Nerve Stimulation Releases Mucosa-derived Inhibitory Factors, Both Prostanoids and Nonprostanoids, in Isolated Ferret Trachea1-3
ANDERS ULLMAN, CLAES-GORAN LCFDAHL, NILS SVEDMYR, and BENGT-ERIC SKOOGH
Introduction Since the initial observation of modulatory mechanisms in the airway mucosa and epithelium regulating bronchomotor tone and bronchial responsiveness (1, 2), much interest has focused on this phenomenon. The release of modulatory factors derived from the epithelium and/or other elements of the mucosal tissue has been demonstrated in vitro in a number of species, including the human (3-12). In many of these reports only the basal release of these factors has been studied, although an activation of the release by p-adrenoreceptor agonists, for example, has been suggested (13). Greater knowledge of the mechanisms regulating the release of these factors is important when it comes to understanding the physiologic and possible pathophysiologic role of these phenomena. We have previously demonstrated the release of a mucosa-derived factor(s), inhibiting as much as some 50010 of the contractile responses to supramaximal cholinergic nerve stimulation in isolated ferret trachea (12). The release was activated by phasic stimulations with alternating electrical field stimulation (EFS) and direct vagal nerve stimulation (DNS), indicating an interaction between the neuralsystem and the airway mucosa (12). One factor mediating this effect is obviously a prostanoid as it is significantly reduced, but not abolished, by indomethacin·(12). If there is a neural activation of this mucosa-dependent inhibition, there may be a difference in the magnitude of the inhibition if the stimulation is given only by DNS of the vagal nerve compared with the more general and postganglionic nerve stimulation by EFS. Another possibility that should be considered is that the regional differences in the neural anatomy of the airways influence the inhibition, which could be one explanation of the previously reported observations of dif748
SUMMARY The Influence of an Intact mucosa layer In Isolated ferret trachea on the contractile responses to repeated, short-lasting (20 s, every 2 min) nerve stimulations was studied In a nervemuscle preparation stimulated to chollnerglc-evoked contractions by either direct vagal nerve stimulation (DNS)or transmural electrical field stimulation (EFS). The contractile responses were monitored by three strain gauges connected to the proximal, middle, and distal segments of the trachea. The mucosa was either left Intact (M +)or removed from the membranous part by dissection (M-). The successive decrease In contractile responses was studied for 60 min using repeated DNS and for a further 60 min using alternating EFS and DNS. During the first period, In which DNS alone was used, there was a significantly more pronounced decrease In the M+preparation compared with the M-. This effect was most prominent In the proximal part of the trachea and was not blocked by Indomethacin. During the subsequent stimulation period In which alternating EFS and DNS were used, the rate of decrease was significantly greater In all segments. In this phase, however, the Inhibitory mucosa-dependent effect was significantly attenuated by Indomethacin treatment. In conclusion, this stUdy demonstrates that the mucosa-dependent Inhibition or relaxation In ferret trachea Is mediated by both a prostanold and a nonprostanold factor. The nonprostanold factor but not the prostanold factor can be activated by DNS,whereas the release of the prostanold Is activated by EFS, which Indicates that the activation of the prostanold and nonprostanold factor Is controlled by different mechanisms. AM pI=''' RESPIR DIS 1990; 141:748-751
ferent magnitudes of the mucosa-dependent inhibition at different levels in the airways (11, 13, 14). The aim of the present study was to characterize the mechanisms that regulate the release of the mucosa-derived inhibitory factor in ferret trachea. Wetherefore tried to determine whether the release of an inhibitory factor, prostanoid and/or nonprostanoid, could be activated by DNS alone. If this were so, we also set out to examine whether the inhibition varies between the proximal and distal parts of the trachea. Methods The study was performed in vitro using a nerve-muscle preparation of ferret trachea (15-17). Male ferrets were rendered unconscious by electric shocks and were killed by exsanguination. The study was approved by The Gothenburg University Ethics Committee for Animal Experiments. The trachea, with intact right recurrent and vagus nerves, was rapidly removed and immersed in an organ bath filled with 200 ml Krebs Ringer solution (KR) of the following composition (in mM): NaCI 118, KCI 5.9, CaCI 2.5, MgS0 4 1.2, NaH 2P04 1.2, NaHC03 25.5, and glu-
cose 5.6, maintained at 38° C and gassed with 94070 O 2 and 6070 CO 2 • Ascorbic acid was added to the KR in a concentration of 0.3 mM as an antioxidant. The bath was continuously flushed with prewarmed KR, 10 ml/min. In the organ bath the trachea was split along the anterior long axis and opened. One side was attached by pins, and the opposite side was connected to three strain gauges (Grass FT 03) to record isometric muscle tension at optimum resting tension (7 to 10g). The contractile responses wererecorded from the three strain gauges, one distal, one middle, and one proximal. Comparisons were made for the three segments separately. In some preparations the mucosa covering the tracheal muscle was removed by cutting
(Received in originalform February 7, 1989 and in revised form August 9, 1989) 1 From the Departments of Clinical Pharmacology and Pulmonary Medicine, Gothenburg University, Sweden. 2 Supported by grants from the Swedish Association Against Heart and Chest Diseases. 3 Correspondence and requests for reprints should be addressed to Dr. A. Ullman, Department of Clinical Pharmacology, Sahlgrenska Hospital, S-41345 Gothenburg, Sweden.
749
NERVE STIMULATION RELEASES MUCOSA-DERIVED INHIBITORY FACTORS IN ISOLATED FERRET TRACHEA
along the borders to the cartilaginous part on both sides of the membranous part, after which the mucosa could easily be peeled off. Care was taken not to damage the mucosa in other parts of the trachea. Phasic contractions were evoked by nerve stimulation for 20 severy 2 min. This was done either by direct nerve stimulation alone, or by alternating electrical field stimulation and DNS. EFS was given via two platinum electrodes, one of which was placed above and one below the tracheal muscle. DNS was applied to the vagal nerve and to the laryngeal end of the recurrent nerve simultaneously via two suction electrodes.The electrical impulses were biphasic square waves with a duration of 0.25 ms (DNS) or 0.5 ms (EFS). The frequencies used were 2 and 12Hz for DNS and 12 Hz for EFS. These DNS frequencies induce contractile responses of about 50 and 90010, respectively, of the maximum response to DNS, which is achieved at 24 Hz. The current for DNS was supramaximal (22 rnA). The current for EFS was submaximal (600 to 1000 rnA) and was adjusted so that the responses matched the initial responses to 12 Hz DNS. For statistical evaluation the response to 2 Hz DNS was used as primary effect parameter. In each experiment two stimulation sequences wereused. Initially stimulations were given by "DNS only" (figure 1) to establish the baseline response during 30 min and for 60 min further. Thereafter, the stimulation sequence was changed to also include EFS (EFS/DNS, figure 1). Experiments werecarried out both on preparations with intact mucosa (M+) and on preparations with the mucosa removed over the tracheal muscle (M-). M+ experiments were performed in the presence and absence of indomethacin (10 J,1M). Indomethacin per se does not affect the baseline response or the degree of decay in mucosa-denuded preparations in this experimental model (12). A control series was performed to evaluate the degree of spontaneous decay of contractile responses. In this series the mucosa was left intact. Every 16 min one 12-Hz and one 2-Hz DNS stimulation was given. After an equilibration period of 60 min the electricalstimulations (DNS alone) werestart-
Period 1
Period 2
''DNS only"
"EFS/DNS"
=ig. 1. The stimulation sequences during the first and second period of each experiment. EFS, electrical field stlmulation; DNS, direct nerve stimulation of the vagal
ierve,
TABLE 1 BASELINE CONTRACTILE RESPONSES TO 2-Hz DIRECT NERVE STIMULATION Mean (g)
SEM
M+ MM+/lndo
28 27 25
3.2 3.5 1.8
Proximal
M+ MM+/lndo
32 32 31
1.8 2.3 2.7
Middle
M+ MM+llndo
30 27 24
5.5 4.3 1.1
Distal
M+ MM+llndo
21 19 21
3.3 3.5 2.7
Nerve All segments n =6 n =5 n =6
Definition of abbreviations: intact mucosa = M+; removed mucosa = M-; Indo = indomethacin, 10 JIM.
ed. The contractile responses 30 min after the start of stimulation wereused as baseline. The successive decrease in contractile force was then observed for 60 min after baseline measurements, after which stimulations weregiven in the second sequence (alternating DNS and EFS) for a further 60-mln period. The results are expressed as the mean ± SEM. For unpaired comparisons a two-way analysis of variance was used, followed by Fisher's least-significant difference test. For paired comparisons Student's t test for paired samples was used; p < 0.05 was considered significant.
Results
icantly enhanced the rate of decrease in the contractile responses. The decrease in indomethacin-treated preparations with intact mucosa was similar to that in mucosa-denuded preparations but significantly (p < 0.01) less pronounced than in untreated preparations with intact mucosa (figure 2). Thus DNS activates a mucosa-dependent inhibitory factor that is not blocked by indomethacin, whereas EFS releases an additional mucosa-dependent inhibitory factor that can be blocked by indomethacin. To elucidate possible segmental differences in the distribution of the inhibitory effects we also evaluated the contractile responses separately for the three segments. The proximal segment showed a larger baseline contractile response compared with the distal segment in both preparations with intact and with removed mucosa. Despite this finding, the mucosa-dependent inhibition of the contractile responses activated by DNS alone was more pronounced in the proximal segment compared to the distal in both indomethacin-treated and untreated preparations (p < 0.05). The percentage decrease in mucosa-denuded preparations was similar in the proximal and distal segments (figure 3). In the middle segment the magnitude of the mean decrease was between that of the proximal and distal segments, but with a considerably greater variation. Af-
The analysis that follows is based on the response to 2 Hz DNS as the primary ef100 fect parameter. The baseline contractions in preparations with and without mucosa as well as with intact mucosa after indomethacin treatment weresimilar (table 1). In control preparations with intact ~ mucosa only stimulated by DNS every := 80 16 min, the successive decrease during a ~ period of 60 min was only 3 ± 2070, there- ~ by indicating low spontaneous decay. f+-I During the first stimulation period 0 with DNS-only (2 and 12Hz every2 min), ~ ~' 60 the successive decrease in the contractile responses was significantly (p < 0.01) more pronounced in preparations with mucosa (19 ± 30/0) compared with mucosa-denuded preparations (8 ± 2070). "DNS only" " "EFSjDNS" 40 This difference was not significantly afBaseline 60 min 120min fected by indomethacin (figure 2). In mucosa-denuded preparations, the Fig. 2. Contractile response to 2-Hz DNS (mean ± addition of EFS (12 Hz) had little effect SEM, n = 5-6) as a percentage of baseline response on the decrease rate of the contractions, in preparations with intact mucosa (M+;e), removed mucosa (M-;O), and indomethacin-treated preparations compared with the rate when stimulated with intact mucosa (M +lIndo; .A). Stimulations were givby DNS alone. In preparations with in- en with DNS only during the first period and by EFSlDNS tact mucosa, the addition of EFS signif- alternating during the second period.
ULLMAN, LOFDAHL, SVEDMYR, AND SKOOGH
750
er magnitude than the inhibition by the combination of DNS and EFS (12). This study was performed to characterize the mechanisms by which the mucosa-dependent inhibition is activated. A question that should be raised is whether the release is due to the muscle contractions per se or to a direct effect of the DNS and EFS. A contraction-depen60 dent release from the mucosa could be explained by a displacement of the mucosa over the tracheal muscle. This explanation is less likely as isometric contrac"DNS only" "EFSjDNS" tions were studied in this experimental 40 [35NS~1 "EFSjDNS" I i i i setup. Another piece of evidence against 60 min 120min Baseline 120min 60 min Baseline this hypothesis is the finding of the more Fig. 3. Contractile response to 2-Hz DNS in the distal and proximal segments (mean ± SEM, n = 5-6) as a pronounced decay when the stimulations percentage of baseline response in preparations with intact mucosa (M+;e), removed mucosa (M-; J), and were evoked by EFS/DNS compared with indomethacin-treated preparations with intact mucosa (M +lIndo; .). DNS alone although the baseline contractions were of a similar amplitude. ter the addition of EFS the mucosa-de- tissue, whereas the inhibition in stimu- Thus, the difference in the magnitude of pendent decay was pronounced in all lated tissues was considerably higher. the inhibitory effect between EFS/DNS segments. Thus in our preparation EFS and DNS and DNS alone cannot be explained by Thus, the mucosa-dependent inhibi- not only cause cholinergic contractions, . a contraction-dependent mechanism. tion of cholinergic contractions activat- but also activate a mucosa-dependent in- The release of inhibitory or relaxant ed by DNS alone was more pronounced hibition of the responses. mucosa-derived factors appears to be in the proximal part of the trachea comThe nature of what has been called neurally mediated. pared with the distal part. the epithelium-derived relaxant factor As the release of the mucosa-depen(EpDRF) has been discussed, and con- dent cyclooxygenase product is activatflicting data have been presented. Sever- ed by EFS but not by DNS, it is possible Discussion al studies have shown that this effect is to speculate that nonneural mechanisms Wehave previously demonstrated that re- due to cyclooxygenase products (4, 12, regulate this release. EFS with the parampeated DNS and EFS activate the release 22-25), whereas in other studies no or eters used produces a specific activation of a mucosa-derived inhibitory factor(s), only partial blockage was seen with the of postganglionic fibers, and the contracinhibiting the contractile responses. This inhibition of prostanoid synthesis (1, 2). tile responses can be totally blocked by factor(s) has been possible to transfer be- It is obvious from the data in this study atropine. However, an additional unspetween organ baths (12). The present study that there are at least two different mu- cific effect of the electric field on other demonstrates that a mucosa-dependent cosa-dependent inhibitory mechanisms cells, such as epithelial cells, cannot be inhibition of cholinergic contractions can in ferret trachea, one that can be blocked totally excluded. This study further shows that there is be activated by DNS alone. However, this by indomethacin and one that is not afinhibition evoked by DNS was not affect- fected by indomethacin. In an earlier a segmental variation in the ferret traed by indomethacin, whereas the more study we showed that the cyclooxygenase- chea when it comes to the effect of mupronounced inhibition by EFS was to a dependent component is released in the cosa-dependent inhibitory factors, since major part blocked by indomethacin. The organ bath and could be transferred by the inhibition by DNS alone was signifiinhibition evoked by DNS alone was sig- the fluid to a recipient preparation and cantly more pronounced in the proximal nificantly more pronounced in the proxi- denuded mucosa (12). In these experi- part of the trachea. One could argue that mal part of the trachea, whereas the EFS- ments we found that the addition of EFS the differences between the proximal and evoked inhibition showed a more homo- to the stimulation sequence caused a the distal parts of the trachea were exgeneous distribution along the trachea. marked inhibition that was significantly plained by different baseline contracMost studies of the interactions be- counteracted by indomethacin. Thus we tions. However, the stronger contraction tween the airway epithelium and smooth. have an indomethacin-sensitive release of in the proximal part of the trachea was more effectively inhibited than the weakmuscle responsiveness have been based a relaxant factor in this preparation. However, stimulation by DNS alone er contraction in the distal part, and the primarily on basal levels and/or a spontaneous release of relaxant factors (1-4, induces an inhibition of contractions in small decay in conjunction with DNS in 9, 10, 18, 19) or, in some studies, on the the presence of intact mucosa that is not the mucosa-denuded preparations did presence of enzymatic degradation sys- blocked by indomethacin. The magni- not show any regional variation. After tems in the mucosa (20,21). We have stud- tude of the two mucosa-dependent in- EFS the percentage inhibition on the ied the relaxant or inhibitory mechanisms hibitory factors seen in the ferret trachea proximal and distal parts was also simiactivated by repeated electrical stimula- has not been compared exclusively.How- lar, which further supports the idea that tions. This study reveals the difference ever, if the present data are compared the difference in baseline contractions between spontaneous and activated re- with our earlier studies, it is evident that does not explain the difference between lease as there was a mean decrease in con- the effect of the nonprostanoid compo- the regional distribution of the DNS eftractions of only 3070 in "unstimulated" nent activated by DNS alone is of a low- fect. The study was performed with the 100
NERVE STIMULATION RELEASES MUCOSA-DERIVED INHIBITORY FACTORS IN ISOLATED FERRET TRACHEA
whole preparation in one bath. This reduces the opportunity to evaluate the segmental distribution of a factor that is released into the fluid in a high concentration. We have previously demonstrated that it was possible to transfer the prostanoid factor between organ baths (12). This indicates that the prostanoid factor is released into the bath fluid in a fairly high concentration, which makes it unlikely that regional differences in the release could be evaluated for this factor. The difference between the proximal and distal segments in the release of the nonprostanoid factor must therefore be explained either by a high tissue affinity or by a short half-life of the factor in the fluid. Our data indicate that neural stimulation causes the release of inhibitory or relaxant factors. This study was not designed to evaluate the site of action for these factors. In an earlier study wefound no effect on the parasympathetic ganglionic transmission by the prostanoid factor (12). This finding may not necessarily be true for the nonprostanoid factor. This gives several possible mechanisms of action for these factors, such as a direct effect on the smooth muscle or effects on neural transmission. The latter mechanism could involve inhibition of cholinergic transmission as well as a facilitation of inhibitory nerves [e.g., nonadrenergic-noncholinergic (NANC)-inhibitory nerves]. The difference in magnitude between the distal and proximal segments of the trachea was pronounced for the nonprostanoid component. The mechanism behind this phenomenon is still unclear, but severalpossible explanations must be considered. If the effect is mediated by a receptor activation, a difference in the distribution of the receptors along the trachea is possible. The inhibitory effect is activated by DNS, and another possibility to consider is that the distribution of certain nerve fibers explains the regional distribution of the inhibition. One may also speculate that the inhibitory effect is mediated by nerve fibers of both a vagal and a spinal origin, as has been suggested for peptidergic nerves in guinea pig airways (26). In the nerve-muscle preparation of ferret trachea there is a marked portion of adrenergic nerves that can be activated by EFS, but only a minor adrenergic effect is achieved by direct stimulation of the vagal nerve (unpublished observation). EFS also activates inhibitory nonadrenergic-noncholinergic nerves, whereas pNS causes only a mi-
nor activation of inhibitory NANC fibers (27). With the limited knowledge of the relativeimportance of these neural mechanisms, we can only speculate on the mechanisms of interaction between airwaynerves and mucosa. Our results demonstrate, however, that the presence of an intact mucosa can regulate the contractile response to direct vagal nerve stimulation. Interactions between the epithelium and peptidergic nerves have previously been suggested in guinea pig (7). It has also previously been reported that the presence of intact epithelium lowered the response to EFS in dog bronchi (2), although other observations suggested no effect of mucosa removal on the frequency-response curve to EFS in bovine trachea (1). In conclusion, the present study demonstrates that neural stimulation can cause the release of mucosa-dependent relaxant or inhibitory factors that counteract the contractile responses. Both prostanoid and nonprostanoid factors are involved. The release of these factors can be an important feedback mechanism for regulation of airway tone, which may be disrupted in epithelial damage, such as that seen in asthma. Acknowledgment The writers wish to thank Lena Bernsten for excellent technical assistance. References 1. Barnes PJ, Cuss FM, Palmer JB. The effect of airway epithelium on smooth muscle contractility in bovine trachea. Br J Pharmacol1985; 86:685-91. 2. Flavahan NA, Aarhus LL, Rimele TJ, Vanhoutte PM. Respiratory epithelium inhibits bronchial smooth muscle tone. J Appl Physiol 1985; 58:834-8. 3. Aizawa H, Miyazaki N, Shigematsu N, Tomooka MA. Possible role of airway epithelium in modulating hyperresponsiveness. Br J Pharmacol1988; 93:139-45. 4. Butler GB, Adler KB, Evans IN, Morgan OW, Szarek JL. Modulation of rabbit airway smooth muscle responsiveness by respiratory epithelium. Am Rev Respir Dis 1987; 135:1099-104. 5. Farmer SO, Fedan JS, Hay DWP, Raeburn D. The effects of epithelium removal on the sensitivity of guinea-pig isolated trachealis to bronchodilator drugs. Br J Pharmacol 1986; 89:407-14. 6. Finnen MJ, Flower RJ, Lashenko A, Williams KI. Airway epithelium influences responsiveness of guinea pig tracheal strips. Br J Pharmacol1986; 88:407. 7. Frossard N, Muller F. Epithelial modulation of tracheal smooth muscle response to antigenic stimulation. J Appl Physiol 1986; 61:1449-56. 8. Montano LM, Selman M, Ponce Monter H, Vargas MH. Role of airway epithelium on the reactivity of smooth muscle from guinea pigs sensitized to ovalbumin by inhalatory method. Res Exp Med (Bed) 1988; 188:167-73. 9. Raeburn D, Hay DWP, Farmer SO, Fedan JS. Epithelium removal increases the reactivity of hu-
751 man isolated tracheal muscle to methacholine and reduces the effect of verapamil. Eur J Pharmacol 1986; 123:451-3. 10. Raeburn D, Hay DW, Robinson VA, Farmer SG, FlemingWW, Fedan JS. The effect of verapamil is reduced in isolated airway smooth muscle preparations lacking epithelium. Life Sci 1986; 38:809-16. 11. Stuart-Smith K, Vanhoutte PM. Airway epithelium modulates the responsiveness of porcine bronchial smooth muscle. J Appl Physiol 1988; 65:721-7. 12. Ullman A, Lofdahl CO, Svedmyr N, Bernsten L, Skoogh BE. Mucosal inhibition of cholinergic contractions in ferret trachea can be transferred between organ baths. Eur Respir J 1988; 1:908-12. 13. Stuart-Smith K, Vanhoutte PM. Heterogeneity in the effects of epithelium removal in the canine bronchial tree. J Appl Physiol1987; 63:2510-5. 14. Hay OW, Raeburn D, Fedan JS. Regional differences in reactivity and in the influence of the epithelium on canine intrapulmonary bronchial smooth muscle responsiveness. Eur J Pharmacol 1987; 141:363-70. 15. Skoogh BE. Parasympathetic ganglia in the airways. Bull Eur Physiopathol Respir 1986; 22: 143-7. 16. Skoogh BE. Transmissionthrough airwayganglia. Eur J Respir Dis [Suppl] 1983; 131:159-70. 17. Skoogh BE, Holtzman MJ, Sheller JR, Nadel JA. Barbiturates depress vagal motor pathway to ferret trachea at ganglia. J Appl Physiol1982; 53: 253-7. 18. Hay DW, Farmer SO, Raeburn D, Robinson VA, Fleming WW, Fedan JS. Airway epithelium modulates the reactivity of guinea-pig respiratory smooth muscle. Eur J Pharmacol 1986; 129:11-8. 19. Goldie RO, Papadimitriou JM, Paterson JW, Rigby PJ, Self HM, Spina D. Influence of the epithelium on responsiveness of guinea-pig isolated trachea to contractile and relaxant agonists. Br J Pharmacol 1986; 5:5-14. 20. Raeburn D, Sequeira DJ, Backes WL. Possible involvement of the CYtochrome P-450 in the epithelium-modulated response to methacholine in guinea pig trachea. Biochem Pharmacol1987; 37: 573-6. 21. Sekizawa K, Tamaoki J, Oraf PD, Basbaum CB, Borson OB, Nadel JA. Enkephalinase inhibitor potentiates mammalian tachykinin-inducedcontraction in ferret trachea. J Pharmacol Exp Ther 1987; 243:1211-7. 22. Brunellieschi S, Haye-Legfrand I, Labat C, Norel X, Benveniste J, Brink C. Platelet-activating factor-acether-induced relaxation of guinea pig airway muscle: role of prostaglandin E2 and the epithelium. J Pharmacol Exp Ther 1987;243:356-63. 23. Farmer SO, Hay DW, Raeburn 0, Fedan JS. Relaxation of guinea-pig tracheal smooth muscle to arachidonate is converted to contraction following epithelium removal. Br J Pharmacol 1987; 92:231-6. 24. Nijkamp FP, Folkerts O. Reversal of arachidonic acid-induced guinea-pig tracheal relaxation into contraction after epithelium removal. Eur J Pharmacol 1987; 131:315-6 25. Tschirhart E, Frossard N, Bertrand C, Landry Y. Arachidonic acid metabolites and airway epithelium-dependent relaxant factor. J Pharmacol Exp Ther 1987; 243:310-6. 26. Saria A, Martling CR, Dalsgaard CJ, Lundberg JM. Evidence for substance P-immunoreactive spinal afferents that mediate bronchoconstriction. Acta Physiol Scand 1985; 125:407-14. 27. Skoogh BE, Lofdahl CO, Lotwall J, Svedmyr N. Non-adrenergic inhibitory nerves to ferret trachea do not go with cholinergic nerves (abstract). Am Rev Respir Dis 1986; 133(Part 2:A116).
ANDERS ULLMAN, CLAES-GORAN LCFDAHL, NILS SVEDMYR, and BENGT-ERIC SKOOGH
Introduction Since the initial observation of modulatory mechanisms in the airway mucosa and epithelium regulating bronchomotor tone and bronchial responsiveness (1, 2), much interest has focused on this phenomenon. The release of modulatory factors derived from the epithelium and/or other elements of the mucosal tissue has been demonstrated in vitro in a number of species, including the human (3-12). In many of these reports only the basal release of these factors has been studied, although an activation of the release by p-adrenoreceptor agonists, for example, has been suggested (13). Greater knowledge of the mechanisms regulating the release of these factors is important when it comes to understanding the physiologic and possible pathophysiologic role of these phenomena. We have previously demonstrated the release of a mucosa-derived factor(s), inhibiting as much as some 50010 of the contractile responses to supramaximal cholinergic nerve stimulation in isolated ferret trachea (12). The release was activated by phasic stimulations with alternating electrical field stimulation (EFS) and direct vagal nerve stimulation (DNS), indicating an interaction between the neuralsystem and the airway mucosa (12). One factor mediating this effect is obviously a prostanoid as it is significantly reduced, but not abolished, by indomethacin·(12). If there is a neural activation of this mucosa-dependent inhibition, there may be a difference in the magnitude of the inhibition if the stimulation is given only by DNS of the vagal nerve compared with the more general and postganglionic nerve stimulation by EFS. Another possibility that should be considered is that the regional differences in the neural anatomy of the airways influence the inhibition, which could be one explanation of the previously reported observations of dif748
SUMMARY The Influence of an Intact mucosa layer In Isolated ferret trachea on the contractile responses to repeated, short-lasting (20 s, every 2 min) nerve stimulations was studied In a nervemuscle preparation stimulated to chollnerglc-evoked contractions by either direct vagal nerve stimulation (DNS)or transmural electrical field stimulation (EFS). The contractile responses were monitored by three strain gauges connected to the proximal, middle, and distal segments of the trachea. The mucosa was either left Intact (M +)or removed from the membranous part by dissection (M-). The successive decrease In contractile responses was studied for 60 min using repeated DNS and for a further 60 min using alternating EFS and DNS. During the first period, In which DNS alone was used, there was a significantly more pronounced decrease In the M+preparation compared with the M-. This effect was most prominent In the proximal part of the trachea and was not blocked by Indomethacin. During the subsequent stimulation period In which alternating EFS and DNS were used, the rate of decrease was significantly greater In all segments. In this phase, however, the Inhibitory mucosa-dependent effect was significantly attenuated by Indomethacin treatment. In conclusion, this stUdy demonstrates that the mucosa-dependent Inhibition or relaxation In ferret trachea Is mediated by both a prostanold and a nonprostanold factor. The nonprostanold factor but not the prostanold factor can be activated by DNS,whereas the release of the prostanold Is activated by EFS, which Indicates that the activation of the prostanold and nonprostanold factor Is controlled by different mechanisms. AM pI=''' RESPIR DIS 1990; 141:748-751
ferent magnitudes of the mucosa-dependent inhibition at different levels in the airways (11, 13, 14). The aim of the present study was to characterize the mechanisms that regulate the release of the mucosa-derived inhibitory factor in ferret trachea. Wetherefore tried to determine whether the release of an inhibitory factor, prostanoid and/or nonprostanoid, could be activated by DNS alone. If this were so, we also set out to examine whether the inhibition varies between the proximal and distal parts of the trachea. Methods The study was performed in vitro using a nerve-muscle preparation of ferret trachea (15-17). Male ferrets were rendered unconscious by electric shocks and were killed by exsanguination. The study was approved by The Gothenburg University Ethics Committee for Animal Experiments. The trachea, with intact right recurrent and vagus nerves, was rapidly removed and immersed in an organ bath filled with 200 ml Krebs Ringer solution (KR) of the following composition (in mM): NaCI 118, KCI 5.9, CaCI 2.5, MgS0 4 1.2, NaH 2P04 1.2, NaHC03 25.5, and glu-
cose 5.6, maintained at 38° C and gassed with 94070 O 2 and 6070 CO 2 • Ascorbic acid was added to the KR in a concentration of 0.3 mM as an antioxidant. The bath was continuously flushed with prewarmed KR, 10 ml/min. In the organ bath the trachea was split along the anterior long axis and opened. One side was attached by pins, and the opposite side was connected to three strain gauges (Grass FT 03) to record isometric muscle tension at optimum resting tension (7 to 10g). The contractile responses wererecorded from the three strain gauges, one distal, one middle, and one proximal. Comparisons were made for the three segments separately. In some preparations the mucosa covering the tracheal muscle was removed by cutting
(Received in originalform February 7, 1989 and in revised form August 9, 1989) 1 From the Departments of Clinical Pharmacology and Pulmonary Medicine, Gothenburg University, Sweden. 2 Supported by grants from the Swedish Association Against Heart and Chest Diseases. 3 Correspondence and requests for reprints should be addressed to Dr. A. Ullman, Department of Clinical Pharmacology, Sahlgrenska Hospital, S-41345 Gothenburg, Sweden.
749
NERVE STIMULATION RELEASES MUCOSA-DERIVED INHIBITORY FACTORS IN ISOLATED FERRET TRACHEA
along the borders to the cartilaginous part on both sides of the membranous part, after which the mucosa could easily be peeled off. Care was taken not to damage the mucosa in other parts of the trachea. Phasic contractions were evoked by nerve stimulation for 20 severy 2 min. This was done either by direct nerve stimulation alone, or by alternating electrical field stimulation and DNS. EFS was given via two platinum electrodes, one of which was placed above and one below the tracheal muscle. DNS was applied to the vagal nerve and to the laryngeal end of the recurrent nerve simultaneously via two suction electrodes.The electrical impulses were biphasic square waves with a duration of 0.25 ms (DNS) or 0.5 ms (EFS). The frequencies used were 2 and 12Hz for DNS and 12 Hz for EFS. These DNS frequencies induce contractile responses of about 50 and 90010, respectively, of the maximum response to DNS, which is achieved at 24 Hz. The current for DNS was supramaximal (22 rnA). The current for EFS was submaximal (600 to 1000 rnA) and was adjusted so that the responses matched the initial responses to 12 Hz DNS. For statistical evaluation the response to 2 Hz DNS was used as primary effect parameter. In each experiment two stimulation sequences wereused. Initially stimulations were given by "DNS only" (figure 1) to establish the baseline response during 30 min and for 60 min further. Thereafter, the stimulation sequence was changed to also include EFS (EFS/DNS, figure 1). Experiments werecarried out both on preparations with intact mucosa (M+) and on preparations with the mucosa removed over the tracheal muscle (M-). M+ experiments were performed in the presence and absence of indomethacin (10 J,1M). Indomethacin per se does not affect the baseline response or the degree of decay in mucosa-denuded preparations in this experimental model (12). A control series was performed to evaluate the degree of spontaneous decay of contractile responses. In this series the mucosa was left intact. Every 16 min one 12-Hz and one 2-Hz DNS stimulation was given. After an equilibration period of 60 min the electricalstimulations (DNS alone) werestart-
Period 1
Period 2
''DNS only"
"EFS/DNS"
=ig. 1. The stimulation sequences during the first and second period of each experiment. EFS, electrical field stlmulation; DNS, direct nerve stimulation of the vagal
ierve,
TABLE 1 BASELINE CONTRACTILE RESPONSES TO 2-Hz DIRECT NERVE STIMULATION Mean (g)
SEM
M+ MM+/lndo
28 27 25
3.2 3.5 1.8
Proximal
M+ MM+/lndo
32 32 31
1.8 2.3 2.7
Middle
M+ MM+llndo
30 27 24
5.5 4.3 1.1
Distal
M+ MM+llndo
21 19 21
3.3 3.5 2.7
Nerve All segments n =6 n =5 n =6
Definition of abbreviations: intact mucosa = M+; removed mucosa = M-; Indo = indomethacin, 10 JIM.
ed. The contractile responses 30 min after the start of stimulation wereused as baseline. The successive decrease in contractile force was then observed for 60 min after baseline measurements, after which stimulations weregiven in the second sequence (alternating DNS and EFS) for a further 60-mln period. The results are expressed as the mean ± SEM. For unpaired comparisons a two-way analysis of variance was used, followed by Fisher's least-significant difference test. For paired comparisons Student's t test for paired samples was used; p < 0.05 was considered significant.
Results
icantly enhanced the rate of decrease in the contractile responses. The decrease in indomethacin-treated preparations with intact mucosa was similar to that in mucosa-denuded preparations but significantly (p < 0.01) less pronounced than in untreated preparations with intact mucosa (figure 2). Thus DNS activates a mucosa-dependent inhibitory factor that is not blocked by indomethacin, whereas EFS releases an additional mucosa-dependent inhibitory factor that can be blocked by indomethacin. To elucidate possible segmental differences in the distribution of the inhibitory effects we also evaluated the contractile responses separately for the three segments. The proximal segment showed a larger baseline contractile response compared with the distal segment in both preparations with intact and with removed mucosa. Despite this finding, the mucosa-dependent inhibition of the contractile responses activated by DNS alone was more pronounced in the proximal segment compared to the distal in both indomethacin-treated and untreated preparations (p < 0.05). The percentage decrease in mucosa-denuded preparations was similar in the proximal and distal segments (figure 3). In the middle segment the magnitude of the mean decrease was between that of the proximal and distal segments, but with a considerably greater variation. Af-
The analysis that follows is based on the response to 2 Hz DNS as the primary ef100 fect parameter. The baseline contractions in preparations with and without mucosa as well as with intact mucosa after indomethacin treatment weresimilar (table 1). In control preparations with intact ~ mucosa only stimulated by DNS every := 80 16 min, the successive decrease during a ~ period of 60 min was only 3 ± 2070, there- ~ by indicating low spontaneous decay. f+-I During the first stimulation period 0 with DNS-only (2 and 12Hz every2 min), ~ ~' 60 the successive decrease in the contractile responses was significantly (p < 0.01) more pronounced in preparations with mucosa (19 ± 30/0) compared with mucosa-denuded preparations (8 ± 2070). "DNS only" " "EFSjDNS" 40 This difference was not significantly afBaseline 60 min 120min fected by indomethacin (figure 2). In mucosa-denuded preparations, the Fig. 2. Contractile response to 2-Hz DNS (mean ± addition of EFS (12 Hz) had little effect SEM, n = 5-6) as a percentage of baseline response on the decrease rate of the contractions, in preparations with intact mucosa (M+;e), removed mucosa (M-;O), and indomethacin-treated preparations compared with the rate when stimulated with intact mucosa (M +lIndo; .A). Stimulations were givby DNS alone. In preparations with in- en with DNS only during the first period and by EFSlDNS tact mucosa, the addition of EFS signif- alternating during the second period.
ULLMAN, LOFDAHL, SVEDMYR, AND SKOOGH
750
er magnitude than the inhibition by the combination of DNS and EFS (12). This study was performed to characterize the mechanisms by which the mucosa-dependent inhibition is activated. A question that should be raised is whether the release is due to the muscle contractions per se or to a direct effect of the DNS and EFS. A contraction-depen60 dent release from the mucosa could be explained by a displacement of the mucosa over the tracheal muscle. This explanation is less likely as isometric contrac"DNS only" "EFSjDNS" tions were studied in this experimental 40 [35NS~1 "EFSjDNS" I i i i setup. Another piece of evidence against 60 min 120min Baseline 120min 60 min Baseline this hypothesis is the finding of the more Fig. 3. Contractile response to 2-Hz DNS in the distal and proximal segments (mean ± SEM, n = 5-6) as a pronounced decay when the stimulations percentage of baseline response in preparations with intact mucosa (M+;e), removed mucosa (M-; J), and were evoked by EFS/DNS compared with indomethacin-treated preparations with intact mucosa (M +lIndo; .). DNS alone although the baseline contractions were of a similar amplitude. ter the addition of EFS the mucosa-de- tissue, whereas the inhibition in stimu- Thus, the difference in the magnitude of pendent decay was pronounced in all lated tissues was considerably higher. the inhibitory effect between EFS/DNS segments. Thus in our preparation EFS and DNS and DNS alone cannot be explained by Thus, the mucosa-dependent inhibi- not only cause cholinergic contractions, . a contraction-dependent mechanism. tion of cholinergic contractions activat- but also activate a mucosa-dependent in- The release of inhibitory or relaxant ed by DNS alone was more pronounced hibition of the responses. mucosa-derived factors appears to be in the proximal part of the trachea comThe nature of what has been called neurally mediated. pared with the distal part. the epithelium-derived relaxant factor As the release of the mucosa-depen(EpDRF) has been discussed, and con- dent cyclooxygenase product is activatflicting data have been presented. Sever- ed by EFS but not by DNS, it is possible Discussion al studies have shown that this effect is to speculate that nonneural mechanisms Wehave previously demonstrated that re- due to cyclooxygenase products (4, 12, regulate this release. EFS with the parampeated DNS and EFS activate the release 22-25), whereas in other studies no or eters used produces a specific activation of a mucosa-derived inhibitory factor(s), only partial blockage was seen with the of postganglionic fibers, and the contracinhibiting the contractile responses. This inhibition of prostanoid synthesis (1, 2). tile responses can be totally blocked by factor(s) has been possible to transfer be- It is obvious from the data in this study atropine. However, an additional unspetween organ baths (12). The present study that there are at least two different mu- cific effect of the electric field on other demonstrates that a mucosa-dependent cosa-dependent inhibitory mechanisms cells, such as epithelial cells, cannot be inhibition of cholinergic contractions can in ferret trachea, one that can be blocked totally excluded. This study further shows that there is be activated by DNS alone. However, this by indomethacin and one that is not afinhibition evoked by DNS was not affect- fected by indomethacin. In an earlier a segmental variation in the ferret traed by indomethacin, whereas the more study we showed that the cyclooxygenase- chea when it comes to the effect of mupronounced inhibition by EFS was to a dependent component is released in the cosa-dependent inhibitory factors, since major part blocked by indomethacin. The organ bath and could be transferred by the inhibition by DNS alone was signifiinhibition evoked by DNS alone was sig- the fluid to a recipient preparation and cantly more pronounced in the proximal nificantly more pronounced in the proxi- denuded mucosa (12). In these experi- part of the trachea. One could argue that mal part of the trachea, whereas the EFS- ments we found that the addition of EFS the differences between the proximal and evoked inhibition showed a more homo- to the stimulation sequence caused a the distal parts of the trachea were exgeneous distribution along the trachea. marked inhibition that was significantly plained by different baseline contracMost studies of the interactions be- counteracted by indomethacin. Thus we tions. However, the stronger contraction tween the airway epithelium and smooth. have an indomethacin-sensitive release of in the proximal part of the trachea was more effectively inhibited than the weakmuscle responsiveness have been based a relaxant factor in this preparation. However, stimulation by DNS alone er contraction in the distal part, and the primarily on basal levels and/or a spontaneous release of relaxant factors (1-4, induces an inhibition of contractions in small decay in conjunction with DNS in 9, 10, 18, 19) or, in some studies, on the the presence of intact mucosa that is not the mucosa-denuded preparations did presence of enzymatic degradation sys- blocked by indomethacin. The magni- not show any regional variation. After tems in the mucosa (20,21). We have stud- tude of the two mucosa-dependent in- EFS the percentage inhibition on the ied the relaxant or inhibitory mechanisms hibitory factors seen in the ferret trachea proximal and distal parts was also simiactivated by repeated electrical stimula- has not been compared exclusively.How- lar, which further supports the idea that tions. This study reveals the difference ever, if the present data are compared the difference in baseline contractions between spontaneous and activated re- with our earlier studies, it is evident that does not explain the difference between lease as there was a mean decrease in con- the effect of the nonprostanoid compo- the regional distribution of the DNS eftractions of only 3070 in "unstimulated" nent activated by DNS alone is of a low- fect. The study was performed with the 100
NERVE STIMULATION RELEASES MUCOSA-DERIVED INHIBITORY FACTORS IN ISOLATED FERRET TRACHEA
whole preparation in one bath. This reduces the opportunity to evaluate the segmental distribution of a factor that is released into the fluid in a high concentration. We have previously demonstrated that it was possible to transfer the prostanoid factor between organ baths (12). This indicates that the prostanoid factor is released into the bath fluid in a fairly high concentration, which makes it unlikely that regional differences in the release could be evaluated for this factor. The difference between the proximal and distal segments in the release of the nonprostanoid factor must therefore be explained either by a high tissue affinity or by a short half-life of the factor in the fluid. Our data indicate that neural stimulation causes the release of inhibitory or relaxant factors. This study was not designed to evaluate the site of action for these factors. In an earlier study wefound no effect on the parasympathetic ganglionic transmission by the prostanoid factor (12). This finding may not necessarily be true for the nonprostanoid factor. This gives several possible mechanisms of action for these factors, such as a direct effect on the smooth muscle or effects on neural transmission. The latter mechanism could involve inhibition of cholinergic transmission as well as a facilitation of inhibitory nerves [e.g., nonadrenergic-noncholinergic (NANC)-inhibitory nerves]. The difference in magnitude between the distal and proximal segments of the trachea was pronounced for the nonprostanoid component. The mechanism behind this phenomenon is still unclear, but severalpossible explanations must be considered. If the effect is mediated by a receptor activation, a difference in the distribution of the receptors along the trachea is possible. The inhibitory effect is activated by DNS, and another possibility to consider is that the distribution of certain nerve fibers explains the regional distribution of the inhibition. One may also speculate that the inhibitory effect is mediated by nerve fibers of both a vagal and a spinal origin, as has been suggested for peptidergic nerves in guinea pig airways (26). In the nerve-muscle preparation of ferret trachea there is a marked portion of adrenergic nerves that can be activated by EFS, but only a minor adrenergic effect is achieved by direct stimulation of the vagal nerve (unpublished observation). EFS also activates inhibitory nonadrenergic-noncholinergic nerves, whereas pNS causes only a mi-
nor activation of inhibitory NANC fibers (27). With the limited knowledge of the relativeimportance of these neural mechanisms, we can only speculate on the mechanisms of interaction between airwaynerves and mucosa. Our results demonstrate, however, that the presence of an intact mucosa can regulate the contractile response to direct vagal nerve stimulation. Interactions between the epithelium and peptidergic nerves have previously been suggested in guinea pig (7). It has also previously been reported that the presence of intact epithelium lowered the response to EFS in dog bronchi (2), although other observations suggested no effect of mucosa removal on the frequency-response curve to EFS in bovine trachea (1). In conclusion, the present study demonstrates that neural stimulation can cause the release of mucosa-dependent relaxant or inhibitory factors that counteract the contractile responses. Both prostanoid and nonprostanoid factors are involved. The release of these factors can be an important feedback mechanism for regulation of airway tone, which may be disrupted in epithelial damage, such as that seen in asthma. Acknowledgment The writers wish to thank Lena Bernsten for excellent technical assistance. References 1. Barnes PJ, Cuss FM, Palmer JB. The effect of airway epithelium on smooth muscle contractility in bovine trachea. Br J Pharmacol1985; 86:685-91. 2. Flavahan NA, Aarhus LL, Rimele TJ, Vanhoutte PM. Respiratory epithelium inhibits bronchial smooth muscle tone. J Appl Physiol 1985; 58:834-8. 3. Aizawa H, Miyazaki N, Shigematsu N, Tomooka MA. Possible role of airway epithelium in modulating hyperresponsiveness. Br J Pharmacol1988; 93:139-45. 4. Butler GB, Adler KB, Evans IN, Morgan OW, Szarek JL. Modulation of rabbit airway smooth muscle responsiveness by respiratory epithelium. Am Rev Respir Dis 1987; 135:1099-104. 5. Farmer SO, Fedan JS, Hay DWP, Raeburn D. The effects of epithelium removal on the sensitivity of guinea-pig isolated trachealis to bronchodilator drugs. Br J Pharmacol 1986; 89:407-14. 6. Finnen MJ, Flower RJ, Lashenko A, Williams KI. Airway epithelium influences responsiveness of guinea pig tracheal strips. Br J Pharmacol1986; 88:407. 7. Frossard N, Muller F. Epithelial modulation of tracheal smooth muscle response to antigenic stimulation. J Appl Physiol 1986; 61:1449-56. 8. Montano LM, Selman M, Ponce Monter H, Vargas MH. Role of airway epithelium on the reactivity of smooth muscle from guinea pigs sensitized to ovalbumin by inhalatory method. Res Exp Med (Bed) 1988; 188:167-73. 9. Raeburn D, Hay DWP, Farmer SO, Fedan JS. Epithelium removal increases the reactivity of hu-
751 man isolated tracheal muscle to methacholine and reduces the effect of verapamil. Eur J Pharmacol 1986; 123:451-3. 10. Raeburn D, Hay DW, Robinson VA, Farmer SG, FlemingWW, Fedan JS. The effect of verapamil is reduced in isolated airway smooth muscle preparations lacking epithelium. Life Sci 1986; 38:809-16. 11. Stuart-Smith K, Vanhoutte PM. Airway epithelium modulates the responsiveness of porcine bronchial smooth muscle. J Appl Physiol 1988; 65:721-7. 12. Ullman A, Lofdahl CO, Svedmyr N, Bernsten L, Skoogh BE. Mucosal inhibition of cholinergic contractions in ferret trachea can be transferred between organ baths. Eur Respir J 1988; 1:908-12. 13. Stuart-Smith K, Vanhoutte PM. Heterogeneity in the effects of epithelium removal in the canine bronchial tree. J Appl Physiol1987; 63:2510-5. 14. Hay OW, Raeburn D, Fedan JS. Regional differences in reactivity and in the influence of the epithelium on canine intrapulmonary bronchial smooth muscle responsiveness. Eur J Pharmacol 1987; 141:363-70. 15. Skoogh BE. Parasympathetic ganglia in the airways. Bull Eur Physiopathol Respir 1986; 22: 143-7. 16. Skoogh BE. Transmissionthrough airwayganglia. Eur J Respir Dis [Suppl] 1983; 131:159-70. 17. Skoogh BE, Holtzman MJ, Sheller JR, Nadel JA. Barbiturates depress vagal motor pathway to ferret trachea at ganglia. J Appl Physiol1982; 53: 253-7. 18. Hay DW, Farmer SO, Raeburn D, Robinson VA, Fleming WW, Fedan JS. Airway epithelium modulates the reactivity of guinea-pig respiratory smooth muscle. Eur J Pharmacol 1986; 129:11-8. 19. Goldie RO, Papadimitriou JM, Paterson JW, Rigby PJ, Self HM, Spina D. Influence of the epithelium on responsiveness of guinea-pig isolated trachea to contractile and relaxant agonists. Br J Pharmacol 1986; 5:5-14. 20. Raeburn D, Sequeira DJ, Backes WL. Possible involvement of the CYtochrome P-450 in the epithelium-modulated response to methacholine in guinea pig trachea. Biochem Pharmacol1987; 37: 573-6. 21. Sekizawa K, Tamaoki J, Oraf PD, Basbaum CB, Borson OB, Nadel JA. Enkephalinase inhibitor potentiates mammalian tachykinin-inducedcontraction in ferret trachea. J Pharmacol Exp Ther 1987; 243:1211-7. 22. Brunellieschi S, Haye-Legfrand I, Labat C, Norel X, Benveniste J, Brink C. Platelet-activating factor-acether-induced relaxation of guinea pig airway muscle: role of prostaglandin E2 and the epithelium. J Pharmacol Exp Ther 1987;243:356-63. 23. Farmer SO, Hay DW, Raeburn 0, Fedan JS. Relaxation of guinea-pig tracheal smooth muscle to arachidonate is converted to contraction following epithelium removal. Br J Pharmacol 1987; 92:231-6. 24. Nijkamp FP, Folkerts O. Reversal of arachidonic acid-induced guinea-pig tracheal relaxation into contraction after epithelium removal. Eur J Pharmacol 1987; 131:315-6 25. Tschirhart E, Frossard N, Bertrand C, Landry Y. Arachidonic acid metabolites and airway epithelium-dependent relaxant factor. J Pharmacol Exp Ther 1987; 243:310-6. 26. Saria A, Martling CR, Dalsgaard CJ, Lundberg JM. Evidence for substance P-immunoreactive spinal afferents that mediate bronchoconstriction. Acta Physiol Scand 1985; 125:407-14. 27. Skoogh BE, Lofdahl CO, Lotwall J, Svedmyr N. Non-adrenergic inhibitory nerves to ferret trachea do not go with cholinergic nerves (abstract). Am Rev Respir Dis 1986; 133(Part 2:A116).
