Control of the Systemic Heart and the Portal Heart of Myxine Glutinosa

Several investigations in the past have shown that the systemic heart of hagfish behaves as if it is aneural (Augustinsson et al. 1956; Jensen, 1965), even though nerves in the heart and adjacent to it have been reported for the Pacific hagfish Eptatretus stoutii (Hirsch et al. 1964). Experiments performed on the isolated systemic heart of the Atlantic hagfish Myxine glutinosa have shown that drugs known to have strong effects on the hearts of other vertebrates (catecholamines, acetylcholine) have almost no effect (Östlund, 1954; Fänge and Östlund, 1954). However, in vivo adrenaline application increased systemic heart rate, stroke volume and blood pressure and application of the β-adrenoceptor antagonist sotalol decreased heart rate, which suggests the presence of a tonic β-adrenergic stimulation of the systemic heart (Axelsson et al. 1990).

In addition to the systemic heart, hagfish have a portal heart, an accessory heart situated on the hepatic portal vein (Fänge et al. 1963) and unique among the vertebrates. Both this and the systemic heart contain high levels of catecholamines (Östlund, 1954; Östlund et al. 1961; Bloom et al. 1961; von Euler and Fänge, 1961). Jensen (1961) suggested that the hagfish systemic heart may have an endocrine function. The basis for this suggestion was the finding that the systemic heart of E. stoutii was insensitive to an extract made from hagfish systemic hearts and thus containing catecholamines. The insensitivity to catecholamines and the effect of sotalol can be explained if there is a continuous saturated stimulation of the β-adrenoceptors which makes the heart unresponsive to further stimulation although still sensitive to the antagonistic action of sotalol (Axelsson et al. 1990). This would be consistent with the observation that reserpine, which depletes the catecholamine content of the heart, had a negative chronotropic effect on the perfused systemic heart of Myxine glutinosa, but subsequent addition of catecholamines to the perfusion fluid caused the heart to resume beating (Bloom et al. 1963). Release or leakage of catecholamines could constitute a means of regulating or maintaining heart function and/or could act on structures downstream (Forster et al. 1991). Perry et al. (1993) have shown that the cholinoceptor agonist carbachol can elicit catecholamine release in an in situ saline-perfused hagfish heart, although some other mechanism, possibly involving pituitary hormones, may operate in vivo.

Increased filling pressure/volume has been observed to accelerate the systemic heart of M. glutinosa and E. stoutii (Johansen, 1960; Jensen, 1961; Bloom et al. 1963; Chapman et al. 1963), although information on pressure is limited. In a study of the isolated perfused systemic heart of E. cirrhatus, heart rate remained unchanged as either preload or afterload was varied (Forster, 1989).

Some aspects of possible control mechanisms of the systemic heart and the portal heart were investigated in the present study. The effects of preload and afterload on the performance of the systemic heart were studied using an in situ perfusion technique which allowed accurate pressure measurements and calculations of power generation. The blood pressure in the posterior cardinal vein, which supplies the systemic heart, was measured in resting animals, and the chronotropic effects of different agonists and antagonists were investigated in isolated systemic and portal hearts.

Hagfish (Myxine glutinosa L.) were caught in the Gullmars Fiord on the Swedish west coast and were transported to Göteborg. Animals of either sex, weighing between 20 and 100 g, were used in the experiments. Prior to use, animals were kept in aquaria with circulating sea water at 10–11 °C. The animals were killed by decapitation.

A large ventral incision revealed the systemic heart and connecting vessels of freshly killed animals. Saline-filled cannulae (PE200) were advanced into the ventral aorta and the posterior cardinal vein (PCV) and firmly secured. The cannulae were of the same construction as those used by Franklin and Axelsson (1994), making pressure measurements close to the heart possible. This is an advantage compared with methods in which the pressure is measured in a connecting tube some distance from the preparation and where the actual pressure must be calculated by taking into consideration factors such as the resistance of the connecting tubes. In this study, pressure was measured using a PE90 catheter tipped with a short piece of PE10 tubing inserted into the PE200 cannulae. The PE90 tubing was connected to a Honeywell pressure transducer (model 156PC06GW2).

The inflow cannula was connected to a reservoir containing Ringer’s solution (perfusion fluid) and a constant-pressure device which could be raised or lowered to set the input pressure. The Ringer’s solution had the following composition (g l^-1): 27.7 NaCl, 0.6 KCl, 0.73 CaCl₂.2H₂O, 0.72 MgSO₄.7H₂O, 1.82 MgCl₂.6H₂O, 0.08 NaH₂PO₄, 1.26 NaHCO₃, 1 glucose. Noradrenaline (3 nmol l^-1) and adrenaline (1 nmol l^-1) were added to the perfusion fluid to provide physiological catecholamine levels (see Perry et al. 1993). The Ringer was bubbled with 0.3 % CO₂ in air and kept at 10 °C. The outflow cannula was connected to a tube which could be adjusted vertically to alter the afterload. Input and output pressures and heart rate were displayed on a chart recorder (Grass Polygraph model 79). Heart rate was derived from the pulsatile output pressure signals using a Grass 7P44 tachograph unit. AD/DATA software (P. Thorén, Hässle AB, Sweden) was used to sample and store the data on a personal computer. The digital display of the computer allowed accurate calibration and pressure measurements. The pressure transducers were calibrated against a static water column, where zero was set to the level of Ringer’s solution in the organ bath in which the preparation was immersed.

Starling curves were constructed by gradually increasing the inflow pressure until no further increase in flow rate was seen. Flow rate was determined gravimetrically. The output pressure was kept constant close to physiological values for ventral aortic pressure (see Axelsson et al. 1990). Power curves could be determined by keeping the inflow pressure constant at a nearly physiological level and gradually increasing the output pressure. The β-adrenoceptor antagonist sotalol (10^-5 mol l^-1) was added to the perfusing Ringer’s solution and the effects of preload and afterload were investigated during sotalol treatment, i.e. each preparation acted as its own control.

Individual curves where cardiac variables (stroke volume, flow rate and power output) were plotted against input or output pressures, and fitted to the data using a third-order polynomial, made it possible to construct composite graphs with data generated at set pressure intervals (see Franklin and Axelsson, 1994). Stroke volume and flow rate were normalised per kilogram body mass (BM) and power output was normalised per gram ventricle mass (VM).

The Wilcoxon signed-ranks test was used to evaluate statistically significant differences (P<0.05). Comparisons were made between the value at an input pressure of 0.02 kPa and the value where the input pressure corresponded to the mean in vivo venous blood pressure (preload curves) or the value where the output pressure corresponded to the peak of the power output curve (afterload curves). Comparisons between untreated and sotalol-treated preparations were made at the same points. Data are presented as means ± S.E.M.

Hagfish were anaesthetised in a mixture of 0.4 g l^-1 benzocaine and 0.4 g l^-1 MS222. A small ventral incision close to the tail exposed the PCV underneath the intestine. The vein was cannulated with PE50 tubing filled with 3 % NaCl containing 50 i.u. of heparin. Animals were allowed to recover for 24 h prior to venous pressure measurement. The animals were transferred to an experimental tank and the cannula was attached to a Honeywell pressure transducer (model 156PC06GW2) calibrated against a static water column, where zero was set to the water level in the experimental tank. Pressures were recorded on a Grass recorder. Data are presented as means ± S.E.M.

The systemic heart and the portal heart were removed from freshly killed animals and suspended in organ baths containing the Ringer’s solution described above. The isolated hearts were attached to Grass isometric force transducers (FT03) connected to a Grass Polygraph recorder system (model 79). Calibration of the system was made using a 1 g weight. The rate of contraction was obtained from the spontaneously contracting hearts using a Grass 7P44 tachograph unit. AD/DATA software was used to sample data.

Chronotropic effects of adrenaline and acetylcholine before and after sotalol (10^-5 mol l^-1) treatment were investigated by cumulative addition of these agonists (10^-9 to 10^-5 mol l^-1) to the organ baths, except for the addition of acetylcholine after sotalol treatment (10^-5 mol l^-1), when only one concentration was tested (10^-5 mol l^-1). Chronotropic effects of carbachol before and after atropine (10^-5 mol l^-1) treatment were investigated by cumulative addition of carbachol (10^-9 to 10^-3 mol l^-1). Chronotropic effects of adrenocorticotropic hormone (ACTH) were investigated by cumulative addition of ACTH (10^-8 to 10^-6 mol l^-1) to the organ baths. The Wilcoxon signed-ranks test was used to evaluate statistically significant differences (P<0.05). Data are presented as means ± S.E.M.

The following drugs were used: adrenaline bitartrate (Sigma), arterenol bitartrate (Sigma), sotalol hydrochloride (Bristol Laboratories, Bristol-Myers Squibb), carbamylcholine chloride (Sigma), atropine sulphate (Sigma) and porcine adrenocorticotropic hormone (Sigma). The drugs were dissolved in distilled water to the appropriate concentrations.

Incre. ased input pressure (P_in) produced an increased flow rate (Q̇) through the systemic heart by an increase in both. stroke volume (VS) (Fig. 1) and heart rate (fH) (Fig. 2). Q̇ was 29.4±3.0 ml min^-1 kg^-1 BM and VS was 1.3±0.2 ml beat^-1 kg^-1 BM in untreated preparations at a P_in of 0.1 kPa, which corresponds to the mean in vivo venous blood pressure (0.10±0.02 kPa). Table 1 summarizes the results from venous pressure measurements and gives mean values for cardiovascular variables at physiological pressures. Treatment with the β-adrenocepto.r sotalol (10^-5 mol l^-1) resulted in a significant decrease in Q̇ to 19.1±2.4 ml min^-1 kg^-1 BM.

Comparison of minimum and maximum values of fH occurring during the course of P_in elevation for each preparation shows that both the untreated and sotalol-treated preparations are pressure-sensitive at relatively low pressures. The lowest fH values recorded were 23.0±1.0 beats min^-1 at a P_in of 0.020±0.002 kPa for the untreated preparations and 20.5±1.5 beats min^-1 at a P_in of 0.020±0.002 kPa for sotalol-treated hearts. The highest fH values were 26.7±1.2 beats min^-1 at a P_in of 0.10±0.02 kPa and 24.4±1.5 beats min^-1 at a P_in of 0.15±0.02 kPa, respectively, for untreated and sotalol-treated preparations. Sotalol treatment significantly decreased the fH by 2.8±1.3 beats min^-1. The P_in ranges where minimum and maximum fH were observed were not very different between untreated and sotalol-treated heart preparations (Fig. 2).

Power output increased in response to elevated P_in (Fig. 1) in both untreated and sotalol-treated preparations, although not to the same extent. There was a significant difference between the untreated preparations, in which the power output was 0.50±0.06 mW g^-1 VM, and the sotalol-treated preparations, in which the power output was 0.30±0.04 mW g^-1 VM, at an inpu. t pressure of 0.1 kPa.

Q̇ decreased significantly as a result of a decrease in VS when the output pressure (P_out) was raised (Fig. 3) and, in sotalol-treated hearts, a small decrease in fH (not shown). At a P. ^out of 1.8 kPa, the value at which power output peaked, Q̇ was 16.5±1.8 ml min^-1 kg^-1 BM and VS was 0.68±0.09 ml beat^-1 kg^-1 BM in the untreated preparations. Sotalol. treatment produced a significant decrease in Q̇ (10.5±2.1 ml min^-1 kg^-1 BM) due to a decrease in fH by 4.8±1.1 beats min^-1 (N=11), while VS remained unchanged. Power output was substantially lower in the sotalol-treated heart (0.31±0.07 mW g^-1 VM) than in the untreated heart (0.54±0.07 mW g^-1 VM).

fH for the systemic heart and the portal heart were 19.3±1.5 beats min^-1 (N=8) and 19.7±1.0 beats min^-1 (N=9), respectively, before addition of adrenaline. Adrenaline had no effect on fH while the β-adrenoceptor antagonist sotalol decreased systemic fH to 13.5±0.8 beats min^-1 (N=7) and portal fH to 15.5±1.0 beats min^-1 (N=8). Fig. 4 illustrates the effect of sotalol added to the Ringer’s solution bathing the preparations expressed as mean values taken from individual preparations over time. Moderate concentrations (10^-9 to 10^-6 mol l^-1) of adrenaline after sotalol treatment did not elicit any response, although a small acceleration could be observed in both the systemic heart and the portal heart at adrenaline concentrations between 10^-6 and 10^-5 mol l^-1.

Acetylcholine application at 10^-6 and 10^-5 mol l^-1 decreased portal fH by 3.0±0.7 and 2.1±0.6 beats min^-1, respectively (N=8), and also decreased systemic fH by 1.8±0.4 beats min^-1 (N=8), but in the latter case the difference was significant only at an acetylcholine concentration of 10^-6 mol l^-1 (results not shown). Acetylcholine (10^-5 mol l^-1) application after sotalol treatment had no significant effect on portal fH, but produced a significant decrease (1.6±0.4 beats min^-1) in systemic fH (N=8). Neither the cholinergic agonist carbachol nor the cholinergic antagonist atropine had any effect on the fH of the systemic heart (N=12) or the portal heart (N=8). A small but significant negative chronotropic effect on the systemic heart was caused by carbachol (10^-3 mol l^-1) after atropine treatment (N=10). ACTH (10^-8 to 10^-6 mol l^-1) had no effect on the fH of either heart (N=8).

This study is the first evaluation of the power-generating capacity of the Myxine glutinosa systemic heart. The experiments investigating the effects of different agonists and antagonists on the systemic heart are not without precedent, although previous reports are inconsistent and, in the case of the portal heart, scarce.

The systemic heart and the portal heart of Myxine glutinosa are comparatively insensitive to adrenergic and cholinergic agonists. It has been suggested that the chromaffin cells of the systemic heart are able to release catecholamines in response to cholinergic agonists, and possibly ACTH (Perry et al. 1993); however, it was not possible to detect any effects of endogenous catecholamines on the isolated heart preparations in the present study.

The effects of the cholinergic agonists are contradictory. Acetylcholine at high concentrations decreases fH of the systemic heart by 9 % and of the portal heart by 15 %, while carbachol has no effect. The results of Östlund (1954) and Fänge and Östlund (1954) on the systemic heart showed no effect of acetylcholine, which is in agreement with the lack of innervation of the systemic heart (Augustinsson et al. 1956). The physiological relevance of the chronotropic effects of acetylcholine in this study and the finding that catecholamine release can be cholinoceptor-mediated remain unresolved.

In accordance with in vivo studies on the systemic heart (Axelsson et al. 1990), sotalol decreased systemic fH by 30 % and portal fH by 21 % in the isolated heart preparations and decreased fH by 12 % in the in situ perfused systemic heart. This provides further evidence for a tonic β-adrenoceptor stimulation of the systemic heart and also demonstrates that β-adrenoceptors on the portal heart are tonically stimulated. Endogenous catecholamine stores appear to saturate cardiac β-adrenoceptors as no stimulatory effect of adrenaline could be detected on the isolated hearts.

The previously reported accelerating effect of adrenaline on the systemic heart in vivo (Axelsson et al. 1990) is not consistent with the insensitivity to adrenaline demonstrated in this study and with observations by Östlund (1954) and Fänge and Östlund (1954). Adrenaline may cause vasoconstriction and thereby augment venous return to the heart in vivo and, by means of a pressure-sensitive mechanism, increase fH. Intrinsic rate regulation has been shown in elasmobranchs (Jensen, 1970), although information concerning teleosts is scarce (Farrell, 1984). Heart rate in E. cirrhatus increases during swimming, possibly caused by an increased venous return (Forster et al. 1988). In the present study, an increase in fH of 14 % was found when P_in was raised within the range of venous blood pressures observed in resting hagfish. In the in situ perfused portal heart of E. cirrhatus, an increase in P_in caused a β-adrenoceptor-dependent increase in the rate of contraction (Johnsson et al. 1996). The systemic fH of E. cirrhatus did not change in response to increased preload (Forster, 1989), possibly because the preparation did not include the sinus venosus, which is the pacemaker region in vivo (Davie et al. 1987).

The power output at a P_in of 0.1 kPa, which was the mean venous pressure in resting animals (Table 1), and at a P_out close to mean aortic pressure was 0.50±0.06 mW g^-1 VM. This is considerably higher than the reported maximal power output achieved by the systemic heart of E. cirrhatus (0.37±0.03 mW g^-1 VM) (Forster, 1989). Estimates of resting power output for Myxine glutinosa by Driedzic et al. (1987) (0.05 mW g^-1 heart mass) and Axelsson et al. (1990) (0.15 mW g^-1 VM) are lower than the present results, probably because of the different methods employed. Sotalol treatment reduced power output by 40 % and flow rate by 35 %, which again supports the suggestion that β-adrenoceptor stimulation is vital for the normal function of the hear.t (Axelsson et al. 1990). There was a fourfold increase in Q̇ for the systemic heart preparation when P_in was elevated from 0.02 to 0.1 kPa. This indicates that augmented venous pressure could modulate VS as well as fH. Wh.en afterload was increased from 0.2 to 1.8 kPa (P_in≈0.05 kPa), Q̇ decreased by 33 % due to a 31 % decrease in VS, although the systemic heart was still able to pump approximately 17 ml min^-1 kg^-1 BM and to develop a VS of approximately 0.7 ml beat^-1 kg^-1 BM.

In conclusion, endogenous catecholamines are important for the normal activity of the systemic heart and the portal heart of Myxine glutinosa, as illustrated by the effect of sotalol on fH, Q̇ and power output. The previously reported increase in fH and Q̇ caused by adrenaline in vivo (Axelsson et al. 1990) was probably a secondary effect due to an increased venous return to the heart. T.he systemic heart responds to increased preload b.y increased Q̇ in accordance with Starling’s law of the heart. Q̇ was augmented by means of an increased VS and a β-adrenoceptor-independent pressure-sensitive acceleration of the heart.

This study received financial support from the Swedish Natural Science Research Council, the Forestry and Agriculture Research Council, the Adlerbertska Research Foundation, the von Beskow Foundation, the Helge Ax:son Johnson Foundation, the Hierta-Retzius Foundation for Scientific Research, the Royal and Hvitfeldska Foundation, the Lars Hierta Memorial Foundation and the Wilhelm and Martina Lundgren Research Foundation. The authors wish to thank Professor Stefan Nilsson for helpful comments on the manuscript.