Longitudinal metabolomic profiling across three age points reveals key metabolites associated with body weight, serum immunoglobulins, and biochemical parameters in pigeons

Abstract

Post-hatch development in pigeons involves dramatic shifts in nutrition and metabolism. However, the underlying systemic metabolic reprogramming remains poorly characterized. Thus, longitudinal analyses of body weight (BW), serum immunoglobulins, biochemical parameters, and untargeted metabolomics at three key developmental stages: postnatal day 5 (PND 5, crop milk dependence; N = 10), PND 15 (transition to grains; N = 10), and PND 25 (independent grain intake; N = 10) were conducted. The BW increased significantly between PND 5 and PND 15 (P < 0.05), concomitant with elevated serum glucose, albumin, total cholesterol, and high-density lipoprotein. Immunoglobulin (Ig) dynamics revealed a significant decline in IgA at PND 15 and PND 25 relative to PND 5 (P < 0.05). The IgG showed a temporary significant decrease at PND 15 compared to PND 5 (P < 0.05), returning to PND 5 levels by PND 25. Metabolomics demonstrated dynamic pathway alterations. Comparing PND 15 to PND 5, differential metabolites were significantly enriched in five pathways (P < 0.05), most prominently alanine, aspartate and glutamate metabolism (P < 0.001) and purine metabolism (P = 0.003). The PND 15 to PND 25 transition featured prominent shifts, notably in glycerophospholipid metabolism (P < 0.001) and tricarboxylic acid (TCA) cycle continuation (P = 0.026). Crucially, PND 25 vs PND 5 analysis identified seven remodeled pathways, with core reprogramming involving alanine, aspartate and glutamate metabolism (P < 0.001), arginine biosynthesis (P < 0.001), and the TCA cycle (P = 0.002), which emerged as a central metabolic hub. The K-means clustering of 25 hub metabolites and physiological parameters delineated seven co-regulation patterns. Notably, BW and nutritional markers (albumin, cholesterol) correlated positively with TCA intermediates (citrate, α-ketoglutarate, malate; P < 0.05), while showing inverse associations with purine catabolites (e.g., guanine, xanthosine) and bile acids. Conversely, immunoglobulins correlated positively with purine metabolites and bile acids. This study identifies TCA cycle intermediates and purine metabolites as dual biomarkers regulating growth and immune function during pigeon development. These findings provide a foundation for targeted nutritional strategies that require adjustments as a function of aging, such as key metabolite supplementation and phospholipid modulation, to optimize pigeon management practices.

Introduction

Pigeon meat is prized as a delicacy for its low fat and high protein content, gaining popularity among consumers in Europe, the United States, and China (Kokoszyński et al., 2020; Pomianowski et al., 2009). Driven by rapid pigeon breeding industry expansion, China has emerged as the world's largest pigeon producer, with annual production and sales approaching 700 million squab pigeons, supplying 80 % of global pigeon meat (Jiang et al., 2019). Unlike other poultry, pigeons are altricial birds. During early post-hatch development, squabs rely exclusively on parental "crop milk", transitioning to a mixture of "crop milk" and soaked grains, and ultimately to independent grain consumption (Gillespie et al., 2013; Shetty et al., 1992). As altricial birds, squabs require prolonged parental care, which extends the breeding cycle per batch and reduces the annual number of breeding batches for commercial flocks. Additionally, crop milk synthesis demands extra nutritional input from parent pigeons, increasing feed costs while lowering the input-output ratio. This developmental pattern prolongs breeding cycles and compromises production efficiency in commercial squab operations. The rising global demand for pigeon meat is primarily concentrated in China, parts of Southeast Asia (including Vietnam and Thailand), and select Middle Eastern regions. This demand is predominantly for human consumption, driven by its integration into traditional dietary practices (culinary cultures in certain Chinese regions and festive customs in parts of the Middle East) as well as its nutritional attributes (high protein, low fat, low cholesterol) aligning with contemporary nutritious diet preferences. Additionally, commercial pigeon farms use controlled environmental management (temperature-humidity regulation, ventilation) analogous to chicken farms, and prevent diseases through vaccination, regular quarantine, and biosecurity (including personnel and vehicle disinfection) (Adawy and Abdel-Wareth, 2023). Consequently, a significant disparity exists between the rising global demand for pigeon meat and the comparatively underdeveloped production technologies relative to other poultry sectors. A deeper understanding of the biological mechanisms governing growth, immune maturation, and metabolic homeostasis during key developmental stages is therefore critical for optimizing pigeon health, welfare, and productivity (Fu et al., 2023). While physiological parameters (e.g., body weight, serum immunoglobulins, biochemical markers) offer valuable phenotypic insights, they provide limited resolution into the underlying molecular drivers of age-associated changes (Orakpoghenor et al., 2021).

Metabolomics serves as a powerful system biology tool, capturing the dynamic biochemical phenotype through comprehensive profiling of low-molecular-weight metabolites (<1500 Da) within biological systems (Muthubharathi et al., 2021; Zhang et al., 2025a). These metabolites represent the terminal outputs of genomic, transcriptomic, and proteomic activity and are directly responsive to environmental and physiological influences. As such, the metabolome provides a sensitive readout of developmental progression, nutritional status, and pathophysiological states (Feng et al., 2024; Zhang et al., 2025a). Although metabolomic studies have elucidated metabolic pathways in chickens, particularly in contexts of nutrition and disease (Cônsolo et al., 2020; Shao et al., 2018; Van Every and Schmidt, 2021), a significant knowledge gap persists regarding the longitudinal evolution of the serum metabolome across critical pigeon developmental stages and its systematic integration with fundamental physiological traits.

Previous avian metabolomics researches have often focused on single time points or isolated systems, leaving a critical void: understanding how the pigeon serum metabolome dynamically remodels across key developmental transitions (e.g., squab to fledgling) and functionally integrates with core physiological parameters. Identifying metabolites linked to growth, immunity, and metabolic homeostasis could yield valuable biomarkers for precision breeding programs and optimized management strategies. To address this gap, a longitudinal serum metabolomics study at three pivotal developmental stages was conducted, concurrently measuring body weight, immunoglobulin profiles (IgG, IgM, IgA), and serum biochemistry. The present study hypothesizes that each developmental stage exhibits a distinct metabolic signature that not only correlates with but may functionally contribute to critical phenotypic transitions. Through this integrated analysis, the present study aimed to both elucidate the metabolic mechanisms underlying squab development and identify novel biomarker associated with pigeon production efficiency and metabolism, which can inform future strategies for nutritional regulation.

Materials and methods

Ethics statement

All animal procedures were performed according to the National Standards of the People's Republic of China for Laboratory Animal Welfare and were approved by the Animal Care and Use Committee (ACUC) of Nanchang Normal University.

Animal and sample collection

A total of N = 30 healthy male Wahui pigeons (identified by farm technical staff via physiological traits) were obtained from the Wahui Pigeon Breeding Farm (Jinxian County, Nanchang, Jiangxi Province, China). Pigeons were selected at three developmental stages (Fig. 1): postnatal day 5 (PND 5, N = 10), PND 15 (N = 10), and PND 25 (N = 10). All pigeons were maintained under identical environmental condition with ad libitum access to feed and water. Following body weight measurement, 3 mL of blood were collected using non-heparin-coated tubes. Samples were centrifuged at 3,000 × g for 10 min. The serum supernatant was aliquoted into 1.5 mL sterile microcentrifuge tubes, immediately flash-frozen in liquid nitrogen, and stored at −80°C until subsequent serum biochemical analysis and untargeted metabolomics profiling.

Fig. 1.

Pigeons at three developmental stages.

Serum biochemistry and serum immunoglobulins measurements

Serum biochemical parameters including glucose, total protein (TP), urea nitrogen, albumin, triglycerides, total cholesterol (TC), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) were quantified using the standardized commercial assay kits (Nanjing Jiancheng Bioengineering Institute Inc., Nanjing, China) in accordance with the instructions.

Serum immunoglobulin levels (IgA, IgG, IgM) were measured using commercial ELISA kits according to the manufacturer’s protocols (Jiangsu Meimian Industrial Co., Ltd, Yancheng, China).

Liquid chromatography-mass spectrometry (LC/MS) analysis

Serum metabolomic profiling was conducted using an ultrahigh-performance liquid chromatography system (Vanquish, Thermo Fisher Scientific) coupled to a high-resolution Orbitrap Exploris 120 mass spectrometer (Thermo Fisher Scientific). Following established methodologies (Guo et al., 2019), 100 μL serum aliquots were mixed with 300 μL methanol (Merck, Germany) and 10 μL l-2-chlorophenylalanine (internal standard), vortexed for 30 s, and centrifuged at 13,800 × g for 10 min at 4°C prior to supernatant transfer to injection vials. Chromatographic separation employed a Waters ACQUITY UPLC BEH Amide column with mobile phases consisting of (A) 0.1 % formic acid in water and (B) 0.1 % formic acid in acetonitrile under gradient elution (0-1 min: 95 % B; 1-14 min: 95→65 % B; 14-16 min: 65→40 % B; 16-18 min: 40 % B; 18-18.1 min: 40→95 % B; 18.1-23 min: 95 % B) at 0.3 mL/min flow rate with 2 μL injection volume and 4°C autosampler temperature. Mass spectrometry operated in ESI positive/negative mode with spray voltages of +3.8 kV/−3.4 kV, sheath/aux gas flows of 50/15 arb, and capillary temperature at 320°C. Raw data conversion to mzXML format utilized ProteoWizard (v3.0), with subsequent metabolite identification via an in-house R package referencing the BiotreeDB (V3.0) database.

Statistical analysis

Body weight, serum immunoglobulin levels, and biochemical parameters were analyzed using SPSS (version 25.0; IBM Corp., Armonk, NY) with one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. Results are expressed as mean ± SEM, with statistical significance defined as P < 0.05 and high significance as P < 0.01. Metabolomics data underwent multivariate analysis including principal component analysis (PCA) and orthogonal partial least squares-discriminant analysis (OPLS-DA) to identify age-dependent metabolic variation. Differential metabolites were identified based on thresholds of P < 0.05, fold change (FC) > 1.5 or < 0.67 (PND 15/PND 5, PND 25/PND 15, PND 25/PND 5), and OPLS-DA variable importance in projection (VIP) > 1. Metabolic pathway enrichment analysis was performed via the MetaboAnalyst web platform (https://www.metaboanalyst.ca). Spearman's rank correlations between differential metabolites and phenotypic markers (body weight, immunoglobulins, biochemical parameters) were computed and visualized using custom R scripts.

Results

Analysis of body weight, serum immunoglobulins, and biochemical parameters

Body weight exhibited significant age-dependent progression (F_{2, 27} = 1005.93, P < 0.001, ${\eta }_P^2$ = 0.98; Fig. 2A), with the most rapid increase occurring between PND 5 and PND 15. Although significant weight gain continued from PND 15 to PND 25 (P < 0.05), the growth rate markedly decelerated compared to the earlier developmental phase. The curve model, Gompertz, can be fitted with the growth curve of the pigeons (R² > 0.98). The equation of Gompertz model was Y = 466.58 × e^{-4.94exp(−0.24t)} (Y = BW of pigeon (g); t = time (day)), with the terminal BW, inflection weight and age in this model being 466.6 g, 171.6 g, and 6.8 day, respectively (Fig. S1; Table S1).

Fig. 2.

Comparison of body weight, serum biochemical parameters, and immunoglobulins at different ages (PND 5, PND 15, and PND 25) in Pigeons. A, body weight. B, glucose. C, total protein. D, albumin. E, urea nitrogen. F, triglyceride. G, total cholesterol. H, high-density lipoprotein. I, low-density lipoprotein. J, immunoglobulin (Ig)A. K, IgG. L, IgM.

Serum biochemical parameters analysis (Fig. 2B-I) revealed glucose levels significantly elevated on PND 15 and PND 25 compared with PND 5 (F_{2, 27} = 7.55, P = 0.002, ${\eta }_P^2$ = 0.36), while maintaining comparable concentrations between PND 15 and PND 25 (P > 0.05). The TP demonstrated progressive accumulation across timepoints, achieving statistical significance on PND 25 compared to PND 5 (F_{2, 27} = 4.50, P = 0.021, ${\eta }_P^2$ = 0.25). Correspondingly, albumin showed significant increases on PND 15 and PND 25 compared to PND 5 (F_{2, 27} = 26.71, P < 0.001, ${\eta }_P^2$ = 0.66), with no inter-stage difference observed between these later timepoints (P > 0.05). There was no significant difference in serum urea nitrogen content among the three time points (F_{2, 27} = 1.78, P = 0.188, ${\eta }_P^2$ = 0.12). There is a significant difference in serum triglyceride concentration among the three stages (F_{2, 27} = 11.10, P < 0.001, ${\eta }_P^2$ = 0.45), which increased from PND 5 to PND 15 (P < 0.05) before declining by PND 25 to a level not significantly different from the start (P > 0.05). The TC level in serum on PND 15 and PND 25 was significantly higher than that on PND 5 (F_{2, 27} = 17.83, P < 0.001, ${\eta }_P^2$ = 0.60). The HDL displayed a progressive and statistically significant increase across all three stages (F_{2, 27} = 35.01, P < 0.001, ${\eta }_P^2$ = 0.72). The LDL concentration was significantly altered by the stage (F_{2, 27} = 5.90, P = 0.007, ${\eta }_P^2$ = 0.30), with levels increasing significantly from PND 5 to PND 15 (P < 0.05) before declining to an intermediate value on PND 25 that was not different from either of the earlier timepoints (P > 0.05).

Immunoglobulin profiling indicated IgA was significantly reduced on PND 15 and PND 25 compared to PND 5 (F_{2, 27} = 16.29, P < 0.001, ${\eta }_P^2$ = 0.55), with comparable levels maintained between PND 15 and PND 25 (P > 0.05). There is a significant difference in IgG among the three stages (F_{2, 27} = 3.73, P = 0.037, ${\eta }_P^2$ = 0.22), which displayed a significant decrease on PND 15 compared to PND 5 (P < 0.05), but recovered by PND 25 to concentrations statistically indistinguishable from both previous stages (P > 0.05). The IgM levels remained consistent throughout development with no significant temporal variations observed (F_{2, 27} = 2.15, P = 0.136, ${\eta }_P^2$ = 0.14) .

Overview of serum metabolome at different developmental stages

To comprehensively characterize metabolic profile dynamics during pigeon development, non-targeted metabolomics was employed across three key developmental stages: PND 5 ("crop milk" intake), PND 15 (mixed "crop milk" and soaked grains), and PND 25 (independent grain intake). The PCA revealed distinct serum metabolite profiles at each timepoint (Fig. 3A), with clear separation between PND 5-PND 15 and PND 15-PND 25 clusters, though the latter showed reduced intergroup distance compared to the former. A total of 1547 metabolites spanning 11 distinct classes were identified in combined anion/cation modes (Fig. 3B), dominated by lipids and lipid-like molecules (34.97 %), organic acids and derivatives (20.43 %), and organoheterocyclic compounds (15.58 %). The OPLS-DA confirmed significant metabolic differences between all pairwise comparisons (Fig. 3C–E), with models demonstrating high explanatory (R²Y > 0.94) and predictive power (Q²Y > 0.69; Table S2). Permutation tests yielded negative Y-axis intercepts (Fig. 3F–H), validating model robustness against overfitting and confirming data reliability.

Fig. 3.

Multivariate statistical analysis of serum metabolomic data based on LS-MS/MS. A, principal component analysis (PCA) of serum metabolomic profiles in Wahui pigeons at different ages. B, composition analysis of all identified metabolites in serum metabolomic profiles of Wahui pigeons at different ages. C, D, and E, are the orthogonal partial least squares–discriminant analysis (OPLS-DA) score plots of PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively. F, G, and H, are the permutation test results of OPLS-DA of PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively.

Using thresholds of VIP > 1.0, P < 0.05, and FC > 1.5 or FC < 0.67, the present study identified 258 (136 upregulated, 122 downregulated), 235 (101 upregulated, 134 downregulated), and 329 (138 upregulated, 191 downregulated) differential metabolites in the PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5 comparisons, respectively (Fig. 4A). These comparisons revealed 24 shared differential metabolites with 49, 70, and 89 unique DEMs for PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively (Fig. 4B). In order to further reveal the metabolic pathways that play a key role in pigeon growth, these differential metabolites were used for KEGG pathway analysis. The Fig. 4F–H illustrate the metabolic pathways enriched by the differential metabolites screened out by comparing PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively. The results indicate that the differential metabolites in the PND 15 vs PND 5 are significantly enriched in 5 pathways, namely alanine, aspartate and glutamate metabolism (P < 0.001, pathway impact = 0.156), purine metabolism (P = 0.003, pathway impact = 0.022), arginine biosynthesis (P = 0.016, pathway impact = 0.143), primary bile acid biosynthesis (P = 0.032, pathway impact = 0.076), and citrate cycle (TCA cycle, P = 0.037, pathway impact = 0.149). The PND 25 vs PND 15 comparison revealed enrichment in glycerophospholipid metabolism (P < 0.001, pathway impact = 0.339), arginine biosynthesis (P = 0.011, pathway impact = 0.011), TCA cycle (P = 0.026, pathway impact = 0.088), glutathione metabolism (P = 0.048, pathway impact = 0.030), and alanine, aspartate and glutamate metabolism (P = 0.048, pathway impact = 0.050). As for the PND 25 vs PND 5, the different metabolites are significantly enriched in 7 pathways: alanine, aspartate and glutamate metabolism (P < 0.001, pathway impact = 0.356), arginine biosynthesis (P < 0.001, pathway impact = 0.218), TCA cycle (P = 0.002, pathway impact = 0.223), histidine metabolism (P = 0.011, pathway impact = 0.123), purine metabolism (P = 0.013, pathway impact = 0.042), glyoxylate and dicarboxylate metabolism (P = 0.014, pathway impact = 0.140), and ascorbate and aldarate metabolism (P = 0.034, pathway impact = 0.405). The functional interactions of these pathways were analyzed by MetaboAnalyst web server, demonstrating that the TCA cycle as a central hub connecting these enriched pathways (Fig. 5).

Fig. 4.

Overview of differential metabolites and their enriched differential metabolic pathways in serum metabolomics at different developmental stages in Wahui pigeons. A, the number of differential metabolites between each two-time-point in pigeon serum. B, common and unique differential metabolites identified in the comparison between each two time points in pigeon serum. C, D, and E, are volcano plots of PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively. F, G and H, are the metabolic pathways enriched by the differential metabolites screened out by comparing PND 15 vs PND 5, PND 25 vs PND 15, and PND 25 vs PND 5, respectively.

Fig. 5.

Network map of differential metabolites enriched pathways. Each point represents a pathway. The color represents the significant P value of enrichment. The redder the point represents, the more significant the enrichment of the pathway. The size of the circle represents the number of metabolites annotated to the pathway (count), and the larger the circle is, the higher the count.

To identify key regulatory metabolites, the present study analyzed compounds within enriched differential pathways as hub metabolites, identifying 25 candidates clustered into three distinct categories (Fig. 6). K-means clustering further revealed coordinated trends between these hub metabolites and physiological parameters (body weight, biochemical indices, serum immunoglobulins) across development (Fig. 7, Table 1). The analysis delineated seven co-expression clusters: Cluster 1 (3 members) exhibited initial decline followed by increase; Cluster 2 (6 members: body weight, albumin, glucose, TC, citrate, 4‑hydroxy-2-oxoglutarate) showed continuous increase; Cluster 3 (11 members: IgG, IgA, IgM, 8-oxoadenine, guanine, xanthosine, N-acetyl-l-aspartate, inosine, guanosine, cholic acid, glycocholic acid) decreased then stabilized; Cluster 4 (3 members: urea nitrogen, trans-urocanate, PA(18:3(9Z,12Z,15Z)/0:0)) remained stable before declining; Cluster 5 (3 members: triglycerides, LysoPC(20:4(8Z,11Z,14Z,17Z)/0:0), 5-oxoproline) increased then decreased; Cluster 6 (9 members: HDL, TP, 3-methylhistidine, fumaric acid, glucaric acid, PC(20:4(5Z, 8Z, 11Z, 14Z)/15:0), d-glucuronic acid, l-malate, and alpha-ketoglutaric acid) displayed slow then rapid increase; Cluster 7 (2 members: LDL and argininosuccinic acid) demonstrated rapid initial increase followed by gradual decline, with distinct temporal dynamics across developmental stages.

Fig. 6.

Heat map of hub metabolites involved in enrichment of differential metabolic pathways.

Fig. 7.

K-means clustering analysis of differential metabolites, body weight, serum immunoglobulins, and biochemical parameters in Wahui pigeons. The X-axis represents three key time points in the growth and development of pigeons, and the Y-axis represents the levels of metabolites and physiological parameters. Additionally, sub-class indicates metabolites and physiological parameters exhibiting similar trends. Clusters 1 to 7 have 3, 6, 11, 3, 3, 9, and 2 metabolites or physiological parameters, respectively.

Table 1.

Cluster differential metabolites and physiological parameters statistics.

Cluster 1	Cluster 2	Cluster 3	Cluster 4	Cluster 5	Cluster 6	Cluster 7
Beta-Hydroxymyristic acid	Body weight	IgG	Urea nitrogen	Triglycerides	HDL	LDL
Taurine	Albumin	IgA	Trans-urocanate	LysoPC(20:4(8Z,11Z,14Z,17Z)/0:0)	TP	Argininosuccinic acid
Gamma-Glutamylcysteine	Glucose	IgM	PA(18:3(9Z,12Z,15Z)/0:0)	5-oxoproline	3-Methylhistidine
	TC	8-Oxoadenine			Fumaric Acid
	Citric Acid	Guanine			Glucaric acid
	4-Hydroxy-2-oxoglutarate	Xanthosine			PC(20:4(5Z,8Z,11Z,14Z)/15:0)
		N-Acetyl-l-aspartic acid			D-Glucuronic acid
		Inosine			L-Malate
		Guanosine			alpha-Ketoglutaric acid
		Cholic acid
		Glycocholic acid

Correlation analyses revealed significant associations between physiological parameters and key metabolites (Fig. 8). Body weight, TC, HDL, and albumin showed strong positive correlations with citrate, glucarate, l-malate, and α-ketoglutarate (P < 0.05), while exhibiting significant negative correlations with 8-oxoadenine, guanine, xanthosine, inosine, guanosine, cholate, and glycocholate (P < 0.05). Serum glucose demonstrated positive correlations with fumaric acid, PC(20:4(5Z,8Z,11Z,14Z)/15:0), l-malate, and alpha-ketoglutaric acid (P < 0.05), but negative correlations with 8-oxoadenine, guanine, xanthosine, N-acetyl-l-asparitic acid, and glycocholic acid (P < 0.05). Urea nitrogen levels correlated positively with trans-urocanate, cholic acid, and glycocholic acid (P < 0.05), while showing inverse relationships with beta-hydroxymyristic acid, citric acid, glucaric acid, PC(20:4(5Z, 8Z, 11Z, 14Z)/15:0), d-glucuronic acid, and alpha-ketoglutaric acid (P < 0.05). Immunoglobulins (particularly IgA) were positively associated with 8-oxoadenine, guanine, xanthosine, N-acetyl-l-asparitic acid, inosine, guanosine and glycocholic acid (P < 0.05), and negatively correlated with LysoPC(20:4(8Z,11Z,14Z,17Z)/0:0), citric acid, glucaric acid, argininosuccinic acid, and 4‑hydroxy-2-oxoglutarate (P < 0.05).

Fig. 8.

Heatmap of correlations between differential metabolites and physiological parameters. The redder the color, the larger the positive correlation coefficient, and the bluer the color, the larger the negative correlation coefficient. Significance is indicated by asterisks, *** for P value ≤ 0.001, ** for P value ≤ 0.01, * for P value ≤ 0.05.

Discussion

Pigeon meat is increasingly popular among consumers, driving rapid market growth. However, research on pigeon production technology lags behind other poultry like chickens and ducks (He et al., 2025; Zhang et al., 2025b). While studies have characterized growth curves and serum parameters in pigeons (Gao et al., 2016; Orakpoghenor et al., 2021), deeper understanding of growth-metabolic markers remains limited. In order to further explore the changes in the metabolic profile of pigeons during growth and development and their close correlation with physiological parameters, non-targeted metabolomics was employed to detect metabolites at three key developmental time points, namely PND 5 ("crop milk" intake period), PND 15 (mixture of "crop milk" and soaked grains period), PND 25 (independent grain intake period) in the present study.

Body weight serves as a crucial health and management indicator (Yu et al., 2025). Our results show rapid weight gain from PND 5 to PND 15 (460 g final weight), slowing thereafter. This is consistent with previous studies that the inflection age of pigeons is around 8th day and the mature body weight is around 500 g at 28th day (Gao et al., 2016; Xiang and Wang, 2000). Since crop milk plays an indispensable role from day 1 to day 8 (Gao et al., 2016), the timepoints in the present study effectively represent the transition from crop milk dependence to independent feeding.

Serum biochemical indicators can reflect the nutritional status of animals. Serum glucose levels increased significantly from PND 5 to PND 15 and then remained stable until PND 25, which is consistent with previous study reporting that blood glucose levels in pigeons stabilized from day 14 (Gao et al., 2016). The present study found that serum TP and albumin levels continued to increase with age, which is in line with previous findings in pigeons and pheasants (Gao et al., 2016; Kececi and Çöl, 2011). This could be explained by the high metabolic rate and rapid growth rate of birds, which require high blood protein levels to be used for protein synthesis during growth. Moreover, as a marker of protein catabolism in birds (Singer, 2003), serum urea nitrogen levels did not change significantly with age, but decreased numerically—a pattern that likely reflects balanced protein synthesis and utilization, consistent with the steady weight gain observed, and supporting the absence of metabolic disruption in healthy squabs under controlled conditions. This also supports the view that pigeons' ability to synthesize and utilize protein gradually matures. Similarly, Gao et al. (2016) found that serum urea nitrogen levels in pigeons decreased from day 1 to day 7 and then remained stable. Triglycerides, TC, HDL, and LDL are the core indicators in blood lipid testing, and they are closely related to each other and together reflect the metabolic status of lipids in the body. Consistent with the study of Xu et al. (2020), the highest concentration of triglycerides appeared on the 14th day. This study also found that triglycerides first increased and then decreased, with the highest content on the 15th day. This should be due to the consumption of crop milk, which is rich in lipids. The serum TC content increased significantly from the 5th to the 15th day and then remained stable, which may be the combined effect of changes in HDL and LDL content. Both HDL and LDL increased significantly from the 5th to the 15th day, but from the 15th to the 25th day, HDL continued to increase, while LDL fell back, ultimately resulting in the TC content remained stable from the 15th to the 25th day. Generally, HDL is considered good cholesterol, while LDL is considered bad cholesterol (von Eckardstein et al., 2023), so the above changes in serum cholesterol in pigeons with age are beneficial for growth.

Despite some differences in chemical structure, the biological role of antibodies is the same in birds and mammals, which is to eliminate extracellular invaders from the body (Tizard, 2002). Pigeon milk, in addition to being rich in protein and fat, also contains considerable amounts of immunoglobulin IgA (1.45 mg/mL) and IgG (0.34 mg/mL) (Ge et al., 2020; Wang et al., 2023), which is derived from the parental blood, most likely through selective active transfer. The IgA and IgG in the crop milk are absorbed in the intestine of the young pigeons by endocytosis and transferred to the blood in a manner similar to mammalian colostrum (Tizard, 2002). Therefore, the finding in this study that the levels of IgG and IgA in pigeon serum decreased significantly from PND 5 to PND 15 may be related to the gradual reduction in crop milk intake.

The metabolomics results showed that although the three time points could be distinguished, indicating that the metabolites in the pigeon serum changed significantly with age, the distance between the PND 5 and the PND 15 was greater than the distance between the PND 15 and the PND 25, indicating that the changes in serum metabolites from the PND 5 to the PND 15 were greater than those between the PND 15 and the PND 25. One plausible explanation for this observation is that, firstly, a significant dietary shift occurs between PND 5 and PND 15, transitioning from exclusive reliance on crop milk to a mixture of crop milk and grain feed. In addition, the pigeons are in a phase of rapid development during this period, characterized by more vigorous metabolic activity. The number of differential metabolites was also greater in PND 15 vs PND 5 than in PND 25 vs PND 15 (258 vs 235). It is of note that when comparing the differential metabolite compositions of PND 15 vs PND 5 and PND 25 vs PND 15, there are also great differences, with only 79 differential metabolites being shared (Fig. 4B).

The temporal analysis of metabolic pathway enrichment across key developmental stages (PND 5, PND 15, and PND 25) in pigeons reveals dynamic shifts in metabolic priorities that align with physiological maturation. During the rapid growth period from PND 5 to PND 15, differential metabolites were significantly enriched in alanine-aspartate-glutamate metabolism and TCA cycle. These metabolic shifts may reflect an urgent need of young pigeons for protein synthesis substrates and energy supply (Peng et al., 2018; Poortmans, 2004; Seol et al., 2006; Sunny, 2007) and these micro-pathway activations likely serve as the biochemical drivers for the gradual increase in serum TP (a macro-indicator of protein status) observed in this study. At the same time, the activation of purine metabolism and arginine biosynthesis suggests the enhancement of nucleic acid synthesis (such as cell proliferation) (Rengaraj et al., 2013; Weinstock et al., 1972; Welch and Rudolph, 1982) and nitrogen metabolism (such as muscle growth) (Castro and Kim, 2020; Fathima et al., 2024; Liao et al., 2025; Wang et al., 2022), which is consistent with the characteristics of rapid organ development and weight growth at this stage. The enrichment of the primary bile acid pathway may be associated with the establishment of lipid digestion function (Ge et al., 2019; Krogdahl, 1985). These pathways support foundational tissue development and nutrient utilization in juvenile birds. By the mid-development transition (PND 25 vs PND 15), metabolic emphasis shifts toward structural and protective functions. The dominance of glycerophospholipid metabolism reflects membrane biogenesis and remodeling critical for organ maturation, a process that occurs without disrupting serum lipid homeostasis (e.g., total cholesterol stability after PND 15) as the body balances lipid synthesis and utilization, while persistent arginine biosynthesis and TCA cycle activity sustain growth demands. Notably, the emergence of glutathione metabolism suggests increasing oxidative stress regulation, potentially linked to weaning or immune system activation (Cao et al., 2024; Deschutter and Leeson, 1986; Farooqui et al., 2000; Surai et al., 2019; Yin et al., 2023). Across the entire developmental process (PND 25 vs PND 5), the number of differential metabolite enrichment pathways was the largest, and the core pathways continued to strengthen: amino acid metabolism (glutamate, histidine) and arginine synthesis jointly drove protein accumulation, and the TCA cycle served as an energy hub throughout the development process. In addition, the emergence of glyoxylate and dicarboxylic acid metabolism was particularly critical, suggesting that pigeons may optimize the supply of dicarboxylic acids (such as oxaloacetate) through this pathway to maintain TCA cycle flux (Van Every and Schmidt, 2021), make up for the limitation of the lack of glyoxylate cycle in animals, and support efficient energy generation. The activation of ascorbic acid metabolism enhances the redox balance ability to meet the high activity requirements of subadults (Coudert et al., 2023; Van Hieu et al., 2022). Therefore, pigeon development from PND 5 to PND 25 undergoes: 1) coordinated activation of amino acid and energy metabolism, 2) emphasis on membrane remodeling and oxidative stress defense, and 3) comprehensive maturation of metabolic networks.

The trajectory of 25 hub metabolites across pigeon development reveals a tightly coordinated metabolic network that interfaces dynamically with physiological maturation. Clustering analysis delineates seven distinct metabolic-physiological modules, each exhibiting stage-specific behaviors: metabolites driving energy and biomass accumulation, such as citric acid, alpha-ketoglutaric acid, and l-malate (Clusters 2 and 6), demonstrate sustained positive correlations with body weight, serum albumin, and lipoproteins, underscoring the centrality of TCA cycle flux in powering growth through mitochondrial bioenergetics and cholesterol-dependent membrane expansion (Adhikari, 2017; Neuwirth et al., 2023; Ojano-Dirain et al., 2007). Conversely, purine derivatives (e.g., guanosine, 8-oxoadenine) and bile acids (e.g., glycocholic acid) in Cluster 3 decline post-weaning yet maintain strong positive associations with immunoglobulins (IgA/IgG/IgM), suggesting their dual role in early immune cell proliferation and later mucosal immunity maintenance (Jia and Dong, 2024). This metabolic-immune crosstalk contrasts sharply with antagonistic relationships between TCA intermediates and purine metabolites, implying a developmental trade-off wherein resources shift from immune priming toward somatic growth (Rauw, 2012). Concurrently, stress-adaptive metabolites like 5-oxoproline (glutathione precursor) (Zhou et al., 2019) and argininosuccinic acid (nitric oxide regulator) (Ouchi et al., 2021) exhibit transient peaks (Clusters 5 and 7), synchronizing with lipid remodeling and vascular development phases, while late-stage declines in urea nitrogen and urocanate (Cluster 4) signal improved nitrogen utilization efficiency. Critically, phospholipid speciation bifurcates functionally: the negative correlation of LysoPC(20:4(8Z,11Z,14Z,17Z)/0:0) with immunoglobulin hints at inflammatory signaling (Li et al., 2024) during rapid growth, whereas structural PC(20:4(5Z,8Z,11Z,14Z)/15:0) aligns with serum glucose content, optimizing membrane integrity for nutrient transport. Collectively, these patterns depict a hierarchical metabolic program, prioritizing immune investment in juveniles (PND 5-PND 15), energy-intensive growth in subadults (PND 15-PND 25), and metabolic stabilization in near-fledging birds (PND 25).

Importantly, the hub metabolites exhibiting strong temporal trends and physiological correlations emerge as potential candidates for future biomarker development and precision nutrition strategies in pigeon rearing though validation in larger cohorts and diverse farming contexts is needed. The consistent positive association of TCA cycle intermediates (citrate, alpha-ketoglutarate, malate) with body weight gain and serum albumin positions them as potential indicators of growth efficiency. Their decline could signal metabolic stress or nutrient insufficiency, enabling early intervention. Concurrently, the inverse relationship between purine catabolites (guanine, xanthosine) and lipoproteins, but positive linkage to immunoglobulins, suggests their dual utility as sensitive reporters of immune activation status, with potential for modulation via dietary nucleotide supplementation (Hess and Greenberg, 2012). For metabolic health monitoring, 5-oxoproline’s peak in early growth (Cluster 5) may reflect glutathione synthesis burden, serving as an oxidative stress biomarker to guide antioxidant provision, while the late-stage drop in urea nitrogen paired with urocanate (Cluster 4) could inform protein requirement adjustments to minimize nitrogen waste. Notably, phospholipid speciation offers preliminary actionable insights that elevating PC(20:4(5Z,8Z,11Z,14Z)/15:0), which correlates with glucose uptake and HDL, via omega-3/6-fortified feeds may enhance membrane function and energy homeostasis (Chilton et al., 2017), whereas suppressing LysoPC(20:4(8Z,11Z,14Z,17Z)/0:0) levels (linked to immunoglobulin suppression) might mitigate inflammation through conjugated linoleic acid supplementation. Collectively, these metabolite-physiological networks provide a preliminary molecular blueprint for stage-optimized feeding programs with further research needed to validate which metabolite changes translate to long-term growth, immune competence, or metabolic resilience.

This study is subject to several inherent limitations that warrant consideration. The analytical framework, constrained by three selected time points, provides insightful yet potentially discontinuous snapshots of metabolic dynamics. Furthermore, the sample size, while adequate for a primary metabolomic survey, may limit the extrapolation of findings and constrain sophisticated statistical modeling. Critically, the identified correlations between serum metabolites and physiological parameters are associative and do not imply causality. Although these findings generate compelling hypotheses regarding metabolic priorities during development, definitive functional validation and assessment of the predictive utility of these candidate biomarkers necessitate future investigations with expanded cohorts, denser temporal sampling, and interventional designs.

Conclusion

This study deciphers stage-specific metabolic reprogramming during pigeon development, revealing TCA intermediates correlate with growth, while purine metabolites align with immune modulation. Their inverse correlations demonstrate a resource trade-off between anabolism and immunity. Key metabolites such as citric acid, α-ketoglutarate, guanosine serve as dual biomarkers with causal links yet to be established. These findings provide a benchmark for future research into metabolic regulation mechanisms.

CRediT authorship contribution statement

Yuanfei Li: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Qiao Xu: Resources, Investigation, Funding acquisition, Conceptualization. Changzheng Guo: Validation, Methodology, Formal analysis. Jishang Gong: Investigation, Data curation, Conceptualization. Keke Che: Visualization, Resources. Zheng Jiao: Visualization, Methodology. Xuewen Chai: Resources. Rendian Zhang: Investigation. Jiguo Xu: Supervision. Xinwei Xiong: Resources.

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.