Lactobacillus amylovorus extracellular vesicles mitigate mammary gland ferroptosis via the gut-mammary gland axis

Zhang, Qianzi; Chen, Dongpang; Ding, Hanting; Li, Qihui; Yuan, Siyu; Li, Haobin; Guan, Wutai; Zhang, Shihai

doi:10.1038/s41522-025-00752-4

Download PDF

Article
Open access
Published: 21 June 2025

Lactobacillus amylovorus extracellular vesicles mitigate mammary gland ferroptosis via the gut-mammary gland axis

Qianzi Zhang¹,
Dongpang Chen¹,
Hanting Ding¹,
Qihui Li¹,
Siyu Yuan¹,
Haobin Li¹,
Wutai Guan^1,2,3 &
…
Shihai Zhang ORCID: orcid.org/0000-0002-5397-3354^1,2,3

npj Biofilms and Microbiomes volume 11, Article number: 113 (2025) Cite this article

2512 Accesses
Metrics details

Subjects

Abstract

Lactation is essential for supporting neonatal growth and development, and its regulation is influenced by the gut microbiota. However, the role of gut microbes in lactation under conditions of oxidative stress remains unclear. In this study, we identify a novel function for gut microbiota in regulating maternal lactation through the modulation of ferroptosis in the mammary gland under oxidative stress. We identify Lactobacillus amylovorus (L. amylovorus), enriched in mothers with low oxidative stress, as negatively correlating with both oxidative stress and ferroptosis. In a mouse model, L. amylovorus alleviates mammary ferroptosis and promotes lactation. In addition to producing of short-chain fatty acids, L. amylovorus secretes bacterial extracellular vesicles (BEVs) enriched in oleic acid, a monounsaturated fatty acid that can be transferred to the mammary gland. Mechanistically, the accumulation of oleic acid in mammary epithelial cells enhances their resistance to ferroptosis, thereby supporting milk production. These findings highlight the potential of L. amylovorus and its BEVs as therapeutic tools to counteract oxidative stress-induced lactation decline.

Amino acid utilization allows intestinal dominance of Lactobacillus amylovorus

Article 27 July 2022

Gut microbiota regulates acute myeloid leukaemia via alteration of intestinal barrier function mediated by butyrate

Article Open access 09 May 2022

Probiotic Lactobacillus fermentum strain JDFM216 improves cognitive behavior and modulates immune response with gut microbiota

Article Open access 10 December 2020

Introduction

Oxidative stress plays a crucial role in maternal reproductive disorders and is strongly linked to negative pregnancy outcomes, including preeclampsia, gestational hypertension, gestational diabetes, preterm birth, and fetal growth restriction¹. The lactation period, a critical phase in the maternal reproductive cycle, relies on the functional integrity of mammary epithelial cells to ensure normal milk production. Excessive oxidative stress can damage these cells, leading to reduced milk yield and altered milk composition, which negatively impacts the health of newborns.

Research has shown that host oxidative stress is closely linked to environmental factors², dietary composition³, and different physiological stages, particularly pregnancy and lactation. In recent years, gut microbiota have also been identified as key players in the modulation of oxidative stress⁴. While substantial evidence indicates that gut microbiota primarily reside in the intestines and regulate local intestinal functions⁵, the relationship between gut microbiota, their metabolites, and mammary gland function remains largely unexplored. Some studies suggest that a small number of gut microbes can translocate to the mammary glands via the bloodstream, influencing mammary function and being transmitted to the next generation through milk⁶. However, such remote translocation is relatively rare due to the gut barrier. More commonly, gut microbiota affect distant organs by secreting metabolites that influence the function of the brain, liver, and heart^7,8,9.

In addition to metabolites, bacterial extracellular vesicles (BEVs) serve as an important mechanism for long-distance communication by gut microbiota, garnering increasing research attention¹⁰. In mouse models, BEVs have been found to accumulate significantly in tissues such as joints and mammary glands, suggesting their regulatory roles in these organs. For instance, studies have demonstrated that Fusobacterium nucleatum in the gut can deliver BEVs to the joints, triggering local inflammation and thus contributing to exacerbating rheumatoid arthritis¹¹.

In this study, we aimed to investigate whether the levels of Lactobacillus amylovorus (L. amylovorus) in the maternal gut are associated with oxidative stress and ferroptosis in the mammary glands. Further investigations revealed that L. amylovorus alleviates oxidative stress-induced ferroptosis by transferring oleic acid, a monounsaturated fatty acid, into mammary epithelial cells via BEVs. This enrichment enhances the synthesis capacity of milk lipids in the maternal mammary glands and improves the growth performance of offspring under oxidative stress conditions. These findings highlight the protective role of L. amylovorus in mitigating oxidative stress-induced ferroptosis and suggest its potential as a clinical therapeutic approach to improve maternal lactation decline caused by oxidative stress.

Results

Oxidative stress-driven ferroptosis impairs milk fat synthesis in lactating sows

Pigs serve as an excellent comparative model for studying human reproductive physiology because their mammary development, anatomical structure, and lactation mechanisms, particularly in milk fat synthesis and alveolar formation, closely resemble those in humans. In this study, we analyzed 63 recently farrowed sows based on their oxidative stress index (OSI) (Fig. 1a). The top and bottom 10 sows with the highest and lowest OSI were categorized into high oxidative stress (HOS) and low oxidative stress (LOS) groups, respectively, to assess the impact of oxidative stress on lactation (Fig. 1b). Significant differences were observed between the HOS and LOS groups in total oxidative stress (Fig. 1c), total antioxidant status (Fig. 1d), and the OSI (Fig. 1e) (P < 0.001). Blood analyses showed that sows in the HOS group had significantly lower levels of glutathione (GSH) (Fig. 1f) and malondialdehyde (MDA) (Fig. 1g) (P < 0.01).

**Fig. 1: Oxidative stress induces ferroptosis in sow mammary glands and inhibits milk fat synthesis.**

Correlation analyses indicated a strong negative relationship between milk fat synthesis and OSI (P < 0.001), whereas milk protein and milk lactose levels were negatively correlated with OSI, but not significantly (P > 0.05) (Fig. 1h–j). The HOS group showed a significant reduction in milk fat content compared to the LOS group (P < 0.01), while milk protein and milk lactose levels did not differ significantly (P > 0.05) (Fig. 1k–m). To further investigate the effect of oxidative stress on milk fat synthesis, we isolated mammary cells from colostrum samples of sows in both groups. The HOS group exhibited significantly lower expression of key enzymes involved in lipid synthesis (ACACA, FASN, DGAT1), regulatory factors (SREBP1, PPARγ), and fatty acid transport proteins (CD36, FATP4, FABP4) compared to the LOS group (Fig. 1p, r) (P < 0.05). Similar reductions were observed at the gene expression level (Fig. 1q, s) (P < 0.05). These findings indicate that oxidative stress impairs lactational performance by decreasing the efficiency of milk fat synthesis and transport.

Oxidative stress is commonly associated with intracellular accumulation of ROS and disruption of iron homeostasis, both of which are key triggers of ferroptosis. Analysis of ferroptosis-related indicators revealed significant differences in GSH and MDA levels (Fig. 1n, o) (P < 0.01). The HOS group showed increased expression of ferroptosis marker genes and proteins, such as PTGS2, while the expression of genes and proteins involved in maintaining redox balance, including GPX4 and FTH, decreased (Fig. 1t, u) (P < 0.05). These results confirm that ferroptosis is significantly elevated in sows experiencing high oxidative stress. To further explore the connection between oxidative stress and ferroptosis, we developed an in vitro oxidative stress model using sow mammary epithelial cells. Treatment with ferroptosis inhibitor Fer-1 significantly reduced cell death caused by oxidative stress (Supplementary Fig. 1a–c) (P < 0.01). Additionally, intracellular reactive oxygen species (ROS) levels decreased markedly (Supplementary Fig. 1d, e), lipid peroxidation (LPO) was reversed (Supplementary Fig. 1f, g), and mitochondrial structure was preserved (Supplementary Fig. 1h) (P < 0.01). MDA concentrations decreased (Supplementary Fig. 1i), while GSH levels increased significantly (Supplementary Fig. 1j) (P < 0.01). The expression of ferroptosis marker genes (PTGS2) was downregulated, and genes (Ferritin and GPX4) involved in ferroptosis suppression were upregulated (Supplementary Fig. 1k) (P < 0.05). These results demonstrate that Fer-1 effectively alleviates ferroptosis induced by oxidative stress. Moreover, cells treated with Fer-1 showed larger intracellular lipid droplets (Supplementary Fig. 1l, m), higher extracellular and intracellular triglyceride levels (Supplementary Fig. 1n, o), and increased expression of genes involved in milk fat synthesis and transport (Supplementary Fig. 1p, q) (P < 0.05). These results indicate oxidative stress impairs mammary lipid synthesis and transport by increasing ferroptosis.

Distinct gut microbiota profiles correlate with ferroptosis and milk fat synthesis in sows

To examine whether gut microbiota affect oxidative stress, ferroptosis, and milk fat synthesis and transport, we compared the fecal microbiota of sows in the HOS group with those in the LOS group. Principal Coordinate Analysis (PCoA) revealed distinct β-diversity patterns between the two groups (Fig. 2a), and notable differences in α-diversity (Shannon, Chao1, Simpson Indices) were also observed (Fig. 2b–d) (P < 0.05). At the phylum level, the Firmicutes-to-Bacteroidetes ratio was significantly higher in the LOS group, suggesting that these sows were generally healthier than those in the HOS group (Fig. 2e).

**Fig. 2: Maternal oxidative stress-induced ferroptosis and is closely associated with gut microbiota composition.**

Further genus-level analysis showed a markedly higher abundance of Lactobacillus in the LOS group (Fig. 2f, g) (P < 0.05), a finding corroborated by a differential abundance heatmap (Fig. 2h). Functional profiling using Tax4Fun indicated that the LOS group exhibited stronger associations with lipid metabolism (Fig. 2i). Moreover, Spearman correlation analysis of the relative abundances of different genera against key indicators in colostrum, such as milk fat, lactose, protein content, and ferroptosis-related genes expression, demonstrated that Lactobacillus was positively correlated with enhanced milk fat synthesis and lower ferroptosis levels (Fig. 2j) (P < 0.05).

To identify the predominant Lactobacillus species in these sows, we conducted quantitative assessments of the seven most abundant Lactobacillus species in the pig gut¹². Among them, Lactobacillus reuteri, Lactobacillus plantarum, Lactobacillus amylovorus, Lactobacillus johnsonii, and Lactobacillus salivarius were significantly enriched in the LOS group (Fig. 2k–q) (P < 0.05), suggesting that these species hold strong potential for reducing oxidative stress and ferroptosis while promoting milk fat synthesis.

Gut microbiota transplantation from sows modulates ferroptosis and milk fat synthesis in mice

To determine whether gut microbiota regulate milk fat synthesis by influencing oxidative stress and ferroptosis, we conducted an in vivo experiment by transplanting fecal microbiota from sows in the HOS and LOS groups into mice, respectively (Fig. 3a). On day 14 post-transplantation, we analyzed the cecal microbiota of the dams. The LOS group formed a distinct cluster in β-diversity analysis (Supplementary Fig. 2a) and differed in α-diversity (Supplementary Fig. 2b–d). At the phylum level, the LOS group showed a higher Firmicutes-to-Bacteroidetes ratio (Supplementary Fig. 2e). At the genus level, Lactobacillus abundance was notably higher in the LOS group (Supplementary Fig. 2f, g) (P < 0.001), mirroring the composition observed in the transplanted sow microbiota, with Lactobacillus amylovorus as the most prevalent species. Similarly, among seven common Lactobacillus species, Lactobacillus amylovorus, Lactobacillus plantarum, Lactobacillus johnsonii, and Lactobacillus reuteri were significantly more abundant in the LOS group (Supplementary Fig. 2j–p) (P < 0.01). And the Spearman analysis revealed that, similar to the results from the sow experiment, the Lactobacillus had a strong association with the iron-related death genes, such as PTGS2, GPX4, and Ferritin (Supplementary Fig. 2h) (P < 0.05). These results indicate that the microbiota transplantation from sows to mice was successful.

**Fig. 3: Impact of maternal gut microbiota transplantation on ferroptosis in mammary glands and lactation performance.**

The experiment lasted from day 0 of lactation to day 14, the peak of lactation, in the mice. We standardized the litter size to 13 pups per dam at birth and adjusted pup weights to ensure uniform growth. Under these conditions, pups in the HOS group had significantly lower body weights than those in the LOS group (Fig. 3b), and differences in body size between the two groups were readily visible (Fig. 3c). To minimize individual and environmental variability, we conducted cross-fostering starting on lactation day 0 and day 8. Pups cross-fostered on day 0 showed similar growth trends to those in the non-cross-fostered group, indicating minimal impact from external factors (Fig. 3d). This suggests that differences in offspring growth were mainly shaped by the maternal microbiota introduced by HOS or LOS sows. Moreover, pups cross-fostered on day 8 revealed that dams carrying LOS-derived microbiota produced more milk, supporting faster growth in their offspring (Fig. 3e).

We next examined how microbial transfer affected ferroptosis in the dams by analyzing their mammary gland tissues. Transmission electron microscopy showed extensive mitochondrial damage in the HOS group (Fig. 3f). These dams also had reduced GSH content (Fig. 3g) and elevated MDA levels (Fig. 3h) (P < 0.001), all indicative of enhanced ferroptosis. Ferroptosis marker genes and proteins (e.g., PTGS2) were upregulated, while genes (GPX4 and Ferritin) involved in maintaining redox balance were downregulated (Fig. 3i, j) in the mammary gland (P < 0.05), confirming that HOS-derived microbiota significantly increased ferroptosis in the dams.

To compare mammary gland development in the HOS and LOS groups, we performed whole-mount and H&E staining. LOS dams displayed more densely packed secretory lobules (Fig. 3k) and larger alveolus areas (Fig. 3l, p) (P < 0.001), both favorable for milk secretion and storage. Pups from LOS dams also showed larger gastric milk curds (Fig. 3m), suggesting greater milk intake and production. Oil Red O staining (Fig. 3n, q) and measurement of the milk fat percentage (Fig. 3o, r) further confirmed that LOS dams produced milk with higher lipid content (P < 0.01), promoting faster pup growth. At the molecular level, key enzymes and regulatory factors for lipid synthesis (ACACA, FASN, DGAT1, SREBP1, PPARγ) as well as fatty acid transporters (CD36, FATP4, FABP4) were significantly upregulated in the mammary gland of LOS group (Fig. 3s–v) (P < 0.05). Intriguingly, Tax4Fun analysis linked the LOS microbiota to stronger lipid metabolism functions (Supplementary Fig. 2i). These findings indicate that compared with HOS-derived microbiota, LOS-derived microbiota protected the dams from ferroptosis and promoted milk fat synthesis and transport in the mammary gland.

LOS-derived microbiota alleviates ferroptosis and maintains mammary lactation under oxidative stress

To determine whether LOS-derived microbiota protect mammary glands from ferroptosis and impaired lactation under oxidative stress, we established an oxidative stress model in mice. We then treated the mice either with a ferroptosis inhibitor or with fecal microbiota transplanted from LOS sows (Fig. 4a). To accurately assess lactation, we standardized each litter size to 13 pups and ensured uniform pup weights at birth. Both ferroptosis inhibition and LOS microbiota transplantation markedly reduced oxidative stress, evidenced by steady weight gain (Fig. 4b) and clear physical improvements compared with the oxidative stress (OS) group (Fig. 4c). Consistent with these outcomes, mice treated with ferroptosis inhibitors or LOS microbiota had higher GSH levels (Fig. 4e) and lower MDA levels (Fig. 4f) (P < 0.05). Furthermore, mitochondrial structure was better preserved (Fig. 4d). The expression of ferroptosis marker PTGS2 (both protein and gene) declined, while redox-regulating genes and proteins (GPX4 and FTH) were upregulated (Fig. 4g, h) (P < 0.05), indicating a substantial reduction in ferroptosis.

**Fig. 4: Maternal microbiota transplantation reduces ferroptosis in mammary glands and enhances lactation performance under oxidative stress.**

Mice in the ferroptosis inhibitor and LOS microbiota groups also had healthier mammary glands with a higher abundance of milk fat. These improvements were reflected by more compact alveolar structures (Fig. 4i), larger alveolus areas (Fig. 4j, n) (P < 0.05), and greater milk production (Fig. 4k). Oil Red O staining (Fig. 4l, o) and milk fat percentage measurements (Fig. 4m, p) confirmed higher fat content in these groups (P < 0.05). Additionally, enzymes involved in lipid synthesis (ACACA, FASN, DGAT1), key transcription factors (SREBP1, PPARγ), and fatty acid transporters (CD36, FATP4, FABP4) were significantly upregulated in the ferroptosis inhibitor and LOS microbiota groups (Fig. 4q–t) (P < 0.05). Taken together, these findings demonstrate that supplementing mice with LOS-derived microbiota alleviates ferroptosis and preserves milk fat synthesis in the mammary gland under oxidative stress.

L. amylovorus emerges as the key LOS-derived Lactobacillus for counteracting oxidative stress-induced ferroptosis and supporting milk fat synthesis

To identify which Lactobacillus species in the LOS group efficiently alleviates oxidative stress-induced ferroptosis and improves milk fat synthesis, we established an in vitro oxidative stress model in HC11 mammary epithelial cells using H₂O₂. We then prepared cell-free supernatants from five Lactobacillus species (Lactobacillus amylovorus, Lactobacillus johnsonii, Lactobacillus reuteri, Lactobacillus plantarum, and Lactobacillus salivarius) enriched in the LOS group and screened their antioxidant and anti-ferroptosis effects (Supplementary Fig. 3a). Among these five species, L. amylovorus demonstrated the strongest ability to reduce excessive free radicals and lipid peroxides induced by oxidative stress (Supplementary Fig. 3b–e) (P < 0.01). Furthermore, L. amylovorus supernatant significantly increased the lipid droplet content secreted by mammary epithelial cells compared with the other four species, as evidenced by larger lipid droplet areas after Oil Red O staining (Supplementary Fig. 3f, g) and higher extracellular and intracellular triglyceride levels (Supplementary Fig. 3h, i) (P < 0.05). These findings suggest that L. amylovorus may be a key gut microbe for mitigating ferroptosis induced by oxidative stress and enhancing lactation capacity.

L. amylovorus extracellular vesicles as primary mediators of protection against oxidative stress-induced ferroptosis

To determine whether Lactobacillus amylovorus protects mammary glands from oxidative stress, we conducted an in vivo mouse experiment. Previous research has shown that gut microbes can influence distant tissues and organs by releasing bioactive factors, including bacterial extracellular vesicles. We hypothesized that L. amylovorus exerts its effects primarily through metabolites and/or vesicle secretion. In this study, the extracellular vesicle inhibitor GW4869 was used to assess whether inhibiting L. amylovorus vesicle release diminishes its protective effects on ferroptosis and milk fat synthesis (Fig. 5a and Supplementary Fig. 4). The results from in vivo treatment of mice with the GW4869 inhibitor alone show that GW4869 has no significant effect on offspring growth performance (Supplementary Fig. 5a–c), maternal mammary ferroptosis levels (Supplementary Fig. 5d–g), or lactation performance (Supplementary Fig. 5h–q) (P > 0.05). Similarly, in vitro treatment of HC11 cells with GW4869 inhibitor also reveals no significant impact on ferroptosis (Supplementary Fig. 6a–d) or milk fat synthesis (Supplementary Fig. 6e–h) in mammary epithelial cells (P > 0.05). These findings collectively indicate that host-derived BEVs do not contribute to the observed phenotypes.

**Fig. 5: *Lactobacillus amylovorus* reduces oxidative stress-induced ferroptosis in mammary glands and enhances lactation performance via metabolites and BEVs.**

Quantification of L. amylovorus in the maternal gut confirmed successful colonization of the administered strain (Supplementary Fig. 7). In vivo experiments further revealed that L. amylovorus significantly mitigated oxidative stress-induced damage in pups, and this benefit was greatly reduced upon vesicle inhibition (Fig. 5b, c). As illustrated in Fig. 5d, mice gavaged with L. amylovorus exhibited more intact mitochondrial structures in their mammary glands, whereas mice treated with L. amylovorus with GW4869 continued to show severe mitochondrial damage. This trend was consistent with changes in GSH and MDA in the blood (Fig. 5e, f) (P < 0.05). Ferroptosis-related assessments (Fig. 5g, h) indicated that mice receiving L. amylovorus had substantially lower ferroptosis compared with the oxidative stress group (P < 0.05). While partial protection remained in the vesicle-inhibited group, it was noticeably weaker than the normal L. amylovorus treatment.

Similarly, mice supplemented with L. amylovorus had tightly arranged alveoli (Fig. 5i), larger alveolus areas (Fig. 5j, n), increased pup milk intake (Fig. 5k), and more lipid droplet secretion (Fig. 5l, o) (P < 0.05). Centrifugation of milk samples revealed a higher milk fat percentage in the L. amylovorus group (Fig. 5m, p) (P < 0.05). Enzymes and genes involved in milk fat synthesis and transport were significantly upregulated (Fig. 5q–t) (P < 0.05), demonstrating L. amylovorus’s capacity to alleviate oxidative stress-induced damage to milk fat synthesis. While minor improvement persisted with bacterial extracellular vesicle inhibition, it was markedly reduced compared with complete L. amylovorus treatment. These results, consistent with previous in vitro findings, suggest that L. amylovorus combats oxidative stress-induced ferroptosis and maintains milk fat synthesis largely through bacterial extracellular vesicle secretion, while additional metabolites may also play a role.

In order to further validate the function of L. amylovorus BEVs, we isolated and purified them by ultracentrifugation (Supplementary Fig. 8a). Transmission electron microscopy revealed that the vesicles had a disc-like shape with an average diameter of about 100 nm (Supplementary Fig. 8b), and nanoparticle tracking analysis (NTA) confirmed a predominant size of approximately 111.6 nm (Supplementary Fig. 8c). After labeling these vesicles with Dil dye, we co-cultured them with HC11 mouse mammary epithelial cells. Fluorescence microscopy showed that they entered the cells and accumulated around the nucleus (Supplementary Fig. 8d). In an H₂O₂-induced oxidative stress model in HC11 cells, L. amylovorus BEVs significantly reduced ROS (Supplementary Fig. 9a, b) and LPO levels (Supplementary Fig. 9c, d) (P < 0.05), providing protection comparable to that of ferroptosis inhibitors and L. amylovorus cell-free supernatant. L. amylovorus BEVs also increased lipid droplet formation (Supplementary Fig. 9e, f) and both extracellular and intracellular triglyceride levels (Supplementary Fig. 9g, h) (P < 0.05). In contrast, L. amylovorus pretreated with GW4869 (a vesicle secretion inhibitor) showed a markedly weaker protective effect, consistent with our in vivo findings. Short-chain fatty acids (SCFAs) are another major metabolite of L. amylovorus and have been reported to mitigate ferroptosis. We hypothesized that SCFAs may partly alleviate oxidative stress in mammary epithelial cells when BEV secretion is inhibited. Treatment with SCFA mixture (acetic, propionic, and butyric acids in a 65:20:15 ratio) decreased ROS (Supplementary Fig. 10a, b) and LPO (Supplementary Fig. 10c, d), while enhancing triglyceride levels (Supplementary Fig. 10e, f) and lipid droplet formation (Supplementary Fig. 10g, h). However, no statistically significant differences were observed (P > 0.05). Although SCFAs exhibited some protective effects, they provided less protection compared to L. amylovorus BEVs. Collectively, these in vitro results support the conclusion that L. amylovorus mitigates oxidative stress-induced ferroptosis primarily through BEV secretion, with a smaller contribution from SCFAs alone.

L. amylovorus BEVs mitigate oxidative stress-induced ferroptosis and preserve mammary function in mice

To evaluate the effectiveness of L. amylovorus BEVs in vivo, we directly treated BEVs into mice under oxidative stress to assess their impact on ferroptosis and mammary gland function (Fig. 6a). We first confirmed whether these vesicles could reach the mammary gland by labeling them with Dil dye and administering them via intraperitoneal injection. In vivo imaging at 1, 6, and 12 h revealed that fluorescence initially remained near the injection site but later appeared in both the abdominal cavity and mammary glands (Fig. 6b). Dissection and organ imaging at 12 h further confirmed fluorescence signals in the mammary gland (Fig. 6c), indicating that L. amylovorus BEVs can circulate to the mammary gland.

**Fig. 6: *Lactobacillus amylovorus* alleviates oxidative stress-induced ferroptosis in mouse mammary glands and enhances lactation performance via bacterial extracellular vesicles.**

Under oxidative stress, mice receiving L. amylovorus BEVs showed marked improvements in mammary function, as evidenced by higher pup weights (Fig. 6e) and larger overall size (Fig. 6d). These mice also displayed intact mitochondrial structures in the mammary gland (Fig. 6f), increased levels of GSH (Fig. 6g), and reduced levels of MDA (Fig. 6h) (P < 0.01). Surprisingly, ferroptosis-related proteins (Fig. 6i) and genes (Fig. 6j) did not change significantly (P > 0.05), suggesting that BEVs reduce ferroptosis through mechanisms beyond direct gene or protein regulation. In addition, BEV-treated mice exhibited tighter alveolar structures (Fig. 6k), larger alveolus areas (Fig. 6l, p), increased milk content in pups (Fig. 6m), elevated lipid droplet formation (Fig. 6n, q), and higher milk fat percentage (Fig. 6o, r) (P < 0.05). Genes and proteins linked to milk fat synthesis and transport were also significantly upregulated (Fig. 6s–v) (P < 0.05). These findings illustrate that L. amylovorus BEVs substantially mitigate oxidative stress-induced ferroptosis, thereby preserving milk fat synthesis and transport.

Lipid components of L. amylovorus BEVs govern ferroptosis protection and milk fat preservation

Having confirmed that L. amylovorus BEVs alleviate oxidative stress-induced ferroptosis and enhance mammary function, we next examined their underlying mechanism. Since BEVs are composed of proteins, lipids, and nucleic acids, we investigated which biomolecules drive their protective effects. First, we treated the BEVs with DNase I, RNase A, or Proteinase K, then examined their ability to mitigate ferroptosis in oxidative stress-induced HC11 cells (Supplementary Fig. 11a). Removing nucleic acids or proteins did not diminish L. amylovorus BEV-mediated protection, as indicated by unchanged ROS (Supplementary Fig. 11b, c), LPO (Supplementary Fig. 11d, e), lipid droplet formation (Supplementary Fig. 11f, g), and extracellular/intracellular triglyceride levels (Supplementary Fig. 11h, i) (P > 0.05). These results suggest that nucleic acids and proteins do not play the primary protective role, prompting us to focus on BEV lipids. We then extracted the lipid fraction from the BEVs (Fig. 7a) and separated it into hydrophilic and lipophilic phases. Only the lipophilic phase effectively protected HC11 cells from oxidative stress-induced ferroptosis, promoting greater lipid droplet secretion (Fig. 7b, c), increasing extracellular and intracellular triglycerides (Fig. 7d, e), and significantly reducing ROS (Fig. 7f, g) and LPO (Fig. 7h, i) (P < 0.05). These effects confirm that L. amylovorus BEVs primarily function through their lipid components under oxidative stress.

**Fig. 7: *Lactobacillus amylovorus* BEVs mitigate oxidative stress-induced ferroptosis and enhance milk fat synthesis through lipophilic substances.**

Lipidomics analysis demonstrated that monounsaturated fatty acids (MUFAs) accounted for approximately 49.74% of the total fatty acids in the BEVs, with oleic acid representing 50.55% of the MUFAs (Fig. 8a). Compared to the other four Lactobacillus species, Lactobacillus amylovorus exhibited the highest proportion of MUFAs (Fig. 8b–e). Monounsaturated fatty acids can replace polyunsaturated fatty acids (PUFAs) in cell membranes, reducing membrane fluidity and oxidation sensitivity, thereby decreasing the risk of ferroptosis. To validate this, we added oleic acid, the most abundant MUFA in the BEVs, to mammary epithelial cells under oxidative stress and ferroptosis conditions. As expected, oleic acid treatment substantially decreased ROS (Fig. 8f, g) and LPO (Fig. 8h, i), while boosting extracellular and intracellular triglyceride content (Fig. 8j, k) and lipid droplet formation (Fig. 8l, m) (P < 0.05). These observations indicate that oleic acid alleviates oxidative stress-induced ferroptosis and maintains milk fat synthesis without altering ferroptosis-related gene or protein expression.

**Fig. 8: *Lactobacillus amylovorus* BEV-derived oleic acid mitigates oxidative stress-induced ferroptosis and enhances milk fat synthesis in mammary epithelial cells.**

Discussion

While numerous studies have established a close relationship between gut microbiota and host oxidative stress, the underlying mechanisms remain to be fully elucidated^13,14. In this study, we identified oxidative stress-induced ferroptosis as a central factor leading to decreased milk lipid synthesis in the mammary glands. We observed a negative correlation between the levels of L. amylovorus in the maternal gut and both oxidative stress and ferroptosis in the mammary glands. Importantly, our findings demonstrate that L. amylovorus mitigates oxidative stress-induced ferroptosis by delivering oleic acid, a monounsaturated fatty acid, to mammary epithelial cells via BEVs. This enrichment of oleic acid enhances the cells’ resistance to ferroptosis, boosts the synthesis capacity of milk lipids in the mammary glands, and subsequently improves the growth performance of offspring under oxidative stress conditions.

L. amylovorus is an acidophilic lactic acid bacterium predominantly found in the intestines of humans and various animals, particularly enriched in the colon¹⁵. It efficiently ferments starch, produces beneficial metabolites, and maintains gut microbiota balance, thereby promoting nutrient absorption and immune function in the host¹⁶. L. amylovorus has shown potential in regulating several gut-related diseases, such as inflammatory bowel disease (IBD) and irritable bowel syndrome (IBS)^17,18. Additionally, BEVs from Lactobacillus species have been implicated in regulating multiple physiological functions. For instance, Lactobacillus johnsonii and Lactobacillus murinus BEVs activate M2 polarization in intestinal epithelial cells, thereby enhancing the barrier function of the gut^19,20. Similarly, BEVs from Lactobacillus plantarum influence skin health by affecting M2 macrophage polarization²¹, and BEVs from Lactobacillus druckerii reduce hypertrophic scar formation by inhibiting the expression of Collagen I/III and α-SMA²². Furthermore, Lactobacillus rhamnosus BEVs interact with bone physiology by promoting angiogenesis and osteogenesis, thereby reducing osteonecrosis of the femoral head²³. These studies collectively suggest that BEVs derived from L. amylovorus may possess significant biological value.

In this research, we observed that BEVs secreted by L. amylovorus accumulate extensively in the mammary glands. We also demonstrated that these BEVs are capable of transporting substantial amounts of oleic acid into mammary epithelial cells. Although the specific mechanisms facilitating the accumulation of BEVs in the mammary glands remain unclear, our evidence strongly supports the protective role of L. amylovorus BEVs against oxidative stress in the mammary glands. Another key finding of our study is that oxidative stress-induced decline in mammary gland lactation is primarily due to ferroptosis. Ferroptosis is a form of programmed cell death characterized by iron-dependent lipid peroxidation, leading to the accumulation of lipid hydroperoxides containing highly unsaturated fatty acids such as arachidonic acid and linoleic acid^24,25. A hallmark of ferroptosis includes significant mitochondrial shrinkage, increased membrane density, and the reduction or disappearance of cristae structures²⁶. In this study, we observed substantial mitochondrial abnormalities and changes in key genes that exacerbate ferroptosis in mammary epithelial cells under oxidative stress. Previous research has shown that dietary monounsaturated fatty acids (MUFAs) can increase the ratio of MUFA to polyunsaturated fatty acids (PUFAs) in membrane lipids, thereby protecting cells from oxidative lipid damage^27,28. This aligns with our findings that BEVs enriched with MUFAs from L. amylovorus can alleviate oxidative stress-induced ferroptosis.

In conclusion, this study reveals that ferroptosis is a critical mechanism by which oxidative stress impairs lactation performance in the mammary glands. Furthermore, L. amylovorus and its BEVs offer a promising therapeutic strategy to counteract oxidative stress-induced lactation decline. These insights not only enhance our understanding of the gut-mammary axis in maternal health but also pave the way for developing novel interventions to support lactation under stress conditions.

Methods

Animal experiments with sows

A total of 63 Landrace × Yorkshire sows (first day of lactation) with comparable backfat depths and parity numbers were selected. All sows underwent routine health checks and were found to be in good physical condition. They were housed under standard swine production conditions and had ad libitum access to feed and water. The diet was formulated according to NRC (2012) guidelines (see Supplementary Table 1). Blood samples were collected to measure the oxidative stress index (OSI), calculated as total oxidant status (TOS) divided by total antioxidant status (TAS) (Jiancheng Bioengineering Institute, Jiangsu, China). The 10 sows with the highest OSI were designated as the high oxidative stress (HOS) group, and the 10 with the lowest OSI were designated as the low oxidative stress (LOS) group. Subsequent procedures, including fecal sampling (for microbiota preservation) and colostrum collection (for analyzing milk composition and isolating mammary epithelial cells), were conducted according to these groupings.

Animal experiments with mice

Twenty specific-pathogen-free (SPF) female ICR mice (6 weeks old) were obtained from Guangdong Sijia Jingda Biotechnology Co., Ltd. (Guangdong, China). They were randomly divided into two groups with equal distribution and housed under SPF conditions (12-h light/dark cycle, 24 °C, 55–60% humidity) with free access to food and water. In this fecal microbiota transplantation study, mice were pair-mated at 8 weeks of age, with vaginal plug confirmation establishing gestational timelines. Ten days prior to parturition, dams underwent gut microbiota depletion via broad-spectrum antibiotics (1 g/L each of neomycin, ampicillin, gentamicin, metronidazole, and 0.5 g/L vancomycin; Sigma-Aldrich, St. Louis, MO, USA) administered in drinking water. Postpartum litters were standardized to 13 pups per dam and orally gavaged daily with 200 μL of fecal suspensions (10⁹ CFU/mL) derived from either high oxidative stress sows (HOS) or low oxidative stress sows (LOS) until lactation day 14 (peak milk production). At termination, dams were euthanized via intraperitoneal sodium pentobarbital overdose (100 mg/kg), followed by cervical dislocation. Mammary gland tissue and cecal contents were aseptically collected, flash-frozen in liquid nitrogen, and stored at −80 °C for downstream omics analyses.

Four-week-old SPF female ICR mice were intraperitoneally injected daily with D-galactose (150 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) for 42 consecutive days to induce oxidative stress, while control groups received equivalent volumes of sterile saline²⁹. The OS + Fer-1 group was co-administered D-galactose and 10 mg/kg Ferrostatin-1 (Fer-1; Selleck Chemicals, Houston, TX, USA) daily³⁰. At 10 weeks of age, mice were pair-mated, with vaginal plug detection used to establish parturition timelines. Ten days prepartum, dams underwent gut microbiota depletion via a broad-spectrum antibiotic cocktail in drinking water. Postpartum litters were standardized to 13 pups per dam, and interventions were administered from lactation day 0 to 14 as follows: LOS sow-derived fecal microbiota (oral gavage), GW4869 inhibitor (5 μM; oral gavage; Selleck Chemicals, Houston, TX, USA), Lactobacillus amylovorus (1 × 10⁹ CFUs; oral gavage), GW4869-pretreated L. amylovorus³¹ (1 × 10⁹ CFUs; oral gavage), or L. amylovorus-derived BEVs (2.6 × 10¹⁰ particles/kg body weight in 20 μL PBS; intraperitoneal injection). Control dams received sterile saline via matched routes. All groups included 8 biological replicates (n = 8 dams/group). On lactation day 14, dams were euthanized via an overdose of sodium pentobarbital (100 mg/kg), followed by cervical dislocation, and mammary tissue was aseptically collected for downstream analyses. 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).

Cell culture experimental design

Mouse mammary epithelial HC11 cells were maintained in DMEM/F12 medium supplemented with 10% FBS, 5 μg/mL IGF-1, 10 ng/mL EGF, 5 μg/mL ITS, and 1% penicillin-streptomycin (MCE, Monmouth Junction, NJ, USA) at 37 °C and 5% CO₂. To induce differentiation, the cells were first cultured in a medium devoid of epidermal growth factor (EGF) for 24 h. Following this, they were transferred to DIP medium, which contained 1 μM dexamethasone, 5 μg/mL insulin, and 5 μg/mL prolactin. When cells reached ~70% confluence, they were exposed to 200 μM hydrogen peroxide (H₂O₂) for 12 h to induce oxidative stress³². In an additional group (OS + Fer-1), the cells were pretreated with 10 μM Fer-1 (Selleck Chemicals, Houston, TX, USA) for 6 h prior to H₂O₂ exposure³³. Following stress induction, cells were treated for 24 h with one of the following: cell-free supernatants from various Lactobacillus strains (3%), supernatants of L. amylovorus pretreated with GW4869 (3%), L. amylovorus-derived BEVs (1.3 × 10¹⁰ particles/mL), different BEV fractions, short-chain fatty acids (SCFAs; acetate: propionate: butyrate = 65%:20%:15%, 2.5 mM), or oleic acid (50 μM). Intracellular reactive oxygen species (ROS), lipid peroxidation (LPO), and lipid accumulation were then quantified.

Oxidative stress index (OSI) measurement

The TOS and TAS in the blood of 63 sows on the first day of lactation were measured (Jiancheng Bioengineering Institute, Jiangsu, China), and these were then used to calculate OSI (OSI = TOS/TAS). The sows were ranked based on their OSI values to categorize their oxidative stress status.

Measurement of glutathione (GSH), malondialdehyde (MDA), and triglycerides (TG)

Blood or serum samples from sows and mice were used to measure GSH, MDA, and triglyceride (TG) levels using commercial kits (Jiancheng Bioengineering Institute, Jiangsu, China). Cultured cells were also analyzed similarly after treatment; cells were washed, lysed, and processed for GSH, MDA, or TG measurement.

Milk composition analysis in sows

From each of the 63 sows, 50 mL of colostrum was obtained on lactation day 1 and immediately frozen at −80 °C. Fat, protein, lactose, and solids-not-fat were measured using a milk composition analyzer (Ekomilk Bond, Espoo, Finland).

Isolation of mammary epithelial cells from sow colostrum

Colostrum was mixed with PBS in a 1:1 ratio and subjected to centrifugation at 870 × g for 20 min at 4 °C. The resulting pellet was resuspended in 5 mL PBS and centrifuged at 490 × g for 5 min. This washing procedure was repeated twice to isolate mammary epithelial cells.

Preservation of fecal microbiota from lactating sows

Fecal samples (100 g) were taken from each of the 63 sows on lactation day 1, mixed with 300 mL sterile saline, homogenized, and filtered. The filtrate was centrifuged at 600 × g for 15 min, and pellets were resuspended in sterile saline with 10% glycerol. These suspensions were frozen at −80 °C for future experiments.

16S rRNA sequencing and data analysis

Microbial composition was analyzed via 16S rRNA next-generation sequencing (Biomarker Biotechnology, Beijing, China). Total DNA was extracted from sow fecal samples (A260/A280 ratio 1.8–2.0). Primers F (5′-ACTCCTACGGGAGGCAGCA-3′) and R (5′-GGACTACHVGGGTWTCTAAT-3′) were used to amplify the V3 + V4 regions. After purification, qPCR products were sequenced on an Illumina NovaSeq 6000 (San Diego, CA, USA).

Quantification of bacterial load

Total DNA was extracted from fecal samples using a commercial fecal DNA extraction kit (Tiangen Biotech, Beijing, China). DNA concentration was measured with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA) and normalized using nuclease-free water prior to RT-qPCR analysis. Bacterial load was quantified by comparing Ct values to a standard curve generated from known concentrations of bacterial DNA. Primer sequences specific to the target bacterial species are provided in Supplementary Table 3.

Bacterial plate counting of fecal samples

Fresh fecal samples were collected from mice into sterile tubes, homogenized in sterile saline, and serially diluted up to 10⁷-fold. Aliquots of each dilution were spread onto autoclaved, non-selective plate count agar (HKM, Guangzhou, China) using disposable sterile spreaders. Plates were incubated at 37 °C for 48 h in a constant-temperature incubator, and bacterial colonies were then counted to determine viable cell numbers.

Mitochondrial morphology analysis

On day 14 of lactation, mice were euthanized, and mammary gland tissues were collected, cut into 1 mm³ pieces, and fixed in 2.5% glutaraldehyde at 4 °C overnight. After dehydration, infiltration, and embedding, ultrathin (70 nm) sections were examined under a Talos L120C transmission electron microscope (TEM; Thermo Fisher Scientific, Eindhoven, Netherlands) to observe mitochondrial size, shape, and density. Mammary epithelial cells underwent a similar fixation and preparation process.

Whole-mount staining

On lactation day 14, mammary gland tissues from euthanized mice were washed and fixed in Carnoy’s fixative for 12 h. Tissues were stained overnight with Carmine Alum (Sigma-Aldrich, St. Louis, MO, USA), then dehydrated in ethanol (75%, 95%, and 100%) and cleared in xylene³⁴.

Mammary tissue H&E staining

Also on lactation day 14, mammary gland tissues were harvested, washed, and fixed in 4% paraformaldehyde before paraffin embedding and sectioning. After deparaffinization in xylene and rehydration, sections were stained with hematoxylin and eosin to visualize alveolar structures, then dehydrated and mounted.

Oil Red O staining

Mammary gland tissues were sectioned at 8 μm thickness and embedded in OCT (Sangon Biotech, Shanghai, China), then stained with filtered Oil Red O (Sangon Biotech, Shanghai, China). The nuclei were counterstained with hematoxylin (Sangon Biotech, Shanghai, China). For cell cultures, HC11 cells were fixed using 4% paraformaldehyde, stained with Oil Red O, washed, and counterstained with hematoxylin to visualize lipid droplets.

Analysis of murine milk fat percentage

Mice were separated from pups overnight before milk collection on day 14 of lactation. Oxytocin (4 IU; MCE, Monmouth Junction, NJ, USA) was administered intraperitoneally, and milk was collected 30 min later. Samples were diluted 1:2 with PBS and centrifuged at 3000 rpm for 20 min at 4 °C. The volume of the fat layer that separated after centrifugation was measured, and the fat percentage was calculated by comparing the volume of the fat layer to the total milk volume in the capillary tube³⁵.

Purification and characterization of bacterial extracellular vesicles (BEVs)

L. amylovorus was cultured in MRS medium at 37 °C for 72 h. The culture supernatant was sequentially centrifuged (360 × g, 15 min; 3000 × g, 15 min; 10,000 × g, 15 min) to remove debris, then filtered through a 0.22 μm membrane. The filtrate was ultracentrifuged at 180,000 × g for 90 min at 4 °C (Beckman Optima XE-100, Beckman Coulter, Brea, CA, USA)³⁶. The obtained BEV pellet was resuspended in PBS, washed, and then centrifuged at 180,000 × g for 90 min at 4 °C. The resulting crude BEVs were filtered through a 0.22 μm syringe filter and subsequently purified and concentrated using the ExoSure™ Exosome Isolation Kits (GeneCopoeia, Rockville, MD, USA) to isolate bacterial extracellular vesicles ranging from 30 to 150 nm. The concentration of bacterial extracellular vesicles was determined using a protein assay kit (Beyotime, Jiangsu, China). For morphological analysis, BEVs were negatively stained with phosphotungstic acid (PTA), air-dried, and visualized by transmission electron microscopy (TEM). BEV particle size and concentration were measured with a NanoSight NS300 system (Malvern Panalytical, Malvern, UK)³⁷.

Inhibition of BEVs production

Lactobacillus amylovorus was inoculated into autoclaved MRS medium at a concentration of 10⁷ CFUs/100 mL, with or without 10 μM GW4869, an inhibitor of extracellular vesicle secretion. Cultures were incubated at 37 °C with shaking at 120 rpm for 72 h. Following incubation, BEVs were isolated as described above, and their protein content was measured using a BCA protein assay kit (Beyotime, Jiangsu, China). Bacterial viability under GW4869 treatment was assessed by colony-forming unit (CFU) counts on MRS agar plates (HKM, Guangzhou, China) and by measuring optical density at 600 nm (OD600) using a spectrophotometer (Evolution 300, Thermo Fisher Scientific, Wilmington, DE, USA).

Evaluation of BEVs biodistribution in tissues

On lactation day 14, Dil-labeled (MCE, Monmouth Junction, NJ, USA) BEVs were injected intraperitoneally into mice. Fluorescence signals were recorded at 1, 6, and 12 h with an IVIS Lumina III system (PerkinElmer, Waltham, MA, USA). Twelve hours post-treatment, the mice were euthanized, and the major organs (heart, liver, spleen, lungs, kidneys, and mammary glands) were harvested and analyzed to assess BEV distribution.

Western blot

On lactation day 14, mammary glands were collected from euthanized mice, washed, minced, and lysed in RIPA buffer (Beyotime, Jiangsu, China) containing protease and phosphatase inhibitors (MCE, Monmouth Junction, NJ, USA). Homogenates were centrifuged at 12,000 × g for 15 min; supernatants were quantified by a BCA assay (Beyotime, Jiangsu, China). For HC11 cells, they were lysed in RIPA buffer with protease and phosphatase inhibitors. Lysates were centrifuged at 12,000 × g for 15 min at 4 °C, and protein concentrations were measured using the BCA assay (Beyotime, Jiangsu, China). Equal protein amounts (20–40 µg) were separated by 10–12% SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was blocked with 5% non-fat milk in TBST for 1 h and incubated overnight at 4 °C with primary antibodies. After washing, the membrane was incubated with secondary antibody for 1 h, followed by chemiluminescent detection (Tanon 5200 imaging system, Tanon Science & Technology, Shanghai, China). Bands were quantified using ImageJ software. Protein levels were normalized to β-actin. All the antibody information used above can be found in Supplementary Table 4.

RNA isolation and real-time qPCR

On lactation day 14, mammary gland tissues were dissected and washed with PBS. Total RNA was extracted (EZBioscience, Roseville, MN, USA), and for cultured cells, the cell RNA purification kit (EZBioscience, Roseville, MN, USA) was used. RNA concentration and purity were confirmed by Nanodrop (Thermo Fisher Scientific, Wilmington, DE, USA); cDNA was synthesized using 4× EZscript reverse transcription mix II (EZBioscience, Roseville, MN, USA). A 20 μL SYBR qPCR Master Mix reaction (EZBioscience, Roseville, MN, USA) was run on a QuantStudio 3 system (Applied Biosystems, Foster City, CA, USA), with β-actin as the reference gene. The relative expression levels of target genes were determined using the 2^−ΔΔCT method, with primer sequences provided in Supplementary Table 2.

Analysis of cell viability, intracellular ROS, and LPO levels

HC11 cells were plated in 12-well plates and cultured until they reached ~70% confluence. Following treatments, cells were stained with 2 μM Calcein AM and 5 μM PI (MCE, Monmouth Junction, NJ, USA) to assess viability via fluorescence microscopy (NIKON ECLIPSE TiE, Nikon Instruments, Tokyo, Japan). For ROS detection, 10 μM DCFH-DA (MCE, Monmouth Junction, NJ, USA) was applied; for LPO, 2 μM BODIPY 581/591 C11 (MCE, Monmouth Junction, NJ, USA) was added. After staining, cells were visualized under a fluorescence microscope, and the intensity of the relevant fluorescence was quantified with ImageJ.

Preparation of cell-free supernatants of Lactobacillus

Freeze-dried powders of L. amylovorus, L. johnsonii, L. plantarum, L. reuteri, and L. salivarius (BNCC, Beijing, China) were revived and adjusted to 1 × 10⁵ CFU/mL before storage at −80 °C. Each strain (2% inoculum) was cultured in 200 mL MRS at 37 °C under anaerobic conditions for 72 h. Cultures were centrifuged twice at 1,000 × g for 15 min (4 °C). Supernatants were filtered through 0.22 μm membranes and used for cell treatments.

Detection of BEVs uptake by cells

After HC11 cells were treated, 5 μM Dil-labeled BEVs were introduced to the cells and incubated at 37 °C for 6 h. After incubation, the cells were washed, fixed with 4% paraformaldehyde for 30 min, and then washed again with PBS. DAPI mounting medium was then applied, and cells were viewed under an inverted fluorescence microscope. Red Dil fluorescence indicated BEV uptake, while nuclei were stained blue.

Removal of BEV proteins and nucleic acids

BEVs isolated by ultracentrifugation at 4 °C were resuspended (1.3 mg/mL), subjected to six freeze-thaw cycles, and then treated with 30-second sonication. The samples were then treated with RNase A (1 U/μL), DNase I (2 U/μL), or proteinase K (40 μg/mL; MCE, Monmouth Junction, NJ, USA) to selectively remove RNA, DNA, or proteins, respectively. After 30-min incubation at 37 °C, reactions were terminated with PMSF (for proteinase K inactivation) or EDTA (for DNase I/RNase A inactivation) as appropriate, followed by two PBS washes prior to cellular processing.

Lipid extraction from BEVs

The BEVs isolated by ultracentrifugation were resuspended in deionized water (1.3 mg/mL). A 200 μL aliquot of the BEVs suspension was mixed with 375 μL of chilled methanol, followed by addition of 3 volumes (1125 μL) of methyl tert-butyl ether (MTBE; Sigma-Aldrich, St. Louis, MO, USA) and vigorous shaking for 90 min. After adding 650 μL deionized water and incubating at room temperature for 15 min, the mixture was centrifuged at 10,000 × g for 10 min to separate the upper phase (lipophilic fraction) from the lower phase (hydrophilic fraction). The lipophilic fraction was concentrated by liquid nitrogen evaporation to obtain purified lipids for cellular treatment (20 μL/mL culture medium).

Lipidomic analysis of BEVs

The BEVs were mixed with chloroform: methanol (2:1), followed by the addition of an equal volume of distilled water. After vortexing, the mixture was subjected to centrifugation at 3000 × g for 15 min, and the lower chloroform fraction was collected, dried under nitrogen, and resuspended in a suitable organic solvent (e.g., methanol or chloroform-methanol). Lipid species were profiled and fatty acid composition was determined using liquid chromatography-mass spectrometry (LC-MS; Suzhou PANOMIX Biotechnology Co., Ltd., Guangdong, China).

Statistical analysis

Data analysis was performed using GraphPad Prism 10.1.2. For comparisons between two groups, an unpaired two-tailed Student’s t-test was employed, while one-way ANOVA with Tukey’s post hoc test was used for comparisons involving three or more groups. Results are expressed as mean ± SEM, unless stated otherwise. Statistical significance was determined at *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and ns (not significant) for P > 0.05.

Data availability

The 16S sequencing data from this study are available in the NCBI Sequence Read Archive under the accession number PRJNA1210767. All data presented in this work are available within the article and the supplementary files. Any additional requests can be addressed to the corresponding authors.

References

Agarwal, A., Gupta, S., Sekhon, L. & Shah, R. Redox considerations in female reproductive function and assisted reproduction: from molecular mechanisms to health implications. Antioxid. Redox Signal. 10, 1375–1404 (2008).
Article CAS PubMed Google Scholar
Samet, J. M. & Wages, P. A. Oxidative stress from environmental exposures. Curr. Opin. Toxicol. 7, 60–66 (2018).
Article PubMed PubMed Central Google Scholar
Tan, B. L., Norhaizan, M. E. & Liew, W.-P. -P. Nutrients and oxidative stress: friend or foe?. Oxid. Med. Cell. Longev. 2018, 9719584 (2018).
Article PubMed PubMed Central Google Scholar
Gu, F. et al. The hindgut microbiome contributes to host oxidative stress in postpartum dairy cows by affecting glutathione synthesis process. Microbiome 11, 87 (2023).
Article CAS PubMed PubMed Central Google Scholar
Fan, Y. & Pedersen, O. Gut microbiota in human metabolic health and disease. Nat. Rev. Microbiol. 19, 55–71 (2021).
Article CAS PubMed Google Scholar
Zhang, J., Xia, Y. & Sun, J. Breast and gut microbiome in health and cancer. Genes Dis. 8, 581–589 (2021).
Article PubMed Google Scholar
Rooks, M. G. & Garrett, W. S. Gut microbiota, metabolites and host immunity. Nat. Rev. Immunol. 16, 341–352 (2016).
Article CAS PubMed PubMed Central Google Scholar
Wang, H. et al. The microbial metabolite trimethylamine N-oxide promotes antitumor immunity in triple-negative breast cancer. Cell Metab. 34, 581–594.e588 (2022).
Article CAS PubMed Google Scholar
Cui, W. et al. Gut microbial metabolite facilitates colorectal cancer development via ferroptosis inhibition. Nat. Cell Biol. 26, 124–137 (2024).
Article CAS PubMed Google Scholar
Toyofuku, M., Nomura, N. & Eberl, L. Types and origins of bacterial membrane vesicles. Nat. Rev. Microbiol. 17, 13–24 (2019).
Article CAS PubMed Google Scholar
Hong, M. et al. Fusobacterium nucleatum aggravates rheumatoid arthritis through FadA-containing outer membrane vesicles. Cell Host Microbe 31, 798–810.e797 (2023).
Article CAS PubMed Google Scholar
Wylensek, D. et al. A collection of bacterial isolates from the pig intestine reveals functional and taxonomic diversity. Nat. Commun. 11, 6389 (2020).
Article CAS PubMed PubMed Central Google Scholar
Qiao, Y. et al. Alterations of the gut microbiota in high-fat diet mice is strongly linked to oxidative stress. Appl. Microbiol. Biotechnol. 97, 1689–1697 (2013).
Article CAS PubMed Google Scholar
Mossad, O. et al. Gut microbiota drives age-related oxidative stress and mitochondrial damage in microglia via the metabolite N 6-carboxymethyllysine. Nat. Neurosci. 25, 295–305 (2022).
Article CAS PubMed Google Scholar
Konstantinov, S. R. et al. Specific response of a novel and abundant Lactobacillus amylovorus-like phylotype to dietary prebiotics in the guts of weaning piglets. Appl. Environ. Microbiol. 70, 3821–3830 (2004).
Article CAS PubMed PubMed Central Google Scholar
Omar, J. M. et al. Lactobacillus fermentum and Lactobacillus amylovorus as probiotics alter body adiposity and gut microflora in healthy persons. J. Funct. Foods 5, 116–123 (2013).
Article Google Scholar
Hao, Y. et al. Limosilactobacillus mucosae and lactobacillus amylovorus protect against experimental colitis via upregulation of colonic 5-hydroxytryptamine receptor 4 and transforming growth factor-β2. J. Nutr. 153, 2512–2522 (2023).
Article PubMed Google Scholar
Fang, S. et al. β-fructosidase FosE activity in Lactobacillus paracasei regulates fructan degradation during sourdough fermentation and total FODMAP levels in steamed bread. LWT 145, 111294 (2021).
Article CAS Google Scholar
Fan, J. et al. Novel mechanism by which extracellular vesicles derived from Lactobacillus murinus alleviates deoxynivalenol-induced intestinal barrier disruption. Environ. Int. 185, 108525 (2024).
Article CAS PubMed Google Scholar
Tao, S. et al. Extracellular vesicles derived from Lactobacillus johnsonii promote gut barrier homeostasis by enhancing M2 macrophage polarization. J. Adv. Res. 69, 545–563 (2024).
Article PubMed PubMed Central Google Scholar
Kim, W. et al. Lactobacillus plantarum-derived extracellular vesicles induce anti-inflammatory M2 macrophage polarization in vitro. J. Extracell. Vesicles 9, 1793514 (2020).
Article CAS PubMed PubMed Central Google Scholar
Han, F. et al. Extracellular vesicles from Lactobacillus druckerii inhibit hypertrophic scar fibrosis. J. Nanobiotechnol. 21, 113 (2023).
Article CAS Google Scholar
Chen, C.-Y. et al. Glucocorticoid-induced loss of beneficial gut bacterial extracellular vesicles is associated with the pathogenesis of osteonecrosis. Sci. Adv. 8, eabg8335 (2022).
Article CAS PubMed PubMed Central Google Scholar
Dixon, S. J. et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell 149, 1060–1072 (2012).
Article CAS PubMed PubMed Central Google Scholar
Mortensen, M. S., Ruiz, J. & Watts, J. L. Polyunsaturated fatty acids drive lipid peroxidation during ferroptosis. Cells 12, 804 (2023).
Article CAS PubMed PubMed Central Google Scholar
Jiang, X., Stockwell, B. R. & Conrad, M. Ferroptosis: mechanisms, biology and role in disease. Nat. Rev. Mol. Cell Biol. 22, 266–282 (2021).
Article PubMed PubMed Central Google Scholar
Papsdorf, K. et al. Lipid droplets and peroxisomes are co-regulated to drive lifespan extension in response to mono-unsaturated fatty acids. Nat. Cell Biol. 25, 672–684 (2023).
Article CAS PubMed PubMed Central Google Scholar
Magtanong, L. et al. Exogenous monounsaturated fatty acids promote a ferroptosis-resistant cell state. Cell Chem. Biol. 26, 420–432.e429 (2019).
Article CAS PubMed PubMed Central Google Scholar
Liang, X. et al. Peroxiredoxin 4 protects against ovarian ageing by ameliorating D-galactose-induced oxidative damage in mice. Cell Death Dis. 11, 1053 (2020).
Article CAS PubMed PubMed Central Google Scholar
Xiao, Z. et al. Ferrostatin-1 alleviates lipopolysaccharide-induced cardiac dysfunction. Bioengineered 12, 9367–9376 (2021).
Article CAS PubMed PubMed Central Google Scholar
Hoshino, D. et al. Exosome secretion is enhanced by invadopodia and drives invasive behavior. Cell Rep. 5, 1159–1168 (2013).
Article CAS PubMed Google Scholar
Coyle, C. H. et al. Mechanisms of H₂O₂-induced oxidative stress in endothelial cells. Free Radic. Biol. Med. 40, 2206–2213 (2006).
Article CAS PubMed Google Scholar
Wu, C. et al. Induction of ferroptosis and mitochondrial dysfunction by oxidative stress in PC12 cells. Sci. Rep. 8, 574 (2018).
Article PubMed PubMed Central Google Scholar
Sternlicht, M. D. Key stages in mammary gland development: the cues that regulate ductal branching morphogenesis. Breast Cancer Res. 8, 1–11 (2005).
Article Google Scholar
Schwertfeger, K. L. et al. Expression of constitutively activated Akt in the mammary gland leads to excess lipid synthesis during pregnancy and lactation. J. Lipid Res. 44, 1100–1112 (2003).
Article CAS PubMed Google Scholar
Konoshenko, M. Y., Lekchnov, E. A., Vlassov, A. V. & Laktionov, P. P. Isolation of extracellular vesicles: general methodologies and latest trends. Biomed. Res. Int. 2018, 8545347 (2018).
Article PubMed PubMed Central Google Scholar
Yáñez-Mó, M. et al. Biological properties of extracellular vesicles and their physiological functions. J. Extracell. Vesicles 4, 27066 (2015).
Article PubMed Google Scholar

Download references

Acknowledgements

This study was financially supported by the National Key R&D Program of China (2021YFD1300700 and 2024YFD1301004), the Guangdong Basic and Applied Basic Research Foundation (2023A1515012098), the Science and Technology Program of Guangzhou (202102020056), and the Talent Project of the Wen’s Science and Innovation Center (SCAU). All the experimental illustrations and the graphical abstract were created using BioRender.com (Permissions for use have already been obtained).

Author information

Authors and Affiliations

Guangdong Province Key Laboratory of Animal Nutrition Control, College of Animal Science, South China Agricultural University, Guangzhou, China
Qianzi Zhang, Dongpang Chen, Hanting Ding, Qihui Li, Siyu Yuan, Haobin Li, Wutai Guan & Shihai Zhang
College of Animal Science and National Engineering Research Center for Breeding Swine Industry, South China Agricultural University, Guangzhou, China
Wutai Guan & Shihai Zhang
Guangdong Laboratory for Lingnan Modern Agriculture, South China Agricultural University, Guangzhou, China
Wutai Guan & Shihai Zhang

Authors

Qianzi Zhang
View author publications
Search author on:PubMed Google Scholar
Dongpang Chen
View author publications
Search author on:PubMed Google Scholar
Hanting Ding
View author publications
Search author on:PubMed Google Scholar
Qihui Li
View author publications
Search author on:PubMed Google Scholar
Siyu Yuan
View author publications
Search author on:PubMed Google Scholar
Haobin Li
View author publications
Search author on:PubMed Google Scholar
Wutai Guan
View author publications
Search author on:PubMed Google Scholar
Shihai Zhang
View author publications
Search author on:PubMed Google Scholar

Contributions

S.Z. and W.G. designed the research. Q.Z. and Q.L. wrote the initial manuscript. Q.Z., Q.L., S.Y., and H.L. performed the experimental work and collected data. Q.Z. and D.C. analyzed the data. Q.Z., D.C., and H.D. designed the figures. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Shihai Zhang.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary File-Western Blot Bands

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zhang, Q., Chen, D., Ding, H. et al. Lactobacillus amylovorus extracellular vesicles mitigate mammary gland ferroptosis via the gut-mammary gland axis. npj Biofilms Microbiomes 11, 113 (2025). https://doi.org/10.1038/s41522-025-00752-4

Download citation

Received: 17 February 2025
Accepted: 04 June 2025
Published: 21 June 2025
DOI: https://doi.org/10.1038/s41522-025-00752-4