From maternal microbes to offspring development: gut microbiota-derived thiamine regulates the gut microbiota and drives the placental Notch pathway to coordinate angiogenesis and nutrient transport Authors: Guodong Sun, Ze Wang, Xing Guo, Haoyang Sun, Teng Teng, Baoming Shi Published: 2025 ABSTRACT BackgroundThe maternal gut microbiota can modulate host physiological homeostasis through metabolites. Maternal reproductive potential hinges on placental angiogenesis and nutrient transport efficiency, directly determining fetal developmental outcomes. However, the specific molecular mechanisms by which microbial metabolites influence reproductive potential remain to be elucidated. This study aimed to clarify the mechanisms by which maternal gut microbiota affects reproductive potential.ResultsWe initially analyzed the metabolic profiles by untargeted metabolomics and the fecal microbiota by 16S rRNA sequencing in sows with different reproductive potential. Sows with high reproductive potential exhibited elevated plasma arginine and fecal thiamine levels. Meanwhile, Lactococcus was enriched in the feces of sows with high reproductive potential. Subsequently, we evaluated the effects of thiamine (a signature metabolite identified) on maternal reproductive potential, gut microbiota, placental angiogenesis, and nutrient transport capacity using a rat model. The results showed that thiamine supplementation in pregnant rats effectively promoted offspring growth and enhanced transplacental thiamine metabolism. Moreover, thiamine modulated maternal gut microbiota composition, increased the abundance of Prevotellaceae Ga6A1 group and Bacteroidale RF16 group unclassified, and promoted butyrate production. We found that thiamine improved placental function by enhancing thiamine-related metabolic enzymes and acetyl-CoA content. It also promoted the migratory capacity of pTr cells. Importantly, thiamine facilitated placental angiogenesis by activating Notch signal transduction, which in turn initiated the PI3K/AKT signaling cascade. Ultimately, this cascade regulated the efficiency of placental nutrient metabolism and the expression of nutrient transporters.ConclusionsCumulatively, the gut microbiota regulates early offspring development through metabolite-mediated host interactions. This study provides new evidence that maternal gut microbiota-derived thiamine activates placental Notch signaling to coordinate angiogenesis and nutrient transport, thereby improving pregnancy outcomes. These findings provide novel perspectives and potential actionable strategies for maternal microbial regulation of maternal–fetal health during gestation.Graphical Abstract Video AbstractSupplementary InformationThe online version contains supplementary material available at 10.1186/s40168-025-02317-1. Introduction Emerging evidence suggests that the host's reproductive potential is profoundly impacted by the composition of its gut microbiota, though the underlying mechanisms remain incompletely understood. Research shows that sows with high reproductive potential exhibit lower microbial diversity pre-farrowing but higher diversity post-farrowing compared to sows with low reproductive potential [1]. Similarly, a separate study identified notable variations in fecal microbiota abundance and community structure during gestation in sows with different reproductive potential [2]. Furthermore, it has been shown that fecal microbiota transplantation from high-reproductive-potential sows to low-reproductive-potential sows enhances endometrial gland development in recipient sows [3]. Recent studies have demonstrated that maternal-origin Lactobacillus vaginalis can enhance uterine receptivity, thereby supporting embryo survival [4]. Collectively, these findings demonstrate the modulatory role of gut microbiota; however, the precise molecular pathways and physiological mechanisms underlying this microbial-reproductive axis require further investigation. Imbalances within the mammalian gut ecosystem instigate the development of metabolic disorders such as fatty liver disease, obesity, and inflammatory bowel disease [5]. Currently, a variety of strategies are available to modulate the gut microbiota (e.g., fecal microbiota transplantation, probiotic supplementation, dietary modification) and thereby alleviate diseases caused by disorders of the gut microbiota [6–8]. Numerous studies have established that fecal microbes are able to modulate host metabolism through their metabolites. Although these studies are significant for elucidating both pregnancy physiology and pregnancy-associated pathologies, research specifically examining the mechanistic links between microbial metabolites and sow reproductive potential remains relatively limited. Pregnancy complications are recognized as one of the central challenges to global public health in the twenty-first century, posing particularly severe risks to women in low- and middle-income countries [9]. According to World Health Organization modelling, global maternal mortality from pregnancy complications is projected to reach 70 per 100,000 by 2030 [10]. Some complications, including preeclampsia, gestational hypertension, and gestational diabetes mellitus, not only induce fetal developmental abnormalities but also compromise maternal long-term health [11]. Pigs are frequently employed as model organisms in biomedical research, offering distinct advantages over other animal species due to their close anatomical and physiological resemblances to humans [12, 13]. Selective breeding in pigs confers a genetic resistance to type 2 diabetes development [14, 15]. Moreover, sows with higher reproductive potential exhibit superior placental efficiency and enhanced nutrient transport capacity, which are directly related to improved fetal outcomes [16, 17]. Given the physiological similarities between pigs and humans, we speculate that sows with high reproductive potential may possess specialized regulatory mechanisms that improve maternal–fetal outcomes. Elucidating these mechanisms is essential for optimizing health management during pregnancy. As the central mediator of maternal–fetal exchange, the placenta critically regulates fetal organogenesis, growth, and metabolic homeostasis [18]. The Notch pathway, an evolutionarily conserved key regulator, coordinates essential cellular processes [19]. The Notch pathway operates through receptor-ligand interactions between adjacent cell membranes, triggering downstream target gene activation. Activated during placental morphogenesis, the Notch signaling cascade orchestrates angiogenic programming by coordinating trophoblast invasion and spiral artery remodeling [20]. Further clarification is needed regarding how the Notch pathway regulates maternal reproductive potential, despite its established role as a central orchestrator of placental angiogenesis. In animals, thiamine predominantly exists and functions in its bioactive form as thiamine pyrophosphate (TPP). Serving as a coenzyme for pyruvate dehydrogenase (PDH), transketolase (TKT), and the α-ketoglutarate dehydrogenase complex (α-KGDHC), thiamine regulates key metabolic pathways in animals [21]. It is also crucial for maintaining acetyl-CoA synthesis and modulating histone acetylation processes [22]. Notably, recent studies have shown that women with gestational diabetes mellitus have lower serum thiamine levels, which show a strong inverse correlation with glucose intolerance [23]. In contrast, elevated antenatal thiamine intake reduces gestational diabetes mellitus risk, whereas deficiency compromises maternal health and impairs fetal gastrointestinal, cardiac, and neurological development [24, 25]. Nevertheless, the specific function of thiamine in placental development remains largely unexplored, despite its established functions in regulating metabolism and supporting healthy pregnancy. In this study, we characterized the fecal microbiota and metabolic profiles of sows with differing reproductive potential. Subsequently, the promotive effects of thiamine—a key metabolite biomarker of high reproductive potential—on placental angiogenesis and fetal development were validated using a pregnant rat model. Our findings provide new evidence addressing a critical gap in the regulation of host reproductive potential by the microbiota and provide novel insights into improving placental and fetal development through maternal nutritional interventions. Materials and methods Animal subjects and sample collection Twelve healthy Landrace × Large White crossbred sows (parity 2–3) with similar body weight (average body weight: 228.71 ± 5.12 kg) and estrus dates were used in this study. Once the sows came into estrus, they were artificially inseminated three times at 8 h intervals using semen from Duroc boars with an identical genetic background. The semen exhibited good sperm motility (> 70% linear movement). Each insemination dose contained approximately 2.5–3.0 × 109 active sperm. The total litter size, live births, stillbirths, and mummified piglets were recorded for each sow after delivery. Sows were allocated to high (H) or low (L) reproductive potential groups based on live-born litter size (Table S1). Sows with more than 14 live births were assigned to the high reproductive potential group, while those with fewer than 11 live births were assigned to the low reproductive potential group. Sows received NRC (2012)-compliant basal diets (Table S2). After farrowing, blood was collected from the auricular vein using sodium heparin vacutainer tubes. The blood was centrifuged at 1000 g for 15 min to obtain plasma, which was then stored at −80℃ for subsequent analysis. Immediately after farrowing, fresh fecal samples were collected from sows right after defecation. To prevent environmental contamination, the middle portion of freshly excreted feces was directly collected using sterile gloves and promptly transferred into sterile cryovials. All samples were snap-frozen in liquid nitrogen immediately after collection and stored at −80℃ for subsequent microbiome and metabolomic analyses. Next, a high-dose thiamine supplementation experiment was performed. 12-week-old female SD rats (281.28 ± 2.63 g) and sexually mature male SD rats (402.32 ± 3.51 g) were selected for mating (purchased from Liaoning Changsheng Biotechnology Co.). Successful conception was confirmed by vaginal plug examination on the day following mating, with the detection day designated as embryonic day 0 (E0). Subsequently, rats underwent randomization into control (C) or high-dose thiamine (HT) groups. Based on previous studies, the HT group received thiamine-supplemented water (200 mg/kg BW) throughout gestation and lactation [26]. The thiamine concentration in drinking water was adjusted to 1.82 mg/mL, calculated based on an estimated daily water intake of 22 mL per 200 g rat. During the experimental period, rats were individually housed with ad libitum water and a commercial diet (Liaoning Changsheng Biotechnology, Liaoning, China). Thiamine was purchased from Jiangxi Tianxin Pharmaceutical Inc. (99%, Jiangxi, China). Euthanasia was conducted via CO₂-induced unconsciousness followed by cervical dislocation. At day E18, three rats per group were subjected to cesarean section under anesthesia to collect the placenta. The placenta was fixed in 4% paraformaldehyde for histological processing (HE staining and immunohistochemical analysis), whereas the remaining tissues were immediately snap-frozen in liquid nitrogen with long-term storage at −80℃. Following parturition, litter size was documented, and offspring body weight was measured. At postnatal day 1 (P1) and P22, one male and female offspring per litter were randomly selected and euthanized for plasma and tissue collection. Each litter was standardized to eight pups, consisting of 4 males and 4 females. At day P22, blood samples from maternal and male offspring, along with maternal jejunal tissue, ileal tissue and colon contents, were collected. Blood samples were collected from rats by decapitation into blood collection tubes. The blood was allowed to clot at room temperature for 30 min, followed by centrifugation at 1000 g for 15 min to separate the serum. The serum was stored at −80°C until analysis. Microbiota sequencing and data analysis of fecal samples The CTAB method was employed to extract total DNA from feces and colon contents. Primers 341F (5'-CCTACGGGGNGGCWGCAG-3') and 805R (5'-GACTACHVGGGGTATCTAATCC-3') combined with barcode sequence to amplify the V3-V4 region of bacterial 16S rDNA. AMPure XT beads purified the samples (Beckman Coulter Genomics Inc., Danvers, MA, USA). The purified products underwent Qubit (Invitrogen, USA) quantification. Constructed amplicon pools were verified for size and quantified before NovaSeq PE250 sequencing. The sequences obtained from sequencing were filtered for quality control. Use Divisive Amplicon Denoising Algorithm to dereplicate to obtain Amplicon Sequence Variant (ASV) feature sequences and ASV abundance tables. Feature abundance was normalized to relative abundance using SILVA (version 138). Species annotation via SILVA database. Fecal microbiota phenotypes are predicted using the BugBase platform (http://bugbase.cs.umn.edu). Detection and bioinformatics analysis of plasma and fecal metabolites Metabolites were extracted from fecal and plasma samples using precooled 50% methanol. Then, the mixture was centrifuged at 4000 g for 20 min, and the resulting supernatant was collected for analysis. Subsequently, the supernatant was analyzed by LC–MS according to the instructions. Chromatographic separations employed a Vanquish Flex UPLC system (ThermoFisher Scientific Inc., Bremen, Germany). Reversed-phase separation used an ACQUITY UPLC T3 column (100 mm*2.1 mm, 1.8 umWaters, Milford, USA). Metabolite composition was analyzed using a high-resolution tandem mass spectrometer Q-Exactive (ThermoFisher Scientific Inc., Bremen, Germany). Metabolites were annotated using KEGG and HMDB databases. Metabolite differences were assessed by Student's t-test with FDR correction using the programming language R (version 4.2.1, The Comprehensive R Archive Network, USA). Detection of blood biochemical indicators Serum glucose, cholinesterase (CHE) and alkaline phosphatase (ALP) were measured using a Beckman Coulter analyzer (Unicel D × C 800 Synchron Clinical System). Hematoxylin and eosin-stained sections Fixed placental tissues underwent ethanol dehydration, paraffin embedding, sectioning, and H&E staining. Morphological assessment was then performed using a digital slide scanning system (SNIOI240I20, Winmedic Tech Co., Ltd., Shandong, China). The total area of the placenta, spongiotrophoblast area, labyrinth area and blood sinus area was measured using ImageJ (version 1.8.0, NIH, USA). Immunohistochemistry The immunohistochemical procedure followed established protocols [27]. Placental sections were sequentially incubated with primary antibody (4℃, overnight) and secondary antibody (30℃, 60 min). Freshly prepared DAB solution was used for color development. Sections were then counterstained with hematoxylin for 3 min and dehydrated. Sections were examined using a digital slide scanning system (SNIOI240I20, Winmed Tech Co., Ltd., Shandong, China) and analyzed using ImageJ (version 1.8.0, NIH, USA). Detection of thiamine, PDH, TKT, α-KGDHC, acetyl-CoA, and antioxidant function-related index levels A 10% placental homogenate was prepared with saline to evaluate the levels of thiamine, PDH, TKT, α-KGDHC, acetyl-CoA, and related antioxidant function indices. Thiamine levels in serum, placenta, intestine and feces were determined using a thiamine detection kit (Beijing Leagene Biotechnology Inc., Beijing, China). The levels of PDH, TKT, α-KGDHC, and acetyl-CoA in placental tissues were quantified by commercial ELISA kits (Shanghai Yuanju Biotechnology Inc., Shanghai, China). Total antioxidant capacity (T-AOC), superoxide dismutase (SOD), catalase (CAT), malondialdehyde (MDA), glutathione peroxidase (GSH-Px), total glutathione (T-GSH), oxidized glutathione (GSSG), and reduced glutathione (GSH) levels in placental tissues were measured using commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Cell culture, wound healing assay, and cell viability assay Porcine placental trophoblast (pTr) cells were provided by the School of Animal Science and Technology, Northeast Agricultural University. pTr cells were cultured in DMEM (Gibco, 12500062, USA) culture medium supplemented with 10% fetal bovine serum (Gibco, 10100147, USA) and penicillin–streptomycin (100 U/ml penicillin and 100 µg/ml streptomycin; Beyotime, Shanghai, China). Cell viability was measured using the MTT assay. 100 μL of medium containing pTr cells at a density of 1 × 105 cells/mL was added to a 96-well plate. When the cells reached 80–90% confluency, the culture medium was aspirated. The cells were cultured for 24 h in 100 μL of serum-free medium containing varying concentrations of thiamine (0, 0.2, 0.4, 0.8, 1.6, 3.2, 6.4, and 12.8 mmol/L). After removing the medium, 100 μL of prepared MTT solution (5 mg/mL) was added, and the plate was incubated at 37 ℃ for 4 h. The solution was then aspirated, and DMSO was added to completely dissolve the purple formazan crystals. The absorbance of each well at 570 nm was measured using a microplate reader (AMR-100, Hangzhou Allsheng Instrument Co., Ltd., Zhejiang, China). For wound healing assays, pTr cells in 6-well plates were cultured for 24 h, wounded with 200μL tips, and then incubated for 24 h in serum-free medium ± 0.8 mmol/L thiamine. Scratch areas were observed at 0h and 24 h using a cell imaging instrument (Thermo Fisher Scientific Inc., Bremen, Germany). Scratch areas were analyzed using ImageJ. RT-qPCR Total RNA isolated from the placenta with TRIzol (Thermo Fisher Scientific, Bremen, Germany) was quantified using a nanophotometer (Implen GmbH, Munich, Germany). cDNA was synthesized from RNA with PrimeScript™ FAST Kit (Takara Biomedical Technology Inc., Beijing, China). RT-qPCR mixtures prepared with TB Green (Takara Biomedical Technology Inc., Beijing, China) were analyzed on Q2000B System (Hangzhou LongGene Scientific Instruments Inc., Hangzhou, China). Quantitative PCR analysis using the 2−ΔΔCt method with β-actin as housekeeping gene. For each sample, at least two technical replicates were conducted. Table S3 details the primer sequences. Western blot Placental proteins were isolated using RIPA lysis buffer (Epizyme Biotech Inc., Shanghai, China) with phosphatase inhibitors and PMSF. Proteins were SDS-PAGE-separated and electroblotted to PVDF membranes. After blocking with protein-free buffer (Epizyme Biotech Inc., Shanghai, China), membranes were incubated overnight at 4℃ with primary antibodies. Subsequently, the membranes were incubated with secondary antibodies at 35℃ for 1 h. Protein bands were detected using an ECL Kit (Beyotime Biotechnology Inc., Shanghai, China) and quantified with ImageJ. Antibody details are in Table S4. Detection of short-chain fatty acids (SCFAs) The concentration of SCFAs in feces was determined as previously described [28]. Briefly, the sample was diluted with water and incubated at 4℃ for 48 h. Subsequently, a 25% (w/v) metaphosphoric acid-crotonic acid mixture buffer solution was added, followed by centrifugation at 13000 g for 10 min. The supernatant was subjected to membrane filtration, and this process was repeated three times. The filtrate was then analyzed using a Shimadzu GC2010 high-performance gas chromatography system, with SCFAs concentrations quantified via a pre-established standard curve. Statistical analysis Data were analyzed in SPSS 25.0 (IBM-SPSS Inc., Chicago, USA). Statistical significance was assessed using a two-tailed unpaired t-test or Welch's t-test when the data were normally distributed. A nonparametric test (Mann–Whitney U test) was employed when the data were not normally distributed. Data visualization used GraphPad Prism v8.0.2 (GraphPad Software Inc., San Diego, USA). The graphical abstract was plotted using BioRender. Results expressed as mean ± SEM. The correlation of fecal microbiota with plasma and fecal metabolites was analyzed using the MetOrigin [29]. The microbiota-metabolite correlation analyses were conducted on all individual samples, with FDR correction applied. Results Fecal microbiota analysis in sows with different reproductive potential We compared the fecal microbiota of sows with different reproductive potential (Fig. 1A). As shown in Fig. 1B, the L group exhibited 1395 ASVs, and the H group had 1217 ASVs, with 1099 ASVs shared between groups. No significant differences in α and β diversity were found (P > 0.05; Fig. 1C-G). Figure 1H displays phylum-level microbial abundance via a chord plot, whereas Fig. 1I shows the 30 most abundant genera via a chord plot. The relative abundances of Sphingomonas, Cellvibrio, SBR1031 unclassified, Coriobacteriaceae UCG-002, Methylococcaceae unclassified, Chelativorans, Mitochondria unclassified and AKAU3644 unclassified were significantly decreased in the H group (P < 0.05, Fig. 1J). Lactococcus and Actinomarinales unclassified were enriched in the H group (P < 0.05, Fig. 1J). Linear discriminant analysis effect size (LEfSe) analysis indicated that Lactococcus and Actinomarinales unclassified could serve as biomarkers for high reproductive potential (Fig. 1K). The metabolic phenotypes of fecal microbiota of sows with different reproductive potential showed no major changes (P > 0.05, Fig. 1L). Furthermore, Mantel test analyses demonstrated a significant positive correlation between live births and both Lactococcus and Actinomarinales unclassified (P < 0.05, Fig. 1M). PICRUSt analysis based on the COG database revealed significantly higher abundances in the H group of site-specific DNA-adenine methylase and phosphopantetheinyl transferase (P < 0.05, Fig. S1A). Concentrations of SCFAs in feces of sows with different reproductive potential did not change significantly (P > 0.05, Fig. S1B-H).Fig. 1Differences in fecal microbiota composition of sows with different reproductive performance. A Examination of fecal microbiota and fecal & plasma metabolomics in sows with different reproductive potential. B Venn diagram. C-E Comparison of Shannon (C), Chao1 (D) and Observed species (E) indices. F Principal coordinate analysis (PCoA) plot of fecal microbiota. G The Bray–Curtis rank within each group. H Relative abundance and composition of fecal microbiota at the phylum level. I Relative abundance and composition of the top 30 genus levels of fecal microbiota abundance. J Significantly different fecal microbes at the genus level. K LEfSe analysis based on significantly different genus. L Prediction of microbial metabolic phenotypes using BugBase platform. M Correlation analysis of reproductive potential with significantly different microbiota. n = 6 Differences in fecal microbiota composition of sows with different reproductive performance. A Examination of fecal microbiota and fecal & plasma metabolomics in sows with different reproductive potential. B Venn diagram. C-E Comparison of Shannon (C), Chao1 (D) and Observed species (E) indices. F Principal coordinate analysis (PCoA) plot of fecal microbiota. G The Bray–Curtis rank within each group. H Relative abundance and composition of fecal microbiota at the phylum level. I Relative abundance and composition of the top 30 genus levels of fecal microbiota abundance. J Significantly different fecal microbes at the genus level. K LEfSe analysis based on significantly different genus. L Prediction of microbial metabolic phenotypes using BugBase platform. M Correlation analysis of reproductive potential with significantly different microbiota. n = 6 Differences in fecal and plasma metabolite composition in sows with different reproductive potential and their correlation with fecal microbiota To further analyze metabolic differences between sows with different reproductive potential, we performed untargeted metabolomics analysis of feces and plasma (Figs. 2 and 3). 415 and 323 differential ions were identified in feces and plasma, respectively (Figs. 2A, C, 3A and C). Based on secondary mass spectrometry data, fecal and plasma metabolite ions were matched to an in-house standards database. 23 significantly different secondary metabolites (thiamine, choline, 5-Acetamidovalerate, etc. (Table S5)) and 9 significantly different secondary metabolites (citric acid, DL-Arginine, lysoPE 20:4, etc. (Table S6)) were identified in feces and plasma, respectively. Furthermore, the actual concentration of thiamine in fecal samples from the H group was significantly higher than that in the L group (P < 0.01, Fig. S1I).Fig. 2Differences in fecal metabolomes of sows with different reproductive potential. A Heat map showing distribution of significantly different metabolites. B PLS-DA analysis of fecal metabolomic differences in sows with different reproductive potential. C A volcano plot was used to show the differential metabolites between the L and H groups. D KEGG enrichment analysis of significantly different metabolites. E, F Fecal metabolites were classified by origin using MetOrigin. G Host-microbe shared metabolic pathways from metabolic pathway enrichment analysis based on fecal metabolites. H Correlation network diagram of microbiota and fecal metabolites. n = 6Fig. 3Differences in plasma metabolomes of sows with different reproductive potential. A Heat map showing distribution of significantly different metabolites. B Partial Least Squares Discriminant Analysis (PLS-DA) analysis of plasma metabolomic differences in sows with different reproductive potential. C A volcano plot was used to show the differential metabolites between the C and HT groups. D Correlation coefficients for significantly different metabolites. Positive and negative correlations are denoted by red and blue, respectively. Circle color depth and size are directly proportional to Spearman's rank correlation value. E KEGG analysis of significantly different metabolites. n = 6. *P < 0.05, **P < 0.01 Differences in fecal metabolomes of sows with different reproductive potential. A Heat map showing distribution of significantly different metabolites. B PLS-DA analysis of fecal metabolomic differences in sows with different reproductive potential. C A volcano plot was used to show the differential metabolites between the L and H groups. D KEGG enrichment analysis of significantly different metabolites. E, F Fecal metabolites were classified by origin using MetOrigin. G Host-microbe shared metabolic pathways from metabolic pathway enrichment analysis based on fecal metabolites. H Correlation network diagram of microbiota and fecal metabolites. n = 6 Differences in plasma metabolomes of sows with different reproductive potential. A Heat map showing distribution of significantly different metabolites. B Partial Least Squares Discriminant Analysis (PLS-DA) analysis of plasma metabolomic differences in sows with different reproductive potential. C A volcano plot was used to show the differential metabolites between the C and HT groups. D Correlation coefficients for significantly different metabolites. Positive and negative correlations are denoted by red and blue, respectively. Circle color depth and size are directly proportional to Spearman's rank correlation value. E KEGG analysis of significantly different metabolites. n = 6. *P < 0.05, **P < 0.01 Partial Least Squares Discriminant Analysis (PLS-DA) revealed distinct variation in fecal and plasma metabolites in sows with different reproductive potential (Figs. 2B and 3B). We analyzed correlations among significantly different fecal metabolites (Fig. S2A). Thiamine was significantly correlated with 5-acetamidovalerate (P < 0.05), whereas choline showed no significant correlations with any fecal differential secondary metabolites (P > 0.05). We analyzed the functional changes of differential metabolites in sows with different reproductive potential using KEGG enrichment analysis (Fig. 2D). The results showed significant enrichment of pathways, including choline metabolism in cancer and glycerophospholipid metabolism. Correlation analysis of significantly different secondary metabolites in plasma showed that DL-arginine was significantly negatively correlated with cis-8,11,14,17-eicosatetraenoic acid and citric acid (P < 0.05, Fig. 3D). KEGG analysis of plasma metabolites revealed enriched pathways including Glycerophospholipid metabolism and Citrate cycle (Fig. 3E). To analyze the origin of metabolites in feces and plasma, we used MetOrigin to integrate metabolomics and microbiomics data. Analysis showed 245 metabolites of host origin, 339 metabolites of microbiota origin, 970 metabolites of food-related origin, and 612 metabolites of other origins (drugs or environmental, etc.) in feces (Fig. 2E). A total of 208 host-microbiota co-metabolites were identified in fecal metabolites (Fig. 2F). We obtained eight metabolic pathways associated with host-microbiota co-metabolism (Fig. 2G). Microbiota-fecal metabolite interaction networks revealed positive correlations between the thiamine metabolism pathway and two metabolites, along with multiple microbes (Fig. 2H). In plasma, 143 metabolites of host origin, 200 metabolites of microbiota origin, 377 metabolites of food origin, and 308 metabolites from other sources (drugs or the environment, etc.) were detected (Fig. S2B). Among the plasma metabolites, there were 21 host-specific metabolites, 78 microbiota-specific metabolites, and 122 host-microbe co-metabolites (Fig. S2C). Subsequently, we obtained 12 metabolic pathways associated with host-microbiota co-metabolism (Fig. S2D). The microbiota-plasma inter-metabolite network plot showed that arginine was associated with multiple metabolic pathways and positively correlated with Lactococcus only (Fig. S2E). The characteristic metabolite thiamine enhances reproductive performance in rats Subsequently, we supplemented rats with thiamine, a metabolite characteristic of sows with high reproductive potential, to systematically evaluate its regulatory effects on reproductive performance (Fig. 4A). Thiamine supplementation had no significant impact on body weight in pregnant and lactating rats (P > 0.05, Fig. 4B). However, it significantly enhanced feed intake at E3, E5, E9, E12, P8, P10, and P11 (P < 0.05, Fig. 4C and Fig. S3C), while showing a tendency towards increased average daily feed intake (ADFI) during gestation (P = 0.080, Fig. S3A). At the same time, we compared the water intake of the two groups. Water intake and average daily water intake (ADWI) showed no significant difference between pregnancy and lactation with thiamine supplementation (P > 0.05, Fig. S3E-H). Supplementation with thiamine did not significantly alter total litter size, the number of male and female offspring, or sex ratio in rats (P > 0.05, Fig. 4D). Furthermore, thiamine had no apparent effects on either placental efficiency (P > 0.05, Fig. S3B). Notably, thiamine supplementation significantly enhanced litter weight at P15 and P22 (P < 0.05, Fig. 4E), concomitant with improved average daily gain (ADG) in offspring (P < 0.05, Fig. 4H). Additionally, it significantly increased individual body weight at P8, P15, and P22 (P < 0.01, Fig. S3J-L), and a tendency toward increased litter weight was observed at P8 (P = 0.090, Fig. 4E). It also did not affect organ indices in offspring at birth and weaning (P > 0.05, Fig. 4F, 4G, S3M-N).Fig. 4Thiamine improves reproductive performance in rats. A Maternal rats were supplemented with thiamine throughout gestation and lactation, and placentas were collected on E18. B Maternal weight changes during pregnancy and lactation, n = 10. C Maternal feed intake during pregnancy, n = 10. D Maternal litter size, n = 10. E Litter weight of offspring at different times after birth, n = 10. F Organ index of male born offspring, n = 10. G Organ index of male offspring at weaning, n = 10. H ADG of offspring during lactation, n = 10. *P < 0.05, **P < 0.01 Thiamine improves reproductive performance in rats. A Maternal rats were supplemented with thiamine throughout gestation and lactation, and placentas were collected on E18. B Maternal weight changes during pregnancy and lactation, n = 10. C Maternal feed intake during pregnancy, n = 10. D Maternal litter size, n = 10. E Litter weight of offspring at different times after birth, n = 10. F Organ index of male born offspring, n = 10. G Organ index of male offspring at weaning, n = 10. H ADG of offspring during lactation, n = 10. *P < 0.05, **P < 0.01 Thiamine supplementation modulates the composition of maternal colonic microbiota As shown in Fig. 5A, the C group exhibited 2954 ASVs, and the HT group had 3366 ASVs, with 1425 ASVs shared between groups. No significant differences in α and β diversity were found (P > 0.05; Fig. 5B-F). Figure 5G displays phylum-level microbial abundance via a chord plot, whereas Fig. 5H shows the 30 most abundant genera via a chord plot. The relative abundances of Globicatella, Family_XIII_AD3011_group, Atopostipes and Rothia were significantly decreased in the HT group (P < 0.05, Fig. 5I). Janibacter, Ruminococcus torques group, Duncaniella, Bacteroidales_RF16_group_unclassified, Pseudomonas, DNF00809, Prevotellaceae_Ga6A1_group, Acinetobacter and Faecalimonas were enriched in the HT group (P < 0.05, Fig. 5I). LEfSe analysis indicated that Prevotellaceae_Ga6A1_group and Bacteroidales_RF16_group_unclassified could serve as biomarkers for the HT group (Fig. 5J). Thiamine did not significantly alter the metabolic phenotype of the rat colonic microbiota (P > 0.05, Fig. 5K). PICRUSt analysis based on the COG database revealed significantly higher abundances in the HT group of Threonine/homoserine/homoserine lactone efflux protein and C-terminal processing protease CtpA (P < 0.05, Fig. S4A). In addition, thiamine significantly increased the butyrate levels in the maternal colonic contents (P < 0.05, Fig. 5N), while it did not significantly affect the levels of other SCFAs (P > 0.05, Fig. 5L-M, O-R).Fig. 5Thiamine modulates the composition of the maternal colonic microbiota at weaning. A Venn diagram, n = 6. B-D Comparison of Shannon (B), Chao1 (C) and Observed species (D) indices, n = 6. E PCoA plot of fecal microbiota, n = 6. F The Bray–Curtis rank within each group, n = 6. G Relative abundance and composition of fecal microbiota at the phylum level, n = 6. H Relative abundance and composition of the top 30 genus levels of fecal microbiota abundance, n = 6. I Significantly different fecal microbes at the genus level, n = 6. J LEfSe analysis based on significantly different genus, n = 6. K Prediction of microbial metabolic phenotypes using BugBase platform, n = 6. L-R Levels of SCFAs in feces, n = 8. *P < 0.05 Thiamine modulates the composition of the maternal colonic microbiota at weaning. A Venn diagram, n = 6. B-D Comparison of Shannon (B), Chao1 (C) and Observed species (D) indices, n = 6. E PCoA plot of fecal microbiota, n = 6. F The Bray–Curtis rank within each group, n = 6. G Relative abundance and composition of fecal microbiota at the phylum level, n = 6. H Relative abundance and composition of the top 30 genus levels of fecal microbiota abundance, n = 6. I Significantly different fecal microbes at the genus level, n = 6. J LEfSe analysis based on significantly different genus, n = 6. K Prediction of microbial metabolic phenotypes using BugBase platform, n = 6. L-R Levels of SCFAs in feces, n = 8. *P < 0.05 Thiamine supplementation elevates maternal-offspring thiamine levels and promotes placental development There were no major changes in plasma glucose, CHE, or ALP levels in maternal or offspring at weaning between the HT and C groups (P > 0.05, Fig. 6A-B). At weaning, the HT group had markedly higher thiamine levels in the maternal jejunum and ileum (P < 0.01, Fig. 6C). Additionally, their offspring showed elevated serum thiamine concentrations both at birth and at weaning (P < 0.01, Fig. 6D). To systematically evaluate the effect of thiamine on placental development, we conducted hematoxylin and eosin staining followed by histomorphometric analysis of the placenta (Fig. 6E, F). The results showed that thiamine supplementation did not remarkably alter the total placental area, spongiotrophoblast area, or labyrinth area (P > 0.05, Fig. 6G-I). Notably, thiamine significantly increased the ratio of labyrinth area to total placental area (P < 0.05, Fig. 6J), with a concurrent tendency to increase the area of blood sinus in the labyrinth (P = 0.056, Fig. 6K). Furthermore, our findings revealed no apparent alterations in mRNA expression related to the nuclear factor erythroid 2 p45-related factor 2 (Nrf2)/Kelch-like ECH-associated protein l (Keap1) signaling pathway in the placenta (P > 0.05, Fig. S5A-B). Analysis of antioxidant function showed that thiamine supplementation also failed to alter mRNA levels of antioxidant enzymes or the levels of indicators related to placental antioxidant function (P > 0.05, Fig. S5C-K).Fig. 6Thiamine promotes placental development in rats. A, B Maternal (A) and offspring (B) serum levels of glucose, CHE and ALP at weaning, n = 8. C Thiamine content of maternal jejunum and ileum, n = 8. D Serum thiamine content at birth and weaning of offspring, n = 8. E, F HE-stained section of the placenta; scale bar: 1000 μm (E), 50 μm (F), n = 4. G Total area of the placenta, n = 4. H Area of placental spongiotrophoblast, n = 4. I Area of placental labyrinth, n = 4. J Ratio of labyrinth to total placental area, n = 4. K Area of blood sinus in labyrinth, n = 4. *P < 0.05, **P < 0.01 Thiamine promotes placental development in rats. A, B Maternal (A) and offspring (B) serum levels of glucose, CHE and ALP at weaning, n = 8. C Thiamine content of maternal jejunum and ileum, n = 8. D Serum thiamine content at birth and weaning of offspring, n = 8. E, F HE-stained section of the placenta; scale bar: 1000 μm (E), 50 μm (F), n = 4. G Total area of the placenta, n = 4. H Area of placental spongiotrophoblast, n = 4. I Area of placental labyrinth, n = 4. J Ratio of labyrinth to total placental area, n = 4. K Area of blood sinus in labyrinth, n = 4. *P < 0.05, **P < 0.01 Thiamine enhances placental barrier function and improves nutrient transport and utilization As shown in Fig. 7A-C, thiamine significantly enhanced ZO-1 transcription and protein synthesis in the placenta (P < 0.05). Thiamine promoted the mRNA and protein expression of glucose transporter type 1 (GLUT1) and fatty acid transport protein (FATP) 1 (P < 0.05, Fig. 7D-E), and tended to increase the mRNA expression of sodium-coupled neutral amino acid transporter 1 (SNAT1) (P = 0.086, Fig. 7D) and FATP4 (P = 0.071, Fig. 7D). Additionally, thiamine upregulated the mRNA expression of genes related to lipid metabolism (hormone-sensitive lipase (HSL), lysosomal acid lipase (LAL) and endothelial lipase (EL)) and glucose metabolism (glycogen phosphorylase M (PYGM) and Glycogen synthase 1 (Cys-1)) in the placenta, thereby enhancing the ability of the placenta to utilize lipids and glucose (P < 0.05, Fig. 7F-G).Fig. 7Thiamine strengthens placental barrier integrity and facilitates nutrient utilization. A‚ B Representative images (A) and quantitative analysis (B) of ZO-1 IHC in the placenta, n = 4; scale bar: 50 μm. C The mRNA expression of placental tight junction proteins, n = 8. D Placental mRNA expression of nutrient transporter genes, n = 8. E Protein content of GLUT1 and FATP1 in the placenta, n = 8. F Placental mRNA expression of lipid metabolism genes, n = 8. G Placental mRNA expression of glucose metabolism genes, n = 8. *P < 0.05, **P < 0.01 Thiamine strengthens placental barrier integrity and facilitates nutrient utilization. A‚ B Representative images (A) and quantitative analysis (B) of ZO-1 IHC in the placenta, n = 4; scale bar: 50 μm. C The mRNA expression of placental tight junction proteins, n = 8. D Placental mRNA expression of nutrient transporter genes, n = 8. E Protein content of GLUT1 and FATP1 in the placenta, n = 8. F Placental mRNA expression of lipid metabolism genes, n = 8. G Placental mRNA expression of glucose metabolism genes, n = 8. *P < 0.05, **P < 0.01 Thiamine promotes placental angiogenesis and increases the levels of thiamine-related metabolic enzymes Figure 8 illustrates the expression levels of placental angiogenesis-related markers. Thiamine supplementation upregulated the mRNA and protein expression of angiogenesis-related genes (placental growth factor (PLGF), vascular endothelial growth factor receptor (VEGFR) 2, Hand1, Hand2 and fos-like antigen 1 (Fosl1)) in the placenta (P < 0.05, Fig. 8A-C). Intriguingly, thiamine treatment significantly enhanced the migratory capacity of pTr cells (P < 0.05, Fig. 8D). Meanwhile, 0.8 mmol/L thiamine treatment did not significantly impact the survival rate of pTr cells (P > 0.05, Fig. S5L). Thiamine similarly enhanced succinate dehydrogenase subunit A (SDHA) mRNA expression (P < 0.05, Fig. 8E). Placental mRNA levels of SLC19A2, SLC25A19, and thiamine pyrophosphokinase1 (TPK1) were markedly higher in the HT group (P < 0.05, Fig. 8F). Moreover, thiamine dramatically raised the levels of thiamine, PDH, α-KGDHC and acetyl-CoA in the placenta (P < 0.05, Fig. 8G-I, K).Fig. 8Thiamine enhances placental angiogenesis and increases the levels of enzymes related to energy metabolism. A, B Placental mRNA expression of angiogenesis genes, n = 8. C Placental expression of angiogenesis-associated proteins, n = 8. D Analysis of wound healing assay with pTr, n = 3; scale bar: 200 μm. E Placental mRNA expression of mitochondrial respiratory chain genes, n = 8. F Placental mRNA expression of thiamine transport and metabolism genes, n = 8. G Thiamine levels in the placenta, n = 8. H-J The levels of PDH (H), α-KGDHC (I), TKT (J) and acetyl-CoA (K) in the placenta, n = 8. *P < 0.05, **P < 0.01 Thiamine enhances placental angiogenesis and increases the levels of enzymes related to energy metabolism. A, B Placental mRNA expression of angiogenesis genes, n = 8. C Placental expression of angiogenesis-associated proteins, n = 8. D Analysis of wound healing assay with pTr, n = 3; scale bar: 200 μm. E Placental mRNA expression of mitochondrial respiratory chain genes, n = 8. F Placental mRNA expression of thiamine transport and metabolism genes, n = 8. G Thiamine levels in the placenta, n = 8. H-J The levels of PDH (H), α-KGDHC (I), TKT (J) and acetyl-CoA (K) in the placenta, n = 8. *P < 0.05, **P < 0.01 Thiamine activates the Notch-PI3K/AKT pathway in the placenta We investigated whether thiamine activated the Notch signaling pathway to promote placental angiogenesis (Fig. 9). Our data demonstrated that thiamine significantly upregulated the mRNA expression levels of Notch receptors (Notch1, Notch2, Notch3 and Notch4), ligands (Jagged1 and Delta-like (DLL) 4) and their downstream target genes (hairy and enhancer of split (Hes) 1, Hes5, Hes-related with YRPW motif (Hey) 1) in the placental (P < 0.05, Fig. 9A-C). Additionally, thiamine enhanced the protein abundance of Notch1, Notch intracellular domain (NICD), Jagged1, DLL4, and Hes1 (P < 0.05, Fig. 9D). Notably, thiamine elevated the mRNA and protein levels of insulin-like growth factor 1 receptor (IGF-1R) in the placenta (P < 0.05, Fig. 9E-G). Thiamine treatment markedly upregulated the mRNA levels of Phosphatidylinositol-3-kinase (PI3K), AKT, and cyclin D1 in the HT group (P < 0.05, Fig. 9G). Furthermore, thiamine enhanced the placental protein content of PI3K, 3-phosphoinositide-dependent protein kinase-1 (PDK1), and cyclin D1, accompanied by an elevated phosphorylated AKT (P-AKT)/total AKT (T-AKT) ratio (P < 0.05, Fig. 9H).Fig. 9Thiamine facilitates placental development by activating the Notch-PI3K/AKT signaling pathway. A Placental mRNA expression of Notch receptor genes, n = 8. B Placental mRNA expression of Notch ligand genes, n = 8. C Placental mRNA expression of Notch pathway downstream targets, n = 8. D Placental expression of Notch pathway-associated proteins, n = 8. E, F Representative images (E) and quantitative analysis (F) of IGF-1R IHC in the placenta, n = 4; scale bar, 50 μm. G Placental mRNA expression of PI3K-Akt signaling pathway genes, n = 8. H Placental levels of PI3K-Akt pathway-associated proteins, n = 8. *P < 0.05, **P < 0.01 Thiamine facilitates placental development by activating the Notch-PI3K/AKT signaling pathway. A Placental mRNA expression of Notch receptor genes, n = 8. B Placental mRNA expression of Notch ligand genes, n = 8. C Placental mRNA expression of Notch pathway downstream targets, n = 8. D Placental expression of Notch pathway-associated proteins, n = 8. E, F Representative images (E) and quantitative analysis (F) of IGF-1R IHC in the placenta, n = 4; scale bar, 50 μm. G Placental mRNA expression of PI3K-Akt signaling pathway genes, n = 8. H Placental levels of PI3K-Akt pathway-associated proteins, n = 8. *P < 0.05, **P < 0.01 Discussion The extensive microbiota in the mammalian gut regulates host physiological functions by its metabolites [30]. Evidence indicates that the maternal gut microbiota regulates embryonic development [31, 32]. Here, we assessed the impact of sow fecal microbiota with varying reproductive potentials on host metabolic profiles. We further employed multi-omics integration to deconstruct how lineage-specific metabolites coordinate developmental crosstalk at the maternal-placental-fetal interface. Metabolomic analysis of feces revealed significantly elevated thiamine levels in sows with high reproductive potential. KEGG enrichment analysis indicated activation of the thiamine metabolic pathway in sows with high reproductive potential. Furthermore, the actual concentration of thiamine in the feces of sows with high reproductive potential was also found to be significantly increased. Although this study did not detect significant differences in overall fecal microbial diversity or the abundance of dominant genera among sows with differing reproductive potentials, we observed a notable enrichment of Lactococcus in the H group. This low-abundance genus could serve as a potential biomarker for high reproductive potential, aligning with emerging evidence that a microbe's functional importance in regulating host phenotypes is not dependent on its relative abundance [33]. The ability of Lactococcus to produce various beneficial metabolites through proteolysis is well-documented [34]. Due to the evolutionary loss of biosynthetic pathways, mammals are obligatorily reliant on dietary or microbially-derived thiamine [35]. Notably, previous studies have demonstrated that Lactococcus exhibits the capacity for de novo synthesis of thiamine [36, 37]. In microbial ecosystems, certain low-abundance microorganisms are considered keystone taxa because they influence the structure and stability of microbial communities through the production of metabolites [38]. This may explain why Lactococcus, despite its low abundance, is associated with high reproductive potential. In this study, we maintained sows under strictly uniform dietary conditions, thereby effectively eliminating dietary intake as a major source of variation in fecal thiamine levels. Thus, the significantly elevated thiamine levels in high-reproductive-potential sows are more likely due to the metabolic activities of their gut microbiome. It is important to emphasize that microbial metabolic functions, particularly complex multi-step processes like thiamine biosynthesis, often depend on collaboration and interaction at the community level rather than being achieved by a single microbial strain alone [39, 40]. Such intricate community-level effects are challenging to replicate fully in simplified in vitro culture systems. Therefore, the proposed hypotheses require further validation through experiments such as fecal microbiota transplantation to establish precise causal relationships. Collectively, these findings propose a novel microbiota-mediated mechanism: gut microbes may enhance host reproductive potential by providing thiamine. Notably, we did not observe a concurrent increase in plasma thiamine. This is consistent with thiamine's rapid uptake by erythrocytes and tissues upon absorption [35], suggesting that the microbiome-derived thiamine is likely efficiently utilized by the host, rather than accumulating in circulation. Arginine has been shown to enhance placental blood flow and promote placental development [41, 42]. We noticed a marked enrichment of arginine in plasma of sows with high reproductive potential. Arginine and thiamine may collectively contribute to enhanced reproductive potential, as the improvement in placental development mediated by arginine may facilitate thiamine transfer to the fetus and enhance the placenta's utilization of thiamine. Conversely, our data indicated reduced citrate concentrations in the plasma of the H group. This finding aligns with KEGG enrichment analysis showing significant activation of the citrate cycle. As citrate is a key intermediate rapidly consumed in the citrate cycle, its depletion indicates enhanced metabolic flux through this cycle, which potentially reflects accelerated energy production needed to sustain reproductive processes. Together, these coordinated metabolic alterations reflect an adaptive physiological mechanism that supports the elevated demands of gestation in high-reproductive-potential sows. Subsequently, this study aimed to verify the direct effects of thiamine on maternal reproductive potential and its underlying mechanisms. For this purpose, a relatively high supplementary dose (200 mg/kg BW) was selected for the rat model. The choice of this dose was primarily based on a previous study [26], which confirmed that thiamine had a high safety margin in lactating rats, with no adverse effects observed even at elevated doses. Maternal thiamine supplementation enhanced litter weight on days 15 and P22 and improved ADG throughout the birth-to-weaning period. These findings contrast with a prior study reporting that lactational thiamine supplementation did not affect offspring body weight in rats [26]. This divergence may reflect differences in supplementation timing. To investigate the underlying mechanisms mediating these beneficial effects on offspring development, we focused on maternal gut microbiota. Thiamine, an essential micronutrient for microorganisms, plays a crucial role in maintaining gut microbial balance. Previous studies have demonstrated that high-dose thiamine supplementation alleviates dysbiosis of the gut microbiota induced by high-fat/high-fructose diets in mouse models [43]. We found that thiamine significantly enriched Prevotellaceae Ga6A1 group and Bacteroidale RF16 group unclassified, which served as potential biomarkers for the HT group. The Prevotellaceae Ga6A1 group is widely recognized as a beneficial bacterium and has been linked to elevated butyrate production [44, 45]. Correspondingly, we observed significantly increased levels of butyrate in maternal colonic contents. It has been shown that butyrate facilitates offspring development, enhances gut health, inhibits histone deacetylase activity, and suppresses inflammatory responses in the host organism [46–48]. Furthermore, we observed a significant upregulation of the C-terminal processing protease CtpA in the HT group. Research indicates that elevated levels of the C-terminal processing protease CtpA may reduce the virulence potential of pathogenic bacteria [49]. The upregulation of C-terminal processing protease CtpA suggests a mechanism by which thiamine may mitigate the adverse effects of potentially pathogenic gut microbiota on maternal health, thereby contributing to a healthier maternal environment. These data suggest that maternal thiamine supplementation supports offspring growth and maturation by modulating the maternal gut microbial community, leading to an improved maternal physiological state conducive to fetal and early postnatal development. The placenta regulates fetal-maternal substance exchange via selective permeability mechanisms, effectively blocking maternal-derived pathogens, toxins, and harmful compounds, while facilitating bidirectional exchange of oxygen, nutrients, and metabolic wastes. Our data showed that supplementation with high doses of thiamine significantly enhanced its retention within the maternal-placental-fetal system, although thiamine has a relatively short metabolic half-life (< 12 h) in animals [50]. This finding indicated that maternal thiamine supplementation facilitates transplacental transfer of thiamine to the fetus. Placental nutrient transport capacity primarily depends on placental total area, blood flow volume, and vascular density. Structural abnormalities in the placental labyrinth impair nutrient transport efficiency, leading to placental dysfunction and significantly elevated risks of intrauterine growth restriction in offspring [51, 52]. Our data revealed that thiamine markedly elevated the labyrinth/total placenta area ratio and showed a trend toward expansion of the placental blood sinus area, suggesting that thiamine enhances maternal–fetal nutrient exchange to meet the fetus's escalating nutritional demands. However, it should be noted that the analysis of placental structural parameters, including the total placental area and the labyrinth area, was conducted with a relatively small sample size. This may have limited the statistical power of these morphological assessments and increased the risk of Type II error. Consequently, the actual promoting effects of thiamine on placental development, particularly structural adaptations such as the labyrinth/total placenta area ratio and the placental blood sinus area, might have been underestimated or obscured in our statistical evaluation. During gestation, the progressive rise in metabolic activity drives the gradual buildup of reactive oxygen species (ROS). This ROS overload induces late-gestational oxidative stress, precipitating maternal physiological dysregulation and oxidative damage [53, 54]. Our data demonstrated that maternal thiamine supplementation failed to activate the Nrf2/Keap1 signaling pathway in the placenta, nor did it induce notable alterations in antioxidant biomarkers. Compromised placental barrier function disrupts the exchange of nutrients and metabolic wastes, increases the risk of fetal exposure to toxins, and may culminate in adverse pregnancy outcomes [55]. ZO-1 localizes to placental trophoblast cell junctions and is essential for placental barrier integrity. ZO-1 knockout mice exhibit severe defects in placental development [56]. Our results revealed that thiamine enhances the placental barrier by upregulating ZO-1 expression in the placenta. Glucose, fatty acids, and amino acids constitute essential nutrients required for placental and fetal development, whose transplacental transfer efficiency critically regulates fetal development. GLUT1 acts as the master regulator of transplacental glucose shuttling, whereas the FATP family not only mediates fatty acid transfer from mother to fetus but also regulates angiogenesis in the placenta [57, 58]. Our data suggested that thiamine enhances the transport efficiency of glucose and fatty acids by upregulating both mRNA and protein expression of placental GLUT1 and FATP1. Additionally, thiamine dynamically sustains fetal energy supply by coordinately activating placental lipid and glycogen metabolic networks. Specifically, thiamine markedly upregulates the mRNA expression of HSL, LAL and EL, which could promote triglyceride hydrolysis into free fatty acids and glycerol to enhance lipolytic efficiency. Concurrently, thiamine enhanced the mRNA expression of PYGM and GYS1, implying a reinforcement of glycogen turnover. This coordinated regulation of lipid and glycogen metabolism likely enhances placental energy production and metabolic flexibility. To elucidate the mechanisms by which thiamine promotes placental development, we conducted comprehensive analyses of thiamine transporters and metabolizing enzymes in the placenta. As anticipated, thiamine promoted mRNA expression of SLC19A2, SLC25A19, and TPK1 in the placenta. In animals, thiamine is transported into cells via SLC19A2 and phosphorylated to TPP (the predominant bioactive form) by TPK1 [35]. Following import into mitochondria by SLC25A19, TPP participates in the catalytic activities of PDH and α-KGDHC. These enzyme complexes catalyze the oxidative decarboxylation of pyruvate to acetyl-CoA and α-ketoglutarate to succinyl-CoA, respectively [59]. Interestingly, we observed that thiamine raised the levels of PDH and α-KGDHC in the placenta, leading to a concomitant elevation in acetyl-CoA production. Acetyl-CoA has been found to enhance the metabolic flux of the citrate cycle [60, 61]. The NADH and FADH₂ produced by the citrate cycle fuel oxidative phosphorylation through the mitochondrial electron transport chain, which ultimately converts ADP to ATP [62]. Critical for mitochondrial energy metabolism, succinate dehydrogenase uniquely links the citrate cycle and respiratory chain [63]. Our findings demonstrated that thiamine enhances SDHA mRNA expression in the placenta, promoting oxidative phosphorylation via enhanced electron transport chain activity. Overall, maternal thiamine supplementation boosts placental thiamine levels and activates associated metabolic enzymes, which synergistically enhance citrate cycle activity and promote mitochondrial energy production. Emerging evidence highlights the role of thiamine in epigenetic regulation through acetyl-CoA-mediated histone acetylation. Specifically, thiamine elevates cellular acetyl-CoA content (via TPP-dependent metabolic pathways), which serves as a substrate for histone acetyltransferases to enhance histone acetylation levels [23, 64]. Notably, acetyl-CoA derived from glycolytic metabolism promotes syncytialization of human trophoblast stem cells by activating histone acetylation [65]. Angiogenesis in the placenta is co-regulated by growth factors such as VEGF, VEGFR and PLGF [66]. VEGFA mainly binds to VEGFR1 and VEGFR2, while VEGFB and PLGF interact with VEGFR1. This interaction activates downstream signaling pathways, thereby promoting both vasculogenesis and angiogenesis [67]. Studies found that benfotiamine (a thiamine analog) promoted myocardial angiogenesis in diabetic mice by enhancing the phosphorylation of VEGFR2 and Akt [68]. Previous studies indicated that histone acetylation upregulates VEGFR2 in endothelial cells, thereby driving angiogenesis [69]. Our findings revealed that thiamine enhanced the mRNA and protein expression of VEGFR2 and PLGF, thereby promoting placental angiogenesis, which may be mediated through acetyl-CoA-dependent histone acetylation. Hand1 and Hand2 are essential fundamental helix-loop-helix transcription factors that regulate placental angiogenesis. It has been found that Hand1 deficiency during early pregnancy results in embryonic lethality and destroys the development of the labyrinth and syncytialization, whereas deficiency in late pregnancy leads to abnormal placental vascular dilation and impaired remodeling [70]. Hand2 regulates pregnancy length at the maternal–fetal interface through modulation of IL-15 expression levels [71]. Previous studies have shown that prenatal acrylamide exposure suppresses Hand1 and Hand2 expression, thereby impeding placental angiogenesis [72]. Fosl1 promotes placental maturation by increasing the expression of genes related to placental development [73]. Our data suggested that thiamine enhanced the levels of mRNA and protein expression of Hand1, Hand2, and Fosl1 in the placenta. In addition, our in vitro experiment further validated that thiamine enhances the migration capability of pTr cells, contributing to placental folding. Collectively, these findings suggest that thiamine coordinately promotes placental angiogenesis and vasculogenesis through multiple pathways. The Notch pathway promotes placental angiogenesis through coordinated proliferation of endothelial and trophoblast cells [74]. This pathway comprises transmembrane receptors (Notch1-4), ligands (DLL and Jagged families), and downstream target genes. Ligand engagement triggers Notch receptor restructuring, permitting γ-secretase cleavage and subsequent NICD release [75]. The NICD then translocates to the nucleus, binding to the DNA-binding protein CSL, and activating transcription of target genes [75]. In placental vascular development, vasculogenesis and angiogenesis are orchestrated by tip cells (pioneering cells that guide angiogenic sprouting) and endothelial stalk cells (trailing cells that stabilize nascent vessels), which collectively establish functional vascular networks [76]. A critical regulatory loop involves VEGFA binding VEGFR2 to drive DLL4 expression and tip cell sprouting. Crucially, tip cell-derived DLL4 activates Notch signaling in neighboring stalk cells, leading to Jagged1 upregulation. Jagged1 then competitively inhibits DLL4-Notch1 binding in stalk cells, preventing signal propagation back to tip cells [19]. This ensures sustained VEGF receptor expression specifically in tip cells, facilitating placental angiogenesis. Additionally, acetylation levels within cells help promote the ongoing activation of the Notch signaling pathway [77]. A clinical study showed that the inhibition of Notch1 and Jagged1 expression in the placenta mediates the occurrence of fetal growth restriction [78]. In our study, thiamine increased mRNA levels of Notch and enhanced protein expression of Notch1, DLL4, and Jagged1 in the placenta. This thiamine-driven Notch pathway activation stimulated downstream target transcription. Our findings indicate that thiamine activates the Notch pathway through VEGFR2 activation and elevated acetyl-CoA levels, thereby enhancing intracellular acetylation to sustain pathway activity. This coordinated mechanism ultimately stimulates sprouting of tip cells, maintains endothelial homeostasis, and supports functional vascular network development. The PI3K/AKT pathway drives energy metabolism through the regulation of phosphorylation levels in nutrient transport proteins and metabolic enzymes [79]. Upon binding of growth factors to transmembrane receptors, the membrane-localized PI3K is activated to convert phosphatidylinositol 4,5-bisphosphate (PIP2) to phosphatidylinositol-3,4,5-trisphosphate (PIP3). Then, PIP3 activates PDK1, which phosphorylates AKT, thereby initiating a signaling cascade that activates downstream effectors to modulate nutrient metabolism and cellular proliferation [80]. Beyond its direct regulation by growth factors, the PI3K/AKT pathway is activated by upstream signaling, including VEGFR2 (which indirectly activates PI3K) and Notch signaling (which upregulates IGF-1R to activate PI3K and enhance glycolytic flux, meeting embryonic energy demands) [67, 81, 82]. Intriguingly, our data reveal that thiamine upregulates placental IGF-1R expression. At the same time, thiamine activates the PI3K/AKT pathway by elevating the expression of PI3K and AKT. Furthermore, thiamine induced a marked elevation in PDK1 protein expression, which facilitates AKT phosphorylation. Cyclin D1 governs placental mitogenesis as a PI3K/AKT pathway downstream mediator [83]. We observed that thiamine also enhanced cyclin D1 expression in the placenta, promoting placental cell proliferation. Our results indicate that thiamine synergistically activates the PI3K/AKT pathway through VEGFR2 and Notch signaling, thereby orchestrating placental nutrient trafficking and cellular proliferation to support placental and fetal development. While our findings elucidate a potential mechanistic link between gut microbiota-derived thiamine and placental and fetal development, it is important to note several limitations of this study. First, although our multi-omics analysis revealed a significant correlation between maternal gut microbiota-derived thiamine and improved reproductive outcomes, the precise causal relationship has not been fully elucidated. Second, the validation in this study primarily relied on a rat model. While rats represent a well-established and widely used model in reproductive research, there are inherent interspecies differences in placental structure and function compared with pigs or humans. Therefore, the mechanisms identified here must be interpreted with caution when extrapolated to other species and require further validation through cross-species comparative studies. Finally, the intervention dose of 200 mg/kg BW far exceeds the physiological increases observed in the feces of high-performance sows. It should be emphasized that the mode of action under this high-dose intervention may extend beyond purely physiological regulation, partially revealing amplifying effects at pharmacological concentrations. Consequently, translating these findings into general dietary recommendations requires caution. Future studies are needed to establish the thiamine dose–response relationship, aiming to determine an optimal, physiologically relevant dosage that maximizes health benefits for both mother and offspring without reaching pharmacological levels. Conclusion In conclusion, this study characterized the distinct metabolic profiles and microbial compositions in sows with different reproductive potential. We identified thiamine as a key metabolite and validated its role in promoting placental angiogenesis and fetal development. Specifically, our findings demonstrate that thiamine enhances offspring development by modulating the maternal gut microbiota composition, which subsequently facilitates the production of SCFAs. Furthermore, thiamine improves placental barrier integrity and nutrient transport efficiency in pregnant rats, consequently improving offspring growth. Mechanistically, thiamine promotes placental angiogenesis by activating the Notch signaling pathway and enhancing placental metabolic plasticity. These findings provide novel insights supporting thiamine supplementation as a potential nutritional strategy to optimize reproductive outcomes. Supplementary Information Additional file 1: Supplementary information includes five figures and six tables. Figure S1. Functional prediction and analysis of the fecal microbiota of sows with different reproductive performance. Figure S2. Traceability analysis of metabolites in sows with different reproductive performance and functional analysis based on metabolite origin by MetOrigin. Figure S3. Effect of thiamine on the reproductive performance of rats. Figure S4. Functional prediction and analysis of the microbiota of rat colon contents. Figure S5. Effect of thiamine on antioxidant function and Nrf2 signaling pathway-related genes in the placenta. Table S1. Total litter size, live births, stillbirths, and mummified piglets in sows with different reproductive potential. Table S2. Composition of experimental diets. Table S3. The real-time PCR primers. Table S4. Information for primary and secondary antibodies. Table S5. Significantly different secondary metabolites in faces. Table S6. Significantly different secondary metabolites in plasma. Additional file 1: Supplementary information includes five figures and six tables. Figure S1. Functional prediction and analysis of the fecal microbiota of sows with different reproductive performance. Figure S2. Traceability analysis of metabolites in sows with different reproductive performance and functional analysis based on metabolite origin by MetOrigin. Figure S3. Effect of thiamine on the reproductive performance of rats. Figure S4. Functional prediction and analysis of the microbiota of rat colon contents. Figure S5. Effect of thiamine on antioxidant function and Nrf2 signaling pathway-related genes in the placenta. Table S1. Total litter size, live births, stillbirths, and mummified piglets in sows with different reproductive potential. Table S2. Composition of experimental diets. Table S3. The real-time PCR primers. Table S4. Information for primary and secondary antibodies. Table S5. Significantly different secondary metabolites in faces. Table S6. Significantly different secondary metabolites in plasma.