Extracellular Vesicles From Limosilactobacillus johnsonii Enhance Milk Fat Synthesis by Inducing CD36 Dynamic Palmitoylation and Activating PPARγ Signalling

ABSTRACT

Mammals support offspring survival through efficient milk production, ensuring the transfer of essential nutrients and energy. Extracellular vesicles (EVs) released by gut microorganisms function as signalling molecules that influence host physiology. In this study, we observed an association between gut microbiota and lactation performance, with Limosilactobacillus johnsonii showing potential in promoting milk fat synthesis. Using a mouse model, we demonstrated that L. johnsonii‐derived EVs enhance mammary gland function, leading to increased milk fat content and improved pup growth. Mechanistically, palmitic acid (C16:0) from L. Johnsonii EVs was found to induce the dynamic changes in CD36 palmitoylation in mammary epithelial cells, thereby facilitating fatty acid uptake as substrates for milk fat synthesis. Additionally, the increased availability of fatty acids further promotes the activation of peroxisome proliferator‐activated receptor‐γ (PPARγ), reinforcing its role in regulating milk fat synthesis. These findings provide new insights into the gut‐mammary gland axis and its role in lactation regulation.

Extracellular vesicles from Limosilactobacillus johnsonii reach the mammary gland through the gut‐mammary axis, where their palmitic acid (C16:0) induce the dynamic changes in CD36 palmitoylation in mammary epithelial cells, thereby facilitating fatty acid uptake as substrates and enhance milk fat synthesis in HC11 cells via PPARγ signalling.

1. Introduction

Breast milk is a primary source of nutrition for mammalian offspring during early life (Petersohn et al. 2023). Lactation performance is influenced by multiple factors, including maternal nutrition, hormonal regulation, genetics, environmental stressors and the presence of specific metabolites or signalling molecules such as cytokines and growth factors. Several signalling pathways have been identified as key regulators of lactation. Among them, the Janus kinase 2‐signal transducer and activator of transcription 5 (JAK2‐STAT5) pathway plays a crucial role in casein protein gene transcription (Villarino et al. 2015; Yang et al. 2015). Additionally, the phosphorylation of the mammalian target of rapamycin (mTOR) and its downstream effectors, including eukaryotic initiation factor 4E binding protein 1 (4EBP1) and ribosomal protein S6 kinase 1 (S6K1), is essential for initiating milk protein synthesis (Appuhamy et al. 2011). Furthermore, peroxisome proliferator‐activated receptor gamma (PPARγ) serves as a key regulator of lipid metabolism (Bionaz and Loor 2008). It promotes de novo fatty acid synthesis by upregulating the expression of diacylglycerol O‐acyltransferase 1 (DGAT1), acetyl‐CoA carboxylase alpha (ACACA), fatty acid synthase (FASN) and fatty acid‐binding protein 4 (FABP4) (Shi et al. 2013). Together, these pathways form a complex regulatory network that governs lactation.

Recent studies have linked the maternal gut microbiota to metabolic changes during pregnancy (Huang et al. 2024; Tian et al. 2023). The gut microbiota undergoes significant shifts between early pregnancy and late pregnancy or lactation, promoting fat deposition and improving insulin tolerance in the host (Koren et al. 2012). Furthermore, the gut microbiota has been shown to affect milk composition. In dairy cows, a greater abundance of Prevotella in the rumen correlates with increased milk fat content (Wu et al. 2021), whereas cows with higher milk protein yield (>3.20%) have a greater presence of Bacteroidetes than do those with lower protein yield (<2.90%) (Xue et al. 2019). Additionally, the rumen microbiome influences the levels of odd‐chain and polyunsaturated C18:0 fatty acids in Holstein cow milk (Buitenhuis et al. 2019). Exogenous probiotics have also been shown to improve lactation performance. Supplementing sows with a diet containing Lactobacillus and yeast fermentation broth during late pregnancy significantly increased milk protein, milk solids and milk fat content (Ma Cui et al. 2020). Additionally, probiotics such as Clostridium butyricum (Cao et al. 2019), Lactobacillus plantarum (Betancur et al. 2021), Bacillus subtilis (Zhang et al. 2020) and Enterococcus faecium (Wang et al. 2021), have been reported to improve milk quality and elevate immunoglobulin levels, a key protein component of colostrum. These findings highlight the critical role of gut microbiota in lactation. However, the precise mechanisms by which gut microbes regulate lactation remain largely unexplored.

Under normal conditions, gut microbes are confined to the intestinal tract by physical and chemical barriers, preventing direct migration to distant organs. In this context, extracellular vesicles (EVs) have emerged as signalling molecules mediating microbiota‐host interactions (Wang et al. 2022; Hezaveh et al. 2022). Microbiota‐derived EVs carry bioactive components, including peptidoglycans, lipids, proteins and nucleic acids (Toyofuku et al. 2019), and have been implicated in various physiological processes and disease regulation (González‐Lozano et al. 2022). Recent studies have highlighted diverse functions of probiotic‐derived EVs. For example, L. johnsonii EVs activate M2 macrophages by inhibiting ERK signalling, alleviating ETEC K88‐induced intestinal inflammation (Tao et al. 2025). Ackermannia derived‐EVs accumulate in bone tissue, mitigating osteoporosis by enhancing osteogenesis and inhibiting osteoclast formation (Liu et al. 2021). Additionally, Lactobacillus EVs have been shown to protect against alcohol‐induced liver injury by activating the Nrf‐2 signalling pathway (Jiao et al. 2025). Despite these findings, the role of probiotic‐derived EVs in mammalian mammary gland function remains largely unexplored.

In this study, we investigated the role of the gut microbiota in lactation and identified L. johnsonii as a potential enhancer of maternal lactation. Mechanistically, palmitic acid (C16:0) carried by L. johnsonii EVs triggered dynamic shifts in CD36 palmitoylation in mammary epithelial cells, promoting fatty acid uptake to support milk fat synthesis. The accumulated fatty acids further active the PPARγ signalling pathway, promoting milk fat synthesis. These findings suggest that L. johnsonii EVs function as molecular messengers linking the gut microbiota to mammary gland function, providing new insights into the gut‐mammary axis and its role in lactation regulation.

2. Materials and Methods

2.1. Sows and Piglets

Forty‐seven Landrace × Yorkshire sows with similar backfat thicknesses and parities were selected on gestation day 85 (G85). All sows were fed the same diet and met the NRC (2012) nutrient requirements (Table S1). The litter size was adjusted to 12 ± 1 piglets within 24 h of birth. On the basis of piglet weight gain from day 1 to 14 (L1–L14), sows were ranked in terms of lactation performance. The top six and bottom six sows were categorised as high‐lactation (HLS) and low‐lactation sows (LLS), respectively. At L21, 12 piglets with similar weights were randomly selected from both the HLS and LLS groups. Half of the piglets in each group were injected intraperitoneally with 10 mg/kg LPS (n = 6). Twelve hours later, all the piglets were euthanised via sodium pentobarbital injection (50 mg/kg body weight), and duodenum, jejunum and ileum samples were collected.

2.2. Mice

Specific‐pathogen‐free (SPF) female ICR mice were obtained from Guangdong Sijia Jingda Biotechnology Co., Ltd. (Guangdong, China). At 8 weeks of age, the mice were mated and divided into seven groups, with six mice per group. After 10 days of broad‐spectrum antibiotic treatment, the dams were gavaged with faecal microbiota from HLS or LLS sows, L. johnsonii (BNCC 135265) (1 × 10⁹ CFUs), GW4869‐treated L. johnsonii (BNCC 135265) (1 × 10⁹ CFUs), L. johnsonii (BNCC 135265)–derived EVs (2.6 × 10¹⁰ particles/kg body weight diluted in 20 µL PBS) or L. plantarum (BNCC 364479)–derived EVs (2.6 × 10¹⁰ particles/kg body weight diluted in 20 µL PBS) every other day from L0 to L14. Each female mouse nursed a litter of 13 pups. The control dams received saline. On L14, mammary gland samples were collected, and the pups (6 per group) were injected with 10 mg/kg LPS. Colon samples were collected 12 h later, and the mice were euthanised via sodium pentobarbital injection, followed by cervical dislocation. The colon injury score was based on the Chiu Intestinal Mucosal Injury Score. The mice were housed under pathogen‐free conditions with a 12‐h light/dark cycle at 24°C and 55%–60% humidity, with food and water provided ad libitum.

L. johnsonii transplantation were performed as previously described. Briefly, pregnant mice received a broad‐spectrum antibiotic cocktail in their drinking water, comprising ampicillin (1 g/L), gentamicin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L) and metronidazole (1 g/L), from gestational day 7 (G7) to 17 (G17) to deplete the intestinal microbiota. The antibiotic solution was replaced every 3 days to maintain drug stability. After the 10‐day treatment, mice were provided with autoclaved purified water for 72 h to eliminate residual antibiotics. Beginning on lactation day 0 (L0), mice received daily oral gavage of L. johnsonii at a dose of 1 × 10⁹ CFU/day for 15 consecutive days, ending on lactation day 14 (L14).

2.3. Cell Culture

Mouse mammary epithelial cells (HC11) were used to investigate the effects of L. johnsonii and its EVs on milk fat synthesis. HC11 cells were cultured in DME/F12 medium supplemented with 10% fetal bovine serum (FBS), 5 µg/mL insulin‐like growth factor‐1 (IGF‐1), 10 ng/mL epidermal growth factor (EGF), 5 µg/mL insulin‐transferrin‐selenium (ITS) and 1% penicillin‒streptomycin. For cell differentiation, 2 µg/mL prolactin was used to treat the cells for 2 days. The cell culture incubator was set at 37°C with 5% CO₂. After overnight incubation to allow for cell attachment and reaching approximately 80% confluence, the cells were treated with L. johnsonii (BNCC 135265)–derived cell‐free supernatant (3%) (The supernatant was adjusted to pH 7.2 ± 0.2 with 1 M NaOH and incubated at 4°C for 10 min to ensure pH stability), L. johnsonii (BNCC 135265)–derived EVs (1.3 × 10¹⁰ particles/mL), EVs components (the remaining components after treatment with DNase I, RNase A or proteinase K, respectively), or C16:0 (10 µM), C18:0 (10 µM) for 24 h (0.1% DMSO was used as the vehicle for C18:0 and C16:0). Lipid droplet formation and the expression of lipid synthesis‐related genes and proteins were assessed.

To deplete specific contents from L. johnsonii (BNCC 135265)–derived EVs (DNA, RNA, protein), EVs (1.3 × 10¹⁰ particles/mL) were treated with DNase I, RNase A or proteinase K, following previously described methods (Hong et al. 2023). Briefly, EVs solutions (1.3 × 10¹⁰ particles/mL) were frozen and thawed six times, followed by sonication for 30 s. RNase A (1 U/µL), DNase I (2 U/µL) or proteinase K (40 µg/mL) was added, and the mixture was incubated at 37°C for 30 min. For HC11 cell treatments, 20 µL of EVs lysis buffer was added per millilitre of culture medium.

2.4. Lipid Extraction From EVs

Lipids were extracted from L. johnsonii (BNCC 135265)‐EVs via methyl‐tert‐butyl ether (MTBE). Briefly, 375 µL of methanol and 1125 µL of MTBE were added to 500 µL of EVs solution (1.3 × 10¹⁰ particles/mL) and agitated for 1.5 h. After 650 µL of deionised water was added and incubated for 10 min at room temperature, the mixture was centrifuged at 10,000 × g for 10 min to separate the organic and inorganic phases. The organic phase was concentrated under liquid nitrogen for further use. For HC11 cell treatments, 20 µL of the organic phase or inorganic phase was added per millilitre of culture medium. HC11 cells were also subjected to combinations of organic phase and GW9662 (10 µM) (Selleck Chemicals LLC) added simultaneously.

2.5. Analysis of Sow Milk

Colostrum (20 mL) was collected from each sow within 12 h postpartum, and milk (20 mL) was collected on L14 after a 20 IU intramuscular oxytocin injection. Milk components, including protein, fat, lactose and non‐fat solids, were analysed via a milk analyser (Ekomilk Bond). Immunoglobulin (IgG, IgA and IgM) concentrations in colostrum were measured via ELISA kits (Shanghai Enzyme‐linked Biotechnology, China). The milk yield of sows was calculated according to the following formula: total milk yield (kg) = 4 × average weight gain of piglets × litter size × lactation days (Miao et al. 2019).

2.6. Isolation of Milk Cells

Sow colostrum was diluted 1:1 with phosphate‐buffered saline (PBS) and centrifuged at 870 × g for 20 min at 4°C. The supernatant was discarded, and the pellet was washed twice. The pellet was resuspended in 5 mL of PBS and centrifuged at 490 × g for 5 min to isolate milk‐separated cells.

2.7. 16S rRNA Sequencing and Data Analysis

To assess the composition and diversity of the intestinal microbiota, we performed 16S rRNA gene sequencing (Biomarker Biotechnology, Beijing, China). Genomic DNA was extracted from faecal samples, and the V3‐V4 regions of the 16S rRNA gene were amplified using specific barcoded primers: F: ACTCCTACGGGAGGCAGCA; R: GGACTACHVGGGTWTCTAAT. The amplified products were purified, prepared for library construction, and sequenced using the Illumina NovaSeq 6000 platform. Data processing was conducted using QIIME2 2020.6, which included sequence denoising, chimera removal and operational taxonomic unit (OTU) clustering. Taxonomic classification at the genus and species levels was performed using the Silva 138 database, with sequence alignment conducted via the BLAST algorithm. Microbial diversity was assessed using α‐diversity indices, including the Chao1 and Simpson indices. Differences in bacterial composition among groups were evaluated through Principal Coordinates Analysis (PCoA) based on distance matrices. Linear Discriminant Analysis (LDA) was applied to identify enriched bacterial taxa in each group, with a significance threshold of LDA > 2.7. Data visualisation and statistical analyses were performed using the Biomarker Cloud Platform. The 16S sequencing data from this study have been deposited in the NCBI Sequence Read Archive: # PRJNA1176286.

2.8. Microbial Preservation of Sow Faeces

Fresh sow faecal samples (100 g) were homogenised in 300 mL of sterile saline and filtered through double‐layer sterile gauze to remove large particles. The filtered suspension was centrifuged at 600 × g for 15 min, and the pellet was subsequently resuspended in saline containing 10% sterile glycerol for storage at −80°C.

2.9. RNA Extraction and Quantitative RT‐PCR

Total RNA was extracted from HC11 cells and piglet jejunum using the Tissue RNA Purification Kit PLUS (EZB‐RN001‐plus) and reverse‐transcribed into cDNA via the 4× EZscript Reverse Transcription Mix II (EZB‐RT2GQ). Real‐time PCR was performed with SYBR qPCR Master Mix (EZB, A0012‐R2) on a QuantStudio 3 system, and the results were normalised to β‐actin expression. The relative abundance of genes was calculated via the 2^−∆∆CT method. The primers used are listed in Table S2.

2.10. Quantitative RT‐PCR of Lactobacillus

Microbial DNA was extracted from the faecal samples via a faecal genomic DNA extraction kit (Tiangen Biotech, China). DNA samples (100 ng/mL) were diluted 4‐fold for quantitative RT‐PCR to quantify the relative abundance of Lactobacillus species. The primers used are listed in Table S3. The relative abundance of Lactobacillus was calculated via the 2^−∆∆CT method (Hong et al. 2023).

2.11. Mouse Milk Sampling and Analysis

At L14, the pups were separated from the dams for 5 h before milk collection. Each dam was injected intraperitoneally with 4 IU of oxytocin and anesthetised with tribromoethanol (AibeiBio, Nanjing, China). Milk was aspirated via a 1 mL syringe and stored at −20°C. For fat analysis, 20 µL of milk was diluted with 40 µL of PBS and centrifuged at 3000 rpm for 20 min, and the top fat layer was collected for fat content estimation (Schwertfeger et al. 2003).

2.12. Cell‐Free Supernatant Preparation

Limosilactobacillus reuteri (BNCC 186563), L. plantarum (BNCC 364479) and L. johnsonii (BNCC 135265) (Table S4) were activated and adjusted to a concentration of 1 × 10⁵ CFU/mL. Each strain was inoculated into MRS medium at a 2% inoculum size. L. reuteri (BNCC 186563) and L. johnsonii (BNCC 135265) were cultured anaerobically at 37°C for 24 h, while L. plantarum (BNCC 364479) was cultured aerobically at 37°C for the same duration. After centrifugation at 1500 × g for 10 min at 4°C, the supernatant was collected and centrifuged again under the same conditions. The resulting supernatant was filtered through a 0.22 µm filter to remove any residual bacteria and debris. Cell viability was assessed using the Cell Counting Kit‐8 (HY‐K0301, Medchemexpress, China), and a 3% concentration of cell‐free supernatant was used for treatment, as it was determined to be suitable without negatively affecting HC11 cell viability.

2.13. Purification and Characterisation of EVs

L. johnsonii (BNCC 135265) was cultured anaerobically in MRS medium at 37°C for 72 h. Before inoculation, the medium was centrifuged at 180,000 × g for 3 h and filtered through a 0.22 µm filter to ensure a purified culture environment. To obtain the cell‐free supernatant, the culture was sequentially centrifuged at 8000 × g for 30 min and 10,000 × g for 45 min, followed by filtration through a 0.22 µm filter to remove dead cells and cellular debris. EVs were then isolated via ultracentrifugation at 180,000 × g for 1.5 h at 4°C using a Beckman Optima XE‐100 ultracentrifuge. The supernatant was discarded, and the EV pellet was resuspended in 5 mL of PBS. To further purify the EVs, the solution was ultracentrifuged again at 180,000 × g for 1 h, and the final pellet was resuspended in 2 mL of PBS. EVs were further purified and concentrated using the ExoSure Exosome Isolation Kit (GeneCopoeia, EP002), which combines precipitation and size exclusion chromatography, yielding a final EV solution volume of 200 µL. The same method was applied for isolating EVs from L. plantarum (BNCC 364479).

The protein concentration of the isolated EVs was determined using a BCA protein assay kit (Beyotime, Jiangsu, China) according to the manufacturer's instructions. For structural analysis, EVs were subjected to negative staining and visualised using transmission electron microscopy (Talos F200S). The particle size distribution of the EVs was analysed using nanoparticle tracking analysis (NTA) with a Zetaview (Particle Metrix) system, following a 200‐fold dilution in PBS.

2.14. Inhibition of EVs Secretion

To assess whether GW4869 inhibits EV production in L. johnsonii (BNCC 135265), the strain was cultured anaerobically at 37°C in 300 mL of MRS medium with shaking at 120 rpm. The bacterial suspension was adjusted to 10⁷ CFUs/100 mL, and GW4869 (10 µM, Selleck Chemicals LLC) was added to the culture medium. After 3 days of incubation, the bacterial cultures were resuspended in fresh MRS medium and continued to grow for another 3 days to evaluate the lasting inhibitory effects of GW4869 on EVs secretion. The protein content of the isolated EVs was quantified using a BCA Protein Assay Kit. Additionally, bacterial viability was assessed using both the MRS agar plate colony counting method and optical density measurements at 600 nm (OD600).

2.15. Tissue Distribution of EVs

To track the distribution of EVs in vivo, EVs were labelled with Dil (MedChemExpress) for fluorescence imaging. Briefly, EVs suspensions were mixed with 5 µM DiI solution and gently vortexed to ensure thorough staining. The mixture was incubated at 37°C for 30 min, followed by ultracentrifugation at 100,000 × g for 1 h to remove excess dye. The labelled EVs were then resuspended in PBS. The concentration of labelled EVs was quantified using a BCA assay, based on the total protein content of the purified EVs sample. The mice were gavaged with DiI‐labelled EVs (2.6 × 10¹⁰ particles/kg body weight) and fluorescence signals were recorded at 1, 6 and 12 h. The control mice were gavaged with equal volume PBS. After the final time point, the mice were sacrificed by cervical dislocation. Fluorescence signals in the mammary tissue, liver, kidney, spleen, lungs and heart were captured via the IVIS Lumina III imaging system.

2.16. EVs Uptake Assay

To assess the ability of HC11 cells to take up EVs, the cells were first starved in serum‐free medium for 8 h. Then, DiI‐labeled EVs were cocultured with the cells for 6 h. To evaluate the effect of endocytosis inhibition, HC11 cells were pretreated with Dynasore (12.5 µM) for 3 h prior to the addition of DiI‐labelled EVs, followed by 6 h of coculture. For the label‐only control, an equivalent amount of DiI (used for EV labelling) was subjected to the same labelling procedure without EVs, then incubated with the cells to determine whether free dye contributed to fluorescence. The control cells were cocultured with equal volume PBS as the EVs were resuspended in PBS. The cells were then fixed with 4% paraformaldehyde for 30 min. After nuclear staining with DAPI, fluorescence microscopy was performed to visualise the uptake of EVs.

2.17. EVs Untargeted Lipidomics

EVs from L. johnsonii (BNCC 135265) were isolated as described in Section 2.13. Multiple rounds of ultracentrifugation were performed, integrating precipitation and size exclusion chromatography to achieve the highest possible purity. To further minimise lipid interference from MRS medium, an equal volume of culture medium supernatant was collected after the initial ultracentrifugation and processed using the same EVs purification method as a control.

Untargeted lipidomics analysis was conducted using an ultra‐performance liquid chromatography‐tandem mass spectrometry (UPLC‐MS/MS) platform with a QExactive Plus mass spectrometer (Thermo Scientific). Lipid identification and data pre‐processing were performed using MSDIAL software. For sample preparation, EV samples were thawed at 4°C, and 200 µL was transferred into an EP tube for freeze‐drying. Subsequently, 150 µL of pre‐chilled butanol/methanol (1:1, v/v) solution containing 10 mmol/L ammonium formate was added, followed by thorough vortex mixing. Samples underwent ultrasonic extraction in an ice bath for 60 min, then centrifuged at 16,000 × g for 20 min at 4°C. The supernatant was collected for analysis. Chromatographic separation was performed using a SHIMADZU‐LC30 ultra‐high‐performance liquid chromatography (UHPLC) system equipped with a Hypersil GOLD C18 column (2.1 × 100 mm, 3 µm). Detection was conducted in both positive and negative ion modes using electrospray ionisation (ESI). After UHPLC separation, mass spectrometry analysis was performed on a QExactive Plus mass spectrometer (Thermo Scientific). Lipid identification and quantification were processed using MSDIAL software (ver5.2.240424.3‐net472). We are extremely grateful to Shanghai Bioprofile Technology Co.Ltd. for its support of our work.

2.18. EVs Fatty Acid Analysis

Fatty acid composition of EVs was analysed using gas chromatography‐mass spectrometry (GC‐MS). Gas chromatographic separation was performed on a TRACE 1300 gas chromatograph (Thermo Fisher Scientific, USA), while mass spectrometric detection was conducted using an ISQ 7000 mass spectrometer (Thermo Fisher Scientific, USA). Fatty acid quantitation of EVs were normalised to the protein content of samples.

2.19. Oil Red O Staining of Cells

Oil Red O staining was used to evaluate lipid droplet synthesis in HC11 cells. Briefly, the cells were washed three times with PBS and then fixed with 4% paraformaldehyde at room temperature for 30 min. After fixation, the cells were washed twice with PBS and stained with filtered Oil Red O solution (Sangon Biotech, Shanghai, China) for 2 h at room temperature. The staining solution was then removed, and the cells were rinsed with 60% isopropanol for 20 s. Next, the cells were washed five times with PBS, and lipid droplets were observed under an inverted microscope.

2.20. Mammary Gland Whole‐mount Staining

Mammary gland tissue was placed on a glass slide and fixed with Carnot's fixative for 12 h. The tissue was then stained with carmine overnight. After staining, the tissue was dehydrated through a graded ethanol series and finally cleared with xylene to achieve transparency.

2.21. Mammary Gland Oil Red O Staining

Fresh mammary gland tissue was used to prepare frozen sections. The sections were immersed in Oil Red O solution for 10 min, followed by nuclear staining with haematoxylin. After staining, the sections were gently washed with PBS to remove any excess dye and observed under an optical microscope.

2.22. Triglyceride Content Determination

A triglyceride assay kit (Jiancheng Bioengineering Institute, Nanjing, China) was used to measure both the intracellular and extracellular triglyceride concentrations. For extracellular triglycerides, the culture medium was aspirated, thoroughly mixed and analysed. The remaining cells were washed twice with PBS and lysed with RIPA lysis buffer (Beyotime, Jiangsu, China). The cells were then placed on a medium‐speed shaker for 15–20 min to ensure complete lysis. The resulting cell suspension was collected for protein concentration and intracellular triglyceride measurements. The triglyceride content was calculated by measuring the optical density (OD) of the samples at 510 nm, following the manufacturer's instructions.

2.23. CD36 Palmitoylation Assay

CD36 palmitoylation was assessed using the acyl‐biotin exchange (ABE) method. Briefly, total proteins were extracted from HC11 cells using a lysis buffer containing N‐ethylmaleimide (NEM) to block free thiol groups. CD36 was then enriched via immunoprecipitation using an anti‐CD36 antibody and magnetic beads (Millipore). To remove palmitate modifications, the samples were treated with a hydroxylamine (HAM) buffer (Millipore) and incubated at room temperature for 40 min. The exposed cysteine residues were then labelled by incubating the samples with biotin‐BMCC buffer at 4°C for 50 min. After labelling, proteins were eluted from the beads by boiling, followed by centrifugation to remove cell debris. The samples were subjected to SDS‐polyacrylamide gel electrophoresis (SDS‐PAGE), and the proteins were transferred onto a PVDF membrane. To detect palmitoylated CD36, the membrane was incubated with HRP‐conjugated anti‐streptavidin at 37°C for 1 h.

2.24. Surface Biotinylation

To assess surface protein expression, HC11 cells were incubated with sulfo‐NHS‐SS‐biotin (0.5 mg/mL) in PBS for 1 h on ice. Following incubation, cells were washed five times with PBS containing 50 mM glycine to quench unreacted biotin. Cells were then lysed using RIPA buffer, and the resulting lysates were incubated with streptavidin‐conjugated magnetic beads (Beyotime, P2151) at 4°C for 12 h. After incubation, the beads were resuspended in 0.1% SDS and heated at 95°C for 3 min. Beads were separated using a magnetic rack for 1 min, and the supernatant containing biotinylated surface proteins was collected for Western blot analysis.

2.25. Western Blot

The samples were lysed in RIPA buffer containing protease and phosphatase inhibitors (Beyotime, Jiangsu, China). After centrifugation at 12,000 × g for 15 min, the protein concentrations in the supernatants were determined via a BCA protein assay kit (Beyotime, Jiangsu, China). Equal amounts of protein were separated by 10% SDS‒PAGE and transferred onto PVDF membranes. The membranes were incubated overnight with the primary antibody, followed by incubation with the secondary antibody at room temperature for 90 min. Protein detection was performed via the Tanon 5200 imaging system. The grayscale values of the bands were analysed via ImageJ software. All the antibodies used in this study are shown in Table S5.

2.26. Quantification and Statistical Analysis

All the data were analysed via GraphPad Prism 8.0.2. A two‐tailed unpaired Student's t‐test was performed to evaluate the differences between the two groups. One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. The data are presented as the means ± SEM unless indicated otherwise, with * indicating statistical significance (p < 0.05), ** indicating highly significant differences (p < 0.01) and ns indicating no significant differences (p > 0.05).

3. Results

3.1. The Maternal Gut Microbiota Is Closely Related to Lactation Performance

Pigs are an ideal model for studying lactation because their physiological processes closely resemble those of humans. In this study, the lactation performance of sows was ranked on the basis of the weight gain of piglets from days 1 to 14 (Figure 1A). Piglets nursed by the HLS gained an average of 3.38 kg over the 14‐day period, whereas those nursed by the LLS gained 2.29 kg on average (Figure 1B). Owing to cross‐fostering within 24 h of birth, there were no significant differences in litter size or average birth weight between the two groups of piglets (Figure 1C,D). During the trial, no creep feed was introduced to the piglets, indicating that their growth during the first 14 days relied entirely on maternal milk. HLS resulted in greater milk production during the first 14 days (Figure 1E). Next, we compared the milk compositions of HLS and LLS. HLS produced milk with relatively high levels of fat and lactose in both colostrum and L14 milk (Figure S1A–F), with a particularly pronounced increase in fat content (Figure 1F,G). Additionally, the levels of IgA and IgM in HLS colostrum were significantly greater than those in LLS colostrum (Figure S1G–I). Piglet weight gain over the 14‐day period was positively correlated with milk non‐ fat solid and lactose contents (Figure S1J–M), with the strongest correlation observed between piglet weight gain and milk fat content (Figure 1H,I). Besides, a multiple‐regression model incorporating both milk yield and milk fat demonstrated that, after adjusting for milk yield, milk fat remained an independent positive predictor of piglet growth.

FIGURE 1.

Sow lactation performance and 16S rRNA amplicon sequencing of the faecal microbiota on gestation day 109. (A) Experimental design of the sow study. (B) Piglet weight gain over 14 days. (C) Litter size within the 14‐day period. (D) Birth weights of the piglets. (E) Sow milk yield during the first 14 days (n  =  6). (F) Colostrum fat composition of sows (n = 6). (G) D14 milk fat composition of sows (n  =  6). (H and I) Correlation analysis between colostrum (H) and D14 (I) milk fat composition and piglet weight gain. (J and K) Comparison of alpha diversity indices (Chao1 and Simpson indices). (L) Principal coordinate analysis (PCoA) based on Bray‒Curtis distances. (M) Relative abundance of the top 10 phyla in the faecal microbiota. (N) Relative abundance of the top 10 genera. (O) Microbial significance statistics at genus level. (P) RT‐qPCR of comparison of Lactobacillus abundance between HLS and LLS. (Q) Linear discriminant analysis (LDA > 2.7, p < 0.05) of the most differentially abundant bacteria between the HLS and LLS, with Spearman's correlations between the selected bacteria and milk components. A two‐tailed unpaired Student's t‐test was performed to evaluate the differences between the two groups. Statistical significance: *p < 0.05; **p < 0.01; ns: not significant (p > 0.05). HLS: High lactation sows. LLS: low‐lactation sows. LLO: offspring of low‐lactation sows. MLO: offspring of medium‐lactation sows. HLO: offspring of high‐lactation sows.

To further confirm that HLS has a greater capacity for milk fat synthesis, we isolated mammary epithelial cells from sow milk samples (Figure S2A). Compared with those in the LLS, the phosphorylation levels of key signalling proteins, including JAK2, STAT5, mTOR, S6K1 and 4EBP1, were significantly greater in the HLS (Figure S2B–E). Similarly, the expression of key enzymes involved in de novo lipogenesis (FASN, ACACA, DGAT1), regulatory factors (SREBP1, PPARγ) and fatty acid transport proteins (CD36, FABP4) was significantly elevated in the milk‐separated HLS cells compared with the LLS cells (Figure S2F–G). These findings indicate that the efficiency of milk fat synthesis in the mammary glands is greater in HLS than in LLS.

To explore the potential relationship between the gut microbiota and milk synthesis, we compared the faecal microbiotas of HLS and LLS. Although there was no significant difference in α diversity between the two groups (Figure 1J,K), principal coordinate analysis (PCoA) revealed a tendency for β diversity to segregate between the HLS and LLS (Figure 1L). At the phylum level, Firmicutes, Bacteroidetes and Proteobacteria were the dominant phyla in faecal microbiota samples from both groups (Figure 1M). Compared to LLS, HLS exhibited a higher relative abundance of Firmicutes and a lower abundance of Bacteroidetes. At the genus level, HLS showed an increased a tendency of abundance of Prevotella and Lachnospiraceae, while the relative abundance of Streptococcus, Christensenellaceae and Coprostanoligenes was reduced. However, only Lactobacillus achieved a statistical difference, which was significantly enriched in gut of HLS (Figure 1N–O). In addition, we further verified the high abundance of Lactobacillus in the HLS by RT‐qPCR quantification (Figure 1P). LDA combined with correlation analysis revealed a potential association between Lactobacillus and lactation performance. Specifically, Lactobacillus was positively correlated with milk fat, lactose and milk yield (p < 0.05) (Figure 1Q).

3.2. HLS‐Derived Intestinal Microbiota Stimulates Milk Synthesis in Mice

To assess whether the intestinal microbiota can influence maternal lactation performance, faecal microbiota transplantation (FMT) was performed. Faecal microbiota from LLS and HLS were transplanted into antibiotic‐treated mouse dams for 15 days (from L0 to L14) (Figure 2A). Broad‐spectrum antibiotic treatment significantly reduced the intestinal microbial abundance in the mice (Figure S3). After FMT, there was no significant difference in α diversity between the two groups of mice (Figure S4A). However, significant differences in β‐diversity were observed between the two groups (Figure S4B,C), with Lactobacillus being notably more abundant in the guts of the mice transplanted with HLS faecal microbiota (THM) (Figure S4D). LDA confirmed the enrichment of Lactobacillus in the THM group compared with the LLS‐transplanted (TLM) group (Figure S4E,F). The significant correlation of Lactobacillus abundance between sows and mice further suggested successful colonisation of the maternal mouse gut by sow‐derived Lactobacillus (Figure S4G).

FIGURE 2.

Impact of faecal microbiota transplantation on the lactation performance of ICR mice. (A) Experimental design schematic. (B–D) Body weights of pups (n = 6): (B) nursed by their biological mothers, (C) cross‐nursed by foster mothers starting at L0, and (D) cross‐nursed by foster mothers starting at L8. Litter sizes were adjusted to 13 for each mother on L0. The data are presented as the means ± SDs. A nonpaired t test was used to determine statistical significance between the OTHM and OTLM groups (*) and between the OTLM group and the control group (*). (E) Correlations between Lactobacillus abundance in the gut and pup body weight (n = 6). (F) Milk clots in the stomachs of pups at L3 (left), developmental progression of pups (centre) and body size at L14 (right). Scale bar: 10 mm. (G) Whole‐mount carmine staining of mammary glands at L14 (n = 3). (H) H&E‐stained mammary gland sections at L14, n = 3. Scale bar: 200 µm. (I) Oil Red O staining of milk fat in mammary gland sections, n = 3. Scale bar: 200 µm. (J) Representative image of milk fat after centrifugation. (K) Quantitative analysis of the alveolar area (n = 6), lipid droplet area (n = 3) and milk fat content (n = 3). (L‐M) Western blot bands (L) and quantification of the fold changes (M) in the expression of JAK2‐STAT5 signalling‐related proteins in mammary glands, n = 3. (N‒Q) Western blot bands (N) and quantification of fold changes (O‒Q) for milk fat synthesis‐related proteins in mammary glands, n = 3. One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. Statistical significance: *p < 0.05; ** p < 0.01; ns: not significant (p > 0.05). Spectrum antibiotics include ampicillin (1 g/L), gentamicin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L) and metronidazole (1 g/L). TLM: mice receiving faecal microbiota from low lactating sows. THM: mice receiving faecal microbiota from high lactating sows. OTLM: offspring of mice receiving faecal microbiota from low lactating sows. OTHM: offspring of mice receiving faecal microbiota from high lactating sows.

To evaluate lactation performance, the pup body weight was homogenised, and the litter size was adjusted to 13 at birth. By L5, pups nursed by THM presented significantly greater body weights, and their growth rate during the entire lactation period was significantly greater than that of pups nursed by TLM or the control group (Figure 2B). Additionally, THM‐nursed pups had more milk in their stomachs by L3 and significantly larger body sizes by L14 than did TLM‐nursed pups and control pups (Figure 2F). These results suggest improved lactation performance in THM dams. A positive correlation was observed between Lactobacillus abundance in the mouse gut and the body weight of pups, indicating that Lactobacillus may increase lactation performance (Figure 2E). To further investigate the impact of the gut microbiota on lactation, we conducted cross‐fostering experiments, switching pups born to THM/TLM dams to be nursed by TLM/THM dams, starting at either L0 or L8. Pups born to TLM but fostered by THM from L0 showed greater weight gain, more milk in their stomachs by L3, and a larger body size by L14 (Figure 2C). Even pups initially nursed by TLM experienced increased weight gain when fostered by THM starting at L8 (Figure 2D). These results demonstrate that the HLS‐derived gut microbiota significantly contributes to enhanced maternal lactation performance.

We then analysed the mammary glands of the dams via whole‐mount and haematoxylin‒eosin (H&E) staining. Whole‐mount analysis revealed that the mammary glands of the THM group had denser secretory lobuloalveolar structures than did those of the TLM group (Figure 2G). Higher magnification revealed enlarged mammary alveoli in the THM, which is indicative of milk accumulation (Figure 2H). Furthermore, Oil Red O staining (Figure 2I) and milk centrifugation (Figure 2J,K) revealed a greater abundance of lipid droplets in the THM mammary glands. At the molecular level, the phosphorylation levels of JAK2 and STAT5 were significantly greater in the mammary glands of the THM group than in those of the TLM group (Figure 2L,M). Similarly, the expression of key enzymes involved in de novo lipogenesis (FASN, ACACA), regulatory factors (SREBP1, PPARγ) and fatty acid transport proteins (CD36, FABP4) was significantly greater in the THM group than in the TLM and control groups (Figure 2N–Q). These findings suggest that the presence of Lactobacillus in the gut has a substantial effect on maternal milk fat synthesis.

3.3. Identification of Specific Lactobacillus Species Contributing to Milk Fat Synthesis

To further identify the specific Lactobacillus species contributing to milk fat synthesis, we assessed the relative abundance of Lactobacillus species in the sow gut. A recent large‐scale genomic analysis identified several dominant, pig‐specific bacterial genera in the gut, including Lactobacillus, Streptococcus, Clostridium, Desulfovibrio, Enterococcus and Fusobacterium (Wylensek et al. 2020). Building on these findings, we focused our analysis on the five most prevalent Lactobacillus species present in the sow gut, namely L. johnsonii, L. plantarum, L. reuteri, Lactobacillus acidophilus and Lacticaseibacillus paracasei. The relative abundances of L. plantarum, L. johnsonii and Limosilactobacillus ruminis were significantly greater in HLS than in LLS, whereas no significant differences were detected for L. acidophilus and L. paracasei (Figure S5A‐E). Given that Lactobacillus can influence the host through the secretion of bioactive metabolites and EVs, we prepared cell‐free supernatants from MRS media cultured with these Lactobacillus species and added them to HC11 cells culture media (Figure S5F). After 72 h of incubation, the L. johnsonii–derived cell‐free supernatant significantly increased lipid droplet formation and triglyceride content in both the HC11 cells and the culture media compared with the other Lactobacillus species tested (Figure S5G–M). This effect was accompanied by a significant upregulation of key lipogenic genes, including FASN, ACACA, DGAT1, PPARγ, CD36 and FABP4 (Figure S5N‒S). These findings suggest that the increased abundance of L. johnsonii in the maternal gut may play a key role in improving lactation performance.

Consistent with the results of the in vitro experiments, L. johnsonii transplantation (LJM) resulted in lactation capabilities comparable to those of THM (Figure 3A,B). Compared with the control offspring, the offspring of LJM (OLJM) had more milk in their stomachs at L3 and greater body sizes from L5 to L14 (Figure 3C). Additionally, the faecal microbiota from LLS combined with L. johnsonii transplantation resulted in offspring weight gain similar to that observed with THM (Figure S6). Structurally, the secretory lobuloalveolar units in the LJM mammary glands were notably denser than those in the control group (Figure 3D). Moreover, the LJM mammary glands presented greater milk filling and lipid droplet contents (Figure 3E–H). At the molecular level, the phosphorylation of JAK2 and STAT5, as well as the expression of lipogenesis‐related proteins, was elevated in the mammary glands of LJM mice compared with those of control mice (Figure 3I–N).

FIGURE 3.

L. johnsonii enhanced lactation performance in ICR mice. (A) Experimental design schematic. (B) Body weights of the pups (n = 6). The litter size was adjusted to 13 for each mother on L0. The data are the means ± SD. A nonpaired t test was used to determine statistically significant differences between the treatment group and the control group. (C) Milk clots in the stomachs of pups at L3 (left), developmental progression of pups (centre) and body size at L14 (right). Scale bar: 10 mm. (D) Whole‐mount carmine staining of mammary glands at L14 (n = 3). (E) H&E‐stained mammary gland sections at L14, n = 3. Scale bar: 200 µm. (F) Oil Red O staining of milk fat in mammary gland sections, n = 3. Scale bar: 200 µm. (G) Representative image of milk fat after centrifugation. (H) Quantitative analysis of the alveolar area (n = 6), lipid droplet area (n = 3), and milk fat content (n = 3). (I and J) Western blot bands (I) and quantification of the fold changes (J) in the expression of JAK2‐STAT5 signalling‐related proteins in mammary glands, n = 3. (K–N) Western blot bands (K) and quantification of the fold changes (L–N) in milk fat synthesis‐related proteins in mammary glands, n = 3. One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. Statistical significance: *p < 0.05; **p < 0.01; ns: not significant (p > 0.05). Spectrum antibiotics include ampicillin (1 g/L), gentamicin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L) and metronidazole (1 g/L). LJM: mice receiving L. johnsonii transplantation. THM: mice receiving faecal microbiota transplantation from highly lactating sows. OTHM: offspring of mice receiving faecal microbiota transplantation from high lactating sows. OLJM: offspring of mice receiving L. johnsonii transplantation.

3.4. L. Johnsonii Enhances Mammary Milk Fat Secretion Through EVs

To investigate the interaction between L. johnsonii and the host mammary glands, we hypothesised that L. johnsonii regulates mammary epithelial function through EVs. EVs were isolated and purified from L. johnsonii (Figure 4A), and transmission electron microscopy revealed their saucer‐like morphology (Figure 4B). Nanoparticle tracking analysis revealed that the EVs had a diameter of 143 nm (Figure 4C). After coculture of EVs with HC11 cells for 6 h, the red‐Dil signal was absorbed by the cells and accumulated in the perinuclear region (Figure 4D). EVs uptake can occur through multiple pathways, and previous studies have shown that the dynamin‐2 inhibitor dynasore effectively blocks this process (Menck et al. 2015). In our study, pretreating HC11 cells with dynasore for 3 h significantly reduced the internalisation of DiI‐labelled EVs (Figure 4D). Both the label‐only and vehicle (PBS) controls showed no detectable fluorescent signals, confirming that the observed fluorescence was due to the uptake of intact EVs rather than free dye. Next, we examined the functional impact of L. johnsonii‐derived EVs on HC11 cells. Compared with vehicle‐treated cells, exposure to L. johnsonii EVs led to a marked increase in lipid droplet accumulation (Figure 4E,F), elevated intracellular and extracellular triglyceride levels (Figure 4G,H), and upregulated expression of genes and proteins involved in milk fat synthesis (Figure 4I–O). These results suggest that L. johnsonii‐derived EVs actively promote lipid synthesis in mammary epithelial cells. To validate this finding in vivo, we administered L. johnsonii pretreated with GW4869, a known inhibitor of EV secretion, to lactating dams (the LGM group) (Figure 5A). While GW4869 treatment did not impair bacterial viability, it significantly reduced EVs production (Figure S7). Notably, offspring nursed by GW4869‐treated dams exhibited reduced growth performance compared to those in the LJM group (Figure 5B,E). In cross‐fostering experiments, offspring born to LJM but fostered by LGM from L0 had lower weight gain, less milk in their stomachs and lower final weights at L14 (Figure 5C). However, when offspring born to the LGM were fostered by LJM starting at L8, their weight gain improved (Figure 5D). Additionally, compared with LJM, LGM resulted in lower alveolar filling, fewer lipid droplets and reduced expression of milk fat synthesis‐related proteins in the mammary gland (Figure 5F–M).

FIGURE 4.

Extracellular vesicles (EVs) derived from L. johnsonii stimulate milk fat synthesis in HC11 cells. (A) Schematic representation of the experimental design. (B and C) Transmission electron microscopy (TEM) (B) and nanoparticle tracking analysis (NTA) (C) of EVs isolated from cultured L. johnsonii. Scale bars: 50 nm. (D) Fluorescence microscopy of HC11 cells treated with DIL‐labelled EVs for 6 h. Scale bar: 200 µm. (E and F) Oil Red O staining of HC11 cells (E) and quantification of the lipid droplet area (F) (n = 3). Scale bar: 200 µm. (G and H) Triglyceride concentrations in HC11 cells (G) and culture medium (H) (n = 6). (I–K) mRNA expression of milk fat synthesis‐related genes in HC11 cells, n = 4. (L–O) Western blot bands of milk fat synthesis‐related proteins (L) and fold change quantification (M–O) (n = 3). A two‐tailed unpaired Student's t‐test was performed to evaluate the differences between the two groups. Statistical significance: *p < 0.05; **p < 0.01. Dynasore (12.5 µM): EVs uptake inhibitor.

FIGURE 5.

Inhibiting extracellular vesicle secretion diminishes the stimulatory effect of L. johnsonii on lactation performance in ICR mice. (A) Schematic representation of the experimental design. (B–D) Body weights of pups nursed by biological mothers (B) and cross‐nursed by foster mothers at L0 (C) and L8 (D) (n = 6). Litter sizes were adjusted to 13 for each mother on L0. The data are presented as the means ± SDs. An unpaired t test was used to evaluate the statistical significance between OLJM and OYGM. (E) Milk clots in the stomachs of pups at L3 (left), developmental progression of pups (centre) and body size at L14 (right). Scale bar: 10 mm. (F) Whole‐mount carmine staining of mammary glands at L14 (n = 3). (G) H&E‐stained sections of mammary glands at L14 (n = 3). Scale bar: 200 µm. (H) Oil Red O staining of milk fat in mammary gland sections (n = 3). Scale bar: 200 µm. (I) Representative image of milk fat after centrifugation. (J) Quantitative analysis of the alveolar area (n = 6), lipid droplet area (n = 3) and milk fat content (n = 3). (K‒M) Western blot bands of milk fat synthesis‐related proteins (K) and fold change quantification (L and M) (n = 3). One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. Statistical significance: *p < 0.05; **p < 0.01. Spectrum antibiotics include ampicillin (1 g/L), gentamicin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L) and metronidazole (1 g/L). LJM: mice receiving L. johnsonii transplantation. LGM: mice receiving L. johnsonii pretreated with GW4869 (a selective inhibitor of EV secretion). OLJM: offspring of mice transplanted with L. johnsonii. OLGM: offspring of mice transplanted with GW4869‐treated L. johnsonii.

To determine whether the observed effects were specific to L. johnsonii‐derived EVs, we included L. plantarum as a comparative control, given its probiotic relevance and EVs production capacity. DiI‐labelled EVs from both strains were orally administered to lactating dams to assess their systemic biodistribution (Figure 6A). The characterisation of L. plantarum‐derived EVs is shown in Figure S8. Six hours post‐administration, fluorescence signals were detected in the mammary glands, demonstrating that EVs from both L. johnsonii and L. plantarum were systemically distributed and successfully reached the mammary tissue (Figure 6B,C). Compared to offspring of L. plantarum EVs‐treated dams (OMPEVs) and control dams, offspring of L. johnsonii EV‐treated dams (OMJEVs) exhibited significantly greater weight gain (Figure 6D,E). This improvement was associated with increased milk yield, as evidenced by more expanded alveoli and higher milk fat content (Figure 6F,J). Consistently, milk fat synthesis‐related proteins were upregulated in the mammary glands of MJEVs dams (Figure 6K,L), but not in MPEVs or CON dams. These results suggest that L. johnsonii‐derived EVs significantly contribute to enhanced maternal lactation performance.

FIGURE 6.

Extracellular vesicles derived from L. johnsonii enhance lactation performance in ICR mice. (A) Schematic representation of the experimental design. (B and C) Localisation of EVs in the body detected by in vivo imaging (B) and within the mammary gland via ex vivo fluorescence imaging (C). (D) Milk clots in the stomachs of pups at L3 (left), developmental progression of pups (centre), and body size at L14 (right). Scale bar: 10 mm. (E) Body weights of pups nursed by their biological mothers (n = 6). Litter sizes were adjusted to 13 for each mother on L0. The data are presented as the means ± SDs. A nonpaired t test was used to assess statistical significance. (F) Whole‐mount carmine staining of mammary glands at L14 (n = 3). (G) H&E‐stained mammary gland sections at L14 (n = 3). Scale bar: 200 µm. (H) Oil Red O staining of milk fat in mammary gland sections (n = 3). Scale bar: 200 µm. (I) Representative image of milk fat after centrifugation (n = 3). (J) Quantitative analysis of the alveolar area (n = 6), lipid droplet area (n = 3), and milk fat content (n = 3). (K and L) Western blot bands of milk fat synthesis‐related proteins (K) and fold change quantification (L) in mammary glands (n = 4). One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. Statistical significance: *p < 0.05; **p < 0.01. Spectrum antibiotics include ampicillin (1 g/L), gentamicin (1 g/L), vancomycin (0.5 g/L), neomycin (1 g/L) and metronidazole (1 g/L). MJEVs: mice gavaged with L. johnsonii EVs. OMJEVs: offspring of mice gavaged with L. johnsonii EVs. MPEVs: mice gavaged with L. plantarum EVs. OMPEVs: offspring of mice gavaged with L. plantarum EVs.

3.5. L. Johnsonii EVs‐Derived Lipids Stimulate HC11 Milk Fat Synthesis

To identify the specific components in L. johnsonii‐EVs responsible for stimulating milk fat synthesis, we treated HC11 cells with L. johnsonii‐EVs and L. johnsonii‐EVs pre‐treated with DNase, RNase or proteinase. The elimination of DNA, RNA and proteins did not diminish the stimulatory effect of L. johnsonii‐EVs on milk fat synthesis (Figure S9). We then isolated the organic phase from the L. johnsonii‐EVs (Figure 7A) and found that, compared with the inorganic phase, treating HC11 cells with the organic phase significantly enhanced milk fat secretion. This was evidenced by increased lipid droplet formation, elevated triglyceride concentrations, and increased expression of lipogenesis‐related proteins (Figure 7B–N).

FIGURE 7.

Lipids from L. johnsonii extracellular vesicles stimulate milk lipid synthesis in HC11 cells. (A) Schematic representation of the lipid extraction process from L. johnsonii EVs. (B and C) Oil Red O staining of HC11 cells (B) and quantification of the lipid droplet area (C) (n = 3). Scale bar: 200 µm. (D and E) Triglyceride concentrations in cells (D) and culture medium (E) (n = 6). (F–I) mRNA expression levels of milk fat synthesis‐related genes in HC11 cells (n = 4). (J–N) Western blot bands of milk fat synthesis‐related proteins (J) and quantification of fold changes (K–N) (n = 3). One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. Statistical significance: *p < 0.05; **p < 0.01; ns: not significant (p > 0.05).

3.6. L. Johnsonii‐Derived EVs Fatty Acids Stimulate Lipid Internalisation via CD36 Palmitoylation and Promote Milk Fat Synthesis in HC11 Cells

To identify the lipid components in L. johnsonii‐derived EVs that may regulate milk fat synthesis, we performed lipidomic analysis. Consistent with previous studies (Skotland et al. 2019; Haraszti et al. 2016), L. johnsonii‐derived EVs were enriched in glycerophospholipids (GP), sphingolipids (SP) and sterol lipids (ST) (Figure 8A). At the subclass level, L. johnsonii‐EVs showed a higher abundance of neutral lipids such as triglycerides (TG) and diacylglycerol (DG), along with structural phospholipids including phosphatidylinositol (PI), phosphatidylcholine (PC) and phosphatidylserine (PS), as well as bioactive sphingolipids like ceramides (Cer) (Figure 8B). The complete untargeted lipidomics data are provided in Table S6. It is important to note that lipids carried by EVs, such as glycolipids and phospholipids, can be enzymatically hydrolysed either extracellular or within recipient cells, releasing free fatty acids (FFAs) that act as bioactive signalling molecules. Based on this rationale, we analysed the FFA composition of L. johnsonii‐EVs. Our results showed that these EVs were predominantly enriched in saturated FFAs, with palmitic acid (C16:0) and stearic acid (C18:0) being the most abundant species (Figure 8C). These data are consistent with the elevated levels of PC and DGDG species containing C16:0 and C18:0 in L. johnsonii‐EVs (Table S6). The complete profile of FAs is provided in Table S7. Subsequently, we further evaluated the specific roles of C16:0 and C18:0 through in vitro experiments. Our results demonstrate a clear dose‐dependent effect of palmitic acid (C16:0) and stearic acid (C18:0) on lipid secretion in HC11 cells, with the most pronounced response observed between 10 and 15 µM (Figure 8D). Quantitative analysis revealed that EVs at a concentration of 1.3 × 10¹⁰ particles/mL contained approximately 12.8 µM of C16:0 and 13 µM of C18:0, respectively. Based on these findings, 10 µM was selected as the standardised dose for subsequent experiments. As expected, treatment with 10 µM palmitic acid (C16:0) significantly enhanced milk fat synthesis in HC11 cells compared to stearic acid (C18:0), as evidenced by increased lipid droplet accumulation and elevated triglyceride levels (Figure 8E,F). Palmitoylation of CD36 has been identified as a key post‐translational modification regulating its membrane localisation and fatty acid transport activity (Hao et al. 2020). Since CD36 must be depalmitoylated to initiate fatty acid uptake, we examined its palmitoylation dynamics following C16:0 treatment. As shown in Figure 8G,H, CD36 palmitoylation began to decline 30 min after C16:0 exposure, reaching its lowest level between 60 and 120 min. Concurrently, Caveolin‐1 (CAV1), a structural protein of caveolae, was observed to co‐migrate with CD36, indicating a potential role in CD36 trafficking. Notably, the reduction in palmitoylation corresponded with internalisation of CD36 from the plasma membrane (Figure 8I,J). To determine whether CD36 could be recycled back to the cell surface after fatty acid withdrawal, we removed palmitic acid (C16:0) from the culture medium and monitored its localisation over time. Within 30 min, CD36 began to re‐palmitoylate and gradually reappeared at the plasma membrane. By 4 h, most of the internalised CD36 had returned to the membrane, as confirmed by surface biotinylation assays (Figure S10). These results suggest that dynamic palmitoylation of CD36 is tightly linked to its membrane trafficking and may serve as a regulatory switch in fatty acid uptake during milk fat synthesis.

FIGURE 8.

Palmitic acid (C16:0) promote CD36 palmitoylation, facilitating fatty acid uptake and enhancing milk fat synthesis in HC11 cells. (A and B) Lipid composition of L. johnsonii EVs (n=3). (C) Top five most abundant saturated fatty acids monounsaturated, and polyunsaturated in L. johnsonii EVs (n=3). (D) The analysis of the dose‐dependent effects of each fatty acid on milk fat synthesis in HC11 cells. (E and F) Oil Red O staining of HC11 cells (E) and quantification of lipid droplet area (F) (n=3). Scale bar: 200µm. (G and H) CD36 palmitoylation levels at different time points following C16:0 treatment. (I and J) The CD36 internalisation levels. (K) Triglyceride concentrations in HC11 cells after inhibiting the internalisation of CD36 (n=6). (L and M) Oil Red O staining of HC11 cells (L) and quantification of the lipid droplet area (M) (n=3). One‐way ANOVA followed by Tukey's multiple comparisons test were used to evaluate the differences more than two group. (0.1% DMSO as the vehicle for C18:0 and C16:0). Statistical significance: *p<0.05; *p<0.01; ns: not significant (p>0.05). The dosages of PP2, ML348 and piceatannol are 20, 10 and 40 µM, respectively. C16:0: Palmitic acid. C18:0: Stearic acid. PP2: the inhibitor of LYN. ML348: the inhibitor of APT1. Piceatannol: the inhibitor of SYK.

CD36 captures extracellular fatty acids at the plasma membrane in its palmitoylated form and initiates endocytosis upon depalmitoylation (Hao et al. 2020). This dynamic cycle is regulated by DHHC5, a palmitoyltransferase inactivated by LYN kinase upon fatty acid binding. Inactivation of DHHC5 enables APT1 to depalmitoylate CD36, which then recruits SYK to activate VAV and JNK, promoting cytoskeletal remodelling and clathrin‐mediated endocytosis (Zeke et al. 2016; Doherty and McMahon 2009; Zhang et al. 2025). Inhibiting LYN, APT1 or SYK disrupts this pathway and blocks lipid droplet formation (Hao et al. 2020). Building on these insights, we investigated whether C16:0 promotes milk fat synthesis primarily by enhancing CD36‐mediated fatty acid transport. When HC11 cells were cultured in serum‐free, fatty acid‐depleted medium, the addition of C16:0 alone led to only modest lipid accumulation. However, in complete medium containing abundant extracellular fatty acids, supplementation with C16:0 resulted in a marked increase in milk fat synthesis (Figure 8K–M). This effect was significantly diminished by pre‐treatment with a CD36 internalisation inhibitor, suggesting that the presence of extracellular fatty acids is necessary for C16:0 to fully exert its effect. These findings indicate that C16:0 does not serve as the main lipid substrate itself but rather functions as a signalling molecule that activates CD36‐dependent endocytosis, thereby enhancing the uptake of ambient fatty acids and promoting milk fat synthesis in mammary epithelial cells. Besides, the accumulation of fatty acids promotes lipid droplet formation in mammary epithelial cells, primarily through the activation of PPARγ (Li et al. 2019; Belal et al. 2018). As expected, in this study, inhibition of PPARγ abolished the stimulatory effects of the organic phase on milk fat synthesis in HC11 cells (Figure S11). These findings indicate that the palmitic acid (C16:0) carried by L. johnsonii EVs facilitates fatty acids uptake by inducing dynamic palmitoylation of CD36, while the accumulated fatty acids further stimulate milk fat synthesis in HC11 cells via PPARγ signalling.

4. Discussion

While growing evidence links the gut microbiota to maternal lactation performance, the underlying mechanisms remain unclear. In this study, we explored the role of L. johnsonii in enhancing lactation. Notably, we observed that L. johnsonii‐derived EVs are critical mediators in the communication between the gut and the mammary gland. Our findings provide new evidence that L. johnsonii promotes milk fat synthesis by enhancing CD36‐mediated fatty acid uptake and activating the PPARγ signalling pathway.

Maternal gut microbiota undergoes significant changes during pregnancy and lactation, profoundly influencing lactation function. Studies have demonstrated that antibiotic‐induced depletion of gut microbiota impairs lactation, highlighting the importance of microbial communities in supporting milk production. Several Lactobacillus species, including L. acidophilus, L. casei, and L. plantarum, have been reported to enhance lactation performance in dairy cows (Boyd et al. 2011), sows (Betancur et al. 2021) and mice (Azagra‐Boronat et al. 2020). Consistent with these findings, we observed a significantly higher abundance of L. johnsonii in the gut of HLS compared to LLS. Additionally, Prevotella and Lachnospiraceae exhibited an increasing trend. Lachnospiraceae and Prevotella encode a large repertoire of carbohydrate‐degrading enzymes and are known producers of short‐chain fatty acids (SCFAs) (De Filippo et al. 2010). Previous studies have shown that cows receiving Prevotella exhibit increased milk fat content, which is associated with elevated acetate and butyrate concentrations in the rumen (Chiquette et al. 2008; Xue et al. 2020). Similarly, specific Prevotella species have been correlated with increased milk yield (Indugu et al. 2017). The milk fat concentration, particularly during mid‐lactation, has a major effect on offspring weight gain and is a key determinant of increased weaning weight (Rempel et al. 2023; Chisoro et al. 2023). Previous studies have shown that newborn piglets have approximately 1% body fat at birth, which increases to 16.4% by day 21 of lactation, reflecting rapid fat deposition (Frondas‐Chauty et al. 2012). Notably, 81.4% of milk fat is directly deposited in piglets, leading to a fivefold increase in body fat during lactation (Chisoro et al. 2023). Identifying gut microbes capable of activating lactation‐associated signalling pathways and enhancing maternal milk fat synthesis could provide valuable strategies for optimising lactation performance and improving offspring growth.

EVs play an important role in gut microbiota‐host communication. Recent studies have demonstrated that EVs can be transported to various organs, including the liver, lungs, brain, heart and kidneys, where they exert diverse biological effects (Tong et al. 2023; Fujita et al. 2018; Zhao et al. 2021; Díaz‐Garrido et al. 2021). Notably, growing research on Lactobacillus‐derived EVs has revealed their broad functional potential, including antimicrobial, immunomodulatory and gut‐protective effects (Tong et al. 2021). EVs derived from L. plantarum (Hao et al. 2021), L. rhamnosus GG (Tong et al. 2021), L. johnsonii (Li et al. 2024), L. fermentum (Wang et al. 2025) and L. paracasei (Choi et al. 2020) have been reported to effectively alleviate DSS/LPS‐induced colitis by downregulating pro‐inflammatory cytokines, enhancing intestinal barrier integrity, and restoring gut microbial homeostasis, underscoring their potential in gut health regulation. Additionally, the combination of L. reuteri‐ and Lc. paracasei‐derived EVs has been shown to enhance bone mass and mitigate the osteoporotic phenotype induced by ovariectomy (Wang et al. 2025). Moreover, EVs from L. paracasei (Kwon et al. 2023), and L. plantarum (Choi et al. 2019) have been implicated in the gut‐brain axis, influencing mood regulation and appetite control in humans. EVs derived from L. johnsonii have been well characterised in terms of their morphology, composition, and function. Scanning electron microscopy has shown that these EVs range from 90 to 125 nm in diameter (Harrison et al. 2021), consistent with our observations. Compositionally, L. johnsonii EVs are rich in phospholipids and contain a signature surface protein, Sdp (Harrison et al. 2021). Notably, IgA and IgG antibodies targeting the Sdp domain have been detected in the plasma of individuals consuming L. johnsonii, suggesting that these EVs can mediate systemic immune responses (Harrison et al. 2021). Functionally, most research has focused on the anti‐inflammatory properties of L. Johnsonii EVs. Recent studies have shown that they promote the polarisation of macrophages toward the anti‐inflammatory M2 phenotype (Teixeira et al. 2022; Tao et al. 2025) and reduce intestinal inflammation by suppressing ERK signalling (Tao et al. 2025). Additionally, L. Johnsonii EVs can protect βlox5 cells from cytokine‐induced apoptosis, potentially through activation of the aryl hydrocarbon receptor (AHR) pathway (Teixeira et al. 2022). These EVs are also internalised by intestinal epithelial cells, where they activate the Nrf2/HO‐1 antioxidant pathway, reducing endotoxin‐induced damage and supporting intestinal barrier integrity (Li et al. 2024). Furthermore, the SH3b2 domain of the Sdp protein, enriched in L. johnsonii EVs, has been shown to inhibit viral infection in vivo (Da Silva et al. 2024), highlighting the broad bioactive potential of these vesicles.

Although substantial evidence supports the role of Lactobacillus in enhancing lactation performance, the precise mechanisms remain largely unexplored. In this study, we detected L. johnsonii‐derived EVs in the mammary glands of mice and found that they enhanced milk fat synthesis efficiency. In contrast, administration of L. plantarum‐derived EVs had no significant effect on lactation performance. Recent study has shown that the proteins, lipids, RNAs (microRNAs) and metabolites components in EVs regulate various physiological functions on the host through transfer by Lactobacillus in the gut. Furthermore, the components and mechanisms of action of Lactobacillus‐derived EVs are gradually becoming more evident. Our findings indicate that L. johnsonii regulates milk fat synthesis primarily through its ester compounds. Lipid analysis of L. johnsonii EVs revealed that palmitic acid, stearic acid, oleic acid and linoleic acid were the most abundant fatty acids, consistent with previous studies (Charlet et al. 2022). These fatty acids are well‐documented for their role in enhancing milk fat synthesis in mammals (Piantoni et al. 2021; Shepardson and Harvatine 2021; Yang et al. 2023; Li et al. 2022). CD36 is a key transporter for fatty acid uptake in mammary gland milk fat synthesis (McManaman and Neville 2003). Palmitoylated CD36 captures fatty acids at the cell surface, while depalmitoylated CD36 facilitates endocytosis, transporting fatty acids into the cell (Zhang et al. 2025; Hao et al. 2020). Additionally, palmitate (C16:0) have been shown to modulate CD36 palmitoylation dynamics, further supporting our findings (Zhang et al. 2025; Hao et al. 2020). Furthermore, long chain fatty acids serve as natural ligands for PPARγ, a key regulator of lipid metabolism (Marion‐Letellier et al. 2016; Wang et al. 2007; Sauma et al. 2006; Zhang et al. 2022; Kliewer et al. 1997; Harvatine et al. 2009; Liu et al. 2016; Bionaz and Loor 2008). In vitro, we also observed that accumulated fatty acids stimulated PPARγ activity in HC11 cells, thereby enhancing milk fat synthesis. However, additional unidentified lipid compounds may also contribute to PPARγ activation, warranting further investigation. Notably, L. plantarum‐EVs contained significantly lower levels of palmitic acid (C16:0) compared to L. johnsonii‐EVs, which may explain their weaker effect on lactation performance (Table S7). These findings suggest that the fatty acid composition of bacterial EVs plays a crucial role in regulating mammary gland function and lactation efficiency.

Alterations in the maternal gut microbiota may impact not only the composition of milk but also the establishment and development of the offspring's gut microbiota, thereby significantly influencing the growth and development of the offspring (Vatanen et al. 2022). Previous studies have shown that the maternal gut microbiota can be transmitted to offspring through multiple pathways, including transplacental, reproductive, milk‐mediated and faecal transmission (Moeller et al. 2016). Consistent with these findings, our study revealed that L. johnsonii, L. reuteri and L. plantarum were significantly enriched in the intestines of piglets from the HLS compared with those from the LLS (p < 0.05) (Figure S12A‐C). Research indicates that Lactobacillus species positively modulate the immune response by balancing proinflammatory and anti‐inflammatory cytokines. For example, L. rhamnosus TL2937 mitigates LPS‐induced inflammation in piglets by downregulating TLR4‐mediated NF‐κB and MAPK signalling (Shimazu et al. 2012). Other species, such as L. salivarius (Sun et al. 2020), L. delbrueckii (Mcilvride et al. 2019), L. johnsonii, L. reuteri (Wang et al. 2020) and L. plantarum (Wang et al. 2019), also reduce LPS‐induced inflammatory responses. Similarly, our study revealed that the colonisation of Lactobacillus in the offspring gut increased intestinal barrier function (Figure S12) and reduced inflammation (Figures S12–S14), thereby promoting improved growth and development in offspring.

5. Conclusions

This study highlights the critical role of gut microbiota in regulating lactation performance, with a particular focus on L. johnsonii. Our findings demonstrate that L. johnsonii enhances mammary gland function through the release of EVs, leading to increased milk fat content and improved offspring growth. The palmitic acid (C16:0) carried by these EVs facilitates fatty acids uptake by inducing dynamic palmitoylation of CD36, while the accumulated fatty acids further stimulate milk fat synthesis in HC11 cells via PPARγ signalling. These results suggest that L. johnsonii‐derived EVs serve as key mediators of gut‐mammary communication, providing new insights into the mechanisms by which intestinal bacteria regulate milk production.

Author Contributions

Qihui Li: Data curation (equal); formal analysis (equal); visualization (lead); writing ‐ original draft (lead). Baofeng Li: Data curation (equal); validation (equal); visualization (equal). Qianzi Zhang: Data curation (equal); validation (equal). Dongpang Chen: Data curation (equal); validation (equal); visualization (equal). Siyu Yuan: Validation (equal). Hanyu Jing: Validation (equal). Haobin Li: visualization (equal). Wutai Guan: Project administration (equal); writing ‐ review and editing (equal). Shihai Zhang: Conceptualization (lead); project administration (lead); writing ‐ review and editing (lead).

Ethics Statement

All the animal experiments were conducted in accordance with the ethical policies and procedures approved by the Animal Care and Use Committee of South China Agricultural University (Guangzhou, China).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Fig.1: Differences in the Milk Composition of High‐ and Low‐Lactation Sows.

Supporting Fig.2: Differences in Milk Fat Synthesis Efficiency between Mammary Glands from High‐ and Low‐Lactation Sows

Supporting Fig.3: Comparison of bacterial colony‐forming units (CFUs) in feces from control and broad‐spectrum antibiotic‐treated mice

Supporting Fig.4: 16S rRNA amplicon sequencing of the mouse gut microbiota at L14

Supporting fig.5: Effect of Lactobacillus‐derived cell‐free supernatant on milk fat synthesis in HC11 cells

Supporting Fig.6: L. johnsonii Supplementation Improves Lactation Performance in Mice Transplanted with Faecal Microbiota from Low‐Lactation Sows.

Supporting Fig.7: GW4869 Inhibits the Secretion of Extracellular Vesicles from L. johnsonii

Supporting Fig.8: Characterization of Lp. plantarum BNCC364479 EVs

Supporting Fig.9: Nucleic Acids and Proteins in L. johnsonii extracellular Vesicles do Not Stimulate Milk Fat Synthesis in HC11 Cells

Supporting Fig.10: R Figure S10. Removal of C16:0 restores CD36 palmitoylation and membrane localization.

Supporting Fig.11: L. johnsonii EVs Lipids Stimulate Milk Lipid Synthesis in HC11 Cells via PPARγ Signaling

Supporting Fig.12: Offspring of Highly Lactating Sows Are Less Susceptible to LPS‐Induced Intestinal Inflammation

Supporting Fig.13: Lactobacillus Mitigates LPS‐Induced Intestinal Morphological Damage

Supporting Fig.14: Offspring of Mice Transplanted with L. johnsonii Show Enhanced Resistance to LPS‐Induced Colitis

Supporting table 1: Composition and nutritional content of the sow diet

Supporting table 2: Primers for real‐time PCR

Supporting table 3: Primers for Lactobacillus species real‐time PCR

Supporting table 4: Information of Lactobacillus species

Supporting table 5: Antibody information

Supplementary Table: jev270143‐sup‐0002‐TableS6.xlsx

Supplementary Table: jev270143‐sup‐0003‐TableS7.xlsx

Data Availability Statement

The accession for the 16s sequencing data in this study is in the NCBI Sequence Read Archive: # PRJNA1176286.