Combined Effects of Environmental P_O2 and Temperature on Ventilation and Blood Gases in the Carp Cyprinus Carpio L.

Fish may encounter large and rapid variations in environmental temperature and in the availability of oxygen. Several regulatory mechanisms, involving various steps in the O₂ transport chain, are utilized to cope with adverse conditions. Responses to acute environmental hypoxia include increased gill ventilation (Johansen, 1982) and improved branchial O₂ diffusing capacity, which is achieved by recruitment of secondary lamellae as well as by secondary lamellar distension owing to an increased perfusion pressure (Soivio and Tuurala, 1981). Oxygen transport by the blood is promoted by an increased O₂ capacity or by an elevation of the O₂ affinity, achieved by modulation of the internal environment of the red blood cells (Tetens and Lykkeboe, 1985).

Information concerning directly measured responses of gill ventilation in teleost fish exposed to hypoxia at different temperatures is scarce, and simultaneously determined blood gas data are not available (Dejours, 1981; Johansen, 1982). Lomholt and Johansen (1979) applied electromagnetic flowmeters to evaluate the effects of considerably decreased water on gas exchange and gill ventilation. Another study utilized the indirect estimation of ventilation by the Fick principle to describe the effects of changes in on ventilation and blood gases in the carp (Itazawa and Takeda, 1978). The degree of hypoxia applied in these studies was considerable. Inspired water was generally reduced to less than 50mmHg (6.7 kPa, Itazawa and Takeda, 1978; Lomholt and Johansen, 1979) or, in some series, even to 25 mmHg (3.3 kPa, Itazawa and Takeda, 1978).

In contrast, the present study has focused on much less severe hypoxia, but has adopted temperature as an additional variable. This strategy was followed to determine whether the ventilatory response to hypoxia was temperature-dependent (Wood, 1982; Glass et al. 1983). Temperature-dependent regulation would be advantageous, given the background of an increasing oxygen uptake at higher temperatures and simultaneously falling blood oxygen-affinity. If conditions for oxygen delivery at the systemic capillary level are to be unchanged, these effects could be compensated for by an increased perfusion rate and, at least partially, also by improved branchial gas exchange conditions. A marked elevation of branchial gas exchange, comparable to the rise in oxygen consumption, could, however, only be brought about by a reduction of the inspired/average gill water difference, induced by an increase in gill ventilation.

In amphibians and reptiles the ventilatory response to hypoxia is markedly decreased at lower body temperatures. This is consistent with a considerably depressed ventilatory response in the air-breathing fish Channa argus during aerial hypoxia at low temperatures (Glass et al. 1986). Similarly, the water-breathing spangled perch Leiopotherapon unicolor essentially lacks any response in breathing frequency at reduced body temperatures (Gehrke and Fiedler, 1988). Breathing frequency alone, however, may be misleading. This is emphasized by studies of ventilation in rainbow trout (Salmo gairdneri), which demonstrated a response to hypoxia mainly by elevation of tidal volume, with essentially no change in breathing frequency (Smith and Jones, 1982). Likewise, no conclusion can be drawn from data obtained in bluegill sunfish Lepomis macrochirus, where breathing frequency increased during moderate hypoxia at 13 and 25 °C, but did not change at 30°C (Spitzer et al. 1969). Uncalibrated estimates of tidal volume indicated that moderate hypoxia was associated with an increased ventilation amplitude at lower temperatures, but not at 30°C.

Accordingly, ventilation was determined in this study by a method allowing simultaneous monitoring of tidal volume and breathing frequency. Gill ventilation together with arterial content and parameters of the acid-base status were measured in carp acclimated to normoxia at two temperatures (10 and 20°C), and during subsequent exposure to acute hypoxia (⁠ of 110 or 75 mmHg/14.7 or 10kPa).

Specimens of carp (Cyprinus carpio, mass 1.3–3.3 kg; 2.3 ± 0.1 kg, x̄±S.E., N=32) were acclimated for 4-8 weeks to 10 or 20°C in large glass aquaria (about 1001 volume per fish) in air-equilibrated water, with a continuous flow-through of fresh Gottingen tap water at a rate of at least 100 ml min⁻¹ animal⁻¹). The fish were fed daily on commercial carp food, and food was withheld only about 48h before experimentation. Blood was sampled from all experimental animals (N=32); gill ventilation was measured in 24 fish.

Anaesthesia was initiated by immersion of the animals in buffered (NaHCO₃) MS 222 (1:10000); after loss of reactivity the anaesthesia and the oxygen supply were maintained by irrigation of the gills with aerated, thermostatted water at a reduced MS 222 concentration (1:15000). A catheter for blood sampling was inserted into the dorsal aorta by direct puncture, applying a modification of the Seidinger technique, and then secured with 3–5 atraumatic surgical polyamide sutures (for details see Claiborne and Heisler, 1984).

Gill ventilation was measured by application of a mask technique. To separate inspired from expired water flows, a funnel was constructed from parts of a surgical rubber glove (Fig. 1), the lumen of which was extended and supported by several rings. The smaller, open end of the funnel was sutured around the mouth of the carp by means of 20–30 small atraumatic polyamide sutures. The larger, open end of the funnel was fitted onto a device holding the electromagnetic flow sensors, so that all inspired water was channelled through the probes. The holding device was constructed of light-weight polymers and suspended from a float by polymer supports. This arrangement allowed the carp a certain degree of freedom in the experimental chamber. After these preparations, the anaesthesia was terminated and when, after a few minutes, the animal resumed spontaneous breathing it was left undisturbed in the experimental apparatus for at least 48 h before the experiment began.

To exclude any disturbing effects of the mask on the establishment of a steady state, the blood gas values of carp fitted with masks were compared with those of specimens equipped with only the dorsal aortic catheter. No differences were found between the two groups.

The experiments were conducted using a recirculation system consisting of a Plexiglas box (≈121), a gas equilibration column (≈61), a particle filter and a circulation pump, similar to the one described in detail by Claiborne and Heisler (1984). The gas equifibration column was supplied with normoxic or hypoxic gas mixtures at a rate of about 71 min⁻¹. Equilibration gases were prepared by a microprocessor-controlled mixing system on the basis of electronic mass flow controllers. Water temperature was regulated to ± 0.1 °C of the desired value (10 or 20 °C) by feedback regulation of thermostatting fluid flow through stainless-steel heat exchangers in the system. The recirculation system was completely flushed frequently with fresh pre-thermostatted and aerated Gottingen tap water to avoid build-up of toxic waste substances.

Normoxic control values of gill ventilation and blood gas variables in animals exposed to well-aerated water were determined starting 48 h after surgery. When repeated sampling indicated that a steady state had been attained, the water was lowered to either 110 or 75 mmHg (14.7 or 10 kPa). Gill ventilation was monitored continuously and the blood variables were measured 1 and 4h after initiation of water hypoxia.

Inspired gill ventilation was measured by means of two ‘cannulating’ electromagnetic flow probes (inner diameter 12 mm), controlled by two synchronized flowmeters (Gould/Statham SP2202). Two probes were used to reduce the breathing resistance. The signal from the flowmeters was recorded directly (Brush mark 260) and in integrated form for determination of ventilation minute volume. Calibration of the flowmeters and integrators was performed by pumping the experimental water at known rates (within the range of the inspired gill flow of carp) through the flow probes. The ventilation reported is based on the average gill ventilation determined during a period from 2 min before to 2 min after blood sampling.

Blood samples (1.5 ml), drawn from the dorsal aortic catheter, were analyzed for and pH by means of microelectrodes (Radiometer, Copenhagen) as described by Claiborne and Heisler (1984). Blood oxygen content was measured by a micro-method according to Tucker (1967). Blood saturation was calculated from O₂ and haemoglobin contents as described by Jensen et al. (1987). Haematocrit was determined in triplicate after centrifugation at 19000 g in micro glass capillaries, and the supernatant plasma was analyzed for total CO₂ content with a Capni-Con III (Cameron Instruments Inc., Port Aransas, Texas, USA). CO₂ solubilities at 10 °C and 20 °C were calculated from the polynomials of Heisler (1984). (Note: the first sign of the last line of the α-formula is misprinted and should read ‘+’.) Blood lactate concentration was measured enzymatically (Boehringer GmbH, Mannheim, FRG), and haemoglobin concentration was determined by the cyanmethaemoglobin method (Jensen et al. 1987).

Ventilation of carp in air-equifibrated water increased markedly when water temperature was raised from 10 to 20°C (Fig. 2). Correspondingly, the ventilatory responses to hypoxia at 20 °C were ‘scaled up’ compared with those at the lower temperature. Referred to baseline values for normoxic conditions, however, the increase in gill ventilation during hypoxia was similar at both temperatures (Fig. 3). The responses to moderate hypoxia were significant at both temperatures and were maintained throughout the 4h period of hypoxic exposure. Gill ventilation rapidly recovered to control values upon return to normoxia.

The breathing pattern of normoxic carp consisted of short breathing episodes of a few breaths alternating with breath-holds (Fig. 4). With rising temperature the number of breathing episodes per unit of time increased considerably. During normoxia the number of breathing episodes per minute was 2.4 ± 0.5 at 10°C (x̄±S.E., N=, and it increased to 6.5±0.8 at 20°C (x̄±S.E., N=6; P<0.01, grouped i-test). The number of breaths per episode remained constant (2.6±0.2 and 2.7±0.3 at 10°C and 20°C, respectively) (control values of 110mmHg/14.7kPa hypoxia series). The tidal volume of normoxic carp was smaller at 10°C than at 20°C, and the breathing frequency fell markedly, so that the changes in these respiratory parameters accounted for the temperature-related fall in normoxic ventilation (Table 1). Increased ventilation during hypoxia resulted partly from a larger tidal volume and partly from an elevated breathing frequency. Adjustment of tidal volume was more important at 10°C, whereas the frequency response dominated at 20°C (Table 1). This frequency response included a reduction in the duration of breath-holds, which often resulted in continuous, non-periodic ventilation at 20°C and inspired ⁠, of 75mmHg (10 kPa).

The ventilatory responses greatly affected the acid-base status of the arterial blood. At an inspired water of 75mmHg (10kPa), was reduced by almost half its normoxic values of about 3.3mmHg (0.44kPa, 10°C) and 4.0mmHg (0.53 kPa, 20°C) (Fig. 5). This fall in resulted in an elevation of dorsal aortic pH by about 0.2 units compared with the normoxic values of 8.07 (10°C) and 8.01 (20°C) (Fig. 6). The rise in pH was accompanied by a compensatory fall in plasma [HCO₃⁻] (Fig. 7). At both temperatures arterial was low during normoxic conditions (approximately 15mmHg, 2kPa). The hypoxia-induced reduction was not consistently significant (Fig. 8). Also, no clear correlation existed between the extent of reduction of inspired ⁠, and the fall in arterial ⁠. The response of arterial to inspired hypoxia varied considerably among individuals, so that in some cases the 4 h value exceeded even the normoxic control ⁠. Hypoxia did not affect arterial O₂ content or O₂ saturation (Table 2). Likewise, arterial haematocrit and haemoglobin concentration were unaffected. Average pooled control values (x̄±S.E., N=12) for haematocrit were 25.0±1.3% at 10°C and 25.7±0.9% at 20°C and average blood haemoglobin concentrations were 1.07±0.08 mmol l⁻¹ (tetramer) at 10°C and 1.21±0.1 mmol l⁻¹ at 20°C. Arterial lactate concentration rose under hypoxic conditions, in spite of the maintained blood oxygen saturation. The increase was most pronounced at 20°C with a rise from 1.36±0.35mmoll^-1 (normoxic controls) to 4.09±1.2mmol l⁻¹ during 4 h at an inspired of 75 mmHg (10 kPa) (x̄±S.E., N=7; P<0.02, paired t-test). At 10°C the same reduction of water resulted in an increase in blood [lactate] from 1.41±0.22 mmol l^−l to 2.10±0.32 mmol l⁻¹ (x̄±S.E., N=8; P<0.01, paired i-test).

Carp maintain a markedly increased gill ventilation, with relatively small and inconsistent decreases in and unchanged arterial O₂ saturation and O₂ content, in the face of hypoxia. These data resemble a pattern described by Hughes et al. (1983), in which of carp changed little during moderate hypoxia and even increased when the of the inspired water was reduced from 130 to 90mmHg (17.3 to 12 kPa). Another finding of the present study is that the ventilatory response to hypoxia was graded according to the degree of decrease in inspired water ⁠, although the effects on did not differ for the two hypoxic conditions. These findings are difficult to interpret owing to a large variation between individuals and correspondingly large standard deviation. Non-arterial chemoreceptors, however, may be involved, a notion supported by direct and circumstantial evidence for the existence of extra-arterial chemoreceptor sites (including venous and external locations) (Randall and Daxboeck, 1984; Milsom and Brill, 1986).

The exact nature of the breathing stimulus in fish exposed to hypoxia is not yet certain. Some data favour the hypothesis that blood O₂ content rather than is the decisive factor (Randall, 1982). Studying ventilatory responses to hypoxia and hypercapnia in the rainbow trout (Salmo gairdneri), Smith and Jones (1982) demonstrated an inverse relationship between arterial O₂ content and gill ventilation. Hypercapnia in the trout resulted in a reduction of O₂ content due to the Root effect, while ⁠, remained at normocapnic levels. Concomitantly, gill ventilation increased, suggesting that the O₂-mediated stimulus to breathing depended on content rather than partial pressure. In contrast, the present data for carp do not favour arterial O₂ content as a key factor, since substantial increases in ventilation occurred, and were maintained, while O₂ content remained at normoxic levels.

Regardless of the precise nature of the stimulus and the location of the receptor system, the regulatory system is capable of preventing a significant fall in and oxygen saturation, relative to the characteristically low values during normoxia (Table 3). Although water exceeded 150 mmHg (20 kPa) in air-equilibrated water, of the carp was only about 15mmHg (2kPa), i.e. the difference between water and arterial oxygen partial pressure was greater than 135 mmHg (18 kPa). The attained permitted an oxygen saturation of barely 50% at 20°C, and somewhat higher at 10°C because of the temperature-related increase in blood O₂-affinity. Below about 50 %, oxygen saturation may approach a critical threshold for complete tissue oxygen supply (Glass et al. 1983). Accordingly, we expected a ventilatory response to decreases of inspired ⁠, although the extent of hypoxia applied in this study was much less severe than those of previous investigations (Table 3). A response was expected particularly because the O₂ extraction of carp is quite high even in normoxia, precluding any substantial increase in this variable to alleviate hypoxic conditions (Lomholt and Johansen, 1979). Consequently, during hypoxia the oxygen demand has to be satisfied by an increase in gill ventilation. In spite of its reputation as being highly tolerant to lack of O₂ (Blazka, 1958), the carp regulates oxygen supply to the tissues to maintain an aerobic metabolic energy production. Curiously, the blood lactate levels, found to be in the same range as in earlier studies on carp and other fish species (Albers et al. 1983; Fuchs and Albers, 1988; Heisler, 1984), increased during hypoxia, although Pa_O2 was little affected and arterial O₂ content remained constant. This pattern is consistent with data for the channel catfish Ictalurus punctatus (Burggren and Cameron, 1980), and may be indicative of a small safety margin for complete oxygen supply in fish.

The lack of any effect of hypoxia on arterial oxygen saturation in carp can partly be explained on the basis of hyperventilation-induced increases in pHa combined with the large Bohr factor (Ultsch et al. 1981; Albers et al. 1983). In addition, the O₂ affinity of carp blood may become considerably elevated owing to the readjustment of the intracellular pH of erythrocytes induced by decreased nucleoside triphosphate (NTP) concentrations (Lykkeboe and Weber, 1978). However, this is unlikely to have been a major factor in the present study. The absence of any increase in haematocrit or haemoglobin concentration during acute moderate hypoxia is consistent with data for rainbow trout, in which more severe hypoxia is required to increase haematocrit, caused both by swelling of erythrocytes and by an elevation in red cell haemoglobin content (Tetens and Lykkeboe, 1985).

Comparatively low values are characteristic of carp, although data in the literature are inconsistent (Table 3). This scatter is probably unrelated to the methodological approach, because differing values have been obtained using identical techniques (Claiborne and Heisler, 1984; present study). The variability may rather be due to unidentified biological influences, as discussed earlier in relation to the effects of hypoxia on O₂ uptake (Hughes et al. 1983). A number of conditions may account for the low of carp. The typically high O₂ extraction from the water in the gills of carp (Saunders, 1962; Lomholt and Johansen, 1979) sets an upper limit for the range of ⁠. Full blood oxygen saturation is not achieved, which could result partly from a mismatch between ventilation and perfusion, or from blood bypassing the functional gas-exchange surfaces (Piiper and Scheid, 1984). Regardless of the mechanisms involved, incomplete blood saturation implies low Pao₂ values because of the high oxygen affinity (low P₅₀) of carp blood (Wood, 1982; Albers et al. 1983).

In contrast to the data on arterial oxygen partial pressure, consistent values for gill ventilation have been reported for carp (Table 3). Indirect estimates (Itazawa and Takeda, 1978) based on the Fick principle (Rahn, 1966) agreed well with direct measurements of ventilation (Lomholt and Johansen, 1979; present study), whereas an indirect approach based on the differences in between inspired and expired water sampled from buccal and opercular catheters, respectively, resulted in a somewhat higher value (Saunders, 1962). This relative consistency between techniques is somewhat fortuitous considering the problems resulting from periodic ventilation, from an uneven distribution in the expired water stream (Garey, 1967) and from the potential cutaneous uptake of oxygen from the water (Feder and Burggren, 1985).

The periodic breathing pattern of carp (Peyraud and Serfaty, 1964; Lomholt and Johansen, 1979; Hughes et al. 1983; present study), with its regular interruptions of breathing activity by periods of apnoea, is characteristic of ectothermic airbreathers. This periodicity also occurs in other teleost fish (Gehrke and Fiedler, 1988; Roberts and Rowell, 1988), and in elasmobranchs during hyperoxia (Heisler et al. 1988), although the rate of ventilation is much greater than in air-breathing lower vertebrates, matching the lower O₂ content of water relative to air (Rahn, 1966). The periodic breathing pattern leaves a considerable margin for periods of elevated ventilation brought about by changes in ventilatory frequency, without the necessity for increasing the flow velocity of the viscous breathing water during individual breaths. Although the ventilatory adjustments described in this study are also mediated by increases in tidal volume, at least at 10°C, tidal volume was rather constant at 20°C. This may be related to the fact that any further elevation of the flow velocity of the comparatively viscous water is energetically impossible on the basis of the much higher normoxic ventilation rate at 20°C (see Table 1). At any rate, these data clearly indicate that measurement of breathing frequency alone is insufficient for an assessment of control of breathing in fish. This is also evident from data in the literature on breathing frequency and O₂ uptake measured during small changes in water temperature occurring over a few minutes (Moffitt and Crawshaw, 1983). The resulting temperature-induced changes in O₂ uptake were accompanied by rapid adjustments of breathing frequency (Q₁₀=2.6, consistent with this study) as well as by a reduction in tidal volume which, however, could not be quantified by the techniques used.

Several recent studies have investigated the effects of temperature on the ventilatory response to hypoxia in ectothermic air-breathers. Comparable data are scarce for fish (Gehrke and Fiedler, 1988), although this aspect of respiratory control is apparently highly relevant, considering the variability of temperature and O₂ availability in the aquatic environment. The data provided by this study do not allow a generalized conclusion to be drawn about the differences in the regulation of ventilation between air-breathing and water-breathing ectothermic vertebrates, but some obvious characteristics may be listed tentatively. In carp, consistent with data for some other teleost fish (Burggren and Cameron, 1980; Rantin and Johansen, 1984), a considerable ventilatory response is provoked by relatively moderate falls in inspired ⁠, whereas several-fold larger decreases in inspired are required for a comparable response in ectothermic air-breathers (Glass et al. 1983). Also, carp retain a relatively large ventilatory response at 10°C, whereas hypoxic responses at similarly low temperatures are virtually absent in reptiles and amphibians (Glass et al. 1983, 1986). In contrast to air-breathers, the more than twofold increase of ventilation in carp exposed to an inspired of 75 mmHg (10 kPa) is barely correlated with small and rather inconsistent reductions in ⁠. These differences are presumably related to the lower oxygen content of the breathing medium at the same level of inspired ⁠, which can barely be completely compensated by generally enhanced levels of ventilation, higher blood oxygen-affinity and larger ratios of perfusion over oxygen consumption of systemic tissues. Accordingly, the safety margin for maintaining complete oxygen supply in fish is probably much smaller than in air-breathers, which must favour development of a predominantly O₂-transport-oriented control of breathing to match the frequently changing and generally lower environmental O₂ availability.

We would like to acknowledge support to MLG by the Alexander von Humboldt-Stiftung and later by the Carlsberg Foundation, and by Deutsche Forschungsgemeinschaft. We also greatly appreciate the excellent technical assistance of Sylvia Glage and Giinther Forcht.