Transcriptomic insights into thermal manipulation and heat stress in skeletal muscles of broiler chickens

Abstract

Heat stress (HS) poses a substantial burden to poultry production, compromising muscle integrity, growth performance, and overall welfare of broiler chickens. Elevated ambient temperatures disrupt energy metabolism, impair muscle function, and activate cellular stress responses. Embryonic thermal manipulation (TM) has emerged as a promising strategy to enhance thermotolerance through developmental reprogramming; however, the molecular basis underlying TM, acute, and chronic HS effects in skeletal muscle remains incompletely understood. In this study, bulk RNA sequencing of the pectoralis major muscle was performed in broilers using a factorial experimental design. Eggs were incubated under thermoneutral (37.8°C, 56% RH) or thermal manipulation (38.5°C, 65% RH, 18 hours per day) conditions, and then the chickens were later subjected to either acute heat stress (35 ± 0.5°C for 12 h on day 22 post-hatch) or chronic heat stress (35 ± 0.5°C for 5 consecutive days from day 18 to 22 post-hatch). Transcriptomic analysis identified distinct molecular profiles associated with TM and HS. Acute HS resulted in pronounced transcriptional changes, including activation of immune signaling, stress-responsive pathways, and suppression of metabolic and energy-related processes, while chronic HS affected subtle transcriptional changes related to immune and inflammatory response. Also, TM induced persistent modulation, altering the muscle response to acute HS, partially preserving metabolism and energy generation with enriched pathways related to cellular organization, ion homeostasis, and muscle maintenance. Integrative functional analyses using over-representation analysis (ORA) and gene set enrichment analysis (GSEA) revealed that TM reshaped the transcriptional landscape by promoting potentially adaptive responses rather than stress-induced disruption. Collectively, our findings demonstrate that embryonic TM impacts long-lasting transcriptional plasticity in skeletal muscle and that acute and chronic HS induced both shared and different expression profiles. This work provides molecular insight into muscle-specific heat adaptation and supports the incorporation of TM-based strategies into sustainable poultry production systems facing warming climates.

Introduction

Skeletal muscle represents the primary edible protein found in livestock meat, and constitutes 45-60% of the adult body mass (Han et al., 2019). In broiler chickens, breast muscle weight is a central trait in production, which has been substantially enhanced through intensive artificial selection (Bailey et al. 2015). Breast muscles currently account for more than 20% of the live body weight in chickens (Petracci et al. 2015). Accordingly, improving muscle growth has become a priority in poultry breeding programs, which continuously seek strategies to optimize broiler development and thermotolerance, especially under rising environmental challenges such as heat stress (HS). HS is one of the most critical constraints in poultry production, impairing growth performance, feed efficiency, metabolic homeostasis, and muscle development due to the limited capacity of chickens to dissipate excess heat (Madkour et al. 2022; Nawaz et al. 2021a).

One promising approach to improve thermotolerance in broilers exposed to heat stress is embryonic thermal manipulation (TM), a strategy that involves modifying incubation temperature and humidity to induce long-term physiological adaptations (Loyau et al. 2015). TM has been reported to enhance broiler chickens’ tissue integrity, oxidative stress response, and immune function, as well as their hatchability, post-hatch body weight, and thermotolerance acquisition, potentially activating tissue-protective mechanisms (Al Amaz and Mishra 2024; Al-Zghoul 2018; Al-Zghoul, Sukker, and Ababneh 2019; Piestun et al. 2008; Yalçin et al. 2008). Therefore, TM has been suggested as an alternative non-genetic solution to improve chicken growth under stressful environmental conditions (de Barros Moreira Filho et al. 2015).

In skeletal muscle, both embryonic TM and HS have been shown to influence complex regulatory networks and post-hatch physiological parameters. HS disrupts muscle metabolic homeostasis, impairing energy metabolism, lipid utilization, mitochondrial activity, and structural integrity of the pectoralis major, as demonstrated by metabolomic and transcriptomic studies (X. Li et al. 2024; Liu et al. 2022; Wu et al. 2024). These alterations contribute to reduced muscle yield and meat quality, highlighting skeletal muscle as a primary target tissue during thermal challenge. Conversely, TM is linked to increased muscle mass and expanded myogenic cell pool in both embryonic and post-hatch growth (Piestun, Yahav, and Halevy 2015). Previous studies suggest that TM modulates skeletal muscle development via changes in growth factor signaling, myogenic marker genes, and mitochondrial activity, with key genes such as TRPV2 (transient receptor potential vanilloid 2), Atrogin-1 (muscle atrophy F-box protein), HSPs (heat shock proteins), and IGFs (Insulin-like growth factors) (Al-Zghoul et al. 2015; Al-Zghoul and El-Bahr 2019; Al-Zghoul and Hannon 2016; Dalab et al. 2021; X. Li et al. 2024). Using microarray analysis, Loyau et al. (2016) showed that embryonic TM induces a stronger effect during subsequent post-hatch heat exposure compared to thermoneutral conditions in skeletal muscles, reshaping metabolic, stress-responsive, and vascular pathways (Loyau et al. 2016). This finding reflects that TM increases developmental plasticity toward future stressors rather than constitutive expression changes. Despite these advances, the responses of broilers’ skeletal muscle to the combined effects of TM and HS remain incompletely understood, particularly how developmental manipulation may reshape muscle transcriptional responses to later acute or chronic thermal challenges. Transcriptomics offers a powerful framework to evaluate gene expression profiles across different thermal conditions and developmental stages.

This study aims to characterize the global transcriptomic landscape in the pectoralis major muscle tissue of broiler chickens subjected to TM, acute, and chronic HS. Our analysis employed a factorial experimental design and an integrative functional enrichment analysis, identifying key genes and biological pathways involved in response to high temperatures.

Materials and methods

Experimental design

Experimental procedures were approved by the Animal Care and Use Committee of the Jordan University of Science and Technology (ACUC number: 16/4/12/265; Date: April 27, 2025). Hatching eggs from 35-week-old commercial ROSS broiler breeders (n = 384) were collected from a hatchery in Irbid, Jordan. The average egg weight was 62.36 ± 3.63 g. Before incubation, eggs were preheated to 25°C and visually checked. Eggs presenting with morphological abnormalities (e.g., misshapen, dirty, cracked, or unusual size) were discarded. The remaining ones were surface-sterilized using 70% alcohol. Eggs were placed with their large ends up on plastic trays and incubated in commercial Type I HS-SF incubators (Masalles Company Inc., Barcelona, Spain). After seven days of incubation, eggs were candled, discarding 20 eggs due to non-fertility or dead embryos.

Until embryonic day 10, the eggs were maintained at standard conditions of 37.8°C and 56% relative humidity. Then, they were randomly divided into two groups with one incubator per group. The groups included a control group (Con-N, n = 182) and a thermally manipulated group (TM-N, n = 182). The Con-N group was kept at 37.8°C and 56% relative humidity. In contrast, during embryonic days 10 to 18, the TM-N group was incubated at 38.5°C for 18 hours and at 65% humidity. The eggs were automatically flipped every hour throughout this time, a practice that has been demonstrated to promote embryonic development (Oliveira et al. 2015). A schematic of the experimental design is illustrated in Fig. 1.

Fig. 1.

| Experimental design of embryonic thermal manipulation and post-hatch heat stress challenges. (A) Fertilized eggs were incubated under either control conditions (37.8°C, 56% relative humidity) or thermal manipulation conditions (TM; 38.5°C, 65% relative humidity) during embryogenesis. After hatching, chicks from each incubation condition were assigned to one of three treatments: thermal neutral conditions (21°C; Con-N and TM-N), acute heat stress (35 ± 0.5°C for 12 h on day 22 post-hatch; Con-AHS and TM-AHS), or chronic heat stress (35 ± 0.5°C from day 18 to day 22 post-hatch; Con-CHS and TM-CHS). At day 22 post-hatch, skeletal muscle samples were collected for bulk RNA-seq analysis. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-AHS, control subjected to acute heat stress; TM-AHS, thermally manipulated subjected to acute heat stress; Con-CHS, control subjected to chronic heat stress; TM-CHS, thermally manipulated subjected to chronic heat stress; h, hours; D, days; RH, relative humidity.

Hatching management and post-hatching rearing

After hatching, chicks were transferred to the Animal House of Jordan University of Science and Technology. They were assigned to cage pens at random. Each cage contained 8 chicks and served as one experimental unit. Consistent management procedures were applied to raise the chickens for 22 days (Aviagen 2009). The temperature was kept at 35°C for the first week, with 23 hours of lighting and 1 hour of darkness. Thus, the temperature gradually decreased. The temperature was kept at 21°C during post-hatch days 19 to 22. Chickens were fed basal diets formulated to meet the standard nutrient requirements recommended by the National Research Council (National Research Council 1994). In particular, we used two feeding phases of corn-soybean meal-based rations: a starter phase, from day 1 to day 10, and a grower phase, from day 11 to day 22. This ensures optimal energy metabolism and crude protein levels of 3,000 and 3,100 kcal/kg and 23% and 20%, respectively. Throughout the experiment, chickens had ad libitum access to food and water.

Acute and chronic heat stress exposure

Chickens were subjected to two independent heat stress (HS) experiments: an acute heat stress (AHS) challenge and a chronic heat stress (CHS) challenge (Fig. 1). In both experiments, birds originated from two primary experimental groups maintained under thermoneutral conditions, namely thermo-manipulated (TM-N) and control (Con-N). Birds from each primary group were randomly allocated to heat stress or thermoneutral subgroups to evaluate the physiological responses to either short-term or prolonged heat exposure.

Acute heat stress (AHS)

The acute HS challenge was initiated on day 22 post-hatch. From each primary group (TM-N and Con-N), 40 chickens were randomly allocated into two subgroups (n = 20 per subgroup), with each subgroup housed in four pens containing five chickens per pen. This resulted in four subgroups in the acute heat stress experiment: Con-AHS and TM-AHS, which were exposed to acute heat stress, and their corresponding thermoneutral controls (Con-N and TM-N). The heat-stressed subgroups (Con-AHS and TM-AHS) were exposed to 35 ± 0.5°C for 12 h on day 22 post-hatch, whereas the thermoneutral subgroups (Con-N and TM-N) were maintained at 21 ± 0.5°C.

Chronic heat stress (CHS)

The chronic HS challenge was initiated on day 18 post-hatch. Similarly, 40 chickens from each primary group (TM-N and Con-N) were randomly divided into two subgroups (n = 20 per subgroup), with four pens per subgroup and five chickens per pen. This resulted in four subgroups in the chronic heat stress experiment: Con-CHS and TM-CHS, which were exposed to chronic heat stress, and their corresponding thermoneutral controls (Con-N and TM-N). The heat-stressed subgroups (Con-CHS and TM-CHS) were exposed to a constant temperature of 35 ± 0.5°C for five consecutive days (day 18-22 post-hatch) to mimic prolonged environmental heat exposure, while the thermoneutral subgroups (Con-N and TM-N) were maintained at 21 ± 0.5°C.

Physiological measurements

Body weight (BW) and body temperature (BT) were measured from 20 randomly selected chickens per subgroup at several time points post-hatch from day 1 to day 22, and the same individuals were not repeatedly measured across sampling days. Body temperature was measured rectally by inserting the thermocouple probe approximately 2-3 cm into the cloaca using a J/K/T thermocouple meter connected to a rat rectal probe (Kent Scientific Corporation, CT, USA).

Sample collection and RNA sequencing

On day 22 post-hatch, we collected samples from all experimental conditions (Con-N, TM-N, Con-AHS, TM-AHS, Con-CHS, and TM-CHS). Six chickens per group were euthanized by cervical dislocation. After euthanasia, pectoralis major muscle tissues were excised, snap-frozen in liquid nitrogen, and placed in TRI Reagent tubes (Zymo Research, CA, USA). The tubes were stored in a CryoCube F570 Series Ultra-Low Temperature Freezer (Eppendorf, Hamburg, Germany) at −80°C till further RNA isolation. Our sampling aimed to study the transcriptomic profile at a specific postnatal developmental stage.

We isolated total RNA content from pectoralis major muscle tissues using the RNeasy Plus Kits (QIAGEN, Hilden) after tissue homogenization with the Bead Ruptor Elite-Bead Mill Homogeniser (OMNI International in Kennesaw, GA, USA). Afterward, we used three methods to assess the quality of the RNA extracts, including agarose gel visualization, Biotek PowerWave XS2 Spectrophotometer (BioTek Instruments, Inc., Winooski, VT, USA), and Qubit 4 Fluorometer (Thermo Fisher Scientific, MA, USA) using the manufacturer's kit (Qubit™ RNA IQ Assay Kit). For library preparation, three biological replicates (individual birds) per subgroup with RNA integrity number (RIN) greater than 8.0 were selected for RNA sequencing. RNA-seq libraries were generated using the TruSeq Stranded Total RNA Library Prep Gold Kit (Illumina, CA, USA), following the manufacturer’s protocol (TruSeq Stranded Total RNA Reference Guide, 1000000040499 v00). Bulk RNA-seq was performed by Macrogen Inc., Korea, using Illumina sequencer, generating 101 bp paired-end reads.

RNA-seq data processing and quantification

We used Trimmomatic to trim raw RNA-seq results, removing low-quality reads and adapter sequences (Bolger, Lohse, and Usadel 2014). We applied a sliding window of four bases to trim reads when the average quality within the window was < 15, and bases with a Phred score < 3 were trimmed as well from both ends. Also, reads shorter than 36 bp were discarded. Moreover, we used Salmon RNA-seq quantification. Salmon uses a quasi-mapping approach, assigning reads to transcripts without full alignment to the reference transcriptome (Love, Soneson, and Patro 2018; Patro et al. 2017). Salmon is a highly efficient tool, producing accurate results while maintaining a reduced computational time. We obtained the reference transcriptome of chicken from Ensembl (Gallus gallus cDNA GRCg7b release 115: https://ftp.ensembl.org/pub/release-115/fasta/gallus_gallus/cdna/). Finally, we used tximport to infer gene-level quantification (Soneson, Love, and Robinson 2015).

DEGs and functional enrichment

Differential gene expression analysis was performed using the DESeq2 package, identifying differentially expressed genes (DEGs) with a log2 fold change ≥ 1 and q-value < 0.05 (Love, Huber, and Anders 2014). All samples were analyzed using one DESeq2 model, and pairwise comparisons between the six groups were performed using the Wald test. For functional enrichment, the analysis was performed using the clusterProfiler package, with adjusted p-values < 0.05 (Wu et al. 2021). We applied two complementary enrichment approaches: overrepresentation analysis (ORA) and gene set enrichment analysis (GSEA). This aimed to investigate the biological mechanisms with both pronounced and subtle expression changes (Fernandes and Husi 2021). We conducted ORA for the DEGs in each pairwise group comparison using the database of Gene Ontology (GO) Biological Processes (BP) terms (Gene Ontology Consortium 2004). We also conducted GSEA using the GO BP database, accounting for low variability genes and addressing the limitations of ORA, which uses an arbitrary cutoff for DEGs, increasing the sensitivity of our analysis (Laukens, Naulaerts, and Berghe 2015). The background set for the functional enrichment analysis was all genes expressed in our muscle expression matrix, including approximately 16,000 genes.

Finally, we used ViSEAGO to investigate semantic relationships within GO BP terms in ORA, revealing relevant functional clusters (Brionne, Juanchich, and Hennequet-Antier 2019). Gene-to-GO annotations for G. gallus were retrieved from the org.Gg.eg.db package (ID: 9031). Enrichment tests were conducted using topGO in all contrasts, enriched biological processes were merged, and semantic similarity matrices were calculated using the Wang method to account for GO topology (algorithm = elim, statistic = fisher, nodeSize = 10, p-value < 0.01). Hierarchical clustering of enriched terms was performed with Ward.D2 aggregation and dynamic tree cutting. ViSEAGO facilitated the identification of biological modules reflecting coordinated regulatory expression programs. The volcano and dot plots were generated using ggplot2 (Wickham 2016).

qRT-PCR validation

We validated RNA-seq results using Real-Time Quantitative Reverse Transcription PCR (qRT-PCR). We tested seven genes, namely COL1A1, HSP90AA1, HSPH1, FKBP5, MSMB, LECT2, and SFRP4. These genes were selected at random from the top-ranking DEGs list. Briefly, cDNA was synthesized from 2 µg of total RNA using SuperScript IV VILO Master Mix (Invitrogen, Thermo Fisher Scientific, DE, USA) and diluted 1:400. qPCR was performed in duplicate using BlasTaq™ 2X qPCR MasterMix (Applied Biological Materials Inc., BC, Canada) on a Rotor-Gene Q MDx 5plex system (Qiagen, Germany) in 20 µL reactions containing 10.0 µL MasterMix, 0.5 µL of each primer (10 µM), 2.0 µL diluted cDNA, and nuclease-free water.

To normalize gene expression changes, we used two reference genes, β-Actin and GAPDH. Reference gene primers were publicly available (Vitorino Carvalho et al. 2020). The two reference genes yielded comparable results; therefore, only β-Actin was reported. We used NCBI and RefSeq annotation (https://www.ncbi.nlm.nih.gov/nucleotide/) for the seven key DEGs sequences. Also, we used Primer-BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) to design primers using the G. gallus reference genome (taxid: 9031). At least one primer spanned an exon-exon junction for each gene tested. Primer sequences are listed in Table 1.

Table 1.

RT-qPCR primers used to validate DEGs across all experimental groups.

Gene symbol	Forward primer sequence	Reverse primer sequence
COL1A1	TGCTGGAGTAGAGGGTCCTA	TCACCATCATCGCCGTTCTT
HSP90AA1	TGCAAACACAGGACCAACCA	TGATCTTGTCCAGAGCATCAGA
HSPH1	AGTGAGGCTGGAACACAGTC	CGCCTTTCTTTTCACTTGTTTTCT
FKBP5	CATCAAGAGACCCGGGAACG	TTGATCACCTGGCCCTTGC
MSMB	AGCGTGTCAATGTCCAATGC	TTAACTGGTGTTGCATAGGCGG
LECT2	CCTGGTGTCCACTGCTTTTG	TAATTGCCGCAGCCGTATCT
SFRP4	CTCAGTCCCAAGTGCCTCTC	GCAGTCTCTGTTCCCATCTCTTA
GAPDH	TCTCTGTTGTTGACCTGACCTG	ATGGCTGTCACCATTGAAGTC
β-Actin	CAGCCAGCCATGGATGATGA	CATACCAACCATCACACCCTGA

Statistical analysis

All statistical analyses were carried out using R (R Core Team 2023). The hatchability percentage was compared between Con-N and TM-N using a two-sample test for equality of proportions. We compared BW and BT measurements between the experimental groups using a two-way ANOVA test and post hoc Tukey’s test. The tests were performed using individual birds as independent observations, since repeated measurements were not conducted on the same birds over time. For RT-qPCR data, relative expression differences (i.e., 2^-ΔΔCt) were compared between each experimental group and Con-N. Statistical testing was conducted using a simple t-test for the five variables that satisfied the normality assumption (COL1A1, TM-N and Con-AHS; FKBP5, Con-AHS; MSMB; TM-N; HSPH1, TM-AHS) and a Wilcoxon test for all other variables, following a Shapiro-Wilk test. Significance was defined as p-value < 0.05.

Results

Experimental design and BW/BT assessment

Broiler chicken eggs were incubated in control conditions (Con-N) at 37.8°°C or in thermally manipulated conditions (TM-N) at 38.5°C for 18 hours during embryonic days 10 to 18. The hatchability of both groups was high with no statistically significant differences, with 98.3% in Con-N and 96.0% in TM-N (two-sample test for equality of proportions, χ² = 0.96, df = 1, p-value = 0.33).

Body weight (BW) and body temperature (BT) were measured periodically across several developmental time points for all groups (Supplementary Figure 1A-B). From day 1 to day 22 post hatch, BW increased progressively in both TM-N and Con-N, with an overall higher BW in TM-N (p-value = 0.0012). No significant group and day interaction was detected, indicating no clear effect of TM per time point (p-value = 0.89). The average BT in Con-N was 40.73 ± 0.42°C and in TM-N was 40.72 ± 0.46°C. On the other hand, chronic HS significantly decreased BW and increased BT of chickens compared to Con-N (Supplementary Figure 2A-B; BW of Con-N vs Con-CHS: p-value = 0.017; BW of Con-N vs TM-CHS: p-value < 0.0001; BT of Con-N vs Con-CHS: p-value < 0.0001; BT of Con-N vs TM-CHS: p-value < 0.0001). Acute HS slightly increased the average BT in Con-AHS and TM-AHS compared to Con-N, but showed no significant difference in BT and BW. On day 22 post-hatch, average BW for all experimental groups was as follows: Con-N = 1207.50 ± 122.21 g, TM-N = 1221.25 ± 141.21 g, Con-AHS = 1191.88 ± 111.06 g, TM-AHS = 1209.38 ± 89.14 g, Con-CHS = 1111.25 ± 84.67 g, and TM-CHS = 1091.88 ± 84.09 g.

Differential gene expression in muscles following TM and HS

Total RNA was extracted from pectoralis major muscle tissues collected from all groups (Con-N, TM-N, Con-AHS, TM-AHS, Con-CHS, and TM-CHS) on day 22 post-hatch. Bulk RNA-seq was used to investigate the molecular mechanisms underlying embryonic TM, acute HS, and chronic HS, resulting in an overall number of 16,657 genes in the muscle transcriptome. RNA-seq generated high-quality data, with an average of 57.55 ± 4.18 million reads per trimmed FASTQ file, exhibiting a uniform expression profile across different samples (Supplementary Figure 3A-B).

TM affected 44 DEGs, with 30 upregulated and 14 downregulated, when compared to thermoneutral conditions (TM-N vs Con-N; Fig. 2A). In the acute HS groups, 89, 9, and 28 DEGs were identified in comparisons of Con-AHS vs Con-N, TM-AHS vs Con-N, and TM-AHS vs Con-AHS, respectively (Fig. 2B, 2D, 2F). Only a small number of DEGs were detected in the Con-AHS vs TM-N and TM-AHS vs TM-N (3 and 8, respectively; Fig. 2C, 2E). Also, chronic HS groups elicited limited transcriptional changes (Fig. 3). Only 1 was determined as DEG in the comparisons of Con-CHS vs Con-N and TM-CHS vs Con-N (Fig. 3B-C), as well as 7 and 48 in the comparisons of TM-CHS vs Con-CHS and TM-CHS vs TM-N, respectively (Fig. 3E, 3D, 3F). A comparison of chronic and acute HS identified 13 DEGs (Supplementary Figure 3C). There was little to no overlap between the DEGs of the different comparisons, with the largest overlap of 11 DEGs between Con-AHS vs Con-N and TM-AHS vs Con-AHS, and the largest number of unique DEGs of 68 detected in Con-AHS vs Con-N (Fig. 4A-B).

Fig. 2.

| Differential gene expression in skeletal muscle following thermal manipulation and acute heat stress. The volcano plots show differentially expressed genes (DEGs) at day 22 post-hatch across pairwise comparisons: (A) TM-N vs Con-N, (B) Con-AHS vs Con-N, (C) Con-AHS vs TM-N, (D) TM-AHS vs Con-N, (E) TM-AHS vs TM-N, and (F) TM-AHS vs Con-AHS. Vertical dashed lines indicate the ±1 log2 fold change (log₂ FC) threshold, and the horizontal dashed line indicates the adjusted p-value cutoff (q-value cutoff < 0.05). Significantly upregulated genes are shown in green, downregulated genes in purple, and non-significant genes in gray, labeling gene symbols of the top ten up- and downregulated, ranked by FC in each comparison. Numbers in the upper corners indicate DEGs counts (left, downregulated; right, upregulated. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-AHS, control subjected to acute heat stress; TM-AHS, thermally manipulated subjected to acute heat stress.

Fig. 3.

| Differential gene expression in skeletal muscle following thermal manipulation and chronic heat stress. The volcano plots show differentially expressed genes (DEGs) at day 22 post-hatch across pairwise comparisons: (A) TM-N vs Con-N, (B) Con-CHS vs Con-N, (C) Con-CHS vs TM-N, (D) TM-CHS vs Con-N, (E) TM-CHS vs TM-N, and (F) TM-CHS vs Con-CHS. Vertical dashed lines indicate the ±1 log2 fold change (log₂ FC) threshold, and the horizontal dashed line indicates the adjusted p-value cutoff (q-value cutoff < 0.05). Significantly upregulated genes are shown in green, downregulated genes in purple, and non-significant genes in gray, labeling gene symbols of the top ten up- and downregulated, ranked by FC in each comparison. Numbers in the upper corners indicate DEGs counts (left, downregulated; right, upregulated. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-CHS, control subjected to chronic heat stress; TM-CHS, thermally manipulated subjected to chronic heat stress.

Fig. 4.

| Overlap of differentially expressed genes across thermal manipulation and heat stress conditions. Venn diagrams showing the overlap of differentially expressed genes (DEGs) among comparisons related to thermal manipulation and acute HS in (A) and chronic HS in (B). Numbers within each region indicate the number of unique or shared DEGs between the corresponding comparisons, which shows minimal overlap between the experimental conditions. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-AHS, control subjected to acute heat stress; TM-AHS, thermally manipulated subjected to acute heat stress; Con-CHS, control subjected to chronic heat stress; TM-CHS, thermally manipulated subjected to chronic heat stress.

Cluster analysis of biological functions using ViSEAGO

To compare functional pathways patterns across experimental groups, significantly overrepresented GO terms of Biological Processes were clustered using ViSEAGO based on Wang’s distance similarity (Brionne et al., 2019). The resulting heatmaps revealed distinct patterns following TM and heat challenges. Interestingly, TM has induced a unique involvement of several pathways, including negative regulation of extrinsic apoptotic signaling and lysosome organization (TM-N vs Con-N; Fig. 5). Also, TM drove a robust activation of neutrophil chemotaxis, inflammation, and immune responses. These results indicated that embryonic TM induced stable changes in muscle tissues that persisted at least 22 days post-hatch.

Fig. 5.

| Functional enrichment analysis of differentially expressed genes using ViSEAGO hierarchical clustering in acute heat stress comparisons. (A) Heatmap showing enriched Gene Ontology (GO) Biological Process (BP) terms across pairwise comparisons related to thermal manipulation (TM) and acute heat stress (AHS). Significantly enriched GO terms derived from DEG lists were clustered using ViSEAGO based on Wang's semantic similarity. The dendrogram shown on the right represents hierarchical clustering of GO terms, highlighting shared and distinct biological processes across experimental conditions. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-AHS, control subjected to acute heat stress; TM-AHS, thermally manipulated subjected to acute heat stress; IC, information content.

In contrast, acute HS was associated with enrichment of Notch and MAPK signaling pathways (Con-AHS vs Con-N; Fig. 5). Comparisons between acute HS and other experimental groups (Con-N, TM-N, or TM-AHS) showed changes in cell-cell adhesion and establishment or maintenance of epithelial cell apical/basal polarity. In addition, comparing TM-AHS vs Con-N showed enrichment of sulfation and non-canonical Wnt signaling pathways. The latter resulted from the downregulation of a Wnt signaling modulator, SFRP4 (log2 fold change = −2.45). Comparing TM-AHS with Con-AHS showed enrichment in chloride transport, heterophilic cell–cell adhesion via plasma membrane adhesion molecules, and establishment or maintenance of epithelial cell apical/basal polarity as well. The comparison of thermally manipulated groups under normal and acute heat-stressed conditions (TM-AHS vs TM-N) was associated with enrichment in calcium ion homeostasis, muscle contraction, MAPK, and Smoothened (Smo) signaling pathway. Smoothened has previously been implicated in skeletal muscle regeneration and repair (Norris et al. 2023). Several pathways related to steroid biosynthesis and cholesterol metabolism were enriched across multiple comparisons (TM-N vs Con-N, Con-AHS vs Con-N, and TM-AHS vs Con-AHS), indicating that muscle response to thermal exposure involves modulations of lipid metabolism.

On the other hand, chronic HS has not resulted in significant functional pathways compared to the thermoneutral condition (Con-CHS vs Con-N; Fig. 6). This may reflect the limited sensitivity of overrepresentation analysis (ORA) to subtle changes or a restrained transcriptional response consistent with muscle tolerance to prolonged heat exposure. When compared to thermally manipulated conditions (TM-CHS vs Con-CHS), pathways related to the regulation of the cell cycle were enriched, including RNA-templated DNA biosynthesis and mitotic segregation. Comparing TM-CHS and TM-N showed a cluster of pathways related to transmembrane ion transport, including metal ion, calcium ion, monoatomic cation, and anion transmembrane transport (Fig. 6).

Fig. 6.

| Functional enrichment analysis of differentially expressed genes using ViSEAGO hierarchical clustering in chronic heat stress comparisons. (A) Heatmap showing enriched Gene Ontology (GO) Biological Process (BP) terms across pairwise comparisons related to thermal manipulation (TM) and chronic heat stress (CHS). Significantly enriched GO terms derived from DEG lists were clustered using ViSEAGO based on Wang's semantic similarity. The dendrogram shown on the right represents hierarchical clustering of GO terms, highlighting shared and distinct biological processes across experimental conditions. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-CHS, control subjected to chronic heat stress; TM-CHS, thermally manipulated subjected to chronic heat stress; IC, information content.

Gene set enrichment analysis of the muscle transcriptome following thermal exposure

To overcome the limitation of ORA, which can miss low variability genes with a predefined differential expression threshold, we used gene set enrichment analysis (GSEA), which utilizes the overall expression matrix to detect changes in ranked gene sets rather than treating DEGs equally and independently. This was key to enabling complementary characterization of some comparisons not revealed by previous analysis. For instance, chronic HS showed subtle changes with an upregulation of immune system processes, chemotaxis, and cytokine-mediated signaling pathway (Con-CHS vs Con-N; Fig. 7C). These pathways were shared with acute HS as well, indicating common mechanisms for short- and long-term HS adaptation (Con-AHS vs Con-N; Fig. 7B). TM also induced an activation of several immune and inflammatory responses and a suppression of acetylcholine and acetate ester metabolism and protein autoprocessing (TM-N vs Con-N; Fig. 7A).

Fig. 7.

| Gene set enrichment analysis (GSEA) of muscle transcriptomes following thermal manipulation and heat stress. The dot plots show significantly enriched Gene Ontology Biological Process (GO BP; adjusted p-value < 0.05) terms identified by Gene Set Enrichment Analysis (GSEA) across selected key pairwise comparisons, including: (A) TM-N vs Con-N, (B) Con-AHS vs Con-N, (C) Con-CHS vs Con-N, (D) TM-AHS vs Con-N, and (E) TM-AHS vs Con-AHS. The x-axis represents the normalized enrichment score (NES), where positive values indicate gene upregulation in the first condition of each comparison and negative values indicate gene downregulation. Abbreviations: Con-N, thermal neutral control; TM-N, thermally manipulated under neutral conditions; Con-AHS, control subjected to acute heat stress; TM-AHS, thermally manipulated subjected to acute heat stress; Con-CHS, control subjected to chronic heat stress; NES, normalized enrichment score.

Despite that acute HS upregulated immune and inflammatory pathways, it downregulated several metabolic and energy pathways, including amino acid and small molecule metabolism, generation of precursor metabolites and energy, and energy derivation by oxidation of organic compounds (Con-AHS vs Con-N; Fig. 7B). Notably, comparing acute HS under thermally manipulated and thermal neutral conditions, GSEA revealed a reversal of metabolic and energy-producing pathways that were upregulated in TM-AHS, including the generation of precursor metabolites and energy, oxidative phosphorylation, and protein biosynthesis (TM-AHS vs Con-AHS; Fig. 7E). However, other pathways related to development, G protein-coupled receptors (GPCR) signaling, and cell adhesion were downregulated (TM-AHS with Con-AHS; Fig. 7E). Compared with the thermoneutral condition, TM subjected to acute HS resulted in upregulation of immune pathways, downregulation of developmental pathways, and negative regulation of nitrogen, nucleobase-containing compound, and RNA metabolic processes (TM-AHS vs Con-N; Fig. 7D).

RT-qPCR experimental validation of transcriptomic data

We used RT-qPCR to validate RNA-seq analysis for several DEGs across experimental groups (Fig. 8). The tested DEGs were COL1A1, HSP90AA1, HSPH1, FKBP5, MSMB, LECT2, and SFRP4. RT-qPCR showed different expression profiles across groups; for instance, relative to Con-N, HSP90AA1 and HSPH1 were upregulated in Con-AHS but downregulated in TM-AHS.

Fig. 8.

| Validation of selected differentially expressed genes using RT-qPCR. (A) Relative expression levels of COL1A1, HSP90AA1, HSPH1, FKBP5, MSMB, LECT2, and SFRP4, measured by RT-qPCR across all experimental groups. Fold changes were calculated using ΔCt values, then expressed relative to the thermal neutral control (Con-N, 2^−ΔΔCt). Data are presented as mean and SEM (standard error of the mean) in error bars (p-value < 0.001 = ***, p-value < 0.01 = **, p-value < 0.05 = *).

Discussion

In this study, we used bulk RNA-seq to characterize the skeletal muscle transcriptomic profiles of embryonic TM and post-hatch acute and chronic HS in broiler chickens. Overall, TM and HS induced distinct transcriptional responses at day 22 post-hatch. Acute HS elicited a stronger transcriptional response than chronic HS, with more DEGs and more enriched biological pathways. TM altered the transcriptional profile to acute HS, particularly in pathways related to metabolism, energy production, and cellular organization. These transcriptomic insights are relevant to broiler production, as HS is associated with reduced growth efficiency and impaired meat quality traits (Nawaz et al. 2021a). Consistent with the RNA-seq results, RT-qPCR revealed similar changes in gene expression (Fig. 8), demonstrating that RNA-seq data were reproducible and of high quality. The transcriptional profiles of TM, acute HS, and chronic HS were distinct with minimal overlap of DEGs (Fig. 4). These limited intersections suggest that each treatment triggered specific molecular pathways rather than a generalized thermal response, and the high number of unique genes in Con-AHS vs Con-N (68 DEGs) indicates that acute HS induces a strong, treatment-specific transcriptional response. In contrast, embryonic TM can modulate muscle growth trajectories, underscoring the importance of understanding how early-life thermal history shapes muscle function under thermal challenge.

TM during embryogenesis affected several biological pathways in skeletal muscle, indicating long-lasting transcriptomic effects and regulating apoptosis, inflammation, immune responses, and lysosomal organization (TM-N vs Con-N; Fig. 5). Such persistent transcriptomic modulation after thermal manipulation aligns with prior studies that TM can exert stable effects on gene regulation and physiological and metabolic parameters post-hatch (Al-Zghoul 2018; Rocha et al. 2021). For instance, TM protocols have been shown to alter the HSP expression levels and cloacal temperatures later in life, indicating developmental reprogramming rather than a short-term stress response (Al-Zghoul 2018). It is noteworthy that similar long-term responses have been reported across different avian species, suggesting that the embryonic thermal environment represents a conserved biological cue influencing developmental plasticity (Nord and Giroud 2020; Tzschentke and Basta 2002; Vitorino Carvalho et al. 2020). Exposure to thermal signals during critical embryonic windows can drive phenotypic plasticity in gene expression in birds, suggesting shared evolutionary mechanisms by which organisms adjust developmental trajectories in response to thermal cues (Andrieux et al. 2025; Fan et al. 2025; Snead et al. 2025; Vitorino Carvalho et al. 2021). Future comparative studies across avian taxa are required to further investigate the evolutionary dynamics of embryonic modulations. TM modifications can be potentially mediated by epigenetic mechanisms that contribute to enhanced post-hatch adaptability and homeostasis (Kisliouk et al. 2024).

The effects of acute and chronic HS on skeletal muscles displayed both shared and distinct biological features. Both HS protocols activated the immune system and cytokine-related signaling (Con-AHS vs Con-N, Con-CHS vs Con-N; Fig. 7B-C), since thermal challenge disrupts homeostasis and increases systemic inflammation, thereby triggering adaptive immune mechanisms to limit cellular damage and maintain tissue integrity under short-term or prolonged environmental stressors (Lara and Rostagno 2013). Compared to acute HS, chronic HS has not induced additional biological pathways in our analysis, which might indicate that prolonged heat exposure may lead to heat tolerance in muscles at the level of the transcriptome. However, the limited transcriptomic response to chronic HS indicates that integrating additional analyses, including proteomic, biochemical, and growth performance measurements, is needed to fully understand the effects of prolonged heat exposure on skeletal muscle. On the other hand, acute HS was distinguished by the downregulation of metabolic processes, including alterations in energy utilization, small molecule and amino acid metabolism (Con-AHS vs Con-N; Fig. 7B), reflecting a shift from metabolic homeostasis due to stress response pathways to preserve cellular resources under acute thermal challenge. Such shifts are well recognized, whereby energy resources are redirected to support stress mitigation rather than growth or biosynthetic functions (Nawaz et al. 2021b). Also, acute HS involved the regulation of MAPK and Notch signalling (Con-AHS vs Con-N; Fig. 5). These signaling pathways are well-characterized in skeletal muscle stress adaptation, cellular remodeling, and maintenance of tissue integrity (Kramer and Goodyear 2007; Vargas-Franco et al., 2022). MAPK signaling plays a central role in the antioxidation response through ROS-dependent activation of ERK, JNK, and p38 cascades, regulating cell survival and stress-responsive mechanisms in heat-challenged tissues (Aryal et al. 2025). Similarly, Notch signaling is heavily implicated in skeletal muscle development and regeneration via regulating cell fate decisions, cell-cell communications, and coordinating repair processes (Vargas‐Franco et al. 2022). Li et al. (2024) investigated the muscle metabolomics of cyclic chronic exposure to HS and reported disturbances of lipid-related metabolites and pathways, including sphingolipid and ATP-binding cassette (ABC) transporters, accompanied by increased serum corticosterone, free fatty acids, cholesterol, and greater abdominal fat deposition (S. Li et al., 2024). However, in our experiment, consecutive chronic HS has not shown such changes, whereas acute HS and TM groups showed a recurrent enrichment of lipid metabolic pathways (TM-N vs Con-N, Con-AHS vs Con-N, and TM-AHS vs Con-AHS), further supporting cholesterol and lipid metabolism alteration as one of the primary affected pathways in skeletal muscles. Together, these results highlight differences between chronic and acute HS, with chronic exposure characterized by restrained transcriptomic adjustments and acute exposure triggering robust transcriptional changes.

A key finding of this study is that TM modified the skeletal muscle response to acute HS. While acute HS in thermoneutral chickens showed a suppression of peptide biosynthesis and energy generation processes (Con-AHS vs Con-N; Fig. 7B), thermally manipulated chickens exposed to acute HS showed an upregulation of these pathways when compared to Con-AHS (TM-AHS vs Con-AHS; Fig. 7E) and no enrichment when compared to Con-N (TM-AHS vs Con-N; Fig. 7D). This pattern resembles the expression profile of thermoneutral conditions, suggesting partial preservation of metabolism and energy production to mediate HS effects. Moreover, functional pathways related to Smoothened signaling and non-canonical Wnt signaling were enriched in TM-AHS chickens (TM-AHS vs TM-N and TM-AHS vs Con-N, respectively; Fig. 5). Smoothened is a master regulator of tissue repair in several tissues, including the central nervous system and liver, and has been a target for therapeutics development (Del Giovane and Ragnini-Wilson 2018; Michelotti et al. 2013; Ruat et al. 2014). In skeletal muscles, Smoothened signaling regulates satellite cell activation and myogenic progression, playing a critical role in muscle maintenance and regeneration following injury and physiological stress (Palla et al. 2022; Philipp et al. 2008). Non-canonical Wnt signaling, which is a β-catenin-independent Wnt signaling, is involved in the regulation of cytoskeletal organization, cell polarity, and skeletal muscle regeneration after injury (Girardi and Grand, 2018; Kamizaki et al. 2021; Qin et al. 2024). Le Grand et al. (2009) showed the involvement of the Wnt non-canonical pathway in satellite cell self-renewal, regulating the regenerative potential of muscle in adult mammals (Le Grand et al. 2009). While these findings provide mechanistic insights, further studies are required to investigate the cross-species roles of Smoothened and non-canonical Wnt signaling in avian skeletal muscle. These TM-AHS responses were not observed in individual treatment of TM or acute HS (i.e., TM-N and Con-AHS), indicating that they specifically resulted from the interaction between embryonic TM and acute HS. TM may not only change energy allocation dynamics but also impact mechanisms related to muscle maintenance and repair after exposure to acute HS.

One limitation of this study is the relatively small number of biological replicates used for RNA-seq, which may reduce the statistical power to detect subtle transcriptional changes. Additionally, transcriptomic analysis for the chronic HS group was conducted at a single time point. Future studies, including larger sample sizes and multiple sampling time points, will help to further validate and expand our understanding of the transcriptomic responses to TM and HS. The present findings reinforce TM as a practical strategy to enhance post-hatch thermal resilience and optimize chicken rearing conditions, particularly in regions most affected by climate change (Leão et al. 2024; Nawab et al. 2018).

CONCLUSION

In conclusion, embryonic TM induced persistent transcriptional modifications in skeletal muscle and modulated the response to post-hatch HS. TM altered muscle sensitivity to acute heat exposure, supporting improved maintenance of metabolic and cellular stability compared with thermoneutral controls. HS elicited a robust immune and inflammatory response, with acute HS inducing more metabolic and energy disturbances compared to chronic HS. This work provides a molecular basis for understanding heat adaptation in broiler skeletal muscle and supports the integration of TM-based incubation strategies into poultry management systems to promote sustainable production under increasing environmental challenges.

LIST OF ABBREVIATIONS

The following abbreviations are used:

AHS: acute heat stress

BP: biological process

BT: body temperature

BW: body weight

CHS: chronic heat stress

Con-AHS: control subjected to acute heat stress group

Con-CHS: control subjected to chronic heat stress group

Con-N: control thermoneutral group

DEG: differentially expressed gene

GSEA: gene set enrichment analysis

GO: Gene Ontology

GPCR: G protein-coupled receptor

GSEA: gene set enrichment analysis

HS: heat stress

HSP: heat shock protein

ORA: over-representation analysis

RT-qPCR: reverse transcription quantitative polymerase chain reaction

RIN: RNA integrity number

RNA-seq: RNA sequencing

ROS: reactive oxygen species

TM: thermal manipulation

TM-AHS: thermally manipulated subjected to acute heat stress group

TM-CHS: thermally manipulated subjected to chronic heat stress group

TM-N: thermally manipulated thermoneutral group

DECLARATION OF AI AND AI-ASSISTED TECHNOLOGIES IN THE WRITING PROCESS

During the preparation of this work, the authors employed Grammarly to enhance readability and ensure adherence to grammatical and spelling conventions. Subsequently, they meticulously reviewed and made requisite modifications to the content.

DATA AVAILABILITY

The datasets generated and/or analyzed are available in the SRA repository under BioProject number PRJNA1404946, accessed via https://www.ncbi.nlm.nih.gov/sra/PRJNA1404946. The supplementary tables can be accessed on GitHub via https://github.com/shadi-shahatit/TMBroilers_Transcriptomics/tree/main/Muscle.

ETHICAL APPROVAL

Approval was obtained from the Animal Care and Use Committee at Jordan University of Science and Technology (JUST) under approval number (16/4/12/265), which follows the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health (NIH) in the U.S. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

AUTHOR CONTRIBUTIONS

Conceptualization, M.B.Z., S.S., S.H.; Data curation, S.S.; Formal analysis, M.B.Z., S.S., S.H.; Funding acquisition, M.B.Z.; Investigation, M.B.Z., S.S., S.H.; Methodology, M.B.Z., S.S., S.H.; Project administration, M.B.Z.; Resources, M.B.Z., S.S., S.H.; Software, S.S.; Supervision, M.B.Z.; Validation, S.S., S.H.; Visualization, S.S.; Writing – original draft, M.B.Z., S.S., S.H.; Writing – review and editing, M.B.Z., S.S., S.H.

CRediT authorship contribution statement

Mohammad Borhan Al-Zghoul: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization, Data curation. Shadi Shahatit: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Seif Hundam: Writing – review & editing, Writing – original draft, Validation, Resources, Methodology, Investigation, Formal analysis, Conceptualization, Data curation, Visualization.

Disclosures

The authors declare that they have no conflicts of interest regarding the publication of this article.