Haemodynamic Effects of Secretory Agents On the Isolated Elasmobranch Rectal Gland

The relationship between blood flow and the secretory activity of glandular tissue has been the subject of numerous studies in the past and, in the case of mammalian salivary glands, has given rise to much controversy concerning the mechanisms involved. The salt-secreting nasal glands of birds have been studied in some detail from this viewpoint, particularly by Hanwell, Linzell & Peaker (1970, 1971), but no similar investigation has been made on the salt-secreting rectal gland of elasmobranchs. This is despite a suggestion made some 20 years ago by Burger (1962) that vasomotor changes may be a significant factor in controlling secretion rate by the gland. Kent, Peirce & Peirce (1973) published values of blood flow to the rectal gland in free-swimming fish, using labelled microspheres, and recorded a mean value or approximately 52 μl g gland^-1 min^-1 which, in the small specimens used in this study where rectal gland weight is equivalent to 0·08% of body weight (Kent, Peirce & Bever, 1971), converts to a value of 42 μl kg^-1 min^-1. Similarly, in pithed specimens of Squalus, Solomon et al. (1980) obtained a value of 3·31 mlkg^-1 h^-1 or 55μ1 kg^-1 min^-1 for rectal gland blood flow in control fish.

It is known that the elasmobranch rectal gland secretes sodium and chloride ions at a concentration approximately twice that of the plasma (Burger & Hess, 1960). On ap propriate stimulation, such as volume or salt loading, in vivo secretion rate may exceed 50μ1 kg^-1 min^-1 (Burger, 1962), and it is clear that the above flows would be insuf ficient to sustain such rates, implying that blood flow to the gland must increase during secretion. This suggestion is supported by the work of Burger (1962) who, from values of secretion rate and arterial-venous differences in plasma chloride concentration, estimated rectal gland blood flow in vivo and obtained values of 14–114 ml h^-1. Recalcu-lated on a per unit fish weight basis, this is equivalent to a range of 39–657 μl kg^-1 min^-1. Small, but significant increases in rectal gland blood flow on the stimulation of secretion were also observed in pithed preparations by Solomon et al. (1980).

However, in the isolated perfused gland, stimulation of secretion, for example by cyclic AMP + theophylline, is not apparently accompanied by any changes in per-fusion flow rate (Silva et al. 1980; Shuttleworth & Thompson, 1981). In view of the suggested role of circulating catecholamines in maintaining vasomotor tone in elasmobranchs (Opdyke & Holcombe, 1978), it was decided to investigate this paradox by examining the relationship between the haemodynamic effects of catecholamines and known secretory agents on the isolated perfused rectal gland.

Most of the experiments reported here were carried out at Exeter, using the European dogfish, Scyliorhinus canicula, but some initial data were obtained at the Mount Desert Island Biological Laboratory (MDIBL), Maine, U.S.A., using specimens of Squalus acanthias, and have been reported in a preliminary form elsewhere (Shuttleworth & Thompson, 1981). The results obtained with the two species were qualitatively and quantitatively similar, with the notable exception of the effect of vasoactive intestinal peptide (see Results).

Specimens of Scyliorhinus weighing 0·7—1·1 kg were obtained from the Plymouth Laboratory of the Marine Biological Association and were kept in large fibreglass tanks, supplied with sea water at a temperature of 11 °C, until use. Fish were pithed, opened ventrally, and most of the intestine removed to a point approximately 3 cm anterior to the rectal gland. Glands were perfused in situ under constant pressure conditions via a cannula inserted into the single rectal gland artery (modified posterior mesenteric artery). Perfusion flow through the gland was determined by cannulating the rectal gland vein that exits from the posterior end of the gland and immediately reflects anteriorly along the dorsal surface of the gut. Kent & Olson ( 1978) have shown that flow in this vein represents that which has perfused the secretory parenchyma of the gland, as opposed to various non-secretory shunt pathways that generally drain small diffuse vessels in the wall of the intestine. For the measurement of flow, the efferent venous cannula was led to a direct reading rotameter-type flowmeter (Gil-mont). Occasionally an additional flowmeter was inserted into the afferent line, enabling a comparison to be made between the total afferent flow and efferent flow via the rectal gland vein. By subtraction, it was therefore possible to calculate the proportion of the total perfusion flow exiting via the non-secretory vascular shunts.

Afferent perfusion pressure was maintained at 17 mmHg, approximating dorsal aorta blood pressures in this species (Butler & Taylor, 1975), by means of a pump/ overflow system supplied from a reservoir of saline (Table 1) gassed with 0·2% carbon dioxide in oxygen. Efferent venous pressure was maintained at zero with respect to the gland. All experiments were carried out in a constant temperature room at 11 °C. Drugs were added to the saline reservoir in amounts appropriate to give the final desired concentration in the perfusate.

In the preliminary work carried out at MDIBL, specimens of Squalus weighing 3–7 kg were caught in Frenchman’s Bay, Maine, and kept in live cars until use. The technique used was essentially the same as that described above for Scyliorhinus, except that the glands were completely removed from the fish and perfused, in isolation, at a pressure of 20 mmHg with saline gassed with 1 % carbon dioxide in oxygen. The preparation was maintained at a temperature of 13–15 °C by placing it on a thin Perspex plate which formed the roof of a water-cooled chamber. In addition, it is possible to measure secretion rate in the very much larger glands of Squalus (approximately l·0–2·7g wet weight as compared with 0·06–0·12g wet weight in Scyliorhinus). The secretory ducts of isolated perfused Squalus glands were cannulated via their opening on the mucosal surface of the posterior intestine using polyethylene tubing. This cannula was led to a second saline-filled direct reading rotameter flowmeter (Gilmont) for the determination of secretory flow rate.

Drugs used were noradrenaline bitartrate (L-arterenol bitartrate), isoprenaline bitartrate (isoproterenol bitartrate), adenosine and dibutyryl cyclic AMP, (all from Sigma), phentolamine mesylate (CIBA), propranolol hydrochloride (Inderal, ICI), theophylline (BDH), vasoactive intestinal peptide (synthetic porcine, Peninsular Laboratories), and furosemide (generous gift from Hoescht Pharmaceuticals).

Under the conditions used in this study, efferent perfusion flow, as measured in syliorhinus glands, was 8·13 ±0·41 mlg^-1 min^-1 (mean ± s.e., N=25). This constituted 64·6 ± 1·7% of the total afferent flow (mean ± s.e., N= 5), indication that, under control conditions, approximately one-third of the total afferent flow to the gland passed through non-secretory vascular shunts. Despite a very much lower average efferent perfusion flow in isolated Squalus glands of 1·72 ± 0·14 ml g^-1 min^-1 (mean ± s.e., N = 17), this again accounted for approximately two-thirds (70·0 ± 2·1 %, N = 17) of the total afferent flow, as in Scyliorhinus. The weight of the gland in Scyliorhinus is approximately 0·1 g per kg fish, whilst that in the Squalus used was approximately. 0·4–0·5 g per kg fish. When these values are taken into account, the rectal gland flow per weight offish is very similar in the two species (i.e. approximately 0·8 ml kg^-1 min-1 ). The precise significance of this, if any, is not clear, but the similarity, despite the pronounced differences in the weights of the specimens used and the relative sizes of the glands between the species, is certainly of some interest. It is also apparent that this rate of flow is considerably higher than the values of blood flow to the gland in vivo reported by Kent et al. (1973) and Solomon et al. (1980). The fact that it is similar to the maximum value of blood flow calculated from the data of Burger (1962) of 0·657 ml kg^-1 min^-1 (see above) suggests that the isolated gland perfused in vitro may be in a maximally vasodilated state.

Fig. 1 shows the effect of various concentrations of noradrenaline on efferent perfusion flow through the gland of Scyliorhinus. At concentrations above 10−8M, this catecholamine produces a pronounced reduction in perfusion flow indicating an induced increase in vascular resistance within the gland. Statistical analysis of the concentration-response curve indicates a value for the concentration of noradrenaline lving a half-maximal response (ED50) of 1·3 × 10⁻⁷M. Unlike noradrenaline, the specific β-adrenergic agonist isoprenaline, even at concentrations as high as 10⁻⁶ M, was totally without effect on the perfusion flow (Fig. 2). The a-adrenergic antagonist phentolamine, at a concentration of 10⁻⁶ completely reverses the reduction in perfusion flow induced by noradrenaline (Fig. 2), whilst the β-adrenergic antagonist propranolol was without effect (not shown). The data presented in Table 2 show that noradrenaline specifically reduces that proportion of the flow passing through the secretory parenchyma of the gland itself and that the flow through the non-secretory shunts is completely unaffected by the catecholamine.

Normal resting levels of catecholamines in the plasma of Scyliorhinus have been measured by Butler, Taylor, Capra & Davison (1978), who found mean values of 32·1 pmol ml^-1 (3·2 × 10⁻⁸M) for noradrenaline and 25·6pmol ml^-1 (2·56 × 10⁻⁸M) for adrenaline. It can be seen, therefore, that perfusion flow through the rectal gland is sensitive to noradrenaline at concentrations within the physiological range, and this would suggest that blood supply to the gland in vivo may well be restricted under the influence of circulating levels of catecholamines. In this context, it may be of some significance that perfusion flow in the maximally constricted isolated gland is some 5 % of the original flow. For Squalus, this would represent a flow of 86μlg^-1 min^-1 or 35–45 μl kg fish^-1 min^-1, a value very close to those recorded in the in vivo studies of Kent et al. (1973) and Solomon et al. (1980). This suggests that the fish used by these workers may well have been experiencing high levels of circulating catecholamines, probably produced as a result of experimental stress following can nulation or pithing. Various stresses are known to increase plasma catecholamines in other fish species (Nakano & Tomlinson, 1967), and Butler et al. (1978) showed that a hypoxic stress induced greater than ten-fold increases in noradrenaline and adrenaline concentrations in the plasma of Scyliorhinus. It can be seen from Fig. 1 that such concentrations would produce an almost maximum vasoconstriction within the rectal gland.

As previously reported for Squalus (Silva et al. 1980; Shuttleworth & Thompson, 1981), addition of cyclic AMP (cAMP) and theophylline, at concentrations known to produce a pronounced stimulation of secretion, failed to alter perfusion flow in the isolated glands of Scyliorhinus. However, the addition of 5 × 10⁻⁵M-CAMP and 2·5 ×10⁻⁴M-theophylline to glands perfused with noradrenaline reverses the vasoconstriction induced by the catecholamine (Fig. 3). Similarly, the addition of 10⁻⁶M-adenosine, which also produces a large increase in secretion in the isolated glands of Squalus, reverses the vasoconstriction seen in glands perfused with noradrenaline (Fig. 3). In both cases, the reversal of the noradrenaline-induced vasoconstriction is both rapid and virtually complete at the concentrations used and occurs despite the continued presence of the catecholamine in the perfusate. As with the results of noradrenaline alone, the addition of cAMP + theophylline had no effect on the perfusion flow through the non-secretory shunt pathways.

Vasoactive intestinal peptide (VIP) is also a potent secretory agent in the isolated glands of Squalus (Stoff et al. 1979), and in this species low concentrations (2 × 10⁻⁸M) of the hormone produced similar effects to those found with cAMP + theophylline and adenosine (Fig. 4). However, it was found that VIP was completely without effect, under identical circumstances, in the isolated gland of Scyliorhinus, even at concentrations as high as 10⁻⁶ M. The reason for this discrepancy is not clear, but it may be of some significance that VIP also fails to produce any increases in ouabain binding and oxygen consumption in slices of Scyliorhinus glands (T. J. Shuttleworth, unpublished data). Both of these parameters are associated with increases in secretory activity induced by cAMP + theophylline (Shuttleworth & Thompson, 1980a), and the failure of VIP to affect the rectal gland in Scyliorhinus compared to the clear effects at low concentrations in Squalus appears to represent a pronounced and, as yet, unexplained species difference.

Fig. 5 shows full concentration-response curves for noradrenaline in the presence of either cAMP + theophylline or adenosine in the isolated glands of Scyliorhinus. In the presence of the secretory agents, the concentration-response curves are shifted to the right, producing values of the ED50 for noradrenaline of 1·3 × 10⁻⁶M in the presence of adenosine, and 7·8 × 10⁻⁶ M in the presence of cAMP + theophylline. In addition, it can be seen that, with the secretory agents, the maximum effect of noradrenaline, even when used in concentrations as high as 10⁻⁴M, is only approximately 50 % of that seen in the absence of the secretory agents. The β-adrener-gic agonist isoprenaline does not elicit any reversal of the noradrenaline-induced vasoconstriction (Fig. 6) and, furthermore, the normal response to adenosine is not blocked by the β-adrenergic antagonist propranolol (Fig. 6). This indicates that the effective vasodilations induced by the secretory agents do not involve any stimulation of β-adrenergic receptors which, in fact, do not appear to be present in the vasculature of the rectal gland.

It is known that the stimulation of secretion is associated with large increases in oxygen consumption by the gland cells (Shuttleworth & Thompson, 1980b ; Silva et al. 1980). The possibility exists, therefore, that the observed increases in perfusion flow induced by the addition of the adenosine and cAMP + theophylline to glands perfused with noradrenaline may simply be the result of a vasodilation caused by hypoxia subsequent to the stimulation of secretory activity. However, reduction of the P_O2 of the perfusate to 20 % of its original level, by gassing with air instead of 0·2 % carbon dioxide in oxygen, failed to have any effect on rate of perfusion flow in the presence of 5 × 10⁻⁷ M-noradrenaline. This hypothesis was further investigated by studying the effect of the diuretic furosemide on the vascular effects of the secretory agents, and the results are presented in Table 3. Furosemide, at a concentration known to inhibit secretory activity in the gland (Silva et al. 1977; Shuttleworth & Thompson, 1980a), had no effect on the vasoconstriction induced by noradrenaline. More significantly, the diuretic was completely without effect on the inhibition of the noradrenaline-induced reduction in perfusion flow observed in the presence of the secretory agents adenosine and cAMP + theophylline. Furosemide is believed to have its effect by blocking the coupled uptake of sodium and chloride ions across the basolateral membranes of the secretory cells. As such, furosemide inhibits not only secretion but also the increases in secretion-related parameters such as ouabain-sensitive oxygen consumption, ouabain binding, and sodium and chloride uptake into the secretory cells that are seen on the addition of cAMP + theophylline (Shuttleworth & Thompson, 1980a). The fact that the inhibition of secretion and the above secretion-related phenomena was completely without effect on the vasomotor effects of the secretory agents clearly suggests that the vascular and secretory actions these agents are entirely independent of each other.

Adenosine could be exerting its effects on the rectal gland vasculature either intracellularly or extracellularly via specific receptors associated with stimulatory or inhibitory actions on membrane-bound adenylate cyclase (Londos, Cooper & Wolff, 1980). It is known that low concentrations of methylxanthines, such as theophylline, are capable of blocking receptors specific for the ribose moiety of the adenosine molecule (Smellie, Davis, Daly & Wells, 1979). The results in Table 4 show that the observed action of adenosine in blocking the vasoconstriction induced by noradrenaline is itself inhibited by the simultaneous presence of 10⁻⁵ M theophylline. This suggests that ribose-specific stimulatory adenosine receptors (R, receptors) are present on the vascular elements in the gland and that they are responsible for the mediation of the vasomotor effects of the nucleoside. It has been suggested that similar receptors exist on the secretory cells of the rectal glands of Squalus (Erlij, Silva & Rubio, 1980; Forrest, Rieck & Murdaugh, 1980). According to the classification of Londos et al. (1980), such receptors produce their effects by activating adenylate cyclase and increasing intracellular levels of cAMP. This is supported by the similarity seen in the rectal gland between the vasomotor effects of adenosine and the addition of exogenous cAMP together with theophylline, which at a concentration of 2·5 × 10⁻⁴M would act as an inhibitor of phosphodiesterase activity (Smellie et al. 1979). It is known that β-adrenergic-mediated relaxation of smooth muscle frequently involves activation of adenylate cyclase and increases in intracellular cAMP levels (Bar, 1974). Despite the apparent absence of β-adrenergic receptors in the vasculature of the rectal gland, a vasodilatory (i.e. relaxing) mechanism involving cAMP is suggested which may, in this tissue, be activated by adenosine receptors and possibly by other receptors for physiological effectors that are, as yet, unidentified.

To summarize, in the elasmobranch rectal gland, low levels of catecholamines, acting via α-adrenergic receptors, produce a pronounced vasoconstriction reducing flow through the secretory parenchyma of the gland. As this effect occurs within the physiological range of noradrenaline concentrations in the plasma, it is likely that blood flow through the gland is normally restricted in vivo by circulating catecholamines. This vasoconstriction is, however, abolished in the presence of agents known to stimulate secretion by the gland, effectively shifting the normal noradrenaline concentration-response curve so that essentially all significant catecholamine effects are blocked at physiological concentrations. Clearly such action could be the basis of an increase in blood flow to the gland associated with the Simulation of secretion in vivo as observed by Solomon et al. (1980). The evidence suggests that the observed vasomotor activities of the secretory agents are independent of their action on the secretory cells and are mediated via direct effects on vascular smooth muscle in the gland, probably involving changes in intracellular cAMP levels. Nevertheless, it seems likely that the vascular effects of the secretory agents would form an important and integral part of the overall physiological control of secretion rate by the gland in vivo.

The technical assistance of Miss J. L. Thompson is gratefully acknowledged. This work was supported by grants from the Science and Engineering Research Council (GR/B67063) and the Royal Society to the author and N.I.H. Biomedical Research Support Grant 2507 RR05764‐07 to the Mount Desert Island Biological Laboratory.