Hypoxic Vasoconstriction and the Effects of Adrenaline on Gas Exchange Efficiency In Fish Gills

Acute ambient hypoxia is a stress situation calling for compensatory adjustments in exchange and transport of respiratory gases. Adrenergic activity will probably increase in hypoxic stress as in other types of stress.

It is known from perfusion experiments that deoxygenation of perfusion fluid causes gill vascular constriction (Ristori & Laurent, 1977). Fish gills typically have no arterio-arterial by-pass vessels (Laurent & Dunel, 1976; Vogel, Vogel & Kremers, 1973; Vogel, Vogel & Schlote, 1974; Vogel, Vogel & Pfautsch, 1976; Gannon, Campbell & Randall, 1973; Cooke & Campbell, 1980). A preliminary study of the cod, Gadus morhua, has confirmed that no secondary lamellar by-pass shunts are present (S. Nilsson & K. Pettersson, unpublished observations). Exceptions to this generality have been reported for the eel, Anguilla anguilla (Steen & Kruysse, 1964; Laurent & Dunel, 1976), and for the Channel catfish, Ictalurus punctatus (Boland & Olson, 1979). With no by-pass route past the secondary lamellae, hypoxic gill vasoconstriction might therefore be an inappropriate response to a call for increased exchange of O₂, unless the vasoconstriction diverts blood to larger and/or better ventilated areas of the gas exchange surfaces or, alternatively, exerts an influence on the blood to water diffusion barrier.

Adrenaline is known to increase arterial compared to venous outflow from perfused gills (Dunel & Laurent, 1977; Payan & Girard, 1977; Pettersson & Nilsson, 1979 a, b;Rowing, Taylor & Rankin, 1979; Claiborne & Evans, 1980). Adrenaline has also been shown to increase arterial in the eel (Steen & Kruysse,1964; Peyraud-Waitzenegger; 1979).

The objective of the present study was to elucidate, by perfusion techniques, how gill hypoxic vasoconstriction is mediated and whether the response is elicited by blood-borne hypoxia and/or ambient water-oxygen lack, i.e. to elucidate the sites of the gill vascular system that are influenced by hypoxia. Additionally we have tried to resolve the nature and location of an eventual effect of adrenaline on the gill vascular bed during hypoxia.

The Atlantic cod, Gadus morhua, was used in this study. The material comprised both sexes, weighing between 400 and 1600 g. The fish were kept in recirculated aerated sea water at 15 °C until used.

Preparing for an experiment a fish was stunned with a blow on the head, followed by an injection of c. 1000 i.u. of heparin into the caudal vein. Approximately 5 min later the head was removed from the body and a catheter (PE 90) was inserted into the bulbus arteriosus and a slow perfusion was started. We used a filtered saline, containing 1 g glucose I⁻¹ for the perfusion (House & Greene, 1965). Another catheter was placed in the dorsal aorta, while the carotid and coeliaco-mesenteric arteries were ligated. The right side inferior jugular vein was catheterized with PE 50 catheter. The left-side inferior jugular vein was ligated. The head was immersed in saline during the preparation.

Microfil injections through the right-side inferior jugular vein show the gills on both sides to be filled with the cast. Skeletal muscle vessels in ‘the tongue region’ were poorly filled, indicating that vascularization is poor (S. Nilsson & K. Pettersson, unpublished observations). We therefore consider the perfusion fluid collected in the inferior jugular vein to be predominantly branchial venous effluent.

Fig. 1 shows the experimental arrangement used. The gills were ventilated with sea water at a rate of about 700 ml min⁻¹, using an Eheim pump. of the ‘inspired’ water was adjusted by means of a Wosthoff model 301 gas mixing pump. The gills were perfused with saline using a peristaltic pump (Harvard apparatus 6). of the inflow perfusion fluid was adjusted to the desired level by means of Wosthoff gas mixing pumps.

Perfusion pressure in the ventral aorta (PVA) and the dorsal aorta (PDA) were recorded by Statham P23 pressure transducers connected to a Beckman 66A polygraph. Outflow from the dorsal aorta (Q̇) was monitored using a 1 mm cannulating probe and a Statham SP 2202 electromagnetic flowmeter. Outflow from the inferior jugular vein (Q̇), was recorded by means of a Grass photoelectric transducer (drop flow counter) connected to a Gould-Bush Biotach device. Both flow rates were monitored on a Beckman 66 A polygraph. O₂ tensions of both effluents were also continually recorded by two O₂-electrodes (Radiometer E 5046) mounted in flow through cuvettes and connected to Radiometer PHM 71, Mk 2 s, and displayed on a Hewlett-Packard 7132A two-channel potentiometric recorder.

The gills were perfused with a flow rate of about 18 ml kg⁻¹ min⁻¹, a flow within the range described to occur in vivo (Johansen, 1962; Jones et al. 1974; Pettersson & Nilsson, 1980), resulting in a perfusion pressure (PVA) similar to that described in vivo (Helgason & Nilsson, 1973; Jones et al. 1974; Wahlqvist & Nilsson, 1977; Pettersson & Nilsson, 1980). The perfusion pressure remained stable for the first 3 h of perfusion, or increased very slightly. No temporal variations in the response to hypoxia and/or adrenaline were observed.

In some fish head vascular perfusions described earlier (Payan & Girard, 1977; Claiborne & Evans, 1980), the cephalic circulation has not been separated from the venous drainage of the gills, making it impractical to perfuse against an efferent arterial resistance similar to physiological systemic pressure, since this would also expose the venous system to this pressure, and probably cause severe damage to the preparation (Smith, 1977). In our experiments the dorsal aorta outflow catheter was raised until PDA was at least 1 kPa. In the preparation used, it proved impossible to ligate the mandibular, hypobranchial and orbital arteries, originating from the efferent gill arteries of the 1st and 2nd gill arches, and many small arteries originating from the suprabranchial artery, resulting in a leakage from the perfusion circuit. If dorsal aorta pressure was kept at zero this leakage was restricted, but at 1 kPa it increased markedly. This unavoidable and (in the adrenaline experiments) variable leakage represents an obvious drawback of the preparation and the results are carefully interpreted with this deficiency in mind. All earlier perfusions of fish heads or preparations consisting of more than one gill arch have, however, been affected by the same problem. Johansen & Pettersson (1981) have presented evidence suggesting that leakage will occur mainly from the efferent arterial system. They compared gill tissue O₂ uptake from perfused ‘leak free’ single gill arches with head preparations like the one described here, and concluded that leakage occurs mainly from distal arterial vessels of little or no consequence in gill gas exchange and O₂ uptake.

The experiments were performed at 10 °C. The drug adrenaline bitartrate was obtained from Sigma. It was dissolved in saline to a final concentration of 10⁻⁶M. The experiments were performed in two series. In the first series (A) both and were air-saturated and showed if adrenaline caused any effects on when no oxygen tension gradient from irrigating water to perfusion fluid existed. This series served as control to series B, in which was decreased to c. 22 mmHg (Table 2). Wahlqvist & Nilsson (1980) have reported that plasma adrenaline concentration from severely stressed cod can exceed 3 ×10⁻⁷ M. Statistical treatment was performed according to the Wilcoxon matched-pair signed-ranks test.

In the experiments involving the effects of hypoxia on the gill vascular system, perfusions were started using air-saturated perfusion fluid and sea water for irrigation of the gills. After stable conditions were established, the of the perfusion fluid was instanteously decreased from c. 155 to 23 mmHg by shifting to another reservoir. In three experiments the of the irrigating sea water was also decreased (to c. 60 mm Hg), but always after initiation of ‘internal’ hypoxia. The results are summarized in Table 1, and a typical response is illustrated in Fig. 2. A small but significant decrease in dorsal aorta outflow occurred, but no change was apparent in venous outflow rate (Table 1). There was also a significant increase in branchial vascular resistance expressed as an increase in ventral aortic pressure (perfusion flow was kept constant). The successive increments in ventral aortic perfusion pressure with initiation of ‘internal’ hypoxia and secondly ‘external’ hypoxia (as was the case in all experiments performed), testify that the hypoxic vasoconstrictor response can be released by both types of hypoxic stimuli (Fig. 2).

Adrenaline (10⁻⁶M) did not significantly affect dorsal aorta when no oxygen tension gradient existed between ambient water and perfusion fluid (Table 2, Fig. 3 a) Outflow through the dorsal aorta increased, while the perfusion pressure markedly decreased (Fig. 3,a). In another series of experiments when adrenaline was administered while a gradient of about 135 mmHg (from water to perfusion fluid) was present, a marked increase in both dorsal aortic outflow rate and O₂ tension occurred. There was no change in venous outflow rate. Ventral aortic pressure fell more markedly wean in the first series (Table 2, Fig. 3b).

Variable results have been reported on the effect of hypoxia on gill vascular resistance. In intact fish it seems that overall branchial vascular resistance increases during acute exposure to hypoxic water as shown for trout, Salmo gairdneri (Holeton & Randall, 1967). Farrell, Daxboeck & Randall (1979) recorded no vasoconstriction in perfused gills from the lingcod, Ophiodon elongatus, when the perfusion fluid was deoxygenated (the degree of hypoxia is not reported). Ristori & Laurent (1977), however, reported that gills from the rainbow trout showed an increased vascular resistance in response to a change in perfusion fluid from aerated levels to about 20 mmHg in a perfused head preparation. They concluded that the major site of hypoxic vasoconstriction was located distal to the arterio-venous anastomoses (see Fig. 4), since the venous outflow increased while the arterial decreased.

They ascribe the decrease in dorsal aortic outflow to constriction of sphincters at the base of the efferent filamental arteries (EFA, Fig. 4) (Laurent & Dunel, 1976). We believe this explanation must be dismissed, since it would require that the arterial cephalic and/or systemic arterial ‘leaks’ would have to constrict in response to hypoxia as well, since in their constant pressure experiments both dorsal aortic and cephalic/branchial venous flow decreased. In our experimental series on the Atlantic cod no changes in the relative outflow levels occurred; a result suggesting that the vasoconstrictors occurs proximal to the arterio-venous anastomoses (Fig. 4). The slight decrease in dorsal aortic outflow recorded here could be explained by a hypoxic vasodilation of extrabranchial systemic arteries (including the ‘leaks’), which is the typical response to hypoxia of systemic arterioles in other vertebrates. The inferior jugular vein flow is not affected since it consists of almost pure branchial venous outflow.

The perfusion experiments involving a hypoxia stimulus from the external, ambient water side gave, in our experiments, a qualitatively similar response as ‘internal’ hypoxia. This result has two important consequences. Firstly, the vasoconstriction must in part occur distal to the afferent filamental and lamellar arteries, since these sites carjnot be reached by an ‘external’ hypoxic stimulus. Secondly, since the normal gill in an intact fish receives deoxygenated blood (often at O₂ tensions below those employed in others and the present perfusion experiments), a natural hypoxic stimulus should therefore influence vascular channels that are not ‘normally’ hypoxic. These must then be located distally within the secondary lamellae or, more probably, in the efferent lamellar arterioles. This site is normally perfused by oxygenated blood when the gills are functional in gas exchange. ‘External hypoxia’, which is the physiologically significant hypoxia for fish, can be rapidly screened by O₂-sensitive smooth muscle at this site. Vasoconstriction at such sites will, if it affects enough secondary lamellae, cause an increase in afferent lamellar pressure (Fig. 2). This response may in turn cause recruitment of unperfused lamellae (Booth, 1979), thereby increasing the available surface area for gas eschange, and perhaps alter the lamellar size and geometry for more effective diffusion exchange. Only in this way will the hypoxic constriction impart a compensatory advantage to a situation of external O₂ shortage.

The chief response to adrenaline is a decreased resistance to perfusion and has been noted by many others (Krawkow, 1913; Ostlund & Fange, 1962; Reite, 1969; Belaud, Peyraud-Waitzenegger & Peyraud, 1971; Rankin & Maetz, 1971; Wood, 1975; Payan & Girard, 1977; Smith, 1977; D’Amico Martel & Cech, 1978; Pettersson & Nilsson, 1979a, b, 1980; Claiborne & Evans, 1980; Nilsson & Pettersson, 1981). In our experiments dorsal aorta flow increased in all cases, while flow in the inferior jugular vein exhibited great variations. The leak problem associated with head perfusions (discussed above) probably influenced these results. In this way, α-adrenergic constriction of systemic ‘leakage’ vessels increases the outflow rate on the venous side. Meanwhile, α-adrenergic stimulation of the arterio-venous anastomoses (Fig. 4) reduces venous outflow (Payan & Girard, 1977; Nilsson & Pettersson, 1981). The overall combined effect of these counteracting responses is expressed by the large variability in the outflow from the inferior jugular vein catheter, where no significant effects were found (Table 2).

Adrenaline injections in vivo increases dorsal aorta (Steen & Kruysse, 1964; Peyraud-Waitzenegger, 1979). This result has now been confirmed in a perfused preparation, and demonstrated to depend on direct effect of adrenaline on the gills (Table 2, Fig. 36). Adrenaline did not affect dorsal aorta in series A, where no O₂ tension gradient existed between the irrigating water and the perfusion fluid (Table 2, Fig. 3 a), showing that adrenaline does not affect the O₂ requirement of the branchial tissue. The increase in dorsal aorta by adrenaline in series B is, therefore, due to an increased gas exchange efficiency. It is the relatively low flow rate through the venous system that causes the high utilization of O₂ apparent in the venous perfusion fluid (Table 2).

Booth (1979) reported an increase in the number of perfused secondary lamellae following adrenaline injections in the rainbow trout, a finding supported by data from Holbert, Boland & Olson (1979). Our results, which demonstrated an increased O₂ transfer across the gills, support this possibility. Another explanation for the increased efficiency in O₂ transfer from the irrigating water to the perfusion fluid could be increased permeability of the secondary lamellar membrane to lipophilic substances such as oxygen (Haywood, Isaia & Maetz, 1977; Isaia, Maetz & Haywood, 1978).

In conclusion, this paper describes a myogenic vasoconstriction of the branchial vascular bed in response to hypoxia, probably leading to lamellar recruitment and increased functional surface area in the gills. Furthermore, adrenaline enhances oxygen transfer and causes an increased systemic blood flow. The combination of hypoxic vasoconstriction and adrenaline release, such as is likely to occur during reduced ambient O₂ availability, will thus optimize O₂ transfer as well as improve the arterial O₂ transport.