Genetics of Physiological Variation Within and Between Larval Wild-Type AB and Backcrossed NHGRI-1 Zebrafish (Danio rerio)

Abstract

Changes in the environment promote variations in fish physiological responses. Genetic variation also plays a role in physiological variation. To explore the role of genetics in physiological variation, we assessed variation of cardiac function (heart rate, stroke volume, and cardiac output), oxygen consumption, yolk conversion efficiency, and cost of development in embryonic and larval AB wild-type and NHGRI-1 zebrafish (low heterozygosity line backcrossed from AB wild-type) exposed to different temperature and oxygen regimes. Fish were exposed from fertilization to 7 days post-fertilization (dpf) to control conditions (28 °C, 21% O₂) or to low temperature (23 °C, 21% O₂), high temperature (33 °C, 21% O₂), moderate hypoxia (28 °C, 13% O₂), or severe hypoxia (28 °C, 10% O₂). We hypothesized that (1) assessed physiological variables will respond similarly in both fish lines and (2) data variability in the low heterozygosity NHGRI-1 zebrafish will be lower than in AB zebrafish. Cardiac function decreased at lower temperature and in hypoxia in both AB and NHGRI-1 zebrafish. Oxygen consumption was increased by higher temperature and hypoxia in AB fish and by severe hypoxia in NHGRI-1 fish. Yolk conversion efficiency was decreased by lower temperature and hypoxia in AB fish and increased by higher temperature and decreased by hypoxia in NHGRI-1 fish. Cost of development was higher mainly in hypoxia-treated fish. Supporting our hypothesis that genetics contributes to physiological variation, NHGRI-1 zebrafish data showed significantly lower coefficients of variation in 84% of assessed endpoints. We conclude that (1) there is a strong genetic component to physiological variation in fishes and (2) low heterozygosity NHGRI-1 zebrafish are useful models for reducing the ‘noise’ from genetic backgrounds in physiological research in fish, which may aid interpretation of experimental results and facilitate reproducibility.

1. Introduction

Physiological variation—a difference in physiological traits or responses among individuals from a species or population—is an important biological phenomenon because it contributes to the raw material for natural selection and adaptation. Such variation helps to explain species distribution and interactions and influences how populations cope with changes in the environment [1,2]. Physiological variation will reflect both intrinsic and environmental variation, including developmental processes (e.g., variations in growth or development [2]), environmental factors (e.g., induced changes through mechanisms such as phenotypic plasticity [1]), or genetic differences among individuals (e.g., inherited traits may lead to physiological variation [2]). Genetic variability among individuals may produce different physiological responses in a context-dependent fashion, where responses may vary with different environments and even with time of day [3]. Physiological variation is reflected in differences among and within species in metabolic rate [4], temperature tolerance and acclimation ability, swimming performance, hypoxia tolerance, osmoregulatory capacity, and stress responsiveness to name but a few examples [2,3,4,5,6,7,8,9].

Our lack of understanding of genetic variability among individuals (or populations) may constrain interpretation of our results, as well as limit experimental reproducibility [10,11,12,13,14]. Increased physiological variation due to increased genetic variation may also lead to masked-treatment effects, where inter-individual variation may potentially ‘hide’ the effects of experimental treatments, especially if sample sizes are small [15]. Additionally, developmental and sex-specific effects promoted by diverse genetic backgrounds may lead to modified physiological responses [16]. Transgenic or inbred strains/lines of laboratory animals may exhibit distinct responses due to their genetic makeup compared to outbred stocks, where variability is greater [16]. Given these obfuscating issues, the use of inbred strains/lines is a useful experimental strategy for coping with high physiological variation across taxa. In mice and rats, inbred strains/lines show decreased variability in stress responses and metabolic and cardiovascular traits [17,18]. In fishes, studies on inter- and intra-specific variation show differences in brain size and longevity [19], energy requirements [20], basal plasma cortisol levels [21], body morphology [22], and seasonal utilization of food resources influencing coexisting strategies [23] to name a few such effects.

One of the most popular piscine models in research is the zebrafish (Danio rerio)—a native tropical fish from South Asia [24]. Approximately 84% of disease-related human genes have a homologous counterpart in the zebrafish genome [25], and this emulation of human pathologies by the zebrafish makes them ideal animal models [26]. Different responses at different organismal levels occur among and within zebrafish strains/lines and even between the same strains/lines from different laboratories [12]. Different zebrafish strains show different responses in swimming performance and its associated morphological and physiological traits [27], and their genetic variation may impact skeletal development [28]. Our previous study comparing AB and NHGRI-1 zebrafish (a backcrossed line with ~15% the heterozygosity of AB wild-type zebrafish [29]) revealed how variability of morphological responses to changes in temperature and oxygen availability had a strong genetic component, varying between fish lines [30,31].

Here, we have exposed AB and NHGRI-1 zebrafish to control and several experimental conditions—low or high temperature and moderate or severe hypoxia—to assess differences in physiological performance and especially in intrinsic variability. We assessed cardiac function (heart rate, stroke volume, and cardiac output), oxygen consumption, and yolk conversion efficiency from fertilization to 7 days post-fertilization (dpf). We hypothesized that (1) both fish lines will show similar physiological responses to low/high temperature and hypoxia exposure but (2) the decreased heterozygosity in NHGRI-1 fish will have decreased coefficients of variation due to their reduced heterozygosity.

2. Materials and Methods

2.1. Fish Husbandry

Adult AB and NHGRI-1 zebrafish held in the aquatics facility in the Department of Biological Sciences of the University of North Texas were bred to obtain egg clutches from both fish lines. AB and NHGRI-1 fish were originally purchased from Ekkwill Waterlife Resources (Ruskin, FL, USA) and the Zebrafish International Resource Center (ZIRC) (Eugene, OR, USA), respectively. Ten females and five males per fish line (separately) were placed in a Z-Park (Techniplast, West Chester, PA, USA) breeding tank one night prior to breeding. The next morning, one hour after fertilization, fertile eggs were transferred to Petri dishes containing embryo-rearing medium E3 (31–33) and separated into five experimental groups (see next section). Then, Petri dishes were placed in incubators depending on the experimental condition (described below) and maintained in a photoperiod of 14 h light and 10 h darkness. After hatching, larval fish were maintained in E3. Larvae subsisted on their yolk sac reserves through day 7 and were purposefully not fed to minimize variation from exogenous energy sources as well as to avoid the effects of specific dynamic action [32].

2.2. Experimental Design

Three hundred fish (100 per replicate; 3 replicates) were employed from each fish line per experimental condition as follows:

• control group in normoxia (~7.8 mg L⁻¹ O₂) at 28 °C
• low temperature of 23 °C in normoxia
• high temperature of 33 °C in normoxia
• moderate hypoxia of ~4.9 mg L⁻¹ O₂ at 28 °C
• severe hypoxia of ~3.7 mg L⁻¹ O₂ at 28 °C

Fish were exposed to each condition from ~1 hour post fertilization (hpf) to 7 days post fertilization (dpf). Control and high temperature groups were exposed to their corresponding conditions in Hera Cell 240i incubators (Fisher Scientific, Pittsburg, PA, USA). Low temperature embryos and larvae were exposed to room temperature (~23 °C) during experimentation. Moderate and severe hypoxia groups were exposed in MCO-5M-PA incubators (Panasonic, American Laboratory Trading, East Lyme, CT, USA). Incubators were supplied with air lines to deliver either air or a mix of air and nitrogen gas. Gases were regulated with flow meters (Lab-Crest Century 100H, Cole-Palmer, Vernon hills, IL, USA) adjusted to required oxygen levels. A Beckman oxygen analyzer (OM-11, Beckman Coulter Inc., Brea, CA, USA) was used to assess oxygen levels once every 24 h, as well as immediately before experimentation. The oxygen analyzer was calibrated to 100% air saturation before each measurement.

All procedures were approved by the Institutional Animal Care and Use Committee of the University of North Texas (IACUC–24012).

2.3. Physiological Variables

Daily assessments of cardiac function (heart rate, stroke volume, and cardiac output), oxygen consumption, and yolk conversion efficiency were carried out starting at the same time each morning at 9:00 AM and following the same order among variables. The sample size was 10 individuals per endpoint, except for oxygen consumption in NHGRI-1 zebrafish, where the sample size was five. Embryos and larvae from each experimental condition were randomly selected from the three replicates.

Heart rate (f_H). Embryos inside their chorions and/or larvae were placed in 12-well IVF petri dishes (BIRR 4 + 8 Well Dish, IVFSTORE, Alpharetta, GA, USA) containing E3 medium. Ten embryos were placed individually in each well. Then, dishes were placed under a light microscope Nikon Eclipse E200 (Nikon Instruments Inc., Melville, NY, USA) held in a temperature-controlled Styrofoam chamber. The chamber was heated using a 20″ × 20.75″ heating pad (BN-LINK, Santa Fe Springs, CA, USA) regulated with a temperature controller BNQ-T7B(H) (BN-LINK, Santa Fe Springs, CA, USA) set to 28 °C. Embryos/larvae were left 15 min in their well before f_H measurement. Heart rate from each individual was measured by visual observations of the beating heart through the transparent body of the fish and was calculated as the number of beats in two successive intervals of 15 s then averaged and multiplied by 4 to obtain f_H in beats min⁻¹ [33].

Stroke volume (SV). The volume of blood pumped per heart beat (nL beat⁻¹) was calculated using the formula for a prolate spheroid [34]. The volume of the ventricle in diastole and systole was calculated as ${\displaystyle \frac{\pi }{6}}\times a\times b^2$, where a and b are the long and short axis of the ventricle, respectively. Stroke volume was calculated as the difference between calculated ventricular systolic and diastolic volume. To measure these ventricular dimensions, embryos were placed in IVF Petri dishes and observed, as for f_H measurements. For larval fish, each individual was embedded in 3% methylcellulose prepared with E3 and placed in Nunc^TM 35 mm IVF dishes (IVFSTORE, Alpharetta, GA, USA) by gently moving a needle around the individual until movement was restrained [35,36]. Larvae were positioned on their right side to allow better observation of the ventricle. Each SV measurement corresponded to each f_H measurement to calculate the cardiac function of each individual.

Cardiac output (Q). Cardiac output, Q in nL min⁻¹, was calculated from the product of the measured f_H and SV, f_H × SV.

Oxygen consumption (ṀO₂). Individual embryos/larvae were placed in a 24-well respirometry microplate system (OX11900, Loligo Systems, Viborg, Denmark). After calibration, the oxygen partial pressure in each well was measured every 10 min for 2 h. After respirometry trials, embryos and larvae were euthanized with buffered MS-222 (Millipore Sigma, Darmstadt, Germany) and fixed in 10% zinc-buffered formalin (Z-fix, Electron Microscopy Sciences, Hatfield, PA, USA) for 24 h. After fixation, embryos and larvae were rinsed in 1X PBS. Fixed embryos were mechanically dechorionated under a dissection microscope. Both embryos and larvae were weighed to the closest 0.01 mg using a digital scale Mettler Toledo XA105 (Columbus, OH, USA). ṀO₂ was then calculated using oxygen solubility at a given temperature (23, 28, or 33 °C), the decline in the partial pressure of oxygen (ΔPO₂), time in respirometer, volume of the respirometer, and body mass of embryos/larvae to express ṀO₂ as µL g⁻¹ h⁻¹.

Yolk conversion efficiency (YCE). After measuring the mass of each embryo/larvae, microdissections were carried out under a dissecting microscope to separate the embryo/larvae from the yolk sac. Yolk-free embryos were then weighed to the closest 0.01 mg. The mass of the yolk was calculated as the difference between the body mass and the yolk-free embryo mass. Then, both embryos/larvae and yolk sacs were dried for 48 h at 70 °C in a Heratherm IGS60 incubator (Fisher Scientific, Pittsburg, PA, USA). Once dried, yolk-free embryos, larvae, and yolk sacks were weighed to the closest 10 µg. Yolk conversion efficiency was calculated as [37,38].

$$YCE=100\times {\displaystyle \frac{yolk-free dry mass}{average dry yolk mass at 1dpf-dry yolk mass at a given endpoint}}$$

Cost of development (CoD). Cost of development, defined as the necessary oxygen to build a unit of fish mass [37], was determined as the overall ṀO₂ (mM O₂) divided by the dry mass (μg) of the embryo/larvae. Daily mean ṀO₂ was plotted as a function of developmental time and was integrated to calculate the area under the curve to obtain the overall ṀO₂. Obtaining replicates of ṀO₂ measurements from individual fish was not feasible because embryos were preserved before measuring body mass for ṀO₂ calculations and larvae are sensitive to manipulation. Individual measurements of ṀO₂ at each day post-fertilization (n = 10 for AB fish and n = 5 for NHGRI-1 fish) were randomly assigned to 1 of 10 (AB) or 5 (NHGRI-1) replicate relationships of daily mean ṀO₂ and developmental time. Each replicate was integrated to produce 10 (AB) and 5 (NHGRI-1) total values of consumed oxygen. The production of total O₂ values was repeated in triplicate. Then, the average of the overall ṀO₂ values was divided by the dry mass of the embryos/larvae.

2.4. Statistics

Measured variables were assessed as a function of temperature and oxygen availability, so data were separated into two different experiments to assess the effect of each stressor separately. To minimize the number of euthanized fish, we used the same control fish for both experiments.

General linear models (GLMs) were generated for f_H, SV, Q, ṀO₂, YCE, and CoD using time (dpf) and either temperature or oxygen availability as factors. The normality of each model was assessed using Shapiro–Wilks tests. Data that were not normally distributed were rank-transformed. Two-way ANOVAs were carried out to assess the GLM from each variable. Tukey post hoc tests were performed to assess significant differences as a function of each stressor and time. Significant differences were determined when p < 0.05, and non-significant differences when p > 0.05. A summary of detailed statistics is provided in Supplementary Tables S1 and S2. Differences in body mass and other morphological variables between the zebrafish strains have already been published [30]. General linear models, normality tests, data transformation, two-way ANOVAs, and Tukey tests were carried out in RStudio. Graphs were generated in Sigma Plot 15.1.

To quantify the extent of variability between the two different fish lines, the coefficient of variation (CV) associated with the mean values of each endpoint was calculated as $CV={\sigma }^{\prime }÷\overline{x}$, where ${\sigma }^{\prime }$ represents the standard deviation and $\overline{x}$ the mean [39]. Coefficients of variation from 0 to 1 are generally considered to reflect low variation, with a smaller value reflecting a lower amount of variation. Differences between CV from each fish line at a given endpoint were tested for significance using a Z-test. Both CV calculations and Z-tests were performed in Microsoft Excel 2023 [30]. Significant differences are shown as Zcalc < Ztab in the main text. To facilitate reading, Table S3 shows specific Z values for each comparison.

3. Results

3.1. Heart Rate

Heartbeat in control AB fish appeared at 2 dpf with an initial rate of 148 ± 6 beats min⁻¹ (Figure 1A and Figure 2A). At this early developmental stage, no significant differences were observed between control fish and those in the high temperature group (157 ± 5 beats min⁻¹; p > 0.05; Figure 1A). However, fish from low temperature, moderate hypoxia, and severe hypoxia all showed decreased heart rate compared to controls (96.4 ± 4, 103 ± 6, and 79 ± 7 beats min⁻¹, respectively; p < 0.05; Figure 1A and Figure 2A). Notably, heartbeat was observed as early as at 1 dpf in the high temperature group (62 ± 5 beats min⁻¹), reflecting their accelerated development because of higher temperature. Compared to control fish, low temperature fish showed decreased f_H throughout the measured stages (p < 0.05), but fish from high temperature showed no significant difference from controls (p > 0.05; Figure 1A). Significant differences in f_H disappeared after 3 dpf in both moderate and severe hypoxia fish compared to controls (p > 0.05; Figure 2A). In control NHGRI-1 fish, heart rate at 2 pdf was 112 ± 4 beats min⁻¹, which was higher than in low temperature and in moderate and severe hypoxia fish (75 ± 4, 82 ± 5, and 75 ± 5 beats min⁻¹, respectively; p < 0.05) but lower than f_H in high temperature fish (157 ± 2 beats min⁻¹; p < 0.05; Figure 1D and Figure 2D). No significant differences were observed between control and high temperature NHGRI-1 fish at 2 dpf (p < 0.05; Figure 1D), while fish from low temperature had lower f_H throughout the measurement period (p < 0.05; Figure 1D). Significant differences occurred between control and hypoxic groups at 2, 3, and 6 dpf (p < 0.05, Figure 2D).

To summarize heart rate changes, low temperature had a greater effect than hypoxia and high temperature, decreasing heartbeat frequency from 2 to 7 dpf in both fish lines. Additionally, both moderate and severe hypoxia decreased f_H during the first 2 dpf.

3.2. Stroke Volume

Stroke volume (SV) in control AB fish at 2 dpf was 0.24 ± 0.05 nL beat⁻¹. SV in fish from low and high temperature (0.13 ± 0.02, 0.28 ± 0.02 nL beat⁻¹) and moderate and severe hypoxia (0.16 ± 0.3, and 0.13 ± 0.02) was not significantly different from controls (p > 0.05; Figure 1B and Figure 2B). No differences between controls and the rest of the experimental groups occurred across development (p > 0.05; Figure 1B and Figure 2B), except for severe hypoxia fish at 4 dpf (0.30 ± 0.03 vs. 0.19 ± 0.02 nL beat⁻¹; p < 0.05; Figure 2B). In control NHGRI-1 fish, SV was 0.16 ± 0.02 nL beat⁻¹, which was higher than SV from low temperature fish (0.04 ± 0.004 nL beat⁻¹; p < 0.05) and lower compared to SV from high temperature fish (0.23 ± 0.02 nL beat⁻¹; p < 0.05; Figure 1E). In contrast, at 2 dpf, no significant differences were observed when comparing SV from control fish to moderate and severe hypoxia fish (p > 0.05; Figure 2E). Significant differences in SV against control fish were only observed in moderate hypoxia fish at 3 and 5 dpf (p < 0.05; Figure 2E).

To summarize stroke volume changes, few changes were observed as a function of either low or high temperature or hypoxia over the measured developmental range. The most evident alterations were observed at 2 dpf in NHGRI-1 fish, where low temperature decreased SV and high temperature increased SV compared to controls.

3.3. Cardiac Output

Cardiac output (Q) in control AB fish at 2 dpf was 35 ± 7 nL min⁻¹, which was significantly different from Q from low temperature and moderate and severe hypoxia fish (12 ± 2, 15 ± 3, and 10 ± 2 nL min⁻¹, respectively; p < 0.05; Figure 1C and Figure 2C). However, this was not different from high temperature fish (44 ± 7 nL min⁻¹; p > 0.05; Figure 1C). Low temperature, mild hypoxia, and severe hypoxia populations showed decreased Q compared to controls at 3 and 4 dpf (p < 0.05; Figure 1C and Figure 2C). Only at 7 dpf fish from high temperature showed higher Q than controls (p < 0.05; Figure 1C). Control NHGRI-1 fish showed Q = 19 ± 2 nL min⁻¹ at 2 dpf, which was higher (p < 0.05) than Q from low temperature (3 ± 2 nL min⁻¹) and severe hypoxia (9 ± 1 nL min⁻¹; Figure 1F) and lower compared to Q from high temperature fish (36 ± 3 nL min⁻¹; p < 0.05; Figure 2F). Throughout experimentation, Q from control fish was higher than Q from low temperature fish at 5 and 6 dpf and from severe hypoxia at 3 and 6 dpf (p < 0.05; Figure 1F and Figure 2F).

To summarize the cardiac output changes, low temperature and hypoxia decreased Q, especially during the first days of development in both lines, while high temperature induced an increase at 2 dpf in NHGRI-1 fish.

3.4. Oxygen Consumption

Oxygen consumption in control AB fish varied from 18 ± 3 to 39 ± 9 µL O₂ g⁻¹ h⁻¹ throughout the measured developmental range (Figure 3A,B). Lower ṀO₂ than in control fish occurred only in low temperature fish at 3 dpf and moderate hypoxia fish at 1 dpf (p < 0.05; Figure 3A,B). In contrast, increased ṀO₂ compared to controls was observed in high temperature fish at 3, 4, 5, and 7 dpf, in moderate hypoxia fish at 4 and 5 dpf, and in severe hypoxia fish at 1, 3–6 dpf (p < 0.05; Figure 3A,B). In NHGRI-11 control fish, ṀO₂ fluctuated from 5 ± 1 to 12 ± 1 µL O₂ g⁻¹ h⁻¹, exhibiting its highest value at 6 dpf (Figure 3C,D).

To summarize the oxygen consumption changes, ṀO₂ was increased by higher temperature in AB fish and by hypoxia in both lines.

Compared to controls, lower ṀO₂ was observed only in high temperature fish at 1 and 6 dpf and in and moderate hypoxia fish at 6 dpf (p < 0.05; Figure 3C,D). Higher ṀO₂ than in control fish was observed in high temperature and moderate hypoxia fish at 1 dpf and at 3, 5, and 7 dpf in severe hypoxia fish (p < 0.05; Figure 3C,D).

3.5. Yolk Conversion Efficiency

Yolk conversion efficiency (YCE) in control AB fish was 33 ± 3% at 1 dpf, increasing sharply to 114 ± 12% at 7 dpf (Figure 4A,B). Compared to control fish, YCE from high temperature fish showed no significant difference throughout temperature experiment (p > 0.05; Figure 4A). In contrast, YCE from low temperature, moderate hypoxia, and severe hypoxia fish was lower compared to control fish (p < 0.05; Figure 4A,B). In NHGRI-1 fish, YCE from control fish was 44 ± 1% at 1 dpf and increased to 91 ± 4% by 7 dpf (Figure 4C,D). Fish from the low temperature group showed slight but significant differences compared to controls at 3 and 6 dpf (p < 0.05; Figure 4C), while fish from high temperature had higher YCE from 2 to 6 dpf (p < 0.05; Figure 4C). Regarding moderate hypoxia and severe hypoxia groups, YCE was lower compared to controls from 2 to 7 dpf (p < 0.05; Figure 4D).

To summarize the yolk conversion efficiency changes, low temperature decreased YCE in AB fish, high temperature increased YCE in NHGRI-1 fish, and hypoxia decreased YCE in both fish lines.

3.6. Cost of Development

Cost of development (CoD) ranged from ~188 to ~454 μM O₂ mg⁻¹ in control AB fish, which was lower than in low temperature fish at 2 and 5 dpf and high temperature fish at 5 dpf (p < 0.05; Figure 5A). AB fish from moderate hypoxia had higher CoD from 4 to 6 dpf compared to controls, and severe hypoxia fish showed increased CoD from 1 to 6 dpf (p < 0.05; Figure 5B). In NHGRI-1 fish, CoD ranged from 43 to 174 μM O₂ mg⁻¹ in control fish. Cost of development was higher in low temperature fish at 2 and 5 dpf, and it was higher at 1 dpf and lower at 4, 6, and 7 dpf compared to control fish (p < 0.05; Figure 5C). Hypoxia-exposed NHGRI-1 fish had increased CoD from 1 to ~5 dpf compared to controls (p < 0.05; Figure 5D).

To summarize the cost of development changes, low and high temperature had little influence on AB fish, but higher temperature decreased CoD and low temperature decreased CoD in NHGRI-1 fish. Hypoxia increased CoD in both fish lines.

3.7. Coefficients of Variation

Of the 195 total endpoints we analyzed, ~84% had significantly lower coefficients of variation—i.e., lower variability—in the NHGRI-1 zebrafish line compared to the AB line (Z_calc > Z_tab; Figure 6, Figure 7, Figure 8, Figure 9, Figure 10 and Figure 11). There were no significant differences in variability of heart rate between fish lines at 2 dpf in control, low temperature, moderate hypoxia, and severe hypoxia fish (Z_calc < Z_tab); In addition, there were no differences in variability in control fish at 3 dpf and severe hypoxia fish at 4 dpf (Z_calc < Z_tab; Figure 6A–E). However, significant differences in variability of stroke volume from both fish lines occurred in most endpoints (Z_calc > Z_tab), except for 7 dpf in control and high temperature fish (Z_calc < Z_tab) and at 5 dpf in severe hypoxia fish (Z_calc < Z_tab; Figure 7A–E). Similarly, no significant differences in variability of cardiac output were observed at 5 dpf in severe hypoxia fish, at 6 dpf in control fish, and at 7 dpf in high temperature fish (Z_calc < Z_tab; Figure 8A–E). Lack of significant differences in variability of ṀO₂ occurred mainly towards the end of the measurement period at 6 and 7 dpf in control and low temperature fish, at 5 and 7 dpf in high temperature fish, at 5 dpf in moderate hypoxia fish, and at 4 and 6 dpf in severe hypoxia fish (Z_calc < Z_tab, Figure 9A–E). All endpoints for variability in yolk conversion efficiency showed significant differences between AB and NHGRI-1 fish (Z_calc > Z_tab; Figure 10A–E). CoD variability for most endpoints was significantly different between AB and NHGRI-1 fish (Figure 11A–E), except at 7 dpf in control fish, 3 and 5 dpf in low temperature fish, 2, 5, and 6 dpf in high temperature fish, and 2 and 4 to 6 dpf in severe hypoxia fish (Z_calc < Z_tab).

To summarize the variability changes in all measured parameters, NHGRI-1 fish showed smaller coefficients of variation compared to AB fish in ~84% endpoints.

4. Discussion

Values of heart rate, stroke volume, cardiac output, and oxygen consumption from our study are consistent with previously reported values [31,40,41,42,43,44,45,46], especially for AB zebrafish. However, control values for NHGRI-1 zebrafish were typically slightly lower than in control AB zebrafish. Lipid metabolism in the yolk sac [47], mathematical models describing the relationship between yolk absorption and growth [48], and how environmental factors influence the absorption of the yolk [46] have been reported, but we know of no studies of yolk conversion efficiency in zebrafish.

4.1. Physiological Variation Across Taxa

Revealing mechanisms by which physiological variation occurs within and across taxa facilitates our understanding of the contribution of inter-individual variation to the range of variation within a species to the relationship among species and their environment [2]. In invertebrates, major physiological mechanisms (e.g., respiration, digestion, reproduction, behavior, etc.) are highly conserved across taxa with different degrees of complexity, facilitating the investigation of evolutionary adaptations against environmental challenges [49,50] as well as mechanisms associated with plasticity in vertebrates [51,52]. Physiological variation in vertebrates from either natural or laboratory populations influences hypoxia and thermal tolerance [53,54,55,56,57,58], metabolic rate, stress response, developmental plasticity [59], and cardiac function [60]. Essentially, physiological responses are well conserved across the evolutionary scale, and variation in the degree of physiological responses is evident within and among population/species that inhabit different natural environments or when exposed to either similar or completely different controlled laboratory conditions. Moreover, physiological variation is one of the pillars leading to natural selection and evolution [61], even at the molecular level [62]. However, such variation greatly complicates assessment of how animals respond physiologically to environmental stressors. To overcome the obfuscation that variability can bring, the use of research models with decreased variation is a useful strategy to reveal specific biological characteristics [30].

4.2. Physiological Variation Among Zebrafish Strains/Lines

Molecular, physiological, morphological, and behavioral differences have routinely been described among various zebrafish strains/lines—e.g., [63,64,65,66,67,68,69,70,71,72]. However, little is known about physiological variation within strains/lines, as distinct from differences between strains/lines. Among the few such studies, different wild-type zebrafish strains/lines, including AB, Tübingen (TU), Tupfel long fin (TL), and Wild India Kolkata (WIK), differ significantly in, for example, fin length [71,73]. The long fins in TL zebrafish decrease swimming performance compared to AB zebrafish with regular-size fins [71]. In adult zebrafish, swimming performance has a strong correlation with temperature in AB, WIK, and TU zebrafish. WIK zebrafish showed significantly lower swimming capability compared to AB and TU zebrafish [71]. As another example of physiological differences between lines, basal glucose levels in WIK zebrafish are lower compared to TU zebrafish, resulting in growth performance differences between TU and other fish lines [68].

Our study supports our hypothesis that physiological responses in AB wild-type and NHGRI-1 zebrafish will be similar in response to the environment. The evidence is that most of the physiological responses of AB and NHGRI-1 zebrafish were qualitatively similar across time and as a function of changes in temperature and oxygen availability. Consistent with our findings, heat stress affects zebrafish development and growth [74,75] but, again, little information is available about physiological variability among larval zebrafish from different strains/lines in this regard. In embryonic zebrafish, exposure to heat stress delays segmentation to subsequent stages, accelerating development [74]. Heat pulses during embryonic development promote accelerated growth in zebrafish but result in decreased growth in later life stages [75]. Although we documented decreased variation in cardiac function in NHGRI-1 zebrafish, lower temperature decreased heart rate and cardiac output similarly in both AB and NHGRI-1 lines. In contrast, high temperature resulted in little to no significant differences in cardiac function compared to controls. Our data suggest that a relatively high temperature of 33 °C is not a severe enough temperature stressor enough to produce major alterations in cardiac function of embryonic and larval AB and NHGRI-1 zebrafish. A maximum plateau in heart rate where heart rate drops from cardiac heat stress zebrafish has been observed at ~38 °C [76]. Moreover, short-term cardiac responses to heat stress occur mainly in heart rate rather than alterations in ventricular tissue (either hypertrophy or hyperplasia) or contractile function of the heart [76].

As was observed in our temperature experiments, NHGRI-1 zebrafish showed decreased variability compared to AB fish as a function of hypoxia in all analyzed variables. Zebrafish adults and larvae employ aquatic surface respiration (ASR) as a behavioral response to hypoxia, with its onset in larvae occurring as early as 5 dpf [77]. However, this behavior was not observed in our larval fish from both AB and NHGRI-1 lines, probably because larvae were incubated in combined aquatic and aerial hypoxia. Additionally, it is possible that fish may have relied more on aquatic cutaneous respiration, which accounts for the major gas exchange mechanism in larval fish [78,79,80]. Different zebrafish strains/lines react differently under hypoxia exposure. Hif-1α knock out lines exhibit compromised hypoxia tolerance [77], especially when ASR is limited [81]. foxO4 zebrafish have increased hypoxia tolerance and survival and lower oxygen consumption compared to wild-type zebrafish [82]. As another example, HRE:mCherry reporter zebrafish show red fluorescence proteins in tissues that experience low oxygen tension [83]. In our study, at the end of the experiment, most of the assessed physiological variables in hypoxic groups showed no significant differences between strains. It is possible that a series of mechanisms served as self-repair capabilities [84] that minimized differences among hypoxic and control groups. Some self-repair mechanisms may be associated with increased movement and fin agitation to increase water flow thorough respiratory epitheliums and facilitate oxygen diffusion [85,86], enhanced blood–oxygen affinity [87,88], or the modification of cardiac activity [89]. Increased cardiac output occurs in early-stage zebrafish after hypoxia exposure [89], which was only evident in control AB fish compared to their hypoxic counterparts in our study. These differences were driven especially by increased heart rate, because no significant differences in stroke volume were observed in AB fish across experimentation. Not surprisingly, hypoxic groups from both AB and NHGRI-1 fish showed decreased YCE, which may be associated with developmental delays promoted by hypoxia exposure [90].

Changes in the overall oxygen consumption and dry mass of individuals across developmental time can be used to determine the ‘cost’ of development (CoD) [37,91]. Few studies [37,91] have focused on quantification of cost of development in fish in the framework described in the current manuscript, which sets a detailed baseline for assessing oxygen requirements to produce units of mass of zebrafish embryos under different environmental conditions. In our study, CoD in AB fish was but little altered by temperature. However, in NHGRI-1 fish, low temperature increased CoD, while high temperature decreased it. These data are consistent with those from the Australian lungfish Neoceratodus forsteri, where lower temperature increased CoD [91]. In contrast, CoD increased when embryonic lake whitefish Coregonus clupeaformis was incubated in increased constant temperatures, and CoD was also influenced by temperature shifts [37]. Additionally, in our study CoD was more affected by hypoxia than temperature in both AB and NHGRI-1 lines. Hypoxia impairs growth and may affect development by increasing time to hatch in fish, decreasing hatching success, and may produce a high degree of malformations [45,85,90,92,93]. CoD in our hypoxic fish was significantly higher in both fish lines. A strong inverse relationship exists between CoD and yolk conversion efficiency– that is, a higher CoD is correlated with lower YCE, and vice versa. Thus, low temperature and hypoxia not only delayed developmental progression but also development was energetically more costly and less efficient in transferring yolk mass to the embryo (Figure 4 and Figure 5). Interestingly, at the end of the experiments (7 dpf), most values of YCE and CoD were similar among groups in both fish lines. This suggests that fish, even at early developmental stages, may employ self-repair mechanisms [84]. Future studies will assess possible long-term consequences regarding increased cost of embryonic development.

4.3. Zebrafish as a Model for Studying Mechanisms of Physiological Variation

Different zebrafish strains/lines have unique origins that contribute to their observed and quantifiable genetic variation. Some laboratory strains/lines were established from wild isolates through different methods including gynogenetic diploids, half-treated diploids, or inbreeding [71,94]. The need for a robust inbred line with decreased genetic variation led to the development of the NHGRI-1 zebrafish line. NHGRI-1, with only ~15% the heterozygosity observed in the AB fish line, is a useful high-throughput line for genomic and genetic studies [29,95]. The development of this line enhances the reproducibility and reliability of zebrafish studies [95]. In fact, lower morphological variability (associated with lower genetic variability) in NHGRI-1 fish when compared to AB zebrafish was observed for multiple traits including yolk–chorion ratio, whole body mass, yolk-free embryo mass, total length, condition factor, and specific growth rate [30]. Our current results support our second hypothesis—that measured physiological variation will be lower in NHGRI-1 zebrafish compared to AB fish. Indeed, the low intrinsic variability characteristic of NHGRI-1 zebrafish resulted in less variable physiological responses compared to AB zebrafish. Importantly, although both fish lines had similar responses, the coefficient of variation was significantly lower in 163 of 195 (~84%) of all assessed endpoints. These results indicate that the higher heterozygosity of AB zebrafish produced a wider range of physiological phenotypes that increased variability in cardio-respiratory and other responses. A similar conclusion was reached regarding morphological variation in our previous study on these two lines [30].

Although genetic variability leads to different responses in zebrafish, the zebrafish remains an excellent model for studying physiological variation, for several reasons. (1) Accessible genomic resources. Zebrafish have fully sequenced genomes, which facilitates genetic and transcriptomic studies to explore the regulation and expression of genes in response to various physiological conditions [96,97]. (2) Physiological and behavioral traits in a climate change landscape. Despite the fact that zebrafish are routinely reared in laboratories at 27–28 °C, zebrafish are in fact poikilothermic and eurythermal piscine models that can tolerate wide ranges of temperature. Their natural habitat experiences significant temperature fluctuations, allowing the study of adaptive physiological responses associated with climate change and thermal stress [75,97]. (3) Developmental biology. Zebrafish embryos and larvae develop quickly (~48–72 hpf [98]), allowing the assessment of developmental processes and their alterations under different physiological conditions (e.g., embryonic development and growth) [75,76]. (4) Experimental versatility. Zebrafish are animal models that are relatively easy to maintain in laboratory settings. They require less space compared to other model organisms and they have high fecundity, which makes them cost-effective for large-scale studies [76] such as epigenetic inheritance studies where multiple generations are necessary [53].

5. Conclusions

Our results supported our hypotheses that both AB and NHGRI-1 zebrafish lines would show similar physiological responses in response to the environment and that physiological variation would be lesser in NHGRI-1 fish. Supporting evidence is that (1) AB and NHGRI-1 zebrafish had similar physiological responses to different temperatures and oxygen availability and (2) the decreased genetic variability from NHGRI-1 zebrafish led to lower variation in about 4/5 of the assessed endpoints compared to AB zebrafish. The decreased variation in the data from NHGRI-1 fish allowed the statistically significant quantification of physiological differences that were less evident in AB zebrafish due to their higher heterozygosity. The decreased variability in NHGRI-1 fish facilitated the interpretation of physiological responses and their association with the environment. A greater range of phenotypes produces increased genetic variability in AB zebrafish, which can limit not only the interpretation of experimental outcomes but also the experimental reproducibility [10]. Thus, the decreased variability observed in NHGRI-1 fish allows focusing on morphological, physiological, genetic, and genomic key mechanisms without the ‘noise’ of variable genetic backgrounds.