Effects of Preload and Afterload on the Performance of the In Situ Perfused Portal Heart of the New Zealand Hagfish Eptatretus Cirrhatus

The systemic heart of the New Zealand hagfish delivers a cardiac output comparable with that of the hearts of other moderately performing fish species, but at unusually low blood pressures (Forster, 1989; Forster et al. 1991). Venous return in hagfishes is facilitated by the action of a number of accessory pumps. The portal vein heart (portal heart) is one such accessory pump, which forces blood through the liver and onto the systemic heart. Blood is supplied from the supraintestinal vein and the anterior cardinal vein (Fänge et al. 1963). The portal heart has valves at its inlets and outlet to prevent backflow and is composed of cardiac muscle, in contrast to the other venous propulsors whose action depends upon skeletal muscle (Johansen, 1960, 1963; Satchell, 1991).

Contraction of the portal heart, which is not synchronous with the systemic heart (Arlock, 1975; Davie et al. 1987), originates in the widened proximal region of the supraintestinal vein (sinus supraintestinalis) (Müller, 1845; Fänge et al. 1963). Davie et al. (1987) compared the sinus supraintestinalis with the sinus venosus of the systemic heart and argued that it is responsible for the small increase in pressure preceding systole. In teleost fish, atrial contraction fills the ventricle and thereby has a major influence on cardiac output. Mechanisms controlling cardiac function in hagfish are largely unknown. The hagfish systemic heart is notable for its apparent lack of neural regulation (Augustinsson et al. 1956; Jensen, 1965). The systemic and portal hearts of hagfish have internal stores of catecholamines (Östlund, 1954; von Euler and Fänge, 1961; Bloom et al. 1961; Östlund et al. 1961; Laurent et al. 1983; Forster et al. 1991), which may modulate heart function and/or act on structures downstream (Axelsson et al. 1990; Perry et al. 1993).

Hagfish lack a rigid pericardium and must depend upon positive venous pressures to fill the systemic and portal hearts. The only previous study of changes in rate and force development of the portal heart was made on Eptatretus stoutii using a perfusion technique (Chapman et al. 1963). In that study, input and output pressures were not recorded; instead, intracardiac pressure was recorded by insertion of a catheter into the lumen of the portal heart through its wall. No calculations of flow or power output for the portal heart were presented. Chronotropic effects of different filling pressures and volumes have been reported for the systemic heart of Myxine glutinosa and E. stoutii (Johansen, 1960; Jensen, 1961; Bloom et al. 1963; Chapman et al. 1963), although the rate of the isolated perfused heart of E. cirrhatus did not change in response to increased filling pressure (Forster, 1989). The portal heart of M. glutinosa has been observed to accelerate when distended with blood (Johansen, 1960; Fänge et al. 1963).

Cardiac function curves published for fish hearts are the product of at least two muscular, valved chambers in series. The portal heart of the hagfish offers the opportunity to investigate the effects of filling pressure and afterload on the work performed by a heart that is effectively a single chamber, filled at low pressures (Davie et al. 1987).

New Zealand hagfish, Eptatretus cirrhatus Forster, weighing between 810 and 1700 g were used in the experiments. The hagfish were captured in baited traps off Motunau, North Canterbury, New Zealand. They were kept indoors in aquaria supplied with circulating sea water at approximately 16 °C.

Animals were anaesthetized in sea water containing MS222 (0.4 g l^-1 ) and benzocaine (0.4 g l^-1 ). The portal heart was exposed through a large ventral incision. The supraintestinal vein was cannulated with polyethylene tubing (PE200). This inflow cannula was attached to a Marriot bottle containing Ringer’s solution, allowing continuous perfusion of the portal heart as the portal vein was cannulated (PE200). For the recording of input and output pressures close to the inlet and outlet of the heart, a piece of PE10 tubing was inserted into the PE200 cannulae. This short piece of PE10 tubing was connected via PE90 tubing to the pressure transducers (see below and Franklin and Axelsson, 1994).

The head and tail of the hagfish were cut off and the part of the hagfish containing the perfused heart (approximately 20 % of total body mass) was placed in a thermostatted (16–17 °C) organ bath filled with Ringer’s solution. The inflow cannula was attached to a constant-pressure device (see Franklin and Axelsson, 1994) connected to a reservoir containing perfusion fluid. By raising and lowering the constant-pressure device, different input pressures could be achieved. In a similar way, the cannula inserted into the portal vein was connected to a tube which could be moved vertically, thus changing the afterload. The input and output pressures were recorded using a Devices Mx4 recorder and Bell and Howell type 4-327 pressure transducers. The pressure transducers were calibrated against a static water column, where zero pressure was set to the level of Ringer’s solution in the organ bath.

The Ringer’s solution had the following composition (g l^-1 ): 27.7 NaCl, 0.6 KCl, 0.75 CaCl₂.2H₂O, 0.75 MgSO₄.7H₂O, 1.83 MgCl₂.6H₂O, 0.405 NaH₂PO₄, 3.45 NaHCO₃, 1 glucose. Noradrenaline (3 nmol l^-1 ) and adrenaline (1 nmol l^-1 ) were added to provide physiological catecholamine levels (see Perry et al. 1993) in the perfusion fluid. The saline was bubbled with a gas mixture of nitrogen (93.5 %), oxygen (6 %) and carbon dioxide (0.5 %) (see Sundin et al. 1994).

Starling curves were constructed by increasing the preload gradually while maintaining the output pressure at 0.2 kPa. In a similar way, power curves were established by raising the output pressure incrementally until the cardiac output reached a plateau. The input pressure was kept constant at a physiological level (0.04 kPa). Cardiac output was recorded gravimetrically when it had stabilized (1–3 min).

Bolus injections of adrenaline (0.1 ml, 10₋₃ mol l^-1 ) into the inflow tube provided information on chronotropic effects in response to adrenaline. To study the chronotropic effect of the β-adrenoceptor antagonist sotalol, the drug was added to the Ringer’s solution and the effect was recorded 30 min later.

Starling and power curves were also determined following the addition of sotalol. Each preparation acted as its own control for investigating differences in performance after the addition of sotalol.

A ventral incision was made in anaesthetized animals to expose and allow cannulation of one of the small superficial veins situated on the intestine and emptying into the supraintestinal vein. The cannula was advanced into the supraintestinal vein, with its opening towards the portal heart, and firmly secured. After 24 h of recovery, recordings of the supraintestinal blood pressure at rest were made using a Devices Mx4 recorder and Bell and Howell type 4-327 pressure transducers. Calibration of the transducers was performed against a static water column.

Cardiac output, stroke volume and power output were normalised per gram portal heart mass (PHM). Separate curves, in which the different parameters were plotted against filling pressure (preload graphs) and output pressure (afterload graphs), were fitted to the data using a third-order polynomial (see Franklin and Axelsson, 1994). As measured pressures were slightly different in each preparation, this method made it possible to construct composite graphs for which data had been generated at set pressure intervals. Data were calculated and plotted in steps of 0.01 kPa for the preload curves and in steps of 0.02 kPa for the afterload curves.

Data are presented as means ± S.E.M. Wilcoxon matched-pairs signed-ranks test were used to determine statistically significant differences (P<0.05).

Data obtained from the portal heart preparation are summarised in Table 1. The portal heart mass was 0.0104±0.0004 % of body mass (N=11). The supraintestinal vein pressure varied between 0.025 and 0.07 kPa (N=8) (Fig. 1), with a mean pressure of 0.040±0.005 kPa (N=8). Portal heart rate at physiological input and output pressures was 40.8±1.1 beats min₋₁ (N=10). It was observed during the cannulation that when the membrane covering the heart was cut, the heart rate immediately increased and averaged 7.3±1.8 beats min₋₁ higher than before cutting (32.0±4.3 beats min^-1, N=8).

Increased fillin.g pressure caused a marked increase in cardiac output (Q̇), stroke volume (VS) and power output (Fig. 1). At an input pressure of 0.04 kPa, which was the mean venous pressure observed in vivo, VS was

1.5±0.2 ml beat₋₁ g₋₁ PHM (N=10). Maximum VS was rea. ched at 0.1 kPa and was 1.8±0.2 ml beat₋₁ g₋₁ PHM (N=10). Q̇ was 59.7±7.9 ml min₋₁ g₋₁ PHM (N=10) at 0.04 kPa and peaked at 0.1 kPa, measuring 87.5±16.7 ml min₋₁ g₋₁ PHM (N=10). Power output was 0.17±0.03 mW g₋₁ PHM (N=10) at 0.04 kPa and reached a maximum of 0.20±0.03 mW g₋₁ PHM (N=10) at 0.07 kPa. Heart rate (fH) increased between an input pressure of 0.05 kPa (40.8±1.1 beats min^-1 ) and 0.1 kPa (44.7±1.2 beats min^-1 ) (Fig. 2). No pressure-induced rate changes could be detected after sotalol treatment. Sotalol decreased heart rate significantly, but did not change the responses in VS and power output to increased filling pressure (Fig. 1).

Portal heart rate did not change sig. nificantly in response to increased output pressure. VS and Q̇ decreased significantly when afterload increased above 0.22 kPa, while power output peaked (0.16±0.02 mW g₋₁ PHM) at an output pressure of 0.22 kPa and then decreased. Sotalol pretreatment did not alter any of the responses to increased afterload (Fig. 3).

Adrenaline, administered as a bolus injection (0.1 ml, 10₋₃ mol l^-1 ), accelerated the portal heart. Mean heart rate was 42.0±1.5 beats min₋₁ before adrenaline injection and 45.5±1.6 beats min₋₁ (N=9; P<0.05) after treatment. Input and output pressures were kept constant at physiological levels, mean input pressure was 0.037±0.004 kPa and mean output pressure was 0.15±0.01 kPa (N=9). The heart was slowed down by addition of the β-adrenoceptor antagonist sotalol to the perfusate (10₋₅ mol l^-1 ). Post-sotalol mean heart rate was 36.6±1.6 beats min₋₁ (N=9).

The Starling curve for the portal heart (Fig. 1) indicates that, for the range of venous filling pressures measured, output is increased by 36 % owing to an increased stroke volume. Thus, if an increased venous return elevates central venous pressures, the portal heart is capable of responding with an increased output. However, its function is compromised at output pressures exceeding 0.22 kPa (Fig. 3). In the systemic heart of E. cirrhatus, the power output is maximal in the range 1.7–2.0 kPa (Forster, 1989). On a mass-specific basis, the peak power output of the portal heart (approximately 0.2 mW g₋₁ PHM) is not very different from that of the systemic heart (0.37 mW g₋₁ ventricle mass, Forster, 1989), although it operates at about a tenth of the afterload.

The maximum stroke volume of the portal heart, approximately 0.2 ml kg₋₁ body mass, is about 30 % of the maximum stroke volume of the systemic heart in this species (Forster, 1989). It is interesting to note that gut blood flow in teleosts represents approximately 30–40 % of the cardiac output (Axelsson et al. 1989; Axelsson and Fritsche, 1991). The hagfish is the only vertebrate with a specific venous pump supplying its liver. The modest fivefold increase in venous pressure downstream from the portal heart must provide the necessary conditions for adequate flow through the liver and back to the systemic heart. In elasmobranch and teleost fish, the blood pressure is high enough to drive the blood back to the heart without the assistance of a venous pump.

The portal heart rate increased by 10 % when the input pressure was raised from 0.05 to 0.1 kPa (Fig. 2). Swimming has been shown to result in an increase in systemic heart rate in E. cirrhatus, possibly by an augmented venous return (Forster et al. 1988). The venous pressures presented in this study were measured at rest (Table 1). The highest value recorded was 0.07 kPa, and it is possible that swimming can increase venous pressure to well above 0.1 kPa. The present results and observations by Johansen (1960) and Fänge et al. (1963) indicate that the portal heart is sensitive to preload. Intrinsic rate regulation has been shown for elasmobranchs (Jensen, 1970), but information concerning teleosts is limited (Farrell, 1984). Several investigations have indicated the existence of a pressure-sensitive pacemaker in hagfish heart (Johansen, 1960; Jensen, 1961; Bloom et al. 1963; Chapman et al. 1963; Axelsson et al. 1990). Jensen (1961) concluded that diastolic volume governs acceleration in the systemic heart ventricle of E. stoutii. In contrast, the rate of the isolated perfused systemic heart ventricle of E. cirrhatus was not sensitive to changes in preload (Forster, 1989). A possible explanation for these contradictory results is that the preparation used by Forster did not include the sinus venosus, which is the pacemaker region in vivo (Davie et al. 1987). Considering the relatively low venous pressure of the hagfish, it is likely that earlier methods did not allow the monitoring of rate changes in response to small elevations in input pressure within the physiological range.

There was an increase in fH of about 7 beats min₋₁ when a layer of tough membrane covering the heart was cut and the portal heart could distend without the restraint of the tissue. This is further evidence for the existence of a pressure-sensitive pacemaker. Cutting the membrane possibly allowed the heart to distend more than it would in vivo, emphasizing the response to a rise in preload.

Portal heart rate decreased by 13 % when sotalol was added to the perfusion fluid (Fig. 2). Axelsson et al. (1990) showed that the systemic heart of M. glutinosa responded to sotalol with a decrease in heart rate and concluded that endogenous catecholamines are important for the normal activity of the heart. That conclusion was further strengthened by the observation that adrenaline caused an increase in systemic heart rate and stroke volume in vivo. In the present experiment, injection of 0.1 ml of 10₋₃ mol l₋₁ adrenaline prior to sotalol treatment increased portal heart rate by 8 %. Our results show that portal heart rate in hagfish is tonically stimulated by release/leakage of endogenous catecholamines that act on β-adrenoceptors, but is still able to respond to stimulation. A differential distribution of adrenaline and noradrenaline in the walls of the portal heart and systemic heart atrium and ventricle (von Euler and Fänge, 1961) indicates that control of the two hearts may be complex (Forster et al. 1991).

This study was supported by the Swedish Research Council and the Swedish Forestry and Agricultural Research Council.