- Email: [email protected]
Pharmacological characterization of GABA receptors mediating vasodilation of cerebral arteries in vitro
Brain Research, 173 (1979) 89-97 © Elsevier/North-Holland Biomedical Press
89
PHARMACOLOGICAL CHARACTERIZATION OF GABA RECEPTORS MEDIATING VASODILATION OF CEREBRAL ARTERIES IN VITRO
LARS EDVINSSON and DIANA N. KRAUSE
Department of Histology, University of Lund, Lund (Sweden) and Division of Neurosciences, City of Hope National Medical Center, Duarte, Calif. (U.S.A.) (Accepted January 11th, 1979)
SUMMARY
GABA (?,-aminobutyric acid) produced a dose-dependent dilation of isolated cat and dog cerebral artery segments which had been given an active, tonic contraction by either prostaglandin F2a or serotonin. No effect of GABA on extracranial blood vessels was observed. The GABA-induced dilation could be blocked in a dose-dependent manner by either bicuculline or picrotoxin. The latter agent appeared to act as a competitive antagonist. GABA agonists muscimol, imidazoleacetic acid, ~-aminovaleric acid, (-l-)7-amino-fl-hydroxybutyric acid, and fl-alanine also relaxed actively contracted cerebral arteries dose-dependently. The relative potency of these agonists was consistent with that established for GABA receptors on neurons and invertebrate striated muscle. GABA was also tested on two human cerebral arteries and found to cause a small dilation. The results support the existence of a cerebrovascular GABA receptor which may mediate an interaction between GABA and the cerebral circulatory system.
INTRODUCTION
A possible role in the cerebral circulatory system has been proposed recently for the inhibitory neurotransmitter GABA (y-aminobutyric acid). GABA was reported to dilate dog cerebral, but not peripheral, arteries in vitro s. These findings suggest the existence of a cerebrovascular receptor for GABA. An association of GABA with cerebral blood vessels is further indicated by the presence of a non-neuronal form of the enzyme which synthesizes GABA, glutamate decarboxylase (GAD), in relatively high concentrations in pial arteries as compared to peripheral vessels 15,1s. GABA transaminase (GABA-T), the degradative enzyme for GABA also has been demonstrated histochemically in large concentrations around
90 brain vessels, but not around other, e.g. kidney, vessels 24,25. A physiological role for GABA in the regulation of cerebral circulation is supported by a brief Russian report indicating that, in vivo, GABA is capable of decreasing cerebral arterial resistance and increasing cerebral blood flow in cats, rabbits, and dogs 17. In this study, we have further characterized the receptor mediating GABAinduced dilation by examining the effects of various known GABA agonists and antagonists on the mechanical response of cat pial arteries in vitro. For comparison, cerebral arteries from dog and human, as well as extracranial arteries, were also tested for possible GABA responses. MATERIALS AND METHODS
Material Thirteen cats of both sexes (2.0-4.5 kg) and one dog (15 kg) were used. They were anesthetized with sodium pentobarbitol (30 mg/kg i.p.) and killed by exsanguination. The brain was removed, and the middle cerebral, posterior communicating and basilar arteries (300-600/~m in diameter) were dissected out and placed immediately in an oxygenated Krebs-bicarbonate solution. Branches of similar caliber were taken from the external maxillary or lingual arteries. Material from two patients was obtained during neurosurgical tumor operations. (Institutional rules governing the protection of human subjects were followed in obtaining and studying these tissues.) Small pial arteries were removed from macroscopically intact parts of resected brain tissue. The specimens were transported to the laboratory in ice-cold buffer solution.
Vascular preparation The morphology of the vessels and the in vitro conditions for studying their mechanical responses have been determined previously 5,6. Pieces of intra- and extracranial arteries (about 5 mm long each) were used immediately for testing; the remainder were kept in a refrigerator at 4 °C for later use (up to 24 h). The two types of arteries were mounted in a 50 ml temperature-controlled organ bath with two separate systems of L-formed metal holders for recording isometric circular contractions as described previously 5. The bath contained a solution of the following composition (mM): NaC1, 118; KCI, 4.5; CaCI2.2HzO, 2.5; MgSO4.7H~O, 1.0; NaHCOs, 25; KH2PO4, 1.0; and glucose, 6.0. The bath and the stock solution were maintained at 37.5 d: 0.5 °C (range) and aerated continuously with a mixture of 95 % O8 and 5 % COg. Shortly after the arterial preparations had been mounted in the organ bath, each was subjected to a load of 400 dynes and allowed to stabilize; the tension decreased approximately 50--200 dynes. The test drugs were administered after 1.5 h of equilibration. Preliminary tests showed that the GABA agonists dilate cerebral vessels. Under our experimental conditions n, the vessels were almost completely relaxed, and dilator responses were thus difficult to obtain. Therefore vessels were given a resting tone by addition of either 2.5 × 10-6 M prostaglandin F~a (PGF2a) or 3 × 10-s M 5-hydroxytryp-
91 tamine (5-HT). The contraction by either of these two agents, an average of 200 dynes in this study, remained at a steady level for at least 30 min. During this time increasing doses of various GABA agonists were added cumulatively. In blockade experiments, the antagonists were added 15-20 min before the agonist doses and remained in the bath during agonist application.
Analysis of data The effect of the agonists were plotted in terms of response against log dose. The maximum response obtained, EArn (expressed in dynes), and the EDs0 (molar concentration of agonist at which half-maximum response occurs) were used as measures of agonist potency. The results are generally expressed as mean values 4- standard error of the mean. Drugs Bicuculline (Chemicals Procurement Labs), imidazoleacetic acid (Calbiochem), muscimol (Merck Frosst Laboratories), prostglandin F2a (Astra) and taurine (Eastman) were used in this study as well as the following compounds obtained from Sigma: fl-alanine, ?-aminobutyric acid (GABA), (~)?-amino-fl-hydroxybutyric acid, d~aminovaleric acid, glutamic acid, glycine, 5-hydroxytryptamine creatine sulfate (5-HT), and picrotoxin. RESULTS
Response of cat arteries to GABA GABA was found to dilate in a dose-dependent manner intracranial vessels actively contracted by either PGF2a or 5-HT (Fig. 1). This response, however, could GABA concentration (M) 10-8
10-7
10-s
10-s
10-{
O'
102030. 1.0 50 60
I
70. 80 Dilation (dynes) Fig. 1. D o s e - r e s p o n s e curve for the dilator effect o f G A B A tested on segments o f cat middle cerebral arteries actively contracted by either 2.5 × 10 -8 M PGF2a or 3 × 10 -8 M 5-HT. Points represent the m e a n s 4- S.E.M. for 13 experiments.
92
10 8
GABA concentration (M) 10-~ 10-s
10-7
10 ~
10-3
Oo! . . . . . . 40
•
,
°
.
~
•
-
_
6O
80
100
Dilation I%1 Fig. 2. Effect o f p i c r o t o x i n on the dose-response curve f o r the d i l a t o r effect o f G A B A . Isolated cat middle cerebral arteries were pretreated with either 0.3 × I 0 7 M, or 3 :~ l 0 ~ M picrotoxin for 15 min and then given active tone with 2.5 × 10 GM PGE2~ and tested for responses to increasing concentrations of GABA. Each point represents the mean for two experiments.
TABLE I
Dilatory effects o f various G,4BAergic agonists on the cat middle cerebral artery Mechanical responses of isolated segments of cat middle cerebral arteries were recorded as described in Methods. Increasing doses of a G A B A agonist were added cumulatively to the bath of vessel segments which were tonically contracted by approximately 200 dynes with either 2.5 :< 10 6 M PGFe~ or 3 × 10 -.8 M 5-HT. Results are given as means 2_ S.E.M. except for N 2 which is the mean -i standard deviation. N = n u m b e r of experiments, EDs0 dose producing a half-maximum dilation, and Ear, = m a x i m u m dilation.
,4gonist
N
EDso (M)
Muscimol GABA Imidazoleaceticacid 6-Aminovalericacid 7-Amino-fl-hydroxybutyric acid fl-Alanine
6 13 4 2 3 3
3.8 7.7 1.l 2.0 2.3 3.5
2± ± ± ± i
2.5 2.4 0.5 1.4 1.5 3.3
EArn (dynes) x x x × × x
10 -7 10 7 10-~ 10 4~ 10 -G 10 -6
54 59 49 51 47 47
_+ 10 5:I1 5 I1 ~ 4 :L 9 ~: 12
93 not be detected in every vessel; GABA failed to dilate any of the intracranial arteries tested from 4 of the 12 cats studied. In those vessels which did respond (15 out of 33 vessels examined), the GABA dilatory effect, while not strong, was consistent. Characteristically, GABA produced a threshold response at a bath concentration of about 0.33 #M, a half maximum response (EDs0) at 0.8 #M, and a maximum response of approximately 60 dynes dilation at 10-100/~M. Relaxation in response to GABA was reversible and rapid, being complete within 1-2 min. Similar GABA effects were observed with basilar and middle cerebral artery preparations, although most experiments were performed with the latter vessel type. GABA was never found to affect extracranial vessels, nor did it relax intracranial vessels which had been actively contracted with saturated KCI (3 vessel segments tested).
Effect of GABA antagonists The effects of the convulsants picrotoxin and bicuculline, two relatively specific antagonists of GABA responses in neurons3,11,14 and crustacean striated muscle zl, were examined on the GABA-induced dilation of cat middle cerebral arteries. Pretreatment of an intracranial vessel with either picrotoxin or bicuculline at concentrations of 0.3 # M and 3/~M caused a dose-dependent inhibition of the vessel's response to GABA. The antagonists by themselves had no obvious direct effect on the tonic, active contraction induced by PGF2a. Picrotoxin was examined in more detail and appeared to produce parallel shifts in the GABA dose-response curve towards higher agonist doses (Fig. 2), an action characteristic of competitive antagonists.
Effect of GABA agonists Table I summarizes the effects on cat intracranial vessels of several compounds structurally related to GABA that have been well characterized as GABA agonists in neuronal preparations. Muscimol, imidazoleacetic acid, O-aminovaleric acid, (4-)),amino-fl-hydroxybutyric acid, and fl-alanine each caused a dose-dependent relaxation when tested on actively contracted, GABA-responsive vessels at bath concentrations of 10-8-10 -4 M. The maximum dilatory response was similar for all agonists tested, including GABA. When compared on the same vessel preparation, the potency of muscimol was greater than or equal to that of GABA. The other compounds were found to be somewhat less potent than GABA, with fl-alanine generally being the least potent agonist. Other putative amino acid neurotransmitters, glycine, taurine, and glutamic acid, were also tested at 10-4 M on actively contracted vessels, but no response of the vessels could be detected.
GA BA responses in arteries of other species For the purpose of comparison, experiments with GABA were conducted using cerebral arteries obtained from dog and humans. Four dog middle cerebral artery segments contracted by 5-HT were tested, and three exhibited relaxation in response to GABA (10-8-10 -3 M). The responses appeared similar to those observed with cat intracranial vessels. A maximum response of 63 4- 11 dynes dilation was elicited with
94 30-300/~M GABA. Estimated EDs0 values (0.07 ktM, 1.1 #M, 15/zM) varied among the three responding vessels. The dog arteries also dilated in response to muscimol (1 vessel tested) and (-~)7-amino-fl-hydroxybutyric acid (two vessels tested). The potency and maximum dilatory response of both compounds appeared similar to that obtained with G A B A in the same vessel. The effect of G A B A was examined on two human pial arteries (one from each patient) actively contracted by PGFza. A small dilation could be detected in response to GABA at bath concentrations of 0.1-1 #M. One human extracranial vessel was also tested, but no G A B A response was observed. DISCUSSION A recent in vitro study on dog pial arteries 8 suggests that GABA may affect the cerebral vasculature in addition to its known inhibitory neurotransmitter actions on nerve cells and invertebrate striated muscle. Our studies support a direct vasodilatory action of GABA, observed here in isolated cat pial arteries and a small sample of dog and human cerebral arteries. Fujiwara et al. 8 found that most of their isolated dog pial arteries only responded to high G A B A concentrations; however, we obtained from all vessels which responded to G A B A fairly consistent dose-response relationships with an average EDs0 of 7.7 k: 2.4 × 10 -7 M. In our studies the maximum response to GABA was lower in both cat (59 ± 1i) and dog (63 ± 11) vessels than the previously reported value for dog vessels of approximately 200 dynes dilation. In agreement with Fujiwara et al. s, no GABA responses were ever detected in extracranial vessels. Our results are consistent with the presence of a specific GABA receptor on the cerebral vessels which appears analogous to the G A B A receptor found on neurons and invertebrate striated muscle. The G A B A response was blocked in a dose-dependent manner by picrotoxin and bicuculline, known GABA antagonists in vertebrate CNS neurons ~,14 and crustacean neurons 11 and striated muscle 21. Fujiwara et al. s also reported that GABA-mediated dilation in dog was blocked by picrotoxin but not by adrenergic or muscarinic blocking agents. In our studies, picrotoxin appeared to act as a competitive antagonist, producing parallel shifts to the right of the G A B A doseresponse curve. A competitive type of blockade also has been observed for this convulsant in lobster muscle 2~ and the crayfish stretch receptor neuron H, preparations where similar quantitative, physiological dose-response data can be obtained. The existence o f a cerebrovascular G A B A receptor was further confirmed by the GABA-like dilatory action of structurally related compounds which have been characterized as G A B A agonists and which bind directly to the GABA receptor site on neuronal membrane fractions19,2L The relative potency for the agonists tested on the cerebral arteries was consistent with that found for other known GABA receptors. Muscimol was generally the most potent agonist ; this has been demonstrated in many preparations including cat spinal cord neurons3, ~3, crayfish stretch receptor neurons 11, and crustacean neuromuscular junctions ~7. Imidazoleacetic acid, (±)y-amino-fl-hydroxybutyric acid, and 6-aminovaleric acid have been reported to be somewhat less
95 potent than GABA, while fl-alanine has been shown to be an even less effective GABA agonist2,4,9,22, 2a. The cerebral artery response appeared specific for GABA as other putative amino acid neurotransmitters, glycine, taurine, and glutamate had no effect on the vessels. A similar pharmacology is found for the cerebrovascular GABA receptor site when it is identified using [aH]muscimol binding assays (Krause, Wong, Degener and Roberts, in preparation). While the existence of a cerebrovascular GABA receptor suggests a role for this substance in the control of cerebral circulation, the in vitro relaxation produced by GABA and its analogs was not as pronounced as that of other known mediators of vasodilation such as adenosine compounds 10, pH reduction 7, or t-receptor agonists 6. In both our study and that of Fujiwara et al. s, the GABA dilatory effect was undetectable in some of the isolated vessels tested, which may be why Lee et al. failed to observe a response to GABA (3-300/~M) in their isolated cat cerebral arteries 16. The reasons for the lack of response in some vessels is not clear but may relate to possible unfavorable conditions created in vitro or a relatively low number of receptors present in the particular vessels studied. The effectiveness of GABA does not appear to be limited by the GABA-T activity in the vessels, as treatment with a GABA-T inhibitor, aminooxyacetic acid, did not appear to affect the GABA response (unpublished observation). Under certain in vivo conditions however, GABA-mediated vasodilation may be enhanced. GABA has been reported to increase cerebral blood flow in vivo17; and muscimol, when administered intravenously to rats, also produces an increase in local cerebral blood flow as determined by the [14C]ethanol technique (unpublished observation). Cerebral vessels are exposed normally to several sources of GABA in vivo. Brain vessels contain GABA-related enzymeslS,1s,24,~5 and thus are capable of producing GABA themselves. Direct GABAergic innervation of the vessels has never been seen with the specific G A D immunocytochemical technique (Vaughn and Ribak, personal communication), although GABA is released into the extracellular fluid of the brain by neurons and possibly glia2L According to recent reports12, ~s, the cerebrospinal fluid (CSF), which surrounds pial arteries in vivo, contains in humans from about 0.1 to 0.8/zM GABA, concentrations which correspond to the lower range of the doseresponse relationship reported here. Interestingly, elevated CSF levels of GABA have been measured in human patients during certain pathological conditions involving altered blood flow such as migraine headache 26 and cerebral anoxia due to cerebrovascular disease 1. ACKNOWLEDGEMENTS The authors are grateful to Dr. Jan Erik Hardebo and Prof. Christer Owman for helpful discussion and assistance and to Dr. Eugene Roberts for valuable advice and encouragement and for referring us to and kindly translating the Russian articles by Mirzoyan and coworkers. This research was supported by the Swedish Medical Research Council (Grant 04X-732) and USPHS Grants NS-05695 and NS-12116.
96 REFERENCES 1 Achar, V. S., Welch, K. M. A., Chabi, E., Bartosh, K. and Meyer, J. S., Cerebrospinal fluid gammaaminobutyric acid in neurologic disease, Neurology (Minneap.), 26 (1976) 777-780. 2 Bowery, N. G. and Brown, D. A., Depolarizing actions of F-aminobutyric acid and related compounds on rat superior cervical ganglia in vitro, Brit. J. Pharmacol., 50 (1974) 205-218. 3 Curtis, D. R.,Duggan, A. W., Felix, D. and Johnston, G. A. R., Bicuculline, an antagonist of GABA and synaptic inhibition in the spinal cord of the cat, Brain Research, 32 (1971) 69-96. 4 Curtis, D. R. and Watkins, J. C.,The pharmacology ofamino acids related togamma-aminobutyric acid, Pharmacol. Rev., 17 (1965) 347-391. 5 Edvinsson, L., Nielsen, K. C. and Owman, C., Influence of initial tension and changes in sensitivity during amine-induced contractions of pial arteries in vitro, Arch. int. Pharmacodyn., 208 (1974) 235-242. 6 Edvinsson, L. and Owman, C., Pharmacological characterization of adrenergic alpha and beta receptors mediating vasomotor response of cerebral arteries in vitro, Circulat. Res., 35 (1974) 835-849. 7 Edvinsson, L. and Sercombe, R., Influence ofpH and pCO~ on alpha-receptor mediated contraction in brain vessels, Actaphysiol. scand., 97 (1976) 325-331. 8 Fujiwara, M., Muramatsu, I. and Shibata, S., ),-Aminobutyric acid receptor on vascular smooth muscle of dog cerebral arteries, Brit. J. Pharmacol., 55 (1975) 561-562. 9 Godfraind, J. M., Krnjevi6, K., Mareti6, H. and Pumain, R., Inhibition of cortical neurones by imidazole and some derivatives, Canad. J. Physiol. Pharmacol., 51 (1973) 790-797. 10 Hardebo, J. E. and Edvinsson, L., Adenine compounds: cerebrovascular effects in vitro with reference to their possible involvement in migraine, Stroke, 10 (1979) 58-62. 11 Hori, N., Ikeda, K. and Roberts, E., Muscimol, GABA and picrotoxin: effects on membrane conductance of a crustacean neuron, Brain Research, 141 (1978) 364-370. 12 Huizinga, J. D., Teelken, A. W., Muskiet, F. A. J., Meulen, J. V. D., Wolthers, B. G., Identification of GABA in human CSF by gas liquid chromatography and mass spectrometry, New Engl. J. Med., 296 (1977) 692. 13 Johnston, G. A. R., Curtis, D. R., DeGroat, W. C. and Duggan, A. W., Central actions ofibotenic acid and muscimol, Biochem. PharmacoL, 17 (1968) 2488-2489. 14 Krnjevi6, K., Chemical nature of synaptic transmission in vertebrates, Physiol. Rev., 54 (1974) 418-540. 15 Kuriyama, K., Haber, B. and Roberts. E., Occurrence of a new L-glutamic acid decarboxylase in several blood vessels of the rabbit, Brain Research, 23 (1970) 211-123. 16 Lee, T. J.-F., Hume, W. R., Su, C. and Bevan, J. A., Neurogenic vasodilation ofcat cerebral arteries, Circulat. Res., 42 (1978) 535-542. 17 Mirzoyan, S. A. and Akopyan, V. P., The effect produced by gamma-aminobutyric acid on the cerebral circulation and oxygen tension in the brain, Farmakologiya i Toksikologiya (U.S.S.R.), 5 (1967) 572-574. 18 Mirzoyan, S. A.,Kazaran,V.A. andAkopyan, V. P.,Theglutamicdecarboxylaseactivityin blood vessels of the brain, Dokl. Acad. Nauk (U.S.S.R.), 190 (1970) 1241-1243. 19 Olsen, R. W., Greenlee, D., Van Ness, P. and Ticku, M. K., Studies on the gamma-aminobutyric acid receptor/ionophore proteins in mammalian brain. In F. Fonnum (Ed.), Amino Acids ~s Chemical Transmitters, Plenum Press, New York, 1978, pp. 467-486. 20 Sellstr6m,/~. and Hamberger, A., Potassium-stimulated ),-aminobutyric acid release from neurons and glia, Brain Research, 119 (1977) 189-198. 21 Shank, R. P., Pong, S. F., Freeman, A. R. and Graham, L. T., Bicuculline and picrotoxin as antagonists of ~'-aminobutyrate and neuromuscular inhibition in the lobster, Brain Research, 72 (1974) 71-78. 22 Swagel, M. W., lkeda, K. and Roberts, E., Effects of GABA, imidazoleacetic acid, and related substances on conductance of crayfish abdominal stretch receptor, Nature New BioL, 246 (1973) 91-92. 23 Takeuchi, A. and Takeuchi, N., The structure-activity relationship for GABA and related compounds in the crayfish muscle, Neuropharmacology, 14 (1975) 627-634. 24 VanGelder, N. M.,Apossibleenzymebarrier for~,-aminobutyricacidinthecentralnervoussystem, Progr. Brain Res., 29 (1968) 259-268. 25 Waksman, A., Rubinstein, M. K., Kuriyama, K. and Roberts, E., Localization of T-aminobutyrica-oxoglutaric acid transaminase in mouse brain, J. Neurochem., 15 (1968) 351-357.
97 26 Welch, K. M. A., Chabi, E., Bartosh, K., Achar, V. S. and Meyer, J. S., Cerebrospinal fluid ~,aminobutyric acid levels in migraine, Brit. rned. J., 3 (1975) 516-517. 27 Wheal, H. V. and Kerkut, G. A., The action of muscimol on the inhibitory postsynaptic membrane of the crustacean neuromuscularjunction, Brain Research, 109 (1976) 179-183. 28 Wood, J. H., Glaeser, B. S., Enna, S. J. and Hare, T. A., Verification and quantification of GABA in human cerebrospinal fluid, J. Neurochem., 20 (1978) 291-293. 29 Zukin, S. R., Young, A. B. and Snyder, S. H., Gamma-aminobutyricacid binding to receptor sites in the rat central nervous system, Proc. nat. Acad. Sci. (Wash.), 71 (1974) 4802-4807.
89
PHARMACOLOGICAL CHARACTERIZATION OF GABA RECEPTORS MEDIATING VASODILATION OF CEREBRAL ARTERIES IN VITRO
LARS EDVINSSON and DIANA N. KRAUSE
Department of Histology, University of Lund, Lund (Sweden) and Division of Neurosciences, City of Hope National Medical Center, Duarte, Calif. (U.S.A.) (Accepted January 11th, 1979)
SUMMARY
GABA (?,-aminobutyric acid) produced a dose-dependent dilation of isolated cat and dog cerebral artery segments which had been given an active, tonic contraction by either prostaglandin F2a or serotonin. No effect of GABA on extracranial blood vessels was observed. The GABA-induced dilation could be blocked in a dose-dependent manner by either bicuculline or picrotoxin. The latter agent appeared to act as a competitive antagonist. GABA agonists muscimol, imidazoleacetic acid, ~-aminovaleric acid, (-l-)7-amino-fl-hydroxybutyric acid, and fl-alanine also relaxed actively contracted cerebral arteries dose-dependently. The relative potency of these agonists was consistent with that established for GABA receptors on neurons and invertebrate striated muscle. GABA was also tested on two human cerebral arteries and found to cause a small dilation. The results support the existence of a cerebrovascular GABA receptor which may mediate an interaction between GABA and the cerebral circulatory system.
INTRODUCTION
A possible role in the cerebral circulatory system has been proposed recently for the inhibitory neurotransmitter GABA (y-aminobutyric acid). GABA was reported to dilate dog cerebral, but not peripheral, arteries in vitro s. These findings suggest the existence of a cerebrovascular receptor for GABA. An association of GABA with cerebral blood vessels is further indicated by the presence of a non-neuronal form of the enzyme which synthesizes GABA, glutamate decarboxylase (GAD), in relatively high concentrations in pial arteries as compared to peripheral vessels 15,1s. GABA transaminase (GABA-T), the degradative enzyme for GABA also has been demonstrated histochemically in large concentrations around
90 brain vessels, but not around other, e.g. kidney, vessels 24,25. A physiological role for GABA in the regulation of cerebral circulation is supported by a brief Russian report indicating that, in vivo, GABA is capable of decreasing cerebral arterial resistance and increasing cerebral blood flow in cats, rabbits, and dogs 17. In this study, we have further characterized the receptor mediating GABAinduced dilation by examining the effects of various known GABA agonists and antagonists on the mechanical response of cat pial arteries in vitro. For comparison, cerebral arteries from dog and human, as well as extracranial arteries, were also tested for possible GABA responses. MATERIALS AND METHODS
Material Thirteen cats of both sexes (2.0-4.5 kg) and one dog (15 kg) were used. They were anesthetized with sodium pentobarbitol (30 mg/kg i.p.) and killed by exsanguination. The brain was removed, and the middle cerebral, posterior communicating and basilar arteries (300-600/~m in diameter) were dissected out and placed immediately in an oxygenated Krebs-bicarbonate solution. Branches of similar caliber were taken from the external maxillary or lingual arteries. Material from two patients was obtained during neurosurgical tumor operations. (Institutional rules governing the protection of human subjects were followed in obtaining and studying these tissues.) Small pial arteries were removed from macroscopically intact parts of resected brain tissue. The specimens were transported to the laboratory in ice-cold buffer solution.
Vascular preparation The morphology of the vessels and the in vitro conditions for studying their mechanical responses have been determined previously 5,6. Pieces of intra- and extracranial arteries (about 5 mm long each) were used immediately for testing; the remainder were kept in a refrigerator at 4 °C for later use (up to 24 h). The two types of arteries were mounted in a 50 ml temperature-controlled organ bath with two separate systems of L-formed metal holders for recording isometric circular contractions as described previously 5. The bath contained a solution of the following composition (mM): NaC1, 118; KCI, 4.5; CaCI2.2HzO, 2.5; MgSO4.7H~O, 1.0; NaHCOs, 25; KH2PO4, 1.0; and glucose, 6.0. The bath and the stock solution were maintained at 37.5 d: 0.5 °C (range) and aerated continuously with a mixture of 95 % O8 and 5 % COg. Shortly after the arterial preparations had been mounted in the organ bath, each was subjected to a load of 400 dynes and allowed to stabilize; the tension decreased approximately 50--200 dynes. The test drugs were administered after 1.5 h of equilibration. Preliminary tests showed that the GABA agonists dilate cerebral vessels. Under our experimental conditions n, the vessels were almost completely relaxed, and dilator responses were thus difficult to obtain. Therefore vessels were given a resting tone by addition of either 2.5 × 10-6 M prostaglandin F~a (PGF2a) or 3 × 10-s M 5-hydroxytryp-
91 tamine (5-HT). The contraction by either of these two agents, an average of 200 dynes in this study, remained at a steady level for at least 30 min. During this time increasing doses of various GABA agonists were added cumulatively. In blockade experiments, the antagonists were added 15-20 min before the agonist doses and remained in the bath during agonist application.
Analysis of data The effect of the agonists were plotted in terms of response against log dose. The maximum response obtained, EArn (expressed in dynes), and the EDs0 (molar concentration of agonist at which half-maximum response occurs) were used as measures of agonist potency. The results are generally expressed as mean values 4- standard error of the mean. Drugs Bicuculline (Chemicals Procurement Labs), imidazoleacetic acid (Calbiochem), muscimol (Merck Frosst Laboratories), prostglandin F2a (Astra) and taurine (Eastman) were used in this study as well as the following compounds obtained from Sigma: fl-alanine, ?-aminobutyric acid (GABA), (~)?-amino-fl-hydroxybutyric acid, d~aminovaleric acid, glutamic acid, glycine, 5-hydroxytryptamine creatine sulfate (5-HT), and picrotoxin. RESULTS
Response of cat arteries to GABA GABA was found to dilate in a dose-dependent manner intracranial vessels actively contracted by either PGF2a or 5-HT (Fig. 1). This response, however, could GABA concentration (M) 10-8
10-7
10-s
10-s
10-{
O'
102030. 1.0 50 60
I
70. 80 Dilation (dynes) Fig. 1. D o s e - r e s p o n s e curve for the dilator effect o f G A B A tested on segments o f cat middle cerebral arteries actively contracted by either 2.5 × 10 -8 M PGF2a or 3 × 10 -8 M 5-HT. Points represent the m e a n s 4- S.E.M. for 13 experiments.
92
10 8
GABA concentration (M) 10-~ 10-s
10-7
10 ~
10-3
Oo! . . . . . . 40
•
,
°
.
~
•
-
_
6O
80
100
Dilation I%1 Fig. 2. Effect o f p i c r o t o x i n on the dose-response curve f o r the d i l a t o r effect o f G A B A . Isolated cat middle cerebral arteries were pretreated with either 0.3 × I 0 7 M, or 3 :~ l 0 ~ M picrotoxin for 15 min and then given active tone with 2.5 × 10 GM PGE2~ and tested for responses to increasing concentrations of GABA. Each point represents the mean for two experiments.
TABLE I
Dilatory effects o f various G,4BAergic agonists on the cat middle cerebral artery Mechanical responses of isolated segments of cat middle cerebral arteries were recorded as described in Methods. Increasing doses of a G A B A agonist were added cumulatively to the bath of vessel segments which were tonically contracted by approximately 200 dynes with either 2.5 :< 10 6 M PGFe~ or 3 × 10 -.8 M 5-HT. Results are given as means 2_ S.E.M. except for N 2 which is the mean -i standard deviation. N = n u m b e r of experiments, EDs0 dose producing a half-maximum dilation, and Ear, = m a x i m u m dilation.
,4gonist
N
EDso (M)
Muscimol GABA Imidazoleaceticacid 6-Aminovalericacid 7-Amino-fl-hydroxybutyric acid fl-Alanine
6 13 4 2 3 3
3.8 7.7 1.l 2.0 2.3 3.5
2± ± ± ± i
2.5 2.4 0.5 1.4 1.5 3.3
EArn (dynes) x x x × × x
10 -7 10 7 10-~ 10 4~ 10 -G 10 -6
54 59 49 51 47 47
_+ 10 5:I1 5 I1 ~ 4 :L 9 ~: 12
93 not be detected in every vessel; GABA failed to dilate any of the intracranial arteries tested from 4 of the 12 cats studied. In those vessels which did respond (15 out of 33 vessels examined), the GABA dilatory effect, while not strong, was consistent. Characteristically, GABA produced a threshold response at a bath concentration of about 0.33 #M, a half maximum response (EDs0) at 0.8 #M, and a maximum response of approximately 60 dynes dilation at 10-100/~M. Relaxation in response to GABA was reversible and rapid, being complete within 1-2 min. Similar GABA effects were observed with basilar and middle cerebral artery preparations, although most experiments were performed with the latter vessel type. GABA was never found to affect extracranial vessels, nor did it relax intracranial vessels which had been actively contracted with saturated KCI (3 vessel segments tested).
Effect of GABA antagonists The effects of the convulsants picrotoxin and bicuculline, two relatively specific antagonists of GABA responses in neurons3,11,14 and crustacean striated muscle zl, were examined on the GABA-induced dilation of cat middle cerebral arteries. Pretreatment of an intracranial vessel with either picrotoxin or bicuculline at concentrations of 0.3 # M and 3/~M caused a dose-dependent inhibition of the vessel's response to GABA. The antagonists by themselves had no obvious direct effect on the tonic, active contraction induced by PGF2a. Picrotoxin was examined in more detail and appeared to produce parallel shifts in the GABA dose-response curve towards higher agonist doses (Fig. 2), an action characteristic of competitive antagonists.
Effect of GABA agonists Table I summarizes the effects on cat intracranial vessels of several compounds structurally related to GABA that have been well characterized as GABA agonists in neuronal preparations. Muscimol, imidazoleacetic acid, O-aminovaleric acid, (4-)),amino-fl-hydroxybutyric acid, and fl-alanine each caused a dose-dependent relaxation when tested on actively contracted, GABA-responsive vessels at bath concentrations of 10-8-10 -4 M. The maximum dilatory response was similar for all agonists tested, including GABA. When compared on the same vessel preparation, the potency of muscimol was greater than or equal to that of GABA. The other compounds were found to be somewhat less potent than GABA, with fl-alanine generally being the least potent agonist. Other putative amino acid neurotransmitters, glycine, taurine, and glutamic acid, were also tested at 10-4 M on actively contracted vessels, but no response of the vessels could be detected.
GA BA responses in arteries of other species For the purpose of comparison, experiments with GABA were conducted using cerebral arteries obtained from dog and humans. Four dog middle cerebral artery segments contracted by 5-HT were tested, and three exhibited relaxation in response to GABA (10-8-10 -3 M). The responses appeared similar to those observed with cat intracranial vessels. A maximum response of 63 4- 11 dynes dilation was elicited with
94 30-300/~M GABA. Estimated EDs0 values (0.07 ktM, 1.1 #M, 15/zM) varied among the three responding vessels. The dog arteries also dilated in response to muscimol (1 vessel tested) and (-~)7-amino-fl-hydroxybutyric acid (two vessels tested). The potency and maximum dilatory response of both compounds appeared similar to that obtained with G A B A in the same vessel. The effect of G A B A was examined on two human pial arteries (one from each patient) actively contracted by PGFza. A small dilation could be detected in response to GABA at bath concentrations of 0.1-1 #M. One human extracranial vessel was also tested, but no G A B A response was observed. DISCUSSION A recent in vitro study on dog pial arteries 8 suggests that GABA may affect the cerebral vasculature in addition to its known inhibitory neurotransmitter actions on nerve cells and invertebrate striated muscle. Our studies support a direct vasodilatory action of GABA, observed here in isolated cat pial arteries and a small sample of dog and human cerebral arteries. Fujiwara et al. 8 found that most of their isolated dog pial arteries only responded to high G A B A concentrations; however, we obtained from all vessels which responded to G A B A fairly consistent dose-response relationships with an average EDs0 of 7.7 k: 2.4 × 10 -7 M. In our studies the maximum response to GABA was lower in both cat (59 ± 1i) and dog (63 ± 11) vessels than the previously reported value for dog vessels of approximately 200 dynes dilation. In agreement with Fujiwara et al. s, no GABA responses were ever detected in extracranial vessels. Our results are consistent with the presence of a specific GABA receptor on the cerebral vessels which appears analogous to the G A B A receptor found on neurons and invertebrate striated muscle. The G A B A response was blocked in a dose-dependent manner by picrotoxin and bicuculline, known GABA antagonists in vertebrate CNS neurons ~,14 and crustacean neurons 11 and striated muscle 21. Fujiwara et al. s also reported that GABA-mediated dilation in dog was blocked by picrotoxin but not by adrenergic or muscarinic blocking agents. In our studies, picrotoxin appeared to act as a competitive antagonist, producing parallel shifts to the right of the G A B A doseresponse curve. A competitive type of blockade also has been observed for this convulsant in lobster muscle 2~ and the crayfish stretch receptor neuron H, preparations where similar quantitative, physiological dose-response data can be obtained. The existence o f a cerebrovascular G A B A receptor was further confirmed by the GABA-like dilatory action of structurally related compounds which have been characterized as G A B A agonists and which bind directly to the GABA receptor site on neuronal membrane fractions19,2L The relative potency for the agonists tested on the cerebral arteries was consistent with that found for other known GABA receptors. Muscimol was generally the most potent agonist ; this has been demonstrated in many preparations including cat spinal cord neurons3, ~3, crayfish stretch receptor neurons 11, and crustacean neuromuscular junctions ~7. Imidazoleacetic acid, (±)y-amino-fl-hydroxybutyric acid, and 6-aminovaleric acid have been reported to be somewhat less
95 potent than GABA, while fl-alanine has been shown to be an even less effective GABA agonist2,4,9,22, 2a. The cerebral artery response appeared specific for GABA as other putative amino acid neurotransmitters, glycine, taurine, and glutamate had no effect on the vessels. A similar pharmacology is found for the cerebrovascular GABA receptor site when it is identified using [aH]muscimol binding assays (Krause, Wong, Degener and Roberts, in preparation). While the existence of a cerebrovascular GABA receptor suggests a role for this substance in the control of cerebral circulation, the in vitro relaxation produced by GABA and its analogs was not as pronounced as that of other known mediators of vasodilation such as adenosine compounds 10, pH reduction 7, or t-receptor agonists 6. In both our study and that of Fujiwara et al. s, the GABA dilatory effect was undetectable in some of the isolated vessels tested, which may be why Lee et al. failed to observe a response to GABA (3-300/~M) in their isolated cat cerebral arteries 16. The reasons for the lack of response in some vessels is not clear but may relate to possible unfavorable conditions created in vitro or a relatively low number of receptors present in the particular vessels studied. The effectiveness of GABA does not appear to be limited by the GABA-T activity in the vessels, as treatment with a GABA-T inhibitor, aminooxyacetic acid, did not appear to affect the GABA response (unpublished observation). Under certain in vivo conditions however, GABA-mediated vasodilation may be enhanced. GABA has been reported to increase cerebral blood flow in vivo17; and muscimol, when administered intravenously to rats, also produces an increase in local cerebral blood flow as determined by the [14C]ethanol technique (unpublished observation). Cerebral vessels are exposed normally to several sources of GABA in vivo. Brain vessels contain GABA-related enzymeslS,1s,24,~5 and thus are capable of producing GABA themselves. Direct GABAergic innervation of the vessels has never been seen with the specific G A D immunocytochemical technique (Vaughn and Ribak, personal communication), although GABA is released into the extracellular fluid of the brain by neurons and possibly glia2L According to recent reports12, ~s, the cerebrospinal fluid (CSF), which surrounds pial arteries in vivo, contains in humans from about 0.1 to 0.8/zM GABA, concentrations which correspond to the lower range of the doseresponse relationship reported here. Interestingly, elevated CSF levels of GABA have been measured in human patients during certain pathological conditions involving altered blood flow such as migraine headache 26 and cerebral anoxia due to cerebrovascular disease 1. ACKNOWLEDGEMENTS The authors are grateful to Dr. Jan Erik Hardebo and Prof. Christer Owman for helpful discussion and assistance and to Dr. Eugene Roberts for valuable advice and encouragement and for referring us to and kindly translating the Russian articles by Mirzoyan and coworkers. This research was supported by the Swedish Medical Research Council (Grant 04X-732) and USPHS Grants NS-05695 and NS-12116.
96 REFERENCES 1 Achar, V. S., Welch, K. M. A., Chabi, E., Bartosh, K. and Meyer, J. S., Cerebrospinal fluid gammaaminobutyric acid in neurologic disease, Neurology (Minneap.), 26 (1976) 777-780. 2 Bowery, N. G. and Brown, D. A., Depolarizing actions of F-aminobutyric acid and related compounds on rat superior cervical ganglia in vitro, Brit. J. Pharmacol., 50 (1974) 205-218. 3 Curtis, D. R.,Duggan, A. W., Felix, D. and Johnston, G. A. R., Bicuculline, an antagonist of GABA and synaptic inhibition in the spinal cord of the cat, Brain Research, 32 (1971) 69-96. 4 Curtis, D. R. and Watkins, J. C.,The pharmacology ofamino acids related togamma-aminobutyric acid, Pharmacol. Rev., 17 (1965) 347-391. 5 Edvinsson, L., Nielsen, K. C. and Owman, C., Influence of initial tension and changes in sensitivity during amine-induced contractions of pial arteries in vitro, Arch. int. Pharmacodyn., 208 (1974) 235-242. 6 Edvinsson, L. and Owman, C., Pharmacological characterization of adrenergic alpha and beta receptors mediating vasomotor response of cerebral arteries in vitro, Circulat. Res., 35 (1974) 835-849. 7 Edvinsson, L. and Sercombe, R., Influence ofpH and pCO~ on alpha-receptor mediated contraction in brain vessels, Actaphysiol. scand., 97 (1976) 325-331. 8 Fujiwara, M., Muramatsu, I. and Shibata, S., ),-Aminobutyric acid receptor on vascular smooth muscle of dog cerebral arteries, Brit. J. Pharmacol., 55 (1975) 561-562. 9 Godfraind, J. M., Krnjevi6, K., Mareti6, H. and Pumain, R., Inhibition of cortical neurones by imidazole and some derivatives, Canad. J. Physiol. Pharmacol., 51 (1973) 790-797. 10 Hardebo, J. E. and Edvinsson, L., Adenine compounds: cerebrovascular effects in vitro with reference to their possible involvement in migraine, Stroke, 10 (1979) 58-62. 11 Hori, N., Ikeda, K. and Roberts, E., Muscimol, GABA and picrotoxin: effects on membrane conductance of a crustacean neuron, Brain Research, 141 (1978) 364-370. 12 Huizinga, J. D., Teelken, A. W., Muskiet, F. A. J., Meulen, J. V. D., Wolthers, B. G., Identification of GABA in human CSF by gas liquid chromatography and mass spectrometry, New Engl. J. Med., 296 (1977) 692. 13 Johnston, G. A. R., Curtis, D. R., DeGroat, W. C. and Duggan, A. W., Central actions ofibotenic acid and muscimol, Biochem. PharmacoL, 17 (1968) 2488-2489. 14 Krnjevi6, K., Chemical nature of synaptic transmission in vertebrates, Physiol. Rev., 54 (1974) 418-540. 15 Kuriyama, K., Haber, B. and Roberts. E., Occurrence of a new L-glutamic acid decarboxylase in several blood vessels of the rabbit, Brain Research, 23 (1970) 211-123. 16 Lee, T. J.-F., Hume, W. R., Su, C. and Bevan, J. A., Neurogenic vasodilation ofcat cerebral arteries, Circulat. Res., 42 (1978) 535-542. 17 Mirzoyan, S. A. and Akopyan, V. P., The effect produced by gamma-aminobutyric acid on the cerebral circulation and oxygen tension in the brain, Farmakologiya i Toksikologiya (U.S.S.R.), 5 (1967) 572-574. 18 Mirzoyan, S. A.,Kazaran,V.A. andAkopyan, V. P.,Theglutamicdecarboxylaseactivityin blood vessels of the brain, Dokl. Acad. Nauk (U.S.S.R.), 190 (1970) 1241-1243. 19 Olsen, R. W., Greenlee, D., Van Ness, P. and Ticku, M. K., Studies on the gamma-aminobutyric acid receptor/ionophore proteins in mammalian brain. In F. Fonnum (Ed.), Amino Acids ~s Chemical Transmitters, Plenum Press, New York, 1978, pp. 467-486. 20 Sellstr6m,/~. and Hamberger, A., Potassium-stimulated ),-aminobutyric acid release from neurons and glia, Brain Research, 119 (1977) 189-198. 21 Shank, R. P., Pong, S. F., Freeman, A. R. and Graham, L. T., Bicuculline and picrotoxin as antagonists of ~'-aminobutyrate and neuromuscular inhibition in the lobster, Brain Research, 72 (1974) 71-78. 22 Swagel, M. W., lkeda, K. and Roberts, E., Effects of GABA, imidazoleacetic acid, and related substances on conductance of crayfish abdominal stretch receptor, Nature New BioL, 246 (1973) 91-92. 23 Takeuchi, A. and Takeuchi, N., The structure-activity relationship for GABA and related compounds in the crayfish muscle, Neuropharmacology, 14 (1975) 627-634. 24 VanGelder, N. M.,Apossibleenzymebarrier for~,-aminobutyricacidinthecentralnervoussystem, Progr. Brain Res., 29 (1968) 259-268. 25 Waksman, A., Rubinstein, M. K., Kuriyama, K. and Roberts, E., Localization of T-aminobutyrica-oxoglutaric acid transaminase in mouse brain, J. Neurochem., 15 (1968) 351-357.
97 26 Welch, K. M. A., Chabi, E., Bartosh, K., Achar, V. S. and Meyer, J. S., Cerebrospinal fluid ~,aminobutyric acid levels in migraine, Brit. rned. J., 3 (1975) 516-517. 27 Wheal, H. V. and Kerkut, G. A., The action of muscimol on the inhibitory postsynaptic membrane of the crustacean neuromuscularjunction, Brain Research, 109 (1976) 179-183. 28 Wood, J. H., Glaeser, B. S., Enna, S. J. and Hare, T. A., Verification and quantification of GABA in human cerebrospinal fluid, J. Neurochem., 20 (1978) 291-293. 29 Zukin, S. R., Young, A. B. and Snyder, S. H., Gamma-aminobutyricacid binding to receptor sites in the rat central nervous system, Proc. nat. Acad. Sci. (Wash.), 71 (1974) 4802-4807.