Evolution of Control of Epithelial Transport in Vertebrates

Several papers have been presented in this volume concerning the cellular and biochemical mechanisms of water and solute movement across transporting epithelia. This area of research has been studied extensively in recent years and the progress made in this field is impressive. Our research in this area for the past several years has been conducted with a different approach. We have been more concerned with the integration of different physiological processes and their controlling mechanisms in the whole animal. There have been fewer studies in this area of research and, thus, progress has not been as rapid. In this paper, we shall concentrate on the evolution of neural and endocrine control of osmoregulation. Our recent studies suggest that while in mammals hormonal control of epithelial transport may be the main regulatory mechanism, in lower vertebrates (such as fishes and amphibians) neural control may play a more important role. A better understanding of the evolution of these regulatory mechanisms may shed some light on the involvement of neural control in mammalian epithelial transport. In addition to the evolution of these controlling mechanisms, we are interested in the relationship between the vascular and osmoregulatory effects of some hormones and neural transmitters. The evolution of control of osmoregulatory processes is often intimately related to the evolution of control of cardiovascular processes. In some cases, the vascular actions of these regulators may largely explain their effects on the osmoregulatory target organ. In other cases, the vascular actions may significantly modify the direct effects of these regulators on water and electrolyte balance.

Our discussion will concentrate mainly on the control of two physiological processes: renal function and the epithelial transport of water and electrolytes. During vertebrate evolution, the transition from an aquatic to a terrestrial environment required drastic changes in the physiological processes of osmoregulation and their control. In the aquatic environment, respiration is carried out across the extensive respiratory epithelium (i.e., the gills) which is in direct and constant contact with the external medium. Considerable exchanges of water and electrolytes take place across this large surface area. In the osmoregulatory processes of these aquatic animals, many control mechanisms have the gills as their target. In fresh water, the kidneys are also important in removing excess water that gains entry into the body. When the ancestral tetrapods moved onto land, they abandoned the branchial mode of respiration and thus many of the branchial osmoregulatory mechanisms became obsolete. If, indeed, the terrestrial tetrapods were derived from a freshwater ancestor, the renal functions would also have been extensively modified from a water excretory system to a water conserving system. The controlling mechanisms for epithelial and renal osmoregulation would then have to be modified drastically. In fact, when we compare fishes and amphibians, we find that certain endocrine organs involved in water and electrolyte regulation in fishes are absent in amphibians, and vice versa. In addition, the actions and functions of some hormones are very different in fishes and amphibians. The evolution of the controlling mechanisms involved in water and electrolyte homoeostasis during the water to land transition represents perhaps one of the most drastic and significant physiological changes during vertebrate evolution. Based on our own work on fishes and amphibians and studies reported in the literature, we have formulated certain hypotheses on the evolution of these controlling processes. Some of these have been described elsewhere (Pang, Uchiyama & Sawyer, 1982; Pang, Furspan & Sawyer, 1983; Sawyer, Uchiyama & Pang, 1982). In the present paper, we shall discuss some new ideas and expand the existing hypotheses with some of our more recent findings.

Many hormones are known to exert extensive influence on the functions of mam-malian kidneys. The principal ones include antidiuretic hormone (ADH), aldosterone reiotensin and parathyroid hormone. Many other endocrine systems have been implicated. Most of these systems have not been studied much in the lower vertebrates. Thus, the evolutionary significance of these systems cannot be evaluated. The neuro hypophyseal hormone, however, has been studied extensively in most vertebrates. In mammals, ADH is most important in controlling urine volume (see review by Sawyer, 1974). Lack of ADH in man and experimental mammals, as a result of injury to the hypothalamus or a genetic defect, leads to a syndrome known as diabetes insipidus. The affected organism will lose control of the urine volume and will produce a large volume of dilute urine. There is, therefore, an absolute need for ADH in mammalian urine volume regulation. In lower vertebrates, arginine vasotocin (AVT) is the anti-diuretic hormone. Exogenous AVT has been shown to produce antidiuresis in almost all tetrapods thus far studied (see review by Sawyer & Pang, 1979). In chickens, hypothalamic lesions (Koike & Lepkovsky, 1967) and neurohypophysectomy (Shirley & Nalbandov, 1956; Ralph, 1960) produced polydipsia and diuresis which are similar to the symptoms that occur in diabetes insipidus. Similar observations were made in neurohypophysectomized ducks (Wright, Phillips, Peaker & Deaker, 1967; Bradley, Holmes & Wright, 1971). Hypothalamic lesions in the lizard, Amphibolurus omatus, prevented the antidiuretic response normally seen when the animals were challenged with hypertonic saline (Bradshaw, 1975). However, in amphibians, hypothalamic lesions did not produce diabetes insipidus. In Bufo bufo, transection of the preoptic-neurohypophyseal tract did not affect the antidiuretic response to dehydration (Jprgensen, Wingstrand & Rosenkilde, 1956; Jørgensen, Rosenkilde & Wingstrand, 1969). Our recent studies on the bullfrog showed that hypothalamic lesions did not impair the ability of the animals to reduce urine production drastically after dehydra tion as compared with operated controls (Gallardo, Pang & Sawyer, 1980). However, some studies have suggested that AVT might play a modulatory role in the antidiuretic response to dehydration or hypertonic saline injection. Hypothalamic lesions in Bufo marinus attenuated but did not abolish the antidiuretic response to dehydration or hypertonic saline injection (Middler, Kleeman & Edwards, 1967; Shoemaker & Waring, 1968; Bakker & Bradshaw, 1977). These data taken together suggest that in amphibians there is no absolute dependence on AVT for the control of urine volume. One can perhaps postulate an evolutionary trend in the physiological action of AVT in urine volume regulation. The ancestral amphibians were the first vertebrate group venturing onto land. Their most immediate problem, apart from respiration, would be water conservation. The ability to reduce water excretion during dehydration is essential for survival. It is possible that the use of the neurohypophyseal hormone, AVT, was one of the adaptations to terrestrial life. This new antidiuretic action of AVT was perhaps not fully developed in the amphibians, most of which still required access to water. However, reptiles and birds became more fully adapted to the dehydrating terrestrial environment and their ability to utilize AVT to control water excretion was perfected. As a result, they became more dependent on this endocrine system for the control of renal functions. In the latter part of this paper we shall give a fuller description of how this was accomplished.

In our studies of the circulating levels of AVT, we found that in bullfrogs (Rana catesbeiana), as in mammals, dehydration or reduction of blood volume, by haemorage or immersion in hyperosmotic mannitol solution, increased the circulating level of AVT (Pang, 1977). With the rise in plasma AVT levels, the animals exhibit antidiuresis. However, in some animals which were dehydrated or infused with hypertonic saline, antidiuresis was observed without a concomitant increase in plasma levels of AVT. This further supports the idea that in bullfrogs the control of urine produc tion is not completely dependent on the antidiuretic hormone AVT. Although the AVT-antidiuresis system may be incomplete, these amphibians are capable of controlling urine output according to their physiological needs. What then are the controlling mechanisms? We have looked into the possible role of the renin angiotensinsystem. In bullfrogs and Chilean toads (Calyptocephalellacaudiverbera), human angiotensin II (ANG II) produced diuresis and natriuresis at pressor doses and antidiuresis and antinatriuresis at nonpressor doses (Galli-Gallardo & Pang, 1978; Fruchter, Yang & Pang, 1980). The results from other laboratories were less consistent. Tubular diuresis and natriuresis were produced by pressor doses of ANG II in Xenopus laevis (Henderson & Edwards, 1969). However, antidiuresis and antinatriuresis were reported with pressor doses of ANG II in Bufo paracnemis (Coviello, 1969). Moreover, Sokabe et al. (1972) reported that plasma renin activity was low in dehydrated bullfrogs. It is therefore unlikely that the renin angiotensin system plays a significant role in the control of urine production during dehydration in the bullfrog. The other hormone which may play a role in the control of renal function is aldosterone. In amphibians, aldosterone is a circulating hormone and its circulating levels can be lowered by increasing blood sodium levels in Bufo marinus (Garland & Henderson, 1975). However, the effect of aldosterone on amphibian renal function has not been clearly demonstrated. It has been shown to have little (Heney & Stiffler, 1983) or no effect (Garland, Henderson & Brown, 1975).

While searching for the endocrine system responsible for the antidiuresis which is observed in bullfrogs with hypothalamic lesions, we discovered that the nervous system may play a significant role in the control of bullfrog renal function (Pang, 1977; Gallardo et al. 1980). When conscious intact cannulated bullfrogs were dehydrated, the urine volume decreased drastically. A bolus injection of phenoxy - benzamine produced diuresis despite a consistent decrease in systemic blood pressure. The urine volume produced after the administration of this a-adrenergic blocker was similar to the level before dehydration. This gave us the first indication that the a-adrenergic system might be involved in the control of renal function in the bullfrog. Since phenoxybenzamine may have other pharmacological actions, the diuresis we observed may be unrelated to the effect of the a-adrenergic system on the kidney. To show the direct involvement of the nervous system, bullfrog kidneys were perfused in situ with amphibian Ringer’s solution. The perfusion pressure was maintained constant during the experiment. Destruction of the nervous system by pithing in these perfused bullfrogs produced diuresis. Infusion of phenoxybenzamine into perfused kidneys also resulted in diuresis. Norepinephrine elicited a reduction in urine production of the perfused kidneys of the pithed bullfrogs, but not in the kidneys perfused with phenoxybenzamine. These kidneys were viable and responsive to drugs since AVT could produce antidiuresis in them. These data clearly demonstrate the direct antidiuretic action of the α-adrenergic system in the bullfrog kidney.

How does the a-adrenergic system produce antidiuresis? Is this antidiuresis a glomerular vascular effect or a tubular transport effect? From the initial perfusing eddies, we observed that pithing and phenoxybenzamine both produced diuresis and a simultaneous increase in perfusion rate even though the perfusion pressure was held constant. In the pithed bullfrog, norepinephrine produced antidiuresis and a decrease in perfusion rate in kidneys perfused under constant pressure. These data suggest that the a-adrenergic system also has an effect on the glomerular filtration but not necessarily on the tubular reabsorption of fluid. To study this, bullfrog kidneys were perfused in situ. Instead of perfusing only through the arterial supply, the renal portal veins were also perfused separately. In a normal, intact bullfrog, the kidneys receive blood supply from the dorsal aorta and the left and right renal portal veins from the posterior part of the body. The arterial and venous pressures are such that the arterial blood will supply the glomeruli and tubules while the venous blood will irrigate the renal tubules only. This can be achieved in the perfused kidneys by adjusting the perfusion pressures. If the venous perfusion fluid does not reach the glomeruli, inulin put into the venous perfusate, but not into the arterial perfusate, will not appear in the urine. Drugs put into the venous perfusate will then produce effects only in the tubules and not in the glomeruli. Since the two kidneys receive separate venous perfusion, one kidney will serve as the control while drugs can be tested on the other kidney. By adding creatinine to the arterial perfusate which supplies both kidneys from the dorsal aorta, the glomerular filtration rates of the two kidneys can be estimated from the urine collected from the two separate ureters. We used this type of perfusion system in our investigation of the tubular effects of phenoxybenzamine. In all the experiments, inulin was present in the venous perfusate and the perfusion pressures were adjusted so that adequate urine formation could be seen while the urine contained no inulin. Phenol red was introduced into the venous perfusate and its appearance in the urine indicated that the tubules were viable and transporting. Phenoxybenzamine was then introduced into the venous perfusate of only one kidney. The urine volume and glomerular filtration rate (GFR) were determined for the two kidneys separately. Phenoxybenzamine produced diuresis in the treated kidney when compared to the control kidney. No significant persistent changes in GFR could be observed in relation to the diuresis. This indicates that the α-adrenergic system has a direct tubular antidiuretic effect. All these data have been described in Gallardo et al. (1980). It is possible that the tubular diuresis produced by the venous perfusion of phenoxybenzamine is related not to the a-adrenergic system, but to some obscure pharmacological property of phenoxybenzamine. To test this, the experiment was repeated with pithed animals. The tubular diuretic action of phenoxybenzamine could no longer be demonstrated (R. Gallardo & P. K. T. Pang, unpublished data).

We have also tried to show morphologically the presence of innervation of the preglomerular arterioles and tubules. At the light microscopic level, nerve fibres were seen terminating on the preglomerular vessels and renal tubules. At the electron microscopic level, tubular innervation was also observed (Panget al. 1982; R. Gallar-do, J. A. Yee & P. K. T. Pang, unpublished data). All these studies taken together support our hypothesis that the α-adrenergic system is involved in glomerular and tubular antidiuresis in the bullfrog kidney and that these mechanisms are important in the normal control of renal functions in the bullfrog. The neurohypophyseal hor mone (AVT) has only a secondary role. In reviewing the literature, we were unable B find similar work conducted with fishes, reptile or birds. In an extensive review on the functions of the renal nerves (DiBona, 1982), our work on the bullfrog was the only study in lower vertebrates cited.

Recently, we have tried to demonstrate the participation of the nervous system in the control of fish renal function. The trunks of rainbow trout (Salmo gairdneri) were perfused under constant pressure with a trout Ringer’s solution. Destruction of the spinal cord produced an increase in urine production while norepinephrine elicited a decrease in urine volume (P. B. Furspan & P. K. T. Pang, unpublished data). Unfortunately, the anatomy of the fish kidney is such that it is not easy to carry out the same type of studies that we have done in the bullfrog. Nevertheless, further experiments are planned to elucidate the glomerular and tubular action of the nervous system in the control of trout renal function.

In mammals, the dependency of renal functions on ADH is clearly established. The role of the nervous system in the normal physiology of the mammalian kidneys has been the subject of controversy since the time of Homer Smith (Smith, 1937; Schrier, 1974). Recent studies demonstrated the presence of adrenergic innervation of tubular cells (Muller & Barajas, 1972; Barajas & Muller, 1973). Extensive work in recent years also showed that tubular sodium reabsorption was very much influenced by the adrenergic nervous system. Several reviews in the past five years have summarized the findings supporting such a conclusion (DiBona, 1978, 1981; Gill, 1979; Gottschalk, 1979). While the adrenergic influence on mammalian renal function has been demon-strated, the degree of dependency on such a controlling mechanism is still not clear. To quote the concluding remarks of the extensive review by DiBona (1982): ‘While an essential role for the renal nerves can be demonstrated in this and other circum-stances of modest physiological stress, further experimental work is required to define more precisely the functional role of the renal nerves in normal homeostatic regulation’.

Our studies with the bullfrog suggest that the dependence on AVT is minimal. Other investigations reported in the literature also support this hypothesis. However, the nervous control of renal function seems to play a very important role. It is possible that during vertebrate evolution the nervous system initially exerted the main control over renal function with the neurohypophyseal hormone acting as a secondary control. As terrestrial adaptation in higher tetrapods became more complete and the action of the neurohypophyseal hormone became more developed, the role of the nervous system became less and less and the dependence on the neurohypophyseal hormone became greater and greater. This hypothesis is summarized in Fig. 1. While the situation in mammals, as indicated in Fig. IB is well substantiated by published data, the situation in amphibians, as indicated in Fig. 1A, is supported by our own data on the bullfrog. Much more work in many other species of amphibians is required. In addition, the involvement of the nervous system in renal function remains to be demonstrated in fishes, reptiles and birds. The establishment of an evolutionary trend is important not only to our understanding of the evolution of the control of osmore-gulation, but also to our appreciation of the neural control of renal function in mammals including humans.

The above discussion elucidates the integration of the neural and endocrine control the osmoregulatory role of the kidneys. Endocrine control of transport in other vertebrate epithelia has been studied extensively. Is there neural control of transport in such epithelia? Do we know enough to assess the relative significance of the two controlling systems? Our information in this area is quite limited. A brief discussion of the neural influence on branchial, bladder and skin transport will be given here.

As mentioned in the Introduction, fish gills represent an extensive respiratory surface in direct contact with the external medium. In fresh water, the osmotic gradient is such that there can be an extensive influx of water, and in sea water, the reverse is true. To maintain osmolality of body fluids, it is essential to control the water permeability of the gill epithelia. Hormones are known to affect water permeability of gills. An example of such a hormone is prolactin, which reduces water permeability. The hormonal control of gill transport has been reviewed in this sym-posium (Foskett, Bern, Machen & Conner, 1983). Here we shall present evidence supporting the neural control of gill transport. Pic, Mayer-Gostan & Maetz (1974) studied the water fluxes of mullet (Mugil capita) in sea water. Whole animals were injected with tritiated water and the effect of epinephrine on water efflux rate was determined. Within minutes, epinephrine increased water efflux. These investigators suggested that epinephrine caused changes in gill water permeability. Haywood, Isaia & Maetz (1977) perfused isolated trout heads and measured unidirectional water flux. They reported that epinephrine, at a concentration of 10⁻⁶ M, more than doubled the water influx rate. However, epinephrine also caused changes in gill haemodynamics. A change in water flux could be the result of changes in blood flow through the gill filaments. They speculated that the change in flux rate was a result of permeability changes, since epinephrine produced an increase in water influx of unperfused gills. In a similar system, Isaia, Girard & Payan (1978) also observed an increase in water influx when the trout head was perfused with 10⁻⁵M epinephrine. These authors differentiated the haemodynamic and permeability changes and concluded that part of the influx increase was due to actual permeability changes. In perfused heads of toadfish (Opsanus beta) in sea water, Oduleye & Evans (1982) demonstrated an increase in water efflux when 10⁻⁶M epinephrine or 4×10⁻⁷M carbachol was added to the perfusion medium. They also concluded that this change in water efflux was due in part to an increase in water permeability in the gill epithelium. These data showed that catecholamines could be important in the control of gill water movement.

In a review on the hormonal control of fish gill sodium transport, Evans (1978) concluded that epinephrine was important in sodium transport in freshwater and seawater fish. Payan, Matty & Maetz (1975) reported that epinephrine enhanced sodium uptake in the perfused head of freshwater trout. In a later paper, Girard & Payan (1977) observed that this epinephrine effect was specific for sodium but not chloride uptake. In the perfused head of seawater trout, however, epinephrine decreased active sodium efflux (Girard, 1976). The effects of catecholamines on electrolyte transport in the killifish (Fundulus heteroclitus) have also been reported for the opercular membrane which contains transport or chloride cells similar to gills. Degnan & Zadunaisky (1979) observed stimulation of chloride secretion by β-adrenergic agonists and inhibition of chloride secretion by α-adrenergic agonists in the killifish opercular epithelial preparation. In collaboration with investigators at the Zoological Station in Villefranche-sur-Mer, France, we have studied calcium fluxes in fishes. Since there was very little known in this area of research, we started by looking at the uptake of ⁴⁷Ca in various species of fish. We found that several teleost species had similar calcium uptake rates. Our data suggested that fish in calcium deficient environments can transport calcium from the external medium into the internal medium of the fish. Since the intestinal entry of calcium during the flux study period was negligible, we concluded that the fish had active calcium uptake across the gill. We further suggested that in calcium-deficient media the gill permeability to calcium was increased to facilitate this uptake (Pang, Griffith, Maetz & Pic, 1980b). A subsequent more detailed study essentially confirmed the initial studies. Killifish were adapted to low and high calcium fresh water and the branchial calcium uptake was shown to be the same when tested in each adaptation medium (N. Mayer-Gostan & P. K. T. Pang, unpublished data). There are two important aspects of these findings. Firstly, the ability of fish to transport calcium from an external medium to the internal medium across a concentration gradient indicates that, like sodium, environmental calcium can act as a reservoir for regulation of this action within the body. This is important since there has been controversy as to whether bone is an important reservoir for calcium regulation in fishes (Pang, Kenny & Oguro, 1980c). Secondly, this branchial uptake rate can be modified according to the physiological needs of the animal. The hormones involved in fish calcium regulation have been studied extensively (see review by Pang, 1973; Panget al. 1980b). The action of one of these endocrine systems, the Stannius corpuscles, on gill calcium transport has been studied in several laboratories. Isolated eel gill arches were perfused with Ringer’s solution and active uptake of calcium was observed. This uptake of calcium was decreased by the active principle from the Stannius corpuscles (Ma & Copp, 1982; So & Fenwick, 1982). The exact chemical nature of this principle has not yet been determined. Other investigators have reported the action of other hormones such as calcitonin on calcium uptake in the isolated perfused gill (Milhaud, Rankin, Both denson, 1977). In one of our recent studies, we used the perfused trout head model to study the uptake of ⁴⁷Ca by the gills. The gill morphology of the trout used in these studies carried out in Villefranche has been well investigated. The chloride cells were found mainly on the primary gill lamellae at the bases of the secondary lamellae (Sadet, 1980). In the perfused trout head, the perfusion fluid can be collected separately from the arterial and venous outflow. The arterial outflow comes mainly from the irrigation of the secondary lamellae, and the venous outflow from that of the primary lamellae. The calcium uptake across the primary lamellae was forty or more times higher than that of the secondary lamellae. This suggested to us that the active uptake of calcium in the freshwater trout was probably carried out by the chloride cells. In seawater fish, the chloride cells have been shown to excrete sodium and chloride. However, the function of the chloride cells of freshwater fish has not been well documented. In the same study we investigated the effect of epinephrine on gill calcium uptake. Epinephrine, at doses of 10⁻⁶M, increased branchial calcium uptake more than ten-fold, bringing the flux rate to a level similar to that observed in whole animal studies. This increase in uptake can be blocked by propranolol but not phento-lamine. We concluded that the β-adrenergic mechanism was important in calcium transport in the trout (Payan, Mayer-Gostan & Pang, 1981).

In all the fish studies described above, the effects of exogenous catecholamines were studied. Can this be considered to be neural control or control by circulating catecholamines released perhaps from the chromaffin tissue in the head kidneys of the fish? Pettersson & Nilsson (1979) studied the effect of sympathetic nerve stimulation on gill vascular resistance in the cod (Gadus morhua). As will be described later in this paper, exogenous catecholamines produce transport changes and concomitant vascular resistance changes. Pettersson & Nilsson observed that sympathetic chain stimulation elicited vascular resistance changes similar to those observed with exogenous epinephrine. These changes could be demonstrated to have α and β-adrenergic components. This would suggest that neural input into gills is significant in gill physiology. Whether gill transport can be affected by direct innervation has not been demonstrated. The levels of exogenous catecholamines used in most studies were high and it is doubtful whether blood catecholamines from the chromaffin tissues could reach that level. Much work is needed in this area of research. Morphological evidence of innervation of the transport tissue is one approach. Pharmacological and physiological demonstration of neural control of transport is another approach. The interaction between neural and endocrine control of transport may prove interesting.

The control of transport mechanisms in amphibian urinary bladder has been studied extensively. Several papers in this symposium have concentrated on this topic. The hormonal control is well understood, but the neural control has not been established. Neurotransmitters have been shown to alter the transport process. For example, cholinergic agents have been reported to inhibit sodium transport across the isolated toad bladder (Sahib, Schwartz & Handler, 1978; Wiesmann, Sinha, Yates & Klahr, 1978) and turtle bladder (Schilb, 1969). In addition, toad bladder water transport was mildly stimulated by carbachol while the arginine vasopressin stimulated water transport was inhibited by the same agent (Arruda & Sabatini, 1980). Once again, an important approach to this problem is the morphological Notification of direct innervation of the transport epithelial cells. In addition, in vivo and in situ studies will be needed to demonstrate the neural control of transport in a amphibian urinary bladder. The amphibian skin is another important transport surface. Epinephrine or isoproterenol have been shown to stimulate sodium transport in frog skin (Arczynska, Girardier & de Sousa, 1976; Tomlinson & Wood, 1978), while the α-adrenergic mechanism inhibited the epithelial sodium pump (Tomlinson & Wood, 1976). Kuz’min (1975) demonstrated that the epinephrine-stimulated sodium uptake mechanism was different from that of A VP. In the presence of amelioride, isoproterenol induced short-circuit current changes in frog skin, and it was suggested that there was a ouabain-sensitive sodium and chloride secretory mechanism in skin glands of frogs (Thompson & Mills, 1981). Isoproterenol also increased water flow in toad skin (Brown, Grosso & de Sousa, 1980). It has been suggested that chloride transfer in Rana esculenta was localized in a special cell type rich in mitochondria (Kristensen, 1981). Are these cells innervated? While our knowledge of the neural control of transport is scanty and incomplete, data on hormonal control are plentiful. To understand fully the evolution of the controlling systems and the integration of the actions of these systems, we must have a better understanding of the neural control of transport in lower vertebrates.

The neurohypophyseal antidiuretic hormone (AVT or ADH) can elicit antidiuresis in almost all tetrapods thus far studied. Although the physiological role of this hor-mone may be rather modest in some poikilotherms such as the amphibians, the molecule nevertheless possesses such a property. When AVT is administered to fish, the common response is diuresis. This has been shown in goldfish (Maetz, Bourguet, Lahlou & Hourdry, 1964), eels (Chester Jones, Henderson, Wales & Garland, 1971) and lungfishes (Sawyer, 1970; Sawyer, Blair-West, Simpson & Sawyer, 1976; Sawyer & Pang, 1980). In a few cases, AVT was reported to produce antidiuresis at low doses (Henderson & Wales, 1974). How is it possible for the same molecule to produce opposite effects in the same organ in fishes and tetrapods, respectively? We have recently advanced a hypothesis to explain this phenomenon (Pang et al. 1982). In this hypothesis, we suggested that AVT first appeared as a vascular hormone in fishes and that the diuresis produced by AVT was purely a glomerular diuresis. We postulated that the vascular receptors to AVT were either more abundant or more sensitive in the peripheral vasculature as compared to those in the preglomerular circulation. Exogenous AVT would then produce an increase in systemic blood pressure sufficient to overcome the modest preglomerular vasoconstriction, resulting in an increase in urine volume. Two separate lines of evidence were provided from our studies on lungfishes, (i) Conscious free-swimming lungfish responded to injections of AVT with an increase in blood pressure and diuresis. KBIV 24, an analogue of AVT, is a pressor antagonist of AVT. It has no antagonistic action on the tubular action of AVT or ADH. If lungfish were pretreated with KBIV 24, both the pressor and diuretic actions of AVT were abolished, (ii) Lungfish trunks, perfused with Ringer’s solution under constant pressure, responded to AVT with antidiuresis instead of diuression cating the presence of preglomerular vasoconstricting receptors to AVT. When perfused at a constant rate, the perfusion pressure was increased with AVT and diuresis was also observed. This clearly indicates that the diuretic response to AVT in intact lungfish is related to the systemic pressor action, which overcame the modest preglomerular vasoconstriction produced by the same peptide. These unpublished data (P. K. T. Pang & M. Uchiyama) have been described in Sawyer et al. (1982) and Pang et al. (1983). Recently, similar observations were made in perfused trout trunk (P. B. Furspan & P. K. T. Pang, unpublished data). Under constant perfusion pressure, AVT produced dose-related antidiuresis in isolated trout trunk. This renal response was abolished by pretreatment with KBIV 24. This again indicated the presence of preglomerular vasoconstricting receptors for AVT and that the stimulation of this vascular receptor could result in antidiuresis under constant perfusion pressure. When perfused at a constant perfusion rate, AVT produced increases in perfusion pressure and urine volume. These changes could be inhibited by KBIV 24 (P. B. Furspan & P. K. T. Pang, unpublished data). This substantiated the hypothesis that the diuretic action of AVT in intact fishes is a glomerular diuresis as a result of the systemic pressor action.

In the second part of our hypothesis on the evolution of the renal action of AVT, we postulated that when the ancestral tetrapod moved onto land, there appeared an increase in number and/or sensitivity of AVT receptors in the preglomerular circulation as compared to the systemic circulation. AVT would then produce glomerular antidiuresis instead of diuresis. This situation is seen in the mudpuppy (Necturus maculosus), a present-day form. In this aquatic neotenic urodele, AVT produced dose-related antidiuresis and pressor responses. However, if the animals were pretreated with KBIV 24, both the pressor and the antidiuretic actions of AVT were no longer observed. In the perfused mudpuppy kidney in situ, AVT was shown to produce preglomerular vasoconstriction and antidiuresis. When the peptide was introduced into the venous perfusate as described in the bullfrog renal studies, no antidiuretic effect could be seen. All these data suggest that the antidiuresis was purely a preglomerular vascular effect (Pang et al. 1982). That the preglomerular vascular response to AVT is greater than the systemic response was shown in the experiments of Pang, Galli-Gallardo, Collie & Sawyer (1980a). In that study, the kidneys and hind limb of bullfrogs were perfused simultaneously in situ and the vascular resistance to AVT was determined. Data showed that the renal vasculature was more responsive to AVT than was the hind limb.

In the third part of the hypothesis, we postulated that as tetrapods became more adapted to the terrestrial environment, the antidiuretic response to AVT became stronger and that this was accomplished by establishing tubular receptors to AVT. The animals now showed both glomerular and tubular antidiuresis in response to AVT. This was demonstrated in in vivo and in situ studies with the bullfrog (Pang et al. 1983). AVT produced antidiuresis and an increase in blood pressure in conscious bullfrogs. Pretreatment with KBIV 24 abolished the pressor response, but only decreased the antidiuretic response to AVT. In perfused bullfrog kidneys in situ, KBIV 24 inhibited the preglomerular vasoconstriction and reduced, but not abolished, the antidiuretic action of AVT. The tubular antidiuretic effect of AVT was to directly demonstrated when AVT was delivered to the tubules through the renal portal perfusate. This type of response to AVT, a combination of preglomerular vascular and tubular antidiuresis, although clearly demonstrated in our direct bullfrog studies, has also been shown indirectly in reptiles and birds. When mammals became fully adapted to the terrestrial environment and the loop of Henle provided an efficient mechanism to concentrate urine, the tubular action of ADH became the main action. This hypothesis as a whole stresses the relationship between the vascular and osmoregulatory actions of the neurohypophyseal hormone during evolution (Fig. 2). When one looks closely at other osmoregulatory controlling mechanisms, other examples can also be seen. Although most of these studies are incomplete, they do demonstrate the vascular action of controlling factors in their osmoregulatory role. Some examples will be described briefly in the last part of this paper.

In the above discussion, we reviewed the literature and suggested that neuous control of branchial transport may play an important overall role in fish osmoregulation. In most studies, isolated heads or gills were perfused with epinephrine and transport rate changes were observed. However, in all these studies, epinephrine or other neurotransmitters also produced significant vascular changes which affected the pattern of blood flow in the gill filaments. These pattern changes could result in gill lamellar recruitment and/or redirection of blood to or away from the respiratory secondary gill lamellae. Extensive studies have been carried out in this area of research. Some of these studies will be described below.

As early as 1932, Keys & Bateman reported branchial vasodilation produced by epinephrine in the eel. Ostlunde & Fange (1962) observed vasodilation in perfused gills treated with epinephrine. Wood (1974) performed a detailed and critical investigation on the haemodynamics of the perfused trout head and reported that adrenaline decreased vascular resistance in the gills through a β-adrenergic mechanism. Using the perfused trout head as their experimental model, Payan & Girard (1977) analysed the pharmacological nature of the branchial vascular response to epinephrine. Their results suggested that epinephrine produced both α- and β-adrenergic responses in the gill, the a-adrenergic response being vasoconstricting at the arterio-venous anastomosis and the α-adrenergic being vasodilatory at the lamellar arteriole level. The α-adrenergic mechanism was predominant and the net result was vasodilation. Smith (1977) perfused trout gill arches and observed vasoconstriction with acetylcholine. Using ⁵¹Cr-tagged red blood cells, Holbert, Boland & Olson (1979) determined the blood flow of perfused catfish gills. They reported that epinephrine increased while acetylcholine decreased blood flow through the secondary gill lamellae. They suggested lamellar recruitment as the mechanism of action. Most investigations have been conducted with the trout. Some recent studies by Evans and his co-workers have extended these observations to some other marine species. Claiborne & Evans (1980) perfused the head of the marine teleost Myoxo-cephalus octodecimspinosus and reported a decrease in branchial vascular resistance with epinephrine. They analysed the α- and β-adrenergic components of the response and concluded that the β-adrenergic effect was vasoconstricting at the arterio-venous anastomosis and that the β-adrenergic effect was vasodilating at the level of the secondary lamellar arterioles. The perfused head of the marine toadfish (Opsanus beta) also responded to epinephrine in the same manner and to acetylcholine with vasoconstriction (Oduleye, Claiborne & Evans, 1982). All the above mentioned studies were carried out on isolated perfused systems. Booth (1979), using stained red blood cells, studied the branchial blood flow in unrestrained conscious trout. He demonstrated that epinephrine increased secondary lamellar perfusion while acetyl-choline reduced the perfusion.

It is clear that neurotransmitters are important in the control of blood flow in the gill. The changes they produce in gill haemodynamics very much affect the functional area of the exchange surface, which in turn can very much affect the total rate of transport across the gill. The above discussion indicated that these neurotransmitters could change the permeability of the transport epithelium directly and the net result could very well be a summation of the haemodynamic and the permeability effects. However, the question still remains as to whether these changes were produced by calculating catecholamines or direct innervation. The gills are innervated. It is very likely that the haemodynamic effects could be produced by direct neural actionsion in the blood vessels. When Pettersson & Nilsson (1979) stimulated the nerves supplying the gills, they observed vascular responses in cod gills similar to those reported for neurotransmitter perfusion. Whether the permeability changes could be related to direct innervation remains to be demonstrated.

Our brief discussion on the interaction between vascular and osmoregulatory effects of certain controlling mechanisms in kidneys and gills indicates the importance of such a relationship. Whether the osmoregulatory effects of hormones and neural elements on other transport epithelia are also affected by vascular actions of these factors has not been studied extensively. This area of research may be important for our understanding of the mechanism and evolution of these control processes.

The evolution of the controlling mechanisms in osmoregulation is a vast topic. Within the limitations of time and space here, it is impossible to give a detailed and exhaustive evaluation of the present status of the field. In this paper, I have tried to concentrate on two aspects: (i) the interaction between neural and endocrine control and (ii) the contribution of vascular actions of endocrine and neural controlling factors in osmoregulation. These two areas have not received much attention previously. However, they are potentially important in our understanding of the evolution of the control processes in osmoregulation. Throughout the paper, I have tried to point out areas where future investigations would be fruitful and the type of studies that should be done. In any review prepared at this time, it is impossible to cite all the literature. I would like to apologize for any omissions. They occur not because of a lack of importance of these papers but through my personal oversight.

Previously unreported studies were supported by research grants from the National Science Foundation PCM 81–05425 to P. K. T. P. and PCM 80–04076 to W. H. Sawyer. I would like to thank the publishers of Federation Proceedings for permission to reproduce Figs 1 and 2 from Pang, Uchiyama & Sawyer (1982).

Photographs taken at the Discussion Meeting held at Wakulla Springs in March 1983