Intrinsic Mechanisms Controlling Cardiac Stomach Volume of the Rainbow Trout (Oncorhynchus Mykiss) Following Gastric Distension

In higher vertebrates, filling of the stomach leads to a reflex relaxation of the proximal (fundus and corpus) gastric musculature which is believed to be necessary to increase receptive capacity whilst avoiding sharp rises in intragastric pressure. The reflex was called ‘receptive relaxation’ by Cannon and Lieb (1911) since it can be induced by swallowing. The phenomenon depends on extrinsic nerve pathways which suppress excitatory nerves and activate vagal efferent inhibitory tracts (Martinson, 1965; Jansson, 1969; Abrahamsson, 1973; Jahnberg, 1977). Similar extrinsic reflexes can be induced by distension of the oesophagus (Jansson, 1969) or of the stomach itself (Abrahamsson and Jansson, 1973).

However, as gastric volume is deliberately increased following vagotomy, intragastric pressure rises sharply (Jansson, 1969; Staadas and Aune, 1970; Staadas, 1975). Furthermore, distension of the antrum can activate, in addition to vagal reflexes, extrinsic inhibition by way of splanchnic nerves (Abrahamsson, 1973). In addition, chemical or mechanical stimulation of the upper intestine can reflexly inhibit contraction of the stomach by vagal and/or sympathetic nerve pathways (Roman and Gonella, 1987). The term ‘gastric accommodation’ has been reserved for reflex, vagal relaxation induced by distension of the stomach itself, which leads to decreased activity in efferent excitatory nerves (Scratcherd and Grundy, 1982).

There is no direct evidence for similar reflex relaxation mechanisms in fish but field observations indicate that contraction of the stomach may be inhibited following ingestion of food since actively feeding fish of a variety of species have relaxed (‘floppy’) stomachs. The effects of vagal and splanchnic nerve stimulation on gastric muscles of teleosts have been reviewed by Nilsson (1983). Inhibitory vagal effects have been described in Salmo gairdneri (=Oncorhynchus my kiss) and Conger conger, whilst both inhibitory and excitatory tracts occur in Pleuronectes platessa and Lophius piscatorius. Only vagal excitatory influences were reported for Gadus morhua and Platycephalus bassensis. Splanchnic nerves contain excitatory fibres (Salmo, Platycephalus) or both excitatory and inhibitory tracts (Pleuronectes). The possibility exists that these tracts could support gastric receptive relaxation or accommodation in fish, but no experiments have been reported.

The present study has, therefore, been undertaken to examine gastric reflexes following stomach distension in the rainbow trout Oncorhynchus my kiss (=Salmo gairdneri). This species is important commercially in aquaculture but has also been the subject of relatively extensive study in relation to its feeding and gastrointestinal physiology (e.g. Grove et al. 1978; Holmgren and Nilsson, 1981; Holmgren, 1983; Holmgren et al. 1982; Holmgren et al. 1985; Vigna et al. 1985; Thorndyke and Holmgren, 1990; Aldman et al. 1991). It is also possible that whole-organ studies might prove useful in examining the effects of putative neurocrine and endocrine substances that have been found in the rainbow trout gastrointestinal tract. A preliminary account of the study was given in Grove (1986).

Rainbow trout (600-1700 g body mass) were held in the laboratory in circulât-ing, aerated fresh water at 12±1°C and fed every second day with rainbow trout pellets in quantities that represented approximately 2 % of body mass per meal.

For in situ studies, fish (unfed for the previous 24 h) were anaesthetized in MS 222 (1:20000) and held ventral-side-up on an operating table whilst the gills were irrigated with anaesthetic at 12-13.5°C. A balloon (approximately 100ml maximum volume) was passed down the oesophagus into the cardiac part of the stomach. The balloon was connected by a water-filled tube (6mm i.d.) to a wide reservoir (capacity 100ml), suspended from a Grass FT03C strain gauge. Clamping the tube to the movable mounting allowed the reservoir to be raised or lowered whilst the volume of water remaining in the reservoir was recorded on a Grass Polygraph model 7 recorder. The dimensions were such that the reservoir could be raised from stomach level to a height of up to 20 cm, and the outflow of 20-30 ml of water into the balloon would only lower the reservoir level by a small amount (2cm water maximum). This simple arrangement allowed both overall changes in gastric volume as well as reflex gastric contractile activity to be recorded simultaneously. During experiments, steady respiratory movements of the fish continued, indicating that the medulla oblongata was not affected by this level of anaesthesia. Fish could be held in this way for up to 12h and subsequently recovered in the main aquarium.

In some fish, the intestinal branches of the left and right vagus were sectioned near the gills. Such fish were killed at the completion of the test without being allowed to recover from the anaesthetic. Where drugs were tested, appropriate concentrations, dissolved in 0.5-1 ml of saline, were injected into the caudal vein.

For in vitro studies, the stomach was perfused through the coeliac artery with rainbow trout Ringer and drained by the hepatic portal vein (Holmgren et al. 1985). The organ was removed and mounted in an organ bath so that external nerve reflexes were abolished. Test drugs (0.1-1 μmol I⁻¹ in rainbow trout Ringer) were infused into the arterial catheter to determine any effects on the responses to gastric distension. At least 20min exposure to any drug was allowed prior to testing its effect on distension.

When describing and quantifying the responses collected using this distension technique, we have adopted the following terms (indicated in Fig. 1). (1) With the reservoir at stomach height (i.e. the applied hydrostatic pressure set at zero), no water enters the intragastric balloon and the reservoir holds its maximum water mass. The uninflated balloon and intragastric tubing occupied approximately 4ml. (2) On raising the reservoir to apply a stated hydrostatic pressure, rapid entry of water from the reservoir into the balloon partially distended the stomach (phase 1). This volume (V_res) equals the change (decrease) in reservoir mass (which is recorded) and is usually limited by a reflex contraction of the stomach. The water inflow rate in this phase is close to that imposed by the resistance of the balloon itself to distension. (3) Following phase 1, in a stomach subjected to its first distension, a prolonged and slow further increase in gastric volume develops (phase 2). This is typically accompanied by the development of spontaneous contractions of the smooth muscles of the cardiac stomach wall. This expansion (relaxation) occurs at a steadily decreasing rate until volume becomes maximal (Vmax).

The drugs used in the present study were atropine sulphate (Sigma); phentola-mine hydrochloride (Sigma); DL-propranolol hydrochloride (Sigma); MS222 (tricaine methane sulphonate, Sigma); methysergide (1-methyl-lysergic acid butanolamine bimaleate, Sandoz); synthetic somatostatin-14 (Sigma); tetrodo-toxin (TTX, Sigma); natural porcine vasoactive intestinal polypeptide (VIP, V. Mutt, Stockholm).

Gastric distension was achieved by raising the fluid-filled reservoir to a known height above its attached intragastric balloon. The reservoir content was measured continuously. Pressures in excess of 0.2 kPa were required to induce clear responses. For a 630g fish stimulated at high pressure (1.2kPa), an immediate entry of 4-5 ml (=V_res) of fluid occurred, followed by the onset of spontaneous contractions and a slow, subsequent increase in cardiac gastric volume (Fig. 1). The maximum volume accepted (V_max) tended to increase irregularly with fish size. Values were between 10 and 30 ml over the size range 600-1900 g body mass.

Six fish of similar size (600-750 g) were tested using this method, with the following results for the first occasion when distension was induced. In phase 1, V_res increased with pressure at approximately 6-7 ml kPa⁻¹ up to a maximum of 5-8ml. In phase 2, relaxation occurred at an instantaneous rate of 0.03-0.05 min⁻¹, i.e. half of the distension from V_res to V_max (an entry of a further 2.75 ml) required 15-20min.

If a newly distended stomach is deflated and then reinflated, fluid enters the relaxed stomach against a decreased resistance and V_res increases with each subsequent trial (Fig. 2B).

The above observations in situ suggest that gastric distension using a balloon induces a slow relaxation of the gastric musculature. In mammals, this phenomenon depends on a vagal inhibitory mechanism. A preliminary test was undertaken to examine whether a similar mechanism exists in the rainbow trout. After distension, the relaxed stomachs of four rainbow trout were deflated and allowed to rest for 0.5-1 h, during which time each fish was subjected to bilateral section of the intestinal branches of the vagus nerves. These nerves are probably mixed ‘vagosympathetic trunks’ (Nilsson, 1983). After a further 0.5-1 h post-operative rest to ensure that the fish was respiring normally, the effects of gastric distension were again tested (Fig. 3). In all four fish, the rate of distension was rapid and undiminished; little resistance to water inflow was seen. There was no indication that inhibitory activity in the vagal nerves was essential for the maintenance of the relaxed state of the rainbow trout stomach.

Similar studies were then made on five perfused, isolated stomachs removed from rainbow trout (body mass 750-1000g) and suspended in an organ bath. These tests were undertaken to determine responses in the absence of all extrinsic nerves, without prior distension. It was possible, as expected from mammalian studies, that the absence of inhibitory autonomic (vagal or sympathetic) pathways might limit the relaxation observed in vitro. In the event, no such limitation was found. Instead, the patterns of response were similar to those found in situ save that fluid entered the stomach with greater ease. At low pressures, V_res increased by 13 ml kPa⁻¹ ; nearly double the rate found in situ. Above 0.6 kPa, V_res increased to approximately 15-18ml. During phase 2, the instantaneous rate of relaxation was also higher (X’=0.04-0.06min⁻¹). V_max was 30-45 ml in the animals tested (body mass approximately 825 g). All these measures suggest that removal from central influences does not decrease the ability of the cardiac stomach to relax when distended. The pattern of relaxation seen in Fig. 1 appears to depend on changes induced in the cardiac stomach itself, and to be unaffected by extrinsic nerve reflexes. The increased distension rates seen in vitro could be caused by the absence of tonic (excitatory) influences of extrinsic nerves but may simply reflect removal of the stomach from mechanical constraints imposed by containment within the body wall and mesenteries.

Comparison of Figs 1 and 2 shows that, once a stomach has relaxed in response to distension, it retains this relaxed state for some time, even if deflated for a period. To investigate this retained compliance, studies were made in situ (N=4) and in vitro (N=3) in which repeated distensions were made, with interposed deflations, at known time intervals. In Fig. 2A, a perfused preparation was distended for 4min in every 6min at low (but above-threshold) pressures. The successive re-inflations under constant conditions showed that V_res progressively increased. The same pattern occurred in situ (Fig. 2B) even when longer periods (20 min) were left between tests. We found that the stomach remained relaxed (compliant) for some time after stimulation.

In an attempt to measure the duration of such compliance, the tests shown in Fig. 4 were carried out. Fully compliant stomachs were obtained by prolonged distension (approximately 40min) both in situ (N=3) and in vitro (N=3). When such preparations were deflated and then rapidly re-inflated, V_res was maximal and approximately equal to V_max, i.e. there was little resistance to fluid entry. As the duration of rest periods before re-inflation increased, the magnitude of V_res decreased. The rate of recovery of these high values of V_res towards the lower, initial levels was approximately exponential, at instantaneous rates of 0.02-0.03 min⁻¹, indicating that half the recovery towards initial levels is completed in 20-30 min.

Close inspection of records of the type shown in Fig. 4 suggests that two mechanisms are likely to be at work. The magnitude of V_res (end of phase 1) is determined by the onset of reflex contractions of the gut wall in response to distension. However, when the onset of these contractions has returned to the control value (after a rest of 30 min or more between inflations), it is clear that the relaxation rate (K) during phase 2 still remains very high when compared with controls. In Fig. 5, two preparations are shown which were re-inflated after 3h of rest. It can be seen that the relaxation rate in phase 2 is faster even after this long period of rest.

A large number of records were re-examined to find whether any pattern of change in this relaxation rate (K) could be detected with the passage of time for preparations that had been subjected to further distensions after the first test. For clarity, later values of K (K_t) are expressed as a multiple of the initial value (K₀) on first distension for that preparation. The findings are shown in Fig. 6 (with logarithmic axes for convenience of scale) and the main conclusion is that, after 20-30 min of distension, the relaxation rate during phase 2 rises steadily to reach values that typically correspond to K=0.25-0.5 min⁻¹ (cf. 0.03-0.05 min⁻¹ at the start). In no case did we observe a reversal of this trend in K within the duration of our experiments, which typically lasted up to 8 h and with rest periods of up to 4.5 h.

To summarise this part of the study, a stomach with an initial V_res of 10 ml which would distend for 15-20 min during phase 2 to reach a capacity of 30 ml would, when fully compliant, only require 2-3 min to complete phase 2. After a number of trials, we found that it was possible to delay this increase in relaxation rate (K) in phase 2 to allow sequential, comparable responses to a given inflation pressure. Low pressures (approximately 0.5 kPa) applied for a few minutes at intervals no shorter than half an hour usually gave reproducible responses, so that the effects of drugs and other agents could be tested.

The observations described above suggest that the stomach wall possesses its own coordinating systems that (a) induce spontaneous contractions rapidly after distension (Fig. 1); and (b) induce a slowly developing, but large, increase in the rate of relaxation of the stomach wall (Fig. 6).

The stomach wall of the rainbow trout is known to contain a wide range of nerves and secretory cells whose functions include the control of the smooth muscle cells. To unravel the likely control mechanisms involved in receptive relaxation, the next step was to investigate the nerve or endocrine substances involved by perfusing drugs or biochemicals into the isolated stomach by way of its arterial supply.

Fig. 7 shows the invariable response of a newly prepared stomach preparation to brief, low-pressure distension, followed by prolonged rest and perfusion with I0⁻⁶ mol 1⁻¹ tetrodotoxin (N=4). The contraction that ends phase 1 (and determines V_res) is abolished, subsequent spontaneous contractions disappear and the relaxation rate during phase 2 increases to 0.2 min⁻¹. Tests were carried out at earlier times in other preparations and it was found that TTX rapidly resets K to high values, comparable to those found in fully compliant stomachs. It appears that both spontaneous contractions and resistance to distension depend on activation of enteric nerve cells.

The most commonly reported excitatory neurones in the vertebrate gut release either acetylcholine (ACh) or 5-hydroxytryptamine (5-HT), and it was shown that drugs antagonistic to ACh and 5-HT affect the response of the rainbow trout stomach to distension (Fig. 8). The results shown in Fig. 8A were obtained in situ on a fully compliant stomach (i.e. one that had been distended several times at low hydrostatic pressure, 0.4kPa, prior to the tests). Injection of atropine into the caudal vein suppressed the remaining weak, reflex contractions but the further injection of methysergide drastically increased relaxation rate in phase 2. Fig. 8B,C shows the effects on the responses to distension of perfusing the antagonists to 5-HT and ACh into the stomach wall in vitro. After the introduction of methysergide, there was an increase in V_res, but the contraction that ends phase 1 persisted. During phase 2, some reduction in spontaneous contractions occurred, but these were never abolished by this antagonist. Methysergide also induced an increased relaxation rate during this phase as well as an immediate increase in V_max-Addition of atropine, however, abolished both the reflex contraction that terminated phase 1 and any residual contractions in phase 2. In four other preparations tested at low pressures, 5-HT blockade invariably increased the rate of relaxation in phase 2 but the presence of atropine was obligatory to abolish spontaneous contractions. A fully compliant stomach distended rapidly after inflation and consequently methysergide had little effect. Atropine, however, suppressed the persisting, weak reflex contraction that terminates phase 1 (Fig. 8C).

On the assumption that intrinsic cholinergic and tryptaminergic nerves control reflex contractions and gastric muscle tone, respectively, the question remains as to which local agents could suppress their activity when the stomach fills, leading to relaxation and increased compliance. Such agents should have prolonged effects in comparison with ACh or 5-HT if the retention of compliance shown in Figs 3 and 4 is to be explained. In this study, we have concentrated on two polypeptide candidates for this role, somatostatin (SST) and vasoactive intestinal polypeptide (VIP).

The main effect of SST was to suppress both the reflex contraction that terminates phase 1 and the reflex contractions that occur in phase 2 (Fig. 9). In this regard, SST (N=3) resembles atropine, suggesting that it acts to suppress activity in cholinergic nerves that are stimulated in response to distension, without affecting maximum capacity (V_max).

VIP (N=4), when perfused into lightly distended preparations (Fig. 10A), in common with methysergide, partially suppressed contractile activity in phase 2 and increased both K and V_max. It also failed to abolish the contraction that terminates phase 1. When VIP was removed from the perfusate, spontaneous activity in phase 2 rapidly recovered, but the increase in both K and V_max persisted, even after delays of more than 1 h. In more strongly distended stomachs (Fig. 10B), where compliance was high, VIP noticeably increased V_max and this effect persisted for at least 1 h after washing through with Ringer.

In common with other salmonids, the rainbow trout has a well-developed stomach that is divided into a receptive, cardiac region leading into a narrower, muscular pyloric region which coordinates transfer of food to the intestine via the pyloric sphincter. When feeding voluntarily on suitably sized particles, it is not unusual for a hungry rainbow trout of the size used here to consume meals approaching 2.5 % of its body mass. However, satiation times of rainbow trout well in excess of 40 min have been recorded (Ishiwata, 1968; Grove et al. 1978) and this is also true of other salmonids (Brett, 1971; Elliott, 1975). The fish prefer relatively small food items, which they consume one at a time, often retiring to cover or to deeper water to handle and swallow each item before returning for another. The slow onset of full gastric relaxation in this species may be a factor in this slow attainment of satiation. Similarly, gastric emptying requires many hours, and the prolonged relaxation that follows simple, mechanical distension is presumably an essential part of the storage and digestion process.

As a result of our experiments, we propose the following model to explain the likely sequence of mechanical changes that occur in the rainbow trout stomach as feeding begins. The acceptance of food particles during a meal leads to an increasing distension of the stomach wall. This distension is opposed primarily by reflex activation of intramural nerves, which induce muscle contractions and maintain tone. Cholinergic neurones control reflex contractions (which in our recordings terminate phase 1 and determinate V_res). The continuing contractions (seen in phase 2 of our study) probably correspond with existing gastric digestion, where the food is mixed with pepsin and hydrochloric acid in the stomach. Serotonergic nerves, whose presence has been established previously (Holmgren et al. 1985; Kitazawa, 1989), also help to control phase 2 contractions but, in addition, maintain the muscle tone which limits the rate of relaxation (K) towards maximum volume (V_max) in our study.

If gastric distension is maintained (as in feeding), the serotonergic nerve activity may be antagonized by the release of VIP, leading to a fall in tone and to profound gastric relaxation. VIP has been detected in neurones of the rainbow trout stomach (Holmgren et al. 1982) and, in the isolated perfused small intestine of the guinea pig, it has been found to suppress neuronal release of 5-HT following distension (Schwörer et al. 1989). We suggest that it is primarily this long-term suppression of serotonergic stimulation by VIP that is responsible for the high level of compliance which develops in the distended stomach (see Fig. 6).

The possible role of somatostatin in the physiological response to feeding is less clear. Somatostatin may be released from endocrine cells of the stomach wall or from more posterior regions of the gastrointestinal tract and its associated glands as chyme leaves the stomach and enters the intestine. Endocrine cells showing somatostatin-like immunoreactivity are present in the mucosa of the stomach of the rainbow trout (Holmgren et al. 1982), and somatostatin cells frequently occur in pancreatic tissues of salmonid species (Wang et al. 1986; Nozaki et al. 1988). The short-term suppression by somatostatin of enteric nerves that induce contraction in phase 2 would serve as a mechanism to control gastric emptying. Inhibition of the stomach in the rainbow trout may, therefore, include long-term, VIP-induced receptive relaxation, to accommodate a meal, as well as shorter-term, SST-induced inhibition of contractions necessary to control pulsatile gastric emptying.

It must be emphasised that many other factors, as yet unknown, may also be involved in the control of gastric motility in the rainbow trout, since new types of nerves and endocrine cells continue to be detected. The composition of the food is known to affect the motility and gastric emptying. For example, Grove et al. (1978) have shown for the rainbow trout that an increase in dietary energy slows gastric emptying (and consequently limits food intake) by increasing the time between voluntarily ingested meals. However, little is known of the mechanisms of action and which neurones and endocrine cells are mediating the effect.

Furthermore, it cannot be claimed that extrinsic nerves, such as the vagus or sympathetic, have no role to play. The present in situ study used only anaesthetized fish, and the actions of the anaesthetic on central reflexes, if any, are as yet unknown.

The skilled technical assistance of Gunilia Rydgren and preparation of figures by Birgitta Vallander is gratefully acknowledged, as is the support of the Swedish Natural Science Research Council.