Electric catfish hearts are not intrinsically immune to electric shocks

The contraction of the heart is controlled by intrinsic electrical impulses. Cardiac pacemaker cells spontaneously fire impulses that are conveyed to all atrial and ventricular myocytes. The specialized cardiac conduction system ensures the coordinated contraction of the heart and its automatic rhythmic beat (e.g. Reilly, 1998). Cardiac arrhythmias – irregularities in the rate or rhythm of the heartbeat – are one of the leading causes of mortality worldwide (Clauss et al., 2019). The various types of arrhythmias are generally caused by alterations in the myocardial electrical properties, such as excitation or impulse formation, conduction and repolarization (Clauss et al., 2019). Atrial fibrillation is the most common form of arrhythmia and is usually induced by abnormalities in the electrical properties of the atrial tissue itself or by ectopic beats, resulting in premature atrial contractions. Such premature depolarizations mainly originate from activities in the pulmonary myocardial sleeves (Chen et al., 1999; Haïssaguerre et al., 1998) but can also be initiated by external electric shocks applied to the heart, e.g. during an electrical injury (Waldmann et al., 2017). Here, even short electric stimulations with low intensities that occur during vulnerable phases of the cardiac cycle can be sufficient to disturb the cardiac conduction system and to induce ventricular or atrial fibrillations (Reilly, 1998; Wiggers and Wégria, 1940). The severity of such electric shock-induced arrythmias is determined by different factors such as current and voltage intensity or the duration and pathway of current flow. Thus, electric shocks with high voltages cause life-threatening cardiac injuries, such as sudden cardiac standstill or severe myocardial injuries (Waldmann et al., 2018).

The strongly electric catfish (Malapteruridae) produce high-voltage discharges that are generated by specialized electric organs within its body. Each electric organ discharge can reach amplitudes of hundreds of volts and is sufficient to activate motor neurons and induce whole body muscle contractions in other fish (Bauer, 1968; Welzel and Schuster, 2021). By emitting volleys of these high-voltage discharges, electric catfish are able to paralyze and stun prey or foe (Bauer, 1968; Belbenoit et al., 1979; Welzel and Schuster, 2021). We have recently shown that intact electric catfish are remarkably immune, not only to their own shocks but also to external high-voltage shocks delivered by a commercial electrofishing device (Welzel and Schuster, 2021). Here, we examined critical anatomical features of the heart of the electric catfish (Malapterurus beninensis) and analyzed at which current densities its cardiac conduction system is disturbed by electric shocks. For this purpose, we used an ex vivo heart perfusion system to determine sensitivity thresholds of explanted hearts of electric catfish to electric currents and to compare them with goldfish hearts of similar size.

We used 7 electric catfish (Malapterurus beninensis Murray 1865) and 14 goldfish [Carassius auratus (Linnaeus 1758)]. All fish were obtained commercially from an authorized specialist retailer (Aquarium Glaser GmbH, Rodgau, Germany). The electric catfish were kept individually in 60 or 120 l glass aquaria and the goldfish were kept together in a 240 l glass aquarium under a 12 h light:12 h dark cycle (lights on 07:00 h and off at 19:00 h). We used 6 juvenile (7–15 cm, 37–82 g) and 1 adult (20 cm, 167 g) electric catfish [adult M. beninensis are maximal 22 cm in size (Norris, 2002)] as well as different sized goldfish (6–15 cm, 7–155 g). The comparatively large size range of electric catfish and goldfish was chosen to analyse the allometric scaling of stimulation and arrhythmia thresholds with heart size. Also note that fish of similar size were used in a prior study that showed electrical immunity of intact electric catfish (Welzel and Schuster, 2021). All electric catfish were kept for at least 9 months prior to any of the experiments and all voraciously took food during their daily feeding session. Electric catfish were fed with bloodworms and goldfish with flake food. Temperatures (22–24°C), pH (6.5–7.0) and conductivity (250–300 µS cm⁻¹) of the water [based on 50% tap water and 50% demineralized water, enriched with 7% (w/v) NaHCO₃, 4% (w/v) CaSO₄, 1% (w/v) marine salt] were kept constant during the experimental period. Animal care and all experimental procedures were conducted in accordance with the German Animal Welfare Act (Tierschutzgesetz).

A total of 14 goldfish and 7 electric catfish were anesthetized in water containing MS-222 (200–500 mg l⁻¹), weighed and euthanized by decapitation. All fish were placed ventral side up in a dissection dish. The dissection of the fish hearts was performed as described by Cao and Poss (2016). Briefly, the body cavity was opened with a longitudinal cut and the pericardium was opened. The distal end of the bulbus arteriosus was gripped with fine forceps and the hearts were carefully dissected out by severing the ventral aorta and the sinus venosus under a stereomicroscope (Stemi 2000, Zeiss). The hearts were transferred to a 60-mm Petri dish containing oxygenated fish Tyrode solution (132 mmol l⁻¹ NaCl, 2.5 mmol l⁻¹ KCl, 4 mmol l⁻¹ NaHCO₃, 0.33 mmol l⁻¹ NaH₂PO₄, 1 mmol l⁻¹ CaCl₂, 1.6 mmol l⁻¹ MgCl₂, 10 mmol l⁻¹ HEPES, 5 mmol l⁻¹ glucose and 5 mmol l⁻¹ sodium pyruvate; pH adjusted to 7.5 with NaOH) according to Zhang et al. (2011).

To assess the ex vivo sensitivity to electric shocks, we used an isolated heart perfusion system for fish hearts. First, the hearts were transferred to a perfusion dish placed under a stereomicroscope (Stemi 2000-CS, Zeiss) equipped with a camera (Alpha 6300, Sony). A catheter (30–34 gauge, depending on heart size) was fixed at one side of the dish. The hearts were then attached to the needle of the catheter as described by Lin et al. (2015). Briefly, the end of the bulbus arteriosus was carefully gripped with fine forceps and the tip of the needle was inserted into the lumen of the bulbus arteriosus and secured with a suture. We used a pretied knot of a single filament of a braided polyester suture (size 6-0, PS-Eurosutures) to secure the bulbus arteriosus onto the needle. The catheter was connected to a peristaltic pump (Reglo Digital, Ismatec), which provided a constant retrograde flow (0.5–1 ml min⁻¹ depending on the heart size) of oxygenated fish Tyrode solution from the bulbus arteriosus through the ventricle and the atrium. The fish Tyrode solution was placed in a constant-temperature water bath (23.4±0.4°C) and continuously oxygenated by carbogen (95% O₂/ 5% CO₂) during the experiments. Only hearts that showed regular and constant spontaneous beating after a stabilization period of at least 10 min were used for electrical stimulation experiments. The entire procedure from the euthanasia of the fish to cannulation and perfusion of the heart was achieved within 2 min.

Electrical stimulation was applied via a bipolar stimulation electrode consisting of two sintered Ag/AgCl electrodes (540800, AM Systems) that were carefully placed on the epicardial surface by using two mechanical micromanipulators (MM33, Märzhäuser-Wetzlar) so that no change in the visually recorded pattern of the heartbeat occurred. Electrical stimuli were delivered by an isolated constant current stimulator (DS3, Digitimer) driven by Spike2 (Cambridge Electronic Design Limited) software. The batteries of the stimulators were tested at the beginning of each experiment with a digital multimeter according to the manufacturer's instructions. For atrial stimulation (stimulation site 1), the stimulation electrodes were placed on the atrial surface with an interelectrode distance of 2–4 mm. The hearts were additionally stimulated at two other stimulation sites (site 2 and 3). If stimulation site 2 was tested, one stimulation electrode was placed in the middle of the atrium and one in the middle of the ventricle (variable interelectrode distance depending on heart size). If stimulation site 3 was tested, one stimulation electrode was placed at the apex of the ventricle and one at the base of the ventricle near the bulbo-ventricular junction (variable interelectrode distance depending on heart size).

To determine the stimulation threshold, the hearts were stimulated (1 ms pulse duration) in the mid or late atrial diastole with an initial current of 0.05 mA. The amplitude was increased stepwise from 0.05 to 1 mA in steps of 0.05 mA and from 1 to 5 mA in steps of 0.5 mA until atrial and ventricular capture occurred. This current level was called the diastolic stimulation threshold.

To determine the arrhythmia threshold, a 3 s burst of 300 Hz pulses (1.5 ms pulse duration) was delivered to the atrium at stimulation site 1. The current intensity was increased from the stimulation threshold in steps of 0.5–1.0 mA, until any kind of arrhythmia occurred. The minimal current intensity of the burst required to induce any kind of sustained (more than 10 s after the end of the burst) arrhythmias (e.g. atrial fibrillation, atrial flutter, AV block, cardiac arrest) was defined as the arrhythmia threshold.

A pair of custom-made recording electrodes (chloridized silver wires; interelectrode distance 2 mm) was placed on the middle of the atrium near the atrio-ventricular junction to record atrial electrograms. Another pair of recording electrodes was placed on the middle of the ventricle to record ventricular electrograms. Each pair of recording electrodes was carefully placed on the epicardial surfaces by using two mechanical micromanipulators (MM33, Märzhäuser-Wetzlar). Neither the placement of the stimulation electrodes nor the placement of the recording electrodes on the epicardium had any apparent effect on the visually observed (Alpha 6300, Sony) heart function or the heart rhythm. Ag/AgCl pellet electrodes (550008, AM Systems) were used as reference electrodes. Atrial and ventricular signals were amplified (200- to 1000-fold) and low-pass filtered (<100 Hz) by two differential amplifiers (DPA-2FS, npi electronic) and digitized (CED Micro1401-3, Cambridge Electronic Design Limited) at a sampling rate of 20 kHz. All signals were recorded, digitally band pass filtered (5 Hz–50 Hz) and processed using Spike2 software (Cambridge Electronic Design Limited). After filtering, the ECG waves (P, QRS and T) and the PR, QRS and QT intervals were analyzed. The P wave corresponds to the atrial depolarization and was identified as the first abrupt or gradual deviation from the baseline. The QRS complex represents the ventricular depolarization. The T wave represents the ventricular repolarization. The RR interval (heart rate) was defined as the time between the R peaks of two consecutive QRS complexes. The PR interval is the time from the beginning of the P wave to the first deflection of the QRS complex. The QT interval represents the time from the onset of the QRS complex to the end of the T wave. The QRS amplitude was defined as the peak-to-peak amplitude of the QRS complex.

At the end of isolated heart perfusion experiments, the hearts of goldfish and electric catfish were chemically fixed at 4°C using a 4% formaldehyde solution in PBS for 24 h. All samples were rinsed twice in PBS, dehydrated through a series of graded ethanol (70%, 80%, 90% and 3×100% for 1 h each), followed by three incubation steps in xylene (for 1 h each) and embedded in paraffin. The paraffin block was cut on a rotary microtome (RM 2035, Leica) into 10 µm sections, which were mounted on glycerine-coated (P049.1, Carl Roth) glass slides. Paraffin sections were deparaffined (2× xylene for 5 min) and rehydrated in a graded ethanol series (96%, 80%, 70%, 60% and distilled water for 4 min each).

To visualize connective tissues, a Masson's trichrome staining was performed (Masson Goldner Trichrome Staining Kit 3459, Carl Roth). All sections were stained according to the manufacturer's instructions, dehydrated by an ascending alcohol series (70%, 96%, 100% and xylene for 5 min each) and mounted with Eukitt (03989, Sigma). Histological analysis of heart sections was carried out using a microscope (Axio Imager M2, Zeiss) or a stereomicroscope (MZ10F, Leica). Images were acquired using an Axiocam MRc camera (Zeiss) and ZEN2 software (Zeiss).

Statistical analyses were carried out using Graph Pad Prism version 5.01 (Graph Pad Software) and Excel (Microsoft). Normal distribution of data was tested using the D'Agostino-Pearson normality test. Differences in the stimulation and arrhythmia thresholds between explanted goldfish and electric catfish hearts were tested by using the two-tailed Mann–Whitney U-test. Pearson's correlation was used to test the relationship between stimulation or arrhythmia thresholds, respectively, and heart size. Values of P<0.05 were considered statistically significant. The significance level is shown by *P<0.05; **P<0.01; ***P<0.001; ns, not significant. All data are presented as mean± s.e.m. N denotes the number of animals. Movies were edited using Adobe Premiere Pro 2.0 (Adobe Systems Incorporated).

We first tested whether we could identify adaptations or anomalies in the anatomy of the heart of electric catfish. To explore this possibility, we did Masson–Goldner trichrome staining of paraffin sections (Fig. 1A,B). The heart, positioned within the pericardium, is located ventrally to the gills and follows the basic teleost plan (Farrell and Pieperhoff, 2011; Icardo, 2012) with four distinct chambers in series: the sinus venosus, the atrium, the ventricle and the bulbus arteriosus (Fig. 1C). Based on our histological observations, the electric catfish has a typical type II fish heart (Farrell and Jones, 1992; Icardo, 2012; Tota et al., 1983), which is characterized by a thin vascularized outer compact layer of myocardium and an avascular spongy myocardium (Fig. 1B). All cardiac chambers are surrounded by a collagenous epicardium with a thickness comparable to other fish of similar size (Farrell and Pieperhoff, 2011). The heart of the electric catfish has a saccular ventricle and is comparatively small, with a relative heart mass (heart mass as a percentage of body mass) of only 0.049±0.005% (N=7; Fig. 1D). Both characteristics are typical for less active or sedentary fish with low blood pressure (Farrell and Pieperhoff, 2011; Icardo, 2012; Santer, 1985) and would thus seem to be fitting to the lifestyle demands of electric catfish (Bauer, 1968; Norris, 2002). In summary, our findings show no immediately obvious deviations from a type II heart of a less active teleost fish.

We next aimed to directly analyze whether the electrical sensitivity of the explanted heart of an electric catfish is unusually higher than that of other fish. In general, electric shocks have massive effects on the cardiac conduction system and can induce cardiac arrhythmias. Such disturbances of the cardiac cycle are usually detected by recording an electrocardiogram (ECG). A powerful technique in cardiac research to analyse the ECGs and the physiology of hearts is the Langendorff method (Langendorff, 1895). Here, the hearts of mammals are explanted from the body and retrogradely perfused with oxygenated ringer solution via a cannula inserted into the aorta (Bell et al., 2011; Langendorff, 1895). This allows the direct monitoring of the ECG and the investigation of the sensitivity to electrical stimulation and arrhythmogenesis (Bell et al., 2011).

We therefore used a Langendorff-like perfusion system for fish hearts to directly measure the sensitivity of the electric catfish heart to electric shocks (Fig. 2A) while simultaneously recording epicardial surface ECGs of the ventricular and atrial activity (Fig. 2B). In our perfusion system, we cannulated the bulbus arteriosus with a catheter to retrogradely perfuse the isolated hearts from the bulbus arteriosus through the ventricle and the atrium. This ensured that the hearts continued beating for several hours at normal and stable heart rates (Fig. 2C). The heart rates of perfused electric catfish hearts were comparable to both goldfish hearts in the same setup (Fig. 2D) and different teleost fish hearts perfused in other setups (Bielig, 1931; Farrell et al., 1983; Garofalo et al., 2012; Haverinen and Vornanen, 2016; Imbrogno et al., 2001; Overgaard et al., 2004).

We first examined whether the ECGs would indicate any adaptations in the cardiac conduction system. However, the ECG recordings revealed no characteristic deviations (Fig. 2E). Like the ECG of goldfish (Fig. 2F) and other teleost fishes (Farrell and Smith, 2017), the electric catfish ECG is characterized by a distinct P-wave (atrial depolarization), a QRS complex (ventricular depolarization) and a T-wave (ventricular repolarization). The duration of the PR (242.3±12.7 ms, N=7), the QRS (123.9±10.8 ms, N=7) and the QT interval (535.5±25.5 ms, N=7) are within the normal range of teleost fish (PR: 50–300 ms; QRS: 50–300 ms; QT: 200–700 ms) (Badr et al., 2016; Haverinen and Vornanen, 2020; Liu et al., 2016; Milan et al., 2006; Monteiro et al., 2020; Tikkanen et al., 2016) and also comparable to that of the goldfish hearts (Fig. 2G). Also, the amplitude of the QRS complex (1.8±0.3 mV, N=7) is typical for explanted fish hearts (0.6–8 mV) (Haverinen and Vornanen, 2016; Haworth et al., 2014; Stoyek et al., 2016).

In summary, the structure of the electric catfish ECG is similar to that of goldfish and suggests no immediately obvious adaptations. Thus, the similarity in the heart anatomy is also mirrored in the similarity of the electrocardiograms. Furthermore, our ex vivo technique now allowed us to directly compare the electrical sensitivity of the hearts of electric catfish and goldfish.

Next, we used our heart perfusion system to explore whether the electric catfish heart is immune to external electrical stimuli. In general, fish hearts can be stimulated with single square pulses to extrinsically pace their heart rates (Haverinen and Vornanen, 2016; Imbrogno et al., 2001; Lin et al., 2015; Patrick et al., 2010; Stoyek et al., 2016; Vornanen, 1989). Thus, we first examined whether we could stimulate the perfused electric catfish hearts during the diastole with single 1 ms square pulses. For each heart, we determined the diastolic atrial stimulation threshold, which is defined as the minimum stimulus intensity (in mA) that caused a complete heartbeat (Fig. 3A). Surprisingly, even low current intensities were sufficient to stimulate the hearts of electric catfish and to pace their heart rate (Fig. 3B). The mean atrial stimulation thresholds of electric catfish hearts (0.12±0.02 mA, N=7) were not significantly different from that of goldfish hearts (0.12±0.01 mA, N=14, P=0.575, Mann–Whitney U-test) (Fig. 3C). To analyze possible effects of the electrode placement and thus the pathway of the current through the heart (Reilly, 1998), we also compared the stimulation thresholds at two additional stimulation sites (Fig. 3D,E). However, the similarity in the stimulation thresholds can also be verified at these two additional stimulation sites. Furthermore, for both species and stimulation site, the allometric analysis showed no significant correlation between the atrial stimulation thresholds and the mass of the heart, with massive overlap over the full range of masses (Fig. 3F–H). Thus, our findings show that ex vivo the hearts of electric catfish were equally sensitive to electrical stimulation as those of goldfish.

We next examined whether the electric catfish heart is not only sensitive to electrical stimulation but is also susceptible to arrhythmias. We therefore used electrical burst stimulations of the atrium that reliably induce severe atrial or ventricular arrythmias in mammals (Bruegmann et al., 2016; Haugan et al., 2004; Schrickel et al., 2007; Wang et al., 1992; Yang et al., 2016). We again placed the stimulation electrodes on the epicardial surface of the atrium (Fig. 4A) and then applied atrial burst stimulations with a duration of 3 s (900 pulses at 300 Hz). We determined the arrhythmia threshold by increasing the current intensity until any kind of sustained arrhythmia (sustained for more than 10 s after the burst had stopped) occurred. In the goldfish hearts, even at minimum strength, our stimuli induced a variety of different forms of arrhythmia such as cardiac arrest (Fig. 4B), atrial flutter (Fig. 4C) or atrial fibrillation (Fig. 4D). The threshold for producing any of these forms of arrhythmia was 3.61±0.61 mA (N=13, Fig. 4F). Surprisingly, similarly sized currents also induced sustained arrhythmias in each of the explanted hearts of the electric catfish (Fig. 4E and Movie 1). The same spectrum of forms of arrhythmia were found as in goldfish and are shown in Fig. 4E. Moreover, the arrhythmia threshold (3.03±0.96 mA, N=5) was not significantly different (P=0.52; Mann–Whitney U-test) between the electric catfish and the goldfish (Fig. 4F). Also, the time to recovery of the normal heartbeat was comparable. In the electric catfish, 3 of the 5 hearts recovered within a time range similar to that of goldfish hearts (Fig. 4G). Furthermore, in contrast to the stimulation thresholds (Fig. 3F–H), the arrhythmia thresholds correlated clearly with the mass (M) of the heart, following the same relation in both species (Fig. 4H). Our findings thus demonstrate that the electric catfish hearts are clearly not immune at all to electric shocks – in marked contrast to the situation in the intact catfish (Welzel and Schuster, 2021).

Strongly electric fish, such as electric catfish or electric eels, generate high-voltage electric discharges within their bodies that have massive effects on the physiology of prey fish. Each high-voltage discharge is sufficient to activate motor neurons and to induce muscle twitches in prey fish (Catania, 2014; Welzel and Schuster, 2021). By using volleys of these discharges, strongly electric fish induce sustained involuntary muscle contractions (Bauer, 1979; Catania, 2015; Welzel and Schuster, 2021) that immobilize their prey and thus simplify its capture. It is not known how strongly electric fish manage to electrocute their prey but not themselves. In contrast to earlier suggestions (Nelson, 2011), we have recently demonstrated that intact electrical catfish are immune not only to self-produced but also to external high-voltage electric discharges of different waveforms, intensities and durations (Welzel and Schuster, 2021).

Given that cardiac damage and arrhythmia are the most common injuries following electrical shocks in humans (Waldmann et al., 2017, 2018) and since cardiovascular diseases are one of the leading causes of deaths worldwide (World Health Organization: https:/www.who.int/health-topics/cardiovascular-diseases), whether the hearts of strongly electric fish show any intrinsic adaptations that protect them from arrhythmia is a question of special interest. In general, the hearts of most fish are clearly stimulated by high-voltage electric shocks: depending on the strength and duration of electrical stimulation, different effects on the cardiac conduction system can be induced: from extrasystoles (Arbel et al., 1977; Kibler et al., 2021; Preyer, 1933) to arrhythmias and even cardiac arrest (Biswas and Karmarkar, 1979; Schreer et al., 2004). Intrinsic tolerance to high-voltage stimulation could result from changes in various pathways during evolution of the electric catfish. Adaptations in the cardiac conduction system or specialized ion channel compositions of the cardiomyocytes could have caused altered electrical properties of the electric catfish heart. Changes in the electrical excitability, in turn, may have caused lower sensitivity to ectopic beats or external electric shocks, and thus a reduced susceptibility to arrhythmias, such as atrial fibrillations. Nonetheless, our findings indicate no intrinsic specialization of the heart in the electric catfish: taken out of the body, the heart of the electric catfish is not resistant to electrical shocks but, in fact, is equally sensitive as the hearts of a goldfish (Figs 3 and 4) and similar to the hearts of many other fish (Haverinen and Vornanen, 2016; Imbrogno et al., 2001; Lin et al., 2015; Patrick et al., 2010; Vornanen, 1989). Specifically, stimulation with electrical bursts elicited severe and sustained arrhythmias in the hearts of electric catfish (Fig. 4E). Moreover, at any given heart mass, the same currents cause such severe arrythmias in both goldfish and electric catfish (Fig. 4H). The threshold currents increase with heart mass, following the same trend in both species.

It thus seems that the evolution of a strongly electric organ in the African electric catfish did not require the evolution of any specific adaptations in the heart, but that this is reliably protected by other means. We suggest that other electrovulnerable tissues (such as the brain, nerves, sensory cells) also have not evolved any tolerance against high-voltage discharges and are protected in similar ways as the heart. The simplest explanation for our findings would be that either the heart or all electrovulnerable organs are wrapped in insulating tissue. Both would, in principle, be possible in the electric catfish whose electric organ is located below the surface. Hence, all inner organs could be protected by an insulating layer located below the electric organ. Unfortunately, at this stage, all attempts in our lab to detect any conspicuous insulating layers beneath the electric organ have failed and we also did not find any conspicuous insulating layers around the heart. So, either the hypothetical insulators are highly efficient or the actual mechanism of how such layers protect the catfish's heart in vivo is more complex. Given the present findings, it will be interesting to find out what acts so reliably in protecting an otherwise vulnerable heart.

We thank Antje Halwas, Thomas Toesko and Susanne Proschke for excellent technical assistance.