Spatial uniformity of action potentials indicates base-to-apex depolarization and repolarization of rainbow trout (Oncorhynchus mykiss) ventricle

Coordinated contraction of the vertebrate heart is determined by the propagation of action potentials (APs) through all cardiomyocytes of the heart. In fish hearts, electrical excitation is initiated in the sinoatrial (SA) pacemaker at the border zone between the sinus venosus and the atrium (Haverinen and Vornanen, 2007; Jensen, 1965; Mackenzie, 1913; Saito, 1969; Saito, 1973; von Skramlik, 1935; Yamauchi and Burnstock, 1968). In the atrial wall, APs propagate quickly along strands of working cardiomyocytes to the atrioventricular (AV) canal, where the velocity of AP propagation slows down, leaving sufficient time to fill the ventricle with blood (Irisawa, 1978; Sedmera et al., 2003). The cycle of electrical excitation of the heart is terminated by a rapidly advancing ventricular AP, which triggers ventricular myocytes to contract almost simultaneously. The ventricles of endothermic hearts have a specialized endocardial conducting system that originates from the branches of the His bundles and ramifies as Purkinje fibers among the working ventricular myocytes (Gourdie et al., 1993; Jensen et al., 2012; Sedmera and Gourdie, 2014; Szabóa et al., 1986). Unlike mammalian and avian hearts, a histologically specialized conduction path has not been unequivocally demonstrated in fish hearts, although some functional studies suggest the existence of an endocardial conduction pathway in the ventricle of zebrafish (Danio rerio) and African lungfish (Protopterus ethiopicus) (Arbel et al., 1977; Sedmera et al., 2003).

The rhythmic generation of APs and variable rate of AP propagation through the heart are due to the specialized ion current/channel compositions that produce functionally specific APs for each tissue compartment (Chandler et al., 2009; Greener et al., 2009; Greener et al., 2011; Hassinen et al., 2021; Abramochkin et al., 2022). In general, SA and AV nodes, His bundles and Purkinje fibers are histologically identifiable from the adjacent working myocardium with characteristic cellular composition, size and structure, arrangement and relative abundance of connective and muscular tissues, or immunohistochemically by specific molecular markers (Anderson et al., 2009; Boyett, 2009; Jensen et al., 2012; Waller et al., 1993a,b). In mammalian and avian ventricles, APs proceed from the apex to base and from endocardium to epicardium, owing to the His–Purkinje system (Autenrieth et al., 1975). The spread of activation in the fish ventricle is still poorly understood, and it is unclear whether there is a specialized ventricular conduction pathway that quickly transmits APs from the AV canal to the apex of the ventricle. Consistent with the putative existence of a ventricular conduction pathway, apex-to-base depolarization of the zebrafish ventricle has been reported previously (Sedmera et al., 2003; Zhao et al., 2021). However, Jensen et al. (2012) reported that depolarization of the zebrafish ventricle proceeds from base to apex. The reason for the conflicting results is currently unclear. In other studied fish species, including the rainbow trout (Oncorhynchus mykiss), the main vector of activation has been reported to go from base to apex (Kibler et al., 2021; Noseda et al., 1962; Vaykshnorayte et al., 2011, 2018).

Because the wave of depolarization (activation) reaches each myocyte at different times, the sequence of ventricular repolarization must be determined by the activation time and the duration of APs at each locus of the ventricular myocardium. Indeed, in mammalian hearts, there are transmural and apicobasal differences in AP duration and underlying ion currents, which affect the repolarization pattern of the ventricles (Bryant et al., 1997; Cheng et al., 1999; Dean and Lab, 1990; McKinnon and Rosati, 2016; Shipsey et al., 1997; Szentadrassy et al., 2005). It is not known, however, whether there is a similar regional variation in AP morphologies and ion channel/current distribution in the fish ventricle. Because the direction of depolarization (i.e. local activation time) and spatial differences in the durations of ventricular APs set limits for possible patterns of ventricular repolarization, we measured cardiac axis (i.e. direction of depolarization) and regional distribution of AP durations to deduce the pattern of ventricular repolarization and the potential presence of a special ventricular conduction pathway in rainbow trout ventricle. Based on significant histological differences in the ventricular structure of the fish (compact versus spongy myocardium), it was hypothesized that there are spatial heterogeneities of electrical activity between the ventricular regions.

Adult rainbow trout [Oncorhynchus mykiss (Walbaum 1792), male and female, mean±s.e.m. body mass=24.60±1.45 g, n=41] were obtained from the local fish farm (Kontiolahti, Finland). In the animal facilities of the University of Eastern Finland (Joensuu), the fish were reared in 500-liter aquaria with a computer-controlled temperature regulation system (Computec, Joensuu Finland). The temperature of the aquaria was regulated at 12°C and aerated water (O₂ 9.7 mg ml⁻¹) flowed through them. The photoperiod was 12 h:12 h light:dark. The fish were fed five times a week (Ewos, Turku, Finland). Experiments were conducted with the permission of the national animal experimental board in Finland (permissions STH252A and ESAVI/8877/2019).

Each rainbow trout was stunned by a blow to the head and killed by pithing. The heart was quickly excised and rinsed in Ca²⁺-free low-Na⁺ solution containing (in mmol l⁻¹): 100 NaCl, 10 KCl, 1.2 KH₂PO₄, 4 MgSO₄, 50 taurine, 20 glucose and 10 HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid] with pH adjusted to 6.9 at 20°C with KOH. Ventricular myocytes were isolated with enzymatic digestion method described previously in detail (Vornanen, 1997, 1998). First, the heart was perfused for 7 min with Ca²⁺-free low-Na⁺ saline and then enzymatically digested by a 15-min perfusion with the same solution containing collagenase (Type IA; 0.75 mg ml⁻¹), trypsin (Type IX; 0.5 mg ml⁻¹) and fatty-acid-free bovine serum albumin (BSA; 0.75 mg ml⁻¹, all Sigma Aldrich, St Louis, MO, USA). Both solutions were oxygenated with 100% O₂, and the enzyme solution was recycled using a peristaltic pump. After enzymatic digestion, the ventricle was excised and four different regions, the spongy layer and three parts (apex, middle region and base) from the compact layer, were carefully separated as schematically indicated in Fig. 1A. Each part was minced with scissors into small pieces in fresh Ca²⁺-free low-Na⁺ solution and dissociated into single myocytes by agitating them through the opening of a Pasteur pipette. Isolated myocytes were stored at 5°C in Ca²⁺-free low-Na⁺ solution and used within 8 h after isolation.

Ventricular APs were recorded in the current clamp mode of the whole-cell patch-clamp as described earlier (Badr et al., 2017). The compositions of internal (pipette) and external (bath) solutions are shown in Tables 1 and 2. Patch pipettes were pulled from borosilicate glass (King Precision, Claremont, CA, USA) and when filled with the intracellular saline solution, the mean±s.e.m. electrode resistance was 2.92±0.07 MΩ (n=93). The mean±s.e.m. capacitive size of ventricular myocytes was 31.23±0.80 pF (n=93). The temperature of the external solution was regulated to 12°C using a Peltier device (HCC-100A, Dagan, MN, USA). Myocytes were paced at the frequency of 1 Hz. The digitized data were recorded using Clampex 9.2 software (Axon Instruments, Saratoga, CA, USA) and the recordings were analyzed using the Clampex 10.4 software package. The following AP parameters were analyzed off-line: resting membrane potential (V_rest; mV), threshold potential for AP initiation (TP; mV), critical depolarization (CD=TP−V_rest; mV), AP overshoot (OS; mV), AP amplitude (Amp; mV), AP duration 50% repolarization level (APD₅₀; ms), the maximum rate of AP upstroke (+dV dt⁻¹; mV ms⁻¹) and the maximum rate of AP repolarization (−dV dt⁻¹; mV ms⁻¹; Fig. 1B).

Ion currents were recorded in the voltage clamp mode of the whole-cell patch-clamp as described earlier in detail (Badr et al., 2017, 2018). Four major ion currents that regulate the shape of ventricular AP in rainbow trout ventricular myocytes were measured: the inward rectifier K⁺ current (I_K1), the fast component of the delayed rectifier K⁺ current (I_Kr), the L-type Ca²⁺ current (I_CaL) and the Na⁺ current (I_Na). The compositions of internal and external solutions used for ion current measurements are shown in Tables 1 and 2. When patch pipettes were filled with the intracellular saline solution, the mean±s.e.m. pipette resistance was 3.20±0.04 MΩ (n=218). The mean±s.e.m. capacitive size of ventricular myocytes was 38.86±0.69 pF (n=218). Experimental temperature was 12°C. The size of ion currents is given as current densities (pA pF⁻¹).

Total RNA was extracted from the same regions of the ventricle, where myocytes were isolated for patch-clamp recordings (apex, middle and base of the compact myocardium and the spongy myocardium, n=6) by TRI Reagent Solution (Thermo Fisher Scientific, Vilnius, Lithuania). Transcript levels were determined in triplicate from each sample using Maxima SYBR Green qPCR Master Mix (Thermo Fisher Scientific), specific primers and AriaMx Real-Time PCR System (Agilent Technologies, Santa Clara, CA, USA), as reported previously (Hassinen et al., 2021). The mRNA expression of the studied genes was normalized to the transcript abundance of reference gene DnaJA2 using the ΔC_t method (Hassinen et al., 2015).

Electrocardiograms (ECGs) were recorded in vivo from mildly anesthetized fish (n=11) in the aquarium room (16°C) as previously described (Badr et al., 2016; Vornanen et al., 2014). ECG recordings were analyzed off-line using LabChart 7.1 software (ADInstruments). Fish were anesthetized with neutralized tricaine methanesulfonate (MS-222; 0.3 mg l⁻¹; Sigma Aldrich) for less than 3 min. When the fish was completely immobile and did not react to handling, it was gently placed on a damp sponge with the abdomen facing up. Aerated water from a 500-liter aquarium (O₂ tension ∼9 mg l⁻¹) was continuously administered to the gills through an oral tube. To define the electrical axis of the heart (the major direction of the overall electrical activity of the heart in the frontal plane), three thin steel electrodes (7-strand Teflon-coated wire, 0.23 mm in diameter; A-M Systems, Carlsborg, WA, USA) were inserted under the ventral skin of the fish to mimic the three limb leads of the Einthoven's triangle (Fig. 4A). For lead I (0 deg), the bipolar electrodes were oriented along the horizontal axis of the heart at the same level (slightly above lower edge of the opercula), with the negative (−) electrode on the right and the positive (+) electrode on the left. For leads II and III, the bipolar electrodes were oriented along the longitudinal axis of the heart. In lead II (+60 deg), the negative electrode was positioned at the ventricular base on the right and the positive electrode close to the apex of the heart on the left. In lead III (+120 deg), the negative electrode was positioned at the ventricular base on the left and the positive electrode close to the apex of the heart on the right (Dupre et al., 2005). In all leads, the reference electrode was inserted near the anus.

In ECGs, a wave of depolarization traveling toward the positive electrode displays a positive voltage on the ECG tracing. Thus, in the case of ventricular depolarization (QRS complex), if the R wave is larger than the S wave, then depolarization propagates towards the positive electrode, but if the S wave is larger than the R wave, then depolarization propagates away from the positive electrode. If R and S waves are equal in size, then depolarization spreads perpendicular to the vector of that lead (isoelectric). The electrical axis of the fish ventricle can be obtained as a sector overlap in the circle formed by the vectors of the three leads. It is noteworthy that if the T wave (repolarization of the ventricle) is concordant with the QRS complex, i.e. it has the same polarity as the QRS wave, then repolarization occurs in the opposite direction to depolarization in that lead (Dössel et al., 2021; Janse et al., 2012). If QRS and T waves are discordant, then repolarization occurs in the same direction as depolarization.

The results are represented as means±s.e.m. Normality of distribution was tested using Shapiro–Wilk test, and if the data failed the assumption of normality, logarithmic transformations were made. After checking the equality of variances, statistical comparison of mean values of APs, ion currents, mRNA amounts and ECG parameters were performed using one-way ANOVA. Tukey's and LSD (in the case of equal variances) or Dunnett's T3 (unequal variances) post hoc tests were used for paired comparisons. Differences were considered statistically significant when P<0.05.

APs were measured in enzymatically isolated ventricular myocytes from four different regions (spongy myocardium; and apex, base and middle region of the compact myocardium) (Fig. 1A). APs from the spongious myocardium and different regions of the compact ventricular myocardium had similar shapes (Fig. 1B). V_rest, TP, CD, OS, −dV dt⁻¹ and APD₅₀ were not different between the different regions (P>0.05) (Fig. 1C–E). Amplitude of the AP was 5.8% higher in the spongiosa compared with the apex (P<0.05), and the maximum rate of depolarization (+dV dt⁻¹) was slightly higher (22.3–23.7%) in the middle region and spongiosa than in the base and apex (P<0.05; Fig. 1F).

I_K1 was elicited from the holding potential of −80 mV by voltage ramps from +60 to −120 mV at the frequency of 0.2 Hz using the established voltage protocol (Fig. 2A). There were no differences in I_K1 density between different regions of the ventricle either for the inward current (IC) or the outward current (OC) (Fig. 2B). The I_K1 is produced by various Kir2 channels, of which Kir2.4, Kir2.2b and Kir2.1a account for approximately 98% of the I_K1 in rainbow trout ventricular myocytes (Hassinen et al., 2021). Consistent with the I_K1 density, there were no differences in the expression of the three major isoforms of I_K1-producing Kir2 encoding genes between the different regions of the ventricle (Fig. 2C).

I_Kr was elicited from the holding potential of −80 mV using a two-step voltage protocol at the frequency of 0.2 Hz using the established voltage protocol (Fig. 2D). There were no significant differences in I_Kr density between different regions of the ventricle (Fig. 2E). The I_Kr is produced by various erg encoding genes, of which KCNH6a (erg2) alone accounts for approximately 99.7% of the current in rainbow trout ventricle (Hassinen et al., 2021). Consistent with the I_Kr density, there were no differences in the expression of the two isoforms of I_Kr-producing erg genes (KCNH6a and KCNH2bb) between the different regions of the ventricle (Fig. 2F).

Ca²⁺ current (I_Ca) was elicited from the holding potential of −80 mV in the presence of 0.5 µmol l⁻¹ TTX, which completely blocks I_Na (Fig. 3A). The threshold voltage of I_Ca was approximately −30 mV and the peak density of I_Ca occurred at 0 mV for all four regions of the trout ventricle suggesting that major part of the charge was carried by L-type Ca²⁺ current (I_CaL). There were no statistically significant differences in the peak I_CaL density at 0 mV between different regions of the ventricle (Fig. 3B). I_Ca of rainbow trout ventricle is produced by a variety of Ca²⁺ channel genes (Hassinen et al., 2021). L-type Ca²⁺ channel cacna1c was the major component in all regions except the base, where cacna1c and the T-type Ca²⁺ channel cacna1ga were equally expressed (Fig. 3C). Of the minor Ca²⁺ components, L-type Ca²⁺ channel cacna1daa and T-type Ca²⁺ channel cacna1ha were more highly expressed in the base than in other regions of the ventricle (P<0.05), whereas no differences were found in T-type Ca²⁺ channel cacna1hb between different regions of the ventricle. The combined expression level of all Ca²⁺ channels did not differ between the four tissue regions (P>0.05; Fig. 3C).

Na⁺ current (I_Na) was elicited from the holding potential of −120 mV every 1 s (Fig. 3D). The peak current density was between −30 and −20 mV. There were no statistically significant differences in the peak current density of I_Na between different regions of the ventricle (Fig. 3E). I_Na is produced by several Na⁺ channel α-subunits in the rainbow trout ventricle, of which SCN4Aba and SCN5LAbb are the dominant isoforms (Hassinen et al., 2021). Consistent with the I_Na density, there were no statistically significant differences in the combined expression of the four α-subunit isoforms between different regions of the ventricle (P>0.05) (Fig. 3F). However, the expression of SCN4Abb was higher in the base than in other regions, and the expression of SCN5LAba in the base was lower than that in the apex and spongiosa.

The positions of the bipolar electrodes and the vectors of three ECG leads, and typical recordings from leads I, II and III are shown in Fig. 4A,B. The P wave and QRS complex correspond to atrial and ventricular depolarization, respectively, whereas the T wave represents the repolarization of the ventricle. The R wave is bigger than the S wave (positive QRS polarity) in leads II (+60 deg) and III (+120 deg) but smaller than the S wave (negative QRS polarity) in lead I (0 deg) (Fig. 4B,C). Furthermore, the amplitude of the R wave is approximately 50% higher in lead III than in leads I and II (P<0.05; Fig. 4E), suggesting that the vector of lead III is closest to the heart axis. Indeed, the overlap of the lead vectors (shaded areas of circles in Fig. 4F) indicates that the cardiac axis is between +90 and +120 deg, i.e. the main axis of ventricular depolarization is from base to apex. The T wave is discordant with the R wave in the three leads, suggesting that repolarization and depolarization of the ventricle in the three leads occur in the same direction (Fig. 4B–D). Given that lead III is almost parallel to the heart axis, the discordant T wave indicates that the main direction of repolarization in rainbow trout ventricle is from base to apex.

Mean duration of the QRS complex (activation) measured from the three EGC leads was 51.5±1.2 ms. As the total length of the trout ventricle was approximately 6 mm, the rate of AP propagation in the ventricular wall was approximately 0.12 m s⁻¹.

Analysis of APs, ion currents and transcript expression of ion channel α-subunits show little regional heterogeneity in the molecular and cellular basis of electrical excitation in the rainbow trout ventricle. Measurement of APs and ion currents indicate that electrical excitation is homogeneous at the level of isolated ventricular myocytes. This differs from several observations on mammalian cardiac myocytes, which show large regional differences in the shape of ventricular APs, underlying ion currents and ion channel expression (Antelevitch and Fish, 2001; Gaborit et al., 2007; McKinnon and Rosati, 2016; Schram et al., 2002; Szentadrassy et al., 2005; Watanabe et al., 1983). It might be assumed that the situation would be different if APs were measured from an intact fish ventricle. In multicellular tissue, myocytes are electrically coupled to each other, and electrotonic current flow between myocytes will attenuate, not increase, differences in AP shape (Anyukhovsky et al., 1999; Janse et al., 2012).

The conduction of depolarization and repolarization of the endothermic (mammalian and avian) hearts can be reduced to the consensus view that the major direction of depolarization is from apex to base and that the last activated regions repolarize first because they have the shortest APs. Therefore, the main direction of repolarization is from base to apex and from epicardium to endocardium, although this concept has been questioned (Christian and Scher, 1967; Opthof et al., 2017). This excitation pattern is structurally and functionally based on the rapid conduction pathway of the His–Purkinje system and apicobasal and transmural differences of AP durations and ion current densities. Owing to the fast-conducting His–Purkinje system, the effect of activation time to repolarization pattern may be small in endothermic hearts (Ramanathan et al., 2006). In contrast, hearts of embryonic endotherms and adult ectotherms seem to lack an anatomically distinct ventricular conduction pathway similar to the His–Purkinje system, and therefore in these hearts activation may occur from base to apex (Jensen et al., 2012). The present ECG recordings indicate that the main cardiac axis is indeed oriented craniocaudally, suggesting that depolarization of the rainbow trout ventricle occurs from base to apex. This is consistent with activation patterns of cod (Gadus morhua), pike (Esox lucius) and rainbow trout ventricle determined by multielectrode arrays (Azarov et al., 2013; Kibler et al., 2021; Vaykshnorayte et al., 2011, 2022). In these teleost species, the earliest activation occurs in the endocardial surface of the dorsal base near the atrioventricular orifice and spreads from there toward the ventral areas of the base and at the same time to the apex of the ventricle. Optical mapping of activation in the zebrafish heart indicates an activation pattern similar to that of cod, pike and trout (Jensen et al., 2012). However, the results for zebrafish are partly contradictory, as activation has also been shown to propagate from apex to base (Sedmera et al., 2003; Zhao et al., 2021). Two of the zebrafish studies (Jensen et al., 2012; Zhao et al., 2021) used the same voltage-sensitive dye, di-4-ANEPPS, for optical mapping. The fact that a different conclusion was reached using the same recording method suggests that differences in experimental conditions might explain the conflicting results. As the depolarization pattern has only been studied in a few teleost species, it is impossible to make any generalization of the ventricular activation pattern in fish. However, current observations indicate base-to-apex type of activation in most of the species studied.

The direction of ventricular repolarization is dependent on the spatial spread of depolarization (activation time) and AP duration (Janse et al., 2012). The current ECG recordings in rainbow trout reveal that the repolarization proceeds from base to apex. In the three leads, the T wave had an opposite polarity than the R wave, indicating that repolarization proceeded in the same direction as depolarization in these leads. The base-to-apex repolarization is consistent with multielectrode array studies in this species, which showed the earliest repolarization in the ventral base and then in the dorsal base and apex of the ventricle (Kibler et al., 2021). The homogeneity of AP durations throughout the rainbow trout ventricle is consistent with the base-to-apex repolarization pattern. In fact, similarity of AP durations does not allow for apex-to-base repolarization. In rainbow trout ventricle, activation reaches the apex within approximately 52 ms from the first depolarization at the base. During this time, APs of the basal myocytes have already started repolarization, which therefore must proceed towards the apex in the same order as depolarization owing to the spatial similarity of AP durations. The spatial homogeneity of AP durations is supported by similar density of the main depolarizing and repolarizing ion currents and similar transcript levels of ion channel α-subunits in different parts of the ventricle. Consistent with this, the ventricle of summer-acclimatized rainbow trout at 18°C shows little difference in activation time and activation–repolarization interval (Vaykshnorayte et al., 2022). Also consistent with the current AP duration data, a multielectrode array study has shown that repolarization time is similar at the base and apex of the rainbow trout ventricle at the sinus rate (Kibler et al., 2021). Notably, when the combined data from the three regions of the compact myocardium were compared with that of the spongy layer, practically no differences were found in APs, ion currents or ion channel expression (Fig. S1). This strongly suggests the absence of transmural heterogeneity of activation and repolarization in the trout ventricle. However, Vaykshnorayte et al. (2022) found transmural and apicobasal differences in activation time and activation–repolarization interval in the ventricle of winter-acclimatized rainbow trout at 3°C, suggesting that the patterns of activation and repolarization may be temperature-dependent. The repolarization pattern of the ventricular epicardium of the pike heart is shown to occur mainly in the apex-to-base direction, although an additional area of early repolarization was found at the posterior ventricular base, i.e. at the site of early depolarization (Vaykshnorayte et al., 2011). Apex-to-base repolarization requires that the durations of APs become shorter from the base of the ventricle towards the apex of the ventricle. Consistent with this, the repolarization time of the apex is slightly shorter in the apex than base of the pike ventricle (Vaykshnorayte et al., 2011). A similar difference in repolarization as between rainbow trout and pike is found also between canine and human hearts: in human ventricles, activation and repolarization proceed in opposite directions, whereas the activation and repolarization follow the same pattern in dog ventricles. Accordingly, in humans, QRS and T waves are concordant, but in the dog, they are discordant (Janse et al., 2012). It seems that the repolarization of the vertebrate ventricle(s) is more variable than depolarization, both within and between species (Opthof et al., 2017). The functional significance of different repolarization patterns remains to be elucidated.

Mammalian ventricles have two types of cardiac myocytes: working ventricular myocytes and conducting Purkinje fibers, the former showing region-specific characteristics of excitability (epicardial, midmyocardial and endocardial myocytes) (Anyukhovsky et al., 1999). In contrast to mammalian ventricles, the rainbow trout ventricle seems to have only working myocardial cells, which are electrophysiologically homogeneous through the ventricular myocardium. This may suggest that specific conduction fibers are absent and therefore electrical excitation propagates along the working myocardial cells throughout the ventricle, which may be an evolutionally old feature and common to the hearts of all ectotherms (Jensen et al., 2012; Boukens et al., 2019). In trout ventricle at 16°C, the rate of AP propagation was approximately 0.12 m s⁻¹. In mammalian Purkinje fibers, the rate of AP propagation at 37°C is 2 m s⁻¹ or slightly higher (Cranefield et al., 1971; Dominguez and Fozzard, 1970), whereas in working ventricular muscle it is approximately 0.6 m s⁻¹ (Kelly et al., 2013; McIntyre and Fry, 1997). Using a Q₁₀ value of 2.0, the rate of AP propagation in mammalian Purkinje fibers would be approximately 0.47 m s⁻¹, and 0.14 m s⁻¹ in working ventricular myocardium at 16°C. The fact that the rate of AP propagation in the trout ventricle is only about one-fourth of the rate of mammalian Purkinje fibers, but similar to the rate of mammalian working myocardium, suggests that conduction does not involve Purkinje fiber-type myocytes. However, if present, the conducting fibers should be electrophysiologically clearly recognizable (e.g. by large I_Na and fast +dV dt⁻¹) from the working cardiac myocytes (Dangman et al., 1982; Makielski et al., 1987). The present study includes patch-clamp recordings from over 300 myocytes from different regions, which is only a tiny fraction of the total myocyte population of the ventricle, and therefore it is possible that not a single conducting fiber was recorded. When searching for a specialized ventricular conducting system, it should be histologically discrete and insulated by a sheath of fibrous tissue from the adjacent working myocardium (Anderson et al., 2009) and characterized by a high density of molecular markers of a fast-conducting tract (e.g. connexins, sodium channels) (Haufe et al., 2005; Kanter et al., 1993). Indeed, Boukens et al. (2019) hypothesized that the electrical conduction delay in the ventricle of ectotherms compared with mammals might be due to the tortuous structure of the spongious myocardium, whereas the compact architecture of mammalian ventricles accelerates electrical conduction independently of the conduction system, as occurs in the mammalian atria.

To our knowledge, this is the first study to look at the expression of AP waveforms, ion currents and ion channels in the fish ventricle. Findings at all organization levels (APs, ion currents, ion channel expression) indicate regional homogeneity of electrical excitation throughout the rainbow trout ventricle. At the organismal level, these findings are consistent with the base-to-apex spread of both depolarization and repolarization, which does not require a specialized ventricular conduction system. From the perspective of ventricular function, this means that at the onset of contraction, the pressure is directed toward the apex of the ventricle, forcing blood to flow through the network of spongious myocardium. Because the basal area relaxes first, blood flows from the apical ventricle towards the base and into the bulbus arteriosus is facilitated by the early increase in the volume of the basal ventricle. Future studies should investigate whether the absence of specialized ventricular conduction system, characterized by high I_Na density and fast AP upstroke velocity, sets the upper limit to the heart rate of fish at critically high temperatures (Vornanen, 2020).

We thank Anita Kervinen for taking care of fish and preparing solutions for the patch-clamp experiments.