ABSTRACT
Understanding animal thermal tolerance is crucial to predict how animals will respond to increasingly warmer temperatures, and to mitigate the impact of the climate change on species survival. Yet, the physiological mechanisms underlying animal thermal tolerance are largely unknown. In this study, we developed a method for measuring upper thermal limit (CTmax) in larval zebrafish (Danio rerio) and found that it occurs at similar temperatures as in adult zebrafish. We discovered that CTmax precedes a transient, heat-induced brain-wide depolarization during heat ramping. By monitoring heart rate, we established that cardiac function is sub-optimal during the period where CTmax and brain depolarization occur. In addition, we found that oxygen availability affects both locomotor neural activity and CTmax during a heat stress. The findings of this study suggest that neural impairment due to limited oxygen availability at high temperatures can cause CTmax in zebrafish.
Highlights
Larval zebrafish reach their critical thermal limit (CTmax) at similar temperature as adult zebrafish
Acute heat stress causes a brain-wide spreading depolarization near the upper thermal limit
CTmax precedes brain-wide depolarization
Heart rate declines at high temperatures but is maintained during CTmax and brain depolarization
Neural activity is impaired prior to CTmax and brain-wide depolarization
Oxygen availability in the water affects both CTmax and neural activity
INTRODUCTION
Periods of abnormally high temperatures are becoming more numerous and severe as climate change progresses (Seneviratne et al., 2014; Stillman, 2019). As biological rates are highly affected by temperature, animals that hold similar body temperatures as the surrounding environment (i.e. ectotherms) might be particularly vulnerable to extreme heat (Deutsch et al., 2008; Schulte, 2015; Pinsky et al., 2019). Under acute heating, ectotherms reach a critical thermal limit (CTmax), a temperature where their equilibrium is lost and their movement becomes disorganized (Friedlander et al., 1976; Morgan et al., 2018; Jørgensen et al., 2020). The thermal tolerance of ectotherms correlates with their geographical distributions (Perry et al., 2005; Sunday et al., 2012), suggesting that limits in thermal tolerance partly determines distribution ranges. Moreover, rapid onset of heat waves can cause mass mortality in fishes, which suggests that thermal tolerance can be an increasingly important trait in the future (Genin et al., 2020).
Knowledge about the mechanisms controlling thermal tolerance is therefore important to understand how ectotherms may respond to increasingly warmer temperatures and predict ecosystem impacts. Although the physiological mechanisms of thermal tolerance are largely unknown, two major bottlenecks have been proposed; oxygen limitation and neural dysfunction (Friedlander et al., 1976; Pörtner et al., 2017; Jutfelt et al., 2019).
The oxygen limitation view (i.e. the oxygen- and capacity-limited thermal tolerance (OCLTT) hypothesis) suggests that tissue oxygen shortage limits thermal tolerance under acute heat stress (Pörtner & Knust, 2007; Ern, 2019). According to this hypothesis, the cardiorespiratory system in ectotherms is unable to match the increasing oxygen requirement with temperature, leading to tissue hypoxia and thus setting the limits for thermal tolerance. Insufficient cardiorespiratory function at high temperature, or near upper thermal limit, has been reported in insects and fish (Pörtner, 2002; Farrell, 2007, 2009; Ern et al., 2014), but it is maintained in a number of other ectotherms (Ern et al., 2015). In addition, experimental manipulation of oxygen availability during heat stress does not always affect thermal tolerance (McArley et al., 2020). These contrasting results suggest that oxygen limitation can only partly explain thermal tolerance, which is likely limited by different mechanisms in various species and contexts (Ern et al., 2016; Lefevre, 2016; Verberk et al., 2016; Jutfelt et al., 2018, 2019). Furthermore, the OCLTT hypothesis does not specify which physiological mechanisms fail first due to tissue hypoxia (Jutfelt et al., 2018).
Efficient locomotion (which is impaired at CTmax) requires normal neural function that is disrupted at high temperature. Therefore, neural control of locomotion is a heat-sensitive physiological mechanism that could underlie upper thermal tolerance, either via direct thermal effects on neurons (Friedlander et al., 1976; Jutfelt et al., 2019) or via indirect thermal effects from oxygen limitations (Schulte, 2015). Examples of severe heat-induced neural dysfunctions include spreading depolarizations in fruit fly (Jørgensen et al., 2020), loss of rhythmic neural activity in the digestive system of the crab (Marder et al., 2015), and thermogenic seizures in vertebrates (Dube et al., 2000; Shinnar & Glauser, 2002; Hunt et al., 2012). Yet, very few studies have directly investigated how neural dysfunction relates to measures of thermal tolerance. In the fruit fly, spreading depolarizations measured in head-restrained flies occur at similar temperatures as heat-induced coma in freely moving conspecifics (Jørgensen et al., 2020). Similarly, goldfish lose equilibrium at temperatures that induce hyperexcitability in cerebellar neurons of anesthetized conspecifics (Friedlander et al., 1976), and a study on Atlantic cod (Gadus morhua) found that cooling the brain marginally increased CTmax, suggesting a causal link between brain function and thermal tolerance (Jutfelt et al., 2019).
The relative importance of cardiorespiratory and oxygen limitations, versus direct thermal impacts on neuronal function, as the mechanisms limiting thermal tolerance during warming remains unclear (Clark et al., 2013a, 2013b; Jutfelt et al., 2014, 2018, 2019). This is partly due to the challenge of recording brain activity and CTmax in the same freely moving animal. In this study, we solved that challenge by simultaneously measuring brain activity and locomotion during heating using non-invasive fluorescence imaging in transparent larval zebrafish (Danio rerio) expressing a calcium indicator in the brain. We also recorded heart rate simultaneously with neural activity during heating in a separate group of larval zebrafish. Finally, we measured CTmax and neural activity in hypo- and hyperoxic conditions. This allowed us to test two predictions. First, if dysfunction of the central nervous system limits thermal tolerance, neural dysfunction should briefly precede or coincide with CTmax. Second, if neural dysfunction is caused by a cardiorespiratory limitation, we predict that failure of cardiac function precedes neural malfunction, and that water oxygen manipulation affects neural function and CTmax.
RESULTS
CTmax is similar in larval and adult zebrafish
The activity and upper thermal limit (CTmax) were recorded in freely swimming five-day-old larval zebrafish in a custom-designed glass chamber (sFigure 1A, B), in which water temperature increased from 28°C at a rate of 0.3°C/min (heat ramp fish), or was maintained at 28°C (control fish). To measure the activity level of larval zebrafish in our set-up, we recorded swimming speed (Figure S1), episodes of disorganized swimming (spiral swimming, Movie S1) and loss of equilibrium (Movie S2) at selected temperature ranges during the heat ramp. The swimming speed of five-day-old zebrafish decreased on average from 2.9±0.6 mm/s to 1.7±0.3 mm/s during the assay, at a similar rate in the control and the heat ramp fish (linear mixed-effects model, time: β±S.E.=0.3±0.1, t(103)=-2.5, p=0.013; treatment: β±S.E.=0.8±0.6, t(24)=1.2, p=0.23, R2 0.1, Figure S1). Ten of the 12 heat ramp fish, but none of the 14 control fish, displayed episodes of disorganized circular swimming prior to CTmax (Chi-square: X2(df=1,N=26)=15.6, p<0.001; Figure 1C). Animals close to their upper thermal limit often lose equilibrium, a behavioural event that is commonly used to determine CTmax across species (Moyano et al., 2017; Morgan et al., 2018). Here, we found that the heat ramp fish were unable to maintain equilibrium 15.3% of the time, more frequently than control fish, which lost equilibrium 0.8% of the time (Wilcoxon rank sum test, heat ramp: median±IQR=15.3±13.7%, control: 0.8±6.8%, r=0.58, p=0.004, Figure 1D). Yet, half of the control fish lost equilibrium during the assay (Figure 1D), likely because five-day-old zebrafish do not completely master postural stability.
We therefore measured CTmax as the temperature at which zebrafish larvae repeatedly failed to escape touches to the trunk (loss of response, see Methods). Using this loss of response criterion, five-day-old heat ramp fish reached CTmax at 41.4±0.1°C (Figure 1E). Nine-day-old fish tested in the same setup reached CTmax at similar temperatures of 41.3±0.2°C, indicating that the loss of response criterium is stable during larval development (linear regression model, nine-day-old: β±S.E.=-0.1±0.2°C, t(36)=-0.5, p=0.7, Figure 1E). Furthermore, adults from the same line tested using the loss of equilibrium protocol (Morgan et al., 2018), reached CTmax at similar temperatures (41.0±0.1°C) as larvae (adult: β±S.E.=-0.4±0.2, t(36)=-1.9, p=0.07; F(2,36)=1.9, R2=0.1, Figure 1E).
Warming causes a transient brain-wide depolarization near the upper thermal limit
Neurons operate best within a certain thermal range (Robertson & Money, 2012; Tang et al., 2012; Marder et al., 2015). Therefore, we hypothesized that a neural malfunction would occur as the fish’s temperature approaches the upper thermal limit. To record neural activity in the entire zebrafish brain during a heat ramp, zebrafish larvae, expressing the calcium indicator GCaMP6s in all neurons, were mounted in agarose under an epifluorescence microscope (Figure 2A, B). Calcium events were detected in the brainstem of heat ramp and control fish (Figure 2C, D). The frequency of the brainstem calcium events remained stable over time in the control fish (Figure 2F). In heat ramp fish, the frequency of brainstem calcium events increased with temperature, before abruptly declining (Figure 2F) during the 15 minute-period preceding a transient brain-wide depolarization (Figure 2D). The telencephalon and the brainstem, which are normally weakly co-active, both displayed a sharp increase in activity during the depolarization (Figure 2D). The brain-wide spread of the depolarization is further illustrated in laterally mounted fish (Figure 2E, Movie S3): it spread relatively slowly, and reached the dorsal brain regions 10-12 seconds after the onset in the ventral diencephalon (Figure 2E). The heat-induced brain-wide depolarizations started on average at 40.5±0.4°C in agar-embedded fish (Figure 2G), 0.8°C below the average CTmax temperature recorded in freely swimming age-matched fish (41.4±0.1°C, Figure 1E). Our results show that a transient brain-wide depolarization occurs during the heat ramp at temperatures close to CTmax in zebrafish.
CTmax precedes the brain-wide depolarization
Since agar-embedded larvae experienced a brain-wide depolarization at temperatures close to the upper thermal limit of zebrafish, we asked whether the depolarization preceded, or co-occurred with, the loss of response observed during heat ramp in freely swimming fish. To do so, we simultaneously measured CTmax and neural activity in freely swimming zebrafish during a heat ramp, using epifluorescence imaging (Figure 3A, B, (Muto et al., 2017)). CTmax (40.9±0.2°C) preceded the brain-wide depolarization (41.4±0.2°C, Figure 3C) in all individuals by 0.5°C on average (linear mixed-effects model, β±S.E.=0.5±0.1, t(6)=4.5, p=0.004, R2=0.3, Figure 3D). These results from freely swimming fish show that brain-wide depolarizations were not an artefact due to agarose embedding in the previous experiment (Figure 2) and that the brain-wide depolarization does not precede CTmax.
Cardiac function declines before CTmax and the brain-wide depolarization
Hypoxic brain tissues due to a cardiorespiratory limitation, could result in neural dysfunction during heat ramping. We therefore simultaneously measured neural activity and heart-beat frequency in the same fish during a heat ramp (Figure 4A, B) to test if the heart rate was altered prior to the depolarization onset. When the temperature was increased from 28°C to 34°C, the heart rate increased with a mean Q10 value of 1.9, from 4.4±0.1Hz to 5.7±0.1Hz (linear mixed-effects model: β±S.E.=1.3±0.2Hz, t(36)=6.9, p<0.001, R2=0.8 Figure 4C). From 34 to 37°C, the heart rate decreased by 16% to 4.7±0.2Hz (β±S.E.=0.9±0.2, t(36)=-4.9, p<0.001) and then remained stable during the period preceding the brain-wide depolarization (β±S.E.=-0.1±0.2, t(36)=0.6, p=0.6, Figure 4C). We also examined whether heart rate was affected during the brain-wide depolarization but it was unchanged (Figure 4D, E). When the fish were returned to normal holding temperature after the brain-wide depolarization the heart rate decreased by 34% from 4.7±0.2Hz to 3.1±0.3 Hz (β±S.E.=1.6±0.2 Hz, t(36)=-8.5, p<0.001, Figure 4C). Altogether, our data shows a decline in cardiac activity during the heat ramp, but no abrupt change in cardiac function prior to the depolarization onset, during the period when CTmax occurs.
Oxygen availability modulates temperature for CTmax and brain-wide depolarization
Since the heart rate is only 4-5 Hz at the time of the brain-wide depolarization (Figure 4C and 4E), it is possible that the cardiorespiratory system cannot match the increasing oxygen demand of the neural tissue. If oxygen availability limits thermal tolerance, we expect that oxygen manipulations during heat ramping would alter CTmax and neural function.
To test the effect of oxygen availability on CTmax, we first measured the temperature at loss of response in freely swimming larvae zebrafish exposed to hypoxic and hyperoxic conditions during heat ramping, at 60 and 150% oxygen of air-saturated water respectively. Both treatments had on average a lower CTmax than in the previous experiment (Figure 1E) likely due to experimental differences. The average CTmax occurred 0.9°C lower in hypoxia (39.2±0.1°C) than in hyperoxia (40.1±0.3°C; linear regression model, β±S.E.=0.9±0.3°C, t(26)=3.2, p=0.004; F(1,26)=10.2, R2=0.3; Figure 5A), which indicates that oxygen availability affects behavioural thermal tolerance in larval zebrafish.
To test the effect of oxygen availability on neural activity, agar-embedded fish were placed under an epifluorescence microscope and subjected to three oxygen treatments during heat ramping: hypoxia, normoxia and hyperoxia at 60, 100 and 150% oxygen of air saturated water, respectively. The brain-wide depolarization occurred at 40.7±0.4°C in normoxia. It occurred 1.8°C lower in the hypoxia (38.8±0.2°C; linear regression model, β±S.E.=-1.8±0.4, t(20)=-4.8, p<0.001; F(2,20)=40.4, R2=0.8; Figure 5B). On the contrary, the hyperoxia treatment during heat ramping increased the depolarization temperature by 1.3°C compared to normoxia (42.0±0.2°C; β±S.E.=1.3±0.4, t(20)=3.3, p=0.003). These results indicate that oxygen availability strongly influences the onset of brain-wide depolarization in larval zebrafish.
To further determine the effect of oxygen availability on neural activity when zebrafish reach their upper thermal limits, we calculated the frequency of brainstem calcium events during the 15 minutes before the brain-wide depolarization. Neural activity in the locomotor brain center sharply decreased at high temperature in all treatments: almost no calcium events were detected during the minutes preceding the depolarization (Figure 5C). The hyperoxia group retained a higher event frequency than the hypoxia and control groups at the start of this period (linear mixed-effects model, time: β±S.E.=0.08±0.02, t(216)=3.9, p<0.001; hyperoxia*time: β±S.E.=0.09±0.03, t(216)=3.3, p=0.001; R2=0.3, Figure 5C). This indicates that increased oxygen availability partially rescued neural function near zebrafish upper thermal limit.
We also observed an effect of oxygen availability on the temporal dynamic of the brain depolarization. Oxygen availability did not change the amplitude of the brain-wide depolarization (normoxia: 88.2±17.6%ΔF/F0; linear regression model, hypoxia: β±S.E.=-3.1±17.1, t(20)=-0.2, p=0.86; hyperoxia: β±S.E.=-15.4±17.4, t(20)=-0.9, p=0.39; F(2,20)=0.5, R2=-0.05, data not shown). However, oxygen availability significantly improved the depolarization recovery time. Compared to the normoxia group, the hyperoxia group returned to baseline activity after the depolarization 1.3 minutes faster, and the hypoxia group 2 minutes slower (recovery time in normoxia: 3.0±0.4 min; linear regression model, hypoxia: β±S.E.=2.0±0.5, t(19)=4.1, p<0.001; hyperoxia: β±S.E.=-1.3±0.5, t(19)=-2.5, p=0.02; F(2,19)=0.5, R2=0.7, Figure 5D-E).
DISCUSSION
A massive heat-induced global depolarization arose in the larval brain, at similar temperatures as CTmax occurred in freely swimming fish. These brain depolarizations were 4-5 times the magnitude of brainstem locomotor calcium events, and lasted much longer (Figure 2). This abnormally high neural activity in response to heat stress is reminiscent of heat-induced seizures (Dube et al., 2000; Shinnar & Glauser, 2002; Hunt et al., 2012), and of heat-induced spreading depolarizations (Jørgensen et al., 2020), which are slow propagating waves of brain depolarization (Spong et al., 2016). The temporal dynamic of the events measured here is characteristic of the latter. First, the depolarizations spread slowly across the brain (Figure 2E), at a speed consistent with that reported for spreading depolarizations (2-9 mm/min, (Woitzik et al., 2013; Spong et al., 2016)), and slower than seizures (Wenzel et al., 2017; Liu & Baraban, 2019). Second, the long post-depolarization recovery time (Figure 5E) is consistent with that seen after spreading depolarizations (Sawant-Pokam et al., 2016; Spong et al., 2016), and it is much longer than that measured after brain seizures (a few seconds, (Diaz Verdugo et al., 2019; Liu & Baraban, 2019)). Overall, the temperatures at which these spreading depolarizations occurred, as well as their transient nature, made it a plausible candidate to explain the prolonged, yet reversible unresponsive state that we observed at CTmax.
If upper thermal tolerance in zebrafish larvae is caused by a spreading depolarization shutting down central nervous system function, the depolarization should precede or coincide with CTmax measured in the same animal. Surprisingly, larval zebrafish reached CTmax before they developed a brain-wide depolarization. This is in contrast with a series of studies in insects in which measures of upper thermal tolerance correlated with the onset of spreading depolarization in the central nervous system (Andersen et al., 2018; Jørgensen et al., 2020). In these studies, upper thermal limits and spreading depolarization onsets were measured in separate individuals, respectively freely moving flies and central nervous system preparations. Using a similar approach in the first part of this article, we also found overlapping temperature ranges for CTmax and spreading depolarization onset. Only when measuring both parameters within the same individual could we resolve the small but consistent difference between these consecutive events (Figure 3D), and determine that CTmax preceded spreading depolarization. We thus conclude that spreading depolarization is not the cause of CTmax in zebrafish.
The fact that CTmax precedes the spreading depolarization does not rule out neural impairment as a possible mechanism determining the upper thermal limit. Previous studies suggested that neural function is impaired during acute heat stress due to either a direct effect of temperature (Friedlander et al. 1975; Jutfelt et al. 2019), or due to heat-induced tissue hypoxia. Extracellular recordings of the pyloric neuron in the crab stomatogastric nervous system show that the characteristic rhythmic firing crashes at high temperatures (Marder et al., 2015). Similarly, extracellular recordings of cerebellar neurons in anesthetized goldfish indicate a sharp decrease in spontaneous and sensory-evoked firing rates at high temperatures (Friedlander et al., 1976). In addition, a strong reduction in spontaneous neural activity occurs in ischemic brain tissue in rodents (Buzsaki et al., 1989; Barth & Mody, 2011) and in oxygen-deprived cultures of human cortical neurons (le Feber et al., 2016). Such findings indicate that both direct effects of temperature and tissue hypoxia are plausible candidate mechanisms. In line with these previous studies, we found that neural activity was strongly reduced in locomotor brain regions during the minutes preceding the depolarization (Figure 2 and 5C), which we interpret as an early indication of neural malfunction that underlies the degradation of zebrafish locomotion observed during the corresponding period (Figure 1C-D). As this period of neural inactivity in the minutes leading up to CTmax could be due to both direct impacts of temperature and/or oxygen deficit on neural function, we manipulated oxygen water concentrations to test if oxygen availability alters CTmax and neural function.
We found that hyperoxia (150% air saturation) improved both behavioural and neural thermal resilience compared to hypoxia. The fact that neural activity in the locomotor brain centre was more resilient to high temperature in hyperoxia than in hypoxia and normoxia provides a likely mechanism for the improved behavioural resilience of zebrafish in hyperoxic conditions. Furthermore, similarly to earlier measurements in normoxia, neural activity was strongly silenced in the brainstem of fish during the very last minutes preceding the brain depolarization, when fish reach their upper thermal limit (Figure 3D). This shows that central nervous function impairment coincided with the temperatures where CTmax occurs, further suggesting that neural impairment due to a cumulative lack of oxygen and/or accumulation of anaerobic metabolites contributes to the unresponsive state observed at the upper thermal limit.
A few studies have found a positive effect of increased oxygen availability on thermal tolerance, while most fail to find an effect of hyperoxia (McArley et al., 2020). Hyperoxia improved the upper thermal limit by 1.1 °C (loss of equilibrium) in the European perch (Ekström et al., 2016) and in the Common triplefin (McArley et al., 2018). Increased aquatic oxygen availability also extended the survival of several ectotherm species (arthropods, chordates, echinoderms, molluscs, and fish) at extreme high temperatures (Verberk et al., 2018; Giomi et al., 2019). However, hyperoxia did not improve thermal tolerance for a large number of other fish species (McArley et al., 2020). Moreover, the upper thermal limit of the red drum and lumpfish was only reduced under severe hypoxic conditions (Ern et al., 2016). Additionally, the aerobic scope, the aerobic capacity available for non-maintenance activities, remains high in some ectotherms at thermal limits, indicating surplus oxygen transport capacity (Overgaard et al., 2012; Grans et al., 2014; Wang et al., 2014; Brijs et al., 2015; Ekström et al., 2016). Taken together these studies suggest that oxygen limitation is not a general mechanism limiting thermal tolerance across all species and contexts. Our results suggest that inter-species differences in neural tissue resilience to heat and oxygen deprivation might be a critical physiological factor explaining these different reports. It will also be interesting to investigate whether the sensitivity of CTmax to oxygen availability in zebrafish is limited to early life stages or persists throughout ontogeny.
To test whether heat ramping causes cardiac dysfunction we examined heart rate and brain activity simultaneously. The heart rate initially (from 28-34°C) increased as expected with temperature, similar to previous reports (Gollock et al., 2006; Farrell, 2007). This initial increase was followed by a decrease to to 270 beats per minute, a heart rate then sustained upon further heating and throughout the brain depolarization. Similarly, a decrease or plateau in heart rate with increasing temperatures has been reported for salmonids (Farrell, 2009), and European perch (Ekström et al., 2016). The lack of heart rate increase when approaching thermal limits might contribute to mismatch between oxygen delivery and metabolic rate. Such a mismatch may have caused developing tissue hypoxia and accumulation of metabolites, which could have contributed to causing the loss of responses at CTmax.
Animals are adapted to widely different climates, and because of this, the temperatures at which heat-induced behavioral alterations occur are strongly species-specific. Yet, the sequence and nature of heat-induced behavioral alterations seem relatively well conserved across species. For example, a similar sequence of events occur in fruit fly and goldfish subjected to rapidly increasing temperatures: hyperactivity, then loss of coordinated movement, followed by loss of equilibrium, before entering into a heat-coma indicated by an animal lying still and unresponsive to mechanical stimulations (Friedlander et al., 1976; Jørgensen et al., 2020).
To enable comparison with previous studies we quantified the behavioral changes that occurred during heat ramping. Although the swimming velocity of heat ramp fish tended to be higher than that of controls between 31-38 degrees, there was no clear period of heat-induced hyperactivity, possibly due to the high interindividual variability in locomotion in larval zebrafish. Loss of motor coordination (spiral swimming) and loss of equilibrium occurred in most heat ramp larval zebrafish, followed by a state of unresponsiveness that was similar to the heat-coma documented in goldfish and fruit flies. The loss of response endpoint used here was stable throughout larval development and yielded results within the range recorded using the loss of equilibrium criterion in adults of the same Tg(elavl3:GCaMP6s) line (Figure 1E), or in wild-type adult zebrafish (Morgan et al., 2018), thus supporting its use a criterion for CTmax in larvae. In addition, spiral swimming could be used as a novel, specific and early sign of heat-induced locomotor dysfunction before the upper thermal limit is reached.
In conclusion, we show that the acute thermal limits of larval zebrafish are not caused by global brain depolarization, but are instead related to a drop in neural acivity preceding CTmax. Furthermore, in concordance with OCLTT predictions, tissue oxygen availability appears to constrains both brain function as well as the whole animal thermal limits during thermal ramping.
METHODS
Animals and housing
Experiments were conducted on five- to ten-day-old larvae and seven- to ten-month-old zebrafish (Danio rerio). The transgenic line Tg(elavl3:GCaMP6s) (Vladimirov et al., 2014) in nacre/mitfa background (Lister et al., 1999) was used. Eggs were collected in the morning and kept at a density of one per mL in fish water (0.2g of marine salt and 0.04 L AquaSafe per litre of carbon-filtrated water, used for all experiments). After hatching at three days post-fertilization, the larvae were kept in small nursery tanks and fed twice a day with larval food (Tetramin, Tetra) after 5 days post-fertilization. Fish were maintained under standard laboratory conditions (26.8 ± 0.1°C; 12/12-hour light/dark cycle). On the day of the experiment, the fish were fed once after the experiment was finished. All experimental procedures performed on zebrafish were in accordance with the 2010/63/EU directive and approved by the Norwegian Animal Research Authority (Food and Safety Authority; Permit number: 8578).
The experiments consisted of five parts performed on different fish: 1) measuring CTmax in freely swimming zebrafish larvae, 2) recording brain activity during warming in agar-embedded zebrafish larvae, 3) simultaneous recording of CTmax and depolarization in freely swimming zebrafish larvae, 4) simultaneous recording of heart rate and brain activity during heat ramping and 5) recording of CTmax and neural function under oxygen manipulation during warming.
CTmax in freely swimming larvae
CTmax Setup
Larval zebrafish swam in the behavioural arena (central compartment) of a double-walled glass heating-mantle made in the NTNU glass workshop. The dimensions were as follows: outer diameter= 42 mm, inner diameter= 29 mm, outer height= 32 mm, inner height= 19 mm. The fish movement was recorded using a webcam (Logitech C270, 720p) positioned above the arena. The arena was placed above a background illumination light source to enhance contrast. Two outlets connected the glass heating mantle to a water bath. Adjustment of water temperature of the arena was achieved by pumping water from an external heating bath through the heating-mantle surrounding the arena. The heating rate in the arena was 0.3°C per minute as described in Morgan et al., (2018). The water temperature inside the central chamber of the arena was recorded at 1 Hz using two thermocouples (type K, Pico Technology) connected to a data logger (TC-08, Pico Technology).
CTmax assay
The arena was filled with three mL of 28°C water supplied with air bubbling through a modified hypodermic needle fixed to the side wall of the compartment. At the beginning of a CTmax recording, a single larva was carefully transferred to the arena. All individuals were given 15 minutes to habituate to the arena before the recording started. The CTmax assay lasted up to 50 minutes, during which the temperature within the arena increased from 28°C by 0.3°C/min until the larvae reached CTmax. The temperature was kept constant at 28°C for recordings of control fish. The fish behaviour was recorded at 4-7 Hz during four intervals: 28-29°C, 31-32°C, 34-35°C and 37°C-CTmax. For control larvae, recordings of matching durations were taken at corresponding time points. Methods commonly used to determine CTmax in fish were unsuitable for 5 dpf larvae zebrafish: loss of equilibrium (Morgan et al., 2018) also occurred in control fish without heat ramp; and muscular spasms (Lutterschmidt & Hutchison, 2011) could not be reliably quantified due to the larvae’s small size. Thus, CTmax was determined as the temperature at which the animal became unresponsive (Friedlander et al., 1976; Sherman & Levitis, 2003; Jørgensen et al., 2020), which was defined as the first of three consecutive tactile stimulations that did not elicit an escape response (loss of response). The stimulations were applied to the larva’s trunk using the tip of a capillary micro-ladder (VWR), with a minimum of three seconds between consecutive stimulations. Upon CTmax, the fish was transferred to 28°C water and visually monitored. Fish that did not recover normal locomotor activity were rapidly euthanized and were not included in the analysis (this happened in only four out of 44 fish, similar to previous reports (Morgan et al., 2018; Åsheim et al., 2020)).
CTmax assay in adult zebrafish
To compare CTmax throughout development, 16 adult Tg(elavl3:GCaMP6s) zebrafish, aged seven- to ten-month-old, were tested in the CTmax setup described by Morgan et al., (2018). The fish were tested in groups of eight fish in a rectangular acrylic tank (25 cm long, 20 cm wide, 18 cm deep) filled with nine litres of water. Individual CTmax was measured at the temperature of loss of equilibrium determined as two seconds of inability to maintain postural stability (Morgan et al., 2018). The fish were removed immediately after the criterion was reached. All animals survived the test.
CTmax assay data analysis
The fish position was manually detected in MATLAB (MathWorks 2018) at 28-29°C, 31-32°C, 34-35°C, 37-38°C and during the 1°C elevation preceding CTmax. The average speed was calculated during these intervals. Loss of equilibrium and spiral swimming events were manually labelled by an experimenter during the 12 min preceding CTmax in heat ramp fish, and during the corresponding period in control fish. Loss of equilibrium was recorded when the fish tilted to the side for at least one second (Movie S2). Spiral swimming event was recorded when the fish rapidly performed a minimum of two consecutive revolutions with diameter below 5 mm (Movie S1).
Brain activity during warming in agar-embedded larvae
The calcium imaging to record brain activity during warming was executed using the same double-walled glass heating mantle as described for the CTmax assay above.
Epifluorescence calcium imaging in agar-embedded zebrafish larvae
For the imaging of neural activity in embedded fish (Figure 2A, B), a five-day-old larva was embedded at the bottom of the central glass compartment in 2% low-gelling temperature agarose (Merck). The agarose was carefully removed around the eyes and mouth upon hardening and the arena was filled with 3 mL water. Calcium fluorescence was recorded with a custom-made epifluorescence microscope equipped with a 10x water-immersion objective (UMPLFLN, Olympus), a set of GFP emission-excitation filters (FGL400, MD498, MF525-9, Thorlabs) and a mounted blue LED controlled by a driver (MWWHL4, LEDD1B, Thorlabs). Images were collected at 5 Hz via a custom-written Python script using the Pymba wrapper for interfacing with the camera (Mako G319B, Allied vision).
Simultaneous recording of CTmax and depolarization in freely swimming larvae
Calcium imaging in freely swimming fish was used to record CTmax and brain-wide depolarization in the same individuals during one trial. For the imaging of freely swimming fish, the double-walled glass heating-mantle was placed under an epifluorescence microscope (Axiolmager, ZEISS) equipped with a megapixel camera (AxioCam 506, ZEISS). The larva was swimming at the centre of the arena, within a 13 mm diameter region delimited with a nylon mesh, to match the limited field of view of the microscope (14 x 14 mm). Time-lapse of fluorescent images were recorded at 5 Hz using the Zen software (ZEN Lite Blue, ZEISS). Since the recordings were made in the dark, the experimenter watched the live recording display of the recording on the computer screen to visually guide the pokes to the fish tail. The nylon mesh created a small thermal gradient in the arena from the central part to the outer area outside the mesh. During the experimental trials, the temperature was recorded outside the mesh to ensure a full view of the fish. The temperature gradient was thoroughly quantified in an additional experiment over six heat ramps, by recording the temperature in the arena centre and outside the mesh near the outer wall of the arena with two thermocouples at each location. The gradient was accounted for to calculate the water temperature the fish was exposed to.
Calcium imaging data analyses
Calcium imaging recordings on embedded fish were corrected for slow drift due to agarose expansion using the Fiji’s (Schindelin et al., 2012) Template matching plugin (Align slices in stack). The regions of interest corresponding to the whole brain, the telencephalon and the brainstem were manually segmented. The raw calcium fluorescence signal was calculated by averaging all pixels within a region. To reproducibly detect the brain-wide depolarization onset time across fish, the whole brain raw fluorescence signal was first processed to filter out the calcium events using a 2nd order Butterworth filter (low pass, 0.01 Hz cut-off frequency). The average value and standard deviation of the filtered trace’s time derivative were calculated during the 17 min preceding the approximate onset of the depolarization. The depolarization onset was set when the time derivative exceeded the average baseline value by more than five standard deviation. For calcium event detection during the period preceding the brain-wide depolarization, the fluorescence change was calculated for each brain region using a sliding window of the previous two minutes. The calcium events were detected automatically using the MATLAB function findpeak.
Measuring brain-wide fluorescence during the freely swimming fluorescence assay was cumbersome due to the fish moving and tilting. Thus, the depolarization onset temperature was determined by two investigators, who independently selected the time at which the fluorescence increased abruptly within the brainstem/diencephalon. The final depolarization onset value was obtained by averaging both estimates (agreement between experimenters: 19 ±4s out of 55 minutes).
Cardiac function and brain activity during heat ramping in agar-embedded larvae
Heat rate and neural activity were simultaneously recorded at 20 Hz in laterally mounted larvae, under an epifluorescence microscope (AxioImager, ZEISS) equipped with a megapixel camera (AxioCam 506, ZEISS). Recordings were aligned when needed using Fiji’s Template matching plugin. Whole brain and heart regions of interest were selected in Fiji. Fish whose heart was out of focus during more than two recording intervals (1 fish out of 10), and frames where movement artefacts could not be corrected were not included in the rest of the analysis. Heartbeat frequency during a recording interval was calculated using Matlab’s continuous wavelet transform function (cwt).
Oxygen effects on CTmax and neural function during warming
CTmax assay with oxygen level manipulation
The CTmax setup described earlier was used to record the effect of oxygen manipulation on CTmax. The arena was intermittently bubbled with pure oxygen or pure nitrogen to increase or decrease oxygen levels, respectively. The bubbling flow rate was manually adjusted using tubing clamps. A fibre optic oxygen probe (OXROB10, PyroScience) and a temperature probe (TSUB21, PyroScience) connected to an oxygen and temperature meter (FireSting O2, PyroScience) were placed in the chamber to monitor and record the oxygen levels during the assay. The oxygen saturation level was displayed in real time using an oxygen logger software (pyro oxygen logger, PyroScience) and kept at 150 %, or 60%, of air saturation during the whole CTmax assay by manually adjusting the bubbling intensity.
Recording neural activity with oxygen level manipulation
The epifluorescence calcium imaging setup for embedded fish was used to measure neural activity during heat ramping with oxygen manipulation. Fish were visually stimulated using pulses of red light (2 second long, every 30 second) using a red LED positioned in front of the fish. For control treatment, the water was bubbled with air through a modified hypodermic needle and maintained at 100% oxygen of air saturation during the trials. To increase the oxygen level during oxygen manipulation trials, oxygen was gently blown on the surface. To decrease oxygen saturation, the water was bubbled with nitrogen. A bare fibre microsensor (OXB430, PyroScience) and a temperature probe (TDIP15, PyroScience) were placed in the chamber to monitor and record oxygen levels. Oxygen levels were adjusted to either 150%, or 60%, of air saturated water during the high and low oxygen treatments, respectively.
Calcium imaging data analyses
Depolarization onset (Figure 5B) and brainstem calcium events (Figure 5C) were calculated as described above. For calcium events, fish in which the brainstem signal was weak and noisy due to uneven mounting were not included in the analysis. One fish in the hyperoxic group could not be included in the calculation of the recovery time (Figure 5E) because the recording ended before the fluorescence reached baseline after the depolarization.
Statistical analyses
Data are reported in the text as mean and standard error (S.E.), unless stated otherwise. All statistical analyses were conducted in R version 4.0.2 (R Core Team, 2020). We used linear regression models on independent measurements and the assumptions of normality for homoscedasticity of residual variance were assessed visually using residual plots. Alternative models were considered if assumptions were violated. For responses recorded as repeated measures on individual fish linear mixed-effects models were created using the lmer function from the lme4 package v.1.1-23 to account for fish identity as random effect. Akaike information criterion was used to consider interactions in models with two predictor variables and model assumptions were visually assessed using residual plots. The effects of predictor variables are presented with effect size (β), their S.E. estimated by the respective models and the R2 for the fixed effects is presented. The significance of effects was considered with a p-value criterion of p<0.05.
CTmax and behavioural response to heat ramping
To analyse the swimming velocity during heat ramping and in the control group a linear mixed-effects model was performed with treatment and time step as predictor variables and the fish identity as random factor. No interaction was included (ΔAIC=2). Table S1 is a summary of the model results and Figure S1 displays the datapoints. The proportion of spiral swimming events was analysed using a Chi-square test to handle the lack of variance in the control group. A Wilcoxon rank sum test was used to test the treatment effect on loss of equilibrium response. To test the difference in CTmax throughout development, a linear regression model was used with CTmax as a function of the age. The CTmax value of one adult fish was an extreme outlier and was not included in the analysis. Results from models with and without this fish are included in the supplements (Table S2).
Temperature of CTmax and depolarization
The relationship between depolarization and CTmax response in freely swimming fish was analysed with a linear mixed-effects model with temperature as response variable, a categorical fixed effect of response type (CTmax or depolarization) and fish identity as a random factor (Table S3).
Change in cardiac function during heat ramping
Effect of temperature on heart rate was tested with a linear mixed-effects model where fish identity was defined as a random factor. The model was releveled to check for differences between the different temperature steps and the p-value significance criterion was accordingly reduced to p<0.001 for this analysis. Model results are presented in Table S4.
Effect of oxygen manipulation during on CTmax and depolarization
The effect of oxygen level on CTmax was analysed with a linear regression with CTmax as a function of the hypoxia and hyperoxia treatments. Two outliers and two fish who were euthanized after the assay were removed from the hyperoxic group. Table S5 displays model results including these outliers. The effect of oxygen on neural activity was assessed using linear regression analyses. Depolarization temperature, amplitude of depolarization and recovery after depolarization were tested as functions of the treatment groups hypoxia, normoxia and hyperoxia. Frequency of calcium events for the 15 minutes preceding the brain-wide depolarization was analysed with an interaction between time and treatment. One extreme outlier in the hyperoxic group was excluded from the analyses. Results from models made with and without this fish are presented in Table S6.
Authors contributions
AHA, PH, FJ & FK conceived the project. FJ & FK jointly supervised. AHA, PH & FK built the experimental setups. AHA, PH, PK collected the data. FK wrote the software for data acquisition and analyses. AHA, PH & FK analysed the data, AHA and FK prepared the figures and AHA performed the statistical analyses. AHA & FK wrote the manuscript, with contributions from all authors.
Funding
This work was supported by the Research Council of Norway (FK: FRIPRO grant 262698, FJ: 262942)
SUPPLEMENT
Supplementary Movie S1. Spiral swimming in a five-day-old zebrafish heat ramp larva. Speed 1X.
Supplementary Movie S2. Loss of equilibrium in the same heat ramp larva as in Supplementary Movie 1. Speed 1X.
Supplementary Movie S3. Representative illustration of the depolarization spreading in the brain of a five-day-old Tg(elavl3:GCaMP6s) larva exposed to a heat ramp (same fish as in Figure 2E). Speed 1X.
Acknowledgements
The authors wish to thank the staff working at the animal facility for fish care, the members of the Animal Physiology section at the Department of Biology for scientific discussions, and Marius Mæhlum for the heartbeat frequency analysis code.