Intrinsic autoregulation of cardiac output in rainbow trout(Oncorhynchus mykiss) at different heart rates

Intrinsic regulation of the teleost heart is greatly dependent on central venous pressure, which exerts a modulatory role on stroke volume according to the well-known Frank-Starling mechanism. This mechanism is well characterized in situ (Farrell et al.,1982; Farrell,1991) and results from perfused heart studies have been extrapolated to cardiac function in the intact animal.

Heart perfusion studies, however, operate at rates well above the in vivo range of resting heart rate(Altimiras and Larsen, 2000)because the inhibitory cholinergic influence on the heart is abolished. At the lower in vivo heart rates, the loading conditions of the heart will differ due to a longer filling time(Farrell and Jones, 1992),which in turn is expected to increase venous pressure.

This prediction has not been demonstrated experimentally due to technical limitations in recording pressure in the sinus venosus(Psv) in teleosts. Such a study, however, could shed light on two separate aspects of the cardiac physiology of teleosts. First, it could explain the shift from vis-à-fronte atrial filling to vis-à-tergo filling(Farrell, 1991). The unique vis-àfronte filling mechanism observed in some fish species relies on the generation of negative intrapericardiac pressures associated with the elastic recoil of the ventricle during diastole. This mechanism differs from vis-à-tergo filling, which depends on the build-up of pressure in the central veins. Second, it could explain the role of venous pressure in modulating cardiac output through the interdependence between heart rate and stroke volume.

The aim of the study was to validate a new technique for measuring Psv, and to study Psv in trout at varying heart rates. Since truly resting heart rates were unattainable, a novel pharmacological approach to manipulate heart rate using the bradycardic agent zatebradine was developed and validated.

Rainbow trout Oncorhynchus mykiss Walbaum were obtained from a local fish farm. Fish were kept in 2 m³ tanks at 15°C in the semi-open freshwater circulation system of the department and fed twice a week.

Eight fish (580±53 g body mass) were used. The fish were anaesthetized in a solution of MS-222 (100 mg l^-1; Sigma) buffered with sodium bicarbonate (200 mg l^-1) until breathing movements ceased, and placed ventral side up on an operating sling. The gills were continuously irrigated with aerated water containing a diluted solution of MS-222 (75 mg l^-1). A polyethylene cannula (PE50, Clay Adams;Becton Dickinson, Sparks, MD, USA) filled with heparinized (100 i.u. ml^-1) 0.9% NaCl was implanted into the dorsal aorta to measure blood pressure (Pda) as previously described(Hughes et al., 1983). The cannula was secured with a suture on the back of the animal. In order to measure cardiac output Q̇ (=ventral aortic blood flow), the ventral aorta was exposed through an incision on the left side of the isthmus. A cuff-type pulsed Doppler flow probe (Iowa Doppler Products, Iowa City, IA, USA) with an internal diameter of 2.2-3.0 mm was placed around the ventral aorta and the lead from the probe was secured with two skin sutures.

After surgery, the animals were transferred to a holding chamber and allowed to recover for 24-36 h. The flow probes were connected to a Directional Pulsed Doppler Flowmeter (Model 545-4C, Iowa University, USA). The dorsal aortic cannula was connected to a Statham P23 (Hato Rey, Puerto Rico)pressure transducer connected to a bridge amplifier channel of a recorder(Grass Instruments, Quincy, MA, USA). The pressure transducer was calibrated against a static water column. Heart rate (fh) was obtained from the phasic blood flow signal using a Grass tachograph (model 7P44D).

The flow, pressure and tachograph signals were fed into a computer and stored to disk at 10 Hz using a custom-made program (Labview v.5.1, National Instruments, Austin, TX, USA).

The experimental protocol consisted of a control period followed by serial injections of zatebradine hydrochloride (Boehringer Ingelheim,Skärholmen, Sweden), corresponding to cumulative doses of zatebradine of 0.5, 1, 2, 4, 6 and 8 mg kg^-1. A 30 min recording period was taken post-injection. The main decrease in heart rate occurred within the first 10 min post-injection.

Six fish were used (450-550 g in body mass). Fish were killed by a sharp blow to the head and the heart was rapidly excised from the animal and placed on a chilled Petri dish, where three longitudinal strips under 1 mm width were obtained from the ventral ridge of the pyramidal heart.

The strips were mounted for measurements of twitch force and rate of contraction using force transducers (Grass, FT-103) connected to a 4-channel bridge amplifier unit (Somedic, SenseLab 4CHAMP, Hörby, Sweden). The signals were stored digitally at 20 Hz for further analysis using a custom-made program (LabView 5.1, National Instruments). The Ringer solution was identical to the one used in the heart perfusion experiments and was also bubbled with 80 kPaO₂/balance N₂.

The strips were stretched to 90% of maximal contraction, paced at 0.2 Hz throughout the experiment and allowed to stabilize for 1 h. Following a 20 min control recording, the Ringer solution was quickly changed and the first concentration of zatebradine was tested. The three strips were run simultaneously, one being used as control and the other two as treatments. The concentrations of zatebradine tested were 0.5, 1, 2, 5 and 10 mg l^-1.

Ten fish, mass 300-500 g, were used in the experiment. The animals were instrumented with a dorsal aortic catheter for measurement of heart rate and drug injection as described for Series I. Following this procedure, a catheter was non-occlusively inserted in the left ductus of Cuvier (LDC) and forwarded to the sinus venosus. The surgical procedure was as follows. The operculum and the gills were retracted to expose the gill-free V^th branchial arch and a 1 cm incision was made parallel to this arch (see Fig. 1 for a graphical view). The incision was initiated on top of the cleithrum bone and was followed towards the ventral edge of the bone. The LDC was exposed at this location. It was noticed that the motor nerve to the left pectoral fin runs on top of the LDC and this was used as an anatomical landmark. A loose pocket of the LDC wall was gently pulled and held with 3-0 suture thread(Fig. 1C). A loose thread was placed downstream and a small cut was made between the threads. A PE-50 catheter (with a bubble 35 mm from the tip) was inserted in the hole and advanced to the sinus venosus (15-20 mm) before secured tightly in place. The cannula was secured twice on the skin, once adjacent to the bubble on top of the cleithrum bone and once close to the dorsal fin. The position of the catheter tip was verified postmortem.

This new procedure is similar to those previously used to measure the pressure in the ducts of Cuvier (Olson et al., 1997; Minerick et al.,2003), but allows the measurement of pressures in the sinus venosus.

Fish were allowed to recover for 24 h before the effects of three doses of zatebradine (1, 2 and 4 mg kg^-1) on heart rate and venous pressure were recorded.

Nine fish were used (505±24 g body mass, 469±31 mg wet ventricular mass). Fish were killed by a sharp blow to the head and transferred to an operating sling. Heparin (1 ml kg^-1) was injected via the caudal vessels to prevent clotting of the blood during surgery. The heart was perfused in situ following previous protocols with a few variations (Farrell et al.,1986). Briefly, the abdominal cavity was opened, a double-bore cannulae (as shown in Franklin and Axelsson, 1994) was inserted in the sinus venosus through a hepatic vein and the perfusion of Ringer started from a reservoir. Other hepatic veins, if present, were ligated to prevent leakage. The gill arches were cut and Ringer flowed freely through the heart with each heart beat. The isthmus was cut, exposing the ventral aorta and the afferent branchial arteries, and a double bore cannula was inserted through the ventral aorta into the bulbus arteriosus. The entire fish was then transferred to a constant temperature stainless steel trough (15°C) filled with 0.9% NaCl. The cannula to the atrium was connected to two water-jacketed reservoirs with Ringer solution via a constant pressure device. The ventral aortic cannula was connected to an output pressure head. The composition of the perfusion solution was (in mmol l^-1): 104.1 NaCl; 3.1 KCl; 0.9 MgSO₄; 2.5 CaCl₂. The buffer system was a Hepes-Tris mixture (5 mmol l^-1 Hepes adjusted to pH 7.8 with Tris to a final concentration in the perfusate of 5 mmol l^-1). The following metabolic substrates were added: glucose, glutamate, fumarate and pyruvate (5 mmol l^-1 each). The perfusate was gassed with 80 kPa O₂/balance N₂ using a Gas Mixing Flowmeter (model GF-3MP, Cameron Instruments, Port Aransas, TX, USA).

A tonic adrenergic stimulation (5 nmol l^-1 adrenaline bitartrate; Sigma) was maintained by pumping a 2.5 μmol l^-1adrenaline-Ringer solution through a side port of the inflow tubing to the heart using a peristaltic pump (Minipuls 3, Gilson, Villiers le Bel, France). It has been established that tonic adrenergic stimulation is essential for the long-term viability of perfused hearts(Graham and Farrell, 1989). The flow rate of the pump was automatically adjusted to follow changes in cardiac output so that the dilution rate of the adrenaline solution was constant (1:500). The same system was used to deliver a constant zatebradine concentration (1 mg l^-1) to the heart.

Cardiac output was measured using an in-line flow probe (Transonic 2N,Ithaca, NY, USA; 2 mm internal diameter) coupled to a transit-time flowmeter(Transonic, T106). Preload and afterload pressures were measured using disposable pressure transducers (DPT6100, Peter von Berg Medizintechnik,Kirchseeon/Eglharting, Germany) connected to a 4-channel bridge amplifier unit(SenseLab 4CHAMP). Pressures were calibrated against static water columns and the calibration checked every 30 min. The signals were recorded on a Grass recorder (model 7WU), which also logged instantaneous heart rate using a tacograph unit (Grass, 7P44). In parallel, the data was digitally stored using a custom made program (LabView 5.1, National Instruments). Preload and afterload pressures, cardiac output and heart rate were averaged for 5 s and stored.

The experimental protocol consisted of a 30 min stabilization period,followed by a recording period, during which zatebradine-free perfusate(Control group, N=4) or perfusate with 1 mg l^-1zatebradine (ZAT group, N=5) was supplied to the heart. The zatebradine dose was chosen after preliminary experiments. Larger doses (2, 3 and 4 mg l^-1) elicited faster changes in fh but were unsuitable for prolonged exposure because the heart became arrhythmic before reaching a heart rate of 30 beats min^-1, which was the lowest target fh intended in the experiment.

During the stabilization period, afterload was set to 5 kPa and cardiac output was adjusted to 30 ml min^-1 kg^-1viachanges in preload pressure. In the ZAT group, preload pressure was allowed to change as heart rate decreased following administration of zatebradine. The perfused hearts in the Control group were also subjected to the zatebradine after the 2 h control trial.

In Series I, stroke volume (Vs) was calculated as the ratio between Q̇ and fh. Systemic vascular resistance(R_sys) was calculated as Pda divided by Q̇, assuming that venous pressure was zero and did not change during the experimental protocol. It was also assumed that no significant changes in blood viscosity took place during the experiment. Q̇, Vhand R_sys are presented as percent changes from the control value.

The dose-effect relationship of zatebradine (ZAT) on fhresembles that of a Michaelis-Menten enzymatic reaction if the effects are plotted as the decrease in fh from control values. Thus, a Levenberg-Marquardt non-linear fitting procedure was used to adjust the fhversus [ZAT] data for each animal to the Michaelis-Menten equation to obtain the concentration of zatebradine that induces half maximal effects on fh, referred here as[ZAT]_0.5.

Control values of the different cardiovascular parameters were subsequently compared against those values at [ZAT]_0.5 for normalization purposes. If the [ZAT]_0.5 deviated more than 10% of the closest nominal [ZAT] used in the study, the values for [ZAT]_0.5 were linearly interpolated between the two values closest to [ZAT]_0.5.[ZAT]_0.5 was in the range 1.5-3 mg l^-1. Since the window around K_m in the Michaelis-Menten equation is highly linear, the error induced by interpolation is minimized.

In Series II, twitch force and maximal rate of contraction were determined as previously described (Hove-Madsen and Gesser, 1989).

All data are presented as means ± s.e.m.

Wilcoxon's signed-ranks test for paired samples (two-tailed) were used to evaluate the statistical significance of control and [ZAT]_0.5cardiovascular variables in Series I and III and the cardiac variables at different heart rates in Series IV. A Kruskal-Wallis non-parametric paired test was used to compare the effects of zatebradine in Series II. Asterisks in the figures indicate significant changes between paired samples. In the case of repeated tests, a modified Bonferroni procedure was used to reduce the risk of discarding a true null hypothesis(Holm, 1979).

Zatebradine is an effective bradycardic agent in vivo. From a control value of fh=44.4±2.3 beats min^-1(N=14), a dose of 2 mg kg^-1 resulted in a 41% decrease in fh. The effects plateaued at 10.4±2.7 beats min^-1 at the highest dose used (10 mg kg^-1). This fh is outside the physiological range of fh for trout at 15°C(Fig. 2). [ZAT]_0.5was 2.79±0.47 mg l^-1.

The effects of zatebradine were highly significant. At [ZAT]_0.5, fh decreased by 50%, Pda decreased by 14% and Q̇ by 15%, while Vs increased by 66% and R_sys remained unchanged (Figs 2 and 3).

The application of zatebradine to ventricular strips did not show any significant changes on twitch force or on the maximum rate of force development at any of the doses employed(Table 1).

The stepwise decrease in heart rate induced with zatebradine was coupled to a progressive increase in venous pressure. At a control fhof 58.3±3.5 beats min^-1 (N=10), venous pressure was subambient (-0.06±0.04 kPa) and this increased significantly to-0.02±0.04 kPa and 0.07±0.05 kPa after injection of zatebradine(2 mg kg^-1 and 4 mg kg^-1, respectively)(Fig. 4).

The control experiments revealed little deterioration of fh and P_in in the perfused heart after 2 h (Fig. 5). P_in increased from -0.023±0.013 kPa (N=4)to 0.003±0.020 kPa and fh decreased from 69±4 beats min^-1 to 59±3 beats min^-1. Q̇ decreased from 2.49±0.17 mW g^-1 28.1±1.2 ml min^-1 kg^-125.6±1.3 ml min^-1 kg^-1, Vsincreased from 0.41±0.03 ml kg^-1 to 0.44±0.03 ml kg^-1 and power output decreased from 2.52±0.18 mW g^-1 to 2.27±0.17 mW g^-1 (not shown). In comparison, zatebradine-treated preparations (1 mg kg^-1, N=7) showed a significant change in preload and heart rate after 90 min perfusion. The decrease in fh was almost linear down to 40 beats min^-1 and slowly leveled thereafter at 28.3±2.5 beats min^-1 (Fig. 5).

A small but significant change in Q̇and power output of the heart occurred at a heart rate of 30 beats min^-1 in comparison to 60 beats min^-1 (control)(Fig. 6). Q̇ decreased from 29.4±0.2 ml min^-1 kg^-1 (N=7) to 26.8±0.5 ml min^-1 kg^-1 and power output decreased from 2.74±0.21 mW g^-1 to at fh=30 beats min^-1 to and fh=60 beats min^-1,respectively. At the same time, P_in increased significantly from 0.000±0.007 kPa (N=7) to 0.053±0.013 kPa and Vs increased from 0.49±0.00 ml kg^-1 to 0.88±0.02 ml kg^-1(Fig. 6).

The results of the study indicate that zatebradine is a useful cardiac chronotropic molecule with which to manipulate heart rate in fish, as previously shown in mammals. Direct evidence of the link between heart rate and venous pressure is also presented. Lower heart rates attained by zatebradine injection cause an increased pressure in the sinus venosus that,in turn, increases stroke volume through the Frank-Starling mechanism. Such compensation between heart rate and stroke volume contributes to the short-term modulation of cardiac output in vivo.

Zatebradine, a bradycardic agent, was tested as a means of manipulating heart rate within its physiological range(Altimiras and Larsen, 2000). Zatebradine-induced bradycardia occurs as a result of the inhibition of the hyperpolarization-activated current in pacemaker cells, as shown in different mammalian species (Kobinger and Lillie,1984; Schipke et al.,1991).

The suitability of zatebradine as a chronotropic agent also requires that it has no inotropic effects and no effect on peripheral resistance. As shown in Fig. 2, a dose of zatebradine of 2.79±0.47 mg l^-1 decreased heart rate by half the maximum change, which compares well with similar experiments in mammals(Franke et al., 1987; Kalman et al., 1995; Ryu et al., 1996; Schipke et al., 1991). Thus,zatebradine is also a bradycardic agent in trout, indicating that the hyperpolarization current (I_f) is also a component of the pacemaker currents in nodal cells in this species.

The results from the Series II experiments indicate that zatebradine has no significant effect on the peripheral vasculature because systemic resistance is unchanged (Fig. 3). The significant decrease in dorsal aortic pressure is due to the concomitant decrease in cardiac output when heart rate decreases to half the maximum effect (Fig. 3).

Finally, no significant inotropic effects of zatebradine were found in isolated ventricular strips (Table 1), as was the case for mammals(Chen and Slinker, 1992).

Altogether these results indicate that zatebradine is an appropriate pharmacological tool to manipulate heart rate without other significant cardiovascular changes and as such, is well suited to study both in vivo and in vitro how heart rate affects intrinsic cardiac regulation.

Pressure in the sinus venosus in control conditions was negative in seven out of ten fish, with a mean value of -0.06±0.04 kPa(Fig. 4). To our knowledge these are the first pressure measurements in the sinus venosus of a teleost fish and directly corroborate the predictions based on data from perfused hearts (Farrell and Jones,1992). Values for central venous pressure in trout obtained in other studies indicate values slightly above ambient, averaging 0.37 kPa in the Cuverian ducts (at an average heart rate of 62.8 beats min^-1 at 12°C; Olson et al., 1997)or 0.19 kPa in the common cardinal vein (at an average heart rate of 32 beats min^-1 at 10°C; Kiceniuk and Jones, 1977). Collectively, these values portray a pressure gradient from slightly above ambient in the central veins to slightly subambient in the sinus venosus that could support the operation of the vis-à-fronte mechanism for atrial filling(Farrell and Jones, 1992).

Vis-à-fronte filling is allegedly associated with resting conditions and low stroke volumes in active teleosts because filling pressures above ambient (vis-à-tergo filling) are required for normal and elevated stroke volumes (Farrell and Jones, 1992). In resting conditions, however, heart rates are considerably lower. Recent studies in trout using surgery-free methods indicate that true resting heart rates at 15°C are in the order of 30 beats min^-1 (Altimiras and Larsen, 2000), well below the control fh value obtained in this study (58.3±3.5 beats min^-1). Such low heart rates could not be attained in surgically instrumented animals, so zatebradine was used instead. Lowering the heart rate lengthens the filling time and Psv increases (0.07±0.05 kPa at fh=24.6±3.7 beats min^-1; Fig. 4), which questions the existence of negative pressures in the sinus venosus in resting conditions and, at the same time, casts doubts on the in vivo relevance of vis-à-fronte atrial filling. Minerick et al.(2003) have recently reached the same conclusion by demonstrating a dynamic coupling between venous pressure and cardiac output.

The rise of Psv associated to the progressive decrease in heart rate from the Series III experiments also provided a mechanistic explanation of the autoregulation of cardiac output observed in vivo. Thus, the small decrease in Q̇ down to 87.5±4.6% when fh was halved with zatebradine would be explained by a rise in Psv that would allow a concomitant increase in Vs to 165±13%(Fig. 3), according to the Frank-Starling relationship. Complete compensation was not attained, perhaps because maximal stroke volume was reached(Forster and Farrell,1994).

The mechanism of cardiac output autoregulation related to variations in heart rate and filling pressure (denominated time-dependent autoregulation in the rest of the Discussion) was further verified in perfused hearts with the simultaneous measurement of Vs and P_inat varying heart rates. As shown in Fig. 6, the zatebradine-mediated drop in heart rate is coupled to a simultaneous increase in P_in and Vssimilar in magnitude to the in vivo values.

Time-dependent autoregulation enhances the repertoire of intrinsic cardiac regulation to include the effects of varying filling time. Strictly speaking,the mechanism is dependent on the extrinsic neurohumoral modulation of heart rate, but its compensatory role is exerted through changes in stroke volume by riding on the Starling curve of the heart. The impact of time-dependent autoregulation, however, is limited by the interaction with other mechanisms regulating venous pressure and stroke volume. Adrenaline, for instance, is known to increase venous pressure (Zhang et al., 1998).

At high heart rates, the low venous pressures expected from short filling times are likely to be counteracted by the positive inotropic effect related to adrenergic activation. This prediction needs to be confirmed experimentally, but it is already known that energetically demanding conditions such as exercise require elevated cardiac outputs achieved by increasing heart rate and stroke volume simultaneously, and this is incompatible with subambient or low venous pressures(Farrell et al., 1996). In the absence of adrenergic estimulation, a reciprocal relationship between fh and Vs has already been shown in paced trout hearts in vitro(Farrell et al., 1989). Besides the possibility that short filling times limit stroke volume, a reduced force of contraction resulting from the negative staircase effect was also speculated (Farrell and Jones,1992).

At low heart rates, a longer filling time promotes an increase in venous pressure that results in atrial distension. In turn, atrial stretch triggers the release of atrial natriuretic factor (ANF)(Cousins and Farrell, 1996),which might break down the reciprocal fh-Vs relationship because ANF is a potent vasodilator that lowers venous pressure(Olson et al., 1997).

Such a cardioprotective role of ANF(Farrell and Olson, 2000) is not incompatible with the time-dependent autoregulatory mechanism proposed in this study, which is targeted at a more physiological range of sinus venosus pressures. ANF release in freshwater- and seawater-acclimated perfused trout hearts is evident at filling pressures above 0.1 kPa(Cousins et al., 1997), while the results presented indicate the operation of the time-dependent autoregulatory mechanism from subambient to pressures slightly below 0.1 kPa.

List of symbols
 
• fh

  heart rate

 
• Mv

  ventricular mass

 
• Pda

  pressure in dorsal aorta

 
• P_in

  preload pressure

 
• P_out

  afterload pressure

 
• Psv

  pressure in sinus venosus

 
• Q̇

  cardiac output

 
• R_sys

  systemic vascular resistance

 
• Vs

  stroke volume

We acknowledge the help of Gunilla Rydgren, Karin Enarsson and Anne-Maj Samuelsson. Boehringer Ingelheim kindly supplied the zatebradine used in the study. J.A. was supported by a TMR-EU grant (contract ERBFMBICT982940).