Calcium-activated potassium channels mediate prejunctional ...

Proc. Nati. Acad. Sci. USA

Vol. 89, pp. 1325-1329, February 1992

Physiology

Calcium-activated potassium channels mediate prejunctional inhibition of peripheral sensory nerves (charybdotoxin/airways/neurogenic inflammation/C fibers)

DAVID STRETTON, MOTOHIKO MIURA, MARIA G. BELVISI, AND PETER J. BARNES Department of Thoracic Medicine, National Heart and Lung Institute, London SW3 6LY, United Kingdom

Communicated by L. L. Iversen, October 21, 1991 (received for review June 18, 1991)

ABSTRACT Activation of several receptors, including jt-opioid, a2-adrenergic, and neuropeptide Y receptors, inhibits excitatory nonadrenergic noncholinergic (NANC) neural responses in airways, which were mediated by the release of peptides from capsaicin-sensitive sensory nerves. This raises the possibility of a common inhibitory mechanism, which may be related to an increase in K+ conductance in sensory nerves. To examine this hypothesis, we have studied whether K+channel blockers inhibit the effects of neuromodulators of sensory nerves in guinea pig bronchi by using selective K+channel blockers. Charybdotoxin (ChTX; 10 nM), which blocks large conductance Ca2+-activated K+-channel function, completely blocked and reversed the inhibitory effects of a jA-opioid agonist, neuropeptide Y. and an a2-adrenoceptor agonist on excitatory NANC responses. Neither inhibitors of ATP-sensitive K+ channels (BRL 31660 or glibenclamide, both at 10 bsM) nor an inhibitor of small conductance Ca2+-activated K+ channels (apamin; 0.1 ,M) were effective. This suggests that ChTX-sensitive K+-channel activation may be a common mechanism for the prejunctional modulation of sensory nerves in airways. This may have important implications for the control of neurogenic inflammation.

g-opioid agonist [D-Ala2,N-MePhe4,Gly-o15]enkephalin

(DAMGO), the a2-agonist clonidine, and NPY on e-NANC responses in guinea pig bronchi in vitro. These agonists were chosen as they do not have direct effects on airway smooth muscle tone. We also examined effects of other K'-channel blockers such as apamin, a protein extracted from bee venom and a blocker of small conductance K'a (25), and BRL 31660 and glibenclamide, blockers of KATP (26, 27), under the same conditions.

MATERIALS AND METHODS Animal Preparation and Experimental Protocol. Male Dunkin Hartley guinea pigs (200-350 g) were killed by cervical dislocation, and the lungs with trachea and bronchi were rapidly removed. The parenchyma was dissected away to reveal two main bronchi and four hilar bronchi, and the rings of these bronchial tissues were suspended in KrebsHenseleit solution of the following composition: 118 mM NaCl, 5.9 mM KCl, 1.2 mM MgSO4, 2.5 mM CaC12, 1.2 mM NaH2PO4, 25.5 mM NaHCO3, and 5.6 mM glucose. Indomethacin was present throughout at a concentration of 10 AuM, in order to prevent modulation of neural responses by endogenous prostaglandin production. The solution was maintained at 37°C and gassed continuously with a mixture of 95% 02/5% C02, giving a pH of 7.4. The tissues were allowed to equilibrate for 1 h, with frequent washing, under a resting tension of 500 mg, which was found to be optimal for measuring changes in tension in these airways. Isometric contractile responses were measured by using Grass FT.03 force-displacement transducers and recorded on a polygraph (Grass model 7D; Grass). In the presence of atropine and propranolol (both 1 jam), e-NANC bronchoconstrictor responses elicited to EFS (40 V; 0.5 ms; 8 Hz for 15 s) were studied by suspending tissues between platinum wire field electrodes. We chose the frequency of 8 Hz because this frequency causes '50% of maximal responses (Fig. 1) and is appropriate to detect the modulatory effects on e-NANC responses in guinea pig bronchi. Biphasic square wave pulses were delivered from a Grass S88 stimulator. Tissues were stimulated every 30 min or when the response had returned to its baseline level. The contractile responses to EFS were completely abolished by tetrodotoxin (1 AM), confirming that they were of neural origin. Pretreatment of the guinea pig with capsaicin (28) also abolished contractile responses, confirming that it was due to release of transmitter from unmyelinated sensory nerves. Effect of K+-Channel Blockers. Once three identical responses had been obtained, the effect of ChTX alone on e-NANC responses was examined. Second, the inhibitory

Electrical field stimulation (EFS) elicits excitatory nonadrenergic noncholinergic (e-NANC) neural bronchoconstrictor responses in guinea pig airways, which are mediated by the release of tachykinins from unmyelinated sensory nerves (1). This e-NANC response has been shown to be inhibited at a prejunctional level by the activation of several receptors such as ,-opioid (2-5), a2-adrenergic (6), y-aminobutyric acid type B (7), histamine H3 (8), neuropeptide Y (NPY) (9, 10), galanin (11), adenosine A2 (12), and 232-adrenergic (13) receptors. The mechanism by which prejunctional receptor activation leads to inhibition of neuropeptide release is unclear, and the presence of so many inhibitory prejunctional receptors on sensory nerves suggests that there may be a common molecular mechanism of action (14). In the central nervous system, some of these agonists inhibit neuronal function by opening a common K+ channel (15, 16), raising the possibility that a similar mechanism may operate in peripheral nerves. Cromakalim, an activator of ATP-sensitive K+ channel (KATP) (17, 18), is indeed able to inhibit e-NANC responses and this effect is blocked by glibenclamide (19), suggesting that a KA2TP is involved. Another K+-channel blocker, charybdotoxin (ChTX), extracted from the venom of the scorpion Lejurus quinquestriatus var. habraeus, has been shown to block large conductance Ca2+-activated K+ channels (Kca) in various tissues, including nerves (20-23). Furthermore, a K+ channel cloned from a hippocampal neuronal library is sensitive to ChTX (24). In this study, we examined whether ChTX inhibited the neuromodulatory action of the

Abbreviations: EFS, electrical field stimulation; e-NANC, excitatory nonadrenergic noncholinergic; NPY, neuropeptide Y; ChTX, charybdotoxin; KAP, ATP-sensitive K+ channel(s); KL, Ca2+activated K+ channel(s); DAMGO, [D-Ala2,N-MePhe4,Gly-

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

ol5]enkephalin.

1325

1326

Proc. Nati. Acad. Sci. USA 89 (1992)

Physiology: Stretton et al. 100 -

C) 0 0 w

LK

to

,

DAMGO 10 7M

.-

50 -

~~~CT

ao~

J ~~

0

U

E

~

DAMGO 10

~

ChTX

7M

08

10-8M

aS

C x

'u

a

m

0-

Clonidine 10 7M

1

2

4

8

16

32

64

U

ChTj 10 8M

500 mg |

d

-

Stimulating frequencies (Hz) FIG. 1. Frequency-response curves of e-NANC constrictor responses elicited by EFS (40 V; 0.5 ins; 1-64 Hz for 15 s) in guinea pig bronchi in vitro. All responses were obtained in the presence of atropine and propranolol (both at 1 uM). Constrictor responses to each frequency were expressed as a percentage of maximal constriction to 10 mM acetylcholine (ACh). Mean ± SEM values of six airways are shown.

effects of DAMGO, clonidine, or NPY (all at 0.1 ,uM, which has previously been found to have a significant inhibitory effect) on e-NANC responses were examined in the presence or absence of ChTX (10 nM), using an incubation time of 20 min, which had previously been shown to give maximal inhibition. Effects of additional concentrations of ChTX (1 and 30 nM) on DAMGO-induced inhibitory modulation of e-NANC responses were also examined. In other experiments, reversal of the modulatory effects of the agonists on e-NANC responses after the addition of ChTX was examined. The effects of the opioid antagonist naloxone (1 ,M) on the modulatory responses to DAMGO and the a2 antagonist idazoxan (1 ,uM) on the response to clonidine confirmed that opioid and a2-receptors were involved, respectively. In this series of experiments, we also confirmed by using matched control tissues that modulatory effects of DAMGO, clonidine, and NPY persisted for the duration of the experiment. We also studied the effect of the active trans enantiomer of cromakalim [the (-)-3S, 4R form], BRL 38227 (1 ,uM) (29) on e-NANC responses and the effect of KA blockers BRL 31660 and glibenclamide (both 10 ,uM) on this response. We then studied the effect of the same concentration of BRL 31660 and glibenclamide on inhibitory modulation by DAMGO, clonidine, and NPY to e-NANC responses. In another set of experiments, the effect of apamin (0.1 ,uM) on DAMGO-induced e-NANC modulation was also examined. Contrtle-Response Studies. To confirm that modulatory responses obtained in the present study are not due to postjunctional mechanisms, the effects of ChTX, BRL 31660, DAMGO, clonidine, and NPY on the contractile responses induced by exogenous substance P (1 nM to 0.1 mM) were also studied. Drugs. Drugs and chemicals were obtained from the following sources: indomethacin, NPY, substance P, DAMGO, atropine sulfate, propranolol hydrochloride, apamin, glibenclamide, naloxone, idazoxan, tetrodotoxin (Sigma), ChTX (Receptor Research Chemicals, Baltimore), clonidine (Boehringer Ingelheim). BRL 31660 and BRL 38227- were kindly supplied by SmithKline Beecham.

5min K

a

a

NPY 10 7M

ChTX 10 8M

e FIG. 2. Typical e-NANC bronchoconstrictor responses elicited by EFS (n) (40 V; 0.5 ms; 8 Hz for 15 s) in guinea pig bronchi in vitro. (a) Lack of effect of ChTX (10 nM) on e-NANC responses. (b) Lack of effect of the g-opioid agonist DAMGO on e-NANC responses in the presence of ChTX (10 nM). (c) Prejunctional inhibition of e-NANC responses produced by DAMGO (0. 1.uM) and reversal by ChTX (10 nM). (d) Inhibitory effect of clonidine (0.1 ,uM), reversal by ChTX (10 nM). (e) Inhibitory effect of NPY (0.1 1&M), reversal by ChTX (10 nM). Added drugs were continuously present after each arrow.

by tetrodotoxin. DAMGO, clonidine, and NPY (all at 0.1 ,uM) attenuated the EFS-induced e-NANC response by 52.6% ± 7%, 54.3% ± 8%, and 25.3% ± 2%, respectively (n = 7; Fig. 3). ChTX (10 nM) alone had no effect on e-NANC responses (Fig. 2a) but prevented the inhibition of e-NANC responses by DAMGO (Fig. 2b), clonidine, and NPY and also completely reversed the inhibition induced by these three compounds (Fig. 2 c-e). In the presence ofChTX (10 nM), the agonists did not show any significant inhibitory nerve modulation (Fig. 3). However, ChTX (1 nM) had no significant effect on the modulation of e-NANC responses by DAMGO; inhibitions of e-NANC responses by DAMGO in the presence of 1, 10, and 30 nM were 45.9% ± 9.6% (n = 3; P > 0.05), 5.8% ± 4.5% (n = 7; P < 0.005), and 4.4% ± 3.4% (n = 5; P < 0.005), respectively (Fig. 4). 0 .°

60-

40. c

0

o

40-

z

z I

0O o

20

C

0 .0

C

0 v_

RESULTS In the presence of atropine and propranolol (both at 1 I.M), EFS of guinea pig bronchial rings gave reproducible e-NANC bronchoconstrictor responses (Fig. 2), which were inhibited

I

-

I1= [L

DAMGO

CLONIDINE

NPY

FIG. 3. Histograms showing percentage inhibition of e-NANC responses by DAMGO, clonidine, and NPY (all at 0.1 ,M) in the presence (hatched bars) and absence (open bars) of ChTX (10 nM). Mean values + SEM of seven airways are shown. Asterisks indicate significant difference from values without ChTX (***, P < 0.005).

Physiology: Stretton et al. 0

Proc. Natl. Acad. Sci. USA 89 (1992) a

60 -

.° 80-

I

4_

L.

C

0

0

0 0

40 -

60-

IL

0

z

z

z

o Z

6 40

T

0

20 0

42 0

.0 _E

._ ._

_

1327

**

._

r-0V-

0

_HC

C

_

0

BRL38227

0

0

0

The KATP activator BRL 38227 (1 ,uM) produced a 68.0% ± 8.3% (n = 7) inhibition of the e-NANC response and this

attenuation was completely inhibited and reversed by BRL B6

b

DAMGO

CLONIDINE

C) -C

FIG. 4. Histograms showing percentage inhibition of e-NANC responses by DAMGO (0.1 !LM) in the presence (hatched bars) and absence (open bar) of various concentrations of ChTX (1, 10, and 30 nM). Mean values + SEM of five to seven airways are shown. Asterisks indicate significant difference from values without ChTX (***, P < 0.005).

a

-j

+7A

-

0-

FIG. 6. Histograms showing percentage inhibition of e-NANC responses by BRL 38227 (1 juM), DAMGO (0.1 ,AM), and clonidine (0.1 AM) in the presence (hatched bars) and absence (open bars) of ATP-sensitive K+-channel blocker BRL 31660 (10 AM). Mean values ± SEM of seven airways are shown. Asterisks indicate significant difference from values without BRL 31660 (**, P < 0.01).

31660 (10 AM) (Figs. 5 a and b and 6) and glibenclamide (10 ,uM). BRL 31660 and glibenclamide (both at 10 AM) had no effect on e-NANC responses. The inhibitory effects of DAMGO, clonidine, and NPY were unaffected by BRL 31660 (Figs. 5 c and d and 6) and glibenclamide, suggesting that the prejunctional inhibition of the e-NANC response induced by BRL 31660

BRL 38227 106M + ~ ~~ ~ ~ ~ ~~~~m |5

Smin

C

d_

nAMArfr

\

Clonidine

IX

10

7M

7M

BRL 31660 10 5M

BRL 31660 10 5M

Naloxone 10o6M

Idazoxan 106 M

Naloxone 10 e L6 -7 Apmi 1-.7M

FIG. 5. Typical traces showing the effect of the ATP-sensitive K+-channel activator BRL 38227 (1 ,M) and blocker BRL 31660 (10 AM) and small conductance K&a blocker apamin (0.1 ILM) on the e-NANC bronchoconstrictor responses elicited by EFS (m) (40 V; 0.5 ms; 8 Hz for 15 s) in guinea pig bronchi in vitro. (a) Inhibition of the e-NANC response by BRL 38227 and reversal by BRL 31660. (b) Lack of effect of BRL 38227 on e-NANC responses in the presence of BRL 31660. (c and d) Inhibition of e-NANC responses by DAMGO and clonidine, respectively, showing the inability of BRL 31660 to reverse these effects. (e) Inhibition of e-NANC responses by DAMGO, showing the inability of apamin to reverse this effect. Inhibitions by DAMGO and clonidine persisted sufficiently during the experiment and were reversed by the respective antagonists naloxone and idazoxan (both 1 A&M). Added drugs were continuously present after each arrow.

1328

Physiology: Stretton et al.

Proc. Natl. Acad. Sci. USA 89 (1992)

100

c) 0 0 4-1 C c

0 m

50

-

E x

0-1 FIG. 7. C

-9

-8

-5 -7 -6 log [Substance P (M)]

-4

oncentration-response curves constructed to exoge-

nous substan< in absence (ope r circles) of ChTX (10 M)tguresence (spolid Circles) and

Each point r(epresents a mean value ± SEM from seven animals. ACh, acetylc boline.

these three agonists does not occur through the activation of a common]KATP on sensory nerves. Apamin (I0.1 AuM) was also without effect on the inhibition of e-NAN( responses induced by DAMGO (Fig. 5e). DAMGO-in duced inhibition of e-NANC responses in the presence of apamin was 49.8% ± 8.3% (n = 5), which was not significantly) V different from that without apamin (P > 0.05). ChTX (10 nM) had no effect on the concentration-response relationship constructed to exogenous substance P (1 nM to 0.1 mM) (Fiig. 7); mean ± SEM values of EC50 were 3.2 ± 0.2 ,uM withoutt and 2.7 ± 0.2 AM with ChTX (n = 7; P > 0.05). Similarly, nieither BRL 31660, DAMGO, clonidine, nor NPY had any effe*ct on the concentration-response relationships to substance P',with EC50 values of 2.6 ± 0.3, 3.2 ± 0.3, 3.0 ± 0.3, and 2.5 ± 0.2 ,uM, respectively (n = 5).

DISCUSSION We have d(emonstrated that ChTX blocks prejunctional inhibitory efifects of a ,u-opioid, a2-agonist, and NPY on e-NANC br7onchoconstriction, suggesting that, at least in the airwavs. nr-eiunctional inhibition of the e-NANC resnonse following prejunctional receptor occupation on sensory nerves is achieved through the coupling of prejunctional receptors to a common ChTX-sensitive K+ channel. This report demonstrates the presence of a common pathway in inhibitory modulation of peripheral sensory nerves. It has been reported that in the central nervous system the stimulation of more than two receptors coexisting on the same neuron causes an increase in K+ conductance through the common intracellular mechanism (15, 16). The present study provides functional evidence that the activation of a large conductance K'a is involved in inhibitory modulation of sensory nerves. ChTX is a relatively selective antagonist of large conductance K&a (20), but it might have additional effects on other channels. A neuronal K+ channel that is sensitive to ChTX has been cloned and is presumably the channel relevant to the neuromodulatory effects described here (24). It has been reported that the KATP channel opener cromakalim inhibits e-NANC responses prejunctionally in vivo in guinea pigs (19). We have now confirmed this in vitro by using the active enantiomer of cromakalim, BRL 38227 (lemakalim). This inhibitory effect of BRL 38227 was completely blocked by the K'TP antagonist glibenclamide (27) and BRL 31660(26), indicating that KATP are able to modulate e-NANC responses. However, the present study also dem-

onstrates that KLTP are not involved in inhibitory neural modulation by agonists such as A-opioids, a2-agonists, and NPY. The presence of apamin-sensitive K+ channels (small conductance Ka) in nerves has also been reported (30). The fact that apamin at a relevant concentration failed to reverse the inhibitory modulation of DAMGO on e-NANC responses indicates that small conductance K&a are not involved in this common mechanism of inhibitory modulation. We chose to examine those agonists that have previously been found to inhibit e-NANC responses in guinea pig bronchi without any effect on airway smooth muscle responses, so that an effect on sensory nerves was not in doubt (2-5, 9). Recently, opioids have been shown to directly inhibit release of substance P from airway sensory nerves (31). None of these agonists had any effect on substance P-induced bronchoconstriction, and ChTX was also without effect on airway smooth muscle contraction. Sensory neuropeptides in airways mimic many of the pathological features of bronchial asthma, including airway smooth muscle contraction, edema, mucus hypersecretion, and plasma extravasation (32), and there is circumstantial evidence that they may contribute to inflammation in asthmatic airways (33). A common mechanism of prejunctional inhibition of sensory nerves may therefore be of potential therapeutic importance in diseases such as asthma and rhinitis, where the exaggerated effects of sensory neuropeptides released via an axon reflex mechanism have been implicated (14). Moreover, the observation obtained in the present study may be relevant to other sensory nerves outside the respiratory tract. Substance P, which is one of the sensory neuropeptides in airways, may be released from the terminals of primary sensory neurons in the spinal cord (34). Opiate analgesics are able to suppress the stimulus-evoked release of substance P in spinal cord (35). Since opiates are known to increase K+ conductance of nerve cells in the central nervous system (15, 16), it is possible that the same type of K+channel activation is involved in the modulation of pain sensation. Similarly, release of neuropeptides from sensory nerves may be important in inflammatory diseases of other peripheral organs, including joints, gastrointestinal tract, skin, eyes, and bladder (36, 37). The demonstration of a common intracellular mechanism for inhibition of neuropeptide release from sensory nerves may therefore have therapeutic implications for a variety of human diseases. 1. Grundstrom, N., Andersson, R. G. G. & Wikberg, J. E. S. (1981) Acta Pharmacol. Toxicol. 49, 150-157. 2. Frossard, N. & Barnes, P. J. (1987) Eur. J. Pharmacol. 141, 519-522. 3. Belvisi, M. G., Chung, K. F., Jackson, D. M. & Barnes, P. J. (1988) Br. J. Pharmacol. 95, 413-418. 4. Bartho, L., Amann, R., Saria, A., Szolcsanyi, J. & Lembeck, F. (1987) Naunyn-Schmiedebergs Arch. Pharmakol. 336, 316320. 5. Belvisi, M. G., Stretton, C. D. & Barnes, P. J. (1990) Br. J. Pharmacol. 100, 131-137. 6. Grundstrom, N., Andersson, R. G. G. & Wikberg, J. E. S. (1984) Acta Pharmacol. Toxicol. 54, 8-14. 7. Belvisi, M. G., Ichinose, M. & Barnes, P. J. (1989) Br. J. Pharmacol. 97, 1225-1231. 8. Ichinose, M. & Barnes, P. J. (1989) Eur. J. Pharmacol. 174, 49-55. 9. Matran, R., Martling, C.-R. & Lundberg, J. M. (1989) Eur. J. Pharmacol. 163, 15-23. 10. Stretton, C. D., Belvisi, M. G. & Barnes, P. J. (1990) Neuropeptides 17, 163-170. 11. Giuliani, S., Amann, R., Papini, A. M., Maggi, A. & Meli, A. (1989) Eur. J. Pharmacol. 163, 91-96. 12. Verleden, G. M., Belvisi, M. G., Stretton, D. & Barnes, P. J. (1991) Am. Rev. Respir. Dis. 143, A357 (abstr.). 13. Verleden, G. M., Belvisi, M. G., Rabe, K. & Barnes, P. J. (1991) Am. Rev. Respir. Dis. 143, A357 (abstr.).

Physiology: Stretton et al. 14. Barnes, P. J., Belvisi, M. G. & Rogers, D. F. (1990) Trends Pharmacol. Sci. 11, 185-189. 15. North, R. A. (1986) Trends Neurosci. 9, 114-117. 16. North, R. A., Williams, J. T., Surprenant, A. & Christie, M. J. (1987) Proc. Nati. Acad. Sci. USA 84, 5487-5491. 17. Cook, N. S. & Hales, C. N. (1984) Nature (London) 311, 271-273. 18. Arch, J. R. S., Buckle, D. R., Bumstead, J., Clarke, G. D., Taylor, J. F. & Taylor, S. G. (1988) Br. J. Pharmacol. 95, 763-770. 19. Ichinose, M. & Barnes, P. J. (1990) J. Pharmacol. Exp. Ther. 252, 1207-1212. 20. Gimenez-Gallego, G., Navia, M. A., Reuben, J. P., Katz, G. M., Kaczorowski, G. J. & Garcia, M. L. (1988) Proc. Natl. Acad. Sci. USA 85, 3329-3333. 21. Miller, C., Moczydlowski, E., Latorre, R. & Phillips, M. (1985) Nature (London) 313, 316-318. 22. Hermann, A. & Erxleben, C. (1987) J. Gen. Physiol. 90, 27-47. 23. Reinhart, P. H., Chung, S. & Levitan, I. B. (1989) Neuron 2, 1031-1041. 24. Christie, M. J., Adelman, J. P., Douglass, J. & North, R. A. (1989) Science 244, 221-224. 25. Banks, B. E. C., Brown, C., Burgess, G. M., Burnstock, G.,

Proc. Nati. Acad. Sci. USA 89 (1992)

26. 27.

28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

1329

Claret, M., Cocks, T. M. & Jenkinson, D. H. (1979) Nature (London) 282, 415-417. Taylor, S. G., Foster, K. A., Shaw, D. J. & Taylor, J. F. (1989) Br. J. Pharmacol. 98, 881P. Cook, D. L. & Hales, C. N. (1984) Nature (London) 311, 271-273. Lundberg, J. M., Brodin, E. & Saria, A. (1983) Acta Physiol. Scand. 119, 243-252. Black, J. L., Armor, C. L., Johnson, P. R. L., Alouan, L. A. & Barnes, P. J. (1990) Am. Rev. Respir. Dis. 142, 1384-1389. Blatz, A. L. & Magleby, K. L. (1986) Nature (London) 323, 718-720. Ray, N. I., Jones, A. J. & Keen, P. (1991) Br. J. Pharmacol. 102, 797-800. Barnes, P. J. (1991) Am. Rev. Respir. Dis. 143, s28-s32. Barnes, P. J. (1986) Lancet i, 242-245. Konishi, S., Tsunoo, A. & Otsuka, M. (1981) Nature (London) 294, 80-82. Yaksh, T. L., Jessell, T. M., Gamse, R., Mudge, A. W. & Leeman, S. E. (1980) Nature (London) 286, 155-157. Holtzer, P. (1988) Neuroscience 24, 739-768. Maggi, C. A. & Meli, A. (1988) Gen. Pharmacol. 19, 1-43.