An X-Ray Study of the Influence of Vasoactive Intestinal Polypeptide and Substance p on the Secretion of gas Into the Swimbladder of a Teleost Gadus Morhua

The teleost swimbladder keeps the fish in neutral bouyancy in the water. When a fish is descending in the water, the gas content of the swimbladder is compressed, and gas secretion has to be increased to compensate for the increased hydrostatic pressure from the outside. This is achieved by enhanced metabolic activity of the gas gland cells, giving an increased concentration of oxygen in the blood of the rete mirabile (see Fange, 1953, 1983; Steen, 1963). This oxygen can diffuse into the swimbladder once the oxygen tension in the blood exceeds that in the swimblad-der.

Previous studies have concluded that gas secretion is controlled mainly by cholinergic fibres running in the vagus nerve (Fange, 1953; Fange et al. 1976), but an additional, unidentified mechanism was suggested by Fange (1973). Resorption, in contrast, is controlled by adrenergic nerves that contract the secretory mucosa and relax the resorptive mucosa, thus increasing the resorbing area (Fange, 1953; Nilsson, 1971; Ross, 1978). In addition, adrenergic drugs constrict the small arterial vessels in the submucosa of the secretory part and dilate the vessels in the resorbent part, an effect also seen after stimulation of the ramus intestinalis of the vagus (Stray-Pedersen, 1970).

The teleost swimbladder mucosa is innervated by nerve fibres containing non-adrenergic, non-cholinergic substances (see Lundin and Holmgren, 1984,1989). In the cod, vasoactive intestinal polypeptide (VlP)-like and substance-P-like immu-noreactive fibres innervate the swimbladder wall, and VIP-like immunoreactive fibres are present in the gas gland, particularly close to the vessels of the rete mirabile. VIP-like immunoreactivity was also observed in the vagus nerve and in the vagus branch leading to the swimbladder, ‘the swimbladder nerve’ (Lundin and Holmgren, 1984).

In vitro studies of isolated strips from the cod swimbladder mucosa have shown that VIP relaxes and that substance P and 5-hydroxytryptamine (5-HT) constrict the smooth muscles (Lundin, 1991). The present study was performed to elucidate the effects of these putative neurotransmitters on gas secretion into the swimblad-der in vivo, using an X-ray method where changes in the cross-sectional area/volume of the swimbladder can be measured over time. The X-ray method is a convenient method that requires little handling of the animal. The Atlantic cod Gadus morhua was chosen as a model because of its relatively fast gas secretion rate, its availability throughout the year, its suitable size and because much of the current information about the mechanisms and control of swimbladder function have been obtained from studies in the cod.

Atlantic cod, Gadus morhua, 100-300g, caught in cages at a depth of 2-4 m off the Swedish west coast, were used. The fish were kept in recirculating, aerated sea water at 10°C for 1-3 weeks before use.

The fish were anaesthetized with MS 222 (tricaine methanesulphonate; 100 mg 1^-1, Sigma) until breathing movements just ceased, and were kept with aerated sea water containing MS 222 (50 mg 1^-1) flushing over the gills during the operation. A catheter for injection of drugs (PP25 or PE 50), filled with heparinized (50i.u.ml^-1) 0.9% NaCl was inserted into the efferent branchial artery of the third gill arch. The catheter was secured by a ligature around the gill arch and with a skin suture along the body wall. The end of the catheter was sealed by melting. In some fish, the intestinal branches of the vagus nerves were cut on both sides through small incisions made immediately posterior to the gills.

The fish were allowed to recover for approximately 20 h in a seawater tank. In this tank, with a water volume of 0.4 m³ (depth 0.6 m), the fish could move freely throughout the experiment.

Each fish was lightly anaesthetized, weighed, individually marked and X-rayed (=starting value). The fish were then given an injection (1.0 ml kg^-1) of the drug to be tested or NaCl (0.9%) through the gill catheter, and the swimbladder gas was removed through the body wall with a syringe. Some groups were also given atropine (1.2 mg kg^-1) at the start of the experiment. All fish were again X-rayed (=0 h). At 3 and 6 h the fish were again injected with the drugs and X-rayed. At 10, 24 and sometimes 48 h they were X-rayed only. X-rays were taken with a ZAR/PLH-1 veterinary X-ray unit.

The area of the swimbladder on the X-ray photographs (see Fig. 1) was measured with a PC computer program especially designed for this purpose. A sequence of points defining the outline of the swimbladder was entered into the program using a mouse. The area was calculated from these points using a standard area formula for polygons.

The starting value, i.e. before removal of the gas, was set at 100%, and the subsequent measurements (at 0, 3, 6, 10, 24 and 48h) were calculated as the percentage refilled. Mean curves from 5-9 fishes were plotted for each group (see Figs 3 and 4). Fishes that had a zero volume of more than 30% (i.e. that had not been properly emptied) were excluded.

To determine the relationship between the area of the swimbladder on the X-ray photographs and the volume of the swimbladder in the fish, 12 fishes were anaesthetized and X-rayed. A syringe was then inserted into the swimbladder through the body wall, 1 ml of gas was withdrawn, and the fish was X-rayed again. This was repeated until the swimbladder was empty. The total volume withdrawn was noted, and the volume was plotted as a function of the area (Fig. 2), showing that, for specimens of this size, there is a linear relationship between the volume and the area on the photographs. This means that the value obtained for the relative degree of refilling is independent of whether area or volume is measured. The changes in the laterally projected area of the swimbladder are thus considered to be a good representation of the changes in volume in this study.

The mean filling rate between 0 and 24 h (in mlkg^-1h^-1) was calculated for all fishes using the regression line from the area-volume plot (Fig. 2) to convert the measured increase in area to an increase in volume. Mean values and standard error of the mean (S.E.M.) were calculated for each group (see Table 1). These filling rates were used for the statistical evaluation.

The significance of observed effects of the drugs injected or vagotomy was analysed using the Mann-Whitney U -test (two-tailed). The level of significance was set at P≤0.05 divided by the number of comparisons made, and a sequentially rejective Bonferroni test was used (Holm, 1979).

The drugs used were 1.2 mg ml’¹ or 3.6 mg ml^-1 atropine sulphate (Sigma), 10⁻⁶moll^-1 carbamylcholine chloride (carbachol) (Sigma), 10⁻⁶moll^-1 5-hydroxytryptamine (serotonin) creatinine sulphate (Sigma), 10⁻⁵moll^-1 DL-noradrenaline/metanephrine HC1 (Sigma), 10⁻⁵moll^-1 substance P (Cambridge Research Biochemicals) and 10⁻⁵moll^_1 synthetic porcine vasoactive intestinal polypeptide (Cambridge Research Biochemicals).

The antisera used were the VIP antibody MI-VIP (MILAB), diluted 1:200, and the substance P antibody G1O (a gift from Dr A.-C. Jonsson), diluted 1:100.

All drugs were dissolved in 0.9 % NaCl. For the control groups 0.9 % NaCl was used. 0.1ml per 100g body mass was given at each injection.

The mean rate of refilling was calculated over the period 0-24 h. There was great variability in the secretion rate until 10 h, probably because of the frequent handling of the fish during this period.

After removal of gas, the swimbladders of fish that had not been injected or vagotomized (N =9) were refilled to 95 % after 24 h (Fig. 3) at a mean filling rate of 1.47±0.13mlkg^-1h^-1. Injections of VIP (N =6) significantly reduced the mean rate of filling to 0.88±0.20mlkg^-1h^-1 (Table 1 and Fig. 3). No significant effect on the secretion rate could be seen in control fishes injected with carbachol (N = 6), substance P (N =5), noradrenaline (N =5) or 5-HT (N =5) (see Table 1).

To block endogenous substance P and VIP, antisera against them were injected into a few control fishes, but no significant changes in the mean filling rate could be seen (Table 1).

Atropine (1.2mgkg^-1), caused a statistically significant inhibition of the gas filling rate (Table 1 and Fig. 3), giving a mean filling rate of 1.01±0.09ml kg^-1 h^-1. Injections of a higher dose of atropine (3.6 mg kg^-1) did not increase the degree of inhibition. Atropinized fish subsequently injected with VIP (N =6) or substance P (N=6) at 0, 3 and 6h increased their secretion rate back to the control level, with mean filling rates of 1.39±0.11 and 1.39±0.06mlkg^-1h^-1, respectively (Table 1 and Fig. 4). Injections of noradrenaline (N=5) or 5-HT (N=6) had no measurable effect on the secretion rate in atropinized fish (Table 1).

Fishes were vagotomized, ‘stimulated’ by withdrawal of gas and injected with 0.9% NaCl (N=7). The secretion was almost completely inhibited by the vagotomy (Table 1 and Fig. 3), with a mean filling rate of 0.18±0.04mlkg^-1h^-1. Injections of carbachol (N =4), VIP (N =3) or substance P (N =3) had no effect on the gas secretion rate in vagotomized fish when injected either separately or simultaneously (N =3, Table 1).

This study of secretory mechanisms in the teleost swimbladder uses a non-invasive method to measure the volume of the swimbladder in vivo during the experiment. The results confirm the findings of many previous studies on the swimbladder using other methods, where the involvement of the vagus nerve and of cholinergic and adrenergic mechanisms in the control of gas secretion and resorption have been demonstrated (Fahlén et al. 1965; Fange et al. 1976; Nilsson, 1971, 1972; Nilsson and Fange, 1967; Ross, 1978; Stray-Pedersen, 1970). It also presents new results concerning the non-adrenergic, non-cholinergic innervation of the teleost swimbladder.

The rate of filling of the swimbladder measured in this study (1.5 ml kg^-1 h^-1) corresponds well with calculated values for gas secretion in the cod [Enns et al. 1967 (2.4 ml kg^-1 h^-1); Tytler and Blaxter, 1973 (1.67-2.5 ml kg^-1 h^-1); Harden Jones and Scholes, 1985 (4.8 ml kg^-1 h^-1)], especially when the fact that the rate of filling is the net effect of secretion and resorption of gas in the bladder is taken into consideration. The rate of secretion under the experimental conditions of the present study may be nearly maximal for the gas gland of the cod, since it was not increased by carbachol, VIP and substance P.

It is likely that the resorption rate during the experiment was enhanced by circulating catecholamines. Although the fish were allowed to move freely in their tank to minimize stress and to let gas secretion proceed at a maximal rate throughout the experiment (see Copeland, 1952), the handling during the X-ray photography inevitably results in increased blood levels of catecholamines and other ‘stress’ hormones. In addition, it has been demonstrated that the anaesthetic MS 222 has several side effects, one being the release of catecholamines (Iwama et al. 1989). High levels of catecholamines in the fish during these experiments are indicated by the observation that noradrenaline did not decrease the filling rate, although the swimbladder is innervated by noradrenergic nerves, which normally stimulate resorption (Nilsson, 1971; Abrahamsson and Nilsson, 1976).

It has been proposed that MS 222 depresses neuronal activity for a short period (see Houston et al. 1971). However, the filling of the swimbladder seemed to continue at a steady rate until it was full (see Figs 3 and 4), suggesting that the impairment of neuronal function was of marginal importance during the experiments.

5-HT-like material has been demonstrated immunohistochemically in the swimbladder mucosa of several teleost species, both in nerves (goldfish Carassius auratus; K. Lundin, unpublished observations) and in endocrine/paracrine cells [eel Anguilla anguilla and rainbow trout Oncorhynchys mykiss (Lundin and Holmgren, 1989)]. However, 5-HT could not be detected in the cod swimbladder by the same antisera that reveal immunoreactive nerves and endocrine cells in the gut of the cod (Jensen and Holmgren, 1985; Lundin and Holmgren, 1989; K. Lundin and S. Holmgren, unpublished observations). 5-HT constricts the smooth muscles of both cod and eel swimbladder in vitro, possibly mimicking a hormonal effect (Lundin and Holmgren, 1989), but its effects on filling rate in the present study are inconsistent, with large variations between individuals. This large variation in vivo may be the result of interactions between general effects of 5-HT on the fish (e.g. on heart rate and blood flow) and more specific effects on the swimbladder mucosa and gas gland, and prevents any further elucidation of the mechanisms of action of 5-HT in the teleost swimbladder from the results of this study.

The cholinergic (muscarinic) antagonist atropine produced only a partial blockade of gas secretion, implying that additional mechanisms are involved. One possible candidate is the VIP-like peptide that is present in large numbers of nerve fibres in the swimbladder mucosa and in fewer nerve fibres in the gas gland and vagus nerve (Lundin and Holmgren, 1984, 1989). In teleosts, VIP is a potent vasodilator (Jensen et al. 1991; Lundin and Holmgren, 1984), and usually relaxes smooth muscles, including the muscularis mucosae of the swimbladder (Aidman and Holmgren, 1987; Lundin, 1991; Lundin and Holmgren, 1984, 1986). Cod VIP differs from porcine VIP in five positions but is virtually equipotent with mammalian VIP in bioassays on mammals (Thwaites et al. 1989).

The potent inhibitory effect of VIP in control fishes disappeared after atropine treatment, indicating that VIP is modulatory on the cholinergic neurones of the gas gland and normally has an inhibitory effect on them. This is in agreement with findings in mammals, where VIP inhibits neurotransmission in the cholinergic pathways controlling the smooth muscle of the guinea pig airway (Ellis and Farmer, 1989; Martin et al. 1990) and is thought to inhibit the release of acetylcholine in the uterine cervix (Stjernqvist and Owman, 1987). However, the most common effect of VIP on secretory events in mammals is a stimulation (see Dimaline, 1989), and the possibility that porcine VIP acts as a partial agonist in the fish tissue cannot be completely excluded.

The inhibitory effect of atropine on the gas secretion was removed by VIP. When the cholinergic muscarinic pathways are blocked by atropine, other effects of VIP, such as a relaxation of the smooth muscles of the swimbladder mucosa and an increase in the blood flow to the gas gland (Lundin and Holmgren, 1984) may dominate, which would facilitate the filling of the swimbladder. It is also possible that endogenous cod VIP has a dual effect on gas secretion: a stimulatory effect directly on the secretory cells or indirectly via non-muscarinic pathways and an inhibitory effect on cholinergic neurones acting on muscarinic receptors. Under experimental conditions, the direct effect is revealed only after atropinization, when the gas gland cells are not maximally stimulated by the cholinergic neurones.

Using immunohistochemistry, substance-P-like material has been found in only a few nerve fibres in the cod swimbladder mucosa, and none has been found in the gas gland (Lundin and Holmgren, 1989). Ligation of the vagus nerve causes accumulation of a substance-P-like peptide (J. Jensen, A.-C. Jonsson and S. Holmgren, unpublished results), but its transport along the branch to the swimbladder has not been studied. Immunoreactive endocrine cells occur frequently in the intestine of the cod (Jensen et al. 1987), and it is possible that at least part of the effect of substance P on the swimbladder mimics an endogenous hormonal mechanism.

The stimulatory effect of substance P revealed after atropine treatment is unlikely to be via the cholinergic muscarinic pathways, but may be directly on the swimbladder tissues or indirectly via non-muscarinic pathways. Exogenous substance P is a potent vasodilator in the cod gut (Jensen et al. 1991), but is otherwise constrictory on teleost smooth muscle (Holmgren, 1983; Jensen and Holmgren, 1985; Kitazawa et al. 1988), including those of both the resorbent and the secretory parts of the swimbladder (Lundin, 1991). Dilation of the blood vessels to the swimbladder and constriction of the resorbent part of the swimbladder by substance P would facilitate filling.

The blockade of secretion obtained by vagotomy could not be reversed by carbachol, VIP or substance P, separately or in combination, indicating that there are additional, unidentified fibres in the vagus that must be intact for gas secretion to occur. Nevertheless, the fish is affected in many ways by vagotomy, which may also impair normal function at the receptor level. It is also possible that the failure of carbachol to stimulate gas secretion in the vagotomized fish is due to simultaneous effects on the whole body, such as lowering of the heart rate and cardiac output, resulting in a reduced blood flow to the gas gland.

In conclusion, it is probable that the peptides VIP and substance P (in combination with cholinergic and adrenergic mechanisms) are involved in the control of secretion of gas into the cod swimbladder, and that they employ separate mechanisms. In addition, the activity of unidentified ‘secretory’ fibres in the vagus nerve may be necessary for secretion.

The authors wish to thank Dr David Grove and Professor Stefan Nilsson for valuable suggestions and ideas and Dr Magnus Lundin for constructing the computer program used in this study for the area measurements. This work was supported by grants from the Helge Ax:son Johnson, Anna Ahrenberg and Wilhelm and Martina Lundgren foundations and from the Swedish Natural Science Research Council.