Maternal obesity modifies the vertical transmission of gut microbiota and metabolites: Insights from mother-infant dyads in a mexican cohort

Abstract

Maternal obesity is increasingly recognized as an important modulator of early-life microbial and metabolic environments. This study investigates the association between maternal body mass index (BMI) and the microbiota and metabolite profiles of colostrum and neonatal feces in a Mexican mother–infant cohort. Milk and fecal samples were collected from dyads of obese and normal-weight mothers. Bacterial microbiota composition was characterized by sequencing the 16 S rRNA gene (V3 region) using Ion Torrent technology, and metabolomic profiling was performed using Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry. The results provide evidence consistent with vertical microbial and metabolic transmission, with Firmicutes and Patescibacteria predominating in Colostrum. Neonates born to obese mothers exhibited reduced relative abundances of Lactobacillus in neonatal fecal samples, alongside increased levels of Lactobacillus and Staphylococcus in both colostrum and mother´s feces. Overall microbial diversity across maternal stool, colostrum, and neonatal stool samples was not significantly associated with maternal BMI; however, distinct metabolite signatures linked to maternal obesity, like oligopeptides, glycoside-related compounds, Phosphatidic Acid (PA) Derivatives, bioactive molecules such as enkephalinamide derivatives, were observed. These findings highlight the role of breastfeeding as a key interface in shaping the neonatal gut microbiota and metabolome and suggest potential pathways linking maternal metabolic status with early-life microbial and metabolic programming. This study advances understanding of maternal–infant microbial ecology and supports further investigation into the long-term health implications of maternal obesity.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-49183-x.

Introduction

Maternal obesity is a major public health concern in Mexico, where prevalence among women of reproductive age remains among the highest worldwide. National data from the Encuesta Nacional de Salud y Nutrición (ENSANUT) indicate that approximately 40% of women are obese, reflecting a substantial burden during the reproductive period¹. This epidemiological context is further complicated by the coexistence of food insecurity and malnutrition, creating a complex and heterogeneous nutritional landscape. Maternal obesity has been consistently associated with increased risk of cardiometabolic diseases, including type 2 diabetes and cardiovascular disorders, as well as adverse reproductive outcomes such as infertility and pregnancy-related complications². In addition, the high prevalence of childhood overweight and obesity in Mexico—affecting approximately 37.3% of children aged 5 to 11—raises concerns about intergenerational transmission of metabolic risk³. These observations underscore the importance of maternal metabolic status during pregnancy as a key determinant of maternal health and early-life developmental trajectories, in line with the developmental origins of health and disease (DOHaD) framework^4–6.

The establishment of the neonatal gut microbiome is a critical process influencing immune maturation, metabolic homeostasis, and long-term health outcomes. Multiple maternal and perinatal factors, including maternal microbiota, delivery mode, environmental exposures, and infant feeding practices, shape early microbial colonization. Among these, vertical transmission of microorganisms and metabolites from mother to infant represents a primary mechanism of early-life microbial establishment, occurring during pregnancy, delivery, and breastfeeding^7–9. Human milk, particularly colostrum, plays a central role in this process by serving as a key biological interface for maternal–infant transfer. In addition to providing essential nutrients, human milk contains diverse microbial communities and a wide array of bioactive compounds that contribute to microbial colonization and functional maturation of the neonatal gut. These include human milk oligosaccharides (HMOs), short-chain fatty acids (SCFAs), and immune-modulating molecules that support gut barrier integrity, regulate host immune responses, and shape microbial community structure^10,11.

The composition and functional potential of the milk microbiome and metabolome are influenced by maternal characteristics, including body mass index (BMI), diet, and metabolic status. Previous studies have demonstrated that maternal obesity is associated with alterations in microbial diversity, shifts in specific bacterial taxa, and changes in metabolite profiles that may affect neonatal gut colonization and immune development^12,13. These alterations may influence not only microbial transmission but also the biochemical environment encountered by the neonate, thereby contributing to early-life metabolic programming. In this context, human milk represents a dynamic system that integrates maternal physiological and environmental signals, with potential implications for offspring health^14–16.

In addition to maternal metabolic status, diet is a major environmental factor shaping the composition and activity of the gut microbiota. Dietary patterns rich in fiber, complex carbohydrates, and plant-derived compounds promote the growth of beneficial commensal bacteria such as Bifidobacterium, Lactobacillus, and Faecalibacterium, which produce SCFAs including acetate, propionate, and butyrate. These metabolites play essential roles in maintaining intestinal barrier function, modulating immune responses, and regulating host energy metabolism^17–19. Conversely, dietary patterns characterized by high fat, high refined sugar, and low fiber intake—commonly associated with Westernized diets—have been linked to reduced microbial diversity and increased abundance of pro-inflammatory taxa, contributing to metabolic dysfunction and insulin resistance^17,20. These dietary influences extend beyond microbial composition to affect microbial metabolic outputs, including bile acid transformation, amino acid metabolism, and lipid metabolism, thereby shaping the overall metabolomic landscape^21,22.

Disruptions in microbial composition and metabolite production, particularly in the context of maternal obesity, may have important implications for neonatal health. Obesity-associated dysbiosis, often characterized by reduced microbial diversity and enrichment of pro-inflammatory taxa, can alter vertical transmission processes and impair neonatal microbiome development^23–25. Additionally, alterations in key metabolites such as SCFAs, bile acids, and branched-chain amino acids (BCAAs) may influence immune development, metabolic regulation, and gut barrier function in early life^{13,19,26–28}. These changes may increase susceptibility to metabolic and immune-related disorders, highlighting the importance of understanding how maternal metabolic and dietary factors shape early-life microbial and metabolic environments.

Despite increasing recognition of the importance of maternal–infant microbial and metabolic interactions, significant gaps remain in understanding how maternal BMI influences the integrated dynamics of the microbiome and metabolome across multiple interconnected compartments. Most existing studies have focused on individual biological matrices or single omics layers, limiting the ability to capture coordinated interactions across systems. In particular, the combined analysis of maternal stool, colostrum, and neonatal stool within the same mother–infant dyads remains limited. Furthermore, evidence from underrepresented populations, including Mexican cohorts, is scarce despite the high prevalence of maternal obesity and its potential intergenerational impact. A comprehensive, multi-compartment and multi-omics approach is therefore needed to better characterize how maternal BMI is associated with microbiome structure, metabolite composition, and cross-compartment relationships at the maternal–neonatal interface.

In this study, we investigated associations between maternal BMI and microbial as well as metabolite profiles across maternal stool, colostrum, and neonatal stool samples in a Mexican mother–infant cohort. Using 16 S rRNA gene sequencing (V3 polymorphic region), untargeted metabolomics, SourceTracker analysis, and multi-omics integration, we aimed to characterize cross-compartment interactions and identify patterns consistent with early microbial transfer and metabolic programming. Through this integrative approach, we sought to provide a more comprehensive understanding of how maternal metabolic and dietary factors are associated with early-life microbial and metabolic environments.

Materials and methods

Study framework and participant selection criteria

This study utilized a retrospective, cross-sectional, and case-control design (Fig. 1), focusing on Mexican pregnant women and their infants receiving care at the “Instituto Nacional de Perinatología-Isidro Espinosa de los Reyes,” a tertiary care governmental facility in Mexico City (19°25′22″ N 99°12′21″ W). Data were collected between June 2022 and August 2023, with ethical approval granted by the Investigation Committee and Bioethics Committee of the National Institute of Perinatology (INPER-2018-1-172, June 22, 2019) and the Bioethics Committee for Research on Human Beings of Cinvestav (COBISH, 2021-06-03-Dictamen COBISH-078_2021). Participants were recruited based on confirmed Mexican ancestry for at least two generations. Eligible mothers were between 18 and 45 years of age, including women with obesity, provided that their infants were born at term and free of congenital malformations. Deliveries included either spontaneous vaginal birth or non-elective cesarean section. Strict exclusion criteria were applied. Women were not enrolled if they followed a vegan diet, consumed probiotics, alcohol, or tobacco, or had a history of pre-existing or gestational diabetes. Additional exclusions included recent or regular use of anti-inflammatory drugs, laxatives, or antacids, as well as hormone replacement therapy. Mothers with gastrointestinal diseases or other medical conditions known to alter the microbiota were also excluded. Despite defining gestational diabetes mellitus (GDM) as an exclusion criterion, one participant with a prior diagnosis of GDM during pregnancy was retained in the dataset. This participant exhibited normal glycemic parameters during the postpartum sampling period, which represents the time point analyzed in this study. All participants gave written informed consent in accordance with the principles of the Declaration of Helsinki.

Fig. 1.

Overview of study design and analytical workflow. Schematic representation of the study design. Mothers were stratified according to pre-pregnancy BMI into normal weight (NW, n = 10), overweight (OW, n = 18), and obese (OB, n = 14) groups. Biological samples, including Mother stool (MS), colostrum (CO), and neonatal stool (NS), were collected within 0–48 h postpartum. Microbiota profiling was performed using 16 S rRNA gene sequencing (V3 region), while metabolite profiling was conducted using untargeted metabolomics. Downstream bioinformatic analyses included diversity assessment, differential abundance testing, source tracking, and multi-omics integration.

Specimen acquisition and data collection

The study cohort consisted of healthy lactating women and women with obesity, alongside their exclusively breastfed, full-term infants, totaling 42 mother-infant dyads. The Colostrum (CO) samples analyzed in this study were collected from participants within 1–2 days following delivery. Concurrently, fecal samples were systematically obtained from both mothers and infants to facilitate comprehensive analysis. All participants in the study were delivered via non-elective cesarean section, resulting in a uniform mode of delivery across the cohort. In total, 34 Mother stool (MS), 55 Neonatal stool (NS), and 35 CO samples were obtained. These samples were categorized based on maternal Body Mass Index (BMI): Normal Weight (NW), Overweight (OW), and Obese (OB). CO was manually collected into sterile glass tubes without prior breast sanitization to accurately represent the microbial exposure encountered by nursing infants. All samples were preserved at − 70 °C to maintain microbial integrity. Demographic and clinical data were obtained through a structured questionnaire, accompanied by anthropometric and biochemical measurements. Each participant also completed a 1-week food frequency questionnaire to elucidate dietary patterns and eating behaviors.

DNA extraction

DNA was extracted from 200 mg of MS and NS samples using the FavorPrep™ Stool DNA Isolation Mini Kit (Cat. FASTI 001–1, FAVORGEN© Biotech Corporation, Zhunan, Taiwan) as per the manufacturer’s instructions. For CO samples, 1 mL was processed with the FavorPrep™ Milk Bacterial DNA Extraction Kit (Cat. FAMBD 001, FAVORGEN© Biotech Corporation, Zhunan, Taiwan). DNA integrity was confirmed via 0.5% agarose gel electrophoresis, while DNA concentration was determined using 260/280 and 260/230 absorbance ratios measured with a NanoDrop Lite Spectrophotometer (Thermo Scientific, Waltham, MA, USA). A negative control consisting of an empty extraction column was processed alongside samples through DNA extraction, PCR amplification, and sequencing to monitor potential contamination.

Amplification of the V3 region of the 16 S rRNA gene in bacterial species

Microbiota composition was determined by sequencing the V3 polymorphic region of the bacterial 16S rRNA gene in each sample. The ~ 281 bp PCR product was amplified using the V3–341F forward primer (set of barcodes) complementary to positions 340–356 (5’-CCT ACG GGA GGC AGC AG-3’), and the V3–518R reverse primer complementary to positions 517–533 (5’-ATT ACC GCG GCT GCT GG-3’) of the Escherichia coli 16 S rDNA molecule rrnB (GenBank J01859.1)²⁹. The thermocycling program was set as follows: 3 min at 98 °C, followed by 25 cycles (12 s at 98 °C, 15 s at 62 °C, and 10 s at 72 °C), and a final extension of 5 min at 72 °C. The PCR product was visualized on a 2% agarose gel. Amplicon quantities were estimated by densitometry using the Image Lab v.4.1 program, and a final library was constructed by pooling equal concentrations of amplicons^30,31.

Next-generation DNA sequencing (NGS)

The final library was purified using a 2% agarose gel stained with SYBR GOLD DNA (E-Gel™ EX, 2%, Invitrogen™, Cat. G401012, Waltham, MA, USA). The DNA library concentration and final fragment size were measured with a 2100 Bioanalyzer Instrument (Agilent Technologies, Santa Clara, CA, USA), resulting in an average library size of 256 bp. Emulsion PCR was conducted using the Ion PGM HI Q view OT2 Kit (Cat. A29900, Life Technologies, Carlsbad, CA, USA) as per the manufacturer’s instructions. Sequencing was performed using the Ion PGM HQ view SEQ Kit (Cat. A30044, Life Technologies, Carlsbad, CA, USA) and the Ion 318 Kit V1 Chip (Cat. 4488146, Life Technologies, Carlsbad, CA, USA) on the Ion Torrent PGM system v4.0.2. After sequencing, the reads were filtered by the PGM software to remove polyclonal sequences (homopolymers > 6) and low-quality sequences (quality score ≤ 20).

Metabolite profiling utilizing electrospray ionization fourier transform ion cyclotron resonance mass spectrometry (ESI FT-ICR MS)

For fecal samples, approximately 10 mg of dehydrated fecal samples were suspended in 100 µL of 0.01 N HCl and vortexed for 1 min. Subsequently, 800 µL of HPLC-grade methanol was added, and the mixture was vortexed for an additional minute. The extracts were then incubated at − 20 °C for 30 min prior to centrifugation at 10,000 rpm for 10 min. The resulting supernatant was passed through a 0.45 μm PTFE membrane syringe filters (Navigator Lab Instruments Co). For Colostrum, 50 µL of the sample was first diluted with an equal volume of Milli-Q water and vortexed for 1 min. To this, 800 µL of HPLC-grade methanol was added, followed by another 1 min vortex. The mixture was incubated at − 20 °C for 30 min and centrifuged at 10,000 rpm for 10 min. This incubation–centrifugation cycle was repeated until no protein pellet was visible. The clarified supernatant was then filtered through a 0.45 μm PES membrane syringe filters (Navigator Lab Instruments Co). Filtered extracts were analyzed on a SolariX XR Fourier Transform Ion Cyclotron Resonance Mass Spectrometer (FT-ICR MS) equipped with a 7 Tesla superconducting magnet and an Apollo II electrospray ionization (ESI) source (Bruker, Bremen, Germany). External calibration in both positive and negative ionization modes was performed using a sodium trifluoroacetate solution. Mass spectra were processed in DataAnalysis v5.2 (Bruker) with a signal-to-noise threshold of 6 to generate peak lists. Metabolite annotation was carried out in Compass MetaboScape 2022b v9.0.1 (Bruker)^30,31.

Bioinformatic analysis

Amplicon Sequence Variants (ASVs) were identified from quality-filtered reads using the QIIME2-2023.5 pipeline³², with the “–p-trunc-len” parameter set to 173 nucleotides. Representative sequences were taxonomically annotated using the Greengenes2 database³³. Subsequent data processing and statistical analyses were conducted in R version 4.4.2 (2024-10-31 ucrt)³⁴ within RStudio 2025.05.1 + 513 IDE³⁵. Data were imported into R using the qiime2R 0.99.6 package, and microbial community analysis was facilitated using the phyloseq 1.46.0 package^36,37. Intra-sample diversity was assessed through the calculation of Observed number of species, and Shannon, Simpson, and Fisher diversity indices, while inter-sample diversity was evaluated via UniFrac distance and Non-Metric Multidimensional Scaling (NMDS) ordination using the vegan 2.6-4 package. Core microbiota were visualized through a heatmap (30% prevalence, 5% detection), and Spearman’s rank correlation of microbiota composition with anthropometric, biochemical, and dietary variables was performed using the microbiome 1.22.0³⁸ and ComplexHeatmap 2.16.0 packages³⁹. Ven Diagram analysis was performed via VennDiagram package⁴⁰. Differential abundance was analyzed with ALDEx2 3.22⁴¹, and sample-type similarities were assessed using the Sourcetracker2 Gibbs sampler⁴².

For metabolite peak selection, Bruker MetaboScape 2022b software v.9.0.1 was employed with the T-Rex 2D algorithm, applying a 60% minimum feature presence threshold across sample types and a signal noise intensity threshold of 500,000. Metabolite annotation was conducted using the Human Metabolome Database (HMDB). Further metabolomic analyses were carried out with the RforMassSpectrometry 0.1.4 package⁴³. Data were normalized using Variance Stabilizing Normalization (vsn 3.68.0)⁴⁴ with zero values imputed to the minimum value. Metabolites with ≥ 55% prevalence across samples were selected for hierarchical clustering heatmap generation using the QFeatures 1.11.2 package. Principal Component Analysis (PCA) was performed on the normalized data⁴⁵. Optimal clustering and partitioning around medoids were determined using the fpc 2.2–11 package⁴⁶. Data management was supported by the tydiverse 2.0.0 package⁴⁷, and visualizations were generated using ggplot2 3.4.40, ggpubr 0.6.0⁴⁸, and ggdendro 0.1.23. Statistical significance was defined as p- and q-values < 0.05. Metabolome and microbiota data were normalized using the Centered Log Ratio (CLR) transformation, which was performed using the miXomics 6.24.0 package⁴⁹. Model tuning and feature selection were performed with the Mfold validation parameter, and repeated cross-validation was applied to estimate the classification error rate based on centroid distance.

Results

Maternal characteristics

The study evaluated 42 mothers stratified into three BMI categories: normal weight (NW), overweight (OW), and obese (OB), according to pre-pregnancy BMI. No statistically significant differences were noted in maternal age (p > 0.05), gestational age at delivery, or parity across the BMI classifications. All deliveries were non-elective cesarean sections. Anthropometric analysis revealed significant differences in BMI between NW and OW groups. Additionally, OB mothers exhibited a significantly higher previous weight when compared to NW mothers. While final weight trended upwards across BMI categories, the differences did not reach statistical significance. Height appeared marginally lower in OB mothers, though this, too, lacked statistical significance. Socioeconomic status and educational attainment showed no significant discrepancies among the groups. However, the proportion of mothers in free union relationships increased with the ascending BMI category. Notably, antibiotic usage was highest in the OB group, although this data was not subjected to statistical analysis (Table 1).

Table 1.

Data of 42 mothers grouped according to BMI level determined according to the pre-pregnancy weight reported by the mother.

Variable		Normal weight	Overweight	Obesity	p-Value	p-Value	p-Value
		(18.5–24.9 Kg/m2)	(25.0-29.9 Kg/m2)	(> 30.0 Kg/m2)	(NW vs. OW)	(NW vs. OB)	(OW vs. OB)
Number of participating mothers (n = 42)		10	18	14	ND	ND	ND
Age (years)		32.9 (± 5.7)	33.9 (± 4.8)	32.4 (± 7.0)	0.5	0.3	0.1
Age range (years)		21 to 40	22 to 43	21 to 42	ND	ND	ND
Current parity	Non-elective cesarean	10 (100%)	18 (100%)	14 (100%)	ND	ND	ND
	Gestational age (weeks)	38.1 (± 1.1)	38.3 (± 1.0)	38.2 (± 1.2)	0.4	0.7	0.8
	Weeks range	36.3–39.5	36.6–40.3	35.5–40.3	ND	ND	ND
Anthropometry	Height (m)	1.6 (± 0.1)	1.6 (± 0.1)	1.5 (± 0.1)	0.3	0.3	0.8
	Previous weight (kg)	54.9 (± 4.0)	65.6 (± 6.6)	83.6 (± 14.6)	0.2	0.04	0.05
	Average BMI	21.9 (± 2.0)	26.8 (± 1.3)	33.0 (± 3.9)	0.03	0.4	0.4
	Final weight (kg)	65.4 (± 5.3)	76.7 (± 8.4)	87.3 (± 15.4)	0.1	0.2	0.3
Socioeconomic data	1	3 (30.0%)	7 (38.9%)	7 (50.0%)	ND	ND	ND
	2	5 (50.0%)	10 (55.6%)	6 (42.9%)	ND	ND	ND
	3	2 (2.0%)	2 (11.4%)	1 (7.4%)	ND	ND	ND
Educational level	Secondary school	3 (30.0%)	5 (27.8%)	3 (21.4%)	ND	ND	ND
	High school	3 (30.0%)	6 (33.3%)	7 (50.0%)	ND	ND	ND
	Technical	0 (0.0%)	1 (5.6%)	2 (14.3%)	ND	ND	ND
	University	4 (40.0%)	5 (27.8%)	1 (7.1%)	ND	ND	ND
	Postgraduate	0 (0.0%)	1 (5.6%)	1 (7.1%)	ND	ND	ND
Marital status	Free union	3 (30.0%)	11 (61.1%)	10 (71.4%)	ND	ND	ND
	Married	6 (60.0%)	3 (16.7%)	2 (14.3%)	ND	ND	ND
	Single	1 (10.0%)	3 (16.7%)	2 (14.3%)	ND	ND	ND
	Divorcee	0 (0.0%)	1 (5.6%)	0 (0.0%)	ND	ND	ND
Antibiotics	Yes	5 (50.0%)	9 (50.0%)	11 (78.6%)	ND	ND	ND
	No	4 (40.0%)	8 (44.4%)	2 (14.3%)	ND	ND	ND
	ND	1 (10.0%)	1 (5.6%)	1 (7.1%)	ND	ND	ND
Blood test	Glucose (mg/dL)	74.5 (± 15.3)	75.9 (± 17.2)	70.6 (± 10.14)	1.0	0.06	0.1
	Cholesterol (mg/dL)	194.8 (± 15.3)	209.4 (± 45.3)	213.9 (± 51.3)	0.9	0.5	0.5
	Triglycerides (mg/dL)	205.3 (± 74.8)	218.8 (± 58.3)	219.0 (± 69.4)	0.215	0.511	0.640
	HbA1c (%)	5.2 (± 0.4)n8	5.1 (± 0.6)*	5.1 (± 0.3)	0.2	0.4	0.01
	HDL (mg/dL)	55.2 (± 14.3)	58.4 (± 16.2)	58.2 (± 14.2)	0.9	0.8	1.0
	LDL (mg/dL)	98.5 (± 42.3)	107.2 (± 38.8)	112.0 (± 45.7)	0.9	0.3	0.3
	Albumin (g/dL)	3.0 (± 0.2)n9	3.0 (± 0.3)	3.0 (± 0.3)	0.2	0.3	0.8
	Phosphatase (U/L)	111.4 (± 29.2) n9	115.5 (± 43.3)	106.5 (± 22.1)	0.3	0.2	0.03
	Fe (µg/dL)	61.7 (± 17.4) n9	54.4 (± 24.7)	78.6 (± 44.9)	0.4	0.08	0.09
	Phosphorous (mg/dL)	4.1 (± 0.7) n9	4.0 (± 0.9)	4.4 (± 0.9)	0.6	0.4	0.8
	BUN (mg/dL)	12.4 (± 2.1)	10.5 (± 2.5)	11.1 (± 2.9)	0.3	0.5	0.9
	UC (mg/dL)	26.5 (± 4.4) n9	22.5 (± 5.3)	23.8 (± 6.2)	0.3	0.5	0.9
Electrolytes	Na (mmol/L)	134.9 (± 2.0)	139.3 (± 10.9)n17	136.6 (± 5.1)	0.04	0.1	0.1
	K (mmol/L)	4.0 (± 0.4)	4.5 (± 0.5)*	4.2 (± 0.3)	0.2	1.0	0.09
	Cl (mmol/L)	110.4 (± 1.9)	113.3 (± 7.0) n17	111.0 (± 3.1)	0.02	0.189	0.05
	iCa (mmol/L)	1.2 (± 0.1)	1.2 (± 0.1) n17	1.2 (± 0.1)	0.4	0.9	0.3
	iMg (mmol/L)	0.5 (± 0.1)	0.5 (± 0.1) n17	0.5 (± 0.08)	0.1	0.2	1.0
Hormone	Insulin (µU/mL)	8.2 (± 6.1)*	10.9 (± 7.6) n16	14.8 (± 10.2) n13	0.3	0.2	0.4
	Osteocalcin (ng/L)	368.5(± 163.8)*	285.3(± 178)n16	304.2 (± 181) n13	0.8	0.9	0.7
	CRP	34.2 (± 24.2)*	25.8 (± 22.0) n16	32.2 (± 19.4) n13	0.7	0.4	0.6

n, sample number; NM, Normal Weight; OW, Over Weight; O; Obesity; Kg, kilograms; mg, milligrams; ng, nanograms; L, liters; dL, deciliters; min, minutes; mM, millimole; µU, microunits; BMI, Body Mass Index; Socio economic data classifies mother’s economic, social, and family situation through a study; HbA1c, glycosylated hemoglobin, HDL, Hight-Density Lipoprotein; LDL, Light-Density Lipoprotein; Fe, Iron; Na, Sodium; K, Potassium; Cl, Chlorine; iCa, ionized Calcium; iMg, ionized Magnesium; CRP, C-reactive protein; BUN, Blood Urea Nitrogen; UC, Calcium Urea; ND, No Determined; Standard deviation is show as ± values; Percentage is shown as %; ^n#, represents the number of participants with whom the statistical analysis was performed; Final weight; post-partum weight; *Anthropometric, biochemical, electrolyte, and hormonal measurements at 0 months were collected during the immediate postpartum period. Parity includes the current pregnancy.

Maternal clinical and biochemical findings

Maternal anthropometric and biochemical characteristics are presented in Table 1. Fasting glucose and lipid profiles did not differ significantly among the groups. In contrast, HbA1c levels showed a statistically significant difference between OW and OB participants, while no significant differences were observed between NW and the other groups (Table 1). Notably, HbA1c variability appeared reduced in the OB group. Electrolyte analysis revealed significantly elevated sodium and chloride levels in the OW group compared to NW. Hormonal markers, including insulin, trended higher in OB mothers but did not reach statistical significance. CRP levels and osteocalcin did not differ significantly across groups. Additional data are presented in the Supplementary Tables (Table S1 and Table S2).

Neonatal outcomes

Neonatal anthropometric and clinical outcomes are summarized in Table 2. Gestational age and sex distribution were comparable across all maternal BMI categories. All infants were delivered via cesarean section. A significant difference was observed in the duration of skin-to-skin contact between OW and OB groups. On regard to lactation initiation, neonates were breast fed immediately after delivery, however, some of them needed to be formula fed since the mother was not capable to express colostrum immediately. In these cases, human lactation was implemented within 1 h. Lactation initiation percentage was lower among the neonates of OB mothers compared to OW and NW. Neonatal anthropometry revealed significant differences in height (NW vs. OW; OW vs. OB), with infants born to OW mothers exhibiting greater height. Neonates of OB mothers showed a significantly lower fat percentage (NW vs. OB) and a higher fat-free mass percentage (NW vs. OB; OW vs. OB). Triceps and subscapular skinfold thickness were significantly greater in OB neonates; alongside increased abdominal circumference compared to OW. Although neonatal glucose levels were higher in the OB group, these differences did not reach statistical significance. In contrast, insulin concentrations were significantly elevated in OB neonates compared to OW. CRP levels were lower in OB neonates relative to OW, and ionized calcium levels were significantly reduced in OB neonates (OW vs. OB). Additional data is provided in Supplementary Table S3.

Table 2.

Data of 42 neonates grouped according to the BMI level of mothers in this study.

Variable		Mothers (NW)	Mothers (OW)	Mothers (OB)	p-value	p-value	p-value
					NW/OW	NW/OB	OW/OB
Number of participating infants (n = 42)		10	18	14	ND	ND	ND
Gestational age (weeks)		38.1 (± 1.1)	38.3 (± 1.0)	38.2 (± 1.2)	0.1	0.7	0.4
Skin-to-skin contact	Yes	6 (60%)	10 (55.6%)	9 (64.3%)	ND	ND	ND
	Average time (min)	25.0 (± 12.3)n6	19.4 (± 21.9)n10	17.6 (± 12.1)n9	0.3	0.3	0.1
	No	3 (30%)	8 (44.4%)	3 (21.4%)	ND	ND	ND
Lactation initiation	Yes	8 (80.0%)n8	17 (94%)n17	11 (78.6%)n11	ND	ND	ND
Gender	Male	5 (50%)	9 (50%)	6 (42.8%)	0.7	0.7	0.9
	Female	5 (50%)	9 (50%)	8 (57.1%)	ND	ND	ND
Anthropometry	Weight (kg)	2.7(± 0.6)	2.9(± 0.6)	2.9(± 0.6)	0.6	0.8	0.4
	Height (cm)	50.0(± 1.9)	50.3(± 2.9)	49.9(± 2.6)	0.01	0.3	0.002
	Fat (%)	7.1(± 4.6)	6.7(± 3.8)	6.2(± 2.4)	0.1	< 0.001	0.04
	Fat-free mass (%)	78.5(± 34.8)	93.4(± 3.8)	93.8(± 2.3)	0.9	0.01	0.01
Circumferences	Head circumference	34.0(± 1.7)	34.2(± 1.5)n15	34.4(± 1.2)n13	0.05	0.2	0.01
	Abdominal circumference	26.8(± 7.2)	28.6(± 5.5)n15	29.2(± 2.8)n13	0.3	0.2	0.01
	Upperarm circumference	9.6(± 0.7)	10.3(± 1.0)n15	9.7(± 1.9)n13	0.1	0.2	0.2
	Femur circumference	13.6(± 3.94)	13.9(± 7.25)n15	10.9(± 1.01)n9	0.7	0.03	0.2
Folds	Triceps skinfold	0.4(± 0.1)	0.4(± 0.1)	0.9(± 1.5)n10	0.3	0.1	0.02
	Subscapular skinfold	0.4(± 0.1)	0.4(± 0.1)n15	0.9(± 1.4)n10	0.4	0.06	0.02
Blood test	Glucose (mg/dL)	56.4(± 13.5)	57.0(± 19.0)	69.7(± 36.2)	0.8	0.2	0.2
	Cholesterol (mg/dL)	49.1(± 8.4)	61.5(± 19.8)	66.9(± 23.6)	0.04	0.06	0.8
	Triglycerides (mg/dL)	31.9(± 14.9)	34.2(± 31.9)	26.6(± 13.0)	0.7	0.3	0.4
	HDL (mg/dL)	25.7(± 7.8)	27.6(± 8.8)	30.5(± 10.2)	0.8	0.4	0.5
	LDL (mg/dL)	17.0(± 7.2)	27.0(± 12.1)	31.0(± 14.0)	0.4	0.4	0.7
	Albumin (g/dL)	3.5(± 0.3)	3.5(± 0.3)	3.7(± 0.3)n12	0.7	0.9	0.5
	Phosphatase (U/L)	151.7(± 48.9)	129.6(± 55.6)	170.0(± 95.1)n12	1.0	0.3	0.2
	Fe (µg/dL)	147.8(± 52.3)	167.1(± 36.7)	163.3(± 29.8)n12	0.2	0.2	0.8
	CRP (mg/L)	0.3(± 0.2)	0.6(± 1.1)n15	0.2(± 0.3)	0.05	0.3	0.04
	Po3 (mg/dL)	5.6(± 0.6)	6.3(± 1.2)	6.1(± 0.9)n12	0.4	0.3	0.8
	BUN (mg/dL)	12.1(± 2.7)	11.6(± 3.2)	10.2(± 2.7)n12	0.8	0.9	0.9
	Urea (mg/dL)	22.0(± 5.4)	24.9(± 6.8)	25.9(± 5.8)	0.9	1.0	0.9
Electrolytes	Na (mmol/L)	129.3(± 2.5)	140.0(± 6.3)	133.4(± 6.00)	0.07	0.2	0.7
	K (mmol/L)	9.0(± 1.8)	9.8(± 2.9)	9.0(± 3.3)	0.2	0.08	0.5
	Cl (mmol/L)	109.1(± 2.2)	111.7(± 4.8)	111.7(± 4.7)	0.03	0.2	0.7
	iCa (mmol/L)	1.3(± 0.1)	1.2(± 0.2)	1.2(± 0.4)	0.2	0.04	0.03
	iMg (mmol/L)	0.5(± 0.2)	0.5(± 0.09)	0.6(± 0.1)	0.3	0.6	0.3
Hormone	Insulin (µU/mL)	5.7(± 4.6)n8	3.6(± 1.2)n16	8.2(± 9.4)n11	< 0.001	0.5	0.01
	Osteocalcin (ng/L)	461.6(± 236.4)	378.2(± 209.7)n15	381.4(± 221.7)n13	0.5	0.8	0.7

n, sample number; Kg, kilograms; mg, milligrams; ng, nanograms; L, liters; dL, deciliters; min, minutes; mM, millimole; µU, microunits; BMI, Body Mass Index; Lactation initiation, neonates were breast fed immediately after delivery, however, some of them received formula since the mother was not capable to express colostrum immediately, in these cases, human lactation was implemented within 1 h. HDL, Hight-Density Lipoprotein; LDL, Light-Density Lipoprotein; Fe, Iron; Na, Sodium; K, Potassium; Cl, Chlorine; iCa, ionized Calcium; iMg, ionized Magnesium; CRP, C-reactive protein; Po3, phosphorus trioxide; BUN, Blood Urea Nitrogen; UC, Calcium Urea; ND, No Determined; Standard deviation is show as ± values; Percentage is shown as %; ^n#, represents the number of participants with whom the statistical analysis was performed.

Microbial composition, diversity, and vertical transmission across maternal and neonatal compartments

High-throughput sequencing of the 16 S rRNA gene (V3 region) generated approximately seven million high-quality reads (mean ~ 40,000 reads per sample; Table S4), Sequence quality filtering was performed based on Phred quality scores. Additionally, low-abundance ASVs (< 1% relative abundance) were excluded to reduce noise, enabling robust taxonomic profiling across mother stool (MS), colostrum (CO), and neonatal stool (NS) samples. Alpha diversity analysis demonstrated clear differences across biological compartments. MS samples consistently exhibited significantly greater richness and evenness compared to CO and NS across all maternal BMI categories. Stratification by maternal BMI revealed modest but consistent reductions in diversity in the overweight (OW) and obese (OB) groups, particularly in CO and NS (Fig. 2A). Beta diversity analysis using Aitchison distance highlighted strong compositional differences between sample types. Principal component analysis showed that the first principal component (PC1) explained 70% of the total variance, with distinct clustering primarily driven by sample type, separating CO and NS communities (Fig. 2B). MS samples from OB mothers exhibited tighter clustering, suggesting reduced inter-individual variability within this group. Comparison of weighted (Fig. S2A) and unweighted (Fig. S2B) analyses indicated consistent clustering patterns, supporting the contribution of both abundant and low-abundance taxa to the observed community structure. Hierarchical clustering further confirmed clear segregation by sample type (Fig. 2C). MS samples displayed greater heterogeneity, whereas CO and NS showed more constrained microbial profiles. Overlay of metadata (BMI, age, sample type) indicated that clustering was primarily driven by biological compartment rather than maternal age. CO samples were enriched in genera commonly associated with milk microbiota, including Streptococcus, Lactobacillus, and Staphylococcus. Variation in the genus Collinsella was observed across BMI categories (Fig. 2C). Analysis of shared and unique taxa using Venn diagrams revealed both compartment-specific and overlapping microbial signatures among MS, CO, and NS across BMI groups. It revealed that the OB group exhibited the highest overlap of bacterial genera across MS, CO, and NS, whereas the NW group showed greater compartment-specific taxa. The OW group demonstrated an intermediate distribution, suggesting a gradual shift in microbial overlap associated with increasing maternal BMI. (Fig. 2D). To investigate potential maternal contributions to neonatal microbial colonization, SourceTracker analysis was performed (Fig. S1). CO emerged as the primary identifiable source of NS microbiota, whereas MS contributed minimally. A substantial proportion of the neonatal microbiome was assigned to “unknown” sources, indicating potential contributions from environmental exposures or uncharacterized maternal niches.

Fig. 2.

Microbial Diversity Metrics and Community Structure. (A) Box plots of alpha diversity indices (Observed species, Shannon, Simpson, Fisher’s) across host classifications normal weight (NW), overweight (OW), obese (OB) and sample types (MS, CO, NS). Sample types are color-coded; group differences are visualized to assess intergroup microbial diversity. (B) Principal Component Analysis (PCA) based on Aitchison distance depicts beta diversity. Shapes represent classification (NW: circles, OW: squares, OB: triangles), and colors reflect sample types (MS: red, CO: green, NS: blue). Contour lines show density of sample clusters, offering insight into community-level microbial variation. (C) Heatmap of bacterial taxa across sample types and BMI classifications. The heatmap displays the relative abundance of selected bacterial genera across maternal stool (MS), Colostrum (CO), and neonatal stool (NS) samples. Color intensity represents abundance levels, ranging from blue (low) to red (high). Hierarchical clustering highlights patterns of similarity across samples and taxa. Top annotation bars indicate sample metadata, including BMI category (NW, OW, OB), maternal age, and sample type (MS, HM, NS). (D) Venn diagrams showing shared and unique bacterial genera across sample types stratified by maternal BMI category (NW, OW, OB). Each diagram represents the overlap of genera detected in Mother stool (MS), Colostrum (CO), and Neonatal stool (NS) within each BMI group. The central intersection indicates taxa shared across all three compartments, while non-overlapping regions represent sample-specific genera.

BMI-associated differences in microbial composition were further explored at the genus level. Genera shown in (Fig. 3A) were selected based on differential abundance results (q < 0.05) and biological relevance. In NS, several genera exhibited significant BMI-associated shifts. Enterococcus_H_360604 was enriched in both OW (LFC = 2.80, q = 0.001) and OB groups (LFC = 3.99, q = 0.002), while Phocaeicola_A_858004 was in OW and OB (LFC = − 1.89, q = 0.014) (Fig. 3A). In MS, Collinsella showed variation across BMI groups, with relatively lower abundance in NW compared to OW and OB individuals (Fig. 3B). Comparative analysis across sample types further demonstrated compartment-specific microbial signatures, with distinct taxa enriched in MS, CO, and NS showed that Staphylococcus and Sphingomonas_L_486704, Corynebacterium, Cutibacterium and Bradyrhizobium among other taxa are abundant in colostrum and neonatal stool regardless of the BMI category of the mother (Fig. 3C). Direct comparisons between MS and NS further supported BMI-dependent differential abundance across compartments, highlighting reductions in Collinsella in MS and shifts in NS, including enrichment of Enterococcus_H_360604, Limosilactobacillus, Klebsiella_724518, and Pectobacterium, alongside depletion of Ruminococcus_D and Phocaeicola_A_858004 (Fig. 3D).

Fig. 3.

Taxon-Specific Microbial Variation Across Groups. (A) Box plot showing the abundance of six selected bacterial genera within NS samples compared across classifications (NW, OW, OB). Clear visual distinctions allow evaluation of weight-associated shifts in taxa. (B) CLR-transformed abundance comparisons for each classification group, accounting for the compositional nature of microbiome data. (C) Bar plot of log fold changes in taxa across CO, MS, and NS groups, emphasizing shifts in microbial communities by metabolic health and symptom status. (D) Focused comparison of microbial taxa between MS and NS, stratified by classification. Bars reflect increase or decrease in specific genera, offering insights into potential biomarkers or condition-specific taxa.

Further genus-level analysis revealed BMI-associated differences in microbial abundance. Specifically, Enterococcus_H_360604 and Limosilactobacillus were enriched in the OB group compared to NW (Fig. S3A). In contrast, Collinsella showed lower abundance in NW individuals relative to OW and OB groups (Fig. S3B). Additionally, in comparisons between OW and NW, genera including Pectobacterium, Klebsiella_724518, and Enterococcus_H_360604 were enriched in the OW group, whereas Ruminococcus_D and Phocaeicola_A_858004 were reduced (Fig. S3C). Across sample types, genus-level distributions showed clear compartment-specific patterns, with taxa such as Staphylococcus and Gemella enriched in colostrum and several genera (e.g., Enterococcus_H_360604, Limosilactobacillus) more abundant in neonatal stool, particularly in OW and OB groups (Fig. S3D).

Metabolomic profiling reveals functional divergence across maternal and neonatal niches

Untargeted metabolomic profiling revealed pronounced compartment-specific organization of metabolic signatures, with additional variation associated with maternal BMI. Hierarchical clustering demonstrated robust segregation of samples primarily by biological compartment MS, CO, and NS, with secondary stratification according to BMI classification. Within this framework, MS samples from obese (OB) mothers exhibited relative enrichment of metabolites such as gambogic acid and phosphatidic acid (PA)–related compounds, whereas NS samples were characterized by higher abundance of peptide-associated metabolites, including enkephalinamide-Leu (Fig. 4A). Principal component analysis further supported this structure, with the first two principal components explaining 70% and 7% of the total variance, respectively. The dominant separation along PC1 corresponded to biological compartment, while BMI-related variation contributed to finer-scale dispersion within clusters (Fig. 4B). Differential abundance testing (ANOVA) highlighted a suite of metabolites with significant variation across MS, CO, and NS, underscoring the metabolic specialization of each compartment (Fig. 4C). In MS, lipid-associated metabolites, including phosphatidic acid (PA) derivatives and compounds such as gambogic acid, showed higher relative abundance, indicating a lipid-enriched metabolic profile. In contrast, NS was characterized by enrichment of peptide-related metabolites, including enkephalinamide-Leu, alongside additional amino acid–derived compounds, suggesting increased representation of peptide metabolism. CO displayed a distinct profile enriched in carbohydrate-associated metabolites, including compounds such as 3-β-galactopyranosyl glucose and related saccharide derivatives, which are consistent with milk-associated bioactive substrates. In addition to compartment-driven variation, BMI-associated differences were observed across multiple metabolites. Several features showed differential enrichment between OW/OB and NW groups, encompassing both lipid- and carbohydrate-related compounds, highlighting the influence of maternal metabolic status on the metabolite landscape. Volcano plot analysis confirmed metabolites significantly associated with maternal BMI, particularly distinguishing OW and OB groups from NW individuals (Fig. S4A), including both enriched and depleted features. Comparative analyses across compartments revealed metabolites uniquely enriched in MS relative to CO and NS, highlighting compartment-specific metabolic specialization. In addition, overlapping metabolite features were observed between CO and NS, indicating partial sharing of metabolite profiles across maternal and neonatal compartments. Distributional analysis further demonstrated consistent shifts in metabolite intensities across BMI categories in all sample types, indicating that maternal BMI is associated with broad alterations in metabolite abundance (Fig. S4B).

Fig. 4.

Metabolite Profiling Across Sample Groups and Classifications. (A) Heatmap illustrates metabolite expression patterns across NW, OW, and OB groups. Sample types (MS, HM, NS) are indicated in the top annotation bar with clustering to reveal sample group similarities. Rows represent individual metabolites; color intensity indicates expression levels. (B) PCA of metabolite profiles, with each point representing a sample. Points are colored by sample type and shaped by classification, with density contours identifying group-specific clustering trends. (C) Likelihood Ratio Test (LRT) bar plot identifying differentially abundant metabolites among MS, CO, and NS. Direction of the bars indicates group-wise enrichment.

Multi-omics integration unveils coordinated microbiota–metabolite–diet networks

To investigate the interplay between maternal diet, gut microbiota, and host metabolic status, we applied a multi-omics integration approach. Feature importance analysis identified key contributors distinguishing maternal BMI categories, including microbial taxa from the phyla Actinomycetota, Bacillota, Firmicutes_A, and Bacteroidetes, metabolite features such as oxidized PA1, and dietary components including whole bread and fresh cream (Fig. 5A). Integrated PCA revealed clear clustering across microbiome, metabolome, and dietary datasets, indicating consistent classification-specific segregation (Fig. 5B). This was further supported by arrow plot integration analysis, which demonstrated coordinated trajectories linking diet, microbiota, and metabolite profiles across BMI groups (Fig. 5C).

Fig. 5.

Multi-Omics Integration and Feature Contribution. (A) Bar plot ranking metabolites, microbes, and dietary features by their contribution to differentiating classifications (NW, OW, OB, “tie” between the three classifications). Features are grouped into diet, metabolome, and microbiota panels. (B) PCA plots across microbial (top row), metabolite (middle row), and dietary (bottom row) data. Panels illustrate how each data type varies by classification, with overlapping ellipses indicating clustering and separation patterns. (C) Arrow plot integrates microbiota, metabolomic, and dietary data, mapping transitions across Dimensions 1 and 2. Arrows represent relationships among blocks (micro, meta, diet); symbols differentiate data types (circles: micro, triangles: meta, minus signs: diet), and colors represent classifications (NW blue, OB orange, OW gray). (D) Network plot connecting dietary items (orange), metabolites (yellow), and microbes (green). Edge colors and thickness reflect interaction strength and direction, helping visualize the complex relationships across these domains. Positive correlation is dark blue and negative correlation is purple color (weight).

Network analysis highlighted dense cross-domain interactions, particularly pronounced in obese mothers, suggesting enhanced connectivity between dietary inputs, microbial taxa, and metabolic outputs (Fig. 5D). Notably, dietary items such as chocolate corn flakes, and mayonnaise were associated with microbial shifts (e.g., Ruminococcus, Blautia) and metabolite features (e.g., glycosides, oligopeptides), indicating coordinated diet–microbiota–metabolite interactions (Fig. 5D). Extended DIABLO analyses (Fig. S5A–C) further supported these findings, revealing robust multi-layered associations and group-specific variation across multiple component spaces. Collectively, these results demonstrate tight integration between maternal diet, microbial ecology, and metabolomic profiles, providing insight into mechanisms that may shape early-life developmental trajectories. In summary, these findings elucidate the tight integration of maternal diet and metabolic characteristics with microbial community dynamics and metabolomic programming at the maternal-neonatal interface, underpinned by statistically robust associations.

Discussion

In this work, we studied the vertical transmission of gut microbiota and metabolites under maternal obesity conditions in mother-infant binomia. This study offers an integrative, multi-omic exploration of how maternal obesity during pregnancy shapes the maternal-neonatal microbiome and metabolome, with implications for intergenerational health trajectories. By analyzing matched mother stool (MS), Colostrum (CO), and neonatal stool (NS) samples, our findings reveal that maternal adiposity profoundly alters microbial composition and function, metabolite distribution, and host-microbe-diet interactions across the perinatal axis. These alterations manifest in both mothers and neonates, suggesting that maternal metabolic status during gestation is a critical determinant of early-life microbial and metabolic programming.

Maternal obesity is associated with altered anthropometric and metabolic profiles in mothers and neonates

Our data reaffirms the developmental origins of health and disease (DOHaD) framework by demonstrating that maternal obesity is associated with adverse metabolic parameters in mothers and altered growth and body composition in their offspring. Obese mothers in our cohort exhibited heightened pre-delivery adiposity, insulin resistance (elevated Homeostatic Model Assessment for Insulin Resistance), hyperinsulinemia, and elevated triglycerides hallmarks of gestational metabolic dysfunction. These patterns are consistent with reports suggesting that maternal obesity exacerbates insulin resistance during pregnancy and fosters a pro-inflammatory, lipotoxic environment that can disrupt placental nutrient signaling and fetal development⁶. Neonates born to obese mothers mirrored this dysmetabolic profile, presenting with increased birth weight, elevated BMI z-scores, greater fat mass, and thicker subscapular and triceps skinfolds. These findings underline the transgenerational impact of maternal overnutrition, possibly mediated by increased placental transfer of glucose and lipids, fetal hyperinsulinemia, and enhanced adipogenesis in utero⁵⁰. Such early-life adiposity may predispose these infants to future metabolic diseases, reinforcing the need for mechanistic insights into the biological vectors of this inheritance.

Obesity-associated gut microbial dysbiosis in mothers

This study provides an integrated multi-omics characterization of the maternal–neonatal interface, demonstrating that maternal BMI is associated with coordinated variation in microbial composition, metabolite profiles, and diet-related features across mother stool, colostrum, and neonatal stool. By jointly analyzing these interconnected biological layers within the same mother–infant dyads, the present work advances current understanding of early-life microbial and metabolic programming. A key observation is the reduction in microbial diversity in colostrum and neonatal stool associated with being overweight and obesity. Early-life microbial diversity plays a critical role in immune system maturation, and reduced diversity has been linked to increased susceptibility to allergic and metabolic diseases^51,52. Previous studies have similarly reported that maternal obesity is associated with altered gut microbiota composition and reduced diversity during pregnancy and postpartum^9,53, supporting the notion that maternal metabolic status can influence the initial microbial environment of the neonate.

Microbial community structure in this study was strongly driven by biological compartment. Mother stool exhibited greater diversity and inter-individual variability, whereas colostrum and neonatal stool showed more constrained microbial profiles. The enrichment of genera such as Streptococcus, Staphylococcus, and Lactobacillus in colostrum is consistent with previous descriptions of early lactation microbiota^11,54. Notably, although Lactobacillus was enriched in maternal compartments (colostrum and Mother stool), its relative abundance was reduced in neonatal stool. This discrepancy suggests that exposure does not necessarily result in successful colonization. Instead, colonization efficiency may be influenced by obesity-associated factors, including altered neonatal gut conditions, microbial competition, and disrupted metabolic signaling, which may limit the establishment of beneficial taxa in early life^13,25,55. Importantly, the present findings are limited to colostrum, as samples were collected within the first 1–2 days postpartum. Given that milk microbiota evolves dynamically across lactation stages, these results may not be directly extrapolated to transitional or mature milk⁵⁶.

Maternal BMI–associated taxonomic differences were observed across maternal and neonatal compartments, particularly involving genera such as Collinsella, Enterococcus, and Klebsiella. Increased abundance of Collinsella has been associated with altered lipid metabolism and insulin resistance^28,57, while enrichment of opportunistic genera such as Enterococcus and Klebsiella has been linked to inflammatory processes and early-life microbial imbalance²⁴. The detection of overlapping microbial signatures between maternal and neonatal samples supports the possibility of maternal contributions to neonate colonization, although the cross-sectional design limits direct inference regarding transmission pathways. This supports the concept of an entero-mammary pathway through which maternal gut microbes may contribute to the milk microbiome and subsequent neonatal colonization through vertical transmission during breastfeeding. Source tracking analysis indicated that colostrum represents a major identifiable contributor to neonatal microbiota, consistent with evidence that breastfeeding facilitates microbial transfer from mother to neonate¹¹. However, a substantial proportion of neonatal taxa were attributed to unknown sources, suggesting additional contributions from environmental exposures, maternal skin microbiota, and other reservoirs⁸. These findings highlight the complexity of early microbial colonization as a multifactorial process.

Obesity rewires the maternal–neonatal metabolome with pro-inflammatory signatures

Metabolomic profiling further revealed distinct compartment-specific and BMI-associated metabolic signatures. Microbial metabolites, including short-chain fatty acids (SCFAs), amino acid derivatives, and lipid-associated compounds, play key roles in regulating host metabolism, immune responses, and gut barrier function^20,58. The enrichment of lipid-associated metabolites, including phosphatidic acid (PA) derivatives and compounds such as gambogic acid, in maternal stool from obese mothers suggests alterations in lipid metabolism associated with maternal adiposity. Previous studies have shown that obesity is linked to dysregulated lipid metabolism and increased circulating lipid intermediates, which may influence both gut microbial composition and downstream metabolite production²⁰. Such lipid-associated metabolic shifts may contribute to a pro-inflammatory metabolic milieu, although causal relationships cannot be established within the current study design. In contrast, neonatal stool was characterized by enrichment of peptide- and amino acid–derived metabolites, including enkephalinamide-Leu, indicating a metabolically distinct early-life gut environment. This pattern is consistent with prior observations that neonatal metabolism is shaped by protein-rich substrates derived from milk and early microbial colonization, with amino acid metabolism playing a central role in early gut development and immune modulation.

Colostrum displayed a metabolite profile enriched in carbohydrate-associated compounds, including 3-β-galactopyranosyl glucose and related saccharide derivatives. These findings are consistent with the known composition of early milk, which is rich in bioactive carbohydrates that function as substrates for microbial growth and modulators of host immunity. Notably, several of the detected metabolites are consistent with pathways related to human milk oligosaccharides (HMOs) and sialylated compounds. Alterations in maternal metabolic status may therefore influence not only microbial composition but also the biochemical environment encountered by the neonate, with potential implications for early-life developmental trajectories. Importantly, several metabolomic features observed in colostrum are consistent with human milk oligosaccharide (HMO)–related and Sialylated compounds, which are central to neonatal nutrition and microbial colonization. HMOs act as selective substrates for beneficial microbes and contribute to immune development and pathogen defense^27,59. Sialylated HMOs additionally play critical roles in neurodevelopment and host–microbe interactions. Emerging evidence suggests that maternal metabolic status influences HMO composition. Notably, a recent study by Corona-Cervantes¹⁰ demonstrated that maternal obesity alters HMO content and is associated with earlier acquisition of late colonizing taxa in the neonatal gut microbiome. Similarly, recent work in comparable populations has highlighted variability in milk sialylation patterns in relation to maternal diet and metabolic health⁶⁰.

Although HMOs were not directly quantified in this study, the detection of metabolites consistent with these pathways suggests that maternal BMI–associated metabolic alterations may extend to functionally important milk components. The detection of overlapping metabolite features between colostrum and neonatal stool suggests potential maternal–neonate metabolic continuity, indicating that maternal metabolic states may influence the neonatal environment not only through microbial transfer but also via shared molecular signatures.

Maternal diet shapes microbiota–metabolite architecture across the dyad

The multi-omics integration analysis revealed coordinated relationships between maternal diet, microbiota composition, and metabolite profiles, supporting the concept of a diet–microbiome–metabolome axis. Dietary factors are well-established drivers of microbial composition and function^21,58, and the observed associations between specific food items, microbial taxa, and metabolite features in this study further support this relationship. Dietary patterns derived from the food frequency questionnaire (FFQ; Table S5) were associated with variation in microbial and metabolic profiles across BMI groups. Increased network connectivity observed in obese mothers may reflect altered ecological interactions or reduced system adaptability; This supports emerging evidence that perinatal dietary quality is a powerful, modifiable determinant of infant metabolic health. Interventions promoting fiber-rich, anti-inflammatory diets during pregnancy may thus recalibrate microbial ecosystems and mitigate intergenerational transmission of obesity risk.

Several limitations should be considered when interpreting these findings. The relatively modest sample size may limit statistical power and generalizability. In addition, the cross-sectional design restricts causal inference, particularly with respect to microbial transmission and microbiota–metabolite interactions. Additionally, variation in early feeding practices during the hospital stay (lactation initiation versus formula supplementation) may have influenced neonatal microbiome composition and represent a potential confounding factor that was not fully controlled. Milk samples analyzed in this study represent colostrum collected in the immediate postpartum period; therefore, the results may not be directly extrapolated to transitional or mature milk, which undergo dynamic compositional changes across lactation. A key strength of this study is the inclusion of colostrum samples, which are rarely available and provide a unique window into early lactational microbial and metabolic composition at the maternal–infant interface. Breast skin was not sterilized prior to sample collection, and thus potential skin-associated microbiota cannot be excluded. Furthermore, taxonomic resolution was limited to the genus level, reflecting inherent constraints of 16 S rRNA gene amplicon sequencing, while platform-specific limitations associated with Ion Torrent technology—particularly homopolymer-related insertion–deletion errors—may have influenced amplicon sequence variant (ASV) accuracy despite stringent quality control measures. Dietary intake was assessed using a 7-day food frequency questionnaire (FFQ) administered in the immediate postpartum period, capturing food consumption during the final week of pregnancy, which may be subject to bias and lacks the precision of short-term dietary assessment methods such as repeated 24-hour recalls. A subset of participants (n ≈ 10) reported antibiotic use during pregnancy, primarily for vaginal infections; however, detailed information regarding dosage, duration, and route of administration was not consistently available, and exposure occurred during the second trimester rather than the peripartum period. Due to the limited and heterogeneous nature of this information, antibiotic use was not included in statistical analyses, although its potential influence on microbiome composition cannot be excluded. Although gestational diabetes mellitus was defined as an exclusion criterion, one participant with a prior history of gestational diabetes was included; however, this individual exhibited normal glycemic status postpartum and is unlikely to have significantly influenced the results. Finally, while mode of delivery was uniform across the cohort (all cesarean section), other potential confounding factors, including environmental exposures, were not fully controlled and may have contributed to the observed variability.

Conclusion

This study provides integrated multi-omics evidence that maternal pre-pregnancy BMI is associated with variations in microbial diversity and metabolite profiles across maternal stool, colostrum, and neonatal stool. By combining microbial, metabolic, and dietary data from matched mother–neonate dyads, we characterize coordinated patterns linking maternal metabolic status with early-life biological environments. Maternal obesity was associated with alterations in gut microbial composition, including reduced abundance of beneficial SCFA-producing taxa and increased representation of taxa linked to inflammatory and metabolic processes. Similar patterns were observed in colostrum, suggesting that maternal metabolic status may influence microbial and molecular exposures during early life. In parallel, metabolomic analyses revealed shifts in stool and milk metabolite profiles, including differences in lipid-, peptide-, and carbohydrate-associated features across compartments. Importantly, maternal dietary patterns were associated with variations in microbiome and metabolome profiles, highlighting diet as a potential modifiable factor in shaping these interactions. However, given the cross-sectional design, these findings should be interpreted as associations rather than causal relationships. Overall, this study contributes to understanding how maternal BMI and diet are linked to early-life microbiome and metabolome configurations, providing a foundation for future longitudinal and interventional studies aimed at improving maternal and neonate health outcomes.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Y.C.-N., M.L.P.-Z., and J.G.-M.; Methodology: Y.C.-N., M.L.P.-Z., J.G.-M., A.K., C.J.J.-C., T.B.G., S.M., and A.P.-E.; Formal analysis and investigation: Y.C.-N., M.L.P.-Z., J.G.-M., A.K., and J.M.V.-I.; Writing - original draft preparation: A.K., and J.G.-M.; Writing - review and editing: Y.C.-N., M.L.P.-Z., J.G.-M., A.K., C.J.J.-C., J.M.V.-I., S.M., A.P.-E., E.R-A., and J.J.C.-A.; Funding acquisition: Y.C.-N., M.L.P.-Z., and J.G.-M.; Resources: Y.C.-N., M.L.P.-Z., J.G.-M., E.R-A., J.J.C.-A., and S.R.-M.; Supervision: Y.C.-N., M.L.P.-Z., J.G.-M., C.J.J.-C., S.M., E.R-A., J.J.C.-A., S.R.-M.

Funding

This work was supported by Secretaría de Ciencia, Humanidades, Tecnología e Innovación (SECIHTI) CONACyT 163235 INFR-2011-01 granted to J.G.-M., CONACyT 302670 INFR-2019 granted to Y.C.-N., and CONACyT FORDECYT-PRONACES/6669/2020_Programa Presupuestario F003-Ciencia de Frontera 2019 granted to M.L.P.-Z., Y.C.-N., and J.G.-M. The sponsors had no role in design and conduct of this study; collection, management, analysis, and interpretation of the data; or in preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data availability

Sequencing data reported in this study have been submitted to the National Center for Biotechnology Information BioProject no. PRJNA1334439, and the link is [https://www.ncbi.nlm.nih.gov/search/all/?term=PRJNA1334439](https:/www.ncbi.nlm.nih.gov/search/all/?term=PRJNA1334439) accessed on October 3rd, 2025.

Declarations

Competing interests

The authors declare no competing interests.

Institutional Review Board

This study was conducted in accordance with the Declaration of Helsinki, and approved by the hospital’s Research, Biosafety, and Ethics Committee with the register number INPER-DG-000419-2019_2018-1-172 (approval date: 22 June 2019).

Informed consent

was obtained from all subjects involved in this study.