Mechanisms Controlling Stomach Volume of the Atlantic Cod (Gadus Morhua) Following Gastric Distension

We have recently reported that distension of the cardiac stomach in the rainbow trout (Oncorhynchus my kiss), caused by inflating an intragastric balloon, led to a sustained increase in compliance of the gastric smooth muscles (Grove and Holmgren, 1992). The large decrease in stomach tone following distension appeared largely to be caused by changes in activity of excitatory neurones within the cardiac stomach wall itself. Reflex, rhythmic contractions (dependent on cholinergic and, to some extent, serotonergic nerves) were readily suppressed by somatostatin (SST). The overall tone of the muscle wall, which limited the rate of distension, depended primarily on activity in serotonergic nerves; vasoactive intestinal polypeptide (VIP) suppressed their effect.

The question remains whether similar processes are found in other fish. We elected to study in vivo and in vitro preparations of the stomach of the cod Gadus morhua, since, like the trout, some information on the functioning of its gastrointestinal system is already available (Nilsson and Fänge, 1969; Larsson and Rehfeld, 1977; Holstein, 1979, 1982, 1983; Holstein and Cederberg, 1984, 1986; Holmgren et al. 1985; Jensen and Holmgren, 1985; Jensen et al. 1987, 1991; Jonsson et al. 1987; Holmgren and Jonsson, 1988). A brief report is also included, for comparison with cod and trout, of in vivo responses to gastric distension obtained from four species of marine flatfish.

The methods used were essentially those reported in our previous study on rainbow trout (Grove and Holmgren, 1992). Atlantic cod Gadus morhua (body mass 280–900 g) were held in the laboratory in recirculating, aerated sea water at 12°C. In addition, several individuals of each of the following flatfish were held at 14°C for preliminary examination: plaice (Pleuronectes platessa, 300–500g), sand dab (Limanda limanda, 200–400 g), brill (Scophthalmus rhombus, 500–600 g) and turbot {Scophthalmus maximus, 500–700 g).

For in vivo study, fish were anaesthetized in MS222 (1:20000) and held on an operating table whilst their gills were irrigated with anaesthetic in sea water. An intragastric balloon, connected to a movable, weighed water-filled reservoir, was inserted into the cardiac stomach by way of the mouth. Vagotomy in the cod was performed by cutting the intestinal branch of the vagus nerve on each side near the gills.

For in vitro study, the stomach and associated blood vessels were removed and mounted in an organ bath. Perfusion of the stomach with cod Ringer at constant pressure (Holmgren and Nilsson, 1974; Holmgren et al. 1985) allowed drugs to be tested, at known concentrations and during distension by the intragastric balloon, in the absence of extrinsic nerve reflexes. Electrical stimulation of the intestinal branches of the vagus nerve was undertaken using a Grass model S7 stimulator with bipolar platinum electrodes.

Calculations were performed as described previously (Grove and Holmgren, 1992). When analyzing the rates of entry of fluid into the intragastric balloon after the hydrostatic pressure had been raised, the slower expansion in phase 2 has been treated as an exponential process between V_res and V_max. Distension rate is expressed as an instantaneous rate (K min⁻¹) to allow comparisons between treatments and with findings on the rainbow trout (Grove and Holmgren, 1992).

Immunohistochemical examination of tissues from the cod gastrointestinal tract to detect serotonergic nerves were made using the whole-mount technique of Costa et al. (1980). The stomach and intestine were cut open, pinned to dental wax and fixed for 18 h by floating in Zamboni’s fixative (2 % formaldehyde and 15 % picric acid in 0.1 mol I⁻¹ phosphate buffer, pH 7.4), rinsed in 80% ethanol, dehydrated, treated with xylene for 30 min and rehydrated. The samples were kept in phosphate-buffered saline (PBS, pH7.4) until further processing (within 24 h). The tissues were peeled so that samples from submucosa, circular muscle and longitudinal muscle could be incubated with primary antiserum (serotonin antiserum 932, Immuno Nuclear Corporation, diluted 1:100) for 18h. They were rinsed in PBS for 1 h and incubated with secondary antiserum (swine anti-rabbit IgG conjugated with fluorescein-isothiocyanate; DAKO Immunoglobulins a/s, Copenhagen, Denmark diluted 1:10) for 1.5 h. After rinsing in PBS for 1 h, the preparations were mounted in glycerol-carbonate buffer (1:1, pH 8.4) and viewed with a Leitz Dialux fluorescence microscope.

Sections of the gut were also examined. After fixation as above, the preparations were put in sodium cacodylate containing 20% sucrose for at least 24 h, embedded in Tissue-Tek (Miles Scientific, Naperville, Illinois) and snap-frozen in liquid nitrogen. Sections (10 μm) were cut on a cryostat, collected on slides and incubated and treated as above.

The drugs used in the present study were atropine sulphate (Sigma); phentolamine hydrochloride (Sigma); MS222 (tricaine methane sulphonate, Sigma); metenkephalin (Sigma); methysergide (1-methyl-lysergic acid butanolamine bimaleate, Sandoz); neurotensin (Sigma); synthetic somatostatin-14 (Sigma); tetrodotoxin (Sigma); natural porcine vasoactive intestinal polypeptide (VIP, V. Mutt, Stockholm).

In Fig. 1, the response of the cod cardiac stomach to inflation at low pressures is shown. Application of minimal hydrostatic pressure (0.2kPa) was followed, as in the rainbow trout, by an immediate inflow (phase 1) of 2–3 ml of water (V_res). Thereafter, contractions of the muscles began and the stomach slowly relaxed (phase 2). Tests showed that rather greater pressures (0.4–0.5 kPa) were required to produce full distension, and these were used throughout subsequent studies. In Fig. 1, this pressure was applied after 14 min of phase 2. It was followed by an increase in both amplitude and frequency of muscular contractions and by the slow relaxation of the stomach towards its maximum volume (V_max). V_max was usually 6–7.5 ml per 100 g body mass, although occasional cod (2 out of 7) had lower cardiac stomach volumes than expected. The stomach volume may reflect the previous feeding history of the fish before study (Flowerdew and Grove, 1979; Treasurer, 1988).

Following the initial distension, the relaxation process in phase 2 took more than 30 min to complete. The instantaneous rate of relaxation (K) in vivo in phase 2 was 0.02–0.07 min⁻¹, values close to those described for the trout. Fig. 1 also shows, from experiments in which deflation of the stomach is followed by reinflation a few minutes later, that the stomach has become fully compliant; it accepts its maximum volume almost as fast as the balloon can inflate (K approaches values of 2.0–3.9 min⁻¹). Even after prolonged rest (several hours in some fish), when the stomach regains its ability to perform large contractions during phase 2, the relaxation rate remains rather high (K =0.4–0.6min⁻¹). Three fish were subjected to vagotomy, but this operation did not affect the responses to gastric distension in vivo.

These results closely resemble those found previously for the rainbow trout. The main differences were that the volume accepted on first distension (V_res) during phase 1 was smaller in the cod (approximately 14–22% of maximum stomach capacity, cf. 30% in trout) and the cardiac stomach volume relative to body mass was higher (6–8ml per 100g, cf. 2–3 ml per 100g in trout).

The body masses of the 39 cod used in this part of the study ranged between 285 and 900g. The mass of the stomach, and the volume it would accept at 0.5 kPa, varied unpredictably - as found in vivo. In general, the stomach, with its appendages and fat stores, formed approximately 3.3% of the body mass. Eighteen of these fish had maximum cardiac volumes comparable to those found in vivo (i.e. 5.5–9ml per 100g), but the remaining fish had probably not been feeding prior to testing and their stomachs accepted smaller volumes, ranging from only 2.5 to 4.9 ml per 100 g.

Fig. 2A shows the response of the isolated, perfused cardiac stomach of the cod to distension. As expected, the previously undistended organ initially resisted the inflow of water. However, several changes were immediately clear when the in vitro preparations were compared with the in vivo preparations. Phase 1 was effectively abolished (V_res was close to zero). Subsequent relaxation during phase 2 was usually somewhat faster than found in vivo (K=0.1–0.2min⁻¹). Furthermore, a delay of only 30 min between a first and second test revealed little sign that compliance had developed.

This preparation was then subjected to repeated distension with short, but varying, periods of rest interposed, during which the stomach was deflated (Fig. 2B). After previous inflation and a rest of only a few minutes (and unlike the response to the first inflation in vitro), phase 1 developed; up to 60% of the maximum volume was accepted before reflex contractions limited inflow. Phase 2 is characterized by rapid, low-amplitude contractions (approximately 2–4min⁻¹). With the passage of time, up to 15 min between deflation and re-inflation, the stomach quickly regained its ability to contract reflexly. It was clear in this and three other preparations (see, for example, Fig. 2C) that the isolated cod stomach retained its compliance after a previous distension for a much shorter period than was found in the cod in vivo or in the trout in vivo or in vitro (see Fig. 4 in Grove and Holmgren, 1992).

The development of compliance in the isolated cod stomach primarily involved a short-term suppression of reflex contractions. The large, up to ten-fold, increase in phase 2 relaxation rate (K) found in vivo no longer occurred. K rarely exceeded 0.4 min⁻¹. During the course of a number of our experiments, resistance of the stomach to distension slowly increased, a phenomenon never seen in the trout, or in the cod in vivo. In such preparations, after approximately 2–3 h, the maximum volume of the cardiac stomach was reduced by 25 %. Preparations showing this trend were not used in the drug tests described below.

Perfusion of the cod stomach (N = 2) with tetrodotoxin (10⁻⁷–10⁻⁶moll⁻¹) to impair the function of enteric neurones produced consistent but unexpected responses when compared to the trout. The resting stomach partially relaxed as the toxin was infused (Fig. 3), indicating that, prior to blockade with TTX, some excitatory nervous activity leading to raised muscle tone was present. This may explain the rather smaller maximum volume observed in many fish in vitro and the tendency for V_max to decrease with time in some of the preparations. On distension in the presence of TTX, a small increase (approximately 8–10%) in V_max was found which, when added to the relaxation before distension was applied, amounted to a final volume some 25 % above that in the control period. Relaxation rate (K) in phase 2 increased in the TTX-treated preparation to reach values close to 0.7min⁻¹. Unlike the situation in trout, spontaneous contractions during distension were not abolished. Instead, the original contractions were replaced by larger, low-frequency contractions (0.7–1 min⁻¹). These may represent myogenic contractions similar to those found in Pleuronectes platessa after treatment with the same toxin (Stevenson and Grove, 1977, 1978).

The major difference between our in vivo and in vitro cod preparations is the absence of extrinsic innervation. Distension of the cod stomach was therefore carried out whilst the intestinal branch of the vagus nerve was stimulated electrically. Stimulation with 1 ms pulses (Fig. 4) partially contracted the isolated stomach (raising the recorded baseline) without abolishing spontaneous contractions. The response to distension was somewhat reduced (Fig. 4A) or unaffected (Fig. 4B). We found no evidence for inhibitory tracts running in the vagus, using this type of stimulation, confirming the findings of Nilsson and Fange (1969). Perfusion with atropine (10⁻⁶moll⁻¹) during stimulation had several effects. The baseline was returned to pre-stimulation levels, spontaneous contractions were abolished (both before and during distension) and relaxation rate (X) increased dramatically (from 0.4 to 1.7min⁻¹, Fig. 4A). Results from similar preparations suggest that the excitatory fibres running in the vagus are cholinergic. Much of the resistance to distension in the isolated stomach, in the absence of extrinsic vagal stimulation, may also depend on reflex activation of excitatory cholinergic nerves in the gut wall.

To examine this hypothesis further, selected antagonistic drugs were perfused into the stomach vasculature in micromolar quantities to test their effects on the responses to distension. The adrenergic antagonist phentolamine (N = 6) sometimes increased resistance to distension (Fig. 5A,D), usually had little effect (Fig. 5B,C) and occasionally contracted the relaxed stomach to raise the baseline (Fig. 5D). The basis for these variable effects are not known. The serotonergic blocker methysergide (N=7), unlike its effects in the trout, where it relaxed the stomach, had little or no effect on the response to distension in the cod (Figs 5B, 6A,B). Atropine (N=7), however, had clear and consistent effects. It abolished increased gastric tone when this was raised by phentolamine (Fig. 5D). It typically increased relaxation rate during phase 2 of distension (Figs 5C, 6A–D), although in one preparation (Fig. 5D) it failed to overcome the effects of phentolamine. In addition, atropine abolished spontaneous contractions, immediately raised distended stomach volume to its maximum (V_max) and dramatically increased relaxation rate (K) to values (1–2 min⁻¹) comparable to those found in the fully compliant stomach in vivo.

These results, especially those following methysergide, strengthen our belief that gastric tone and reflex contractions in the cod are maintained primarily by activity in intramural cholinergic excitatory nerves alone. Unlike the situation in trout, serotonergic nerves play an insignificant part in maintaining the tone and resistance to stretch of the cardiac stomach smooth muscle. Serotonergic intramural neurones have been reported from the cod intestine (Jensen and Holmgren, 1985), but studies on the cod stomach have not been made previously. Immunohistochemical examination of the cardiac stomach of the cod (which is the part recorded from in the present study) failed to reveal the presence of serotonergic neurones, while a few varicose fibres were observed in the submucosa of the pyloric part. However, fibres frequently occurred in the intestine, particularly in the submucosa and the myenteric plexus, confirming previous results (Jensen and Holmgren, 1985). This finding is consistent with our conclusions from drug treatment, that serotonergic nerves are not involved in the control of the smooth muscle of the cardiac stomach of the cod. Accordingly, the great increase in compliance that occurs after distension in vivo in the cod (Fig. 1) requires the intervention of an agent that suppresses cholinergic activity.

In the trout, vasoactive intestinal polypeptide (VIP) proved to be a powerful stomach relaxant, but its action appeared to be directed primarily towards suppression of the effect of intramural, serotonergic nerve activity, which maintains gastric muscle tone. In the cod stomach, where serotonergic mechanisms seem to be weak or absent, perfusion with VIP (10⁻⁹–10⁻⁷ moll⁻¹, N=3) had little detectable effect on the response to distension (Fig. 7A). Negative results were also obtained following perfusion with met-enkephalin (10⁻⁹–10⁻⁷moll⁻¹, N=3, Fig. 7B) or neurotensin (10⁻¹¹–10⁻⁷ molL⁻¹, N=4, Fig. 7C). These two peptides are known to have inhibitory effects on gastric motility in mammals (Hellström et al. 1982; Garzon et al. 1987; Okamoto et al. 1988).

Cholinergic intramural neurones in the trout were primarily associated with the control of reflex contractions following distension, even as gastric tone fell. This reflex recovers relatively quickly after distension but was readily suppressed by somatostatin (SST). Similar reflexes occur in the cod (Fig. 2) so the effects of this peptide (10⁻⁹–10⁻⁷moll⁻¹, N=3) were also tested (Fig. 8). Somatostatin (10⁻⁷moll⁻¹) abolished spontaneous contractions in the resting stomach and decreased muscular tone (cf. the effect of tetrodotoxin). It did not, however, abolish spontaneous contractions following distension or increase compliance.

In general, the studies on isolated cod stomachs suggest that removal from extrinsic control (whether nervous or endocrine) prevents the long-term relaxation/inhibition that follows distension.

To test whether other species, besides rainbow trout and cod, also show increased compliance after distension, we made a preliminary examination of the in vivo responses of four marine teleosts to gastric distension. All were flatfish, but they differed to some extent in feeding habits and in the relative size of their stomachs. The plaice (Pleuronectes platessa) has a relatively small stomach (approximately 3–4 ml per 100 g body mass). Its natural diet primarily includes small molluscs and polychaete worms, although larger shrimps are taken as its size reaches 1.5 kg. The records show (Fig. 9A) that, when distended at pressures of 0.5 kPa, regardless of any previous stimulation, the stomach always distends rapidly and with little resistance (K = 2.5–3.5 min⁻¹). The results resemble those from fully compliant stomachs of other species. The plaice, however, is known to have a reduced stomach, part of an evolutionary trend that has led to loss of the stomach in some of the other Pleuronectidae (Grove and Campbell, 1979). Its close relative Limanda limanda has a larger stomach (approximately 8 ml per 100 g body mass) and ingests a wider range of forage species, including shrimps, crabs, molluscs, worms and brittle-stars. Fig. 9B clearly shows that, after resistance during the initial distension, compliance develops and lasts for at least 1 h (K increases from 0.3 to 1.5 min⁻¹). Two members of the Bothidae, Scophthalmus rhombus and S. maximus, were similarly tested. These fish consume fish as a major part of their diet and have even larger stomachs (8–10 ml per 100 g body mass). In both species (Fig. 9C,D), repeated stimulation increased relaxation rate from initial values of approximately 0.6 min⁻¹ to rates usually three times higher; these persisted for at least 30min. It appears that the responses to increased gastric compliance following distension in cod and trout are also found in other fish that possess well-developed stomachs.

The cod is an active predator in the sea, consuming a variety of relatively large organisms, including crustaceans, molluscs and fish (Daan, 1973). Compared with the rainbow trout, it has a relatively large stomach, eats larger organisms and can feed to satiation within a few minutes, rather than over a period of approximately Ih, when suitable food is available. Stomach distension in vivo, to mimic the arrival of food, leads to a profound and long-lasting increase in compliance similar to that found in rainbow trout.

However, several major differences have been established between these two species as a result of our studies. The trout depends on intramural serotonergic neurones to maintain gastric tone, whilst both serotonergic and cholinergic nerves are involved in rhythmic, reflex contractions induced by distension. In isolated stomachs, VIP opposes serotonergic nerve effects whilst SST suppresses cholinergic activity. The cod lacks the serotonergic mechanism and VIP has no effect on the isolated stomach. Instead, responses to stretch mainly depend on the reflex activation of cholinergic nerves and (after prolonged distension in vivo) the suppression of their activity. SST failed to suppress contractile activity following distension, although spontaneous contractions of the undistended stomach were abolished and its tone reduced.

It is important to bear in mind that the cod and trout are widely separated in teleost taxonomy. Failure of the cod to react to a mammalian peptide (such as VIP) might mean that its equivalent peptide is different in primary structure from that infused, and that its receptors reflect such changes. However, the structure of cod VIP has been described (Thwaites et al. 1989) and it shows little difference from mammalian and avian VIP molecules in a variety of assay systems.

The isolated, perfused trout stomach reacts in a similar fashion to recordings made in vivo, under anaesthesia. However, the isolated cod stomach, unlike the in vivo preparation, loses compliance rapidly when allowed to deflate. This suggests that factors extrinsic to the cod stomach are essential to promote prolonged receptive relaxation. These factors may be humoral or may involve the splanchnic nerves.

In neither the anaesthetized cod nor trout were we able to demonstrate changes in gastric reactions following vagotomy. It may be inappropriate to conclude that extrinsic nervous reflexes, similar to those found in mammals (Cannon and Lieb, 1911; Abrahamsson, 1973; Abrahamsson and Jansson, 1973), play no part in gastric receptive relaxation or accommodation in these fish. Instead, clarification of possible extrinsic reflexes, involving vagal and splanchnic pathways, should involve studies of gastric motility in conscious, actively feeding fish (e.g. Edwards, 1973) rather than tests in the presence of anaesthetic. There is currently a rapid expansion in the number of neurocrine and endocrine substances known to be present in fish and that are likely to affect gastrointestinal muscles. Full understanding of the coordination of stomach motility, and receptive relaxation, during feeding and digestion will depend on such developments. The factor that promotes prolonged gastric relaxation in the intact cod (Fig. 1) remains to be found.

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