Thyroid hormone regulates muscle function during cold acclimation in zebrafish (Danio rerio)

Outside an optimal range, temperature can severely limit performance by compromising the biochemical reaction rates that support whole-animal physiology (Somero, 1995; Guderley, 2004a; Seebacher and Franklin, 2011; Deslauriers and Kieffer, 2012; Herrel and Bonneaud, 2012). The thermal breadth of performance can be shifted and broadened by acclimation, which has been described in many ectothermic vertebrates (Hwang et al., 1990; Johnson and Bennett, 1995; Johnston and Temple, 2002; Seebacher et al., 2003; Guderley, 2004b; O'Brien, 2011). We have shown previously that thyroid hormone (TH) influences locomotor performance during cold acclimation in zebrafish, and that its actions are temperature sensitive (Little et al., 2013). An important outstanding question now is how TH affects locomotion.

TH is ubiquitous across vertebrates (Hulbert and Else, 1981; Norris, 2007), but most research has focused on its thermoregulatory role in mammals. In non-mammalian vertebrates, TH is predominately recognized as a regulator of growth and development (Norris, 2007; Tata, 2011), but there is growing evidence that TH also stimulates metabolism in adult ectotherms (Bouzaffour et al., 2010; Goglia, 2005; Hulbert, 2000; Johnson and Lema, 2011; Nelson and Habibi, 2009). In mammals, TH regulates muscle function by influencing proteins involved in calcium cycling and the contractile machinery of the muscle. It is unknown whether these muscle-specific roles are also conserved in ectotherms (Izumo et al., 1986; Simonides et al., 2001; Forini et al., 2001; Ketzer et al., 2009; Simonides and van Hardeveld, 2008). Our aim was to determine whether TH controls proteins involved in muscle contraction and relaxation, and swimming kinematics during cold acclimation in fish.

Muscle performance is a function of excitation–contraction coupling and relaxation, and the energy (ATP) available for cross-bridge formation and calcium (Ca²⁺) pumping by sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA). Low temperatures can reduce overall muscle function by reducing ATP synthesis, muscle power output and fatigue resistance (Prezant et al., 1990; Johnston and Ball, 1997; Guderley, 2004b; Allen et al., 2008; Seebacher and James, 2008; James et al., 2012). However, many fish can overcome these negative effects of temperature by acclimating muscle function. Cold-acclimated fish can have contraction velocities up to double those of warm-acclimated fish, when tested at a common temperature (Heap et al., 1985; Fleming et al., 1990; Woytanowski and Coughlin, 2013). Increased muscle function during cold acclimation could be a result of increases in the activity of the myofibrillar ATPase (Hwang et al., 1990; Sänger, 1993) and changes to the primary structure and/or isoform composition of the myosin heavy chain proteins (Hwang et al., 1990) that determine muscle fibre type. The calcium-handling proteins that regulate contraction could also drive increases in muscle performance. SERCA activity increased ~60% in carp acclimated to cold conditions (Fleming et al., 1990). Regulation of SERCA has been linked to enhanced swimming performance in fish (Seebacher and Walter, 2012). It is therefore conceivable that SERCA activity may be regulated to maintain muscle function during cold acclimation. In rat cardiac and skeletal muscle, TH increases the expression and alters the isoform composition of SERCA while reducing concentrations of its inhibitor, phospholamban (Carr and Kranias, 2002; Simonides and van Hardeveld, 2008). Increased SERCA activity can accelerate muscle relaxation by increasing the rate of post-contraction sarcoplasmic reticulum (SR) Ca²⁺ sequestration (Seebacher et al., 2012). Additionally, SERCA replenishes SR Ca²⁺ stores to support subsequent contractions and promote fatigue resistance (Carr and Kranias, 2002). TH also regulates myosin heavy chain expression levels, and typically shifts isoform composition toward a fast-twitch phenotype (Izumo et al., 1986; Weiss and Leinwand, 1996; Simonides and van Hardeveld, 2008). Myosin filaments can have different catalytic activities depending upon the heavy chain isoforms expressed. For example, in mammals the fast myosin heavy chain isoforms (FMHCs) confer three to five times faster contraction velocities, but are less efficient than the slow myosin heavy chain isoforms (Purcell et al., 2011).

In the present study, we tested the three hypotheses that hypothyroidism (1) reduces SERCA and myofibrillar ATPase mRNA concentrations and activities. As a consequence, muscle force production and relaxation rate would decrease, which would (2) reduce tail beat frequencies and amplitudes. SERCA activity is modulated by phospholamban, so we tested the hypothesis (3) that TH regulates the mRNA expression of phospholamban, and muscle-specific transcription factors. Kruppel-like factor 15 (KLF-15), a transcription factor important in muscle metabolism, has recently been shown to regulate muscle performance during exercise in mice (Haldar et al., 2012). Some members of the Kruppel-like family of transcription factors are also responsive to TH (Pei et al., 2011), thus we were also interested in whether TH regulates KLF-15 muscle mRNA levels in response to cold exposure. We used a multifactorial experimental design in which we induced hypothyroidism, followed by supplementation with T₂ and T₃ (+ controls) in zebrafish exposed to different chronic and acute temperature combinations.

Zebrafish [Danio rerio (Hamilton 1822)] were purchased from commercial suppliers (Kim's Aquatic World, Sydney, NSW, Australia; Livefish, Bundaberg, QLD, Australia). After a settling-in period of 10 days at 23°C, fish were split into two temperature treatments – a cold acclimation group at 18°C, and a warm acclimation group at 28°C – and held at these temperatures (±0.5°C) for 3 weeks. Within acclimation groups, fish were separated into control and hypothyroid treatment groups. Within the cold-acclimated hypothyroid group, fish were further divided into three treatment groups: (1) fish supplemented daily with 3,5,3′-triiodothyronine (T₃; Sigma-Aldrich, Sydney, Australia), (2) fish supplemented daily with 3,5-diiodothyronine (T₂; Sigma-Aldrich) and (3) fish given the ethanol vehicle. There were five replicate tanks per treatment with 12–15 fish per tank. We induced hypothyroidism by maintaining tank water at 0.3 mmol l⁻¹ propylthiouracil (Sigma-Aldrich), which inhibits TH production, and 5 μmol l⁻¹ iopanoic acid (Thermo Fisher Scientific, Sydney, Australia; administered daily) to inhibit deiodinase activity (Little et al., 2013). All experiments were carried out with the approval of the University of Sydney Animal Experimentation Ethics Committee (approval number L04/4-2012/1/5733).

Critical sustained swimming speed (U_crit) (Brett, 1965) was measured according to published protocols (Little et al., 2013). We used a variable DC power source (MP3090, Powertech, Osborne Park, WA, Australia) to adjust the flow speed through a Perspex swimming flume. Flow speed was measured using a flow meter (FP101, Global Water, Gold River, CA, USA) and the voltage output of the power source was calibrated for flow speed. U_crit was determined according to published protocols (Seebacher and Walter, 2012), with a time interval between speed increments of 600 s, a speed increment of 0.06 m s⁻¹ and an initial flow rate of 0.2 m s⁻¹. Animals were swum until fatigued, which was defined as the time when fish could no longer hold their position in the water column (Brett, 1965). Each fish was swum at 18 and 28°C in random order, with at least 24 h between swim trials. U_crit is reported as body lengths (BL) per second.

Burst velocity was measured according to published protocols (Seebacher et al., 2012). Fish were introduced into a tank (405×600 mm) filled with water to a depth of 25 mm and filmed from above using a Casio Exilim EX F1 camera recording at 30 frames s⁻¹. A 10 cm ruler placed on the bottom of the tank was used as a scale. When at rest, fish were startled by lightly tapping their tail with a stick. The ensuing escape response was filmed, and the fastest speed recorded in five escape responses was used as the maximum burst velocity. Videos were analyzed in Tracker Video Analysis and Modeling Tool software (Open Source Physics, www.opensourcephysics.org).

Tail beat amplitude and frequency were recorded using a Casio Exilim Ex F1 camera filming at 120 frames s⁻¹ while fish were swimming steadily in the swimming flume. Fish were swum steadily at two speeds, 13.2 and 26.4 cm s⁻¹, which correlated to 30 and 60%, respectively, of the average U_crit velocities of all groups. We measured amplitude as the maximal displacement of the tail at the peduncle during one tail beat cycle. Similarly, we determined period as the time taken for the tail to complete one cycle, and calculated frequency as 1/period. These swimming kinematics measures were averaged from three independent tail beat events at 28°C.

For enzyme assays, fish were killed by anaesthesia in buffered MS 222 (0.4gl⁻¹; Sigma-Aldrich) followed by decapitation. Mixed tail muscle was collected by transecting the fish from the anterior border of the anal fin to the posterior border of the dorsal fin. The fins were removed and tissue was transferred into liquid nitrogen and stored at −80°C for later analysis. We measured the activities of SERCA and myofibrillar ATPase at 18 and 28°C. SERCA activity was determined according to James et al. (James et al., 2011). For the myofibrillar assay, we isolated myofibrils according to Ball and Johnston (Ball and Johnston, 1996) followed by an ATPase assay using ammonium molybdate as a colour indicator as described in Walter and Seebacher (Walter and Seebacher, 2009).

Muscle was dissected and stored in RNAlater (Ambion, Austin, TX, USA) at −20°C. RNA was extracted from samples using TRIreagent (Molecular Research Centre, Cincinnati, OH, USA), following the manufacturer's instructions. RNA concentrations and quality were verified using a NanoDrop (NanoDrop Technologies, Wilmington, DE, USA) and a Bioanalyser 2100 (Agilent Technologies, Palo Alto, CA, USA). A 2 μg aliquot of total RNA from each sample was treated with DNase I (Sigma-Aldrich) and reverse-transcribed using RNase HMMLV reverse transcriptase (Bioscript, Bioline, London, UK) and random hexamer primers (Bioline).

Quantitative RT-PCR was performed on an Applied Biosystems 7500 qRT-PCR machine (Applied Biosystems, Sydney, Australia) according to published protocols (Seebacher and Walter, 2012). In short, primers for phospholamban, different ryanodine receptor isoforms, KLF-15 and SERCA isoforms were adopted from a previous publication (Little et al., 2013), or designed according to their respective GenBank or Ensembl sequences (supplementary material Table S1). A consensus primer set, based on sequence data acquired from Ensembl (supplementary material Table S1), was created to exclusively amplify the FMHCs. Real-time PCR reactions contained 1× SensiMix SYBR (Bioline), 4.5 mmol l⁻¹ MgCl₂, 50–900 nmol l⁻¹ primer and ~100 ng cDNA. The cycle consisted of 95°C for 7 min, 40 cycles of 95°C for 20 s, 58°C for 1 min. Dissociation curve analysis was performed after the amplification step to verify the presence of only a single PCR product. Transcript expression levels of the 11 target genes in each treatment group were normalized to elongation factor 1-alpha as per recent recommendations for zebrafish housekeeping genes (McCurley and Callard, 2008), and expressed relative to the warm-acclimated control treatment for the warm/cold hypothyroid experiment, and to the cold-acclimated control treatment for the T₂ and T₃ supplementation experiment.

Data are presented as means ± s.e.m. Data sets were analyzed by permutation ANOVA (PERMANOVA) in Primer 6 (PRIMER-E, Plymouth, UK). In the case of interactions involving acclimation temperature and hypothyroidism, a priori planned Monte Carlo contrasts were conducted in the PERMANOVA software package in Primer 6 to determine the main effects of acclimation temperature in the control treatments, and of thyroid status in the cold and the warm acclimation treatments. For measures that were not sensitive to hypothyroidism, we did not test the effect of T₂ or T₃ supplementation. We analysed data for the supplementation treatments that were carried out at 18°C (or 26.4 m s⁻¹ in the case of tail beat frequency) only, because there were no significant interactions between hypothyroidism and test temperature or test speed. The effects of supplementation with T₂ or T₃ were compared with control and hypothyroid treatments by one-way ANOVA.

The effects of acclimation treatment are presented as absolute values for U_crit, burst speed, tail beat amplitude, tail beat frequency, maximal enzyme activities and relative values for mRNA transcript levels. For hypothyroid treatments, the data are presented as percent change from the controls, i.e. [(control value – hypothyroid treatment value)/control value] × 100. For T₂ and T₃ supplementation treatments, the data are presented as percent recovery from hypothyroid treatment, i.e. [(supplemented treatment value – hypothyroid treatment value)/(control value – hypothyroid treatment value)] × 100.

Cold acclimation significantly increased U_crit at both 18 and 28°C acute test temperatures (Fig. 1A, Table 1). Cold-acclimated fish compensated for the limiting effect of low temperature, and swam as well at the 18°C acute test temperature as warm-acclimated fish did at a test temperature of 28°C.

There was no significant effect of acclimation treatment on burst speed or tail beat amplitude (Fig. 1B,C, Table 1). However, cold acclimation significantly increased tail beat frequency at both 13.2 and 26.4 cm s⁻¹ swimming speeds (Fig. 1D, Table 1). There was no significant effect of acclimation temperature on maximal SERCA activity (Fig. 1E, Table 1), but maximal myofibrillar ATPase activity significantly decreased with cold acclimation (Fig. 1F, Table 1).

There was an interaction between acclimation treatment and hypothyroidism in determining mRNA levels of SERCA 1 (Fig. 1G, Table 2), which significantly increased in response to cold acclimation (post hoc tests). mRNA transcript levels for the FMHCs were significantly lower in response to cold acclimation (Fig. 1G, Table 2). KLF-15 mRNA levels did not change with temperature acclimation.

Hypothyroid treatment significantly decreased U_crit in the cold-acclimated fish but had no effect on U_crit in the warm-acclimated fish (Fig. 2A; significant interaction followed by post hoc comparisons, Table 1). Hypothyroidism significantly lowered the burst speeds of both warm- and cold-acclimated fish (Fig. 2B, Table 1). Hypothyroidism did not affect tail beat amplitude in either acclimation treatment (Fig. 2C), but it significantly reduced tail beat frequency in the cold-acclimated fish at both speeds (Fig. 2D, Table 1).

In parallel with changes in whole-animal performance, hypothyroidism significantly reduced maximal SERCA activity in the cold-acclimated fish, but had no significant effect on the warm-acclimation treatment (Fig. 2E, Table 1). Interestingly, hypothyroidism significantly increased maximal myofibrillar ATPase activity in the cold-acclimated fish, but had no significant effect in the warm-acclimated fish (Fig. 2F, Table 1). Hypothyroidism significantly decreased mRNA transcript levels of the FMHCs and SERCA 3, and significantly reduced mRNA levels of SERCA 1 in cold-acclimated fish but increased SERCA 1 mRNA concentrations in response to warm acclimation (Fig. 2G; significant interaction followed by post hoc tests, Table 2). KLF-15 mRNA levels did not change with hypothyroidism.

Supplementation of hypothyroid fish with either T₂ or T₃ resulted in a significant recovery of U_crit, burst speeds, tail beat frequencies and SERCA activities to those characteristic of cold-acclimated control fish (Fig. 3A–D, Table 3)

There was no significant effect of T₂ or T₃ supplementation on maximal myofibrillar ATPase activity (Fig. 3E, Table 3). Both T₂ and T₃ supplementation promoted the recovery of mRNA levels for SERCA 3, SERCA 1 and the FMHCs (Fig. 3F, Table 3).

Very little is known about the pathways that coordinate local responses with changes in temperature during thermal acclimation, and whether these pathways are universal among tissues and conserved across taxonomic groups. Here we show that TH is a regulator that links changes in acclimation conditions to muscle function and swimming performance.

Locomotor performance is important because it underlies many behaviours that are central to animal fitness (Clobert et al., 2000; Husak et al., 2006; Husak et al., 2008; Eliason et al., 2011). While thermal acclimation of locomotor performance is often attributed to changes in metabolic capacity (Guderley, 2004a), many vertebrates also regulate muscle contractile function and biomechanics in response to temperature (Johnston and Temple, 2002; Seebacher and James, 2008). Previously we found that TH regulates metabolic capacity and locomotor performance in zebrafish, but only during cold exposure (Little et al., 2013). Here we show that the effect of TH can explain changes in locomotor performance through its influence on proteins that function in contraction and calcium handling of the muscle cell. In zebrafish, TH increased levels of SERCA 1, which may lead to a faster muscle phenotype, and which resembles its role in mammals (Simonides et al., 2001). Increased SERCA activity in cold-acclimated fish is paralleled by higher tail beat frequencies, which is likely to be the result of enhanced calcium reuptake into the SR. Frequency is a function of the rates of muscle contraction and relaxation, which are determined by Ca²⁺ release and resequestration into the SR. Hence, the similar responses in SERCA activity and tail beat frequency in response to hypothyroidism are likely to reflect a causal relationship. The faster replenishment of SR Ca²⁺ stores promotes quicker muscle relaxation rates, and restores SR [Ca²⁺] for subsequent contractions. Depletion of SR Ca²⁺ stores is a major cause of muscle fatigue, so that the influence of TH on SERCA is likely to increase fatigue resistance (Schiaffino and Reggiani, 2011; Seebacher et al., 2012) and improve sustained locomotion. The TH-mediated decrease in myofibrillar ATPase activity levels during cold acclimation is more difficult to reconcile with our whole-animal performance measures. Typically, myofibrillar ATPase activity has been shown to increase or remain unchanged in response to cold acclimation (Johnston et al., 1975; Hwang et al., 1990). Myosin ATPase capacity and crossbridge formation are regulated at many cellular levels, including isoform expression profiles, conformational states and ion concentrations (Cooke, 2011; Schiaffino and Reggiani, 2011). TH transcriptionally regulates all myosin heavy chain isoforms in mammals (Izumo et al., 1986), but whether this role is conserved in fish is not known. Nor is it known how the numerous fish myosin heavy chain isoforms functionally compare with their mammalian orthologues. KLF-15 is essential in regulating muscle performance in mice (Haldar et al., 2012). However, its mRNA concentrations in zebrafish did not respond to either thermal acclimation or thyroid hormone, so that at least at the level of mRNA it does not regulate changes in locomotor function in response to these variables.

TH enhances energy metabolism (Little et al., 2013) and swimming performance during cold exposure, both of which depend on a reliable supply of oxygen. Interestingly, components of the cardiovascular system in mammals also have been shown to respond to TH (Izumo et al., 1986; Carr and Kranias, 2002; Wang et al., 2003; Kahaly and Dilmann, 2005; Iordanidou et al., 2010). Enzymes that regulate calcium handling and contraction are the same in skeletal and cardiac muscle, but their isoform compositions can differ greatly between these tissues. It would be of interest to determine whether TH also modulates cardiac performance during cold exposure to determine whether the TH-mediated increase in locomotor performance also relies on cardiac thermosensitivity to TH.

In addition to influencing pathways that are homologous to those that control thermogenesis in mammals (Little et al., 2013), TH regulates SERCA1 during cold exposure in fish. Futile Ca²⁺ cycling by SERCA 1 has become an important model to explain thermogenesis in skeletal muscle (or skeletal muscle derivatives such as brown adipose tissue) of mammals (Clausen et al., 1991; Bal et al., 2012), birds (Bicudo et al., 2001) and some fish lineages (Block and Finnerty, 1994; Tullis and Block, 1996; Morrissette et al., 2003). Our findings that TH upregulates both SERCA activity and SERCA 1 mRNA expression levels in cold-acclimated zebrafish suggest that the role of TH in the thermal responses of early vertebrates may have important implications for the evolution of endothermy. The TH-mediated increase in SERCA1 expression and activity is a mechanism that improves locomotor performance in the cold, but would simultaneously promote heat production (Arruda et al., 2007). Endothermic billfishes and their ectothermic ancestors would provide an ideal system to assess the role of TH in the evolution of thermogenesis. Whether TH promotes thermogenesis in billfishes, and muscle function in response to cold temperatures in close relatives such as carangids (Little et al., 2010), would not only help provide an evolutionary trajectory for thermogenesis in fish, but also may implicate TH as a precursor to the evolution of thermogenesis across vertebrates.

 
• FMHC

  fast myosin heavy chain

 
• KLF-15

  Kruppel-like factor 15

 
• SERCA

  sarco/endoplasmic reticulum Ca²⁺-ATPase

 
• SR

  sarcoplasmic reticulum

 
• TH

  thyroid hormone

 
• U_crit

  critical sustained swimming speed