FLOYD
21.
22.
23.
24.
25.
26.
Stein L, Wise CD, Belluzzi JD: Effects of benzodiazepines on central serotonergic mechanisms, in Mechanism of Action of Benzodiazepines. Edited by Costa E, Greengard P. New York, Raven Press, 1975, pp 29-44 Baumgarten HG, Bjorklund A, Lachenmayer L, et a!: Longlasting selective depletion of brain serotonin by 5,6-dihydroxytryptamine. Acta Physiol Scand Supplement 373, 1971 , pp 1-15 Aprison MH, Ferster CB: Neurochemical correlates of behavior: II. Correlation of brain monoamine oxidase activity with behavioral changes after iproniazid and 5-hydroxytryptophan. J Neurochem 6:350-357, 1961 Goldberg ME, Manian AA, Efron DH: A comparative study of certain pharmacological responses following acute and chronic administration of chlordiazepoxide. Life Sci 6:481-491, 1967 Margules DL, Stein L: Increase of “antianxiety” activity and tolerance of behavioral depression during chronic administration of oxazepam. Psychopharmacologia 13:74-80, 1968 Young AB, Zukin SR, Snyder SH: Interaction of benzodiazepines with central nervous system glycine receptors: possible mechanism of action. Proc Natl Acad Sci USA 71:2246-2250, 1974
Neural
Mechanisms
BY FLOYD
E. BLOOM,
APPENDIX 1 Evidence Linking 1 . Benzodiazepines
2. Benzodiazepine a.
5-HT
b. c.
5-HT 5-HT mine).
reduce
to Brain Serotonin 5-HT
BLOOM
(5-HT)
turnover.
antipunishment effects antagonists (methysergide,
are mimicked cinanserin,
by: BOL-
148)
3.
4.
synthesis nerve
inhibition (PCPA) terminal damage (5,6-dihydroxytrypta-
Punishment effects are intensified by: a. 5-HT agonists (a-methyltryptamine) b. 5-HT precursor (5-hydroxytryptophan plus hibitor) c. 5-HT (administered in lateral ventricle) d. Carbachol activation of 5-HT cells in dorsal cleus. Benzodiazepine potentiated
of Benzodiazepine
effects are by norepinephrine.
antagonized
by
5-HT
MAO
in-
raphe
nu-
but
are
Actions
M.D.
DESPITE THE WELL-DOCUMENTED clinical usefulness of the benzodiazepines in the treatment of anxiety, seizures, and muscular spasms, the neuronal Sites and mechanisms of action from which these therapeutic gains derive remain unknown. Solving the enigma of how and where these drugs act selectively on discrete neurons to alter whole animal or human behavior in specific ways requires detailed descriptions of benzodiazepine actions on many levels of research. The purpose of this report is to examine critically the present status of cell-specific effects of benzodiazepines in the brain and to suggest some areas worthy of continued scrutiny as research progresses.
Currentexperimental paradigms emphasize neurotransmitter-specific interactions to explain the behavioral effects ofbenzodiazepines. According to this approach the broad range ofeffects observed suggests the involvement ofseveral transmitter systems without rigorously establishing that any single transmitter system or physiological synaptic function is either necessary or sufficient to express all benzodiazepine actions. Among the effects that occur, potentiation ofamino acid-mediated presynaptic inhibition in the spinal cord and postsynaptic inhibitions elsewhere in the brain are attractive testable hypotheses. However, direct physiological evidence that benzodiazepines simulate yaminobutyric acid (GABA) or other amino acids specifically and exclusively is needed to corroborate this view.
ORGANIZATION
Revised version of a paper presented the American Psychiatric Association, 14, 1976.
at the 129th annual meeting Miami Beach, Fla. , May
OF
NEURON-DRUG
INTERACTIONS
General
Dr. Bloom is Professor and for Behavioral Neurobiology,
Benzodiazepines
E.
of 10-
Director, Arthur Vining Davis Center Salk Institute, La Jolla, Calif. 92037.
Drug
Actions
The awesome structural and functional complexity of the brain handicaps the analysis of specific drug effects. The situation is further compounded by the rapidity of cellular interactions and by the lack of understanding of many fundamental details underlying such interactions. One simplifying approach is to organize Am J Psychiatry
134:6,
June
1977
669
NEURAL
MECHANISMS
the possible modes of drug interaction with neurons for purposes of detecting general events for further analysis (see figure 1). As a first step, some benzodiazepine actions are generalizable to all nerve cells, e.g. effects on general energy metabolism (2) and maintenance of membrane integrity and transmembrane ionic equilibria (3). For the most part, actions affecting common biological properties of all neurons occur at relatively high dose levels in vitro and are not the drug actions we are now trying to comprehend. ,
Transmitter-Specific
Drug
Actions
Chemical dynamics. A second category of drug actions pertains to selective effects on neurons grouped together by their specific common use of the same chemical synaptic transmitter. Here the potential molecular mechanisms for drug specificity are more easily envisioned because of the numerous points within the life cycle of transmitters that have been uncovered through biochemical pharmacology (see figure 1). Although biochemically detected changes in levels and turnovers of several biogenic amines have been reported (4-13), the data available are insufficient to account for the unique patterns of behavioral responses because similar biochemical and histochemical effects have been reported after other sedative drugs, such as ethanol and barbiturates (5), that clearly differ behaviorally and clinically from the benzodiazepines. Synaptic receptors. Drugs may also act directly at postsynaptic receptor sites. Binding studies of transmitter receptors with benzodiazepines are reviewed by Snyder and associates in this Special Section of the Journal. Neuronal
Integration
Mechanisms
Another mode of categorizing drug-neuron interactions can also be applied to studies of benzodiazepines. Under this category are included studies in which parenterally injected or topically applied drugs were used to influence evoked responses or single unit activity in relationships in which the chemical nature of the synaptic interactions being tested remained undefined. Ifbroadly and consistently applied, such test paradigms could still be used to establish dose-response and time-response curves by which the most rapid and most sensitive effects of the drug on key target cells could be discerned. The available evidence from such experiments is restricted now to three brain regions: the cerebellum, the spinal cord, and the limbic system, especially the hippocampal formation. Very low doses of benzodiazepines have also been reported to antagonize the production of posttetanic potentiation (VFP) in the sympathetic ganglion of the bullfrog in vitro (14); however, since these same drugs do not interfere with FTP in the neuromuscular junction or at cholinergic receptors in the spinal cord (see references in [15] and [16]), this interesting observa670
Am
J Psychiatry
134:6,
June
1977
tion ther
in the sympathetic ganglion cannot be carried furconceptually at this time. Cerebellum Since the cerebellum is generally regarded as inhibitory to the motor cortex, the enhancement of cerebellar activity by benzodiazepines could implement an anticonvulsant action. Purkinje cells in the cerebellar cortex inhibit the neurons of the deep cerebellar nuclei, which project to the cerebral cortex. Therefore, consistent facilitation ofPurkinje cells should inhibit the deep nuclei and disinhibit (i.e. facilitate by removing tonic inhibition) the cortex. Although facilitation of Purkinje cell firing was observed to occur when cats were administered diazepam (17), the opposite effect (which would logically be more “anticonvulsant”) has been reported for the effects of diazepam on both cats and rats (18). However, a direct antagonistic action against y-aminobutyric acid (GABA) here (as recently reported by Steiner and Felix [19]) would, ofcourse, produce the opposite effect. The findings of Suria, Costa, and Mao and associates (14, 15, 20) that diazepam is extremely potent in lowering cyclic 3 ‘-5 ‘guanosine monophosphate (GMP) levels in the cerebellum may relate to the anticonvulsant effects of benzodiaiepines, since several convulsants selectively elevate cerebellar cyclic GMP. Spinal cord. The advantages of studying drug actions in the spinal cord are that one can know with some certainty what is being stimulated and what is being recorded. By recording from and stimulating various combinations of the dorsal or ventral roots of the isolated spinal cord, accurate dose-response estimates of effects at definable endpoints can be obtained. These technical advantages have been applied with considerable impact to the effects of benzodiazepines (16). Because of the functional significance of depolarizing the presynaptic primary afferent fiber, a great deal of emphasis has been placed on changes in the primary afferent terminals as a result of stimulation of an adjacent dorsal root or ipsilateral ventral root. This presynaptic sensory fiber is simultaneously postsynaptic to other spinal interneurons (see figure 1). Activation of these interneurons depolarizeS the primary afferent terminal and decreases the release of excitatory sensory transmitter onto the motoneurons. Therefore, a standard stimulating volley applied to a dorsal root whose primary afferent terminals are depolarized by appropriate spinal interneurons will produce less excitation of the motoneurons. The motoneurons are therefore inhibited presynaptically but without any of the membrane potential shifts associated with conventional postsynaptic inhibition. Several groups of researchers (I 1 18) have reported that benzodiazepines enhance the dorsal root reflex elicited either by ventral root stimulation (depolarizing dorsal root terminals through a set of spinal interneurons believed to release either /3-alanine or taurine [21]) or by adjacent dorsal root stimulation (depolarizing the primary afferent terminals through spinal interneurons releasing GABA [21]). .
,
,
FLOYD
FIGURE 1 Potential Actions of Benzodiazepines Functions*
on General or Specific
Neuronal
E.
BLOOM
have been taken to support the view (15, 18) that GABA-mediated sites of presynaptic inhibition are the major site of selective effects (i.e. an enhancement) produced by benzodiazepines. These views are further strengthened by the related observations that pharmacological elevation of spinal cord GABA levels (through blockade of catabolism) enhances the effects of benzodiazepines on the dorsal root depolarization but blockade of GABA synthesis prevents the enhancement of dorsal root depolarization (1 1 18). On the other hand, chemical axoaxonic presynaptic inhibitory mechanisms cannot account for the effects of benzodiazepines on cerebellum or on hippocampus for two reasons: 1) such Synaptic structures have not been detected there and 2) GABA clearly mediates postsynaptic inhibitions in these brain regions (23). Moreover, although benzodiazepines potentiate dorsal root depolarization, this response is not necessarily an exclusive GABA-mediated event; either taurine or /3alanine (in the frog [21] and the cat [23]) can mediate similar effects, and barbiturates, whose effects are also reversed by drugs that antagonize GABA (24), can enhance the depolarization response as well as the benzodiazepines can. Furthermore, the glycine antagonist strychnine, which is used as a tool for the detection of the glycine-binding protein (25), will also antagonize the ability of both taurine and /3-alanine to depolarize the dorsal root fiber (21) and will partially antagonize these amino acids at postsynaptic receptors elsewhere in the forebrain (23). This crossover of antagonist effects between amino acids suggests that some caution must be used in interpreting the uniformity of transmiller-specific sites at which these convulsant antagofists can reverse the effects of benzodiazepines. Limbic system. The limbic system has figured prominently in occasional abortive attempts to characterize the anxiolytic effects of the benzodiazepines because of the suspicions that this brain formation has affective functions (26). When chronically implanted wire electrodes were used to record from units in the hippocampus, thalamic reticular formation, or anterior hypothalamus of unrestrained rats over long periods of time, only hippocampal neurons responded specifically to benzodiazepines (27). Under these experimental conditions, hippocampal units consistently showed a dose- and time-dependent reduction of unit discharge rates. In other studies, the poorly understood polysynaptic connections between the amygdaloid complex and the hippocampus were also shown to be sensitive to low doses of benzodiazepine (28). The molecular nature and behavioral significance of this otherwise isolated effect therefore remain unknown. Like the actions of benzodiazepines on cerebellar cyclic nucleotides described above, a muscarinic cholinergic mechanism is known to excite hippocampal neurons (29); such receptors are also associated with cyclic GMP synthesis in other neurons (30). Extended analysis of benzodiazepine actions on cyclic GMP levels in the limbic system might also be informative with respect to anxiolytic actions. ,
,
#{149}The enumerated steps indicate possible points in the biological cycle of a central neuron at which benzodiazepines and other psychoactive drugs might interact. Step 1: axoplasmic transport via microtubules. Step 2: electrical cxcitable components of the axonal and perikaryonal membrane. Step 3: organdes and enzymes present within preterminal axons, synaptic boutons, and some dendritic processes for the synthesis, storage, catabolism, and release of the transmitter as well as for the active reuptake and restorage of the transmitter after release. Step 4: glia-neuronal interaction processes, which are
presently
not well characterized but might assist in policing the extrasynaptic to maintain minimal concentrations of active transmitter molecules. Step 5: postsynaptic receptors that trigger events within the postsynaptic cell initiated by the release of transmitter. Step 6: intracellular neuronal organdIes that may mediate the postsynaptic response in some of the subsequent metaspaces
bolic
aspects,
ar interactions bly sensitive
such
as ribosomal
protein
synthesis.
Step
7: cytoplasmic-nucle-
that regulate the expression ofgenetic information that is possito transmitter-regulated events. Step 8: intrasynaptic events of desensitization, habituation, and other physiological modula-
sensitization, tions of transsynaptic function. Step potentials from various convergent
9: integration of simultaneous synaptic excitatory and inhibitory transmitters.
Step 10: discharge ofthe target cell and propagation ofa signal down its axon. Step 1 1 : a presynaptic element contacts one of the afferent terminals onto the target cell and regulates transmitter release presynaptically. synaptic autoreceptors by which the release of a transmiuer terminal may be modified by the transmitter it has previously data for this figure are from Cooper and associates (I).
Step 12: prefrom a synaptic released. The
The ability of benzodiazepines to potentiate the depolarizing response of primary afferent terminals can be antagonized to a considerable degree by picrotoxin and by bicuculline, two convulsants that antagonize postsynaptic inhibitory actions ofGABA, as well as by some of the actions of f3-alanine and taurine (see [16] and [21] for discussion). Similar observations on sites of presumed amino acid-mediated presynaptic inhibition in the cuneate nucleus (22) and postsynaptic inhibition in the striatonigral pathway (18) can also be enhanced by benzodiazepines. These data collectively
Am
J Psychiatry
134:6,
June
1977
671
NEURAL
MECHANISMS
J Pharm
CONCLUSIONS
Almost all of the benzodiazepine actions currently available for review are probably compatible with an umbrella theory focused on GABA-mediated potentiation of pre- and postsynaptic inhibition (15, 18), including the moderate decreases in biochemical indices of activity in catecholamine, acetylcholine, and Serotonin neurons. However, the nonspecificity of the drugs used to manipulate tissue GABA levels, without comparably sensitive measurements of other natural amino acids, weakens this approach. The antagonistic actions reported in the cerebellum maintain the association with GABA in the opposite direction (19). The continual emergence of newer transmitter materials, such as substance P (3 1) and the morphine-like peptides (32), suggests the need for prudence and caution before wholehearted acceptance of a monolithic explanation of benzodiazepine actions through a single transmitter substance. One need only recall how the specific transmitter-related theories of behavioral depression induced by reserpine were forced into continual revisions by the discoveries that not only serotonin but norepinephrine and dopamine as well were profoundly depleted (1) to imagine how the present interpretive status of benzodiazepine-neuronal interactions may be equally shattered by further studies on substance P and related peptides. For example, recent data suggest that p-chlorophenyl GABA (lioreseal), a drug studied as a GABA-synergist (33), is equally well characterized as a substance P antagonist (34).
REFERENCES
I. Cooper JR. Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology. New York, Oxford University Press, 1970, pp 191-202 2. David LF, Gatz EE, Jones JR: Effects of chlordiazepoxide and diazepam on respiration and oxidative phosphorylation in rat brain mitochondria. Biochem Pharmacol 20:1883-1887, 1971 3. Ueda I, Wada T, Ballinger CM: Sodium and potassium-activated ATPase of beef brain-effects of some tranquilizers. Biochem Pharmacol 20:1697-1700, 1971 4. Fuxe K, Agnati LF, Bolme P, et al: The possible role of GABA mechanisms in the action of benzodiazepines on central catecholamine neurons. Adv Biochem Psychopharmacol 14:45-62,
12.
Corrodi H, Fuxe stress and central
K, Lindbrink catecholamine
P. et al: Minor neurons. Brain
tranquilizers, Res 29:1-16,
15.
take
SB, Renyi AL: In vivo inhibition of H3-noradrenaline by mouse brain slices in vitro. J Pharm Pharmacol
7.
R: The
effect
of chlordiazepoxide,
9.
10.
672
Chase TN, Katz RI, Kopin IJ: Effect intracistemally injected serotonin-C14. 9:103-108, 1970 Saad SF: Effect of diazepam on GABA
Am
J Psychiatry
134:6,
June
of diazepam on Neuropharmacology
1977
of mouse
Julien
E: Action
A, Bloom
of the
site
RM:
FE:
Ladinsky
21.
Action
Adv
of diazepam,
of
of
the
Biochem
ben-
Psycho-
chlordiazepoxide, and histamine in Edited by GaratRaven Press, 1973,
dibutyryl
cGMP,
in the electrophysiological
of benzodiazepines.
14:93-102, Cerebellar
in the
zodiazepines 260:346-347,
20.
H:
system.
Problems
of action
and
on
Adv
anal-
Biochem
Psy-
1975 involvement
central
Psychopharmacol 19. Steiner FA, Felix
actions
in the
anti-epileptic
of benzodiazepines.
14:131-152, 1975 D: Antagonistic effects vestibular
action
of
11:683-691, 1972 H, et al: Possible involvement
and
Adv
of GABA
cerebellar
neurones.
of
Biochem
and
ben-
Nature
1976 Mao CC, Guidotti A, Costa E: Inhibition by diazepam of the tremor and the increase in cerebellar cyclic GMP elicited by harmaline. Brain Res 83:516-519, 1975 Barker JL, Nicoll RA, Padjen A: Studies on convulsants in the isolated
frog
spinal
cord:
II.
Effects
on root
potentials.
J Physiol
(Lond) 245:537-548, 1975 22. Davidson N, Southwick CA: Amino acids and presynaptic inhibition in the rat cuneate nucleus. J Physiol (Lond) 2 19:689-708, 1971 23. Curtis DR. Johnston OAR: Amino acid transmitters in the mammalian central nervous system. Rev Physiol Biochem Pharmacol 69:97-188, 1974 24. Nicoll RA: Presynaptic action of barbiturates in frog spinal cord. Proc NatI Acad Sci USA 72:1460-1463, 1975 25. Young A, Zukin SR. Snyder SH: Interaction of benzodiazepines with central nervous system glycine receptors: possible mechanism of actions. Proc Nail Acad Sci USA 71:2246-2250, 1974 26. Douglas Ri: The development of hippocampal function: implications for theory and for therapy, in The Hippocampus, vol 2: Neurophysiology
27.
28.
up-
30.
diaze-
31.
brain.
S.
on spinal Naunyn
on presynaptic nerve terminals in bullfrog sympathetic Brain Res 87:102-106, 1975 E, Guidotti A, Mao CC, et al: New concepts on the mechof action of benzodiazepines. Life Sci 17:167-186, 1975
GABA
32.
33.
of
and
Behavior.
Edited
by
Isaacson
RL,
Pm-
bram KH. New York, Plenum Press, 1975, pp 327-362 Olds ME, Olds J: Effects of anxiety-relieving drugs on unit discharges in hippocampus, reticular mid-brain, and pre-optic area in the freely moving rat. International Journal of Neuropharmacology 8:87-103, 1969 Jalfre M, Monachon MA, Haefely W: Effects on the amygdalohippocampal evoked potential in the cat of four benzodiazepines and some other psychotropic drugs. Naunyn Schmiedebergs Arch Pharmacol 270: 180-191 , 1971 Teitelbaum H, Lee JF, Johannessen JN: Behaviorally evoked hippocampal theta waves: a cholinergic response. Science 188:1114-1116, 1975 Bloom FE: The role of cyclic nucleotides in central synaptic function. Rev Physiol Biochem Pharmacol 74:2-103, 1975 Takahashi T, Konishi 5, Powell D, et aJ: Identification of the motoneuron-depolarizing peptide in bovine dorsal root as hypothalamic substance P. Brain Res 73:59-69, 1974 Hughes J, Smith TW, Kosterlitz HW, et al: Identification of two related tivity.
34. content
Garattini
on the cholinergic 14:63-80, 1975
diazepam. Neuropharmacology 18. Haefely W, Kulcsar A, Mohler
18:322-
fate
S,
chopharmacol
29.
on catecholamine metabolism in regions of rat brain. Eur J Pharmacol 8:296-301 , 1969 8. Bartholini G, Keller H, Pieri L, et al: The effect ofdiazepam on turnover ofcerebral dopamine, in The Benzodiazepines. Edited by Garattini 5, Mussini E, Randall LO. New York, Raven Press, 1973, pp 236-240
GABA ganglia. Costa anism ysis
323,
1966 Taylor KM, Laverty pam, and nitrazepam
1972
Taylor KM, Laverty R: The interaction diazepam and nitrazepam with catecholamines regions of rat brain, in The Benzodiazepines. tini 5, Mussini E, Randall LO. New York,
16. Padjen
17.
24:839-840,
H, Haefely W: The effect of diazepam possible sites and mechanism of action. Arch Pharmacol 284:3 19-337, 1974
pp 191-202 14. Suria A, Costa
1971
6. Ross
Consolo zodiazepines
pharmacol 13.
1975 5.
Pharmacol
11. Polc P, Mohier cord activities: Schmiedebergs
pentapeptides from Nature 258:577-579,
brain 1975
with
potent
opiate
agonist
ac-
Pierau FK, Zimmermann P: Action of a GABA-derivative on postsynaptic potentials and membrane properties of cats’ spinal motoneurons. Brain Res 54:376-380, 1973 Saito K, Konishi 5, Otsuka M: Antagonism between lioreseal and substance P in rat spinal cord. Brain Res 97:117-180, 1975
21.
22.
23.
24.
25.
26.
Stein L, Wise CD, Belluzzi JD: Effects of benzodiazepines on central serotonergic mechanisms, in Mechanism of Action of Benzodiazepines. Edited by Costa E, Greengard P. New York, Raven Press, 1975, pp 29-44 Baumgarten HG, Bjorklund A, Lachenmayer L, et a!: Longlasting selective depletion of brain serotonin by 5,6-dihydroxytryptamine. Acta Physiol Scand Supplement 373, 1971 , pp 1-15 Aprison MH, Ferster CB: Neurochemical correlates of behavior: II. Correlation of brain monoamine oxidase activity with behavioral changes after iproniazid and 5-hydroxytryptophan. J Neurochem 6:350-357, 1961 Goldberg ME, Manian AA, Efron DH: A comparative study of certain pharmacological responses following acute and chronic administration of chlordiazepoxide. Life Sci 6:481-491, 1967 Margules DL, Stein L: Increase of “antianxiety” activity and tolerance of behavioral depression during chronic administration of oxazepam. Psychopharmacologia 13:74-80, 1968 Young AB, Zukin SR, Snyder SH: Interaction of benzodiazepines with central nervous system glycine receptors: possible mechanism of action. Proc Natl Acad Sci USA 71:2246-2250, 1974
Neural
Mechanisms
BY FLOYD
E. BLOOM,
APPENDIX 1 Evidence Linking 1 . Benzodiazepines
2. Benzodiazepine a.
5-HT
b. c.
5-HT 5-HT mine).
reduce
to Brain Serotonin 5-HT
BLOOM
(5-HT)
turnover.
antipunishment effects antagonists (methysergide,
are mimicked cinanserin,
by: BOL-
148)
3.
4.
synthesis nerve
inhibition (PCPA) terminal damage (5,6-dihydroxytrypta-
Punishment effects are intensified by: a. 5-HT agonists (a-methyltryptamine) b. 5-HT precursor (5-hydroxytryptophan plus hibitor) c. 5-HT (administered in lateral ventricle) d. Carbachol activation of 5-HT cells in dorsal cleus. Benzodiazepine potentiated
of Benzodiazepine
effects are by norepinephrine.
antagonized
by
5-HT
MAO
in-
raphe
nu-
but
are
Actions
M.D.
DESPITE THE WELL-DOCUMENTED clinical usefulness of the benzodiazepines in the treatment of anxiety, seizures, and muscular spasms, the neuronal Sites and mechanisms of action from which these therapeutic gains derive remain unknown. Solving the enigma of how and where these drugs act selectively on discrete neurons to alter whole animal or human behavior in specific ways requires detailed descriptions of benzodiazepine actions on many levels of research. The purpose of this report is to examine critically the present status of cell-specific effects of benzodiazepines in the brain and to suggest some areas worthy of continued scrutiny as research progresses.
Currentexperimental paradigms emphasize neurotransmitter-specific interactions to explain the behavioral effects ofbenzodiazepines. According to this approach the broad range ofeffects observed suggests the involvement ofseveral transmitter systems without rigorously establishing that any single transmitter system or physiological synaptic function is either necessary or sufficient to express all benzodiazepine actions. Among the effects that occur, potentiation ofamino acid-mediated presynaptic inhibition in the spinal cord and postsynaptic inhibitions elsewhere in the brain are attractive testable hypotheses. However, direct physiological evidence that benzodiazepines simulate yaminobutyric acid (GABA) or other amino acids specifically and exclusively is needed to corroborate this view.
ORGANIZATION
Revised version of a paper presented the American Psychiatric Association, 14, 1976.
at the 129th annual meeting Miami Beach, Fla. , May
OF
NEURON-DRUG
INTERACTIONS
General
Dr. Bloom is Professor and for Behavioral Neurobiology,
Benzodiazepines
E.
of 10-
Director, Arthur Vining Davis Center Salk Institute, La Jolla, Calif. 92037.
Drug
Actions
The awesome structural and functional complexity of the brain handicaps the analysis of specific drug effects. The situation is further compounded by the rapidity of cellular interactions and by the lack of understanding of many fundamental details underlying such interactions. One simplifying approach is to organize Am J Psychiatry
134:6,
June
1977
669
NEURAL
MECHANISMS
the possible modes of drug interaction with neurons for purposes of detecting general events for further analysis (see figure 1). As a first step, some benzodiazepine actions are generalizable to all nerve cells, e.g. effects on general energy metabolism (2) and maintenance of membrane integrity and transmembrane ionic equilibria (3). For the most part, actions affecting common biological properties of all neurons occur at relatively high dose levels in vitro and are not the drug actions we are now trying to comprehend. ,
Transmitter-Specific
Drug
Actions
Chemical dynamics. A second category of drug actions pertains to selective effects on neurons grouped together by their specific common use of the same chemical synaptic transmitter. Here the potential molecular mechanisms for drug specificity are more easily envisioned because of the numerous points within the life cycle of transmitters that have been uncovered through biochemical pharmacology (see figure 1). Although biochemically detected changes in levels and turnovers of several biogenic amines have been reported (4-13), the data available are insufficient to account for the unique patterns of behavioral responses because similar biochemical and histochemical effects have been reported after other sedative drugs, such as ethanol and barbiturates (5), that clearly differ behaviorally and clinically from the benzodiazepines. Synaptic receptors. Drugs may also act directly at postsynaptic receptor sites. Binding studies of transmitter receptors with benzodiazepines are reviewed by Snyder and associates in this Special Section of the Journal. Neuronal
Integration
Mechanisms
Another mode of categorizing drug-neuron interactions can also be applied to studies of benzodiazepines. Under this category are included studies in which parenterally injected or topically applied drugs were used to influence evoked responses or single unit activity in relationships in which the chemical nature of the synaptic interactions being tested remained undefined. Ifbroadly and consistently applied, such test paradigms could still be used to establish dose-response and time-response curves by which the most rapid and most sensitive effects of the drug on key target cells could be discerned. The available evidence from such experiments is restricted now to three brain regions: the cerebellum, the spinal cord, and the limbic system, especially the hippocampal formation. Very low doses of benzodiazepines have also been reported to antagonize the production of posttetanic potentiation (VFP) in the sympathetic ganglion of the bullfrog in vitro (14); however, since these same drugs do not interfere with FTP in the neuromuscular junction or at cholinergic receptors in the spinal cord (see references in [15] and [16]), this interesting observa670
Am
J Psychiatry
134:6,
June
1977
tion ther
in the sympathetic ganglion cannot be carried furconceptually at this time. Cerebellum Since the cerebellum is generally regarded as inhibitory to the motor cortex, the enhancement of cerebellar activity by benzodiazepines could implement an anticonvulsant action. Purkinje cells in the cerebellar cortex inhibit the neurons of the deep cerebellar nuclei, which project to the cerebral cortex. Therefore, consistent facilitation ofPurkinje cells should inhibit the deep nuclei and disinhibit (i.e. facilitate by removing tonic inhibition) the cortex. Although facilitation of Purkinje cell firing was observed to occur when cats were administered diazepam (17), the opposite effect (which would logically be more “anticonvulsant”) has been reported for the effects of diazepam on both cats and rats (18). However, a direct antagonistic action against y-aminobutyric acid (GABA) here (as recently reported by Steiner and Felix [19]) would, ofcourse, produce the opposite effect. The findings of Suria, Costa, and Mao and associates (14, 15, 20) that diazepam is extremely potent in lowering cyclic 3 ‘-5 ‘guanosine monophosphate (GMP) levels in the cerebellum may relate to the anticonvulsant effects of benzodiaiepines, since several convulsants selectively elevate cerebellar cyclic GMP. Spinal cord. The advantages of studying drug actions in the spinal cord are that one can know with some certainty what is being stimulated and what is being recorded. By recording from and stimulating various combinations of the dorsal or ventral roots of the isolated spinal cord, accurate dose-response estimates of effects at definable endpoints can be obtained. These technical advantages have been applied with considerable impact to the effects of benzodiazepines (16). Because of the functional significance of depolarizing the presynaptic primary afferent fiber, a great deal of emphasis has been placed on changes in the primary afferent terminals as a result of stimulation of an adjacent dorsal root or ipsilateral ventral root. This presynaptic sensory fiber is simultaneously postsynaptic to other spinal interneurons (see figure 1). Activation of these interneurons depolarizeS the primary afferent terminal and decreases the release of excitatory sensory transmitter onto the motoneurons. Therefore, a standard stimulating volley applied to a dorsal root whose primary afferent terminals are depolarized by appropriate spinal interneurons will produce less excitation of the motoneurons. The motoneurons are therefore inhibited presynaptically but without any of the membrane potential shifts associated with conventional postsynaptic inhibition. Several groups of researchers (I 1 18) have reported that benzodiazepines enhance the dorsal root reflex elicited either by ventral root stimulation (depolarizing dorsal root terminals through a set of spinal interneurons believed to release either /3-alanine or taurine [21]) or by adjacent dorsal root stimulation (depolarizing the primary afferent terminals through spinal interneurons releasing GABA [21]). .
,
,
FLOYD
FIGURE 1 Potential Actions of Benzodiazepines Functions*
on General or Specific
Neuronal
E.
BLOOM
have been taken to support the view (15, 18) that GABA-mediated sites of presynaptic inhibition are the major site of selective effects (i.e. an enhancement) produced by benzodiazepines. These views are further strengthened by the related observations that pharmacological elevation of spinal cord GABA levels (through blockade of catabolism) enhances the effects of benzodiazepines on the dorsal root depolarization but blockade of GABA synthesis prevents the enhancement of dorsal root depolarization (1 1 18). On the other hand, chemical axoaxonic presynaptic inhibitory mechanisms cannot account for the effects of benzodiazepines on cerebellum or on hippocampus for two reasons: 1) such Synaptic structures have not been detected there and 2) GABA clearly mediates postsynaptic inhibitions in these brain regions (23). Moreover, although benzodiazepines potentiate dorsal root depolarization, this response is not necessarily an exclusive GABA-mediated event; either taurine or /3alanine (in the frog [21] and the cat [23]) can mediate similar effects, and barbiturates, whose effects are also reversed by drugs that antagonize GABA (24), can enhance the depolarization response as well as the benzodiazepines can. Furthermore, the glycine antagonist strychnine, which is used as a tool for the detection of the glycine-binding protein (25), will also antagonize the ability of both taurine and /3-alanine to depolarize the dorsal root fiber (21) and will partially antagonize these amino acids at postsynaptic receptors elsewhere in the forebrain (23). This crossover of antagonist effects between amino acids suggests that some caution must be used in interpreting the uniformity of transmiller-specific sites at which these convulsant antagofists can reverse the effects of benzodiazepines. Limbic system. The limbic system has figured prominently in occasional abortive attempts to characterize the anxiolytic effects of the benzodiazepines because of the suspicions that this brain formation has affective functions (26). When chronically implanted wire electrodes were used to record from units in the hippocampus, thalamic reticular formation, or anterior hypothalamus of unrestrained rats over long periods of time, only hippocampal neurons responded specifically to benzodiazepines (27). Under these experimental conditions, hippocampal units consistently showed a dose- and time-dependent reduction of unit discharge rates. In other studies, the poorly understood polysynaptic connections between the amygdaloid complex and the hippocampus were also shown to be sensitive to low doses of benzodiazepine (28). The molecular nature and behavioral significance of this otherwise isolated effect therefore remain unknown. Like the actions of benzodiazepines on cerebellar cyclic nucleotides described above, a muscarinic cholinergic mechanism is known to excite hippocampal neurons (29); such receptors are also associated with cyclic GMP synthesis in other neurons (30). Extended analysis of benzodiazepine actions on cyclic GMP levels in the limbic system might also be informative with respect to anxiolytic actions. ,
,
#{149}The enumerated steps indicate possible points in the biological cycle of a central neuron at which benzodiazepines and other psychoactive drugs might interact. Step 1: axoplasmic transport via microtubules. Step 2: electrical cxcitable components of the axonal and perikaryonal membrane. Step 3: organdes and enzymes present within preterminal axons, synaptic boutons, and some dendritic processes for the synthesis, storage, catabolism, and release of the transmitter as well as for the active reuptake and restorage of the transmitter after release. Step 4: glia-neuronal interaction processes, which are
presently
not well characterized but might assist in policing the extrasynaptic to maintain minimal concentrations of active transmitter molecules. Step 5: postsynaptic receptors that trigger events within the postsynaptic cell initiated by the release of transmitter. Step 6: intracellular neuronal organdIes that may mediate the postsynaptic response in some of the subsequent metaspaces
bolic
aspects,
ar interactions bly sensitive
such
as ribosomal
protein
synthesis.
Step
7: cytoplasmic-nucle-
that regulate the expression ofgenetic information that is possito transmitter-regulated events. Step 8: intrasynaptic events of desensitization, habituation, and other physiological modula-
sensitization, tions of transsynaptic function. Step potentials from various convergent
9: integration of simultaneous synaptic excitatory and inhibitory transmitters.
Step 10: discharge ofthe target cell and propagation ofa signal down its axon. Step 1 1 : a presynaptic element contacts one of the afferent terminals onto the target cell and regulates transmitter release presynaptically. synaptic autoreceptors by which the release of a transmiuer terminal may be modified by the transmitter it has previously data for this figure are from Cooper and associates (I).
Step 12: prefrom a synaptic released. The
The ability of benzodiazepines to potentiate the depolarizing response of primary afferent terminals can be antagonized to a considerable degree by picrotoxin and by bicuculline, two convulsants that antagonize postsynaptic inhibitory actions ofGABA, as well as by some of the actions of f3-alanine and taurine (see [16] and [21] for discussion). Similar observations on sites of presumed amino acid-mediated presynaptic inhibition in the cuneate nucleus (22) and postsynaptic inhibition in the striatonigral pathway (18) can also be enhanced by benzodiazepines. These data collectively
Am
J Psychiatry
134:6,
June
1977
671
NEURAL
MECHANISMS
J Pharm
CONCLUSIONS
Almost all of the benzodiazepine actions currently available for review are probably compatible with an umbrella theory focused on GABA-mediated potentiation of pre- and postsynaptic inhibition (15, 18), including the moderate decreases in biochemical indices of activity in catecholamine, acetylcholine, and Serotonin neurons. However, the nonspecificity of the drugs used to manipulate tissue GABA levels, without comparably sensitive measurements of other natural amino acids, weakens this approach. The antagonistic actions reported in the cerebellum maintain the association with GABA in the opposite direction (19). The continual emergence of newer transmitter materials, such as substance P (3 1) and the morphine-like peptides (32), suggests the need for prudence and caution before wholehearted acceptance of a monolithic explanation of benzodiazepine actions through a single transmitter substance. One need only recall how the specific transmitter-related theories of behavioral depression induced by reserpine were forced into continual revisions by the discoveries that not only serotonin but norepinephrine and dopamine as well were profoundly depleted (1) to imagine how the present interpretive status of benzodiazepine-neuronal interactions may be equally shattered by further studies on substance P and related peptides. For example, recent data suggest that p-chlorophenyl GABA (lioreseal), a drug studied as a GABA-synergist (33), is equally well characterized as a substance P antagonist (34).
REFERENCES
I. Cooper JR. Bloom FE, Roth RH: The Biochemical Basis of Neuropharmacology. New York, Oxford University Press, 1970, pp 191-202 2. David LF, Gatz EE, Jones JR: Effects of chlordiazepoxide and diazepam on respiration and oxidative phosphorylation in rat brain mitochondria. Biochem Pharmacol 20:1883-1887, 1971 3. Ueda I, Wada T, Ballinger CM: Sodium and potassium-activated ATPase of beef brain-effects of some tranquilizers. Biochem Pharmacol 20:1697-1700, 1971 4. Fuxe K, Agnati LF, Bolme P, et al: The possible role of GABA mechanisms in the action of benzodiazepines on central catecholamine neurons. Adv Biochem Psychopharmacol 14:45-62,
12.
Corrodi H, Fuxe stress and central
K, Lindbrink catecholamine
P. et al: Minor neurons. Brain
tranquilizers, Res 29:1-16,
15.
take
SB, Renyi AL: In vivo inhibition of H3-noradrenaline by mouse brain slices in vitro. J Pharm Pharmacol
7.
R: The
effect
of chlordiazepoxide,
9.
10.
672
Chase TN, Katz RI, Kopin IJ: Effect intracistemally injected serotonin-C14. 9:103-108, 1970 Saad SF: Effect of diazepam on GABA
Am
J Psychiatry
134:6,
June
of diazepam on Neuropharmacology
1977
of mouse
Julien
E: Action
A, Bloom
of the
site
RM:
FE:
Ladinsky
21.
Action
Adv
of diazepam,
of
of
the
Biochem
ben-
Psycho-
chlordiazepoxide, and histamine in Edited by GaratRaven Press, 1973,
dibutyryl
cGMP,
in the electrophysiological
of benzodiazepines.
14:93-102, Cerebellar
in the
zodiazepines 260:346-347,
20.
H:
system.
Problems
of action
and
on
Adv
anal-
Biochem
Psy-
1975 involvement
central
Psychopharmacol 19. Steiner FA, Felix
actions
in the
anti-epileptic
of benzodiazepines.
14:131-152, 1975 D: Antagonistic effects vestibular
action
of
11:683-691, 1972 H, et al: Possible involvement
and
Adv
of GABA
cerebellar
neurones.
of
Biochem
and
ben-
Nature
1976 Mao CC, Guidotti A, Costa E: Inhibition by diazepam of the tremor and the increase in cerebellar cyclic GMP elicited by harmaline. Brain Res 83:516-519, 1975 Barker JL, Nicoll RA, Padjen A: Studies on convulsants in the isolated
frog
spinal
cord:
II.
Effects
on root
potentials.
J Physiol
(Lond) 245:537-548, 1975 22. Davidson N, Southwick CA: Amino acids and presynaptic inhibition in the rat cuneate nucleus. J Physiol (Lond) 2 19:689-708, 1971 23. Curtis DR. Johnston OAR: Amino acid transmitters in the mammalian central nervous system. Rev Physiol Biochem Pharmacol 69:97-188, 1974 24. Nicoll RA: Presynaptic action of barbiturates in frog spinal cord. Proc NatI Acad Sci USA 72:1460-1463, 1975 25. Young A, Zukin SR. Snyder SH: Interaction of benzodiazepines with central nervous system glycine receptors: possible mechanism of actions. Proc Nail Acad Sci USA 71:2246-2250, 1974 26. Douglas Ri: The development of hippocampal function: implications for theory and for therapy, in The Hippocampus, vol 2: Neurophysiology
27.
28.
up-
30.
diaze-
31.
brain.
S.
on spinal Naunyn
on presynaptic nerve terminals in bullfrog sympathetic Brain Res 87:102-106, 1975 E, Guidotti A, Mao CC, et al: New concepts on the mechof action of benzodiazepines. Life Sci 17:167-186, 1975
GABA
32.
33.
of
and
Behavior.
Edited
by
Isaacson
RL,
Pm-
bram KH. New York, Plenum Press, 1975, pp 327-362 Olds ME, Olds J: Effects of anxiety-relieving drugs on unit discharges in hippocampus, reticular mid-brain, and pre-optic area in the freely moving rat. International Journal of Neuropharmacology 8:87-103, 1969 Jalfre M, Monachon MA, Haefely W: Effects on the amygdalohippocampal evoked potential in the cat of four benzodiazepines and some other psychotropic drugs. Naunyn Schmiedebergs Arch Pharmacol 270: 180-191 , 1971 Teitelbaum H, Lee JF, Johannessen JN: Behaviorally evoked hippocampal theta waves: a cholinergic response. Science 188:1114-1116, 1975 Bloom FE: The role of cyclic nucleotides in central synaptic function. Rev Physiol Biochem Pharmacol 74:2-103, 1975 Takahashi T, Konishi 5, Powell D, et aJ: Identification of the motoneuron-depolarizing peptide in bovine dorsal root as hypothalamic substance P. Brain Res 73:59-69, 1974 Hughes J, Smith TW, Kosterlitz HW, et al: Identification of two related tivity.
34. content
Garattini
on the cholinergic 14:63-80, 1975
diazepam. Neuropharmacology 18. Haefely W, Kulcsar A, Mohler
18:322-
fate
S,
chopharmacol
29.
on catecholamine metabolism in regions of rat brain. Eur J Pharmacol 8:296-301 , 1969 8. Bartholini G, Keller H, Pieri L, et al: The effect ofdiazepam on turnover ofcerebral dopamine, in The Benzodiazepines. Edited by Garattini 5, Mussini E, Randall LO. New York, Raven Press, 1973, pp 236-240
GABA ganglia. Costa anism ysis
323,
1966 Taylor KM, Laverty pam, and nitrazepam
1972
Taylor KM, Laverty R: The interaction diazepam and nitrazepam with catecholamines regions of rat brain, in The Benzodiazepines. tini 5, Mussini E, Randall LO. New York,
16. Padjen
17.
24:839-840,
H, Haefely W: The effect of diazepam possible sites and mechanism of action. Arch Pharmacol 284:3 19-337, 1974
pp 191-202 14. Suria A, Costa
1971
6. Ross
Consolo zodiazepines
pharmacol 13.
1975 5.
Pharmacol
11. Polc P, Mohier cord activities: Schmiedebergs
pentapeptides from Nature 258:577-579,
brain 1975
with
potent
opiate
agonist
ac-
Pierau FK, Zimmermann P: Action of a GABA-derivative on postsynaptic potentials and membrane properties of cats’ spinal motoneurons. Brain Res 54:376-380, 1973 Saito K, Konishi 5, Otsuka M: Antagonism between lioreseal and substance P in rat spinal cord. Brain Res 97:117-180, 1975
