Regulation of heart rate following genetic deletion of the ß1 adrenergic receptor in larval zebrafish.

AIM: Although zebrafish are gaining popularity as biomedical models of cardiovascular disease, our understanding of their cardiac control mechanisms is fragmentary. Our goal was to clarify the controversial role of the ß1-adrenergic receptor (AR) in the regulation of heart rate in zebrafish.
METHODS: CRISPR-Cas9 was used to delete the adrb1 gene in zebrafish allowing us to generate a stable adrb1-/- line. Larval heart rates were measured during pharmacological protocols and with exposure to hypercapnia. Expression of the five zebrafish adrb genes were measured in larval zebrafish hearts using qPCR.
RESULTS: Compared with genetically matched wild-types (adrb1+/+ ), adrb1-/- larvae exhibited ~20 beats min-1 lower heart rate, measured from 2 to 21 days post-fertilization (dpf). Nevertheless, adrb1-/- larvae exhibited preserved positive chronotropic responses to pharmacological treatment with AR agonists (adrenaline, noradrenaline, isoproterenol), which were blocked by propranolol (general ß-AR antagonist). Regardless of genotype, larvae exhibited similar increases in heart rate in response to hypercapnia (1% CO2 ) at 5 dpf, but tachycardia was blunted in adrb1-/- larvae at 6 dpf. adrb1 gene expression was abolished in the hearts of adrb1-/- larvae, confirming successful knockout. While gene expression of adrb2a and adrb3a was unchanged, adrb2b and adrb3b mRNA levels increased in adrb1-/- larval hearts.
CONCLUSION: Despite adrb1 contributing to the setting of resting heart rate in larvae, it is not strictly essential for zebrafish, as we generated a viable and breeding adrb1-/- line. The chronotropic effects of adrenergic stimulation persist in adrb1-/- zebrafish, likely due to the upregulation of other ß-AR subtypes.

INTRODUCTION

The genomes of most vertebrates contain three ß‐adrenergic receptors (ß‐ARs), ß1‐AR, ß2‐AR, and ß3‐AR, encoded by the adrb1, adrb2, and adrb3 genes, respectively. These paralogs originated from a single ancestral gene following two rounds of whole‐genome duplication early in the evolution of vertebrates. , Different ß‐AR isoforms display characteristic pharmacological properties , , and exhibit tissue‐specific expression patterns, although co‐expression of all three receptors occurs in some organs. In the healthy mammalian heart, ß1‐ARs are predominantly expressed, but it is complemented by expression of ß2‐ARs and ß3‐ARs at least in a subpopulation of cardiomyocytes. , , , During heart failure, ß1‐ARs are downregulated and ß2‐ARs become relatively more abundant and of greater functional relevance. , The relative expression of ß2/ß1 receptors is higher in the cardiomyocytes of the sino‐atrial node (SAN), which harbors the dominant cardiac pacemaker, than surrounding atrial or ventricular myocardium. , ,

Zebrafish are emerging as popular model organisms for understanding heart function, including the effects of ß‐adrenergic stimulation. , , , , , Owing to a third whole‐genome duplication event in early teleost fishes, zebrafish are endowed with a bolstered array of adrb genes: adrb1, adrb2a, adrb2b, adrb3a, and adrb3b. , Evidently, the “second” adrb1 gene was lost, presumably as it was redundant; such “nonfunctionalization” is a common fate of duplicated genes. Previously, pharmacological and morpholino‐based knockdown approaches , were used to characterize the roles of the different ß‐AR isoforms in zebrafish, including their importance in setting heart rate. In general, the results of these studies indicate a leading role for the ß1‐AR in determining the normal positive chronotropic response to adrenergic stimulation. The preferential reliance on the ß1‐AR is consistent with it being the most highly expressed adrb gene in the adult zebrafish heart. , In larval [3 and 4 days post‐fertilization (dpf)] zebrafish, morpholino knockdown of adrb1 reduced heart rate and blunted the positive chronotropic responses induced by adrenaline or isoproterenol (ß‐AR agonists), , while adrb2a and adrb2b knockdown increased baseline heart rate. In 5 dpf larvae, adrb1 knockdown and ß1‐AR‐specific antagonist atenolol abolished the tachycardia in response to elevated CO2 (hypercapnia) exposure while atenolol reduced heart rate and cardiac contractility in adult zebrafish hearts. These data are somewhat surprising as previous studies suggested that ß2‐ARs are dominantly expressed in atrium and ventricle in other teleost fish species. , , , , , Moreover, an immunohistological investigation showed abundant ß2‐AR protein expression across the zebrafish heart, including the SAN region.

Pharmacological and morpholino knockdown approaches can be limited by non‐specific and off‐target effects. The generation of adrb knockout mice has previously proven instrumental in delineating the roles of specific receptors in mammals. Mice lacking the ß1‐AR (adrb1 ) exhibit high embryonic lethality (~70%–90% mortality), but those that do survive to adulthood appear phenotypically normal. Despite evidence for enriched ß2‐AR expression in the sinoatrial node cardiomyocytes of mammals, which has functional significance on the ion currents that determine the rate of spontaneous depolarization, , adrb1 mice lack a chronotropic, as well as inotropic, response to ß‐AR stimulation with isoproterenol.

The goal of the present investigation was to generate an adrb1 zebrafish line to clarify the importance of the ß1‐AR in adrenergic control of heart rate in this popular model species. CRISPR‐Cas9 gene editing was used to delete the single adrb1 exon, after which a suite of protocols was conducted to study heart rate regulation in adrb1 larval zebrafish.

RESULTS

The generation of a viable adrb1 −/− line

CRISPR‐Cas9 was used to delete the single exon of adrb1 in zebrafish (Figures 1 and S1). To establish whether wild‐type and knockout larvae exhibited equivalent rates of survival, four separate F1 adrb1 +/− incrosses (1 male 1 female) were performed and the genotypes of the offspring were determined at 7 and 28 dpf. These results were compared with the expected percentage of each genotype based on Mendelian ratios (adrb1 25%, adrb1 50%, adrb1 25%). At both developmental time points, there were no significant differences from expected Mendelian ratios (Figure S2; p = 0.06 at 7 dpf; p = 0.12 at 28 dpf). Indeed, while at 7 dpf, the adrb1 genotype tended to be underrepresented (13.8%), at 28 dpf it was correspondingly overrepresented (41.7%), apparently due to adrb1 and adrb1 being removed from the population by chance at 7 dpf (Figure S2). As such, when the 7 and 28 dpf were pooled to provide an overview of the general population across larval development (Figure 1C), presuming that the individuals sampled at 7 dpf would have survived in similar proportions to those left until 28 dpf, the distribution of genotypes was very close to theoretical Mendelian ratios (p = 0.43; adrb1 24.0%, adrb1 55.8%, adrb1 20.2%). Anecdotally, in the propagated F2 lines, there was no excess mortality in adrb1 fish, which reached reproductive age (~90 dpf) at the same time as adrb1 siblings.

FIGURE 1

The generation and screening of an adrb1 knockout (adrb1 −/−) zebrafish line. (A) Schematic illustration of the adrb1 gene within chromosome 12 including the single exon, the deleted segment (confirmed by Sanger sequencing), and sequencing primer targets. See Supplementary Material for annotated sequence. (B) Polymerase chain reaction with the sequencing primers generated a ~3000 bp wild‐type band (adrb1 +/+), a ~300 bp mutant band only (adrb1 −/−) or both bands (adrb1 +/−). (C) The results of F1 heterozygote (adrb1 +/) incrosses generated offspring in the expected Mendelian ratio (compared with Chi‐square test, p = 0.43; adrb1 24.04%, adrb1 55.77%, adrb1 20.19%). Note that this figure includes pooled data from the populations sampled at 7 and 28 dpf; for each subsample (age) see Figure S2.

Heart rate during development

Throughout development, there was an overall effect of genotype (F (1, 453) = 136.0; p < 0.001), wherein adrb1 −/− larvae generally exhibited heart rates ~20 beats min−1 lower than adrb1 +/+ larvae (Figure 2). There was an overall effect of age (F (8, 453) = 112.9; p < 0.001) but no genotype*age interaction (F (8, 453) = 1.21; p = 0.29) as both genotypes exhibited parallel triphasic changes consisting of an increase in heart rate from 2 to 5 dpf, a decline from 6 to 11 dpf, and a subsequent increase from 11 to 21 dpf (Figure 2).

FIGURE 2

Heart rate during development in unanesthetized adrb1 +/+ and adrb1 −/− larvae. Heart rates were optically measured from 2–21 days post‐fertilization (dpf). Each individual larva was used at only one time point. Statistical results show the output of a two‐way ANOVA. N values were 16–55 at each time point. Data points show means ± SEM.

Heart rate responses to adrenergic agonists and antagonists

Adrenaline (Figure 3A; F (1, 84) = 65.43; p < 0.001), noradrenaline (Figure 3B; F (1, 84) = 14.93; p < 0.001), and isoproterenol (Figure 3C; F (1, 84) = 66.46, p < 0.001) each caused an increase in heart rate. While in each analysis, there was a significant (adrenaline; p = 0.02) or near significant (noradrenaline and isoproterenol respectively; p = 0.06, p = 0.07) overall effect of genotype, in no case was there a genotype*agonist interaction, which indicates that adrb1 −/− larvae responded to the adrenergic agonists in a similar manner as adrb1 +/+ larvae. Thus, the extent of tachycardia in response to adrenergic agonists was independent of genotype. For example, in adrb1+/+ larvae, untreated larvae had heart rates of 196.0 ± 9.5 beats min−1 (mean ± SEM), which rose to 240 ± 9.0 beats min−1 after isoproterenol (44 beats min−1 increase). By comparison, untreated adrb1−/− larvae had heart rates of 169.5 ± 7.3 beats min−1, which rose to 219.5 ± 9.6 beats min−1 after isoproterenol (50 beats min−1 increase).

FIGURE 3

The effect of different adrenergic agonists (A, adrenaline; B, noradrenaline; C, isoproterenol) and general ß‐AR antagonists (propranolol or sotalol) on heart rate in 7 dpf adrb1 +/+ and adrb1 −/− zebrafish larvae. All drugs were applied at a final concentration of 100 μM. The effects of each agonist were analyzed in separate three‐way ANOVAs. Note that on each panel the “no agonist” bars (non‐colored), including those with antagonists, are repeated to provide control data for comparison with a given agonist; all of the data were collected in the same sessions so the use of the same controls was appropriate. ns, p > 0.05; *p < 0.05; **p < 0.01, ***p < 0.001. N values are shown by individual data points. Bars show means ± SEM.

The adrenergic antagonists had strong and statistically significant overall effects (p < 0.001, providing a source of ~50% of variation in the models), however only propranolol, and not sotalol, appeared to attenuate the responses to the agonists (Figure 3). The mechanistic basis for the superior ß‐blocking action of propranolol than sotalol remains unclear, and should be addressed in future pharmacological investigations. In each analysis, there was a genotype*antagonist interaction (p < 0.05), as propranolol abolished the differences in heart rate exhibited between the genotypes. For instance, after propranolol treatment, heart rate in adrb1 +/+ larvae was 114.0 ± 13.1 beats min−1 and in adrb1 −/− larvae was 125.5 ± 5.8 beats min−1.

DMSO (vehicle) exerted a small (accountable for 6.7% variation in the model), but statistically significant overall negative chronotropic effect on heart rate (Figure 4A; F (1, 51) = 6.60; p = 0.01). Atenolol (ß1 receptor antagonist) had no overall effect (F (1, 46) = 2.26, p = 0.14). The ß2 receptor antagonist ICI‐118551 (F (1, 54) = 79.47, p < 0.001) and the ß3 receptor antagonist SR59230A (F (1, 54) = 63.12, p < 0.001) exerted much more pronounced negative chronotropic effects (each accounting for ~44% variation in model). Isoproterenol again exerted strong and statistically significant (p < 0.001) positive chronotropic effects in both genotypes, yet there were no significant (p > 0.05) genotype*agonist or genotype*antagonist interactions in any of the models. None of the ß‐AR specific antagonists reduced the effect of isoproterenol (agonist*antagonist interactions, p > 0.05). Despite the pronounced lowering of heart rate caused by the ß3 receptor antagonist SR59230A, addition of the ß3 receptor agonist BRL‐37344 was without effect (Figure S3: F (1, 22) = 0.90; p = 0.35) in either genotype (interaction, F (1, 22) = 0.19, p = 0.67).

FIGURE 4

The effects of DMSO vehicle (A) and different adrenergic antagonists (B, atenolol, ß1‐specific; C, ICI‐118551, ß2‐specific; D, SR59230A, ß3‐specific) and the ß‐AR agonist (isoproterenol) on heart rate in 7 dpf adrb1 +/+ and adrb1 −/− zebrafish larvae. All drugs, except SR59230A (10 μM) were applied at a final concentration of 100 μM. The effects of each antagonist were analyzed in separate three‐way ANOVAs. Note that on each panel the four “untreated” bars (i.e., no antagonist) are repeated to provide control data for comparison with a given antagonist; all of the data were collected in the same sessions so the use of the same controls was appropriate. ns, p > 0.05; *p < 0.05; **p < 0.01, ***p < 0.001. N values are shown by individual data points. Bars show means ± SEM.

Heart rate responses to hypercapnia

In 5 dpf larvae, there was a strong effect of hypercapnia over time (F (8, 128) = 104.0, p < 0.001), but no overall effect of genotype (F (1, 16) = 1.68, p = 0.21) or genotype*time interaction (F (8, 128) = 0.13, p > 0.99) (Figure 5A). The maximum effect manifested in the first measurement period after hypercapnia exposure commenced (Figure 5A), resulting in a 75.7 ± 7.7 beats min−1 acceleration of heart rate in adrb1 +/+ and a 79.2 ± 12.6 beats min−1 acceleration of heart rate in adrb1 −/− larvae (Figure 5B). At 6 dpf, there was also a significant effect of hypercapnia exposure over time (F (8, 152) = 87.30, p < 0.001) and additionally an overall effect of genotype (F (1, 19) = 7.884, p = 0.01), whereby starting heart rates were ~20 beats min−1 lower in adrb1 −/− than adrb1 +/+ larvae. Notably, there was also a genotype*time interaction (F (8, 152) = 4.162, p < 0.001) which could be ascribed to the fact that adrb1 +/+ larvae exhibited a greater peak increase (p < 0.05) in heart rate at the onset of hypercapnia than adrb1 −/− (78.3 ± 8.0 beats min−1 vs. 51.0 ± 7.2 beats min−1 increase) (Figure 5D).

FIGURE 5

The effect of sustained hypercapnia (10 min) on heart rate in 5 (A, B) and 6 (C, D) dpf adrb1 +/+ and adrb1 −/− zebrafish larvae. Statistical results are the output of a repeated‐measures two‐way ANOVA (A, C). In panels B and D, the maximum changes in heart rate were calculated for each individual immediately after the onset of hypercapnia; genotypes were compared using unpaired t‐test. ns, p > 0.05; *p < 0.05; **p < 0.01, ***p < 0.001. N = 9–11. Data points show means ± SEM (A, C), bars show means ± SEM and individual points (B, D).

The effects of isoproterenol on isolated hearts in vitro

Our series of in vivo procedures indicated that the hearts of adrb1 −/− larvae retained the capacity to respond to adrenergic stimulation with an increase in heart rate. However, this approach could not distinguish whether the tachycardia arose from a direct effect on the heart, as opposed to affecting vascular tone, which could affect heart rate indirectly through possible changes in cardiac filling. To resolve this, we studied the effects of isoproterenol on heart rate in the spontaneously‐beating isolated heart in vitro. Because it was not possible to reliably perform these experiments on small larval hearts, we instead used hearts from adult (3 month post‐fertilization) adrb1 +/+ and adrb1 −/− zebrafish. There was no significant difference (t = 0.51, df = 10, p = 0.62) in body mass of adult fish between the genotypes (adrb1 +/+, 0.20 ± 0.03 g; adrb1 −/−, 0.24 ± 0.07 g). There was an overall significant positive chronotropic effect of isoproterenol treatment (F (1, 10) = 17.1, p = 0.002) but no effect of genotype (F (1, 10) = 0.01, p = 0.92) or isoproterenol*genotype interaction (F (1, 10) = 0.37, p = 0.56), showing that the isolated adrb1 −/− heart responded to adrenergic stimulation in an identical manner to adrb1 +/+ (Figure 6).

FIGURE 6

The effect of isoproterenol (final concentration 1 μM) on heart rate in spontaneously‐beating isolated hearts from adrb1 +/+ and adrb1 −/− adult (3 months post‐fertilization) zebrafish. Statistical results are the output of a repeated‐measures two‐way ANOVA. ns, p > 0.05; *p < 0.05; **p < 0.01, ***p < 0.001. N = 5–7. Data are presented as individual data points.

Cardiac gene expression in larval zebrafish

To understand how adrb1 −/− larvae were capable of preserved responses to adrenergic agonists and hypercapnia (which was blunted but not abolished at 6 dpf), we measured the expression of each of the five zebrafish adrb genes (adrb1, adrb2a, adrb2b, adrb3a, and adrb3b) in the hearts of 7 dpf adrb1 +/+ and adrb1 −/− larvae (Figure 7). This experiment verified that, as a result of the genomic deletion, adrb1 expression was abolished in the hearts of adrb1 −/− larvae. While expression of adrb2a and adrb3a was unchanged in the adrb1 knockouts, both adrb2b (t = 2.63, df = 10, p = 0.03) and adrb3b (t = 3.04, df = 10, p = 0.01) expressions were increased in adrb1 −/− relative to adrb1 +/+ larvae, demonstrating differential regulation of gene transcription, potentially as a form of genetic compensation, for loss of the ß1‐AR.

FIGURE 7

Expression of five adrb genes in 7 dpf adrb1 +/+ and adrb1 −/− zebrafish hearts analyzed with qPCR. Statistical comparisons made with unpaired t‐test. *p < 0.05; ***p < 0.001. For each sample N = 6 pools of 20 hearts. Bars show means ± SEM and individual data points.

Respirometry

To ascertain whether the reduced routine heart rates of adrb1 −/− larvae (Figure 2) were associated with altered oxygen consumption during normoxia or hypoxia (measured as critical oxygen partial pressure; Pcrit), microrespirometry was conducted on 7 dpf adrb1 +/+ and adrb1 −/− larvae. Neither routine oxygen consumption (t = 0.18, df = 42, p = 0.86) nor Pcrit (t = 1.06, df = 34, p = 0.30) was different between genotypes (Figure 8).

FIGURE 8

Routine oxygen consumption (A) and critical oxygen partial pressure (Pcrit; B) in adrb1 +/+ and adrb1 −/− zebrafish determined with respirometry. Oxygen consumption measurements N = 22, successful Pcrit measurements N = 17–19. There were no significant differences in either parameter between genotypes. Bars show means ± SEM and individual data points.

DISCUSSION

Although the results of some previous studies , , suggested that the ß1‐AR (encoded by adrb1) is largely responsible for positive chronotropic effects of adrenergic stimulation in zebrafish, as is believed to be the case in mammals, this notion is in contrast to the long held dogma that ß2‐ARs dominate in the fish heart. Our results indicate that, while deletion of adrb1 reduces routine heart rate, it is not essential for positive chronotropic responses to pharmacological adrenergic stimulation or environmental stressors that increase adrenergic tone. The preserved adrenergic tachycardia may be facilitated by upregulation of other ß‐ARs, namely those encoded by adrb2b and/or adrb3b.

Unlike in mice, in which deletion of the adrb1 results in high embryonic lethality, we observed that adrb1 −/− zebrafish survived in expected Mendelian ratios and appeared outwardly as phenotypically normal and viable adults. Body mass of 7 dpf larvae (see Respirometry methods) and 3 month post‐fertilization adults (see The effects of isoproterenol on isolated hearts in vitro Results) were identical. Despite exhibiting lower heart rates, 7 dpf adrb1 −/− larvae exhibited equivalent rates of oxygen consumption to their adrb1 +/+ cousins. As internal convection is essential for normal aerobic metabolism, even in early larval zebrafish, the maintained rate of oxygen consumption may be supported by increased stroke volume and/or elevated arterio‐venous oxygen extraction, according to the Fick principle. ,

In agreement with previous work in zebrafish, , heart rate acutely rose from 2 to 5 dpf, a pattern which has been attributed to increased circulating catecholamine levels exerting increasing ß‐adrenergic tone. Thus, a blunting of cardiac acceleration during early development was anticipated in adrb1 −/− larvae, yet they exhibited a parallel rise in heart rate. Either the potentially increased adrenergic tone can stimulate other ß‐ARs or the increase in heart rate may represent intrinsic remodeling on the sino‐atrial pacemaker. There was a clear difference in heart rate between genotypes even at 2 dpf, which is consistent with studies showing that heart rate is increased with isoproterenol treatment at 2 or 3 dpf. , , Importantly, the finding of the current study of reduced heart rate in the adrb1 −/− larvae at 2 dpf indicates that cardiac ß‐adrenergic tone may start earlier than traditionally believed (after 5 dpf) as suggested by onset of propranolol sensitivity. This is consistent with the previous finding that some of the enzymes necessary for catecholamine synthesis (i.e. tyrosine hydroxylase and dopamine ß‐hydroxylase) are expressed by 2 dpf, at which point endogenous whole‐body noradrenaline also becomes measurable in larval zebrafish.

Arguably, the most striking finding of the present study was that deletion of adrb1 did not alter the chronotropic responses to ß‐adrenergic receptor agonists, clearly indicating that other receptors can mediate chronotropic responses to catecholaminergic stimulation in the zebrafish heart. Because the pharmacological treatments may have induced supra‐physiological levels of adrenergic stimulation, we subsequently studied the heart rate responses to an environmentally relevant stressor, hypercapnia, which is believed to increase adrenergic tone. However, adrb1 −/− larvae exhibited the typical increase in heart rate during hypercapnia, identical to adrb1 +/+ at 5 dpf albeit significantly blunted at 6 dpf (Figure 5D). The unaltered response to CO2 at 5 dpf was unexpected given that in a previous study, the hypercapnic tachycardia at this developmental time point was abolished by morpholino knockdown of adrb1 or pre‐treatment with an ß1‐AR‐specific ß‐blocker (atenolol). Because the zebrafish heart can respond to increased stretch (i.e. hemodynamic load) with an increased beating rate, and adrenergic stimulation may increase venous return independently of activating cardiac ß1‐ARs, we considered the possibility that the preserved responses of the adrb1 −/− line to adrenergic stimulation could be attributable to an indirect effect on the heart. However, we showed that the spontaneously beating isolated heart from adult adrb1 −/− zebrafish responded to isoproterenol stimulation in an identical manner to adrb1 +/+, confirming that the knockout indeed displays a preserved intrinsic capacity to respond to adrenergic stimulation with an increase in heart rate (Figure 6). Because this experiment was conducted in adult hearts, it suggests the general mechanisms we have uncovered in larvae are likely translatable to later life stages. However, in the future, dedicated experiments will be required to understand fully cardiac control in adult adrb1 −/− zebrafish.

We hypothesize that the sustained hypercapnic tachycardia is a consequence of genetic compensation, i.e. upregulation of related receptors to mitigate the effect of knockout. For instance, it was previously shown that induction of a deleterious mutation of the unrelated gene, egfl7, but not morpholino knockdown, induced genetic compensation in zebrafish. In support of this hypothesis, we showed that cardiac expression of adrb2b and adrb3b mRNA was increased in adrb1 −/− larvae.

The role of ß3‐ARs in the fish heart is enigmatic. Although we observed marked bradycardia in response to the ß3‐antagonist SR59230A, application of the ß3‐agonist BRL 37344 was without effect. BRL 37344 was previously shown to increase heart rate in trout larvae at the same concentration. It should be noted that the specificity of both of these drugs has been questioned, even in mammals, , where they may exert their effects on other ß‐receptor subtypes; thus, their effects should be interpreted cautiously.

The contribution of zebrafish ß2‐ARs to cardiac control also is ambiguous. Previous studies have suggested that ß2‐ARs exert negative chronotropic effects, given that morpholino knockdown of adrb2a and adrb2b increased heart rate and treatment with reputed ß2‐specific antagonist procaterol decreased heart rate. , The latter finding is dubious given that procaterol affects other zebrafish ß‐ARs non‐selectively. For example, procaterol behaves as an agonist for zebrafish ß1‐ARs in transfected HEK293 cells. There is conflicting evidence for a positive chronotropic action of zebrafish ß2‐ARs. In the current study, it was demonstrated that heart rate was reduced by treatment with a putative (but not yet pharmacologically confirmed in zebrafish) ß2‐specific antagonist ICI‐118551. Importantly, it was shown previously that activation of zebrafish ß2‐ARs increases intracellular cAMP akin to ß1‐AR stimulation. In pacemaker cardiomyocytes, such an increase in cAMP would be expected to activate hyperpolarization‐activated cyclic nucleotide–gated (HCN) channels, which are known to determine heart rate in larval zebrafish, and increase heart rate. The possible stimulatory effect of ß2‐ARs is consistent with decades of research suggesting that ß2‐ARs are primarily responsible for the binding of catecholamines and positive inotropic effect on atrial and ventricular myocardium in other fish species. , , , , , Given that both positive chronotropic and positive inotropic effects largely stem from similar mechanisms dependent on cAMP accumulation, it is reasonable to predict that ß2‐ARs could exert stimulatory effects in pacemaker, atrial, and ventricular cardiomyocytes alike.

This study provides mechanistic insights into a fundamental difference in adrenergic cardiac control between zebrafish and mammals. In zebrafish, the ß1‐AR is dispensable for tachycardia, possibly due to redundancy with other ß‐ARs. With the available evidence, it is not possible to conclude whether upregulation of ß2 or ß3‐ARs (or potentially other unidentified non‐adrenergic receptors) are most important in mitigating the effects of adrb1 knockout. The results of the current study also illustrate the difficult challenge of reconciling knockdown, knockout, and pharmacological approaches in studies of cardiovascular control in zebrafish.

MATERIALS AND METHODS

Experimental animals and husbandry

All experiments were carried out in accordance with animal care guidelines provided by the Canadian Council on Animal Care and with prior approval from the University of Ottawa Animal Care Committee (Protocol BL‐226). Wild‐type zebrafish were sourced from the in‐house stock at the University of Ottawa and maintained under standard conditions in 3 or 10 L tanks in a recirculating system supplied with aerated, dechloraminated City of Ottawa tap water (hereafter “system water”) maintained at 28°C with a 14:10‐h light–dark cycle. Fish were fed once or twice daily with a commercial zebrafish diet. Embryos (first for the development of the knockout line and later for maintenance of the line and to supply larvae for experiments) were attained by collecting small groups of 2–3 male and 2–3 female (except where stated otherwise) adult fish in sloped 2 L static tanks overnight with a perforated base insert to allow egg collection after spawning the following morning. Larval zebrafish were fed brine shrimp nauplii and a standard commercial diet from 7 dpf.

Generation of adrb1 line

We aimed to delete the entire single exon of the zebrafish adrb1 gene with the use of CRISPR/Cas9 by using guide RNAs (gRNAs) targeting sequences both upstream and downstream of the gene (NCBI accessions: Gene ID 557194 within genomic region NC_007123.7, chromosome 12). The guide portions of four gRNAs were designed using the CHOPCHOP program , and four corresponding target (T1‐4) short‐guide oligos were developed (Table 1). A DNA construct for each gRNA was generated using a PCR‐based (cloning‐free) method with a guide constant oligo (Table 1) as described previously. , Each 50 μl PCR reaction consisted of 5 μl of 10× DreamTaq buffer (Thermo Scientific), 0.25 μl of 5 U/μl DreamTaq DNA Polymerase (Thermo Scientific), 1 μl of 10 mM dNTPs, 1 μl of 1 μM guide constant oligo, 1.25 μl of 10 μM gDNA primer pair, 40.5 μl dH2O, and 1 μl of the 1 μM target‐specific short‐guide oligo. The thermal profile of the reaction was 94°C for 3 min, 40 cycles of 94°C for 40 s, 55°C for 1 min, and 72°C for 45 s, followed by 72°C for 10 min. The PCR product was purified with the GeneJet PCR purification kit (Thermo Scientific) following the manufacturer's protocol. gRNAs were synthesized using a HiScribe T7 high‐yield RNA synthesis kit (New England Biolabs, Ipswich, MA) according to the manufacturer's instructions, followed by RNA purification with ethanol precipitation. sgRNA size and quality were verified by gel electrophoresis and concentrations measured using a NanoDrop 1000 spectrophotometer (ThermoFisher, Waltham, MA, USA).

TABLE 1

Oligonucleotides used for the generation of CRISPR guide RNAs or as screening primers

Primer/oligo name	Sequence (5′‐3′)
CRISPR/Cas9 gRNA synthesis
T1	GCG TAA TAC GAC TCA CTA TAG GCG TAA AGT AAA ACC CGA AGG TTT TAG AGC TAG AAA TAG
T2	GCG TAA TAC GAC TCA CTA TAG GAG CCA AGA GCT TGT TTT GGT TTT AGA GCT AGA AAT AGC
T3	GCG TAA TAC GAC TCA CTA TAG GTT TTC AGT TGG TTC AAC GGT TTT AGA GCT AGA AAT AGC
T4	GCG TAA TAC GAC TCA CTA TAG GTT TTC AGT TGG TTC AAC GCG TTT TAG AGC TAG AAA TAG
Guide constant	AAG CAC CGA CTC GGT GCC ACT TTT TCA AGT TGA TAA CGG ACT AGC CTT ATT TTA ACT TGC TAT TTC TAG CTC TAA AAC
gDNA primer 1 (F)	GCG TAA TAC GAC TCA CTA TAG
gDNA primer 2 (R)	AAA GCA CCG ACT CGG TGC CAC
Genomic DNA screening
Screening F1	TAA AGC TTC CCA CGT TCA TGG
Screening R1	CTG GAA GCC AAT GAA GGA AAG

Note: Bolded text represents specific guide portion of target (T1‐4) short‐guide oligos.

A schematic step‐by‐step overview of the protocol used to generate the adrb1 −/− line is provided in Figure S1. Microinjections were performed in 1‐cell stage embryos as described previously. The injection solution comprised of the following: 40 ng/μl for each sgRNA, 20 ng/μl Cas9 protein (New England BioLabs, Ipswich, MA, USA), and 0.1% Phenol Red (Sigma, Burlington, MA, USA).

The primary injected embryos were raised to sexual maturity (i.e. F0 population) and were then individually outcrossed with naïve wild‐types of the opposite sex to identify founder fish carrying a germline deletion of adrb1. Juveniles (~60 dpf) from these crosses were genotyped with screening primers F1 and R1 (Table 1). DNA was extracted from fin clips (collected following brief anesthesia in buffered tricaine mesylate (MS‐222; Syndel Laboratories, Nanaimo, BC, Canada; 100 mg l−1) in 20 μl of 50 mmol l−1 NaOH at 95°C for 10 min followed by neutralization with 2 μl of 1 mol l−1 Tris–HCl (pH 8). PCR amplification proceeded under the following conditions: 200 μmol l−1 dNTPs, 2.5 μl 10× DreamTaq buffer (ThermoFisher), 0.125 μl DreamTaq Hot Start DNA polymerase (ThermoFisher), and 5 μl of DNA digest, with a temperature cycle of 94°C for 3 min, and 35 cycles of 94°C for 30 s, 55°C for 1 min, and 72°C for 45 s, and a final 72°C step for 10 min. This amplified a ~300 bp fragment in fish carrying a whole‐exon deletion of adrb1. A single founder was identified that reliably produced embryos (~15% fertilized eggs) carrying the adrb1 deletion when crossed with a wild‐type. Sanger sequencing (Genome Quebec, McGill University, Montreal, Canada) of the PCR product and alignment with the wild‐type genomic sequence (Supplementary Material) confirmed a 2462 bp deletion encompassing the entire single exon (Figure 1A). Fourteen F1 heterozygotes (adrb1 +/−) containing identical deletions were then raised to sexual maturity and in‐crossed to generate an F2 generation.

For screening of the F2 generation (fin clipping at ~60 dpf), we used Phire Tissue Direct PCR Master Mix kit (Thermo Scientific), according to the manufacturer's instructions. The thermal profile of the reaction was 98°C for 5 min followed by 40 cycles of 98°C for 5 s, 62°C for 5 s, 72°C for 1 min, and finally 72°C for 5 min. The enhanced polymerase of the Phire Hot Start II DNA Polymerase (compared with Taq polymerase) allowed not only amplification of the ~300 bp mutant band but additionally the ~3000 bp amplicon of the wild‐type gene (Figure 1B). We confirmed that DNA extracted from naïve wild‐type fish exhibited the ~3000 bp amplicon only and known F1 heterozygotes exhibited both the mutant (~300 bp) and wild‐type (~3000 bp) amplicons. Screening of the F2 generation allowed us to identify and separate adrb1 −/−, adrb1 +/−, and adrb1 +/+ individuals at 60–90 dpf.

Screening of F1 heterozygous crosses for the Mendelian ratio of offspring

In addition to using F1 heterozygotes (adrb1 +/−) to form a stable F2 adrb1 −/− line, we performed a dedicated experiment to record the ratio of adrb1 −/−, adrb1 +/−, and adrb1 +/+ individuals in the offspring of F1 incrosses. Four separate F1 incrosses, each consisting of one adult male and one adult female, were performed and their offspring were raised separately. All of the fish in this experiment were terminally sampled at 7 and 28 dpf with MS‐222 overdose (300 mg l−1) because it is difficult to ensure full recovery after the invasive fin clipping procedure in sub‐juveniles. Individual larvae were screened using the Phire Tissue Direct PCR Master Mix kit as described above.

Larval heart rate measurements

For all of the experiments on larval heart rates, gene expression, and respirometry, we compared F3 adrb1 and adrb1 +/+ “cousins” generated from incrossing the screened adrb1 −/− or adrb1 +/+ F2 populations (Figure S1). By using adrb1 +/+ siblings as breeders, we ensured that genetic background was controlled (Zimmer et al., 2019 ).

Except where described otherwise (hypercapnia experiment), heart rates were measured optically in zebrafish larvae as described previously using a USB microscope (Firefly GT805; Firefly Global Belmont, MA, USA), connected to a personal computer, at a resolution of 640x480 pixels and a frame rate of 32.8 frames per second. Individual larvae, maintained in a holding vessel in a water bath at 28°C, were gently but rapidly pipetted into capillary tubes and video acquisition commenced immediately. This method has previously been shown to allow the measurement of a “routine” heart rate in larval zebrafish.

Routine heart rate across development

We first measured routine (unanaesthetised) heart rates in adrb1 and adrb1 +/+ larvae from 2 to 21 dpf. At 2 and 3 dpf the larvae were manually dechorionated and allowed at least 30 min to recover. Measurements could not be made after 21 dpf as the heart became too difficult to observe reliably due to reduced transparency of the cuticle. Each larva was measured at only one age (non‐repeated measure design) to avoid risks of habituation.

Heart rate responses to adrenergic receptor antagonists and agonists

The protocols that investigated the cardiac response to exogenously applied adrenergic receptor agonists and antagonists solely employed 7 dpf larvae, an age at which previous studies have revealed a robust chronotropic response to isoproterenol. To facilitate measurements, the larvae were lightly sedated with 50 μg ml−1 MS‐222 during the drug exposures (15 min ± antagonist followed by 10 min ± agonist in continued presence of antagonist) and subsequent heart rate measurement. Larvae were first isolated in open top 1.5 ml centrifuge tubes (holding vessel) containing 1 ml of system water containing the anesthetic.

The first protocol investigated the effects of 15 min pre‐treatment with general ß‐receptor blockers propranolol hydrochloride (P0884 Sigma‐Aldrich Canada) or sotalol hydrochloride (S0278 Sigma‐Aldrich Canada) followed by 10 min treatment with one of three adrenergic receptor agonists: isoproteronol bitartrate (I2760 Sigma‐Aldrich Canada), adrenaline bitartrate (E4375 Sigma‐Aldrich Canada), and noradrenaline bitartrate (A0937 Sigma‐Aldrich Canada). Controls were also run with no antagonist and/or agonist. All drugs were used at 100 μM to be consistent with extensive previous work in zebrafish larvae. , , When administering pharmacological agents in the external water, the final internal concentration reaching cardiac receptors is unknown, and at present it represents an unsurmountable challenge to reliably measure circulating (internal) concentrations empirically given the small body size and minute blood volumes of larval zebrafish. Nevertheless, because the dermis of larval fish is thin and highly vascularized, and provides a relatively large surface area for diffusion, this method provides a convenient and reliable route to administer the drugs. It certainly would not be expected that drug penetration would be different in adrb1 and adrb1 +/+ larvae, so thus remains a valid treatment for the purposes of our study. The drugs were dissolved in dH2O at a stock concentration of 10 mM and 10 μl was added to the 1 ml system water in the holding chamber.

For the second protocol, larvae were pre‐treated for 15 min in the absence or presence of a ß‐AR sub‐type‐specific antagonist; atenolol (β1 antagonist, 100 μM ; A7655 Sigma‐Aldrich Canada), ICI‐118551 hydrocholoride (β2 antagonist, 100 μM ; I127 Sigma‐Aldrich Canada), or SR59230A (β3 antagonist, 10 μM ; S8688 Sigma‐Aldrich Canada). Owing to the low water solubility of atenolol and SR59230A, all of these antagonists were dissolved in dimethylsulfoxide (DMSO, Sigma‐Aldrich Canada) to make a 10 mM stock, or 1 mM for SR59230A, and 10 μl of the stock solution was added to the 1 ml system water in the holding vessel. To account for the effects of the vehicle, a series of experiments was conducted with DMSO only, although it has also previously been shown that 1% DMSO has little effect on heart rate in zebrafish larvae. After the 15 min incubation, isoproteronol (100 μM final concentration provided as 10 μl of 10 mM stock in dH2O to a final volume of 1 ml) was added to half of the larvae. Isoproteronol treatment, in the continued presence of the antagonist where applicable, was again maintained for 10 min before heart rate was measured.

As we observed strong effects of ß3 adrenergic receptor blockade, we also studied the effects of 10 min treatment with 100 μM ß3 adrenergic receptor agonist BRL 37344 (B169, Sigma‐Aldrich Canada). This agonist was dissolved in DMSO to a stock concentration of 20 mM and 5 μl was added to the 1 ml holding vessel. Parallel experiments were conducted with the same volume of DMSO (final concentration 0.5%) alone.

The effect of hypercapnia on larval heart rate

To provide consistency with comparable previous work, this experiment used a different approach to measure heart rate. FIve or six dpf larvae were enclosed in a glass capillary tube, closed at each end with mesh, connected to a reservoir to provide a gravity‐fed flow‐through (1 ml min−1) of system water containing MS‐222 (50 μg ml−1) to attain light sedation. The temperature of the flow‐through water was maintained at 28–29°C, measured at the level of the capillary tube. The heart was observed and recorded with an iPhone SE (Apple, Cupertino, CA, USA) mounted onto the eyepiece of a dissection microscope (Zeiss Discovery V8; Zeiss, Oberkochen, Germany). Larvae were allowed to rest for 10 min in normocapnic system water before the experiment commenced. Heart rate was measured every 2.5 min with 10 s videos. Larvae were first exposed to normocapnic water for 3 min, with measurements taken after 30 s and 2.5 min. They were then exposed to pre‐equilibrated hypercapnic system water (1% CO2 in air) for 10 min and measurements commenced immediately upon exposure to hypercapnia and at 2.5 min intervals thereafter. At the end of the 10 min hypercapnia exposure, the procedure was repeated to reinstate normocapnia for 5 min. Heart rates were analyzed in ImageJ and LabChart as described previously. Time‐matched parallel controls were run with larvae maintained in normocapnic water.

Effects of adrenergic stimulation on isolated heart

Hearts from adult (3 month post‐fertilization) adrb1 and adrb1 −/− zebrafish were harvested to allow the direct effect of adrenergic stimulation to be studied in the spontaneously beating heart, as described previously. Briefly, isolated hearts were video recorded, allowing heart rate to be measured, in a 5 ml Petri dish in physiological saline (mM: NaCl, 150; KCl, 2.5; MgSO4, 1.5; NaH2PO4, 0.4; glucose, 10; HEPES, 10; CaCl2, 1; pH set to 7.7 with NaOH) at 28°C before and 5 min after isoproterenol bitartrate (final concentration of 1 μM) treatment.

Gene expression (qPCR)

Real‐time PCR was performed to assess the expression of adrb genes in the heart of 7 dpf adrb1 and adrb1 −/− zebrafish. Larvae were euthanized by chilling in an ice bath and transferred onto a microscope slide. Using a micro‐scalpel, two perpendicular excisions were made to excise the heart, one following the orobranchial cavity just beneath the brain and the other perpendicular to this excision immediately behind of the heart and in front of the yolk sac. Hearts were pooled and homogenized (20 hearts per sample) using a hand‐held homogenizer in 0.5 ml Trizol (15 596 018 Invitrogen). RNA was extracted from these homogenates following the manufacturer's protocol. RNA concentration was determined using a NanoDrop™ 2000 spectrophotometer (Thermo Scientific). RNA was treated with DNaseI (18 068 015 Invitrogen) prior to cDNA synthesis from 0.5 μg total RNA using a high‐capacity cDNA reverse transcription kit with RNase inhibitor (Applied Biosystems, Cat#: 4374967). Real‐time PCR was performed in 20 μl reactions containing 10 μl of SsoFast EvaGreen supermix (1725201 Bio‐Rad), 5 μl of 2 μM forward and reverse primer pairs (Table 2), and 5 μl of cDNA template (0.5 μl reverse transcribed cDNA added to 4.5 μl of ddH2O). Controls (no template added) were run in every assay, and non‐reverse transcribed samples were run for every primer pair. Reaction efficiency for all primer pairs was between 90% and 110%. Relative expression of all adrb genes was normalized to the geometric mean of actβ1, eef1α1l1, and gapdh (Ctadrb1+/+ = 16.993, Ctadrb1−/− = 16.995 [Ct, cycle threshold]) serving as reference genes and then expressed relative to the expression level of adrb1 in the adrb1 group.

TABLE 2

Primers used for real time quantitative PCR

Gene	Orientation	Sequence
eef1α1l1	F	CGTTGAGAAGAAAATCGGTG
	R	CCAGTCCTTAAGTAGAGTGC
actβ1	F	CCCCTTGTTCACAATAACCT
	R	CCCACATAGGAGTCTTTCTG
gapdh	F	CAACCAAATCAGGCATAATGG
	R	AATCAAGGTCAATGAATGGGT
adrb1 a	F	GCTGGGAATAATCATGGGAA
	R	GCTCCTTATCCACCACTTG
adrb2a a	F	GTCACGCTATCCTAACGTCA
	R	ATTCCTCTTTCGCCAAGTTC
adrb2b a	F	AAGCCTTTGAACCAAGATG
	R	GCCTTTCCAAATATGTCCTG
adrb3a a	F	GTTTCTCATTGCCACACGA
	R	ACTCTTCCTCTTTGCTGTCA
adrb3b a	F	GACTCTTGTGAAATTCCTGAAG
	R	GACTGAAGATGCCCATGATAA

Note: All sequences listed 5′ to 3′ with the reverse primer sequences listed as the reverse compliment of the gene sequence.

Primer sequences obtained from Ref. [21].

Routine metabolic rate (RMR) and critical oxygen tension (Pcrit) were measured in 7 dpf adrb1 +/+ and adrb1 −/− larvae using closed system respirometry. Larvae were placed into individual 80 μl respirometry wells of a 24‐well glass microplate (Loligo Systems, Viborg, Denmark) fitted with O2 sensor spots. Each trial consisted of 11 adrb1 +/+ and 11 adrb1 −/− larvae, along with 2 blanks for background control. The microplate was sealed with adhesive plate seals (AB0580, ThermoFisher Scientific, Mississauga, Canada) and maintained at 28°C using a water bath. The sealed microplate and water bath were attached to an O2 fluorescence sensor (SDR SensorDish Reader, PreSens, Regensburg, Germany), and PO2 levels were measured until they plateaued, at which point the experiment was terminated. Larval fish weights were obtained from the same batches of larvae used for the respirometry experiment. Fish (n = 20) were pooled into a single 20 μm cell strainer (pluriSelect, San Diego, USA) and excess water was removed by centrifugation (500× g). Weights were measured using an analytical balance and six pooled weights were obtained for each genotype. No differences were observed for the weights between adrb1 +/+ (0.161 ± 0.005 mg per larva) and adrb1 −/− (0.161 ± 0.008 mg per larva) larvae, and thus the average weight for both genotypes was used for downstream ṀO2 (μmol g−1 h−1) calculation. ṀO2 and Pcrit data were analyzed using the calc_rate() and pcrit() functions from the respR package in R (https://www.r‐project.org/). ṀO2 was obtained from the linear decline of PO2 during the course of the experiment, beginning from ~15 min into the experiment until ~2 h. Pcrit data are reported as the value obtained from the segmented method in the analysis. In addition, Pcrit regression lines were inspected “blindly” for quality control, and only those deemed to be valid traces were retained. Traces that were excluded either did not have a stable oxy‐regulatory phase in the PO2 versus ṀO2 curve to fit a tightly fitted line or results obtained from the broken‐stick method and the segmented method using the pcrit() function differed substantially.

Statistical analyses

Statistical analysis and graph construction was completed in GraphPad Prism (v. 9.3.1; GraphPad Software Inc., San Diego, CA, USA). Chi‐square tests were used to compare observed ratios of adrb1 −/−, adrb1 +/−, and adrb1 +/+ from the F1 incross with expected Mendelian ratio (25%, 50%, 25%) at 7 dpf, 28 dpf, and the age groups pooled. A two‐way analysis of variance (ANOVA) was used to compare heart rates in adrb1 −/− and adrb1 +/+ larvae (genotype) and age from 2 to 21 dpf. Three‐way ANOVAs were used to analyze the effects of genotype, agonist (adrenaline, noradrenaline, or isoproterenol), and antagonist (propranolol and sotalol) for pharmacological protocol 1, and genotype, agonist (isoproterenol), and antagonist (DMSO vehicle, atenolol, ICI‐118551, or SR59230A) for pharmacological protocol 2. A two‐way ANOVA was used to analyze the effects of genotype and BRL‐37344. A repeated‐measures two‐way ANOVA was used to analyze the effects of hypercapnia over time and genotype. A repeated‐measures two‐way ANOVA was used to analyze the effects of isoproterenol treatment and genotype on in vitro isolated hearts from adult adrb1 −/− and adrb1 +/+ zebrafish, the body masses of which were compared with an unpaired t‐test. Unpaired t‐tests were used to compare the expression of each adrb gene, oxygen consumption, and Pcrit in adrb1 −/− and adrb1 +/+ larvae. Data are presented as means ± standard error of the mean (SEM).

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

Figure S1

Click here for additional data file.

Appendix S1

Click here for additional data file.