Chronic Thermal Stress During Early Zebrafish (Danio rerio) Development Induces Morphological, Molecular, and Liver Histopathological Changes

Abstract

Temperature is a critical abiotic factor mediating the physiological fitness of fish. While the impact of acute high-temperature exposure is well documented in teleost fishes, the effects of chronic thermal stress, especially during early stages, remain poorly understood. This study examined the effects of prolonged exposure to elevated temperatures (34 ºC) on zebrafish (Danio rerio) development, survival, molecular responses and liver histology during pre-independent (24–120 h post fertilisation [hpf]) and independently feeding (240–480 hpf) stages. While survival was not affected by elevated temperature, normal development was significantly impaired in both stages. Compared to control conditions (28 °C), heat exposure (34 ºC) increased the incidence of deformities, including spinal and yolk sac abnormalities during the pre-independent feeding stage, and spinal and growth-related deformities during independent feeding. Heat-induced changes in gene expression were most evident during independent feeding, with the upregulation of heat shock proteins (HSP90AA1, SERPINH1A, SERPINH1B) and downregulation of growth (GH, GHRA, IGF-1) genes. By 480 hpf, pronounced liver changes were also observed in heat-exposed fish, characterised by marked cellular vacuolation and hepatic glycogen accumulation. These results highlight the complex, age-specific responses to chronic thermal stress, reflected in altered heat shock response, development, and hepatocyte morphology. These findings contribute to the assessment of stage-specific responses relevant to biomarker development for prolonged heat exposure in developing finfish.

Introduction

Temperature significantly influences the biochemical and physiological processes essential for normal development in organisms (Afonso 2020). This is particularly concerning for many fish species, as climate change has intensified temperature fluctuations in both freshwater and marine environments (Akbarzadeh et al. 2018). For ectotherms, elevated temperatures beyond optimal ranges can induce thermal stress, triggering responses across all levels of biological organisation and potentially affecting development, growth, performance, and survival (Dimitriadi et al. 2018). Understanding temperature effects across developmental stages is critical, as thermal sensitivity often varies with ontogeny (Álvarez-Quintero et al. 2025). Embryonic and larval stages are particularly vulnerable due to underdeveloped physiology and heightened sensitivity to environmental stressors, with temperature having a stronger impact on their development and tissue formation than in adults, increasing mortality risk (Seebacher & Bamford 2024; de Souza et al. 2025).

Zebrafish (Danio rerio), widely utilised as a model species for investigating temperature stress due to its conserved stress response mechanisms and tolerance to diverse abiotic conditions, remains insufficiently explored regarding the chronic effects of elevated temperatures on early developmental processes and subsequent physiological outcomes (Ribas & Piferrer 2014). Wild zebrafish inhabit thermally variable environments such as ponds, lakes, and rivers across South Asia, where adults have been observed in water temperatures ranging from 12.3 °C to 38.6 °C (Arunachalam et al., (2013). In contrast, adult laboratory-reared zebrafish maintained under optimal conditions (28 °C) may exhibit upper thermal tolerance limits exceeding 40 °C, however, embryos and larvae likely possess narrower thermal ranges (Morgan et al. 2018; Pype et al. 2015). Exposure to 36.5 °C in embryos has been associated with high mortality, while sustained temperatures near 32 °C significantly elevate the incidence of morphological abnormalities, notably impairing cardiac and spinal development including pericardial and yolk sac edema, spinal curvature, and cranial malformations (Pype et al. 2015). Additionally, zebrafish embryos at 72 h post-fertilisation (hpf) reared at 30 °C exhibited a 7% rise in pericardial edema compared to those maintained at 27 °C (Duan et al. 2023). Several toxicological and gene knockdown studies, not directly related to temperature, have implicated aberrant regulation of key developmental genes such as BMP2B, ZHE1, and ACVR in the emergence of these abnormalities (Allen et al. 2020; Priyam et al. 2022; Karas et al. 2020). Nevertheless, studies examining how chronic heat stress responses evolve from the embryonic stage through larval development remain limited.

Although our understanding of the chronic heat stress response is expanding, current knowledge remains mostly derived from studies in salmonids (Akbarzadeh et al. 2018). It is well established that heat stress can compromise protein structure and cellular integrity in fish, thereby activating the heat shock response to maintain protein folding and cellular stability (Madeira et al. 2014). Supporting this, acute heat stress investigations in whole early-stage zebrafish have demonstrated upregulation of heat shock proteins, including HSP70 and SERPINH1 (HSP47) (Long et al. 2012; de Souza et al 2025; Hallare et al. 2005). HSP70, HSP90, and SERPINH1 are conserved markers of thermotolerance and cellular stress in fish, with expression and function varying across tissues and developmental stages (Ignatz et al. 2024; Afonso 2020). As thermal stress can impact physiological, behavioural, and cellular responses, these changes have also been shown to secondarily affect growth and survival (Yan et al. 2024). During early development, fish experience rapid morphological changes that are partially regulated by the GH-IGF-1 axis, in which pituitary-derived growth hormone (GH) binds to hepatic GH receptors (GHR), stimulating the liver to produce insulin-like growth factor 1 (IGF-1) (Nakano et al. 2015). These hormones are temperature-sensitive, with the expression of genes such as GH and IGF-1 shown to be downregulated under heat stress, correlating with reduced growth (Shahjahan et al. 2021). Despite this knowledge of the acute stress response, there is limited research on the effects of chronic exposure to elevated temperatures in early zebrafish and its effects on the GH-IGF-1 axis.

The liver plays a central role in detoxification and metabolic homeostasis and is highly responsive to environmental stressors, including temperature fluctuations (Pham et al. 2017).

Histopathological analysis serves as a valuable method for evaluating the impact of thermal stress on liver health, particularly through assessments of hepatocyte morphology and tissue composition (Han et al. 2023). Both acute and chronic exposure to elevated temperatures have been shown to induce hepatic vacuolation, blood infiltration, and cellular necrosis in juvenile largemouth bass (Micropterus nigricans) and pikeperch (Sander lucioperca), pathological changes indicative of liver damage (Zhao et al. 2022; Liu et al. 2022). Concurrently, thermal stress is known to disrupt hepatic energy metabolism, with prolonged exposure often leading to glycogen mobilisation to meet increased cellular energy demands (Yan et al. 2024). Therefore, investigating whether chronic heat stress drives both structural and metabolic alterations in the liver, reflected in glycogen deposition, may provide valuable insights into hepatic responses to elevated temperature. In zebrafish, the liver becomes fully functional by 96 h post-fertilisation (hpf) (4 days), with independent feeding beginning around 168 hpf (7 days) (Pham et al. 2017). These early developmental stages represent a critical window during which histological approaches may reveal early indicators of hepatic stress or damage in response to chronic thermal exposure.

Accordingly, the present study examined the effects of prolonged thermal exposure (34 °C) on development, survival, molecular responses, and liver histology in zebrafish during both pre-independent feeding (24–120 h post fertilisation [hpf]) and independently feeding (240–480 hpf) stages. It was hypothesised that exposure to elevated temperature (34 °C), relative to the control (28 °C), would significantly modulate the expression of genes associated with heat shock response and growth, and lead to adverse effects on developmental progression, hepatocyte morphology, growth and survival. We also aimed to evaluate the temporal expression patterns of these genes to determine whether they exhibit consistent regulation under chronic thermal stress and therefore be used as reliable molecular markers of chronic thermal exposure.

Materials and Methods

Embryo collection

All procedures in this study were approved by the Deakin University Animal Ethics Committee (Permit No. G08-2023). Adult zebrafish were housed at the Deakin University Aquatics Facility (Waurn Ponds, Australia) in 3 L tanks (30 fish per tank, 28 °C, 80% relative humidity, under a 12 h light:12 h dark photoperiod) with a continuous flow of aerated water (RAS system) (Tecniplast, Italy). Embryos for experimentation were obtained by breeding adult zebrafish (2:3 male to female ratio) using the Mass Embryo Production System (MEPS®) (Pentair Aquatic Eco-Systems, USA). Fertilised eggs were collected with a sieve and transferred to 90 mm plastic petri dishes (100 embryos per dish) containing water from the recirculating aquaculture system (RAS) maintained at 28 °C. Embryos were monitored regularly for dead or unfertilized eggs, and the aquatic water was replaced (100%) every 8 h. Embryos were then incubated at 28 °C and 80% relative humidity in a temperature-controlled chamber (Labec, Australia) under a 12 h light:12 h dark photoperiod.

Experimental design

The experiment was conducted to evaluate the effects of thermal stress at different developmental stages: Stage 1: pre-independent feeding stage (Pre-IF: 24–120 hpf) and Stage 2: independent feeding stage (IF: 240–480 hpf), respectively, to account for differences in feeding ability and rearing conditions (Fig. 1). There were two temperature treatments (28 °C and 34 ºC) in each stage, and six and three sampling times, respectively (Fig. 1). The experiment was replicated twice, including mass breeding events.

Fig. 1.

Experimental design. Stage 1: Pre-Independent Feeding (Pre-IF). Zebrafish embryos at 24 hpf were exposed to either 28 ºC (control, green) or 34 ºC (elevated temperature, red) until 120 hpf (in 6-well plates). Stage 2: Independently Feeding (IF). At 120 hpf, remaining fish were transferred to tanks and maintained at established thermal treatments until 480 hpf. Sampling was conducted at 24 hpf (day 1), 48 hpf (day 2), 72 hpf (day 3), 96 hpf (day 4), 120 hpf (day 5), 240 hpf (day 10), 360 hpf (day 15) and 480 hpf (day 20)

Stage 1: Pre-Independent Feeding (24–120 hpf)

Stage 1 was replicated twice over time using 6-well plates, with six replicate wells per treatment at each sampling time (10 fish per well; n = 12 per treatment). At 24 h post-fertilisation (hpf), ten age-synchronised zebrafish embryos maintained at 28 °C were randomly transferred to each well containing 5 mL of aquarium water (Raghunath and Perumal 2018; Priyam et al. 2022). Immediately after transfer, control plates were placed in an incubator at 28 °C, and heat stress plates were placed in an incubator at 34 °C. Both incubators were maintained at 80% relative humidity with a 12 h light:12 h dark photoperiod. Aquarium water, matched to each treatment temperature, was renewed daily (50%; ~ 2.5 mL per well). Water quality parameters were monitored daily and maintained at pH 7 ± 0.5, nitrates < 40 mg/L, nitrite < 5 mg/L, general hardness < 30 mg/L, and carbonate hardness < 40 mg/L. Plates were inspected daily under a stereoscopic microscope (Nikon SMZ 745) to assess hatch rates and survival. Dead embryos or larvae, identified by an opaque body and absence of a heartbeat, were promptly removed. Fish were sampled at 24 hpf (prior to heat exposure), and at 48, 72, 96, and 120 hpf (Fig. 1). At each time point, ten fish from each of 12 wells per treatment were euthanised in 1% (w/v) MS-222 buffered with 1% (w/v) NaHCO₃ (Priyam et al. 2022). Individual images of all humanely killed fish were captured using Mosaic software v2.0 (Tucsen Photonics Co., China) for morphometric and deformity assessment. Images were examined to identify developmental deformities, including yolk sac edema (YSE), pericardial edema (PE), and spinal curvature (SC) (Park et al. 2020; Priyam et al. 2022; Raghunath and Perumal 2018). The proportions of total deformities and of each specific type were expressed as percentages of the total number of fish per well. Hatch and survival rates were expressed similarly. ImageJ v1.54 (Schindelin et al. 2012) was used to measure total and fork lengths at 72, 96, and 120 hpf (Andrialovanirina et al. 2020). For molecular analyses, fish from four wells per treatment (10 individuals per well) were collected at 24, 48, and 72 hpf, rinsed with 1 × PBS, and preserved in RNAlater (Merck, Germany) at − 80 °C.

Stage 2: Independently feeding (240–480 hpf)

At the completion of Stage 1 (120 hpf), 63 fish per thermal treatment were randomly allocated to three 0.7 L tanks and maintained under their respective thermal regimes (28 °C and 34 °C) in environmental incubators for an additional five days (up to 240 hpf). This stage was replicated twice over time (n = 2 independent trials) (Gamperl et al. 2020; Gerber et al., 2021). Each replicate consisted of three tanks per thermal treatment, with approximately 21 fish per tank, and each tank representing a single sampling time (240, 360, and 480 hpf). At each designated time point, all fish from one tank were sampled. Both incubators were maintained at 80% relative humidity with a 12 h light: 12 h dark photoperiod. Fish were fed ad libitum with rotifers (Brachionus plicatilis) (Ribas and Piferrer 2014). At approximately 288 hpf, fish were transferred to 3 L tanks and maintained at their respective thermal treatments within independent recirculating aquaculture systems (RAS; Tecniplast, Italy) set to 28 °C and 34 °C. Water quality parameters were monitored regularly and maintained at pH 7.0 ± 0.08, electrical conductivity 493–529 µS m⁻¹, dissolved oxygen ≥ 80% saturation, and temperature of 27.9 ± 0.1 °C (control) and 33.8 ± 0.1 °C (heat stress). Fish were sampled at 240, 360, and 480 hpf (Fig. 1). At each sampling point, 21 fish from two replicate tanks per treatment were euthanised in 600 mg L⁻¹ benzocaine (Raghunath 2024). Individual images of euthanised fish were captured using Mosaic v2.0 (Tucsen Photonics Co., China) for morphometric and deformity analyses. Morphological abnormalities, including growth retardation (GR) and spinal curvature (SC), were quantified, and the proportions of deformities and survival were expressed as percentages of the total number of fish per tank. In this study, growth retardation was broadly classified as a spinal-derived abnormality (excluding curvature) that results in a visibly smaller, stunted body size or altered and uncoordinated body proportions relative to normal, healthy fish (Huang et al. 2018; Lantz-McPeak et al. 2015; Yang et al. 2025). Morphometric parameters, including fork length (mm), total length (mm), spinal cord length (mm) and eye area (mm²), were measured from images using ImageJ v1.54 (Schindelin et al. 2012) following Andrialovanirina et al. (2020). For molecular analyses, seven fish per tank (14 fish per treatment) were collected at each sampling time, rinsed in 1 × PBS, and preserved in RNAlater (Merck, Germany) at − 80 °C. At 480 hpf, an additional four fish per tank (eight per treatment) were fixed in 10% neutral buffered formalin for histological examination.

Total RNA extraction and cDNA synthesis

Total RNA was extracted from pools of 10 individuals per treatment from 24 hpf, 48 hpf and 72 hpf (n = 4 pools per treatment) (Pre-IF), and pools of 3 larvae sampled at 240 hpf (n = 4 pools per treatment) (IF) (Priyam et al. 2022). Total RNA extraction at 360 hpf and 480 hpf was conducted on 14 individual larvae per treatment (n = 7 fish per tank) (IF). Samples were first rinsed in 500 µL of nuclease-free water and blotted dry of excess RNAlater on a Kimwipe. The Aurum™ Total RNA Mini Kit (Bio-Rad, USA) was then used to homogenise samples and perform RNA extractions for samples across both stages following the manufacturer’s spin protocol for animal tissue (Brown et al. 2021). Total RNA yield and purity was determined with a spectrophotometer (NanoDrop ND-1000, Thermo Fisher Scientific, USA). RNA integrity of selected samples was further inspected by visualisation on a 1% 1X TAE agarose gel with GelRed (Gene Target Solutions, Australia). Total RNA samples were normalised to 30 ng μL⁻¹ with nuclease-free water using a spectrophotometer. Residual gDNA was then removed using the TURBO DNA-free ™ Kit (Thermo Fisher Scientific, USA) following the manufacturer’s specifications for a ‘routine DNase treatment’ (Brown et al. 2021), and the total RNA concentration determined using a spectrophotometer (standardised to approximately 20 ng μL⁻¹). cDNA was synthesised in a 20 μL reaction using the iScript™ cDNA Synthesis Kit (Bio-Rad, USA) following the manufacturer’s recommendations. The same protocol was followed using No Reverse Transcription (NRT) Supermix (Bio-Rad, USA), although total RNA samples were pooled separately for pre- and post-independent feeding groups. cDNA and NRT samples were then diluted fourfold with nuclease-free water.

Gene selection and primer design for qPCR

To examine the effects of temperature on cellular heat responses and developmental processes, twelve genes were selected and classified into two primary functional categories: heat stress response and growth and development. Primer sequences for ten genes were obtained from previous studies in zebrafish in the literature (Table 1). For primer sequences of two genes not cited from the literature, primers were designed using PrimerQuest Tool and manufactured by Integrated DNA Technologies (Baulkham Hills, NSW, Australia) according to relevant gene sequences provided by http://www.ncbi.nlm.nih.gov for zebrafish. Primer specificity for both designed and cited primer pairs were validated using Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). Cited and designed primer sequences, references, amplicon sizes, and corresponding GenBank accession numbers are provided in Table 1. Expected amplicon size was verified for each primer set by visualising RT-qPCR products on a 2% 1X TAE agarose gel with GelRed alongside a 50 bp, 50 μg ladder (Bio-Rad, USA).

Table 1.

Primers for real-time qPCR assays (*designed primers)

Gene	Abbreviation	Accession no	Amplicon size	Forward Primer	Reverse Primer	Efficiency	Reference
Heat shock response
heat shock protein 70	HSP70	AF210640.1	102	GCACCACCTACTCCTGTGTGGG	CTGTGAAGGCAACATAGCTGGG	98.4%	Mottola et al. 2020
heat shock protein 90	HSP90AA1	NM_131328.1	122	GGGCACCATCGCTAAATC	CTTTCTCAGCCACCAGATAC	104.0%	*
peroxiredoxin 6	PRDX6	NM_200805.1	299	CCCGCTGCGTGTTTGTAGT	CTGCCCTTCAGGTTTCAGTTATG	95.9%	Xu et al. 2021
calmodulin 1a	CALM1A	NM_213351.1	126	GTGGCTTGGGAGTTTCAG	GATCAGAGCCCAGAGAAGA	100.0%	*
serpin peptidase inhibitor H1a	SERPINH1A	NM_001110374.2	208	CCAGTGGTAGTGGCCTCATC	CAGAACTGGGCCCGTAGAAG	100.8%	Lin et al. 2022
serpin peptidase inhibitor H1b	SERPINH1B	NM_131204.2	146	GTGAAAAACACAGACGGGGC	GTGCGATGCATCATTGGGAC	105.6%	Lin et al. 2022
Growth and development
growth hormone	GH	NM_001020492.2	161	TCGTTCTGCAACTCTGACTCC	CCGATGGTCAGGCTGTTTGA	96.9%	Chen et al. 2022
growth hormone receptor a	GHRA	NM_001083578.1	170	GGCCGAAAATTCCTTACTGTT	GCTGGCGTTGCTGATTGT	91.5%	Dang et al. 2018
insulin-like growth factor 1	IGF-1	NM_131825.2	183	CAACGACACACAGATATTCCCAGG	TCGGCTGTCCAACGGTTTCTCTT	107.9%	Horie et al. 2023
zebrafish hatching enzyme 1	ZHE1	NM_213635.2	53	GCCCGGTCTGGAAACCA	GTCCGATCTGCACGTTTTCA	104.5%	Karas et al. 2020
activin type 2 receptor	ACVR2AA	NM_001110278.2	155	AAGGCTGCTGGCTTGATGAT	TGATGTCGTCTGAACTGGCG	105.4%	Che et al. 2023
bone morphogenetic protein 2b	BMP2B	NM_131360.2	145	CGAGATCGACCGACGGAAAT	GACCACTGCCGATTTGCTTG	90.8%	Wu et al. 2022
Reference genes
eukaryotic translation elongation factor 1 alpha 1 like 1	EEF1A1L1	NM_131263.1	200	GGTACTACTCTTCTTGATGCCCTTG	GACTTGACCTCAGTGGTTACATTG	94.10%	Dorts et al. 2016
β-actin 2	ACTB2	NM_181601.4	197	ACTGTATTGTCTGGTGGTAC	TACTCCTGCTTGCTAATCC	100.0%	Luzio et al. 2013

Real-time qPCR

The real-time qPCR assays are reported in accordance with the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) (Bustin et al. 2009). Each 10 μL reaction contained 200 nm (0.2 μL) each of forward and reverse primers, 2.6 μL nuclease-free water, 1 × SsoAdvanced Universal SYBR® Green Supermix (Bio-Rad, USA) (5 μL) and 2 μL of diluted cDNA (or diluted NRT sample). Reactions were loaded in duplicate into a 96 clear-well hard-shell PCR plate, covered with a clear adhesive PCR plate seal and run on a CFX Real-Time PCR Detection System (Bio-Rad, USA). Thermal cycling conditions from Long et al. (2012) were used: 40 cycles of 10 s at 95 ºC and 30 s at 60 ºC. Baseline threshold was set to 140 relative fluorescence units, which was within the exponential phase of each run, and Quantification cycle (Cq) was determined using CFX Manager 3.1 (Bio-Rad, USA). Positive amplification was recorded for a given cDNA sample amplification was detected in both duplicate reactions with less than one cycle of deviation. Specific amplification of the target sequence was confirmed by conducting a melt curve analysis, which following the 40th PCR cycle, applied an incremental increase in temperature from 70 °C to 95 °C (increase of 1 ºC between each cycle) (Long et al. 2012). PCR efficiency for each gene was determined by conducting six twofold serial dilutions of cDNA pooled from all samples (Table 1).

Cq values were determined using the fit point method. The mean of the duplicates was used for relative quantification. Relative expression ratios of test samples were determined following methods described by Hellemans and Vandesompele (2014). GeNorm analysis (www.ciidirsinaloa.com.mx/RefFinder-master) was conducted to determine the most suitable reference genes for this study. Average Cq values from two fish per sampling time (total = 28) were included in the analysis, revealing that EEF1A1L1 and ACTB2 were the most suitable reference genes (M-value < 0.38).

The following equation used:

$$Relativegeneexpression=\frac{{((\text{primer efficiency gene of interest}/100)+1)}^{(\text{Calibrator Cq}-\text{Sample Cq gene of interest})}}{{(\text{GeoMean}(\left(\text{primer efficiency reference gene},/,100,),+,1\right))}^{(\text{Calibrator Cq}-\text{Sample Cq reference})}}$$

The average of the Cq values of the samples from 24 hpf was used to create a control average Cq and functioned as the calibrator sample in the relative quantification of gene expression.

Histology

Histological analysis was conducted to investigate the effects of heat stress exposure on 480 hpf hepatic morphology, cell structure and glycogen accumulation. After whole larvae (4 fish per tank, 8 fish per treatment) were fixed in 10% neutral buffered formalin at room temperature overnight, fish were rinsed twice with 1 × PBS and preserved in 70% ethanol. Paraffin processing, embedding, cutting and staining was performed by the Melbourne Histology Platform, The University of Melbourne (Parkville, Australia) using routine methods. In this study, two types of staining were used to identify changes in hepatocyte morphology (H&E) and glycogen deposition (PAS). For each sample, at least 25 serial 4 μm cross-sections that contained the liver were stained with hematoxylin and eosin (Brown et al. 2021). Another 3 slides were subjected to Picrosirius Red and Periodic acid–Schiff (PAS) staining under the same processing conditions. Stained sections were examined under a light microscope (10-400X) and photographed with an LED-powered fluorescent microscope (ZEISS Axio Imager 2; ZEISS, Germany). Slides were observed and quantified for differences in cell size, morphology, cytoplasmic vacuolation and glycogen accumulation (DeBroy and ImageJ 1997–2011) using ImageJ analysis software v1.54 (Schindelin et al. 2012). A score value was then assigned based on the degree and extent of cellular vacuolation and glycogen deposition as follows: 0 (no significant changes), 1 (mild changes), 2 (moderate changes) and 3 (marked changes) (Schafer et al. 2018). All slides were evaluated by a single investigator under blinded conditions.

Statistical Analysis

Pre-independent and independent feeding stages were analysed separately. R 4.0.2 was used to conduct all components of the statistical analysis (www.rproject.org). Linear mixed models were fitted to average relative expression values (Pre-IF and IF) and length (Pre-IF: 72–120 hpf) to observe significant differences amongst sampling times and thermal treatments (Pinheiro et al. 2017). In each of these models, the effect of sampling time and temperature (and their interaction) was tested with a two-way ANOVA, whereby sampling time and temperature were classed as fixed effects while experimental units (Pre-IF: wells, IF: tanks) represented a random effect (intercept) in the model to explain variation between experimental units. Planned post-hoc tests were used to examine specific primary and interactive effects, implementing a false discovery rate (FDR) to address type I errors during multiple comparisons (Benjamini and Hochberg 1995). Normality was inspected through quantile–quantile plots and Shapiro–Wilk tests on residuals. Homogeneity of variance was determined through subjecting residuals to a Levene’s test of equal variance for each fixed effects variable and visualising the distribution of residuals against fitted values on a scatterplot. Where necessary, transformations were used to improve heteroskedastic variance and non-normal distributions. Non-parametric Kruskal–Wallis tests followed by Dunn's tests were used to observe significant differences amongst sampling times and thermal treatments for survival and mortality rates in the pre-IF stage as well as morphometric parameters in the IF stage. Chi-square tests followed by pairwise Chi-square tests were used to assess differences between pre-IF deformed and non-deformed fish including spinal curvature, pericardial edema and yolk sac edema. Fisher’s Exact tests followed by pairwise Fisher tests were used to examine IF survival and deformity counts. Fisher’s Exact tests were also used to assess differences in hepatocellular vacuolation and positive PAS staining severity grading between treatments (qualitative analysis). Student’s t-test was used to examine differences in vacuole area, while Wilcox signed-rank test was used for assessing glycogen area (quantitative analysis). GraphPad Prism software, Version 10 (GraphPad Software Inc., San Diego, CA, USA) was used to present data as means ± SE and statistical significance was considered at a level of p < 0.05.

Results

Survival

Thermal treatments (28 °C and 34 °C) and age (hpf) did not significantly affect the survival of both the Pre-IF (Fig. 2a; p > 0.05) and IF fish (Fig. 3a; p > 0.05).

Fig. 2.

Pre-independent feeding zebrafish survival (a), hatch rate from 48 to 120 hpf (b), total length from 72–120 hpf (c), overall deformity rate from 72–120 hpf (d), Specific deformity rate for zebrafish 72–120 hpf (SC = spinal curvature, PE = pericardial edema, YSE = yolk sac edema) (e), exposed to 28 ºC (control, green) and 34 ºC (elevated temperature, red). All results are presented as mean (± SE). Lower case letters indicate temporal significant differences within temperature regimens. Asterisks indicate significant differences between thermal treatments within sampling time (p < 0.05) (n = 12)

Fig. 3.

Independently feeding zebrafish survival (a), morphometric parameters (b-e) and deformity rate (f) exposed to 28 ºC (control, green) and 34 ºC (elevated temperature, red) from 240 to 480 hpf. All results are presented as mean (± SE). Lower case letters indicate temporal significant differences within temperature regimens. Asterisks indicate significant differences between thermal treatments within sampling time (p < 0.05) (n = 4)

Hatch rate

Hatching activity was first evidenced at 48 hpf (Fig. 2b). From 48 to 72 hpf, there was a significant increase in hatch rate in both thermal treatments (p < 0.0001). By 72 hpf, approximately 95% of fish had hatched, and there were no significant differences in hatch rate between treatments (p > 0.05).

Morphometric parameters and deformity

During the pre-IF stage, fish significantly increased in length (p < 0.05) independent of the thermal treatment. Fish exposed to 34 °C were significantly longer (p < 0.05) at 72 hpf when compared to fish maintained at 28 °C, but by 96 and 120 hpf, length was no longer significantly different between thermal treatments (p > 0.05) (Fig. 2c).

Exposure to 34 °C significantly increased the incidence of deformities between 72 and 120 hpf (p < 0.05), however, the frequency of abnormalities decreased from 72 to 120 hpf (Fig. 2d). Analysis of specific deformities showed no significant difference in the proportion of individuals exhibiting spinal curvature at 72 hpf (p > 0.05). In contrast, a significant increase was observed in heat-stressed fish at 96 hpf and 120 hpf (p < 0.05). Thermal treatment and age did not significantly affect the number of fish exhibiting pericardial edema in pre-IF fish (p > 0.05). A significant increase (p < 0.05) in the incidence of yolk sac edema was observed in heat-stressed fish at 72 hpf and 120 hpf, but not at 96 hpf (p > 0.05). Within the heat-stressed treatment, significant temporal differences were detected for yolk sac edema, characterised by a decline in incidence from 72 to 96 hpf (p < 0.05).

For the IF stage, morphometric parameters included total length, fork length, spinal cord length and eye area (Fig. 3b-e). Independent of sampling time, the average fork and total length of fish maintained at 28 °C was significantly higher than in fish exposed to 34 °C (p < 0.05). By 480 hpf, fish maintained at 28 °C were on average 3 mm longer than fish exposed to 34 °C. Eye area was significantly larger in control fish compared with heat-stressed fish at all sampling times, although increased significantly over time in both treatments (p < 0.05) (Fig. 3e). Heat stress exposure significantly increased the frequency of deformities in from 240 to 480 hpf (p < 0.05), however, the frequency of deformities in both treatments did not alter significantly over time (p > 0.05) (Fig. 3f). These deformities, which included spinal curvature and growth retardation, were not evaluated as independent endpoints in the statistical analysis (Fig. 4 h, j, l).

Fig. 4.

Deformities observed during pre-independent feeding (a-f) and independently feeding (g-l). Images (left) show unaffected fish exposed to control (28 ºC) conditions and images (right) show fish affected by elevated temperature (34 ºC). SC = spinal curvature, YSE = yolk sac edema, PE = pericardial edema, GR = growth retardation. The scale of the images is 0.5 mm

Gene expression

Heat shock response

During the pre-IF stage (up to 72 hpf), no thermal-specific effects were observed (Fig. 5). During the IF stage, HSP70 was significantly downregulated at 360 hpf in the heat-stressed treatment when compared to the control (p = 0.0007) (Fig. 5a). Between 360 and 480 hpf, the control treatment also showed a significant decrease in HSP70 expression (p = 0.0008). Differential expression between treatments for HSP90AA1, SERPINH1A, and SERPINH1B was only observed in IF fish. HSP90AA1 was significantly upregulated in fish exposed to 34 °C (p = 0.0006) at both 240 hpf and 480 hpf (p = 0.0274) (Fig. 5b). SERPINH1A was significantly upregulated in fish exposed to 34 °C at 240 hpf (p = 0.0113) and 480 hpf (p = 0.0299) (Fig. 5c), while SERPINH1B was significantly upregulated at 240 hpf (p < 0.0001), 360 hpf (p = 0.0010) and 480 hpf (p < 0.0001) when compared to the control treatment (Fig. 5d). CALM1A was significantly upregulated in pre-IF fish in the heat-stressed treatment at 24 hpf (p = 0.0036) and 72 hpf (p = 0.0050), while in IF fish the control treatment exhibited significant increases in expression at 240 hpf and 360 hpf (p = 0.0252) (Fig. 5e). During the pre-IF stage, PRDX6 showed significant upregulation in both treatments at 48 hpf and 72 hpf, relative to 24 hpf (p < 0.05). During the IF stage, fish maintained at 28 °C showed significant upregulation of PRDX6 at 360 hpf (p = 0.0016) and 480 hpf (p < 0.0001) when compared to fish at 34 °C (Fig. 5f).

Fig. 5.

Mean (± SE) relative mRNA expression of heat stress response genes in zebrafish exposed to 2 temperature regimens (28 °C and 34 °C) for 480 h (20 days). Gene expression was analysed for pre-IF fish sampled at 24 hpf (day 1, prior to increase in water temperature), 48 hpf (day 2) and 72hpf (day 3) (n = 4 pools of 10 fish per treatment). IF fish were analysed at 240 hpf (day 10), 360 hpf (day 15) and 480 hpf (day 20) (n = 14 fish per treatment). All results are presented as mean (± SE). Letters (Pre-IF: uppercase, IF: lowercase) indicate temporal significant differences within temperature regimens. Asterisks indicate significant differences between thermal treatments within sampling time. 28 °C treatment is indicated by green bars and 34 °C in red

Growth and Development

During the pre-IF stage, growth hormone (GH) gene expression was markedly upregulated in both treatment groups at 72 hpf, and significantly higher expression in the heat-stressed treatment when compared to the control (p = 0.0158) (Fig. 6a). During the IF stage, GH expression was high in both temperatures, except for a significant downregulation in the heat stressed treatment at 480 hpf, when compared the control (p = 0.0037) (Fig. 6a). In both stages, GHRA was significantly upregulated in the control treatment at both 48 hpf (p = 0.0003) and 480 hpf (p < 0.0001), respectively (Fig. 6b). In pre-IF fish, IGF-1 was significantly upregulated at 72 hpf in both treatments compared to 24 hpf and 48 hpf (p < 0.05), however no differences were observed between treatments (Fig. 6c). But in IF fish, IGF-1 expression in the control treatment increased significantly at 360 hpf (p = 0.0004) and 480 hpf (p < 0.0001) compared to the heat-stressed treatment. During the pre-IF stage, BMP2B expression was significantly upregulated in the heat-stressed group at 24 hpf (p = 0.0261). Conversely, during the IF stage, BMP2B expression at 480 hpf was significantly higher in the control treatment compared with fish reared at 34 °C (p = 0.0036) (Fig. 6d). During the pre-IF stage, ACVR2AA was significantly upregulated at 72 hpf in both treatments relative to 24 hpf (p < 0.05) (Fig. 6e). Additionally, in IF fish, this gene showed higher expression in the control treatment compared to the heat-stressed treatment at 240 hpf and 480 hpf (p < 0.05). The hatching-related gene ZHE1 was significantly upregulated at 48 hpf in the control treatment compared to the heat stressed treatment (p = 0.0061) and within treatment expression at 24 hpf (p = 0.0171) (Fig. 6f), while expression significantly declined in both treatments by 72 hpf (p < 0.05), coinciding with the near-completion of hatching (∼100% of embryos hatched at this stage).

Fig. 6.

Mean (± SE) relative mRNA expression of growth and development genes in zebrafish exposed to 2 temperature regimens (28 °C and 34 °C) for 480 h (20 days). Gene expression was analysed for pre-IF fish sampled at 24 hpf (day 1, prior to increase in water temperature), 48 hpf (day 2) and 72 hpf (day 3) (n = 4 pools of 10 fish per treatment). IF fish were analysed at 240 hpf (day 10), 360 hpf (day 15) and 480 hpf (day 20) (n = 14 fish per treatment). All results are presented as mean (± SE). Letters (Pre-IF: uppercase, IF: lowercase) indicate temporal significant differences within temperature regimens. Asterisks indicate significant differences between thermal treatments within sampling time. 28 °C treatment is indicated by green bars and 34 °C in red

Histopathology at 480 hpf (IF)

Hepatocytes from zebrafish maintained at 28 °C displayed normal cellular architecture with occasional small, cytoplasmic lipid vacuoles (Fig. 7a). Fish exposed to 34 °C presented morphological changes in H&E-stained zebrafish liver tissue (Fig. 7b), characterised by pronounced, clear, feathery, cytoplasmic vacuolation. Qualitative analysis revealed that fish exposed to 34 °C exhibited a significantly higher vacuolation grade (moderate and marked) compared to controls, which only showed mild vacuolation (p < 0.05) (Fig. 7c). This observation was supported by quantitative ImageJ analysis, which revealed that the average vacuolated area in the heat-stressed treatment was significantly increased than that observed in the control (p < 0.05) (Fig. 7d). Hepatocytes from control fish maintained at 28 °C showed minimal PAS-positive staining, indicating mild glycogen accumulation within cells exhibiting normal histological architecture (Fig. 8a). In contrast, heat-stressed fish showed markedly increased PAS staining, likely reflecting extensive glycogen deposition associated with hepatocellular swelling and vacuolation (Fig. 8b). The severity of PAS staining was significantly greater in heat-stressed fish, with a higher proportion exhibiting moderate to marked positive staining compared to controls, which were limited to mild staining (Fig. 8c). Quantitative analysis revealed that the percentage of PAS-positive hepatocytes in heat-stressed fish was approximately fivefold higher than in controls (p < 0.05) (Fig. 8d).

Fig. 7.

Histopathological examination of H&E staining in the liver of 480 hpf zebrafish exposed to 28 ºC (a) and 34 ºC (b) (arrows indicate cellular vacuolation). The scale of the images is 50 µm. c. Hepatic cellular vacuolation assessed by grading severity based on representative H&E-stained images. d. Percentage hepatic vacuole area on representative 40 × magnification H&E-stained images determined by ImageJ software. Graph results are presented as mean (+ SE). Asterisks indicate significant differences between treatments (p < 0.05) (n = 8 fish per treatment)

Fig. 8.

Histopathological examination of PAS staining in the liver of 480 hpf zebrafish exposed to 28 ºC (a) and 34 ºC (b) (arrows indicate positive staining shown in pink). The scale of the images is 50 µm. c. Severity of hepatic PAS-positive staining assessed by grading intensity based on representative PAS-stained images. d. Quantification of positive hepatic glycogen staining on representative 40 × magnification PAS-stained images determined by ImageJ software. Graph results are presented as mean (+ SE). Asterisks indicate significant differences between treatments (p < 0.05) (n = 8 fish per treatment)

Discussion

Temperature plays a central role in regulating the cellular, physiological, and developmental processes of ectothermic fish (Afonso 2020). Existing literature suggests that sudden and extreme fluctuations in water temperature can induce acute stress responses, which are required to maintain biological function under such abrupt changes (Álvarez-Quintero et al. 2025). However, comparatively few studies have addressed the subtler, yet potentially more maladaptive effects of prolonged exposure to moderately elevated temperatures, what might be termed the "silent killers" of biological function and development in teleost fishes (Islam et al. 2019; Alfons et al. 2023; Yan et al. 2024). In this study, we investigated the effects of chronic heat exposure that did not result in mortality, but was associated with alterations in physiological processes relevant to tertiary stress responses over time. Elevated temperatures during early development were accompanied by changes at both phenotypic and gene expression levels, suggesting that sustained exposure to temperatures above optimal ranges can represent a biologically relevant stressor during early developmental stages.

This study found no significant effect of elevated temperature on survival or hatching in pre-IF and IF zebrafish. These results align with previous research showing that mortality rates in zebrafish exposed to temperatures between 30–33 °C did not significantly differ from those reared under control conditions (Hallare et al. 2005; de Souza et al. 2025). In this study, we confirmed that exposure to 34 °C appears to represent a thermal regime capable of eliciting physiological responses without causing mass mortality. Hatching, widely used as an endpoint in investigating the effects of environmental stressors, including temperature, on early fish life, is governed by a complex interplay of overlapping biological processes (Muller et al. 2015). No significant differences in hatch rate were observed between fish reared at 28 °C and 34 °C, with nearly 100% hatching in both groups by 72 hpf. At the molecular level, control fish exhibited a significant increase in zebrafish hatching enzyme 1 (ZHE1) expression at 48 hpf, shortly before hatching, followed by a decrease once hatching was completed. Under optimal conditions, ZHE1 expression begins as early as 11.5 hpf, increases by 24 hpf, and diminishes after hatching is complete (Pamanji et al. 2016). In contrast, heat-stressed fish showed a significant reduction in ZHE1 expression at 48 hpf compared to controls, with levels remaining similar to those observed at 24 hpf. This study is among the first to examine ZHE1 expression in response to thermal stress, demonstrating its downregulation under heat stress. Notably, ZHE1 suppression has also been reported following exposure to heavy metals and nanoparticles (Muller et al. 2015). The two primary mechanisms governing hatching are enzymatic and mechanical processes (Rechulicz 2001; Pamanji et al. 2016). Enzymatic hatching involves the release of proteolytic enzymes (such as zinc metalloproteases encoded by the ZHE1 gene) from mature hatching glands, which degrade the chorion and cause the egg envelope to soften and swell (Kawaguchi et al. 2010). Mechanical hatching occurs when increasing embryonic movement, driven by rising metabolic and oxygen demands during development, ruptures the egg envelope in search of more favourable external oxygen conditions (Cowan et al. 2024). Previous research in ide (Leuciscus idus) showed that hatching glands were larger in embryos maintained at optimal temperature compared to elevated temperature (Rechulicz 2001). In zebrafish, embryos exposed to 36.5 ºC had significantly lower hatching enzyme (cathepsin L) activity than embryos reared at 28.5 ºC (Pype et al. 2015). These findings suggest that hatching activity in fish exposed to 34 ºC is driven by both enzymatic and mechanical processes, with mechanical cues perhaps playing a predominant role due to the suppressed expression of ZHE1.

Exposure to unfavourable environmental conditions during early development can directly affect growth, often leading to developmental abnormalities (Takle et al. 2005). Fish reared under control conditions (28 °C) exhibited minimal abnormalities, whereas those exposed to 34 °C showed evidence of malformations as early as 72 h post-fertilisation (hpf). In addition, across all developmental stages examined (72–480 hpf), the number of fish exhibiting at least one type of deformity was approximately twice as high in the 34 °C treatment group compared to controls. Here we demonstrate that heat stress can influence aspects of normal zebrafish larval development. These findings align with previous research documenting heat-induced developmental abnormalities in zebrafish (Valcarce et al. 2024; Hallare et al. 2005; de Souza et al. 2025), including significant increases in edema and tail malformations in embryos exposed to 36.5 °C at 72 hpf (Pype et al. 2015). During the pre-IF stage, common deformities induced by heat stress included yolk sac edema and spinal curvature, which primarily affected the tail and trunk. Yolk sac edema was the most prevalent malformation at 72 and 120 hpf (affecting 22.8% and 8.7% of fish, respectively). Yolk sac edema, characterised by fluid accumulation in the yolk syncytial layer due to disrupted osmotic gradients, may impair nutrient absorption and disrupt the development of organs like the eyes and spine, ultimately reducing growth in zebrafish (Sant & Timme-Laragy 2018). In this study, we also observed a significant increase in spinal curvature at 96 hpf and 120 hpf in heat-stressed fish, whereby the incidence of this deformity was 10 times as high as that observed in fish reared at optimal conditions. In Atlantic salmon embryos, several studies have reported a marked increase in skeletal deformities, particularly spinal curvature, following exposure to elevated temperatures (Wargelius et al. 2005; Takle et al. 2005). Similarly, research on salmonid embryos and fry has shown that elevated temperatures can disrupt bone formation and mineralisation, resulting in abnormal vertebral development and a higher incidence of skeletal malformations (Ytteborg et al. 2010). However, the incidence of deformities in heat-stressed fish decreased between 72 and 120 hpf, likely reflecting the transient nature of some early developmental malformations. Notably, yolk sac edema declined from 72 to 96 hpf, coinciding with the onset of yolk sac resorption (Kondakova and Efremov 2014). As yolk reserves become exhausted and the yolk sac undergoes physiological degeneration, visible edema diminishes over time, despite no significant effects on survival (Halbach et al. 2020).

During the pre-IF stage, no overall differences in growth were observed, except that fish exposed to heat stress exhibited a significantly greater average body length by 72 h post-fertilisation (hpf) compared to controls. Although the impact of elevated temperature on body length was minimal in pre-IF fish, our findings reveal a distinct disruption of normal embryonic and larval development, marked by an increase in morphological abnormalities.

During the IF stage, sustained exposure to 34 °C, 6 °C above the zebrafish's thermal optimum, resulted in a higher incidence of morphological abnormalities, including abnormal skeletal development and growth retardation, observed consistently across all sampling points. This was evidenced by significantly reduced spinal cord length and eye area compared to fish reared at 28 °C. Additionally, key growth parameters such as standard length and fork length were significantly decreased in fish exposed to elevated temperatures between 240- and 480-h post-fertilisation (hpf). These findings align with previous research reporting reduced growth in Thai pangas fingerlings (Pangasianodon hypophthalmus) and catfish (Clarias gariepinus) chronically exposed to temperatures of 36 °C and 37 °C, respectively (Islam et al. 2019; Khieokhajonkhet et al. 2022), as well as reduced body size in 120 hpf zebrafish following 48-h exposure to 35 °C (Álvarez-Quintero et al. 2025). Reduced growth in response to chronically elevated temperature has not been previously reported in pre-IF zebrafish. Ultimately, the changes observed in this study may reflect temperature-driven alterations in developmental timing and locomotive demands (Álvarez-Quintero et al. 2025). In terms of deformities, exposure to multiple acute stressors, including elevated temperature, has been linked to microphthalmia and impaired cranioencephalic development at 168 hpf (Valcarce et al. 2024). However, recent research on the effects of heat stress on growth and development in IF fish remains limited. As these fish aged, the incidence and severity of deformities declined, indicating that developmental processes in pre-IF fish are particularly sensitive to thermal stress and that the morphological impacts of elevated temperature are stage specific. Pre-IF zebrafish appear more susceptible to developmental deformities, whereas IF fish show greater sensitivity to changes in growth, particularly reduced body size. These results underscore the importance of examining the long-term effects of heat stress exposure to accurately assess its cumulative impact on growth and morphological development across multiple early life stages.

The molecular response to thermal stress in zebrafish was influenced by both the temperature treatment and developmental stage, with some genes responding more strongly during pre-IF (growth and development genes), and others during IF (stress genes). The lack of differential heat shock gene expression observed in pre-IF fish suggests the incomplete activation or maturation of heat stress response pathways. Nonetheless, the present study confirms previous observations of heat shock gene activity (Long et al. 2012; de Souza et al. 2025) and further shows that these genes are already expressed at basal levels during this developmental stage. While the underlying mechanisms remain unclear, this constitutive expression is consistent with their established roles in normal zebrafish development, rather than representing an active heat shock response (Blechinger et al. 2002; Krone et al. 2003).

In contrast, in IF fish, the thermal stress response became more apparent. Consistent with previous larval-stage research, heat shock proteins (HSPs) were upregulated in response to elevated temperatures beginning at 240 hpf, signalling the activation of well-characterised molecular stress pathways (de Souza et al. 2025). HSPs are well-established markers of thermal stress across species, with expression patterns influenced by developmental stage, tissue type, and exposure duration; thus, assessing their expression offers valuable insight into thermal stress levels and the overall wellbeing of fish populations (Afonso 2020; Madeira et al. 2014). While HSP70 was not significantly upregulated during heat stress in our study, HSP90AA1 exhibited a clear temperature-dependent response after 240 (10 days) and 480 hpf (20 days) of exposure, suggesting its role in chronic thermal stress response. HSP90 plays a crucial role in the cellular stress response by stabilising and maintaining signalling pathways through its chaperone activity, which supports protein folding, transport, transcription, and interaction with steroid hormone receptors (Krone et al. 2003). Similar to our findings, increased HSP90A expression in zebrafish was reported after 720 h (30 days) of continuous exposure to 30 ºC initiated at the embryonic stage (de Souza et al. 2025). Of particular note were the activity of SERPINH1 genes SERPINH1A (upregulated at 240 and 480 hpf) and SERPINH1B (upregulated at 240–480 hpf), which showed the strongest and most consistent upregulation across nearly all sampling points during the IF phase. SERPINH1, also known as HSP47, is a stress-inducible molecular chaperone located in the endoplasmic reticulum, where it facilitates the proper post-translational folding and stabilisation of fibril-forming collagens (Ignatz et al. 2024; Beemelmanns et al. 2021). In support of our findings, HSP47 was shown to be upregulated in zebrafish larvae subjected to acute exposure at 34 ºC (Long et al. 2012). We aimed to determine which molecular subunit, SERPINH1A or SERPINH1B, more accurately reflects the heat shock response in early-stage zebrafish. SERPINH1B exhibited a more robust and sustained response to heat exposure compared to SERPINH1A, suggesting that SERPINH1B may serve as a more reliable and predictable biomarker for chronic thermal stress in IF zebrafish. The upregulation of SERPINH1A, SERPINH1B, and HSP90AA1 observed in this study suggests that IF fish, unlike pre-IF fish, were able to mount a more robust and targeted cellular stress response under elevated temperature conditions (34 °C). It is well established that the expression of heat shock proteins is positively correlated with thermotolerance, reflecting an increased resistance to heat stress (Ignatz et al. 2024; Farrell 2002; Basu et al. 2002). The enhanced expression of these molecular chaperones likely supports the maintenance of cellular homeostasis by preventing protein misfolding, aggregation and limiting cellular damage (Jeyachandran et al. 2023). Collectively, these processes may contribute to the thermotolerance and adaptive resilience of IF fish when exposed to thermal stress.

The key oxidative stress-related gene PRDX6 exhibited no temperature-dependent changes in expression during the pre-IF fish, although a significant temporal increase was observed in both treatments. This temporal upregulation likely reflects a developmental enhancement of antioxidant capacity, as this gene plays a vital role in counteracting oxidative damage by preventing cell membrane peroxidation (Chang et al. 2016). In contrast, PRDX6 was downregulated during the IF heat-stressed fish compared with controls, aligning with previous observations in salmonids where this gene is suppressed under both thermal and hypoxic stress (Beemelmanns et al. 2021). The downregulation further implies a reduced capacity to manage potential instances of heat-induced oxidative stress. Furthermore, CALM1A expression was significantly upregulated at 24 and 72 h hpf in heat-stressed fish compared to controls (Pre-IF). In response to stress, fish regulate intracellular calcium levels, which, upon binding to calmodulin (CALM) via its EF-hand domains, activate signal transduction pathways that drive gene expression, enzyme activation, and protein synthesis (Cheung 1980). Elevated CALM1A expression at 24 hpf, despite identical thermal conditions between treatments, likely reflects intrinsic developmental variability or slight asynchrony between clutches, consistent with the dynamic regulation of calcium-signalling genes during early zebrafish development (Webb and Miller 2000). The punctual increased expression of CALM1A in pre-IF fish (72 hpf) likely reflects a greater need for calcium-mediated cell signalling to support molecular and protein functions during rapid development. Under heat stress at 34 °C, this upregulation may help maintain or enhance metabolic activity by meeting elevated signalling demands required to preserve cellular homeostasis. However, in IF fish, CALM1A expression was significantly downregulated in heat-stressed groups at 240 and 360 hpf compared to controls (IF). In juvenile Atlantic salmon (Salmo salar), such downregulation of CALM in the liver during chronic thermal stress has been linked to metabolic suppression, allowing fish to redirect energy toward other essential cellular processes (Beemelmanns et al. 2021). These findings suggest that exposure to elevated temperature (34 °C) triggers an early upregulation of CALM1A to enhance cell signalling during development, while prolonged exposure (240 and 360 h) leads to its suppression, likely as a result of metabolic adjustments and energy reallocation to support long-term adaptation to thermal stress.

Chronic exposure to 34 °C impaired the growth and development of IF zebrafish, concurrent with the significant downregulation of key growth-related genes IGF-1, GH, and GHRA. The GH-IGF-1 axis plays a central role in fish growth and development (Horie et al. 2020). However, few studies have investigated the effects of elevated temperature on GH/IGF system gene expression in fish (Shahjahan et al. 2021). In our study, IGF-1 expression was markedly suppressed at 360 and 480 hpf, up to 100-fold lower than in control fish, corresponding with reduced growth (length) and significant downregulation of GH and GHRA at 480 hpf. Similar hepatic downregulation of IGF-1 and reduced growth rates have been reported in rohu (Labeo rohita) exposed to 36 °C for 30 days, and in coho salmon (Oncorhynchus kisutch) subjected to a 2-h heat shock challenge (only IGF1 downregulation) (Shahjahan et al. 2021; Nakano et al. 2015). In zebrafish, hypoxia alone has been shown to suppress IGF signalling activity, and knockdown of the IGF1 receptor (under normal conditions) leads to reduced growth, impaired development, and increased morphological abnormalities (Kamei et al. 2011; Schlueter et al. 2007). These findings, alongside previous research, indicate that chronic heat stress may disrupt the functionality of the GH-IGF-1 axis, with its suppression directly coinciding with diminished growth in IF zebrafish.

Additionally, development genes ACVR2AA (at 240 and 480 hpf) and BMP2B (at 480 hpf) were also downregulated in the IF fish. In zebrafish, activin A receptors (ACVR) interact with bone morphogenetic proteins (BMP) within the transforming growth factor beta (TGFβ) signalling pathway (Allen et al. 2020). While the roles of these genes in response to heat stress have been less explored, disruptions in BMP signalling gradients have been linked to altered skeletal development in zebrafish (Wu et al. 2022). The downregulation of ACVR2AA and BMP2B coincided with a significant increase in deformities observed at 240 hpf and 480 hpf, including growth retardation and spinal curvature. Therefore, our findings suggest that ACVR2AA and BMP2B are critical for normal skeletal development in IF zebrafish, and their suppression under chronic heat stress may induce sublethal phenotypic effects at the organismal level. Interestingly, an upregulation of BMP2B was observed in the heat-stress treatment of pre-IF fish at 24 hpf. Although both groups were maintained under identical thermal conditions at this stage, this difference likely reflects normal developmental regulation, with variation arising from individual differences in the progression of the ventral-to-dorsal BMP activity gradient which is critical for successful morphogenesis during early embryonic development in zebrafish (Ramel and Hill 2013; Wu et al. 2022). Overall, future work should investigate whether gene expression trends observed in this study correspond to protein-level changes, facilitating a more thorough assessment of potential biomarkers.

From a histopathological perspective, this study demonstrates that elevated heat exposure can significantly alter hepatocyte morphology in early-stage zebrafish. Numerous studies have shown that environmental stressors, including temperature, can induce notable changes in teleost liver structure and function (Alfons et al. 2023; Cardoso et al. 2019; Yan et al. 2024). Fish exposed to 34 ºC for 480 h (20 days) exhibited increased hepatocellular vacuolation, likely resulting from significant glycogen accumulation. Hepatocytes of heat stressed fish showed marked cellular enlargement due to cytoplasmic vacuolation associated with strong-positive, granular PAS staining. Increased hepatocyte vacuolation observed under elevated temperature conditions aligns with previous reports in both adult and juvenile fish, which have interpreted this cellular phenotype as a marker of thermal-induced liver damage and an alteration of metabolic pathways (Madeira et al. 2014; Yan et al. 2024). This may indicate that heat stress gives rise to changes in liver function in early-stage zebrafish, although it is unknown whether such cellular alterations are permanent in nature and will persist into adulthood.

The liver is the main driver of glucose metabolism, production and storage, essentially acting as a “glucostat” to maintain blood glucose levels and energy requirements (Li et al. 2022). As a collective, the literature has presented a general trend of a decrease of glycogen content and increase in blood glucose as a result of exposure to increased temperature in fish as glycogen serves as an important energy source under stressful circumstances (Yan et al. 2024; Islam et al. 2019). In contrast, our study showed an increase of hepatic glycogen in response to heat stress, a finding that has not yet been reported in early-stage zebrafish. In a combined stressor study, Cardoso et al. (2019) reported an increase in the percentage of highly stained PAS hepatocytes in juvenile zebrafish exposed to 30 ºC alone. On a similar note, Woo (1990) and Madeira et al. (2014) indicated a significant increase in liver glycogen content and decrease in glycogenolysis activity after chronic exposure to elevated temperature in juvenile red sea bream (Pagrus major). While the exact pathways attributed to the changes reported in this study are not confirmed, an increase in glycogen after 480 h of exposure suggests that these fish are prioritising the maintenance and storage of glycogen reserves over mobilising liver glycogen for metabolic functioning. Dysregulation of glycogen metabolism results in significant pathological changes in the liver (Zhong et al. 2022). In humans, abnormal glycogen accumulation is associated with diseases such as genetic glycogen storage disorders and type 1 diabetes and is often associated with growth retardation (Soon & Torbenson 2023). Excessive glycogen accumulation within hepatocytes reflects impaired carbohydrate metabolism, a condition known as glycogenic hepatopathy (Li et al. 2022). Histologically, this condition is characterised by swollen hepatocytes with abundant clear cytoplasm due to glycogen accumulation, features consistent with the pathological findings observed in this study (Soon & Torbenson 2023). In largemouth bass (Micropterus nigricans), chronic glycogenic hepatopathy resulting from a high-starch diet has been shown to progress to liver fibrosis and cirrhosis (Zhong et al. 2022). To our knowledge, this is the first study to report significant glycogen accumulation in zebrafish liver chronically exposed to elevated temperature. Further research is needed to clarify the specific glycogenic pathways involved in these early-stage changes, assess whether persistent, prolonged exposure leads to liver fibrosis, and determine the permanence of these alterations.

In this study, we demonstrate that zebrafish chronically exposed to elevated water temperatures from 24 h hpf, without intervals of thermal recovery, exhibit the greatest magnitude of response at 240 h (10 days). However, further research should aim to statistically investigate the differences in responses between the two groups to fully elucidate and confirm these observations. The effects observed during the IF stage may be attributable to cumulative exposure over the entire study period, including the pre-IF stage. Collectively, the heat stress response in IF zebrafish was characterised by growth retardation, activation of the molecular heat shock response, increased incidence of deformities, and changes in hepatocyte composition. In their natural habitats, zebrafish may experience prolonged and unpredictable thermal fluctuations, with water temperatures reaching up to 38 °C, and as eurythermal species, they typically inhabit shallow, tropical, and often stagnant waters that are subject to frequent and sometimes abrupt temperature changes (Álvarez-Quintero et al. 2025; Ribas and Piferrer 2014). We speculated that domesticated, laboratory-reared zebrafish raised under consistently stable thermal conditions may have reduced thermotolerance as a result of prolonged generational acclimation. However, previous studies comparing wild and laboratory-reared zebrafish indicate that thermal tolerance is largely maintained during domestication, supporting the use of laboratory-reared zebrafish as a suitable model species for climate change research (Morgan et al. 2019). Our findings suggest that IF zebrafish displayed more pronounced morphological, histopathological, and molecular alterations compared to pre-IF fish, indicating that the IF stage may be more vulnerable to the effects of elevated temperature. This thermal vulnerability may stem from key physiological transitions occurring during this phase, including changes in feeding behaviour, yolk sac depletion, and swim bladder inflation (Yanagitsuru et al. 2021). At optimal developmental temperatures, embryos and newly hatched larvae utilise endogenous energy reserves (e.g. yolk) more efficiently, thereby enhancing the availability of resources to support the transition to independent feeding (Sant & Timme-Laragy 2018). As zebrafish larvae approach metamorphosis (360–720 h post-fertilisation), they undergo critical developmental processes, including scale formation, unique pigmentation of scales, and complete gut differentiation (Ribas and Piferrer 2014). These transitions likely increase their sensitivity to thermal stress, triggering responses observable across multiple levels of biological organisation. Further research is warranted to determine whether exposure to heat stress during early development in zebrafish has persistent effects on growth, reproductive capacity, and survivorship into juvenile and adult stages.

Conclusion

The present findings indicate that chronic exposure to elevated temperatures can modulate stress responses and may have implications for biological processes spanning multiple levels of organisation. Notably, we observed distinct differences between pre-IF and IF fish, with the latter exhibiting developmental inhibition in the following 360 h (15 days) of cumulative heat stress exposure. Importantly, our results highlight the need to assess thermal stress over extended periods, as the most severe and diverse effects only became evident after at least 240 h (10 days) of continuous exposure. This study underscores the significance of individual variation in shaping heat stress responses over time, an area warranting further investigation. Our results indicate that chronic exposure to elevated temperatures can amplify stress responses, with potential consequences for biological processes spanning multiple levels of organisation. Given its consistent expression pattern, we propose SERPINH1B as a potential reliable molecular biomarker of chronic thermal stress in IF zebrafish. Taken together, our research in zebrafish contributes to a broader understanding of how temperature changes may affect ecologically and economically important fish species during early developmental stages. These findings point to the critical role of managing thermal stress responses in preserving the health and viability of fish stocks in the face of environmental change.

Author Contribution

All authors contributed to the study conception and design. Monique Adzijovski: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization. Aaron G. Schultz: Writing – review & editing, Supervision, Investigation, Conceptualization. Andrew P.A Oxley: Writing – review & editing, Supervision. Jemma Bergfeld: Methodology, Supervision, Investigation. Luis O.B. Afonso: Writing – review & editing, Supervision, Investigation, Funding acquisition, Conceptualization.

Funding

Open Access funding enabled and organized by CAUL and its Member Institutions. This study and Monique Adzijovski were supported by a Deakin University Postgraduate Research Scholarship and by Deakin University LES Strategic Growth Funds (PG00232).

Data availability

The data will be made available upon request.

Declarations

Competing interests

The authors declare no interests.