot accept it for further propagation unless sensitized by the action of the acetylcholine, which appears with its arrival at the nerve ending. Such an hypothesis might be stated in terms of Lapicque’s well-known conception, by supposing that the

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

61

action of acetylcholine shortens the chronaxie of the nerve cell, or of the motor end-plate of the muscle fibre, so that it is momentarily attuned to that of the nerve. Later in 1936 Brown, Dale and Feldberg showed that injection of acetylcholine directly into the empty arteries of a normal mammalian muscle could, if given with adequate rapidity, cause contraction of the muscle at not less that half the speed of a maximal motor nerve twitch (Fig. 3.7B). Furthermore, eserine caused the response to stimulation of the nerve supply to be converted from a simple twitch to a repetitive response with a maximum tension twice that of the normal twitch, a result reproduced and examined further by Bacq & Brown (1937; Fig. 3.7F). No mention is made any longer of Lapicque’s theories, especially given the effect of close intrarterial injection of acetylcholine mimicking the normal twitch response of the muscle to nerve stimulation. Rather they suggested that: …acetylcholine, is liberated by arrival of the nerve impulses at the nerve ending, and destroyed during the refractory period by a local concentration of cholinesterase. The facts supporting this hypothesis are: (1) that acetylcholine, when suitably injected into the muscle, produces the kind of contraction which the transmitter should produce; (2) that acetylcholine, identified as such, is liberated by impulses at motor nerve endings; and (3) that a suitable dose of eserine causes the muscle to give a short, waning, tetanic response to a single, synchronous volley of nerve impulses Many of these actions of applied acetylcholine on muscle were subsequently examined by measuring the rate of impulse firing in the muscle by Brown (1937). He found that close arterial injection of acetylcholine into a denervated muscle gave rise to a quick initial contraction (Fig. 3.7E) that was accompanied by a burst of action potentials in the muscle. He comments that: The facts presented give incidental support to the suggestion of a concentration of cholinesterase at the mammalian motor nerve endings. That this is the case was shown by Couteaux and Nachmansohn (1940) who determined the concentration of cholinesterase in the middle portion of the guinea pig’s gastrocnemius. This work showed that cholinesterase was indeed found at relatively high levels in the central region of the muscle where the nerve endings are located (Fig. 3.7G). By 1940 acetylcholine was believed to be the agent of transmission between motor nerve and muscle. However it should be noted that Eccles, who together with O’Connor (1938) had just recorded for the first time the electrical signs of transmission at the endplate (the so called ‘endplate potential’; see also Gopfert and Schaefer, 1937), did not come around to this opinion until he worked on motor nerve transmission to muscle with Katz and Kuffler a few years later (Eccles et al., 1941). Indeed Loewi at the time of being awarded the Nobel Prize with Dale in 1936 was not persuaded of chemical transmission at the motor endplate. Nevertheless, most students of transmission accepted by 1940 that the ‘receptive substance’ of Langley could be identified as a receptor for acetylcholine, although the full flavour of the opposition to the concept of chemical transmission from the time of Loewi’s experiments after the Great War up to the 1950s can only be gauged by reading a first hand account of the controversies (for example that of Bacq, 1983). 3.10 The discovery of the physiological action of single acetylcholine receptors The study of the physiological action of single acetylcholine receptors began in 1970 with the discovery by Katz and Miledi (1970b) of membrane noise at the endplate in response to the steady action of acetylcholine from a micropipette. They hypothesised that during such a steady application: the statistical effects of molecular bombardment might be discernible as an increase in membrane noise, superimposed on the maintained average depolarisation. This is what in fact they observed with an intracellular electrode as can be readily ascertained by inspection of Fig. 3.8A. A simple relationship was then used that connects the size of the elementary voltage (a) due to the opening of a single channel as a consequence of acetylcholine binding to a receptor to the average depolarisation (V) and the root mean square value of its fluctuation (E), namely . Applied to the results of Fig. 3.8A this gave a value for the elementary event of 0.29 µV. In 1971 the

62

HISTORY OF THE SYNAPSE

Fig. 3.7. The identification of acetylcholine as the transmitter substance at the somatic neuromuscular junction. (A) Strength-capacity curves on a toad’s gastrocnemius; O, before curare; I, II, III, after a small dose of curare. In the upper right-hand corner, variation of the chronaxie with the time in hours as abscissa. (From Fig. 2 in Lapicque, 1934). (B) Spinal cat. Record from denervated gastrocnemius, 14 days after nerve section. Perfusion with Locke’s solution. Effect of 1 µg of Ach. by close arterial injection. No time calibration is given. (From Fig. 5 in Brown et al., 1936). (C) Effect of venous fluids from gastrocnemius muscles of cat, in 75 percent. Dilution, on contraction of eserinized leech muscle. (A) and (B) from denervated muscle, (A) resting, (B) during stimulation; (C) and (D) from normal muscle, (C) resting, (D) during stimulation. No time or contraction calibration is given. (From Fig. 5 in Dale et al., 1936). (D) Contractions of decerebrated cat gastrocnemius 20 days after nerve section, in response to close arterial injections of (A), 2–5 µg Ach., (B), 50µg carabaminoylcholine and (C), 21 min. later 10µg Ach. (From Fig. 5 in Brown, 1937). (E) (A), (B) and (C), decerebrate cat, contractions of denervated (12 days) gastrocnemius in response to 0.25 µg, 1.0 µg and 2.5 µg Ach. by close arterial injection respectively. (D), spinal cat, contraction of denervated gastrocnemius in response to 20 µg Ach. by ‘distant’ injection into inferior mesenteric artery. (From Fig. 1 in Brown 1937). (F) Spinal cat, 9 days after lumbosacral sympathectomy. Contractions of gastrocnemius in response to maximal shocks to nerve at 10 sec. intervals. (A), arterial injection of 5 pig eserine; (B), 10 min. later; (C), arterial injection of 20 µg eserine. No time or contraction calibration is given. (From Fig. 4 in Bacq & Brown, 1937). (G) Concentration of choline esterase in the middle portion of the interior section of a guinea pig’s gastrocnemius cut in 11 slices of similar thickness and weight. Each horizontal line corresponds to one slice and indicates its weight in % of the total weight. Abscissae: Region from which the tissue was obtained in terms of order of consecutive slices. Point 50 corresponds to the center region where the nerve endings are situated. Ordinate: Choline esterase in mg Ach hydrolyzed per hour by 100 mg of fresh tissue. (From Fig.1 in Couteaux & Nachmansohn, 1940).

same authors determined the approximate time course of the elementary conductance change underlying the elementary event by recording the extracellular voltage fluctuations due to the bombardment of receptors with acetylcholine. In this case they

THE DISCOVERY OF ACETYLCHOLINE AND THE CONCEPT OF RECEPTORS AT SYNAPSES

63

Fig. 3.8. Discovery of the physiological action of single acetylcholine receptors at the somatic neuromuscular junction. (A) Intracellular recordings of membrane potential from end-plate region of frog muscle fibre. In each block, the upper trace was recorded on a low gain d.c. channel (10 mV scale); the lower was simultaneously recorded on a high gain a.c. coupled channel (0.4 mV scale). The records in the upper row are controls (no ACh); the lower row shows membrane noise during ACh application, by diffusion from a micropipette. In the lower records, the increased distance between a.c. and d.c. traces shows upward displacement of the d.c. trace because of ACh-induced depolarization. Two spontaneous m.e.p.p.s are also seen. (From Fig. 1 in Katz & Miledi, 1970b). (B) Power spectrum of intracellularly recorded ACh noise. Temperature 5.5°C. Linear plot of E2/ f, (see text), in relative units against frequency in Hz in the upper part; double-log plot in lower part. (From Fig. 1 in Katz & Miledi, 1971 a). (C) Schematic circuit diagram for current recording from a patch of membrane with an extracellular pipette. VC, Standard two— microelectrode voltage clamp circuit to set locally the membrane potential of the fibre to a fixed value. P, Pipette, fire polished, with 3 to 5 mm diameter opening, containing Ringer’s solution and agonist at concentrations between 2×10−7 and 6×10−5 M. d.c. resistance of the pipette: 2 to 5 Mohms. The pipette tip applied closely on to the muscle fibre within 200 µm of the intracellular clamp electrodes. VG, Virtual ground circuit, using a Function Modules Model 380K operational amplifier and a 500 Mohm feedback resistor to measure membrane current. The amplifier is mounted together with a shielded pipette holder on a motor-driven micromanipulator. V.Bucking potential and test signal for balancing of pipette leakage and measuring pipette resistance. (From Fig. 1 in Neher & Sakmann, 1976). (D) Oscilloscope recording of current through a patch of membrane of approximately 10 mm2. Downward deflection of the trace represents inward current. The pipette contained 2×10−7 M SubCh in Ringer’s solution. The experiment was carried out with a denervated hypersensitive frog cutaneous pectoris (Rana pipiens) muscle in normal frog Ringer’s solution. The record was filtered at a bandwidth of 200 Hz. Membrane potential: −120 mV. Temperature: 8°C. (From Fig. 2 in Neher & Sakmann, 1976).

ascertained the power spectrum of the acetylcholine induced noise, that is the relationship between () and the frequency (Fig. 3.8B). This gave an average time constant of the elementary event of about 1 ms when the event is treated as decaying exponentially and a net charge transfer across the open channel of about 5×104 univalent ions, with a channel conductance of about 10−10 Siemens.

64

HISTORY OF THE SYNAPSE

Direct recording of the electrical signs of individual acetylcholine receptor channels was made by Neher and Sakmann in 1976. They introduced the technique of recording from a small membrane area of the muscle, so as to decrease background noise (Fig. 3.8C). The tip of a glass pipette of 3 to 5 mm diameter, with fire polished edges, was connected up in the circuit shown in Fig. 3.8C after being filled with Ringer’s solution and an acetylcholine receptor agonist, in this case suberyldicholine (SubCh). This pipette was then applied to the surface of a muscle fibre, denervated so as to ensure an abundance of receptors all over the surface of the muscle and subjected to enzyme treatment for the digestion of connective tissue and the basement membrane. Discrete conductance changes like those shown in Fig. 3.8D could only be resolved if the conductance between the pipette interior and the bath decreased by a factor of at least four after the pipette came into contact with the muscle membrane. Inspection of Fig. 3.8D shows that the amplitude of the single channel conductance is about 3.4 pA, giving a channel conductance of 28×10−12 Siemens. This conductance is about the same as that determined by Katz and Miledi (1973) for the agonist SubCh using the noise method of Fig. 3.8A, namely 28.6×10−12 Siemens. 3.11 Conclusion The saga of the concept of the receptor has been followed from its beginnings in the hands of Langley and Ehrlich to the triumph of recording the electrical signs of the opening of a single acetylcholine receptor channel. This work took almost exactly a century to accomplish, from the experiments of Langley in 1874 on pilocarpine and atropine to those of Neher and Sakmann in 1976. The structure of the receptor molecule was also opened up by the discovery in the late 1960s by Lee and his colleagues of toxins that could irreversibly bind to the receptor (Lee & Chang, 1966; Lee, 1972), and so allow for its isolation. But that story takes us too far from the main theme of this chapter, which has been the establishment of the reality of the transmitter receptor.

4 The Discovery of Adrenaline and the Concept of Autoreceptors at Synapses

4.1 Introduction: the discovery of noradrenaline as a transmitter The presser effects of suprarenal extract were first shown to exist by Oliver and Schafer (1895) in 1894, with the first full paper describing their results appearing in 1895 (Fig. 4.2A). In 1899 Abel (1899) obtained an extract from the suprarenal gland, which was a mono-benzoyl derivative, that was less active than crude extracts of the gland, and which he named ‘epinephrin’. Later, Abel (Fig. 4.1 A) produced a crystalline epinephrine-hydrate, which he thought had the same action as the gland extract. However, this was not the case, and it was up to an industrial chemist in 1901, Takamine (1902), to take Abel’s hydrate and purify it to the active principal. This he patented as ‘Adrenalin’, which he had Parke, Davis & Co. market. Presumably Takamine chose the word ‘Adrenalin’ as the patent name because the physiological community in Britain had been using the name ‘adrenaline’ to refer to the active principal of the gland for some years. At this time, 1901, Langley showed that stimulation of the sympathetics results in changes of effectors, which in many cases can be mimicked by application of suprarenal extract (Langley, 1901). Then Elliott in 1904, working in Langley’s laboratory, published his much-quoted note suggesting that adrenaline might be the chemical stimulant liberated on each occasion when the impulse arrives at the periphery (Elliott, 1904a). In that same year Stolz (1904) and Dakin (1905) independently synthesised adrenaline. In 1910 Barger and Dale showed that the effects of sympathetic nerve stimulation are more closely reproduced by the injection of sympathomimetic primary amines than by adrenaline or secondary amines (Barger & Dale, 1910) (Fig. 4.2B). They studied a large number of synthetic amines related to adrenaline and termed their action ‘sympathomimetic’, without establishing that noradrenaline is the sympathomimetic transmitter, although coming very close to doing so. After the First World War, in 1921, Cannon (Fig. 4.1B) together with Uridil reported that stimulation of the sympathetic hepatic nerves in mammals releases an adrenaline-like substance (which they called sympathin) that increases blood pressure and heart rate (Cannon & Uridil, 1921) (Fig. 4.2C). However adrenaline when injected into the body was shown to accelerate heart rate and simultaneously dilate some vascular beds while constricting others; on the other hand sympathin accelerated the heart without dilation of vascular beds and produced an increase in total peripheral resistance and diastolic blood pressure. Again the possibility, following the observations of Barger and Dale (1910), that sympathin might actually be demythelated adrenaline (noradrenaline) was not pursued, although Bacq (1974) repeatedly advanced this idea. The identification of the sympathomimetic substance as noradrenaline in mammals was further delayed by Loewi’s observation in 1921 that an accelerator substance is released from the toad’s heart on stimulation of the sympathetic nerves (at least those in the vagus that predominate over the parasympathetics in summer; Fig. 4.2D). He concluded that this might be adrenaline, which it is in the case of toads (Loewi, 1921). It took another twenty-five years before the identity of the sympathomimetic substance in mammals was established by von Euler in 1946. He showed that highly purified extracts of sympathetic nerves and effector organs resemble noradrenaline by all criteria available (Fig. 4.2E) and proposed noradrenaline as the sympathetic transmitter (Euler, 1946). Subsequently Ahlquist in 1948 examined the effects of adrenaline and noradrenaline as well as isoproterenol on a variety of target tissues (Fig. 4.2F). He concluded that the difference in action of these catecholamines could be explained by the presence of two types of receptor for the catecholamines, namely alpha and beta (Ahlquist, 1948). 4.2 Early observations leading to the idea of autoreceptors In 1931 Cannon and Bacq reported an increase in the sympathin in the blood stream, as determined by its effects on the heart, following stimulation of the sympathetic nerves to the colon or to the hind quarters (Cannon & Bacq, 1931). Subsequently Jang showed that the responses of several organs to sympathetic nerve stimulation could be potentiated by appropriate concentrations of the adrenergic receptor blocking drugs 933F, yohimbine and ergotoxine (Jang, 1940). That such blocking drugs may enhance the effects of sympathetic nerve stimulation was further elaborated by Holzbauer and Vogt in 1954, who

66

HISTORY OF THE SYNAPSE

Fig. 4.1. The discovery of adrenaline. (A) John J.Abel (1857–1938). Taken from Dale, 1939. (B) Walter Bradford Cannon (1871–1945). Taken from Finger, 1994.

showed that the responses of the dibenzyline pretreated uterus to exogenous adrenaline were greatly enhanced over control untreated preparations (Holzbauer & Vogt, 1956). However, it was not until 1957 that a thorough analysis was made of this phenomenon whereby alpha adrenergic blocking drugs potentiate the overflow of noradrenaline and the contractile response of organs to sympathetic nerve stimulation. Brown and Gillespie stimulated the sympathetic nerves to the spleen of the cat and determined the amount of noradrenaline in the venous blood during different frequencies of stimulation and in the presence of different alpha adrenergic blocking drugs (Brown & Gillespie, 1957). Following 200 impulses, noradrenaline output could just be detected at 10 Hz, with maximum output occurring at 30 Hz (Fig. 4.3A). They showed that the low level of output of noradrenaline at 10 Hz was not due to the breakdown of the catecholamine by monoamine oxidase. Furthermore the level at 10 Hz could be increased to the high level found at 30 Hz if adrenergic blocking drugs such as dibenamine were present in the perfusion fluid (Fig. 4.3A). This led them naturally to the conclusion that: The constancy of the output at frequencies between 1 and 30/sec after dibenamine would be explicable on the ground that the dibenamine had prevented the destruction of the transmitter. The known effect of dibenamine is to block the tissue receptors for noradrenaline, and we must conclude therefore that combination with the receptors is a necessary prelude to the destruction and removal of liberated noradrenaline. In 1961, an article by Koelle in Nature dramatically challenged the conceptual framework in which the mechanisms of synaptic transmission might be considered. He suggested, on the basis of observations showing that the threshold dose of intra-arterially injected carbamylcholine into the superior cervical ganglion was greatly elevated following denervation of the ganglion, that: …acetylcholine mobilised by a conducted presynaptic impulse may act at two sites: postsynaptically, as outline above, but prior to that at the presynaptic terminal itself, to bring about the release from the terminal of sufficient additional acetylcholine or of some other neurohumoral agent to initiate the characteristic postsynaptic effect (Koelle, 1961). Other neurohumoral agents might include noradrenaline, for at that time a theory was in currency that cholinergic mechanisms might be involved in the release of catecholamines (Burn & Rand, 1959). Fig. 4.3B summarises Koelle’s conjectures: acetylcholine is released from vesicles (open circles) and acts back on the nerve terminal to release more acetylcholine containing vesicles (open circles) in the case of cholinergic synapses; in the case of adrenergic nerve terminals acetylcholine acts back on the terminal to release noradrenaline (filled circles). It is interesting that there is now considerable evidence to support the idea that acetylcholine released at motor-nerve terminals during a train of impulses does act back to accelerate the release of further acetylcholine (Wessler, 1992). In the case of adrenergic nerve terminals, there appear to be both nicotinic and muscarinic receptors that can modulate catecholamine release, either increasing or decreasing the release depending on the organ being considered, although it is not clear if these receptors have a physiological role (Fredholm, 1995). Koelle’s idea that acetylcholine acts as a ubiquitous intermediary in synaptic transmission, independent of the identity

THE DISCOVERY OF ADRENALINE AND THE CONCEPT OF AUTORECEPTORS AT SYNAPSES

67

Fig. 4.2. Identification of adrenaline and noradrenaline as transmitter substances released by sympathetic nerves. (A) Suprarenal extract exerts presser effects. Effect of injecting extract of 0.2 gram of dog-suprarenal into a 16 kilo dog. A, auricle; B, ventricle; C, femoral artery; D, injection of suprarenal (watery extract), the line D also indicates 0 pressure; E, time in 0.5 sec. Note delayed inhibition. Taken from Oliver & Schafer, 1895. (B) Injection of sympathomimetic primary amines closely reproduces effects of sympathetic nerve stimulation. At A, 2 c.c of N/100_ phenylpropylamine injected. At B, 2 c.c. of N/100 phenylethylamine injected. At C 2 c.c. of N/100 p. hydroxyphenylethylamine injected. Taken from Barger & Dale, 1910. (C) Evidence that stimulation of the sympathetic hepatic nerves releases an adrenaline-like substance. Average increments of heart rate in 5 second intervals in 5 cases of hepatic-nerve stimulation for 30 seconds (continuous line). Also average increments of heart rate in the same intervals from reflex adrenal stimulation. Taken from Cannon & Uridil, 1921. (D) Evidence that the perfusate of a toads heart collected during a period of stimulation of the vagus nerve when the sympathetic component dominates gives rise to an increased ionotropic effect. 1. Ringer from a 20 min normal period. 2. Ringer from a 20 min accelerans-stimulation period. 3. Ringer. Taken from Loewi, 1921. (E) Effect of purified extracts of sympathetic nerves resemble effects of noradrenaline. Blood pressure of the chloralosed cat is shown. 1, 0. 06 g splenic nerves, cattle, extract treated with Fuller’s earth. 2, 0.5 µg nor-adrenaline-HCI. 3, 1 mg dihydroxynor-ephedrine. 4, 1 µg adrenaline. Then between 4 and 5, an amount of 0.6 mg dihydro-ergotamine/kg i.v. 5, 0.25 g splenic nerves. 6, 2 µg nor-adrenaline-HCI. 7, 4 µg dihydroxy-nor-ephedrine. 8, 3 µg adrenaline. Time 30 seconds. Taken from Euler, 1946. (F) Comparative action of amines on mean arterial pressure of cats, dogs, and rabbits. Ordinates: pressure in mm. Hg; abscissae: time marks at 10 sec. intervals. The amines were injected intravenously and the key in the box indicated the amines injected. Cats: average results for 6 cats with dosage of 0.1 cc. M/1000 solution per kgm. The box gives the average results for 3 cats after dibenamine, dosage, 0.1 cc per kgm. of the concentrations shown. Dogs: average results from 12 determinations in 6 dogs with dosage of 0.05 cc. of a M/1000 solution per kgm. Rabbits: average results from 10 determinations in 6 rabbits with dosage of 0.1 cc. of a M/1000 per kgm solution. Taken from Ahlquist, 1948.

of the transmitter that finally exert an effect on the postsynaptic receptors (Fig. 4.3B), was not supported by further work, but his challenging concepts did open up the possibility that the action of adrenergic and cholinergic blocking drugs might be on the nerve terminal as well as on the postsynaptic membrane.

68

HISTORY OF THE SYNAPSE

4.3 Direct experimental evidence for autoreceptors The year 1971 saw the publication of a large number of papers on different peripheral and central organs that established the existence of presynaptic receptors that could control transmitter release during a short train of impulses in a nerve terminal. The work of Brown and Gillespie (Fig. 4.3A) was not reinterpreted in the light of Koelle’s conjectures (Fig. 4.3B) as evidence for a presynaptic effect of noradrenaline on its own release as Koelle had emphasised that it was always acetylcholine that exerted this effect. Furthermore, with the subsequent discovery in the 1960s of the catecholamine uptake pumps in the terminals of adrenergic neurones (Iversen, 1967); other interpretations of the observations of Brown and Gillespie were popular. Chief amongst these was that alpha adrenoceptor blocking agents, such as phenoxybenzamine, increased overflow of noradrenaline as a consequence of blocking neuronal uptake of the catecholamine. In addition, such blocking agents maintain blood flow in the spleen during sympathetic nerve stimulation by preventing vasoconstriction; this alone would be expected to enhance the diffusion of noradrenaline from the nerve terminals after it had been liberated and so increase overflow. A direct test of the blood flow argument was made by Kirpekar and Puig in 1971. They perfused spleens with a constant outflow pump during the application of phenoxybenzamine or phentolamine and then stopped the flow for different times to test whether the affects of the drug on overflow of noradrenaline could be attributed to its preventing vasoconstriction. Their argument was that if the increase in transmitter output after phentolamine treatment was mainly due to its action in blocking vasoconstriction and thus improving flow rates, then restriction on flow should cause a much greater percentage reduction in phentolamine treated than in untreated spleens. Fig. 4.3C shows that this was not the case, as restricting the blood flow for 30 s resulted in the removal of about the same percentage of noradrenaline in the control spleens as in those treated with phentolamine. These results lead to the conclusion that the alpha receptor blocking drugs do not enhance overflow as a consequence of blocking vasoconstriction (Kirpekar & Pulg, 1971). Kirpekar and Puig observed that noradrenaline released during nerve stimulation appeared to be removed during the course of the stimulation whereas Blakeley and Brown (1964) showed that the uptake of noradrenaline infused into the spleen was not affected by nerve stimulation. In order to reconcile these observations Kirpekar and Puig suggested that: …the noradrenaline released by nerve stimulation acts on these alpha sites of the presynaptic membrane to inhibits its own release. In the presence of adrenergic blocking agents the transmitter output would be enhanced, since the inhibitory alpha sites on the nerve terminals are blocked. These are probably not the same sites through which noradrenaline is taken up into the nerve endings (Kirpekar & Pulg, 1971). Farnebo and Hamberger independently came to the same conclusion in 1971 as a consequence of their studies of the overflow of noradrenaline from nerve-stimulated rat irides. In this work alpha adrenoceptor antagonists such as phenoxybenzamine and phentolamine, used at concentrations that do not affect amine uptake into nerve terminals, increased the overflow of the catecholamine (Fig. 4.3D), whereas alpha adrenoceptor agonists like clonidine decreased the overflow. These authors carefully excluded the possibility that these agents changed the rate of metabolism of the amine by, for example, monoamine oxidase or catechol-o-methyltransferase. They suggested that a possible explanation for their results was inhibition of noradrenaline release via alpha-adrenoceptors on the nerve terminal (Farnebo & Hamberger, 1971). Starke in the same year reached much the same conclusion from his studies of the effects of alpha-adrenoceptor agonists on the overflow of noradrenaline from the isolated perfused rabbit heart, observations that were particularly interesting because of the lack of postsynaptic alpha receptors in the heart. Phenylephrine produced a dose-dependent inhibition of noradrenaline released by sympathetic nerve stimulation (Fig. 4.3E). This lead Starke to conclude that: …adrenergic nerve terminals are endowed with structures related to the alpha adrenoceptive sites of effector cells. On reaction with the alpha stimulants, e.g. with liberated noradrenaline, these neuronal alpha receptors mediate the inhibition of noradrenaline liberation; in the presence of the alpha blockers, this restriction is attenuated (Starke, 1971). This work on the heart was extended several years later by Langer, Adler-Graschinsky and Giorgi in 1977, in an investigation that took advantage of the absence of postsynaptic alpha receptors to determine if blockade of prejunctional alpha receptors on the sympathetic nerve terminals would lead to an expected increase in the overflow of noradrenaline accompanied in this instance by an increase in the response of the organ. They showed that an alpha-receptor antagonist (phentolamine) could produce both a significant increase in the response of guinea-pig atria to nerve stimulation (see Fig. 4.4F), and at the same time enhance the overflow of noradrenaline fourfold (Langer et al., 1977). Interestingly enough at this time (1971) evidence was also growing for autoreceptors at cholinergic and even GABAergic synapses in the central nervous system. Polak showed that atropine could exert a stimulating effect on the release of acetylcholine due to incubating cortical slices of rat brain in high potassium solutions (Fig. 4.3F). As this stimulating action of atropine could be inhibited by adding muscarinic agonists such

THE DISCOVERY OF ADRENALINE AND THE CONCEPT OF AUTORECEPTORS AT SYNAPSES

69

Fig. 4.3. The existence of receptors on nerve terminals. (A) Discovery that adrenergic blocking drugs enhance the overflow of noradrenaline during stimulation of sympathetic nerves to the spleen. Individual results for the output/stimulus of noradrenaline after dibenamine, superimposed on the graph showing the average output in untreated animals (solid line). Taken from Brown & Gillespie, 1957. (B) The idea, due to Koelle, of nerve terminal autoreceptors and of muscle autoreceptors. Upper diagram shows released noradrenaline acting on a target smooth muscle cell as well as on its own terminal to release more noradrenaline. Note that the smooth muscle cells themselves are shown releasing transmitter onto themselves. Lower diagram shows released acetylcholine acting on a postsynaptic neuron as well as back onto the terminal from which it was released. The autoreceptors shown in both the adrenergic and cholinergic synapses accelerate release. Taken from Koelle, 1961. (C) Evidence that the increased overflow of noradrenaline from the spleen on nerve stimulation in the presence of an adrenergic receptor blocker is not due to decreased vasoconstriction enhancing overflow. Effect of flow-stop on noradrenaline (NA) overflow expressed as a percentage of the control output without flow stop is shown. Left and right hand parts of the figure show experiments in normal spleens and spleens treated with phentolamine. Nerves were stimulated at 10 Hz for 20 seconds. Flow was stopped for the duration of stimulation +30, 60 or 120 seconds. Control output was 208±60 ng in normal spleens. In spleens treated with phentolamine control output without flowstop was 499 ±208 ng. Taken from Kirpekar Pulg, 1971. (D) Effect of alpha adrenoceptor antagonists on the overflow of noradrenaline from nerve-stimulated irides. Efflux of tritium from irides of untreated rats superfused with drug-free buffer solution (left) or buffer solution containing phentolamine 10−6M (right). The irides were preincubated with [3H]-NA, 10−7M, for 30 min before superfusion and stimulation. Each point represents mean±S.E.M. of four experiments. Taken from Farnebo & Hamberger, 1971. (E) Alpha adrenoceptor agonists decrease the noradrenaline overflow from the heart. Influence of phenylephrine on the output of noradrenaline during sympathetic nerve stimulation from isolated rabbit hearts. Means±S.E.M, n=number of experiments. Significance of difference from controls was evaluated by Student’s t-test. Taken from Starke, 1971. (F) Muscarinic antagonists can enhance the release of acetylcholine. ACh content (expressed as nmol/g initial weight of wet tissue) of cortical slices and incubation media after 1 h incubation with and without atropine and oxotremorine. (Hatched histogram bars), ACh extracted from the tissue slices; (open histogram bars), ACh in the incubation medium. The figures on top of the columns give the number of observations from which the mean values were obtained. Note that the concentration of atropine in A is ten times less than in B as indicated in the abscissa. The concentrations of oxotremorine are different for each histogram bar as indicated in the abscissa. Taken from Polak, 1971. (G) Effect of alpha-adrenoceptor antagonists on the release of transmitter by sympathetic nerves as measured by excitatory junction potentials. Effect of phenoxybenzamine (10 µg ml−1) on the amplitude of the excitatory junction potential in the mouse vas deferens as stimulation of the sympathetic nerves at 10 Hz. (a) control (b) in the presence of phenoxybenzamine. Taken from Bennett, 1973.

70

HISTORY OF THE SYNAPSE

Fig. 4.4. Identification of presynaptic autoreceptors as different from postsynaptic receptors. (A) The autoreceptor hypothesis. Negative feedback mechanism for norepinephrine released by nerve stimulation. Response (R) of the effector cell mediated through alpha-receptors ( ), and through beta-receptors ( ). Note that, irrespective of the nature of the postsynaptic adrenergic receptor, the presynaptic receptor involved in the regulation of transmitter release is an alpha-receptor. Taken from Langer, 1974. (B) The result of removing the target organ or its nerve supply on the overflow of noradrenaline on exposure to an alpha receptor antagonist. Effects of phentolamine on the overflow of 3H-transmitter induced by 60 mM K+ in the rat submaxillary gland. Ordinates a and c: Fractional release: total 3H-released divided by that remaining in the tissue at the moment of the addition of K+. For estimation of fractional release the spontaneous outflow was subtracted from the total overflow of radioactivity elicited by exposure to 60 mM K+. Ordinates B and D: Ratio of the fractional release of radioactivity by the second potassium depolarization (S2) to that induced by the first one (S1). Abscissa: S1 and S2 indicate the two depolarization periods induced by exposure to 60 mM K+ with an interval of 37.5 min between the two stimulation periods. (Open histogram bars) normal glands; (Vertical striped histogram bars) atrophied glands (15 days after duct ligation). Where indicated phentolamine (Phent), 3.1 µM was added 30 min before S2. Shown between parentheses are the number of experiments per group. Shown are mean values ±S.E.M. * P