Myocardial Oxygen Consumption in the Sea Raven, Hemitripterus Americanus : The Effects of Volume Loading, Pressure Loading and Progressive Hypoxia

Considerable information exists on oxygen uptake of the mammalian heart, which has a well-developed coronary circulation. In contrast information on of the teleost heart, where coronaries are not always present, is limited. Driedzic, Scott & Farrell (1983) measured myocardial to be about 0·28 μl s^-1 kg^-1 fish weight in isolated, perfused sea raven hearts under conditions of low afterload and a reduced cardiac output ⁠. Energy metabolism in the perfused hearts was highly aerobic and increased with power output of the heart.

In view of the modest amount of information concerning myocardial ⁠, a comprehensive study was initiated using an in situ, sea raven heart preparation. The suitability of the preparation for physiological studies is well established (Farrell, MacLeod & Driedzic, 1982; Farrell, MacLeod, Driedzic & Wood, 1983), in that the in situ heart can generate an in vivo work load and the power output of the heart can be varied by simple changes in preload and afterload. The sea raven heart lacks a coronary circulation and so it derives its O₂ supply from venous blood being pumped through the heart. The present work measured myocardial under different power output regimes and different levels of O2 delivery.

Sea ravens, Hemitripterus americanas Gmelin, were caught by otter trawl in Passamaquoddy Bay off St Andrews, New Brunswick. The fish were held in aerated, recirculating sea water tanks (910 °C) prior to use. A total of 35 fish was used for the study, weighing 0·77–2·05 kg (x̄ = 1·25 kg).

The in situ heart preparation is described in full by Farrell et al. (1982, 1983). In essence, the intact heart received a physiological perfusate at a constant input pressure head via a cannula placed in the hepatic vein. All other veins entering the sinus venosus were ligated. Cardiac output was delivered against an output pressure head via a cannula placed in the ventral aorta. Cardiac output was varied by adjusting the height of the input reservoir (preload). Afterload was varied by adjusting the height of the output pressure head. The nerve supply to the heart was severed and so the intrinsic rhythm of the sino-atrial pacemaker set the heart rate. During the preparation time of 10–15 min, the heart received venous blood or perfusate. The fish was fully immersed in a saline bath which acted as a reference for pressure measurements. The perfusate composition (in mmol I^-1) was NaCl, 150; MgSO₄.7H₂O, 2; KC1, 5; CaCl₂, 2·3; Na₂HPO₄, 2·3; NaH₂PO₄,0·2; dextrose, 16·7; and 10gl^-1 polyvinylpyrrolidone (PVP, M_r = 40 000). Control perfusate was gassed with 0·5% CO2 balance air and, after equilibration, the pH was adjusted to pH 7·9 with the addition of NaHO₃ (approximately 10·7 mmol I^-1). The perfusate reservoirs, delivery lines, and the saline bath containing the preparation were all water-jacketed to maintain the temperature at 10 °C.

Following the cannulation procedures, and mean output pressure were set at approximately 11 ml min^-1 kg^-1 fish weight and 40cmH₂O, respectively. The heart performed under these conditions for 10-20 min. If did not stabilize during this period, the preparation was discarded. Representative traces of the control cardiovascular variables ( ⁠, heart rate, output pressure and input pressure) were collected at the end of this stabilization period. Input perfusate samples were taken from the reservoir to provide three consistent values. Likewise, the perfusate leaving the ventral aorta was sampled via a three-way tap to obtain two consistent values. was the difference between the input and output ⁠.

These experiments (N = 9 fish) examined the effect of changing power output of the heart while changing O₂ delivery to the heart. Control was altered by changing preload; power output and O₂ delivery therefore changed in proportion to since the output pressure head and the perfusate were unchanged. Three changes in were examined: control plus 40%, control plus 60% and control less 40 %. A 3-min stabilization period was allowed at each new level prior to sampling the output perfusate and the cardiovascular variables. Control conditions were restored for at least 3 min between each challenge and cardiovascular and perfusate samples were taken at the end of each control period.

These experiments (N = 15 fish) examined the effect of changing power output of the heart while maintaining oxygen delivery (⁠ and constant). Mean output pressure was varied to alter power output, and preload was left unchanged. At each new level, a 3-min stabilization period preceded the sampling of the cardiovascular variables and the output perfusate. Output pressures of approximately 35, 40, 50 and 55 cmH₂O were used. While these pressures span the physiological range for ventral aortic pressures, the net effect on power was small in comparison with the changes. Control conditions were restored for at least 3 min between each challenge, and samples were taken at the end of this control period.

These experiments (N = 11 fish) examined the effects of a stepwise reduction in the perfusate with constant preload and afterload. In addition to the control (air saturated) perfusate four other levels were examined: 105, 80, 55 and 30Torr. The of the perfusate was constant. At each new level of hypoxia a stabilization period of 8–10 min was allowed prior to simultaneous sampling of the cardiovascular variables, the input and the output If the cardiovascular variables began to decline abnormally fast at any level of hypoxia, the heart was assumed to be dying and the experiment was stopped. In the hearts that survived the 30 Torr exposure, preload was increased to evoke the maximal increase in ⁠.

Cardiac output was measured in the outflow line with a flowthrough electromagnetic flow probe and its associated BL 610 Biotronix flowmeter. Input and output pressures were monitored via saline-filled cannulae with a Micron pressure transducer (Narco Life Sciences, Houston, Texas). The flow and pressure signals were suitably amplified and displayed on a chart recorder (Biotronix BL 882, Kensington, Maryland). The of the perfusate was measured at 10°C with an IL 113 acid-base analyser with associated electrode and water jacket. The electrode was calibrated with water-saturated gases (100 % N2 and 12 % O₂) prior to each experiment and the calibration was rechecked with air prior to each sample. The hypoxic gas mixtures were obtained using a multiple flow controller (Matheson model 8249, East Rutherford, New Jersey).

Mean values for input and output pressures and were obtained from area determinations on the pulsatile traces. The pressures were referenced to the saline level in the chamber and appropriate corrections were made for the pressure drops across the input and output cannulae. All pressures are expressed in cmH₂O (1 cmH₂O = 0·098 kPa). Heart rate was determined from the periodicity of the flow trace. Cardiac output (ml min^-1) = heart rate × stroke volume. Power output of the heart (mW) = (mean output pressure — mean input pressure) ⁠. Oxygen uptake of the heart ⁠, where α = 0·038 ml O₂ml^-1 Torr^-1 partial pressure (Altman & Dittmer, 1971). Mechanical efficiency of the heart (%) = [100] [power (mW) X 0·0498]/ ⁠. Cardiac output was normalized per kg fish weight, and power output and were normalized per g ventricular wet weight. The fish was weighed prior to the experiment and the ventricle was weighed following each experiment (Fig. 1). Each fish acted as its own control and mean values ± S.E. are given where appropriate. Statistical differences (P < 0·05) were determined with either a Wilcoxon signed-rank test or a Student’s t-test.

The control values for the cardiovascular performance and O₂ consumption show good agreement between the three experimental protocols (Table 1). All the experiments were performed during winter months (November-February), except six pressure loading and three volume loading experiments which were performed in early summer (May-June). The summer pressure loading experiments were treated separately in Table 1 since they had a significantly higher intrinsic heart rate, a higher and lower efficiency compared to the winter experiments. There were also small variations in ventricular weight expressed as a percentage of total body weight (Fig. 1 ).

Therefore, the control power output per g ventricle wet weight varied amongst preparations because was set according to body weight.

Using these data, the average ⁠, is 0·25 μl s^-1 for a 1-kg sea raven generating a myocardial power output of 0·85 mW at 10°C.

An increase in preload produced increases in ⁠, and power output (Fig. 2A). All individual fish showed a significant linear correlation between and power output (r > 93 % in all nine fish). was linearly related to power output (Fig. 3A), where ⁠, 50df, P<0·05).

Mechanical efficiency of the heart increased with volume loading. For example, a three-fold increase in power output (0·8 to 2·4 mW g^-1) improved efficiency from 15·3% to 21·3% (Table 2).

For the range used in these experiments, ⁠, decreased with increases in (Fig. 4A). Thus, when power output was increased by increasing ⁠, there was an increase in ⁠, an increase in mechanical efficiency, an increase in oxygen delivery, and a decrease in oxygen removal from the perfusate.

Changes in afterload altered power output without major changes in (Fig. 2B). Overall, a linear relationship existed between the ⁠, and power output (Fig. 3B). Since the winter experiments (N = 9) differed the summer experiments (N = 6), two linear regressions are presented: winter ⁠, P<0·05) and summer ⁠, P<0·05). Thus for the summer experiments, was slightly higher for a given power output, i.e. efficiency was lower, but both data sets had the same gradient.

Mechanical efficiency increased with pressure loading. For example, a three-fold increase in power output (0·8 to 2·4 mWg^-1) improved efficiency from 10·0% to 19·2% and from 11·2% to 17·5% in winter and summer experiments, respectively (Table 2).

increased with pressure loading in order to meet the added O₂ demand, provided (i.e. O₂ delivery) was constant (Fig. 4B). Whenever decreased during pressure loading, increased markedly. This was particularly evident in the three experiments where a mean output pressure could not be increased to 55 cmH₂O and decreased as output pressure was raised beyond 50 cmH₂O (Fig. 4B). Here the reduction in O₂ delivery was not completely offset by the increase in O₂ extraction. Control conditions could be restored subsequently, indicating that the heart may not have been damaged by this challenge.

Thus, when power output is increased with a constant O₂ delivery (⁠ and input constant), an increase in is achieved through improved O₂ removal from the perfusate.

These experiments lasted up to 80 min. At each level of hypoxia the cardiovascular variables were stable for many minutes and so the data summarized in Fig. 5 apply to relatively steady-state conditions.

Power output and were not significantly different from their control levels at an input of 81 Torr, but was significantly reduced (Fig. 5). Maintenance of power output in association with a decrease in resulted in an increase in the apparent aerobic efficiency of the heart. Below an input of 81 Torr, power output and were reduced significantly, even though preload and afterload were unchanged. One heart died during the transition from 81 Torr to 55 Torr and five hearts died during the transition from 55 Torr to 30 Torr.

was reduced almost three-fold at 55 Torr and over five-fold at 30 Torr, yet the decreases in power output were small by comparison (81% and 72% of control, respectively). At these extremes of hypoxia there was an increase in the apparent aerobic efficiency of the heart.

The intrinsic heart rate (39·8 ± 1·9 beats min^-1) did not change significantly with progressive hypoxia. Heart rate was reduced by 3–5 beats min^-1 in three of the four hearts that survived at 30 Torr. One additional fish showed an atypical, progressive decrease in heart rate (33·3 beats min^-1 at 165 Torr to 18·8 beats min^-1 at 30 Torr, even though ⁠, power output, and efficiency were not atypical compared to the other 10 fish.

At a of 30 Torr, only small increases in were possible when preload was raised to evoke a maximal increase in ⁠. Cardiac output increased by 8%, 22%, 46 % and 55 % with no significant change in heart rate in the four fish examined. Such increases in were terminal in that decreased rapidly in three of the hearts after only 20–180 s, and the control could not be restored subsequently. One heart maintained a 22% increase in for 10 min without dying.

This is the first comprehensive study of myocardial in a teleost. Based on the present observations, myocardial is 0·25μl O₂s^-1 for a 1-kg fish with a 0·8-g ventricle generating a power output of 0·85 mW. Driedzic et al. (1983), using isolated sea raven hearts, reported a similar value, but the power output was only 0·2 mW and the mechanical efficiency was 4·2%. The mechanical efficiency of the in situ heart was about 15 %, which is comparable to that of mammalian hearts (rat 11 %, Neely, Liebermeister, Battersby & Morgan, 1967; human 9–11 %, Gibbs & Chapman, 1979). The highest efficiency observed in an in situ sea raven heart with nor-moxic perfusate was 26 %, a value close to the 30 % observed in trained athletes after severe exercise (Gibbs & Chapman, 1979). It seems unlikely that the assumption that the trout heart is 40% efficient (Jones, 1971) will be substantiated.

The present myocardial measurement can be used to estimate what proportion of the O₂ contained in venous blood is consumed by the heart of the intact fish. Each heart stroke supplies about 10μl O₂, assuming stroke volume is 0·34 ml kg^-1 and venous O₂ content is 3 vol%. Yet the myocardium of a 1-kg fish requires about 0·37μ1 O₂ per heart beat, assuming heart rate is 40beatsmin^-1. Thus, myocardial removes less than 4 % of the O2 available in venous blood. Myocardial is also about 0’6 % of the standard O2 uptake of the resting animal. This estimate is based on a standard O₂ uptake measurement of 64μl Chs^-1 kg^-1 for the lingcod, a fish with a similar lifestyle to the sea raven (Farrell & Daxboeck, 1981), and is in the middle of the theoretical range (0·08% to 4%) proposed by Cameron (1975). The above calculations confirm theoretical predictions that the fish heart is efficient (Jones, 1971), has a low metabolic demand in terms of the whole animal (Cameron, 1975), and has a venous O2 supply that is more than adequate for the myocardial demands of resting fish (Jones & Randall, 1978).

is linearly correlated to power output of the heart and mechanical efficiency is improved as power output is increased. These findings are consistent with observations on mammalian hearts (Neely et al. 1967; Gibbs & Chapman, 1979; Suga et al. 1981, 1982) but do not support the assumption made for the trout heart (Jones, 1971) that efficiency is constant over a large range of values.

Using the normalized versus power output regression equations, the present findings appear to indicate that pressure work and stroke work have the same metabolic cost in the fish heart. The increases in were equivalent in both pressure-loaded and volume-loaded hearts and efficiency was improved by 5–6 % in both cases when power output was increased three-fold (Table 2). Observations on mammalian hearts demonstrate that an increase in pressure (pressure work) is more costly than an increase in flow (stroke work). Neely et al. (1967), for instance, demonstrated that if was increased three-fold without a significant change in systolic pressure, i.e. diastolic pressure was lowered, heart work increased two-to five-fold without a significant increase in ⁠. In contrast, increased 65 % with a doubling of heart work when the heart generated a greater aortic pressure and stroke volume was constant. However, when was increased without regulating systolic pressure (diastolic pressure constant), increased significantly (a two-fold increase for a six-fold increase in heart work). Thus one explanation for the apparently similar metabolic cost of volume and pressure loading in sea raven hearts is that the volume-loaded heart performed additional pressure work as systolic pressure rose with stroke volume (diastolic afterload was unchanged).

The present and power output data were normalized because the animal weight varied substantially (see Neely et al. 1967). For the volume-loaded heart, absolute and normalized values give the same relationship between and power output (Table 2). However, with pressure loading, efficiency does not improve appreciably (1–2%, Table 2) when absolute values are used, unlike the 5–6% with normalized values. This difference is probably related to the cost of pressure development in fish of different sizes. Only relatively small changes in power output within individual fish were possible with pressure loading (Fig. 2B) and the maximum power output of the smaller fish did not necessarily overlap with the minimum power output of the larger fish. This separation between the absolute data for large and small fish and the different gradient for versus power output, implies that pressure work is more costly in larger fish. This additional cost may be related to the fact that larger fish have relatively larger (thicker?) hearts (Fig. 1).

A seasonal difference in was apparent. Summer fish had a higher and a lower efficiency. Whether this is related to the higher intrinsic heart rate of summer fish observed here and in a separate study (M. S. Graham & A. P. Farrell, in preparation) is not known.

The O₂ content of the air-equilibrated perfusate was 0·76 vol%, which is about four times lower that that of the venous blood supplying O₂ to the heart. Because of this, the of the perfusate decreases during its passage through the heart, whereas the blood probably does not change significantly. This raises the question whether O₂ delivery to the myocardium was limited by perfusion with aerated saline. If it is assumed that the output is indicative of the O₂ gradient driving diffusion, then O₂ delivery from aerated perfusate was probably not diffusion limited since the output ( > 130 Torr) of the perfusate was always much greater than the venous in the intact sea raven (53 Torr, Farrell & Driedzic, 1980). Consequently, the cardiac performance in situ is probably directly comparable to the in vivo situation despite differences of O₂ content in the perfusion media. This conjecture is also supported by preliminary experiments where oxygenating the perfusate (99·5 % O₂) had no effect on cardiac performance or ⁠, and by the fact that the in situ heart performed physiological workloads and showed no deterioration after 2h of experiments.

Unlike the situation with aerated perfusate, O2 diffusion became limiting during perfusion with hypoxic saline. Power output was maintained with an input of 80 Torr, i.e., when the output of 60–65 Torr was slightly above the normal venous of 53 Torr. However, at an input below 55 Torr (output of 40–45 Torr) the decrease in O₂ removal as declined during hypoxia, which is the opposite of the situation with aerated perfusate (Fig. 3A), was a strong indication that O₂ diffusion became limiting. Perhaps the magnitude of the O2 limitation is better highlighted by the death of some hearts and the poor and often terminal response to preload. Extrapolation of this conclusion to intact fish is restricted by the limited information on the cardiac and venous during hypoxia. Nevertheless it seems reasonable to assume that O₂ diffusion does become limiting in the intact sea raven at a perhaps a few Torr lower than that used in the present work: this would account for the negligible change in venous and the possible importance of facilitated O₂ diffusion by haemoglobin. In other species the limiting will undoubtedly vary because of myocardial myoglobin content, ventricular thickness and the presence of a coronary circulation. In intact lingcod, a water of 25–45 Torr reduces cardiac performance (⁠ and arterial pressure reduced by 31 % and 10% respectively, Farrell 1982), but whether this response reflects an O₂ limitation is unknown. In contrast, the trout heart, which is supplemented by arterial blood, maintains its performance during environmental hypoxia (⁠ of 40 Torr) when the venous and arterial levels are 10 and 22 Torr respectively (Holeton & Randall, 1967; Wood & Shelton, 1980).

Hypoxia had no major effect on pacemaker frequency. Any decreases in rate observed here were in unstable preparations exposed to stressful situations (e.g. excessive work loads). These decreases were irreversible. Consequently, the bradycardia observed in intact fish at extremes of environmental hypoxia (e.g. Smith & Jones, 1978; Daxboeck & Holeton, 1978; Wood & Shelton, 1980; Farrell, 1982) is probably entirely a central reflex.

Hearts receiving input perfusate with a of 80 Torr were able to sustain the same level of performance as control hearts. It has previously been shown that sea raven hearts perfused with air-equilibrated media generate essentially all of their ATP requirements via aerobic metabolism (Driedzicet al. 1983). Thus, a 1-kg fish, with a 0·8-g heart and an oxygen consumption rate of 0·25 μl O₂s^-1 would have an ATP turnover rate of approximately 80 nmol ATPg^-1 s^-1. It may be calculated from the data of Turner & Driedzic (1980) that sea raven hearts subjected to anoxic conditions to stimulate glycolysis have a maximal anaerobic ATP production rate of approximately 20nmol ATPg^-1 s^-1. At an input of 80Torr, ATP demand could be matched closely with the sum of ATP regeneration through aerobic and anaerobic metabolism. Larger decreases in external oxygen availability resulted in a decrease in performance, presumably due to the inability to increase further the rate of ATP production. Only 5 of 11 hearts withstood the 30 Torr exposure and their capacity to increase was reduced. Imposed increases in were also detrimental to the heart’s survival. Only one heart sustained an increase in for longer that 2 min. The rapid collapse of the hearts at high work loads during hypoxia may reflect a problem not only with ATP production but also with the removal of anaerobic end products from the myocardium. Intracellular acidosis is detrimental to cardiac contractility (Gesser & Poupa, 1983 ; Farrell et al. 1983 ; Farrell, 1984) and it is recognised acidosis combined with anoxia impair contractility of ventricular strips considerably more than anoxia alone (Nielsen & Gesser 1983).

In summary, the present measurements of myocardial under various power output regimes and levels of O₂ delivery indicate many similarities between the fish heart and the mammalian heart despite the anatomical differences and lack of a coronary circulation in the fish. Perhaps the most important difference is the relative tolerance of hypoxia by the fish myocardium. This aspect of myocardial metabolism would be worthy of further investigation.

This work was supported in part by NSERC of Canada through grants to APF and WRD and in part by N.B. Heart Association through a grant to WRD. Appreciation is extended to the Director and staff of the Huntsman Marine Laboratory, St Andrews, N.B. for supplying the animals used in this study.