Deciphering the uterine transcriptional mechanisms underlying eggshell strength through multi-stage analysis in laying hens

Abstract

Eggshell strength is a critical economic trait that declines with hen age, yet the molecular mechanisms distinguishing stable genetic determinants from age-responsive pathways remain unclear. This study implemented a multi-stage comparative framework using uterine transcriptomes from Rhode Island Red hens at peak lay (60 weeks) and late lay (90 weeks) to address this question. At each age, hens were stratified into Weak and Strong shell strength groups based on longitudinal records (5 time points for the 90-week cohort), with 5 individuals per group selected for RNA-seq analysis. Uterine transcriptomic analysis revealed a conserved core of 86 differentially expressed genes associated with shell strength at both ages, with 96.5% exhibiting concordant regulation direction and highly correlated fold-changes (r = 0.84). Weighted gene co-expression network analysis identified two pivotal modules: a stable intrinsic module (MElightcyan) at 60 weeks correlated with peak-lay strength, and an aging-amplified module (MEyellow) at 90 weeks whose correlation with strength progressively increased from mid- to late lay (r = 0.57 at 40 weeks to 0.72 at 90 weeks). Functionally, enriched pathways shifted from cellular structure (MElightcyan) to calcium signaling and hormone regulation (MEyellow) with age. Transcriptional network analysis identified 8 conserved transcription factors (including SATB1 and RXRG) orchestrating this core program. Integrative analysis prioritizing differential expression, longitudinal phenotypic correlation, and QTL mapping highlighted high-confidence candidate genes, including CNTNAP5 (peak-lay) and SLCO1C1 (late-lay). We propose a two-tiered regulatory model wherein a stable core genetic program interacts with dynamic, age-adapted effector networks to determine shell strength. This model provides a dual-strategy framework, distinguishing targets for genetic selection (core program) from pathways for precision management (age-adapted networks) to mitigate age-related decline in shell quality.

Introduction

Eggshell breaking strength is a paramount economic trait in commercial egg production, directly determining losses from breakage during handling and transport, while also influencing shelf life and hatchability (Nys and Le Roy, 2018). Beyond its economic impact, shell strength serves as a key indicator of hen health, reflecting the functional integrity of calcium metabolism and uterine biomineralization (Fu et al., 2024). The avian eggshell is a sophisticated bio-ceramic formed in the uterus (shell gland), where the precise deposition of calcium carbonate within an organic matrix is governed by a tightly regulated interplay of ion transport, protein secretion, and cellular activity (Gautron et al., 2021). This process is susceptible to a multitude of factors, with the age of the laying hen being one of the most significant (Feng, et al., 2020). A well-documented decline in shell strength with advancing age poses a major challenge to flock sustainability and profitability (Liu, et al., 2025; Wu, et al., 2022), yet the specific molecular alterations within the aging uterus that drive this decline remain inadequately defined.

Transcriptomic approaches have begun to unravel the genetic basis of eggshell formation. Studies have identified uterine genes associated with eggshell calcification, matrix assembly (e.g., SLIT3 (Gao, et al., 2022), KRT14 (Wu, et al., 2022)), and ion transport (e.g., CACNB2, CACNA1C, ATP2B1) (Zhang, et al., 2025) under various conditions. These valuable studies have established important candidate gene lists. However, interpreting the biological significance of such differential expression can be challenging when analyses are confined to a single comparison, for instance, peak lay with end-of-lay (Feng, et al., 2020; Fu, et al., 2024; Zhang, et al., 2022). A key, yet less explored, question is whether observed transcriptional differences represent stable genetic determinants that persist across the laying cycle, or transient state-dependent responses specific to a particular physiological context, such as advanced age (Han, et al., 2025; Zhang, et al., 2019). Resolving this ambiguity is essential to prioritize targets for genetic selection, which requires stable mechanisms, versus those for management intervention, which may address malleable, age-sensitive pathways.

To address this question, a research design that enables the direct comparison of uterine transcriptomes across multiple, strategically chosen laying stages is needed. Such a comparative multi-stage framework can help disentangle the core regulatory program from age-adaptive changes. The Rhode Island Red breed, a major global layer, provides an ideal model for such a study.

Therefore, the present study was designed to implement this framework. We integrated detailed longitudinal measurements of eggshell strength (five time points from 40 to 90 weeks) with uterine RNA-Seq analysis in Rhode Island Red hens at two critical time points: 60 weeks (peak lay) and 90 weeks (late lay). Our objectives were threefold: (1) to identify both conserved and age-specific differentially expressed genes (DEGs) associated with shell strength; (2) to construct gene co-expression networks and correlate their activity with longitudinal phenotypic performance; and (3) to prioritize high-confidence candidate genes and regulators through integrative bioinformatics analysis. We hypothesized that the decline in shell strength involves a dynamic reorganization of the uterine transcriptome, where a set of conserved core regulators interacts with stage-specific networks, particularly those related to calcium homeostasis and systemic signaling in late lay. This study aims to provide a resolved molecular model of eggshell strength regulation, offering insights for distinct strategies in genetic improvement and precision management to sustain shell quality.

Materials and methods

Ethical statement

This study complied with institutional and national regulations for animal welfare. The experimental design and procedures were approved by the Animal Care and Use Committee of Hebei University of Engineering (Approval No. 2019-017), and animals were handled with care to ensure minimal pain or distress.

Experimental animals, phenotyping, and sample selection

All experimental hens used in this study were obtained from the Rhode Island Red pure line maintained at Beijing Zhongnong Bangyang Poultry Breeding Co., Ltd. Chickens were housed in a closed, environmentally controlled facility with four-tier individual cages and provided ad libitum access to feed and water throughout the experimental period. Egg production traits, including age at first egg and daily laying records, were continuously recorded. Eggshell strength was measured at 40, 60, 70, 80, and 90 weeks of age using a multifunctional egg quality analyzer (DET-6500, NABEL Co., Ltd., Japan), with one egg collected per hen at each time point and measured within 24 h after laying.

For phenotypic stratification, hens were ranked according to eggshell breaking strength using longitudinal records. In the 60-week cohort, a total of 537 hens were evaluated based on shell strength measured at 40 and 60 weeks of age. Individuals were divided into a weak-shell group (W, lower 50%) and a strong-shell group (S, upper 50%), and five hens from each group were randomly selected for downstream analyses. Importantly, the selected individuals showed no significant differences in age at first egg or cumulative egg production at 60 weeks, ensuring that shell strength differences were independent of laying performance. For the 90-week cohort, 389 hens were evaluated using eggshell strength data from five time points (40, 60, 70, 80, and 90 weeks). Hens consistently ranking in the upper or lower extremes across all five time points were selected, and five individuals per group were used for further analyses. Consistent with the 60-week cohort, selected hens showed no differences in age at first egg or cumulative egg production at 90 weeks. This selection strategy ensured that the experimental groups represented stable and intrinsic eggshell strength phenotypes rather than transient or production-related variation, partially compensating for the moderate sample size (n = 5 per group) by minimizing within-group phenotypic variability.

Tissue collection and RNA preparation

To minimize physiological variation, all hens were sampled at the same reproductive stage. At 60 or 90 weeks of age, birds were euthanized approximately 7 h after oviposition, corresponding to the early mineralization stage of eggshell formation in the uterus. The uterine tissue was rapidly dissected, rinsed with ice-cold phosphate-buffered saline, snap-frozen, and stored at −80°C until RNA extraction.

Total RNA was extracted using standard protocols, and RNA quality was assessed using an Agilent 2100 Bioanalyzer. Only samples with RNA integrity number (RIN) ≥ 8.0 were used for library construction. Poly(A)+ mRNA was enriched using oligo(dT) magnetic beads, fragmented, and reverse-transcribed into cDNA using random hexamer primers. After second-strand synthesis, end repair, A-tailing, adapter ligation, and PCR amplification were performed. Library quality was assessed using Qubit 2.0 and Agilent 2100, and libraries with appropriate insert sizes and concentrations were pooled for sequencing. Paired-end sequencing (150 bp) was conducted on an Illumina HiSeq X Ten platform.

RNA-seq data processing and differential expression analysis

Raw sequencing reads were initially evaluated using FastQC (v0.11.8) (Andrews, 2010). Adapter sequences, low-quality reads, and reads containing excessive ambiguous bases were removed using fastp (v0.20.0) (Chen, et al., 2018). Clean reads were aligned to the chicken reference genome (GRCg7b, Ensembl release 112) using STAR (v2.7.1) (Dobin, et al., 2013). Gene-level read counts were obtained with featureCounts (v2.0.0) (Liao, et al., 2014). Genes with extremely low expression (average count per million < 1 in all samples, corresponding to raw read count < 10 across all samples) were excluded from downstream analyses. Downstream expression analyses were performed using the DESeq2 package (v1.38.3) (Love, et al., 2014). DEGs between weak- and strong-shell groups were identified separately using the criteria |log₂FC| > 1 and adjusted P-value (Benjamini-Hochberg correction) < 0.05.

Functional enrichment analysis

Functional enrichment analyses were conducted using the clusterProfiler package (v4.6.2) (Wu, et al., 2021), which utilizes the latest KEGG annotations for Gallus gallus (organism code: gga) downloaded from the KEGG database. Gene symbols were mapped to KEGG orthology identifiers using the bitr function with the 'org.Gg.eg.db' annotation package. KEGG pathway enrichment analysis was performed with P < 0.05 were considered significantly enriched.

Weighted gene Co-expression network analysis

Weighted gene co-expression network analysis (WGCNA) was performed with all genes using the WGCNA package (v1.72) (Langfelder and Horvath, 2008). The analysis was applied to all expressed genes (after filtering low-expression genes as described above), not only DEGs, to capture the full co-expression architecture. Genes were further filtered to retain those with a coefficient of variation > 0.4 and detectable expression (counts per million > 1) in at least 50% of samples (revised from the original 30% to a more stringent threshold). A soft-thresholding power of 12 (60-week cohort) and 14 (90-week cohort) was selected using the pickSoftThreshold function to achieve scale-free topology (R² > 0.8). Co-expression modules were constructed using the blockwiseModules function with a minimum module size of 30, merge cut height of 0.25, and deepSplit parameter of 2.

Correlation analysis between gene expression and eggshell strength

To identify genes directly associated with eggshell strength and to distinguish stable genetic determinants from age-responsive factors, Pearson correlation analysis was performed between gene expression levels (measured at the terminal sampling point) and shell strength measurements at multiple earlier time points. This approach is based on the rationale that if a gene represents a stable genetic determinant, its expression level (as a proxy for genetic potential) should correlate with phenotypic performance throughout the hen's lifetime, including at ages before transcriptome measurement. Conversely, genes that correlate only with terminal phenotype are likely age-responsive factors. For the 60-week cohort, correlations were assessed using shell strength at 40 and 60 weeks, whereas for the 90-week cohort, correlations were evaluated across five time points (40–90 weeks). Genes showing significant correlations (P < 0.05) were retained. Genes consistently correlated with shell strength across all available time points within each cohort were considered core candidate genes representing stable determinants.

Integration of QTL information for prioritization of candidate genes

To prioritize functionally relevant genes associated with eggshell strength, differentially expressed genes (DEGs) identified from the uterine transcriptomes were integrated with known quantitative trait loci (QTLs) related to egg production and eggshell traits. QTL information was retrieved from the Chicken QTL Database (Animal QTLdb; https://www.animalgenome.org/cgi-bin/QTLdb/GG/index), including loci associated with eggshell strength, eggshell thickness, egg production, and egg quality–related phenotypes.

For both the 60-week and 90-week cohorts, genomic coordinates of DEGs were intersected with reported QTL regions to identify candidate genes located within trait-associated loci. DEGs overlapping with QTLs were considered as high-confidence candidates potentially involved in the genetic regulation of eggshell strength. Subsequent analyses focused on comparing expression patterns of these QTL-overlapping genes between weak- and strong-shell groups, and evaluating their correlation with longitudinal eggshell strength measurements across different ages. This integrative strategy allowed the prioritization of genes supported simultaneously by transcriptomic evidence, phenotypic association, and positional information from quantitative trait mapping.

Transcription factor identification and regulatory network construction

To investigate the transcriptional regulatory mechanisms underlying eggshell strength variation, transcription factors (TFs) were obtained from the AnimalTFDB v4.0 database. Differentially expressed TFs were identified from the uterine transcriptomes at 60 and 90 weeks of age and were further analyzed for their regulatory associations with candidate genes.

Candidate gene sets used for network construction included: (i) all DEGs identified at each age, (ii) genes from phenotype-associated WGCNA modules (MElightcyan: 102 genes; MEyellow: 760 genes), (iii) the 17 core genes consistently correlated with eggshell strength, and (iv) QTL-overlapping DEGs (9 at 60 weeks; 19 at 90 weeks). Pearson correlation analysis was performed between TF expression levels and candidate genes within each age group. Only TF–gene pairs with strong correlations (|r| > 0.9) and statistical significance (P < 0.05) were retained for network construction, ensuring a high-confidence core regulatory network.

Age-specific transcriptional regulatory networks were built separately for the 60-week and 90-week cohorts. Conserved TF–target relationships were subsequently identified by intersecting the two networks, allowing the extraction of a core regulatory module shared across ages.

Statistical analysis

All statistical analyses were performed using R (v4.2.1). For comparisons of eggshell strength at individual time points, two-tailed Student's t-tests were used. For longitudinal analysis of eggshell strength trajectories, a mixed-effects model with repeated measures was applied, with hen as random effect and age, group, and their interaction as fixed effects. Post-hoc comparisons were performed using Tukey's HSD test.

Results

Phenotypic stratification identifies hens with stable, intrinsic eggshell strength differences

To elucidate factors contributing to eggshell strength, we categorized hens at 60 and 90 weeks of age (wks) into distinct phenotypic groups based on their measured shell strength. The total population at each age was first ranked by eggshell breaking strength and divided into a Weak shell strength group (W, lower 50%) and a Strong shell strength group (S, upper 50%). For subsequent transcriptome sequencing, five individuals were randomly selected from each of these pre-defined W and S groups at both ages. Analysis of fundamental production traits between these selected subgroups revealed no statistically significant differences in total egg number laid (Fig. S1A, B) or age at first egg (Fig. S1C, D) at either 60 or 90 wks. This confirms that the observed divergence in shell strength is not attributable to gross differences in lay persistency or sexual maturity.

The historical eggshell strength records of these selected individuals clearly validated the effectiveness of our terminal-age phenotypic stratification. As expected, hens assigned to the W group based on their strength at 60 wks consistently exhibited significantly lower eggshell strength compared to their S group counterparts, not only at the terminal sampling point (60 wks, Fig. 1A) but also retrospectively at 40 wks (Fig. S1E). Importantly, the magnitude of the strength difference between the W and S groups remained consistently large and statistically significant across these time points, with no observable trend of the gap widening with age within this 60-week cohort. A parallel pattern was confirmed in the 90-week cohort. Hens classified into the W group at 90 wks demonstrated substantially and significantly lower eggshell strength than the S group at every recorded historical time point-40 wks, 60 wks, 70 wks, 80 wks, and 90 wks (Supplementary Figure S1F-I). Similarly, the difference between groups was consistently large and stable over time, without evidence of progressive divergence from 40 to 90 wks.

Fig. 1.

Phenotypic grouping based on eggshell strength.

(A) Eggshell breaking strength (kg/cm²) of selected hens from the 60-week cohort at 60 weeks (n = 5 per group).

(B) Eggshell breaking strength of selected hens from the 90-week cohort at 90 weeks (n = 5 per group). Individual data points, group mean (horizontal bar), and SEM (error bar) are shown. W: Weak shell strength group; S: Strong shell strength group.

Two-tailed Student's t-test was used for comparisons between W and S groups at individual time points, with P-values adjusted for multiple comparisons using the Benjamini-Hochberg method.

Notably, while the relative strength gap between W and S groups remained constant, the absolute eggshell strength for both groups exhibited a clear declining trend with advancing age. This age-associated deterioration culminated in the mean shell strength of the W group at 90 wks (Fig. 1B) falling below the critical threshold of 2 kg/cm². Collectively, these data demonstrate that the phenotypic classification based on terminal shell strength identifies hens with a stable, intrinsic propensity for either strong or weak eggshells throughout their laying cycle.

Transcriptomic profiling identifies a conserved core gene set associated with shell strength

RNA sequencing generated an average of 48.7 ± 3.2 million raw reads per sample with an average of 46.2 ± 2.9 million clean reads per sample. Clean reads were aligned to GRCg7b with an average mapping rate of 91.3%. A total of 16,847 genes were detected as expressed across all samples. Principal component analysis confirmed clear transcriptomic separation within each age cohort (Supplementary Figure S2).

Differential expression analysis between W and S groups identified 303 DEGs in the 60-week cohort (125 upregulated, 178 downregulated in W vs. S) and 649 DEGs in the 90-week cohort (368 upregulated, 281 downregulated in W vs. S) (Supplementary Table S1). A high degree of conservation was observed: 86 DEGs were common to both age groups (Fig. 2A). The expression fold-changes of these 86 common DEGs were highly correlated between the two ages (Pearson's r = 0.84, p < 2.2e-16; Fig. 2B), and 83 out of 86 genes (96.5%) exhibited concordant regulation direction (up or down in W relative to S) at both time points (Fig. 2C).

Fig. 2.

Identification and conservation analysis of uterine differentially expressed genes (DEGs) associated with eggshell strength.

(A) Venn diagram illustrating the overlap of upregulated (Up) and downregulated (Down) DEGs identified in the 60-week and 90-week comparisons. The 86 genes in the central overlap represent common DEGs.

(B) Correlation of log2 fold-change values for the 86 common DEGs between the 60-week and 90-week comparisons. Each point represents one gene. The Pearson correlation coefficient (r) and associated p-value are shown. The solid line indicates the linear regression fit.

(C) Expression trend analysis of the 86 common DEGs. Genes are categorized based on their consistent upregulation (Up/Up, n = 38, Red), consistent downregulation (Down/Down, n = 45, Green), or discordant regulation (Up/Down or Down/Up, n = 3) in the W group relative to the S group across both ages. The three genes with discordant patterns are highlighted in red.

Functional enrichment analysis of these 86 core DEGs revealed pathways consistently associated with shell strength across both ages, including "Cytoskeleton in muscle cells," "Neuroactive ligand-receptor interaction," and "ECM-receptor interaction" (Supplementary Figure S3, Supplementary Table S2). This suggests that baseline uterine structural integrity and responsiveness to local signals are fundamental components of the conserved shell strength program.

Weighted gene co-expression network analysis identifies intrinsic and aging-amplified modules

To delineate coordinated gene networks governing eggshell strength, we employed Weighted Gene Co-expression Network Analysis (WGCNA) with all genes to correlate uterine transcriptional patterns with longitudinal phenotypic performance.

In the 60-week cohort, the MElightcyan module was identified as the most significant (Fig. 3A). It exhibited strong and consistent positive correlations with shell strength at both 40 and 60 weeks of age (Correlation = 0.68, P < 0.001 and 0.67, P < 0.001, respectively). This stability indicates that MElightcyan represents a core intrinsic network whose expression reflects a hen's inherent capacity for strong eggshells during peak lay.

Fig. 3.

Identification of intrinsic and aging-amplified gene co-expression networks correlated with longitudinal eggshell strength.

(A, B) Module-trait relationship heatmaps from WGCNA for the (A) 60-week and (B) 90-week cohorts. Each row represents a gene module (e.g., MElightcyan, MEyellow). Each column represents eggshell strength measured at a specific week of age. Cell values show the Pearson correlation coefficient (top) and its statistical significance in parentheses (bottom). Red/blue color intensity indicates the strength of positive/negative correlation.

(C, D) KEGG pathway enrichment analysis of the key modules identified in (A) and (B).

(C) Enriched pathways for the MElightcyan module (60-week cohort), representing a core intrinsic network for shell strength.

(D) Enriched pathways for the MEyellow module (90-week cohort), representing an aging-amplified network whose influence on shell strength intensifies with age.

A strikingly different pattern emerged in the 90-week cohort (Fig. 3B). The MEyellow module demonstrated significant positive correlations with shell strength at every measured time point from 40 to 90 weeks (r = 0.57, 0.60, 0.54, 0.67, and 0.72 at 40, 60, 70, 80, and 90 wks, respectively; all P < 0.01). Notably, the correlation strength increased progressively, culminating in the strongest association at 90 weeks (r = 0.72). This reveals MEyellow as an aging-amplified network whose relevance to shell strength becomes increasingly dominant with advancing age.

Functional enrichment analysis revealed distinct biological roles for these modules. The intrinsic MElightcyan module was enriched for pathways related to cellular structure and phagosome (Fig. 3C). In contrast, the aging-amplified MEyellow module was enriched for "Calcium signaling pathway," "Hormone signaling," and "Neuroactive ligand-receptor interaction" (Fig. 3D)—pathways critically relevant to biomineralization and systemic regulation. This functional shift from structural integrity to calcium and endocrine signaling underscores the changing molecular priorities as hens age.

Integrative correlation analysis identifies 17 core genes with time-invariant association to shell strength

To directly link gene expression with phenotypic performance and identify genes whose association with shell strength is time-invariant—indicating their role as stable genetic determinants rather than age-responsive factors—we performed an integrative correlation analysis. For the 60-week cohort, we correlated the expression of DEG with shell strength of the corresponding hen at both 40 and 60 weeks. For the 90-week cohort, we correlated DEG expression with strength at all five time points (40-90 weeks).

The intersection of genes that were significantly correlated with shell strength at all available time points within each cohort yielded a highly stringent set of 17 core genes (Fig. 4C). This list includes genes such as ADRA2A, P2RX7, SATB1, and TCF15. All 17 genes were confirmed to be significantly differentially expressed between W and S groups at both 60 and 90 weeks (Fig. 4D, E).

Fig. 4.

Identification of candidate genes through expression-trait correlation analysis.

(A, B) Summary of correlation analysis between DEGs and longitudinal eggshell strength. Bars indicate the number of DEGs whose expression level showed a significant correlation (P < 0.05) with the shell strength of the same hen at the indicated week of age. Analysis was performed separately for DEGs from the (A) 60-week and (B) 90-week comparisons.

(C) Venn diagram identifying the stringent intersection of candidate genes. The sets are: 1) genes significantly correlated with strength at both 40 W and 60W from (A), and 2) genes significantly correlated with strength at all five time points (40 W, 60 W, 70 W, 80 W, 90 W) from (B). The intersection yields 17 core genes consistently linked to the trait.

(D, E) Bar plots show their log2 fold-change (Weak/Strong) in the uterine transcriptome at (D) 60 weeks and (E) 90 weeks.

To assess the broader relevance of the 86 common DEGs identified earlier, we correlated all of them with eggshell strength. Strikingly, 66 of the 86 common DEGs (76.7%) showed a significant correlation with shell strength (Supplementary Table S3). This high percentage strongly validates the biological relevance of the conserved transcriptional signature. Notably, among the 17 core genes, seven (including SATB1, P2RX7, ADRA2A) belonged to the subset downregulated in the strong shell group, while three (including TCF15) were upregulated (Supplementary Table S3). These genes represent the most promising candidate regulators, as their expression is simultaneously differential between phenotype groups, conserved across ages, and quantitatively correlated with the trait across an individual's entire laying history.

Integration with QTL databases prioritizes functional candidate genes

To prioritize functionally relevant candidate genes, we intersected our DEG lists with known quantitative trait loci (QTL) from the Chicken QTL Database (Animal QTLdb) associated with egg production and eggshell quality traits.

In the 60-week cohort, nine uterine DEGs were located within documented QTL regions (Fig. 5A). Notably, CNTNAP5 was found within a QTL directly associated with eggshell strength. In the 90-week cohort, 19 QTL-overlapping DEGs were identified (Fig. 5B), with SLCO1C1 located within a QTL for eggshell thickness.

Fig. 5.

Prioritization of candidate genes through integration with eggshell-related QTLs.

(A, B) Heatmaps displaying the differential expression status and QTL overlap of key candidate genes for the (A) 60-week and (B) 90-week cohorts. Each column represents a differentially expressed gene (DEG) that is located within a known chicken QTL region related to egg production or eggshell quality.

(C) Validation of 9 genes (identified in A). The panel shows, from left to right: 1) its log2FC (W/S) at 60 weeks; 2) the Pearson correlation coefficient (r-40 W) and its P-value (P-40 W) between its expression and shell strength at 40 weeks; 3) the correlation coefficient (r-60 W) and P-value (P-60 W) for strength at 60 weeks.

(D) Validation of 19 genes (identified in B). The panel shows, from left to right: 1) its log₂FC (W/S) at 90 weeks; followed by 2)-6) the Pearson correlation coefficients (r-40 W to r-90 W) and their corresponding P-values (P-40 W to P-90 W) between its expression and shell strength at each week from 40 to 90 weeks.

Validation of expression patterns confirmed that CNTNAP5 expression was significantly lower in the Strong group at 60 weeks and showed significant correlation with shell strength at both 40 and 60 weeks (Fig. 5C). Similarly, SLCO1C1 was consistently upregulated in the Strong group across all time points in the 90-week cohort, and its expression was significantly positively correlated with shell strength from 40 to 90 weeks (Fig. 5D). The co-localization of CNTNAP5 and SLCO1C1 within eggshell-specific QTLs, coupled with their robust differential expression and strong phenotypic correlation, provides compelling multi-evidence support for their roles as key regulators of eggshell strength-CNTNAP5 during peak lay, and SLCO1C1 in maintaining shell integrity during aging.

Transcriptional regulatory network analysis reveals 8 conserved core transcription factors

To elucidate transcriptional regulatory mechanisms, we focused on transcription factors (TFs) that were themselves differentially expressed. In both cohorts, a remarkably high proportion of all DEGs (90.76% and 98.15%, respectively) showed significant expression correlation (P < 0.05) with at least one differentially expressed TF (Fig. 6A, Supplementary Table S4), indicating a pervasive and central role for TF-mediated regulation.

Fig. 6.

Analysis of transcription factor (TF) regulatory networks associated with eggshell strength.

(A) Bar plot showing the number of DEGs and the subset of DEGs whose expression is significantly correlated (P < 0.05) with at least one differentially expressed TF (sigCorDEG_TF) in the 60-week and 90-week cohorts.

(B) Regulatory network for the 90-week cohort. The network visualizes significant TF-DEG pairs that also have |r| > 0.9 and P < 0.05. Nodes represent genes (TFs and candidate DEGs), and edges represent significant expression correlations. Edge thickness is proportional to the absolute value of the correlation coefficient (|r|). (C) Heatmap of expression correlations between TFs and candidate DEGs in the 60-week cohort. The panel shows only the subset of these pairs with a strong absolute correlation coefficient (|r| > 0.9). Rows represent candidate DEGs, and columns represent differentially expressed TFs.

(D) Conserved core regulatory module. The network shows TF-DEG interactions that are preserved in both the 60-week and 90-week stringent (|r| > 0.9) networks. It includes 8 transcription factors that are differentially expressed and maintain strong correlations with specific target DEGs across both ages.

Note: The candidate DEG sets for network construction included genes from the 86 common DEGs, key WGCNA modules, the 17 core genes, and QTL-overlapping genes. Only differentially expressed TFs were considered for correlation analysis.

We constructed age-specific TF-DEG regulatory networks by correlating the expression of all DEGs identified at each age (303 at 60 weeks; 649 at 90 weeks) with all differentially expressed TFs. This comprehensive approach ensures that the full regulatory landscape is captured. Retaining only TF-DEG pairs with strong correlation (|r| > 0.9, P < 0.05) yielded high-confidence networks for each age 502 significant pairs for the 60-week cohort (Fig. 6C) and 1758 pairs for the 90-week cohort (visualized as a network in Fig. 6B, Supplementary Table S5).

Strikingly, the intersection of the stringent networks from both ages revealed a conserved core regulatory module comprising 8 transcription factors (Fig. 6D) (BHLHE41, TCF15, SATB1, DLX2, POU1F1, RXRG, NR1D1, and NR1D2) that were themselves differentially expressed and formed strong (|r| > 0.9) regulatory correlations with target DEGs in both cohorts. The preservation of this specific TF cohort across ages highlights a stable, core transcriptional regulatory program consistently associated with eggshell strength, irrespective of hen age.

Based on our integrative multi-stage analysis, we propose a two-tiered regulatory model of eggshell strength. The model comprises:

Layer 1: Stable Core Genetic Program – Represented by the 86 common DEGs and 8 conserved transcription factors (e.g., SATB1, RXRG). This layer establishes the intrinsic, genetically determined potential for eggshell strength and operates consistently across the laying cycle.

Layer 2: Age-Adapted Effector Networks – Represented by context-responsive modules such as MElightcyan (dominant at peak lay, enriched for structural pathways) and MEyellow (amplified with age, enriched for calcium and hormone signaling). These networks modulate the output of the core program in response to physiological context and age-related demands.

The interaction between these two layers determines the final eggshell strength phenotype. This model provides a framework for understanding both the heritable basis (Layer 1) and the age-associated decline (Layer 2 dysfunction) of eggshell strength, and it distinguishes targets for genetic selection (Layer 1) from pathways for precision management (Layer 2).

Discussion

Eggshell strength is a critical economic trait whose underlying uterine molecular mechanisms require a systems-level understanding. While previous transcriptomic studies have identified candidate genes associated with shell formation(Gao, et al., 2022; Wu, et al., 2022), insights from a single physiological time point cannot distinguish between stable genetic determinants that persist across the laying cycle and transient, age-responsive regulatory states. This distinction is essential for prioritizing targets for genetic selection versus management intervention. In this study, we implemented a comparative multi-stage framework, analyzing uterine transcriptomes and longitudinal phenotypes from hens at two strategic laying stages: 60 weeks (peak lay) and 90 weeks (late lay). This design enabled us to address two fundamental questions: (1) Do the observed transcriptional differences represent stable genetic determinants or age-specific responses? (2) Can we dissect a conserved core regulatory program from stage-adapted effector networks? Our findings provide affirmative answers to both questions and culminate in a dynamic, two-tiered model of uterine regulation that reconciles genetic stability with age-related phenotypic decline.

The most robust finding is the delineation of a conserved core transcriptional program that operates consistently across the laying cycle. This program is empirically defined by 86 differentially expressed genes (DEGs) common to both ages, with 96.5% exhibiting concordant regulation direction (Fig. 2) and highly correlated fold-changes (r = 0.84). Notably, 76.7% of these core DEGs showed significant correlation with longitudinal shell strength (Supplementary Table S3), underscoring their role as stable determinants.

Functional enrichment of this core program revealed pathways consistently associated with shell strength, including "Cytoskeleton in muscle cells," "Neuroactive ligand-receptor interaction," and "ECM-receptor interaction" (Supplementary Figure S3), aligning with the known roles of uterine contractility (Chen, et al., 2021; Zhang, et al., 2025) and tissue integrity in eggshell formation.

At the regulatory apex, we identified 8 conserved transcription factors (BHLHE41, TCF15, SATB1, DLX2, POU1F1, RXRG, NR1D1, NR1D2) that maintained strong regulatory correlations with target DEGs in both age cohorts (Fig. 6). Their biological functions provide insight into the core program's nature: SATB1, a chromatin organizer, may coordinate uterine genes involved in tissue homeostasis and local immune modulation, as uterine health impacts eggshell quality (Feng, et al., 2023). RXRG, a retinoid X receptor, is a key component of vitamin D signaling; we hypothesize it regulates calcium transporters (e.g., calbindin, TRPV6) in the uterus, directly influencing calcium utilization (Nys and Le Roy, 2018). TCF15 governs uterine structural integrity (Khaltabadi Farahani, et al., 2022). BHLHE41 links core regulation to circadian rhythms, potentially connecting photoperiod to biomineralization efficiency (Xin, et al., 2021)(Liu, et al., 2025). POU1F1 and DLX2 suggest integration of endocrine signaling and tissue patterning (Sah, et al., 2018). The persistence of these TFs indicates they establish the baseline genetic potential for shell strength (Layer 1 of our model), defining targets for genetic selection.

Beyond this stable core, our analysis reveals a dynamic layer of stage-specific adaptation. WGCNA uncovered that while core TFs are stable, their downstream network associations shift with age. The MElightcyan module, dominant at 60 weeks, correlates strongly with shell strength during peak lay and is enriched for structural pathways (Fig. 3A, C). In contrast, the MEyellow module at 90 weeks represents an aging-amplified network whose correlation with strength progressively intensifies (r = 0.57 at 40 w to 0.72 at 90 w) and is enriched for "Calcium signaling" and "Hormone signaling" (Fig. 3B, D). The functional shift from structural integrity (MElightcyan) to calcium and endocrine regulation (MEyellow) mirrors the well-documented physiological challenges of aging hens, including reduced calcium mobilization efficiency (Huang, et al., 2025; Yang, et al., 2025) and altered hormonal profiles (Hu, et al., 2025; Zhang, et al., 2019).

The increasing dominance of MEyellow indicates that the core program is differentially executed through these plastic networks; in late lay, this network becomes rate-limiting for shell strength. Thus, MEyellow represents Layer 2 of our model, identifying pathways (calcium transport, hormone signaling) as prime targets for precision management to mitigate age-related decline.

Our multi-step prioritization (differential expression, longitudinal correlation, QTL mapping) converged on high-confidence candidates with distinct stage-specific associations. The intersection of stringent phenotypic correlation yielded 17 core genes (Fig. 4), including ADRA2A and P2RX7, which mediate communication between systemic physiology and uterine function (Liu, et al., 2024)(Sun, et al., 2013). Integration with QTL data highlighted CNTNAP5 (peak-lay) and SLCO1C1 (late-lay) as functionally contextualized candidates (Fig. 5). We propose the following hypotheses for their uterine roles: CNTNAP5 encodes a cell adhesion molecule; we hypothesize it may facilitate cell-cell communication in uterine epithelium or stroma, supporting tissue integrity under high-frequency reproductive demands. SLCO1C1 transports thyroid hormones (Yue, et al., 2025); we hypothesize it regulates local thyroid hormone availability in the uterus, thereby modulating metabolic activity and calcium flux, particularly critical in aged hens where endocrine regulation becomes limiting. These stage-specific candidates exemplify the dynamic nature of regulation and support the concept of age-adapted effector networks (Layer 2).

Integrating our findings, we propose a two-tiered regulatory model of eggshell strength that directly answers the questions posed in the introduction: Layer 1 (Stable Core Program): Comprising 86 conserved DEGs and 8 core TFs, this layer establishes the intrinsic genetic potential for shell strength, operating consistently across ages. It provides targets for genetic selection. Layer 2 (Age-Adapted Effector Networks): Represented by MElightcyan (structural focus) and MEyellow (calcium/hormone focus), this layer modulates core program output in response to physiological context. Its age-amplified dysfunction explains shell strength decline even in hens with favorable genetics, identifying pathways for precision management. This model provides a conceptual framework for understanding both heritable basis and age-associated deterioration, offering a dual-strategy roadmap for improving shell quality.

We acknowledge several limitations. First, while our stringent longitudinal phenotyping partially compensates for sample size (n = 5 per group), larger cohorts would enhance statistical power. Second, all regulatory relationships are correlational and require experimental validation (e.g., CRISPR, RNAi) for prioritized genes (SATB1, RXRG, CNTNAP5, SLCO1C1). Third, direct longitudinal expression measurement in the same individuals is infeasible in uterine tissue; future studies using less invasive tissues could complement our findings. Despite these limitations, the high concordance of our findings across two independent cohorts, the strong cross-age correlations, and the integration with independent QTL data provide confidence in the robustness of our core conclusions.

Conclusion

In summary, using a two-stage comparative framework, we compared uterine transcriptomes from hens at peak lay (60 weeks) and late lay (90 weeks) and correlated gene expression with longitudinal phenotypic records to distinguish stable from age-responsive transcriptional signatures. This work provides a two-tiered regulatory model explaining both the heritable basis and age-associated decline of eggshell strength. Our findings deliver high-confidence candidate genes for genetic selection and pinpoint pathways for precision management, offering a dual-strategy roadmap to improve shell quality and sustainability in egg production.

Data availability

The transcriptome sequencing data were deposited in the Sequence Read Archive (SRA) database (https://www.ncbi.nlm.nih.gov/sra) of NCBI under the BioProject accession numbers PRJNA1430166 and SAMN 56269786 to 56269805, and the SRA project accession numbers SRR 37453962 to 37453981.

Declaration of competing interest

The authors declare that they have no competing interests.

CRediT authorship contribution statement

Liyuan Wang: Writing – original draft, Formal analysis, Data curation, Conceptualization. Lei Liu: Writing – original draft, Visualization, Software, Resources, Methodology, Conceptualization. Ying Bai: Methodology, Investigation, Funding acquisition, Conceptualization. Yanheng Wang: Visualization, Investigation, Data curation. Chuanwei Zheng: Resources, Project administration. Jinwei Wang: Methodology, Investigation, Formal analysis. Shimin Chang: Supervision, Project administration, Funding acquisition. Zhiqiong Mao: Project administration, Investigation, Data curation. Xiaohui Liu: Writing – review & editing, Project administration. Yahui Gao: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

Disclosures

The authors declare that there are no conflicts of interest, financial or otherwise, that could have influenced the work presented in this study.