Strontium ameliorates tibial dyschondroplasia in broilers through gut microbiota-mediated retinoic acid metabolism Authors: Shu-cheng Huang, Bo-wen Xu, Ya-nan Lu, Tao-jing Yue, Wen-li Ding, Fang-fang Guo, Zi-qun Zhang, Xi-shuai Tong, Cai Zhang, Li Wang Published: 2026 ABSTRACT Tibial dyschondroplasia (TD) is a prevalent skeletal disorder in fast-growing broilers that impairs animal welfare and production efficiency. Although strontium (Sr) is known to support bone-strengthening properties, its protective mechanisms against TD, particularly those involving the gut‒bone axis, remain unclear. This study showed that dietary Sr supplementation (200 mg/kg) significantly improved growth performance, restored the tibial microstructure, corrected plasma calcium/phosphorus imbalances, enhanced chondrocyte alignment, increased the osteoblast density, increased the bone mineral density, and improved vascularization in the proximal tibial growth plate in TD broilers. Moreover, the Sr intervention upregulated the bone morphogenetic protein-mediated osteogenic pathway while suppressing RANKL activity and promoted angiogenesis by activating the VEGFA and PI3K/AKT/eNOS signaling pathways, thereby restoring bone remodeling. Interestingly, Sr remodeled the gut microbiota of TD broilers, enriching the beneficial species Pediococcus pentosaceus and reducing the abundance of Levilactobacillus brevis, changes that were correlated with the level of the osteotoxic metabolite 9-cis-retinoic acid (9-cis-RA). Furthermore, exogenous retinoic acid (RA) administration induced TD-like damage, whereas Sr cotreatment mitigated the RA-induced tibial impairment, vascular defects, and dysregulation of the PI3K/AKT/eNOS pathway. Collectively, these results demonstrated that Sr alleviated TD by modulating the gut microbiota to inhibit 9-cis-RA accumulation, thereby normalizing retinoic acid metabolism, promoting vascularization and enhancing osteogenesis; these findings provide novel insights for preventing leg disorders in poultry. Introduction Broiler chickens, a primary global source of dietary protein, have been intensively bred for rapid growth to meet the increasing demand. While this strategy, alongside high-density rearing, has successfully shortened growth cycles and increased weight gain, the accelerated skeletal muscle development often exceeds skeletal maturation. This disparity causes bones to experience excessive mechanical stress, significantly increasing their susceptibility to tibial dyschondroplasia (TD)1–3. TD is a prevalent nutrient-metabolic bone disorder in fast-growing broilers and is pathologically characterized by avascular, opaque white cartilage plugs in the proximal tibial growth plate (TGP), leading to lameness and potential paralysis2,4,5. This disorder stems from a disruption in endochondral ossification, where chondrocytes fail to undergo normal differentiation, maturation, and calcification, thereby initiating skeletal abnormalities6,7. Since it was first reported in 1965, TD has remained a major threat to poultry health worldwide. Leg diseases affect approximately 12.5 billion birds annually, with TD accounting for up to 30% of cases in broilers8–10; in China, its incidence ranges from 15% to 30%10. Research on TD has evolved from traditional pathological characterization toward elucidating its molecular mechanisms and developing comprehensive interventions. Current investigations focus on genetic susceptibility, nutritional and metabolic imbalances, oxidative stress regulation, and environmental interactions5,11–15. Given the central role of broilers in human nutrition and the substantial economic loss caused by TD, elucidating its pathogenesis and developing effective interventions are paramount for sustainable poultry production. Trace elements, which are present at minute concentrations, are indispensable for skeletal homeostasis. They function synergistically to regulate mineral metabolism, modulate osteoclast (OC) activity, facilitate collagen synthesis, and provide antioxidant protection. Strontium (Sr), an essential trace element, has garnered significant research interest because of its multifaceted biological effects on osteogenesis, angiogenesis, and inflammation suppression16,17. Given its ionic radius, which is similar to that of calcium (Ca), Sr can enter cells via Ca channels and be incorporated into the bone matrix, partially substituting for Ca to regulate bone metabolic homeostasis18. Mechanistically, Sr inhibits key osteoclast enzymes such as tartrate-resistant acid phosphatase (TRAP) and carbonic anhydrase II (CA II), thereby reducing bone resorption and microstructural porosity. Concurrently, it increases the mechanical strength of bone through two effects: direct modulation of osteoclast activity and optimization of mineral phase structure ionic substitution19–21. Furthermore, Sr promotes bone vascularization by upregulating vascular-endothelial growth factor (VEGF), which stimulates endothelial cell proliferation and migration22–25. Despite these well-documented bone-protective properties, the specific efficacy of Sr in mitigating TD in broilers, along with its underlying mechanisms, remain to be elucidated17,26–29. The gut microbiota, the largest microecosystem in animals, influences systemic health, including bone diseases30, through secondary metabolite secretion and circulation-mediated crosstalk with distant organs. This interconnection forms the basis of the various gut–organ axes, making the gut microbiota a promising therapeutic target for poultry bone diseases. Accordingly, this study applied Sr intervention in a broiler model of TD. By performing targeted molecular profiling, multiomics, and histopathological analyses, we aimed to (i) systematically delineate the Sr-mediated regulation of key signaling pathways in TD pathogenesis; (ii) decipher gut‒bone crosstalk mechanisms through microbiome‒metabolism networks; and (iii) characterize spatiotemporal metabolite dynamics during bone development. These investigations will elucidate how Sr protects against tibial injury via the gut‒bone axis and provide novel insights into trace element-based strategies for optimizing bone health in poultry. Results Sr intervention improves growth performance and tibial health in TD broilers We induced a TD model in chickens with thiram exposure to investigate the effects of Sr intervention on growth performance and tibial health (Fig. 1A), and TD broilers exhibited the typical lameness posture and pathological widening of the TGP2,6,31 (Fig. 1E). First, the safety of Sr was assessed, and the results showed that the Sr intervention decreased the number of fat droplets to varying extents compared with that in the TD group (Supplementary Fig. 1A–E) and was coupled with hepatocyte fragmentation and a reduced number of hepatocytes, as shown by the histopathological examination of the liver. Sr treatment also reduced the plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) activities (p < 0.001), confirming that the tested Sr dosage did not induce hepatic damage (Supplementary Fig. 1F). Tibial dissection revealed that 7 days of Sr treatment alleviated the symptoms of shortened bone length, a thinned diaphysis, thickened growth plates, a lack of calcification, and avascular white cartilage plugs at the proximal TGP in TD broilers, and it reduced knee swelling and restored the tibial length and weight (Fig. 1F–I). Specifically, after 14 days of Sr administration, the broilers’ mobility recovered, with body weight and tibial parameters tending toward normal (Fig. 1F–I). Quantitative analyses revealed that the change in the area under the curve (AUC) of daily feed intake in the TD group was significantly reversed by Sr treatment in a dose-dependent manner (p = 0.006 for the Sr200 group compared with the TD group; Fig. 1B), which was consistent with the changes in broilers’ body weight (Day 14: p < 0.003; Fig. 1D), tibia weight (p < 0.001 (Day 7) and p = 0.002 (Day 14), respectively; Sr200 group compared with the TD group; Fig. 1F), tibia length (all p < 0.001; Sr200 group compared with the TD group; Fig. 1G), and tibia diameter (p < 0.001 (Day 7) and p < 0.005 (Day 14), respectively; Sr200 group compared with the TD group; Fig. 1H). In contrast, the feed conversion ratio (FCR) and TGP width (p < 0.001, TD group compared with the Sr200 group) exhibited the opposite patterns (Fig. 1C, I). We next investigated the mineralization status of the tibiae and found that the tibiae of the TD group presented a pale ash color (Fig. 1J). Notably, the results revealed that Sr treatment improved the abnormal Ca (p < 0.001 on Day 14 in the Sr200 group) and P (p < 0.001 on Day 7) contents and the ash color caused by TD and re-established the appropriate Ca/P ratio (Ca/P ratio, p < 0.001), indicating that Sr can increase bone mass by correcting mineralization-related metabolic disorders (Fig. 1K). In addition, we also found that plasma biochemical parameters were linked to changes in the bone ash in the TD group, in which Sr restored the normal levels of Ca, P, and ALP in the blood of the TD group (p < 0.001 on Day 14; Sr200 group compared with the TD group), indicating that Sr coordinates systemic mineral metabolism and local bone formation to maintain bone homeostasis (Fig. 1L). Collectively, these results confirm that the Sr intervention can counteract pathological changes, such as impaired feed intake, reduced feed efficiency, and tibial damage, observed in TD broiler chickens and can restore locomotor function and skeletal morphology while improving growth performance, suggesting the potential of Sr as a therapeutic agent for TD.Fig. 1Sr supplementation improves growth performance and tibial health in TD broilers.A Experimental design and timeline. B Quantification of feed intake and area under the curve (0–14 days; n = 3). C Feed conversion ratio (FCR) at d7/d14 (n = 8). D Body weight gain (BWG) at d7/d14 (n = 8). E Representative images of tibial morphology and growth plate lesions. F Tibial weight (n = 8). G Tibial length (n = 8). H Tibial diameter (n = 8). I TGP width (n = 8). J Tibial ash content (n = 3). K Tibial ash content, Ca content, P content and the Ca/P ratio (n = 3). L Plasma Ca levels, P levels, Ca/P ratio, and ALP levels (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). Sr supplementation improves growth performance and tibial health in TD broilers. A Experimental design and timeline. B Quantification of feed intake and area under the curve (0–14 days; n = 3). C Feed conversion ratio (FCR) at d7/d14 (n = 8). D Body weight gain (BWG) at d7/d14 (n = 8). E Representative images of tibial morphology and growth plate lesions. F Tibial weight (n = 8). G Tibial length (n = 8). H Tibial diameter (n = 8). I TGP width (n = 8). J Tibial ash content (n = 3). K Tibial ash content, Ca content, P content and the Ca/P ratio (n = 3). L Plasma Ca levels, P levels, Ca/P ratio, and ALP levels (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). The Sr intervention improves bone tissue development in TD broilers We directly tested whether the Sr intervention affects bone tissue development in TD broilers by performing a histopathological analysis via H&E staining of the proximal tibiae from the Sr intervention group on Days 7 and 14, with the TD broiler group used as a control. Strikingly, compared with the TD group, the Sr intervention group presented a restored chondrocyte morphology, a reconstructed vascular distribution in the proximal TGP, and a significantly lower proportion of abnormal chondrocytes, such as those undergoing karyolysis, cytolysis, and vacuolization (Sr200 group: p < 0.001 on Day 14; Fig. 2A, D, E), indicating that Sr not only repaired the histological structure of bone cells but also increased the vascular distribution in the tibia. Similarly, toluidine blue (TB) and safranin O–fast green (SO–FG) staining further showed the dynamic balance between chondrocytes and osteoblasts (Fig. 2B, C). Compared with TD treatment, Sr treatment significantly reduced the cartilage area (p < 0.001 for the Sr200 group compared with the TD group), increased the number and proportion of osteoblasts (p < 0.001), and normalized the degree of cellular alignment, confirming that Sr can restore the bone-forming capacity of TD broilers through the regulation of chondrocyte and osteoblast compositions and functions.Fig. 2The Sr intervention mitigates TD pathology and enhances bone development in broilers.A H&E staining of the proximal TGP at 7/14 days (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates cytolysis; the red arrow indicates chondrocyte vacuolization. B TB staining of the TGP (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrows indicate nucleolysis, the red arrows indicate cell vacuolization, and the blue arrows indicate cytolysis. C SO/FG staining of the TGP (n = 3; scale bars = 800 μm, 40 μm, 10 μm). The green arrows indicate normal osteoblasts, and the yellow arrows indicate osteoblasts with a structural disruption that have become flattened and exhibit a disordered arrangement. D, E Quantification of normal cells, chondrocytes, and osteoblasts in histomorphometric regions at Days 7 and 14 (n = 3). F Representative 3D micro-CT reconstructions of proximal tibiae on Day 14 (n = 3). G Microarchitectural parameters of the bone mineral density (BMD), bone surface/bone volume (BS/BV), connectivity density (Conn.Dn), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp) on Day 14 (n = 3). H qRT‒PCR analysis of BMP-2, OPG, RUNX2, and RANKL expression (n = 3). I western blot bands showing OPG, RUNX2, and RANKL protein expression (n = 3). J Densitometric quantification of OPG, RUNX2, and RANKL expression. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). The Sr intervention mitigates TD pathology and enhances bone development in broilers. A H&E staining of the proximal TGP at 7/14 days (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates cytolysis; the red arrow indicates chondrocyte vacuolization. B TB staining of the TGP (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrows indicate nucleolysis, the red arrows indicate cell vacuolization, and the blue arrows indicate cytolysis. C SO/FG staining of the TGP (n = 3; scale bars = 800 μm, 40 μm, 10 μm). The green arrows indicate normal osteoblasts, and the yellow arrows indicate osteoblasts with a structural disruption that have become flattened and exhibit a disordered arrangement. D, E Quantification of normal cells, chondrocytes, and osteoblasts in histomorphometric regions at Days 7 and 14 (n = 3). F Representative 3D micro-CT reconstructions of proximal tibiae on Day 14 (n = 3). G Microarchitectural parameters of the bone mineral density (BMD), bone surface/bone volume (BS/BV), connectivity density (Conn.Dn), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp) on Day 14 (n = 3). H qRT‒PCR analysis of BMP-2, OPG, RUNX2, and RANKL expression (n = 3). I western blot bands showing OPG, RUNX2, and RANKL protein expression (n = 3). J Densitometric quantification of OPG, RUNX2, and RANKL expression. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). In TD broilers, the number of trabeculae in the trabecular bone region of the tibia was significantly reduced (p < 0.001), as evidenced by the disrupted trabecular continuity, increased trabecular spacing, and decreased overall bone density (cancellous bone: p < 0.001); furthermore, the trabecular separation index was significantly increased (p < 0.001; Fig. 2F, G and Supplementary Fig. 2). Micro-CT imaging was employed to assess the effects of the Sr intervention on the histological structure of the tibial bone in 14-day-old broilers with TD. As expected, the Sr intervention, particularly at the 200 mg/kg dose (Sr200), significantly increased the trabecular count, connectivity density, and bone density in cancellous (p < 0.01) bone and reduced the trabecular spacing and separation (p < 0.001). These findings indicate that Sr effectively mitigates TD-induced bone loss and structural degradation, highlighting its critical role in inhibiting tibial bone disease progression. We also measured changes in the expression levels of the osteogenic genes BMP-2, OPG, and RUNX2 and the osteocalcin marker gene RANKL in the TGP between the TD group and the Sr intervention group. The results revealed that BMP-2, OPG, and RUNX2 were significantly downregulated (p < 0.01) and that RANKL was upregulated in the TD group (p < 0.001; Fig. 2H), whereas the Sr intervention, particularly the Sr200 intervention, markedly reversed these trends, significantly increasing OPG and RUNX2 levels (p < 0.001) and decreasing RANKL levels (p < 0.001). Protein analyses further confirmed that the OPG and RUNX2 abundances were reduced (p < 0.001) and that the RANKL level was abnormally increased (p < 0.012; Fig. 2I, J). These results indicate that Sr can correct the activity of the RANKL/RANK/OPG system, thereby reversing the imbalance in bone regulation. The ratio of OPG to RANKL (OPG/RANKL ratio) in TGP tissue was quantified to directly evaluate the regulatory effect on osteoclastogenesis. As shown in Supplementary Fig. 3, compared with the control group, the TD group exhibited a significantly reduced OPG/RANKL ratio at both the mRNA and protein levels (all p < 0.001). These results indicate that pro-osteoclastogenic signaling is predominant under TD conditions. Following Sr supplementation, the OPG/RANKL ratio was restored in a dose-dependent manner. In the Sr200 group, the OPG/RANKL ratios based on mRNA and protein levels were significantly increased compared with those in the TD group (p < 0.001), with the protein levels even returning to those in the control group. These observations indicate that Sr effectively reverses the pro-osteoclastogenic molecular microenvironment induced by TD. The Sr intervention ameliorates impaired proximal tibial angiogenesis in TD broilers The proximal tibial tissues were stained using Goldner’s trichrome method to evaluate the effect of the Sr intervention on vascular alterations in the tibiae of TD broilers. The results revealed that the Sr intervention partially restored the pathological manifestations of TD broiler chickens, such as a sparse, disorganized vasculature accompanied by erythrocyte lysis, rupture, and a loss of the normal boundary structure. Moreover, vascular development was restored in a dose-dependent manner in the Sr intervention group, with hemocytes maintaining healthy morphological and structural integrity (Fig. 3A). The results of the quantitative analysis revealed that the vascular distribution area was significantly smaller in the TD group than in the Con group (p < 0.001 on both Days 7 and 14). Sr treatment, particularly the Sr200 treatment, significantly increased the vascular distribution area (p < 0.001 on Day 7, p < 0.001 on Day 14 in the Sr200 group; Fig. 3B). Tibial immunofluorescence staining revealed that the relative densities of CD31 and CD34 in the TD group were significantly lower (p < 0.001 on Day 14), and after Sr200 treatment, the relative densities were significantly higher (p < 0.001 on Day 14) (Fig. 3C, D). These data indicate that TD can disrupt normal tibial angiogenesis and that Sr supplementation can effectively alleviate this pathology.Fig. 3The Sr intervention promotes tibial vascularization in broilers with thiram-induced TD.A Representative images of Goldner’s trichrome-stained sections of the proximal tibia at 7 and 14 days after treatment (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates broken erythrocytes, the red arrow indicates cell adhesion, and the black arrow indicates normal erythrocytes. B Quantification of the vessel area percentage per field (n = 3). C Immunofluorescence staining showing CD31/CD34-positive microvessels in tibial tissues (n = 3; scale bars = 1000 μm, 20 μm). D Integrated fluorescence density of CD31/CD34 signals (n = 3). E qRT‒PCR analysis of VEGFA and VEGFR2 mRNA expression (n = 3). F Representative western blots showing VEGFR2 and VEGFA expression. G Densitometric quantification of VEGFR2 and VEGFA protein levels (n = 3). H–J qRT‒PCR analysis of PI3K, AKT, and eNOS mRNA expression (n = 3). K Representative western blots showing the levels of p-eNOS, eNOS, p-PI3K, PI3K, p-AKT and AKT. L Densitometric quantification of eNOS, PI3K and AKT protein levels (n = 3). M Densitometric quantification of the p-eNOS, p-PI3K and p-AKT protein levels (n = 3). N The ratio of the level of the phosphorylated protein to the level of the total protein. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). The Sr intervention promotes tibial vascularization in broilers with thiram-induced TD. A Representative images of Goldner’s trichrome-stained sections of the proximal tibia at 7 and 14 days after treatment (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates broken erythrocytes, the red arrow indicates cell adhesion, and the black arrow indicates normal erythrocytes. B Quantification of the vessel area percentage per field (n = 3). C Immunofluorescence staining showing CD31/CD34-positive microvessels in tibial tissues (n = 3; scale bars = 1000 μm, 20 μm). D Integrated fluorescence density of CD31/CD34 signals (n = 3). E qRT‒PCR analysis of VEGFA and VEGFR2 mRNA expression (n = 3). F Representative western blots showing VEGFR2 and VEGFA expression. G Densitometric quantification of VEGFR2 and VEGFA protein levels (n = 3). H–J qRT‒PCR analysis of PI3K, AKT, and eNOS mRNA expression (n = 3). K Representative western blots showing the levels of p-eNOS, eNOS, p-PI3K, PI3K, p-AKT and AKT. L Densitometric quantification of eNOS, PI3K and AKT protein levels (n = 3). M Densitometric quantification of the p-eNOS, p-PI3K and p-AKT protein levels (n = 3). N The ratio of the level of the phosphorylated protein to the level of the total protein. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (Sr200 and Sr400 groups vs. the TD group). We performed a qRT‒PCR analysis of VEGFA and VEGFR2 expression in TGP tissue on Day 14 to elucidate the molecular mechanisms underlying TD-associated angiogenic impairment and the restorative effects of Sr (Fig. 3E‒G). The expression levels of VEGFR2 (p = 0.016) and VEGFA (p = 0.383) in TD broiler chickens decreased, whereas they increased after Sr200 treatment (p < 0.024 and p < 0.001, respectively). This observation was further supported by the relative abundances of both proteins (p < 0.01). These results indicate that the Sr intervention may restore proximal tibial angiogenesis in TD broilers by reactivating the VEGFA/VEGFR2 signaling pathway. A previous study revealed that the PI3K/AKT/eNOS signaling pathway plays an important role in the abnormal development of blood vessels in TD broilers2. We further explored whether the PI3K/AKT/eNOS signaling pathway also plays a regulatory role in the effects of the Sr intervention by assessing the mRNA and protein expression levels of PI3K, AKT, and eNOS in the TGP of 14-day-old broiler chickens (Fig. 3H–L). The expression levels of AKT and eNOS in the TGP of TD broiler chickens were significantly lower (p < 0.001) than those in the Con group. Compared with those in the TD group, the expression levels of PI3K, AKT, and eNOS in the Sr-treated group were significantly higher (all p < 0.01), suggesting that Sr may increase eNOS transcriptional activity by activating the PI3K/AKT pathway. Western blot analysis further revealed that the protein levels of PI3K, AKT, and eNOS were significantly lower in the TD group than in the Con group (p < 0.001). Compared with those in the TD group, the PI3K, AKT and eNOS levels in the Sr intervention group were significantly higher (all p < 0.01). The phosphorylation levels of key proteins in this pathway were assessed to clarify the activation of the PI3K/AKT/eNOS signaling pathway (Fig. 3M, N). Compared with those in the control group, the p-PI3K/PI3K and p-AKT/AKT ratios in the TGP of the TD group were significantly decreased (p < 0.05 and p < 0.01, respectively). The subsequent Sr intervention induced varying degrees of recovery of these ratios, demonstrating that Sr effectively activated the PI3K/AKT/eNOS signaling pathway in the TGP of TD broilers. Pearson’s correlation analyses revealed a significant negative correlation (p < 0.01) between the expression of tibial vascularization-related markers (VEGFR2, eNOS, CD31, and CD34) and that of tibial health-related markers (RANKL, bone ash-P, and plasma-Ca), suggesting that bone health is strongly influenced by the vascularization status (Supplementary Fig. 4). Taken together, these results indicated that the Sr intervention alleviated the pathological process of TD by restoring PI3K/AKT/eNOS signaling activity, increasing eNOS expression, and promoting angiogenesis at the proximal TGP. The Sr intervention reverses gut microbiota dysbiosis in TD broilers We investigated the effect of the Sr intervention on the gut microbiota structure of broiler chickens by performing full-length 16S rRNA gene sequencing to systematically analyze the fecal microbial compositions of TD broilers treated with Sr for 14 days. As shown in Fig. 4A, the alpha diversity indices (Chao1, Simpson, Shannon, and observed species indices) of the TD broiler chickens in the Sr intervention group tended to decrease compared with those in the TD group. Principal coordinate analysis (PCoA) and nonmetric multidimensional scaling (NMDS) based on Bray‒Curtis distance matrices further showed distinct separations among groups in the two-dimensional space (Fig. 4B, C). Notably, the distributions of the samples from the Con and TD_Sr groups substantially overlapped, whereas those from the TD group formed independent clusters, indicating that, compared with the Con and TD_Sr groups, the Sr intervention significantly reversed the changes in the gut microbiota structure of TD broiler chickens (Fig. 4B, C). Collectively, these findings suggest that Sr supplementation partially reversed TD-induced dysbiosis in the gut microbiota. The top 10 phyla, top 15 genera, and top 20 species at the taxonomic level are shown in Fig. 4D. Specifically, at the phylum level (Fig. 4E), compared with the TD group, the TD_Sr group presented a significantly decreased relative abundance of Bacteroidota and increased Bacillota/Bacteroidota (Bacil/Bacte) ratio (p < 0.001 and p = 0.005, respectively). Similar restorative effects were observed in the TD_Sr group at the genus and species levels, where TD-induced aberrations in major bacterial populations were reversed.Fig. 4The Sr intervention restores gut microbiota homeostasis in broilers with TD.A Alpha diversity indices: Chao1 index and observed species indices (richness) and Shannon and Simpson indices (diversity). B Beta diversity analyzed by principal coordinate analysis (PCoA) based on of Bray‒Curtis dissimilarity. C Nonmetric multidimensional sorting (NMDS) of beta diversity via dissimilarity matrices. D Microbial composition shifts at the phylum (top 10), genus (top 15), and species (top 20) levels. E Relative abundances of Bacillota and Bacteroidota and the Bacillota/Bacteroidota ratio. F Intergroup discriminant features identified by linear discriminant analysis effect sizes (LEfSe). G Species-level biomarker ranking via random forest analysis (top 20 features). H Venn diagram of differentially abundant species (intersection: top 20 random forest biomarkers and top 20 LEfSe-discriminant species). I Comparison of the relative abundances of different species. The groups were compared using the Kruskal–Wallis nonparametric test, and two groups were compared using Dunn’s test. #p < 0.05, ##p < 0.01, and ###p < 0.001 (Con group vs. the TD group); *p < 0.05, **p < 0.01, and ***p < 0.001 (TD_Sr group vs. the TD group). The Sr intervention restores gut microbiota homeostasis in broilers with TD. A Alpha diversity indices: Chao1 index and observed species indices (richness) and Shannon and Simpson indices (diversity). B Beta diversity analyzed by principal coordinate analysis (PCoA) based on of Bray‒Curtis dissimilarity. C Nonmetric multidimensional sorting (NMDS) of beta diversity via dissimilarity matrices. D Microbial composition shifts at the phylum (top 10), genus (top 15), and species (top 20) levels. E Relative abundances of Bacillota and Bacteroidota and the Bacillota/Bacteroidota ratio. F Intergroup discriminant features identified by linear discriminant analysis effect sizes (LEfSe). G Species-level biomarker ranking via random forest analysis (top 20 features). H Venn diagram of differentially abundant species (intersection: top 20 random forest biomarkers and top 20 LEfSe-discriminant species). I Comparison of the relative abundances of different species. The groups were compared using the Kruskal–Wallis nonparametric test, and two groups were compared using Dunn’s test. #p < 0.05, ##p < 0.01, and ###p < 0.001 (Con group vs. the TD group); *p < 0.05, **p < 0.01, and ***p < 0.001 (TD_Sr group vs. the TD group). We further investigated the differences in the abundances of gut microorganisms among the different treatment groups and identified differentially abundant taxa (LDA > 2, p < 0.05) via linear discriminant analysis effect size (LEfSe). As shown in Fig. 4F, the Con group was enriched in Pseudomonadota, Enterobacterales, Gammaproteobacteria, Enterobacteriaceae, Shigella flexneri, Pedobacter, Pedobacter sp_An13, Klebsiella pneumoniae, Klebsiella, Bacillus sp_RH219, Sphingobacterium_sp_, Micrococcales, Microbacteriaceae, Microbacterium and Microbacterium testaceum; the TD group featured increased abundances of Desulfonatronospira, Desulfonatronovibrionaceae, Desulfonatronospira thiodismutans, Clostridium_sp_TG60_81, Brevibacillus_sp__20B, Heliobacteriaceae, Arcobacter_sp__HL735A, Arcobacteraceae, and Arcobacter; and the TD_Sr group was characterized by elevated levels of Bacillota, Levilactobacillus brevis, Levilactobacillus, Leptolyngbyaceae, Phormidesmis, Pseudanabaenales, Phormidesmis priestleyi, Oxalobacteraceae, Herbaspirillum_sp__PIV_34_1, and Herbaspirillum. Random forest analysis identified the top 20 species-level biomarkers, with the most important species including Arcobacter sp. HL735A, Shigella flexneri, Escherichia coli, Microbacterium testaceum, and Blautia wexlerae (Fig. 4G). A Venn diagram showing the overlap of the top 20 abundant species and random forest-identified biomarkers revealed 12 strains, including Shigella flexneri, Escherichia coli, Blautia wexlerae, Oceanobacillus picturae, Pediococcus pentosaceus, Levilactobacillus brevis, Pedobacter sp. An13, Latilactobacillus curvatus, Lucifera butyrica, Ligilactobacillus salivarius, Bacteroides sp. and Enterococcus faecium (Fig. 4H). Additional statistical comparisons revealed that the abundances of Pediococcus pentosaceus and Levilactobacillus brevis differed significantly (Fig. 4I). Specifically, the relative abundance of Pediococcus pentosaceus was significantly higher (p < 0.05) and that of Levilactobacillus brevis was significantly lower (p < 0.05) in the TD_Sr group than in the TD group. These integrated analyses demonstrated that Sr supplementation effectively reversed the TD-induced disruption of intestinal microbiota homeostasis by modulating the abundances of key bacterial populations. Sr improves tibial health in TD broilers through the gut microbiota A fecal microbiota transplantation (FMT) experiment was conducted to verify whether the gut microbiota is involved in the protective effect of Sr on the tibia of TD chickens. During mixed antibiotic treatment, body weight gain was transiently inhibited in both groups of mice, with gradual recovery following the cessation of antibiotic intervention (Fig. 5A, B). Upon FMT, although both groups maintained positive weight gain, the FMT-Sr group (Sr-treated FMT group) presented a more stable growth rate and significantly greater cumulative body weight gain than the Con group. Further analysis of plasma metabolic indices revealed significantly elevated Ca levels in the FMT-Sr group compared with those in the Con group (p = 0.007). P concentrations tended to increase, whereas ALP activity was significantly increased (p < 0.001) (Fig. 5C–G). These findings suggest that the gut microbiota may promote bone development by regulating mineralization and metabolism. Compared with that of control mice, the skeletal morphology of FMT-Sr-treated mice was stronger, with significantly increased bone density observed via micro-CT imaging (Fig. 5H). Moreover, the cancellous bone density (p = 0.042), bone volume (p = 0.036) and bone surface area (p = 0.003) were higher in the FMT-Sr group than in the Con group. Similar to the effect on density, the FMT-Sr group presented a significant increase in the number of cancellous bone trabeculae (p < 0.001) and reduced trabecular separation (p = 0.027 for cortical bone; p = 0.004 for cancellous bone), indicating that the gut microbiota derived from Sr-supplemented broilers enhances trabecular structural continuity by optimizing the three-dimensional network connectivity in murine recipients (Fig. 5I). H&E staining revealed a significantly greater proportion of bone trabeculae in the proximal tibia cancellous zone in the FMT-Sr group (p = 0.019), with improved three-dimensional thickness and alignment density, suggesting that intestinal microbiota remodeling may regulate the reconstruction of the bone microstructure via the “gut‒bone axis” (Fig. 5J). SO/FG staining further revealed that the number of osteoblasts per unit of bone surface (Ob.S/BS) in the FMT-Sr group was greater than that in the Con group (p = 0.072), with an orderly arrangement of the cells and a balanced nuclear‒cytoplasmic ratio, which implies that gut microbiota remodeling can potentially facilitate osteoblast recruitment via microbial metabolites or signaling molecules in murine models (Fig. 5K). Immunohistochemical staining confirmed significantly increased osteocalcin (OCN) immunoreactivity in tibial tissues from the FMT-Sr group (p = 0.006), indicating increased osteogenic activation (Fig. 5L). qRT‒PCR revealed the significant upregulation of the osteogenic markers RUNX2 and OPG (both p < 0.01) and significant downregulation of the osteoblastic activator RANKL (p < 0.001) in the FMT-Sr group compared with the Con group. The expression of the vascularization-related gene eNOS was significantly increased (p = 0.007), and VEGFA expression tended to increase (p = 0.093; Fig. 5M). Collectively, these data establish that the Sr-modified broiler microbiota orchestrates bone protection in murine recipients through the synergistic modulation of angiogenesis‒osteogenic crosstalk, suggesting that Sr exerts a bone-protective effect by driving changes in the gut microbiota.Fig. 5Fecal microbiota transplantation from Sr-supplemented broilers confers osteoprotection in mice.A Flowchart of the experimental design. B Body weight dynamics and area under the curve from 028 days (n = 3). C Body weight gain (n = 10). D‒G Plasma Ca levels, P levels, Ca/P ratios, and ALP levels (n = 3). H Micro-CT reconstruction of the femora. I Bone morphometric indices of the bone mineral density (BMD), bone surface/bone volume (BS/BV), trabecular number (Tb.N), and trabecular separation (Tb.Sp). J H&E-stained distal femur sections with the quantification of the trabecular bone area (n = 3; scale bars = 400 μm, 40 μm). K Osteoblast localization by SO/FG staining (n = 3; scale bars = 400 μm, 40 μm) and quantification. L Immunohistochemical detection of OCN in femora (n = 3; scale bars = 400 μm, 40 μm). M qRT‒PCR analysis of OPG, RUNX2, RANKL, VEGFA and eNOS mRNA expression (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: *p < 0.05, **p < 0.01, and ***p < 0.001 (Con group vs. the FMT-Sr group). Fecal microbiota transplantation from Sr-supplemented broilers confers osteoprotection in mice. A Flowchart of the experimental design. B Body weight dynamics and area under the curve from 028 days (n = 3). C Body weight gain (n = 10). D‒G Plasma Ca levels, P levels, Ca/P ratios, and ALP levels (n = 3). H Micro-CT reconstruction of the femora. I Bone morphometric indices of the bone mineral density (BMD), bone surface/bone volume (BS/BV), trabecular number (Tb.N), and trabecular separation (Tb.Sp). J H&E-stained distal femur sections with the quantification of the trabecular bone area (n = 3; scale bars = 400 μm, 40 μm). K Osteoblast localization by SO/FG staining (n = 3; scale bars = 400 μm, 40 μm) and quantification. L Immunohistochemical detection of OCN in femora (n = 3; scale bars = 400 μm, 40 μm). M qRT‒PCR analysis of OPG, RUNX2, RANKL, VEGFA and eNOS mRNA expression (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: *p < 0.05, **p < 0.01, and ***p < 0.001 (Con group vs. the FMT-Sr group). The Sr intervention remodels retinoic acid metabolism in TD broilers The plasma metabolomic analysis revealed the significant separation of metabolite distributions between the Con and TD groups and between the TD and strontium intervention (TD-Sr) groups via a standardized scoring analysis (Z score) (Fig. 6A). A total of 5,568 metabolites were identified in positive/negative ion modes. Among the 5568 metabolites, 907 (424 upregulated and 483 downregulated; |Log2FC | ≥ 1, p < 0.05) and 526 (410 upregulated and 116 downregulated; |Log2FC | ≥ 1, p < 0.05) differentially abundant metabolites were identified in the Con vs. TD groups and the TD vs. TD-Sr groups, respectively (Fig. 6B). The KEGG pathway enrichment analysis revealed that differentially abundant metabolites were coenriched in the PPAR signaling pathway, IgA production in the intestinal immune network, aromatic amino acid biosynthesis, and linoleic acid metabolism pathways (Fig. 6C, D). A further focus on key metabolites revealed that 9-cis-RA, arachidonic acid, and linoleic acid play central regulatory roles in multiple pathways (Fig. 6E). An integrated random forest analysis revealed the top 10 metabolic markers of importance in the comparisons of the Con vs. TD groups and the TD vs. TD-Sr groups, and Venn diagrams showed that 9-cis-RA was the core marker common to both comparisons (Fig. 6E–H). Compared with those in the Con group, plasma 9-cis-RA levels were significantly increased in the TD group (p < 0.01) but significantly decreased after the Sr intervention (p < 0.01), suggesting that 9-cis-RA serves as a key metabolic hub in TD pathogenesis and treatment (Fig. 6I). A systematic association analysis revealed that 9-cis-RA was significantly positively correlated with the abundance of intestinal probiotics (Pediococcus pentosaceus and Levilactobacillus brevis) and the expression of the bone resorption gene RANKL (p < 0.05) and negatively correlated with the expression of the osteogenic genes OPG, RUNX2, and ALP (p < 0.05; Fig. 6J). The ternary analysis further confirmed abnormal 9-cis-RA accumulation in the TD group, which was closely associated with an imbalance in the bone microenvironment (Fig. 6K). Notably, the mRNA expression of the specific 9-cis-RA receptor RXR did not significantly differ among the TD, Con, and TD-Sr groups, suggesting that although 9-cis-RA accumulates in vivo, it is not actively metabolized in TD broilers, indicating that factors driving its abnormal accumulation may contribute to TD pathogenesis. Previous studies have shown that 9-cis-RA acts as an RXR ligand, forming a heterodimer with RA-activated RAR to amplify the transcription of bone resorption-related genes (e.g., RANKL) and that RA competitively binds to the nuclear receptor RXR, whose metabolic disorders are likely to affect 9-cis-RA metabolism32,33. In our study, the variation in 9-cis-RA levels may be competitively regulated by RA. Therefore, we further explored the expression of key genes in the RA metabolic pathway in this study (Fig. 6L). The results revealed that the expression of the RA degradation gene Cyp26a1 (p = 0.043) was significantly reduced in the TD group but increased after Sr200 treatment (p < 0.001; Fig. 6L). Other markers were significantly upregulated in the TD group (except for Aldh1a2 and Aldh1a3, p < 0.001) but were downregulated to varying degrees after the Sr200 intervention (RAR, p < 0.001; Adh1, p = 0.031; Rdh7, p = 0.028; Aldh1a1, p < 0.001; Aldh1a3, p = 0.017). The above results indicate that the abnormal regulation of the RA metabolic pathway leads to changes in 9-cis-RA levels and that the Sr200 intervention can effectively reverse these metabolic abnormalities, suggesting that abnormal RA metabolism may induce TD in broilers.Fig. 6The Sr intervention reshapes the blood metabolome in TD broilers and the identification of key metabolites.A Z score-normalized metabolite distribution. B Volcano plot of differentially abundant metabolites (DAMs; |log2FC | ≥ 1, p < 0.05). C Sankey bubble plots showing the top 10 enriched pathways (bubble size = metabolite count, color = p value). D Venn diagram of DAMs associated with different pathways. E Heatmap of the DAMs enriched in the top pathways. F, G Random forest analysis of the top 10 discriminatory metabolites in the Con group vs. the TD group and the TD group vs. the TD-Sr group. H Venn diagram of the overlapping top-ranked metabolites. I Quantification of key DAMs. J Correlation network of 9-cis-RA with differential genera, vascular markers (CD31 and VEGFA), and bone health parameters (RANKL and OPG). Node connections: line thickness = correlation strength; color = significance (red: p < 0.01). K Ternary diagram showing the enrichment of 9 metabolic indices. L qPCR analysis of RA metabolism-related expression of the RXR, RAR, Adh1, Rdh7, Cyp26a1, Aldh1a1, Aldh1a2 and Aldh1a3 mRNAs (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (TD_Sr group vs. the TD group). The Sr intervention reshapes the blood metabolome in TD broilers and the identification of key metabolites. A Z score-normalized metabolite distribution. B Volcano plot of differentially abundant metabolites (DAMs; |log2FC | ≥ 1, p < 0.05). C Sankey bubble plots showing the top 10 enriched pathways (bubble size = metabolite count, color = p value). D Venn diagram of DAMs associated with different pathways. E Heatmap of the DAMs enriched in the top pathways. F, G Random forest analysis of the top 10 discriminatory metabolites in the Con group vs. the TD group and the TD group vs. the TD-Sr group. H Venn diagram of the overlapping top-ranked metabolites. I Quantification of key DAMs. J Correlation network of 9-cis-RA with differential genera, vascular markers (CD31 and VEGFA), and bone health parameters (RANKL and OPG). Node connections: line thickness = correlation strength; color = significance (red: p < 0.01). K Ternary diagram showing the enrichment of 9 metabolic indices. L qPCR analysis of RA metabolism-related expression of the RXR, RAR, Adh1, Rdh7, Cyp26a1, Aldh1a1, Aldh1a2 and Aldh1a3 mRNAs (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (TD group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (TD_Sr group vs. the TD group). Effects of RA on the growth performance and tibial health of broilers We fed broilers RA to assess tibial lesion development and to further explore the mechanism and role of RA in broiler leg pathologies (Fig. 7A). The RA intervention caused significant motor dysfunction in broilers, which manifested as unsteady standing, walking difficulties, or complete immobility, along with a depressed demeanor, reduced food intake, and delayed body weight gain (Fig. 7D). The morphological results revealed that the tibia length of the broilers in the RA30 group was significantly shorter than that in the Con group. Sr (Sr200) treatment significantly improved motor ability and restored the tibial length to near-control levels (Fig. 7D, G). A quantitative analysis of growth performance revealed that the feed intake of the RA30 group was significantly lower than that of the Con group (p < 0.001), whereas the feed intake of the RA30 + Sr200 group significantly rebounded (p < 0.001; Fig. 7B–C). In terms of body weight gain, the cumulative body weight gain in the RA30 group was significantly lower than that in the Con group (p < 0.001), whereas the Sr treatment tended to increase body weight gain but the difference was not statistically significant (Fig. 7E). The feed conversion ratio differed significantly among the groups, with the Con and RA30 + Sr200 groups showing more concentrated data (Fig. 7F). Compared with the Con group, the RA30 group presented significantly reduced tibia length, diameter and weight (p = 0.006, p = 0.009, and p = 0.026, respectively), whereas these parameters were significantly restored after Sr treatment (p = 0.031; Fig. 7G–I). After body weight correction, the tibial weight index (tibial weight/body weight ratio) in the RA+Sr group was significantly higher than that in the RA-only group (p < 0.001; Supplementary Fig. 5). Notably, the width of the growth plate was significantly wider in the RA30 group than in the Con group (p = 0.029), and this indicator tended to decrease after Sr treatment (Fig. 7J). The histopathological analysis further revealed that the cartilage cells of the proximal tibia in the Con group were neatly aligned and that the blood vessels were densely distributed, whereas those in the RA-exposed group exhibited cell vacuolization, disorganized alignment, and a reduced blood vessel density, which was consistent with the results observed in TD broilers (Fig. 7K). The proportion of normal chondrocytes in the RA30 group was significantly lower than that in the Con group (p < 0.001). Conversely, the proportion of normal chondrocytes significantly recovered following Sr treatment (p < 0.001; Fig. 7L). TRAP staining was performed on bone tissue from the proximal tibia to directly assess bone resorption activity (Supplementary Fig. 6). The quantitative analysis revealed that compared with that in the control group, the osteoclast density on the bone surface (Oc.S/BS) was significantly increased in the RA30 group (p < 0.001). Sr supplementation markedly suppressed osteoclast accumulation, with a significant decrease in the Oc.S/BS in the RA30 + Sr200 group compared with that in the RA group (p < 0.001). These results indicate that RA disrupts skeletal development in broilers by inhibiting feeding, interfering with TGP homeostasis, and inducing chondrocyte damage, and that Sr supplementation reverses the motor dysfunction caused by RA and restores tibial morphology and cytoarchitecture.Fig. 7Detrimental effects of RA on tibial health in broilers.A Experimental design and timeline. B, C Feed intake and area under the curve in broilers from 021 days (n = 3). D Representative clinical signs and tibial length phenotypes on Day 21. E‒I Body weight gain, feed conversion ratio, tibial length, tibial diameter, and tibial weight on Day 21 (n = 10). J TGP morphology on Day 21 (n = 10). K Histopathology of the proximal tibia assessed using H&E staining on Day 21 (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates nucleolysis. L Quantification of normal chondrocytes in the microscopic field (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). Detrimental effects of RA on tibial health in broilers. A Experimental design and timeline. B, C Feed intake and area under the curve in broilers from 021 days (n = 3). D Representative clinical signs and tibial length phenotypes on Day 21. E‒I Body weight gain, feed conversion ratio, tibial length, tibial diameter, and tibial weight on Day 21 (n = 10). J TGP morphology on Day 21 (n = 10). K Histopathology of the proximal tibia assessed using H&E staining on Day 21 (n = 3; scale bars = 800 μm, 40 μm, and 10 μm). The orange arrow indicates nucleolysis. L Quantification of normal chondrocytes in the microscopic field (n = 3). The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). RA induces tibia mineralization defects and a systemic Ca‒P imbalance in broilers We next analyzed the effect of RA on the microstructure of broiler tibial tissue. SO/FG staining revealed that the number of osteoblasts in the RA30-exposed group was significantly lower than that in the Con group (p < 0.001), accompanied by pathological changes such as a disorganized arrangement, blurred boundaries, and cellular adhesions (Fig. 8A). Following strontium (Sr200) treatment, the osteoblast count in the RA30 + Sr200 group significantly rebounded compared with that in the RA30 group (p < 0.001), with the cellular arrangement and morphology approaching normal (Fig. 8B). Moreover, the relative abundance of osteocalcin (OCN) in tibial tissues from the RA30 group was significantly lower than that in the Con group (p < 0.001), whereas the Sr intervention significantly restored its abundance (p < 0.001), indicating that RA disrupted bone matrix mineralization by inhibiting OCN expression and that Sr reversed this process (Fig. 8C, D). Characterization of the bone ash further revealed that the tibial ash from the groups exposed to different doses of RA appeared whiter than that from the Con and RA30 + Sr200 groups, suggesting a reduced vascular number (Fig. 8F). An analysis of the weight of the ash fractions revealed that the tibial ash fraction was significantly lower in the RA15 group (p < 0.05) and higher in the RA60 group (p < 0.001) than in the Con group. However, no significant differences were observed among the Con, RA30, and RA30 + Sr200 groups (Fig. 8G). An analysis of the components in the ash fractions revealed that the Ca and P contents in the tibial ash fraction from the RA30 group were significantly lower than those from the Con group (Ca: p < 0.001; P: p < 0.001), with a significantly decreased Ca/P ratio (p = 0.042). Sr supplementation significantly restored Ca and P levels (both p < 0.001), although the Ca/P ratio did not fully recover (Fig. 8H–J). We also found that the Ca level in the plasma of broiler chickens exposed to RA increased significantly (RA30 group: p < 0.01), whereas the P level decreased significantly (RA30 group: p < 0.001), and the Ca/P ratio increased significantly (p < 0.01; Fig. 8E, K–M). The Sr200 intervention significantly restored the plasma P content of the RA30-exposed group (p < 0.05), and the Ca/P ratio returned to normal (Fig. 8E, K–M). Additionally, the plasma ALP levels in the RA-exposed groups showed a significant decreasing trend compared with those in the Con group (RA30 group: p < 0.001) but significantly rebounded in the RA30 + Sr200 group (p < 0.001). In conclusion, RA disrupts skeletal homeostasis by inhibiting osteoblast function and OCN expression, interfering with bone mineral deposition, and disturbing systemic Ca/P metabolism. Conversely, Sr supplementation restored the osteoblast structure, restored OCN expression, promoted Ca/P deposition, and systematically reversed RA-induced pathological effects by regulating ALP activity and the plasma Ca/P balance.Fig. 8RA-induced tibial impairment and Sr-mediated rescue in broilers.A SO/FG staining of the proximal TGP (n = 3; scale bars = 800 μm, 40 μm, 10 μm). The black/orange arrows indicate normal/damaged osteoblasts. B Quantification of the osteoblast density (n = 3). C Immunohistochemical staining for OCN in the proximal tibia (Con, RA15, RA30, RA60, and RA30 + Sr200 groups; n = 3; scale bars = 400 μm, 40 μm). D OCN fluorescence intensity and area (n = 3). E Plasma ALP activity. F, G Tibial ash content, gross morphology, and ash fraction yield. H‒J Mineral composition of tibial ash in terms of Ca levels, P levels, and the Ca/P ratio. K‒M Plasma levels of the minerals Ca and P and the Ca/P ratio. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). RA-induced tibial impairment and Sr-mediated rescue in broilers. A SO/FG staining of the proximal TGP (n = 3; scale bars = 800 μm, 40 μm, 10 μm). The black/orange arrows indicate normal/damaged osteoblasts. B Quantification of the osteoblast density (n = 3). C Immunohistochemical staining for OCN in the proximal tibia (Con, RA15, RA30, RA60, and RA30 + Sr200 groups; n = 3; scale bars = 400 μm, 40 μm). D OCN fluorescence intensity and area (n = 3). E Plasma ALP activity. F, G Tibial ash content, gross morphology, and ash fraction yield. H‒J Mineral composition of tibial ash in terms of Ca levels, P levels, and the Ca/P ratio. K‒M Plasma levels of the minerals Ca and P and the Ca/P ratio. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). RA impairs tibial health through regulation of the PI3K/AKT/eNOS signaling pathway Given the crucial role of the PI3K/AKT/eNOS signaling pathway in tibial bone injury in TD chickens2, we investigated whether this pathway is regulated by RA. Here, we performed qRT‒PCR to analyze the mRNA expression levels of BMP-2, OPG, RUNX2, and RANKL in the TGP tissues of broiler chickens from all groups (Fig. 9A–D). Compared with those in the Con group, the expression of BMP-2 and OPG was significantly lower (both p < 0.001) and that of RANKL was significantly higher (p < 0.001) in the RA30-exposed group, whereas the BMP-2 and OPG mRNA levels in the RA30 + Sr200 group were significantly higher (p = 0.004 and p = 0.007, respectively), and the RANKL level was lower (p < 0.001) than those in the RA30-exposed group, suggesting that RA interferes with the homeostasis of bone remodeling, thereby exacerbating skeletal lesions. Conversely, Sr supplementation restored the dynamic balance of bone formation and resorption that was disrupted by RA exposure. We next analyzed the expression of vascular development-related factors in tibial tissues (Fig. 9E–I). Compared with those in the CON group, the expression levels of VEGFA (p < 0.001), VEGFR2 (p = 0.038), PI3K (p < 0.001), AKT (p < 0.001), and eNOS (p < 0.001) in the RA30 group were significantly lower. After supplementation with 200 mg/kg Sr, the expression of these genes increased (PI3K, p < 0.001; AKT, p < 0.001; eNOS, p < 0.001; VEGFA, p < 0.001; VEGFR2, p = 0.021), indicating that RA may impair endothelial cell function by inhibiting the PI3K/AKT/eNOS signaling pathway in tibial tissue, whereas Sr reverses the inhibitory effect of RA by activating this pathway. Pearson’s correlation analysis further revealed the potential associations between vascular development markers and bone health markers (Fig. 9J). Significant negative correlations were observed between the expression of vascularization-related markers (VEGFA, VEGFR2, PI3K, AKT, and eNOS) and the expression of RANKL (r = −0.653, p < 0.01; r = −0.391, p < 0.01; r = −0.733, p < 0.01; r = −0.475, p < 0.01; r = −0.426, p < 0.01), highlighting the strong effect of the vascularization status on bone health. Collectively, these results indicate that supplementation with Sr can improve RA-induced tibial injury by reversing changes in the expression of components of the PI3K/AKT/eNOS signaling pathway.Fig. 9RA treatment impairs tibial vascularization in broilers by suppressing the PI3K/AKT/eNOS pathway.A–I mRNA expression levels of bone/vascular development markers (BMP-2, RUNX2, OPG, VEGFA, VEGFR2, PI3K, AKT, and eNOS) determined using qPCR (n = 3). J Pearson’s correlation analysis of vascular development markers and bone health parameters. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). RA treatment impairs tibial vascularization in broilers by suppressing the PI3K/AKT/eNOS pathway. A–I mRNA expression levels of bone/vascular development markers (BMP-2, RUNX2, OPG, VEGFA, VEGFR2, PI3K, AKT, and eNOS) determined using qPCR (n = 3). J Pearson’s correlation analysis of vascular development markers and bone health parameters. The data are presented as the means ± SEMs. Statistical significance was determined as follows: #p < 0.05, ##p < 0.01, and ###p < 0.001 (RA30 group vs. the Con group); *p < 0.05, **p < 0.01, and ***p < 0.001 (RA30 + Sr200 group vs. the RA30 group). Discussion Bone health in poultry is essential for maintaining physiological functions and ensuring optimal production performance, as it provides structural support, facilitates locomotion and feeding, and reduces the risk of injury. These properties are especially critical for Ca metabolism in laying hens and for supporting the rapid growth rate characteristic of broilers. However, osteochondral pathologies such as TD and osteoporosis (OP) can lead to lameness, decreased egg production, and increased culling rates, resulting in substantial economic losses to the industry. Trace elements play an indispensable role in regulating bone formation, mineralization and metabolism, directly influencing bone strength and structural integrity through diverse biological pathways. Therefore, investigating their functions is key to optimizing feed formulations, improving production efficiency, reducing skeletal abnormalities, improving animal welfare, and advancing precision nutrition strategies for sustainable poultry farming. In this study, we employed a multidimensional experimental approach to determine the systemic protective effects of Sr supplementation on thiram-induced TD in broilers. We further elucidated the underlying mechanisms involved in the restoration of bone metabolic homeostasis, reconstruction of the bone vascular network, and remodeling of cross-tissue metabolic regulatory networks. These findings provide a theoretical foundation for developing targeted nutritional interventions to improve skeletal health in poultry. Trace elements such as zinc (Zn), copper (Cu), and Sr play crucial roles in avian bone development and Ca deposition by regulating osteoclastogenesis, promoting collagen synthesis, and facilitating matrix mineralization16,34,35. Among these trace elements, Sr has attracted increasing scientific interest. It can partially substitute for Ca within the hydroxyapatite crystal lattice. This substitution modulates bone cell activity by inhibiting osteoclast differentiation while promoting osteoblast proliferation, which collectively increases bone mineral density (BMD)17. Although dietary Sr supplementation has been shown to improve eggshell quality and tibial strength, excessive intake may disrupt Ca–P metabolism, potentially inducing nephrocalcinosis or skeletal deformities36,37. In the present study, histopathological and micro-CT analyses of TD broilers revealed characteristic pathological changes, including abnormal cartilage expansion, reduced osteoblast numbers, decreased BMD, and trabecular thinning. These alterations reflect not only bone loss but also a severe disruption of osteoblast‒osteoclast homeostasis. The Sr intervention mitigated these defects through the activation of the BMP-2/RUNX2 signaling axis, which enhanced the osteogenic differentiation of tibial mesenchymal cells. Crucially, Sr treatment upregulated OPG expression while it suppressed RANKL expression, thereby rebalancing bone remodeling homeostasis and restoring bone density. These findings are consistent with results showing that OPG sequesters TRAF6 to inhibit RANKL signaling38,39. Notably, in the TD model, we observed elevated RANKL mRNA levels and reduced protein expression, a phenomenon that may involve posttranscriptional regulatory mechanisms, such as suppressed translational efficiency or accelerated protein degradation40. Dithiocarbamate compounds can induce protein carbonylation, inhibit heat shock protein expression, and accelerate the degradation of oxidatively modified proteins via the ubiquitin‒proteasome pathway41,42. Consequently, although RANKL transcription is upregulated through compensatory mechanisms, the net levels of its protein product decrease because of accelerated degradation. These findings highlight the importance of posttranscriptional regulation in the pathological state of TD and provide a new perspective for understanding molecular changes related to bone metabolism. Nevertheless, the ratios of the OPG/RANKL mRNAs and proteins, key determinants of osteoclast differentiation and bone resorption, were consistently and significantly lower in the TD group. Sr supplementation efficiently reversed this imbalance by substantially upregulating OPG expression while simultaneously suppressing RANKL expression, providing a molecular basis for its role in inhibiting osteoclast overactivation. The increased OPG/RANKL ratio likely further suppresses osteoclastogenesis via the NF-κB-mediated activation of NFATc1, thereby blocking osteoclastic precursor fusion43,44. Additionally, Sr treatment restored OCN expression in broilers, a process that promotes hydroxyapatite deposition and may exert endocrine effects45. OCN potentially enhances energy metabolism (e.g., via pancreatic β-cell potentiation), indirectly supporting bone development through an improved nutritional status46–48. Vascular impairment is a recognized hallmark of TD pathogenesis in broilers5. Our study revealed the significant downregulation of VEGFA and VEGFR2 expression in the proximal tibial tissues of TD broilers. The Sr intervention restored the angiogenic capacity via activation of the PI3K/AKT/eNOS signaling axis. Interestingly, the histopathological analysis revealed a significant correlation between neovascularization patterns and osteogenic activity. Goldner’s trichrome staining showed that Sr-induced vessels predominantly colocalized within the osteogenic hotspot, as indicated by SO/FG-stained osteoblastic regions. This regional overlap in histopathology was functionally supported by concurrent increases in the immunohistochemical density of vessel markers (CD31 and CD34) and micro-CT-quantified vascular connectivity indices. These findings collectively indicate functional synergy between vascular remodeling and osteogenic restoration. This coupling is synergized by key signaling factors such as VEGF, FGF-2, and PDGF. VEGF secreted by hypertrophic chondrocytes promotes osteogenesis via β-catenin activation while regulating vascular aggregation through isoforms like VEGF-A49. FGF-2 increases endothelial cell proliferation, and PDGF stabilizes the vasculature to accelerate bone regeneration, as reported in a clinical study50. The hypoxic microenvironment further amplifies angiogenesis via HIF-1α/2α-dependent VEGF induction, with HIF-driven H-type vessels enriching osteoprogenitor cells and directly promoting ossification51,52. Notch signaling bidirectionally regulates this process by restraining excessive vascularization while promoting chondrocyte survival via Noggin secretion, thereby maintaining vascular–bone homeostasis53. Further single-cell transcriptomic studies could further delineate the changes in endothelial subpopulations specifically induced by Sr that regulate the spatial distribution54. Collectively, our findings demonstrate that the Sr-mediated restoration of the VEGFA/VEGFR2 pathway amplifies bone formation by enhancing vascular–endothelial cell–osteoblast crosstalk. This multicellular interaction mechanism helps explain the superior therapeutic efficacy of Sr to Cs supplementation alone in reversing TD-associated bone damage55–57. Disturbed Ca and P homeostasis, a hallmark feature of TD, was effectively corrected by the Sr intervention58. In TD broilers, we observed a characteristic imbalance: a decreased tibial Ca content concomitant with elevated plasma Ca levels, indicating an impaired tissue Ca deposition. Mechanistically, Sr, whose ionic radius is similar to that of Ca, competitively binds to the Ca-sensitive receptor (CaSR). This receptor activates the MAPK/ERK pathway, mimicking low-Ca signaling to stimulate parathyroid hormone (PTH) secretion and promote osteoclastic bone resorption, thereby mobilizing Ca stores59,60. Similarly, the Sr intervention restored gut microbiota stability in broilers with TD, specifically by enriching the abundance of Levilactobacillus brevis strains, a butyrate-producing bacterium. Butyrate is known to upregulate the expression of sodium (Na)–P cotransporter protein (NaPi-II) and increase intestinal P absorption. This “microbiota‒plasma metabolite” axis plays a critical role in restoring P homeostasis61–64. Furthermore, FMT from Sr-supplemented broilers to healthy mice promoted skeletal development, as evidenced by an increased BMD, increased trabecular bone volume, and upregulated OPG and OCN expression. In general, a lower BMD is often associated with a reduced trabecular bone quantity and poor structure, leading to a decreased BMD48. However, in the present study, an integrated analysis of micro-CT images and bone morphometric parameters suggested a different mechanism induced by the Sr intervention. Key parameters such as trabecular separation (Tb.Sp) and a lower BS/BV ratio indicate an optimization of trabecular bone quality rather than a mere increase in quantity. Specifically, the decreased BS/BV ratio implies that Sr promotes thickening and improves the connectivity of existing trabeculae, shifting the architecture from a delicate, high-surface-area network to a more robust and stable structure. This quality-driven enhancement, coupled with increased bone matrix mineralization, fully explains the observed increase in BMD. This interpretation is consistent with reports in the literature that strontium salts, such as strontium ranelate, can “improve bone microarchitecture and increase trabecular thickness”16,20. Collectively, these findings validated the role of Sr in the gut microbiota-mediated osteoprotective mechanism in broilers. The gut microbiota regulates distal organ functions through its metabolites entering systemic circulation65,66. In broilers with TD, integrated metabolomic profiling revealed the marked accumulation of 9-cis-RA, which strongly correlated with the expression of microbial markers and bone resorption marker genes (e.g., TRAP and CTSK). These findings establish a critical “microbiota–metabolite–bone” axis in TD pathogenesis. Interestingly, while the mRNA expression of the 9-cis-RA-specific nuclear receptor RXR remained unchanged in TD broilers, the levels of RAR (a receptor shared by all-trans-RA) were significantly increased. This finding indicates that excess all-trans-RA competitively binds to RAR, reducing the binding efficiency of 9-cis-RA and leading to its accumulation67. Thus, dysregulated RA metabolism appears to have detrimental effects on broiler bone health, whereas Sr may exert protective effects partly by modulating retinoic acid levels32. The blockade of 9-cis-RA/RXR signaling by all-trans-RA exacerbates bone damage through two principal pathways: (i) the abnormal formation of RXR/RAR heterodimers increases the transcriptional activity of osteolytic genes such as RANKL67, and (ii) unmetabolized 9-cis-RA acts as an endogenous ligand that activates peroxisome proliferator-activated receptor γ (PPARγ), which inhibits osteogenesis and promotes adipogenesis when overexpressed68. This metabolic disruption aligns with the enrichment of the PPAR pathway observed in the present study, helping to explain the imbalance in bone marrow stromal cell (BMSC) fate determination in TD broilers69. Importantly, the Sr-mediated regulation of RA metabolic genes (RAR and Cyp26a1) appears to restore physiological 9-cis-RA flux by accelerating all-trans-RA degradation and reducing its competitive occupancy of RXR, a mechanism consistent with the “ligand shunting” paradigm in the nuclear receptor network70. Genes related to RA synthesis (e.g., RAR, Adh1, Rdh7, and Aldh1a1) were significantly upregulated in the TD broiler group compared with the Con group, indicating hyperactive RA metabolism during TD progression. As a vitamin A metabolite, RA is known to impair skeletal health32. For instance, Wang et al.71 induced osteoporosis in mice via RA administration, and Shen et al.72 showed that high-dose RA suppresses growth hormone (GH)–insulin-like growth factor (IGF)-1–IGF-binding protein 3 (IGFBP3) axis activity, disrupting cartilage/bone homeostasis and inhibiting skeletal growth in developing animals. Consistent with the known disruption of chondrocyte hypertrophic zone vascularization by RA, severe leg abnormalities, including growth plate widening, tibial shortening, and motor impairment, were detected in RA-fed broilers73. The Sr intervention effectively restored the tibial length, trabecular architecture, and growth plate vascularization, indicating dual mechanisms of the repair of endochondral ossification: (i) activation of osteogenic marker genes (e.g., BMP-2 and RUNX2) to promote BMSC differentiation into osteoblasts and (ii) suppression of osteoclastogenesis via an increased OPG/RANKL ratio, thereby re-establishing bone remodeling homeostasis. These effects correlated with the restoration of the level of OCN, a key mineralization regulator, by Sr. RA concurrently inhibited proangiogenic pathways (PI3K/AKT/eNOS and VEGFA/VEGFR2), reducing eNOS activity and the vessel density, consistent with RAR-mediated HIF-1α degradation impairing hypoxic signaling74. Conversely, Sr reactivates endothelial PI3K/AKT signaling by restoring VEGFA/VEGFR2 activity, thus increasing nitric oxide (NO) synthesis and neovascularization via vascular‒osteogenic coupling. RA also disrupted Ca/P metabolism, decreasing the tibial Ca content while increasing the plasma Ca level, such as through the RAR-mediated inhibition of renal tubular Ca reabsorption. Sr corrected the Ca/P ratios by increasing ALP activity. The multilevel consistency between the RA-induced tibial damage and thiram-induced TD phenotypes confirms the pivotal role of RA in TD pathogenesis, although specific regulatory nodes (e.g., RAR-mediated inhibition of chondrocyte autophagy) warrant further investigation. Additionally, this study revealed that high-dose RA treatment markedly reduced food intake, which may have induced a systemic catabolic state and contributed secondarily to bone loss75. However, the quantitative analysis revealed that severe skeletal deterioration was due mainly to the direct pharmacological effects of RA on bone. RA-treated broilers exhibited pronounced reductions in the bone volume fraction and trabecular number, together with markedly increased serum levels of osteoclast-related markers, indicating a shift in bone remodeling toward resorption. These findings are consistent with the evidence that excess RA promotes osteoclast formation and activity and impairs osteoblastogenesis via the nuclear RAR signaling pathway, leading to cortical thinning and trabecular rarefaction in rodents and other models33,71. Thus, the direct effects of RA, which stimulate osteoclastogenesis and inhibit osteoblast differentiation, likely represent the primary mechanism driving the observed bone phenotype, whereas the systemic catabolic effects on a reduced appetite may further amplify this osteotoxicity. Future studies using pair-fed or paired-cage designs will be essential to disentangle these components and define the appetite-independent toxicity of RA to the skeleton. Overall, the present results underscore the substantial risk to skeletal health associated with high-dose RA exposure at the whole-animal level. The ash fraction analysis revealed that the ash weight in the RA15 group was lower than that in the control group, whereas the ash weight of the RA60 group was higher. This apparently paradoxical pattern likely reflects qualitatively different mineralized tissues after treatment with low and high doses of RA. Consistent with studies showing that excess RA inhibits osteoblast differentiation and mineralization by suppressing WNT signaling and other pathways76,77, low-dose RA (RA15) may primarily reduce the bone mass and mineral content, yielding a lower ash weight. In contrast, high-dose RA (RA60) not only suppresses osteogenesis but also profoundly disturbs endochondral ossification and growth plate dynamics, which are highly sensitive to RA67. RA can rapidly drive hypertrophic chondrocytes toward matrix mineralization78 and, when dysregulated, promote aberrant mineral deposition in cartilage-like matrices rather than in normal lamellar bone. Thus, RA60 may impair the normal degradation and remodeling of growth plate cartilage, leading to abnormal cartilage retention and subsequent pathological calcification (i.e., “chondrocalcinosis”) rather than physiological bone mineralization 65,76. The accumulation of this pathologically mineralized cartilage would “artificially” increase the ash weight in broilers treated with RA60, despite overall skeletal toxicity. Together, these findings support a bidirectional, dose-dependent effect of RA on skeletal outcomes: at lower doses, a predominant loss of healthy bone occurs; at higher doses, additional components of ectopic or dysorganized calcification are superimposed on bone loss. The RA30 + Sr200 group showed a dissociation between restored feed intake and a lack of weight gain, indicating that Sr can alleviate RA-induced anorexia but not the underlying systemic toxicity. Prior work has shown that RA can modify intestinal morphology without consistently improving body weight and that excessive doses of RA can damage intestinal and hepatopancreatic tissues and impair growth performance79,80. Based on this information, we propose that (i) RA compromises intestinal absorptive efficiency and that (ii) the RA-mediated inhibition of endochondral ossification and systemic anabolic pathways restricts the conversion of ingested nutrients into body mass. Thus, Sr supplementation targeting appetite alone is insufficient to overcome multilevel RA toxicity, highlighting the limits of single-node interventions in promoting true growth recovery. This study clarifies the protective role of Sr in tibial health and reveals its ability to mitigate the detrimental effects of RA via the regulation of the gut‒bone axis. While these findings improve our understanding, several key questions warrant further investigation: (i) the bone health impacts of specific RA isomers require validation through targeted intervention trials; (ii) the RA signaling pathway governing avian osteochondral homeostasis needs deeper mechanistic dissection; and (iii) the precise regulatory nodes through which Sr modulates RA metabolism remain to be elucidated. In summary, this study reveals that dietary Sr alleviates TD in broilers via the gut‒bone axis. Mechanistically, Sr restores the bone microstructure and Ca/P homeostasis, activates osteogenic factors and proangiogenic pathways, and promotes OCN expression while inhibiting osteoblastic activity. Furthermore, Sr remodels the gut microbiota of TD broilers, leading to the downregulation of the expression of the pathogenic metabolite 9-cis-RA and the modulation of the RA signaling pathway. Importantly, Sr rebalances the diet‒microbiota‒metabolite axis by reducing the competitive binding of RA isomers to nuclear receptors, thereby normalizing chondrocyte differentiation and endochondral ossification. These findings identify novel targets for preventing avian skeletal disorders and provide a theoretical basis for developing microbiota-targeted strategies for metabolic bone disorders. Methods Ethics statement All animal experimental procedures were reviewed and approved by the Animal Welfare and Ethics Research Committee, Henan Agricultural University, Zhengzhou, China (Ethical Approval Number: HNND2025030601). We strictly adhered to the “3Rs” principles of animal welfare and institutional guidelines. Broilers and mice were housed under standardized conditions (temperature, 28 ± 7 °C, humidity 60 ± 10%, 12/12-h light/dark cycle) with free access to food and water. On the experimental start date, the animals were randomly assigned to groups based on body weight using a completely randomized method. All terminal sampling was performed following humane euthanasia. The broiler chickens and mice were deeply anesthetized via an intraperitoneal injection of sodium pentobarbital (100 mg/kg body weight). After confirming a loss of reflexes, blood was collected via the carotid artery and cardiac puncture. During the experiments, the data collectors were aware of the group assignments; however, all subsequent laboratory analyses (histology, micro-CT, and molecular detection) were conducted in a blinded manner. Reagents and materials Tetramethylthiuram disulfide (thiram; AR grade, purity ≥97%, #C15916423) was purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). Strontium chloride (SrCl2; AR grade, purity ≥99.5%, #M22HS178637) was obtained from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). One-day-old male Arbor Acres (AA) broilers (initial body weight: 47.82 ± 5.35 g) were obtained from Xingda Poultry Industry Co., Ltd. (Kaifeng, Henan, China). Detailed information regarding the antibodies is listed in Supplementary Table S1. Experimental design and procedures In the Sr intervention trial, the broilers were housed under controlled environmental conditions: a 12 h/12 h light‒dark cycle, ambient temperature maintained within the brooding range (gradually decreased from 35 °C to 33 °C over the first week and then stabilized at 33 °C), relative humidity at 60–70%, and ad libitum access to feed and water. One hundred fifty chicks were randomly allocated into two primary groups: a control (Con) group (n = 30) and a TD group (n = 120). The Con group received a standard basal diet throughout the experiment. From Days 4 to 7 post-hatching, the TD model group received a basal diet supplemented with thiram at a dosage of 100 mg/kg body weight (BW) to induce TD. On Day 8, successful thiram-induced TD was confirmed. Starting on Day 4, the TD group was further randomly subdivided into four treatment groups (n = 30 broilers per group): the TD group (receiving the basal diet only (no Sr supplementation)) and the Sr100/Sr200/Sr400 groups (receiving the basal diet supplemented with strontium chloride (SrCl₂) at 100 mg/kg, 200 mg/kg, and 400 mg/kg BW, respectively). The broilers were continuously treated with Sr for 11 days. For the fecal microbiota trial, all male SPF-grade C57BL/6 mice (22–24 g) used in this study were purchased from Liaoning Changsheng Biotechnology Co., Ltd. (Shenyang, Liaoning, China). The mice were acclimatized under SPF-grade feeding conditions with 12 hours of light exposure per day for one week. The mice were then randomly divided into a blank (Con) group and a colony transplantation (FMT-Sr) group. Both groups were treated with a mixture of antibiotics (200 mg/kg ampicillin, metronidazole, neomycin, and 100 mg/kg vancomycin) by gavage for 6 days to eliminate the intestinal microbiota. After the flora had stabilized for 1 day, the Con group was gavaged with saline, and the FMT-Sr group was gavaged with a bacterial mixture (stools were collected from broilers after 21 days of continuous feeding with Sr, homogenized by adding saline at 50 mg/ml, centrifuged at 100 × g for 2 min at 4 °C, and collected as a suspension) for 3 days, after which each group was gavaged once every two days until the end of the 21-day period. In the retinoic acid (RA) experiment, the animals were housed in the same environment as in the Sr intervention trial. One hundred fifty healthy broilers were randomly divided into five groups: the control group (Con), low-dose RA group (RA15), medium-dose RA group (RA30), high-dose RA group (RA60), and medium-dose RA and Sr200 cogroup (RA30 + Sr200). The Con group was fed a normal diet; the low, medium, and high RA treatment groups were supplemented with 15, 30, and 60 mg/kg RA, respectively, in the normal diet from Day 4 to Day 21; and the medium-dose RA and Sr200 combination treatment group was supplemented with 30 mg/kg RA and 200 mg/kg strontium chloride, respectively, in the normal diet from Day 4 to Day 21. Sample collection In the Sr intervention trial, on Days 7 and 14, ten broilers per group were fasted for 12 hours and euthanized humanely, after which their body weights were recorded. Approximately 3 mL of blood was collected aseptically from the jugular vein into sodium heparin anticoagulant tubes (#20172220312, AOSAITE Company, Shandong, China). The tubes were gently inverted for mixing and centrifuged at 3000 × g for 10 min at 4 °C. The plasma supernatant was aliquoted and stored at −80 °C for further analysis. The liver, kidney and proximal tibia from each broiler were excised. One set of samples was immersion fixed in 4% paraformaldehyde (PFA) for 24–48 h for the histopathological examination. A second set of liver, kidney, and tibia tissue samples and fresh fecal samples collected on Day 14 were rapidly frozen in liquid nitrogen and stored at −80 °C for molecular analyses. Tibiae were dissected to remove the soft tissue. Tibia samples were weighed on an electronic balance with a sensitivity of 0.001 g. The tibial length and diameter were measured using digital Vernier calipers (#A231210151, DELI Company, Zhejiang, China). The tibiae were then longitudinally sectioned along the sagittal plane to expose the growth plate. The width of the growth plate at the proximal tibia was measured using calipers. In the fecal microbiota trial, on Day 28, ten mice per group were fasted for 12 h and euthanized humanely, after which their body weights were recorded. Blood was collected aseptically and loaded into sodium heparin anticoagulation tubes (#20172220312, AOSAITE Company, Shandong, China). The tubes were gently inverted for mixing and centrifuged at 3000 × g for 10 min at 4 °C. The plasma supernatant was aliquoted and stored at −80 °C until further analysis. Femoral tissue from each mouse was excised and immersion fixed in 4% PFA for 24–48 h for the histopathological examination. For the retinoic acid trial, bone and blood samples were collected and processed using the same method as in the Sr intervention trial. Histopathological analysis Liver tissue, proximal tibial tissue and mouse femoral bone tissue fixed with 4% PFA were decalcified in a commercial decalcification solution (the liver tissues did not undergo the decalcification process). Following decalcification, the tissues were processed through a standard histological protocol: dehydration in a graded ethanol series, clearing in xylene, and embedding in paraffin using a fully enclosed tissue processor (ASP300S; Leica Biosystems; Buffalo Grove, IL, USA). Sections (4–5 µm thick) were cut with a microtome and stained with hematoxylin and eosin (HE), toluidine blue (TB), safranin O/fast green (SO/FG), Goldner’s trichrome (GT), and TRAP81–84. The stained sections were examined and imaged under brightfield illumination using a Motic BA410 microscope (Motic China Group Co., Ltd.). Morphometric analyses of the growth plate structure, cartilage proteoglycan content, cartilage matrix, and bone mineralization and vascularization within the proximal tibia were performed using Motic Images Advanced 3.2 software (Motic China Group Co., Ltd.). All bone histomorphometric analyses were performed on the primary cancellous bone region of the proximal epiphysis of the tibia. The standardized method for localizing this region of interest (ROI) is described below. Anatomical localization was performed on sagittal tissue sections, with the growth plate cartilage‒epiphyseal junction serving as the proximal boundary. The ROI was defined as the metaphysis within 0.2 mm distal (toward the diaphysis) from this junction. This area contains a newly formed primary trabecular network resulting from endochondral ossification during active remodeling, representing a characteristic site for assessing osteoblast activity and the early bone formation status. Micro-CT analysis The tibiae from each group of broilers were scanned using a SkyScan 1176 small animal micro-CT system (Bruker, Germany). Using the top as the reference plane, 520 layers downward were counted; layers 440–520 constitute the cortical bone region. Using the top as the reference plane, 350 layers downward were counted; layers 270–350 constitute the ROI for cancellous bone. The thickness of each layer is 17.76 μm. Bone morphometric parameters, including the bone volume/total volume (BV/TV), bone mineral density (BMD), cortical thickness (Ct.Th), cortical area (Ct.Ar), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp) and structural modeling index (SMI), were analyzed in the ROI using CTAn software, with consistent analysis ranges observed across all samples. Three-dimensional reconstructions were generated using CTVOX software. The bone mass of the femurs of the mice was analyzed using the same method. Determination of Ca, phosphorus (P), and alkaline phosphatase (ALP) levels The instructions provided with the kit from Nanjing Jiancheng Bioengineering Co., Ltd., were followed. A microplate enzyme-linked assay was used to determine plasma ALP (ALP/AKP) activity levels, an MTB microplate assay was used to determine plasma Ca levels, and a phosphomolybdic acid assay was used to determine plasma P levels. The sample weight was between 2 and 5 grams per group, defatting treatment was performed, and the weight (accurate to 1 mg) was recorded before the sample was placed in a muffle furnace. The samples were calcined at 1000 °C for 4 h, removed, cooled, and weighed (accurate to 1 mg). The ash weight was recorded. The bone ash content was calculated as follows: bone ash content = ash weight/defatted bone weight. A mixed solution of hydrochloric acid and nitric acid was added, the mixture was boiled for 10 minutes, cooled, transferred to a 100 mL volumetric flask, diluted to volume, and the sample solution was stored at room temperature. The Ca and P contents of the tibia were measured using the ethylenediaminetetraacetic acid (EDTA) complexometric titration method (GBT 6436-2018) and the vanadomolybdenum yellow colorimetric method (GBT 6437-2018), respectively. Tibial growth plate (TGP) isolation and RNA and protein extraction Meticulous dissection was needed to ensure the specificity of the molecular analysis of the TGP. Precise incisions were made below the cancellous bone at the metaphysis and above the epiphysis at the diaphysis to isolate a pure, milky-white, translucent cartilage band, which was immediately frozen. Cartilage tissue was subsequently thoroughly homogenized using TRIzol reagent supplemented with β-mercaptoethanol to extract total RNA. Concurrently, the tissue was lysed with enhanced RIPA buffer containing protease and phosphatase inhibitors to extract total protein. RNA extraction and real-time fluorescence quantitative PCR (RT‒qPCR) Total mRNA was extracted from proximal tibia tissue using TRIzol reagent. mRNA quality was assessed using a nanodroplet spectrophotometer, and cDNA was subsequently synthesized via reverse transcription according to the manufacturer’s protocol. cDNA synthesis was completed using ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotechnology Co., Ltd., China). RT‒qPCR was subsequently performed. The mRNA expression level was determined using the 2-ΔΔCT method. The primers specific for the GAPDH mRNA and target gene sequences (OPG, RUNX2, RANKL, VEGFA, VEGFR2, PI3K, AKT, eNOS, RXR, RAR, ADH1, RDH7, CYP26A1, ALDH1A1 and ALDH1A2) were designed using the NCBI platform and synthesized by Tsingke Biological Technology Co., Ltd. The primers used in this study are listed in Supplementary Tables 2 and 3. The GAPDH gene served as an internal reference. Protein extraction and western blot (WB) analysis Total protein was extracted from TGP tissues using RIPA lysis buffer, and the protein concentration was determined using the BCA method. The required loading volume was calculated based on the measured protein concentration. Protein samples were loaded onto a precast SDS‒PAGE gel for electrophoretic separation. After electrophoresis, the proteins were transferred to a PVDF membrane (0.45 μm) using a transfer apparatus. After transfer, the PVDF membrane was blocked with rapid blocking solution at room temperature for 20 min. The membrane was subsequently washed with PBST and incubated overnight at 4 °C with primary antibodies against the target proteins (OPG, RUNX2, RANKL, VEGFA, VEGFR2, PI3K, p-PI3K, AKT, p-AKT, eNOS, and p-eNOS) and the housekeeping protein GAPDH. After washing with PBST, the membranes were incubated with secondary antibodies. Finally, the membranes were developed, and band intensity values were analyzed using ImageJ software. 16S rRNA sequence analysis Microbial total genomic DNA samples were extracted from feces using the OMEGA Soil DNA Kit (M5635-02, Omega Bio-Tek, Norcross, GA, USA) according to the standard procedure provided by the manufacturer, and the resulting DNA samples were stored at −20 °C for further analysis. The extracted DNA was quantitatively analyzed using a NanoDrop NC2000 UV‒Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), whereas DNA integrity and quality were assessed via electrophoresis on a 1% agarose gel (100 V, 30 min) to ensure that the DNA samples met the subsequent experimental requirements. PCR amplification of the nearly full-length 16S rRNA gene was performed using the bacterial universal primers 27 F (5’-AGAGTTTGATCMTGGCTCAG-3’) and 1492 R (5’-ACCTTGTTACGACTT-3’) and DNA samples with a two-step PCR amplification procedure, and the PCR products were subsequently purified using Agencourt AMPure Beads (Beckman Coulter, Indianapolis, IN) for the purification of the PCR products and quantification with a PicoGreen dsDNA assay kit (Invitrogen, Carlsbad, CA, USA). After the quantification was complete, equal amounts of amplicons were mixed and commissioned for single-molecule real-time (SMRT) sequencing using the PacBio Sequel platform by Shanghai Personal Biotech (Shanghai, China). Sequence data were analyzed mainly using QIIME2 in combination with the dada2 plugin and the R package (v3.2.0). Microbial community diversity and abundance were subsequently assessed by performing α diversity and β diversity analyses. The 16S sequencing data were analyzed using the gene cloud tool (https://www.genescloud.cn). Blood metabolomic analysis Frozen blood samples were thawed at 4 °C, transferred to a precooled methanol/acetonitrile/water mixture (2:2:1, v/v) and subjected to vortexing and sonication for 30 min at 4 °C, followed by a 10-min resting period at −20 °C. The mixture was centrifuged (14,000 × g, 20 min) at 4 °C to remove the precipitate, and the supernatant was collected and vacuum freeze-dried. Prior to the mass spectrometry analysis, the mixture was reconstituted with 100 μL of an acetonitrile–water mixture (1:1, v/v), vortexed at 4 °C, allowed to stand for 15 min, centrifuged (14,000 × g, 15 min, 4 °C), and passed through a 0.22 μm PVDF membrane, after which 300 μL of the supernatant was transferred to the injection vials for the nontargeted metabolomic analysis using UHPLC‒QTOF‒MS. The raw mass spectrometry data were preliminarily analyzed using Compound Discoverer™ 3.3 software (Thermo Fisher Scientific) and then converted to .mzXML files via the ProteoWizard tool (v3.0.21048). The XCMS software package (Bioconductor 3.18) was used for peak alignment, retention time correction and peak area extraction, and the resulting data matrices were subjected to the identification of metabolite structures (based on the self-constructed database of Shanghai Applied Protein Technology Co., Ltd.) and standardized preprocessing, after which data quality control (QC-RLSC algorithm) and a multivariate statistical analysis were performed. The significance threshold was set at |log2(fold change)| ≥ 1 and a corrected p value (FDR) < 0.05. Significantly differentially abundant metabolites were subjected to pathway annotation using the Kyoto Encyclopedia of Genes and Genomes (KEGG, Release 107.0) database, and the metabolic pathway enrichment analysis was completed using the hypergeometric distribution test. The multivariate statistical analysis module (principal component analysis, orthogonal partial least squares discriminant analysis, and volcano plots of differentially abundant metabolites) was implemented using the OmicStudio platform (https://www.omicstudio.cn/tool), and a metabolic interaction network topology analysis and visualization were performed using the GenesCloud system (https://www.genescloud.cn). Metabolic pathway chordal maps and clustering heatmaps were generated with the HiPlot online tool (https://hiplot.com.cn/home/index.html). Statistical analysis The data were statistically analyzed using SPSS 26.0 (SPSS Inc., USA) and GraphPad Prism 8.0.2 (GraphPad Software, USA). After confirming the normality of the data with the Shapiro‒Wilk test, comparisons between two groups were analyzed using independent samples t test. For multiple comparisons, Tukey’s HSD test was uniformly applied after significant results were obtained by ANOVA; for nonparametric data, the Kruskal‒Wallis test was followed by Dunn’s post hoc analysis (with the Bonferroni correction). The associations of metabolites, core microorganisms and related genes were explored by performing Pearson’s correlation analysis and the Mantel test. The data are reported as the means ± standard errors of the means (SEMs), and statistical graphs were generated with GraphPad Prism. p values < 0.05 were considered to indicate statistical significance. Supplementary information Supplementary material Author Checklist Supplementary material Author Checklist