Polystyrene microplastics induce potential toxicity through the gut-mammary axis

Abstract

Microplastics (MPs), as an emerging environmental pollutant, pose a grave threat to food safety and public health. However, studies on MP toxicity to organs other than the intestine remain limited, especially its link to the intestinal microbiota. To address this gap, we evaluated the potential toxicity of polystyrene (PS)-MPs to the gut and mammary glands during lactation exposure in mice. PS-MPs (~1 μm) can disrupt the intestinal barrier and cause colonic inflammation and gut microbiota dysbiosis. Moreover, they can accumulate in mammary tissue and cause inflammatory damage. Transcriptome data suggested that PS-MPs cause maternal mammary lipid metabolism disorders and ferroptosis. Fecal microbial transplant (FMT) was then performed, and it reproduced the observed leakage of the blood-milk barrier and inflammation of the mammary gland. This study demonstrated that MPs induced gut and mammary inflammation and exacerbated inflammatory damage through the gut-mammary axis. In addition, MPs caused mammary lipid disorders and ferroptosis. The findings confirmed that PS-MPs may be transported to mammalian organs other than the intestine (e.g., mammary gland) and revealed the critical role of the intestinal microbiota. These findings will provide guidance for further studies on the potential foodborne risks of MPs.

Introduction

Extensive use of plastic items has contributed to plastic production and social development^1,2. Nevertheless, massive quantities of plastics in the environment are hard to degrade effectively³. These plastics accumulate in the natural environment and spread through the food chain, which poses a great threat to the ecosystem and human health⁴. Russell et al. utilized the word “microplastics” (MPs) for the first time in 2004 to characterize small plastic particles⁵. Plastics with diameters less than 5 mm are defined as MPs, while plastics with diameters smaller than 1 μm are referred to as nanoplastics (NPs)⁶. These smaller-sized particles can readily enter organisms and induce toxicity; moreover, they can interact with contaminants in the natural environment⁷.

Particle size is a key determinant of MPs' ecotoxicity. MPs with particle sizes less than 100 μm can pass through the gut barrier and reach other in the body⁸. Thus, further investigations are required to determine the ecotoxicity and health risks induced by smaller MPs (<5 μm) compared to larger MPs (>100 μm). Many studies have shown that MPs can be seriously toxic to several organs, such as the gut, kidneys, lungs, and liver^9–11. The gut is the major organ for absorption, digestion, and immunity maintenance. The gut mucosa is the intestinal tract’s first line of defense against harmful substances from the outside world, and the group of tiny organisms that colonize the intestines is called the intestinal flora¹². Investigations have demonstrated the importance of the intestinal microbiota in substance secretion, nutrient absorption, barrier maintenance, immunity, and metabolic regulation¹³. Recent studies have shown that MPs can trigger gut ecological dysregulation, barrier dysfunction, and metabolic disorders, and increase the risk of diabetes in mice^14–16. Nevertheless, the influence of exposure to smaller sizes of MPs (~1 μm) on mammalian distal organs, especially mammary tissue, has not been clarified.

The mammary gland is an essential organ for the secretion of milk; its structure and development are crucial for the maintenance of lactation and the baby’s health¹⁷. A recent study reported that the intestinal microbiota may promote mastitis and revealed the notion that the gut-mammary axis¹⁸. Dysregulated gut ecology can compromise the gut barrier and induce systemic inflammation, leading to the migration of gut bacteria and secondary metabolites to distant organs¹⁹. Moreover, intake of MPs triggers gut inflammation and liver metabolic disorders, resulting in severe gut-liver axis disruption²⁰. However, studies on MP-induced toxicity to organs beyond the gut and liver, especially distal organs such as the mammary gland, are still scarce. Thus, whether MPs can directly or indirectly cause damage to mammary tissue remains largely unknown.

Here, we explored the potential toxic impact of PS-MPs through in vivo and in vitro models. The study found that orally ingested MPs can accumulate in various tissues and cause significant harm to the gut and mammary glands. In addition, we explored novel toxicity mechanisms of MPs from the perspective of the gut microbiota by combining microbiomics and transcriptomics analyses. This study provides valuable information for further investigations on MP toxicity and mitigating MP contamination in food products.

Results

Bioaccumulation and characterization of PS-MPs

Scanning electron microscopy (SEM) images showed that normal and fluorescent plastics both appeared to be regularly spherical, and the diameters were 1.15 ± 0.037 μm and 1.02 ± 0.041 μm, respectively (Fig. 1A, B). The zeta potential values ranged from −47.9 mV to −25.2 mv (zeta potential absolute value ≥ 25 mV), indicating that the PS-MPs were stable and had larger zeta potentials in the medium, suggesting that they tend to aggregate in the medium (Table S2). Raman spectroscopy verified that both PS-MPs have similar official groups, suggesting no compositional differences between ordinary and fluorescent PS-MPs (Fig. 1C). In particular, the peak at 620 cm⁻¹ indicated the bending vibration between carbon atoms within the benzene ring, the peak at 795 cm⁻¹ indicated the telescopic oscillation between the benzene ring and the carbon chain atoms, the peak at 1001 cm⁻¹ indicated aromatic ring respiration vibrations, and the peak at 1603 cm⁻¹ indicated the characteristic peaks of the benzene ring²¹. After lactation exposure ended, weak fluorescent signals were observed in the spleen, colon, kidney, mammary gland, liver, and lung tissues (Fig. 1D). Compared to the control group, the body weight in the PS-MP groups did not significantly differ (Fig. 1E, F). In addition, quantitative Raman spectroscopy results showed no remarkable enrichment of PS-MPs in the heart, spleen, kidney, and lung tissues. MPs exhibited high accumulation rates in the colon, liver, and feces (Fig. 1G). The gut and liver, as major digestive organs and metabolic hubs, tended to have high accumulation rates. The highest accumulation rate in the feces indicated that PS-MPs are excreted primarily through the intestinal tract and thus show limited transfer to other organs.

Fig. 1. Characteristics of PS-MPs and bioccumulation of PS-MPs in various tissues.

A, B SEM images and particle size distributions of PS-MPs. C Raman spectral images of PS-MPs and fluorescent PS-MPs. D Accumulation of fluorescent PS-MPs in the mouse heart, colon, kidney, spleen, liver, mammary gland, and lung tissues. E Experimental design: Female mice in the treatment group received PS-MPs (3 mg/L and 30 mg/L drinking water) during lactation, whereas mice in the control group received normal water. F Curves of body weight changes in the control and PS-MPs groups during lactation. G Quantitative analysis of PS-MPs in the heart, liver, colon, kidney, spleen, mammary, lung, and feces of mice (bioaccumulation rate, value normalized) (n = 4). *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, compared to the control.

PS-MPs-induced colonic inflammation and intestinal barrier leakage

Next, we assessed the impact of PS-MPs on the gut. Compared with that in the control mice, the colonic length in the high-dose group was markedly shortened (Fig. 2A, B). The SEM pictures demonstrated that small-sized MPs effectively crossed the intestinal barrier (Fig. 2C). The transmission electron microscopy (TEM) findings revealed that tight junctions (blue arrows) were broken and gut villi (red arrows) were detached in the MP-exposed groups (Fig. 2D). We further observed that inflammatory infiltration increased and the goblet cells number decreased in the MP-treated groups (Fig. 2E, J, K). Moreover, the intestinal pro-inflammatory factors were markedly increased in the MP-treated group with dose effects (Fig. 2G–I), indicating corresponding gut inflammatory damage. The western blot (WB) showed that the TLR4, MyD88, and phospho-NF-κB proteins were remarkably upregulated in MP-treated groups (Fig. 2S), indicating that PS-MPs could trigger gut inflammation through the TLR4-NF-κB pathway. Additionally, PS-MPs decreased the expression level of barrier proteins (Fig. 2F, S, Fig. S1). The qPCR analyses on pro-inflammatory and barrier genes revealed similar trends (Fig. 2T). These results indicated that PS-MPs induce colonic inflammation and intestinal barrier leakage, causing serious dysregulation of intestinal ecology, which is consistent with previous findings^22,23.

Fig. 2. PS-MPs triggered colonic inflammation and disrupted the intestinal barrier.

A, B Representative images of colonic length. C SEM images of the inner lumen of the colon. D TEM pictures of tight junctions and colonic villi. E Periodic acid Schiff (PAS)- and hematoxylin-eosin (HE)-stained images. F Immunohistochemical (IHC) images of tight junction proteins. G–I Measurement of inflammation-related indices. J Histologic scoring of HE-stained sections. K Number of goblet cells in the crypts. L–S Protein expression levels associated with the colonic barrier and TLR4 pathway. T mRNA levels of inflammatory and barrier-related factors. *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, ****p-value < 0.0001, compared to the control.

PS-MPs-induced gut ecological dysbiosis and gut microbiota disruption

Further, we investigated the effects of PS-MPs on gut microbiota. Compared to the controls, no significant differences were observed in the Shannon and Simpson indices in the MP-treated group, while Choa1 indices and the operational taxonomic unit (OUT) abundance were markedly elevated in the low-dose group (Fig. 3A). The principal coordinate analysis (PCoA) analysis showed remarkable differences in the gut microbiota between the groups (Fig. 3F). At the phylum level, Bacteroidota, Proteobacteria, and Firmicutes were the major dominant phyla of the microbial community (Fig. 3B). Compared with the controls, the PS-MP treatment showed no significant change in the proportion of Firmicutes to Bacteroidota (Fig. 3D). Bacterial taxonomic communities between distinct groups were further identified according to the linear discriminant analysis of effect size (LEfSe) (log₁₀LDA score >4) (Fig. 3E, H). The abundance of Dubosiella, Akkermansia, Faecalibaculum, and Bifidobacterium was significantly decreased in the PS-MP group (Fig. 3C, G), while the abundance of Erysipelatoclostridium and Coriobacteriaceae_UCG-002 was significantly increased (Fig. 3G). Short-chain fatty acids (SCFAs) analysis showed that the butyric acid, acetic acid, and propionic acid levels significantly reduced with PS-MP treatment (Fig. 3I). Subsequently, we explored the association among the intestinal microbiota and PS-MP-induced colonic indexes, phenotype, and inflammation by Spearman’s association analysis (Fig. 3J). Results revealed that Akkermansia, Mucispirillum, and Faecalibaculum were positively and significantly associated with the barrier-related indices, number of goblet cells, and length of the colon, and negatively related to the upregulation of pro-inflammatory factors. Moreover, Aerococcus and Erysipelatoclostridium were negatively correlated with barrier-related indices. Overall, these data indicate that the gut microbiota likely plays a major role in PS-MP-induced colonic inflammation and injury.

Fig. 3. PS-MPs might disrupt the intestinal microbiota and further induce gut ecological dysbiosis.

A Simpson, Chao1, Shannon, and OUT indices indicated changes in α-diversity among the control and MP treatment groups. B Abundance of microbes at the phylum level between the two groups. C Abundance of microbes at the genus level between the two groups. D Ratio of Firmicutes/Bacteroidota. E Cladogram of the LEfSe showing that various taxa were enriched in different groups from phyla to genera. Screening for differential microbiota with the Wilcoxon rank-sum test. F PCoA scoring graph for mice fecal samples. G Abundance of Dubosiella, Faecalibaculum, Akkermansia, Bifidobacterium, Erysipelatoclostridium, and Coriobacteriaceae_UCG-002 (Mann−Whitney U-test). H LEfSe showed that distinct bacterial populations were significantly enriched in various groups (log10LDA score >3). I SCFA analyses revealed the relative abundance of every constituent. J Spearman's relevance analysis. Red denotes a positive association, and green denotes a negative association. *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, ****p-value < 0.0001, compared to the control.

PS-MPs damaged the mammary barrier and induced mammary inflammation

Subsequently, we investigated the effects of PS-MPs on the mammary. Notably, the SEM images identified MPs in the follicles of the breast tissues (Figs. 1G, 4B), indicating that MPs passed through the gut barrier and arrived at distant tissues. Compared to the control mouse, the mammary glands of the mouse in the high-dose microplastics (H-MPS) group had significant leukocyte infiltration (Fig. 4A, C). Similarly, the IL-6, TNF-α, and serum LPS levels and MPO levels were markedly increased in the H-MPS group, although the low-dose microplastics (L-MPS) group did not show significant changes (Fig. 4D–G). The pro-inflammatory factor IL-1β level showed a dose effect (Fig. 4H). IHC results indicated that the area positive for barrier-related proteins was remarkably decreased in the H-MPS group than in the controls, but did not significantly differ in the L-MPS group (Fig. 4I, Fig. S2). The WB trend of barrier-related proteins was consistent with the IHC results (Fig. 4J–M). MyD88, TLR4, and phospho-NF-κB protein levels were significantly increased in the PS-MPs group than in the control group, showing that PS-MPs may activate the Toll-like receptor 4 (TLR4) inflammatory pathway (Fig. 4J, N–Q). Furthermore, qPCR-based results of inflammation- and barrier-associated genes supported our suspicions (Fig. 4R). These data suggest that PS-MPs can cross the mammary barrier and cause inflammation of the mammary gland, especially in the high-dose group.

Fig. 4. PS-MPs-induced mild mammary gland inflammation and damaged the blood-milk barrier.

A HE and PAS staining images. Red arrows represent inflammatory cell infiltration. B Images of breast follicles by SEM (H-MPS: 30 mg/L). Red arrows denote PS-MPs particles. C Histologic scoring according to HE-stained mammary sections. D, E Mammary tissue MPO activity and serum LPS levels. F–H Measurement of inflammatory markers in mammary tissues. I IHC pictures of tight junction proteins. J–Q Protein levels associated with the TLR4 pathway and the mammary barrier. R mRNA expressions of inflammatory- and mammary barrier-associated factors. *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, compared to the control.

Transcriptomic analysis revealed PS-MPs-induced ferroptosis and autophagy in the mammary glands

To further reveal the damaging effects of PS-MPs on the mammary gland, transcriptomic analysis of mammary tissues was performed for the H-MPS group. Compared with those in the controls, 944 genes were down-regulated and 91 genes were upregulated in the H-MPS group (Fig. 5A, I). PCA revealed large differences among the samples from the control and PS-MP groups (Fig. 5E). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses revealed alterations in mitochondrial membrane-associated and multiple disease pathways. The PPAR signaling, fatty acid metabolism or degradation, peroxisome, glutathione metabolism, citrate cycle, oxidative phosphorylation, and glycolysis/gluconeogenesis pathways were significantly enriched (Fig. 5B, C). These data suggested that PS-MPs influenced mammary tissue damage and regulated mammary lipid synthesis and metabolism. To further assess the overall biological effects of the transcriptome, we performed Gene Set Enrichment Analyses (GSEAs) for all genes. The GSEA results showed that ferrous iron binding, ferroptosis, TNFR2 non-canonical NF-κB, and autophagy-other pathways were significantly activated (Fig. 5D, F–H). These changes imply that PS-MPs may induce mammary gland injury through iron ferroptosis and autophagy pathways. Furthermore, we observed the ultrastructure of mammary tissues. TEM images showed that the PS-MP group exhibited mitochondrial shrinkage, cristae disruption, and autophagic vacuoles, which suggested the occurrence of ferroptosis and autophagy (Fig. 5J). The expression of genes associated with ferroptosis and autophagy confirmed this hypothesis (Fig. 5K).

Fig. 5. PS-MPs-induced mammary gland inflammation, ferroptosis, and autophagy in mice.

A, B Reactive oxygen species (ROS) levels in breast epithelial cells of the control and treated mice. C, D Detection of mitochondrial membrane potential (MMP) changes (red/green ratios) in different groups of mice mammary epithelial cells by JC-1 probe. E Detection of cellular iron content in different groups. F–H Levels of lipid peroxidation associated indices malondialdehyde (MDA), glutathione (GSH), and superoxide dismutase (SOD). I, J Detection of mammary ferroptosis proteins GPX4, FTH1, ACSL4, and SLC7A11. β-actin was used as the control. K mRNA levels of ferroptosis- and lipid peroxidation-related genes assessed by qPCR. * p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, ****p-value < 0.0001, compared to the control.

PS-MPs-induced ferroptosis and lipid peroxidation in the mammary gland

Next, we tested the indicators related to ferroptosis and peroxidation. Cellular experiments in vitro showed that ROS levels were markedly elevated as the concentration of PS-MPs increased (Fig. 6A, B). The occurrence of iron death is often accompanied by changes in mitochondrial function²⁴. MMP (red/green ratio) was significantly reduced in the MP-exposed groups and showed a dose effect, compared to controls (Fig. 6C, D). Moreover, the total iron content of cells in the PS-MP group was markedly increased (Fig. 6E). These experiments showed that a certain level of PS-MPs-induced ferroptosis in mouse mammary cells. We observed that the lipid peroxidation MDA level was significantly elevated while the GSH and SOD levels were remarkably reduced (Fig. 6F–H), suggesting that PS-MPs damaged the antioxidant defense system of mouse mammary glands, resulting in lipid peroxidation. The WB results demonstrated that MPs could decrease SLC7A11, GPX4, and FTH1 protein levels and increase ACSL4 protein levels in the mammary tissues (Fig. 6I, J). Then, we focused on the expression of genes related to ferroptosis and iron metabolism in the mammary tissues of the mouse. Consistently, the qPCR results supported our speculations (Fig. 6K). In conclusion, in vivo and in vitro investigations indicated that PS-MPs-induced ferroptosis in mouse mammary glands.

Fig. 6. PS-MPs-induced lipid peroxidation and ferroptosis in mice mammary glands.

A, B ROS levels in mammary epithelial cells. C, D Detection of MMP changes in mouse mammary epithelial cells. E Detection of cellular iron contents. F–H Levels of lipid peroxidation- associated indexes GSH, MDA and SOD. I, J Detection of mammary ferroptosis proteins. K mRNA levels of ferroptosis- and lipid peroxidation-related genes. *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, ****p-value < 0.0001, compared to the control.

PS-MPs exacerbated inflammation and disrupted the mammary barrier in mice via the gut microbiome

Next, we investigated whether PS-MPs promote mammary inflammation and ferroptosis by modulating the intestinal microbiota. First, we investigated the microbial compositions of donor-independent and mixed samples. The mixed samples revealed a significant increase in colony abundance and a significant elevation of dominant genera, suggesting further optimization of the recipient transplantation effect (Fig. S3). FMT from donors in the PS-MP group evoked mammary gland damage and elevated levels of inflammatory markers in recipient mice than in the control and CMT groups (Fig. 7C–I). The TLR4 inflammatory pathway was also activated (Fig. 7K, M–O). Furthermore, the IHC and protein findings suggested that in the HMT group, the integrity of the mammary barrier was disrupted (Fig. 7B, K, M, Fig. S4). The expression levels of FTH1, GPX4, and SLC7A11 were unchanged in the HMT group, whereas the ACSL4 protein level was significantly elevated. These changes suggested that the ferroptosis pathway was not activated (Fig. 7J, L). Moreover, we examined the mRNA levels of ferroptosis, inflammation, and blood-milk barrier-associated factors and found that they changed in a manner consistent with our predicted results (Fig. 7P). These results suggested that the FMT from the PS-MP group triggered mammary gland inflammation in recipient mice but had no effect on ferroptosis, which may be due to differences in microbial composition and function.

Fig. 7. FMT from PS-MPs-treated mice induced inflammation in the breast of recipient mice without triggering ferroptosis.

A Scheme of experimental design. B IHC images of tight junction proteins. C Representative HE pictures of mammary sections. D Histologic scores were based on HE-tainted sections. E, F Mammary tissue MPO activity and serum LPS levels. G–I Determination of inflammatory markers. J Detection of mammary ferroptosis proteins. K Expression levels of tight junction and inflammatory proteins. L–O Quantitative results of proteins. P mRNA levels of ferroptosis, inflammation, and blood-milk barrier-related genes. *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001, compared to the control.

Discussion

Due to rapid industrialization, MP pollution has developed into a worldwide environmental difficulty²⁵. However, MP damage to distal organs beyond the gut remains unknown. Furthermore, whether the gut microbiota plays a critical role in MP-induced tissue injury remains unclear. Previous investigations have suggested that MPs might accumulate in the liver, intestines, and kidneys, thereby causing histopathological changes^15,26. We observed significant accumulation of PS-MP in the colonic tissues and differential detachment in the small intestinal villi. Additionally, the Toll-like receptor and NF-κB pathway associated with inflammation were activated. Toll-like receptor represents a particular target of LPS (components of the outer cell wall of gram-negative bacteria)²⁷. LPS could be directly recognized by TLR4 (an important family member of TLRs), which activates the TLR4 pathway²⁸. The TLR4 pathway is strongly related to the NF-κB or JNK signaling pathways²⁹. TLR4 can activate the downstream NF-κB cascade (through the MyD88-dependent pathway), resulting in the release of a large number of pro-inflammatory molecules, such as IL-6, IL-1β, and TNF-α³⁰. Our results imply that PS-MPs-induced colonic inflammation and damaged the intestinal barrier, thereby allowing the passage of LPS.

The intestinal microbiota is an essential driver that regulates the organic immunity system under pathological and physiological conditions^31,32. Based on 16S rRNA sequencing data, we observed that exposure to environmentally relevant doses of PS-MPs impacted intestinal microbial communities and functions. However, exposure to PS-MPs during lactation did not impact the richness of microbial species (alpha diversity) but led to remarkable differences in the composition of microbes (beta diversity) between groups. This finding might indicate that environmentally relevant doses of PS-MPs did not alter the overall microbial abundance in individuals. The abundance of probiotics Dubosiella, Faecalibaculum, Bifidobacterium, and Akkermansia was reduced significantly by PS-MPs, while those of pathogenic Erysipelatoclostridium and Coriobacteriaceae_UCG-002 were markedly elevated. Akkermansia is mainly colonized in the gut mucosal layer, and changes in its population may be related to the development of various illnesses, such as autism, obesity, and Crohn's disease (CD)³³. Increasing the amount of Akkermansia can reduce inflammation and gut permeability³⁴. Dubosiella is strongly associated with barrier and colitis-related parameters³⁵, and Faecalibaculum is an essential butyrate producer, which acts as a significant anti-inflammatory in the colon and helps to maintain Th17/Treg homeostasis³⁶. In addition, the abundance of Coriobacteriaceae_UCG-002 and Erysipelatoclostridium was positively correlated with intestinal injury and inflammation^37,38. These bacterial changes imply that intestinal microbes may mediate PS-MP-induced inflammation of the colon. Particularly, decreases in the probiotic bacteria Akkermansia and Faecalibaculum alter the intestinal metabolic environment and inhibit the synthesis of SCFAs. These bacteria weaken the host’s anti-inflammatory capacity, leading to the translocation of pathogenic bacteria and increasing the risk of disease. Research had suggested that SCFAs are essential in maintaining intestinal health and metabolic homeostasis^39–41. In the present investigation, PS-MPs significantly decreased the levels of butyric, acetic, and propionic acids. Similarly, we observed a significant reduction in the bacteria related to SCFAs production. The above results imply that the deterioration of the intestinal microbiota was closely associated with PS-MP-induced intestinal damage.

We also observed damage to distal organs (mammary glands) outside the gut by PS-MPs. The results showed that PS-MPs disrupted the blood-milk barrier, causing mammary gland inflammation, and activated the TLR4 pathway. Additionally, significant modifications in factors associated with the blood-milk barrier and inflammation were observed in the high-dose group, suggesting a dose effect of PS-MP-induced mammary damage. Recent studies have revealed that dysregulation of gut ecology causes LPS release, which regulates the process of mastitis¹⁸. In the current study, the elevated serum LPS levels hinted that PS-MP caused dysregulation of gut ecology may be related to the process of mastitis. Further transcriptomic data demonstrated macroscopic alterations in mammary genes induced by maternal PS-MP exposure. These differential genes involve various biological processes and critical pathways, including the PPAR signaling pathway, TCA cycle, metabolic processes, mitochondria, fatty acid metabolism, and glutathione metabolism. The PPAR signaling pathway is involved in metabolizing exogenous and endogenous substances⁴², while the TCA cycle is a major energy metabolism pathway in mammals⁴³. These pathways are important components of mammary gland metabolism. Mammary lipid metabolism disorder leads to endocrine disruption and metabolic diseases⁴⁴. Important metabolic routes of the mammary gland become reprogrammed following exposure to PS-MPs during lactation, which may be a major cause of changes in the physiological, biochemical, and histopathological indices in mice. Studies have shown that the perinatal period in animals is a critical period for feeding behavior, adipose tissue differentiation, and energy expenditure^45,46; thus, animals are vulnerable to MPs during this period. Mammals may be vulnerable to environmental chemical disturbances that can affect disease risk, and disorders of mammary metabolism can affect maternal health and infant development¹⁷. Of note, ferroptosis- and autophagy-related pathways were significantly altered, which could be a potential factor underlying mammary gland damage and certain metabolic diseases caused by PS-MPs.

As is well known, ferroptosis is a new mode of cell death associated with iron overload. NPs can cause ferroptosis in ovarian cells⁴⁷, while MPs can induce ferroptosis and aggravate liver injury in mice⁴⁸. Nevertheless, the effect of PS-MPs on ferroptosis in the mammary gland has never been previously reported. We found that PS-MPs caused lipoperoxidation and mammary iron accumulation and promoted the development of ferroptosis through in vivo and in vitro studies. FMT experiments reproduced the weakening of the blood-milk barrier, elevated serum LPS, and mammary gland inflammation caused by PS-MPs. Unlike the results observed in the MP treatment, the FMT experiments failed to reproduce mammary ferroptosis. These findings indicated that the intestinal microbiota was involved in PS-MP-induced mammary gland damage. Our study presents certain limitations. For example, we only used PS-MPs as the experimental materials, whereas actual food sources of microplastics are diverse, including polyvinyl chloride (PVC), polyethylene (PE), and polypropylene (PP). Future studies should include additional polymer types, size ranges, and morphologies to reflect the complexities in real food systems.

Nevertheless, our study highlights the complex mechanisms of MPs-induced toxicity, including dysregulated intestinal ecology, abnormal lipid metabolism, tissue inflammation, and ferroptosis (Fig. 8). These findings indicate that ingestion of MPs poses a potential health risk. Further control of MP contamination in food is needed to promote public health and wellness.

Fig. 8.

Schematic diagram of the mechanism of PS-MP-induced damage to the gut-mammary axis (by biorender.com).

Methods

Chemicals and cell culture

Green fluorescent and non-fluorescent PS-MPs (1 μm) were purchased from the Baseline ChromTech Research Center (Tianjin, China). Two PS-MPs were analyzed using SEM, and the compositions of these plastics were examined using Raman spectroscopy. The particle size and zeta potential were determined using a nano laser particle sizer (Malvern, UK). The environmentally relevant doses in the current study were based on previous studies on human MP intake through food (details are provided in the Supporting Information). For this study, a normal mouse breast epithelial cell line was used (EpH4). EpH4 cells were exposed to 0, 0.15, 0.3, or 0.6 mg/mL PS-MPs for 48 h. Subsequently, the medium was discarded, and cells were cleaned at least 3 times using a sterile PBS solution. Then, the MMP, iron, and ROS levels were detected with relevant kits.

Animal husbandry

ICR mice (7 weeks old) used in this experiment were purchased from the animal experiment center of Northwest A&F University (Yangling, Shaanxi, China). All mice were housed under SPF conditions at 21–24 °C with 12 h/day of light for 1 week. The mice were then grouped in a ratio of one male and two females in cages, and the pregnancy was verified through vaginal pessaries. Then, the male mice were removed. End-of-gestation mice were divided randomly into three groups (Control, L-MPS, and H-MPS), each with seven mice. L-MPS denotes the low-dose microplastics group (3 mg/L), while H-MPS denotes the high-dose microplastics group (30 mg/L). All mice had ad libitum access to drinking water (MPs added to drinking water). After lactation, the mice were euthanized by anesthesia with isoflurane, then colonic length was measured, and serum, feces, and various tissue samples were gathered and stored at -80 °C until assay.

In the FMT experiment, feces from different treatment groups were collected and processed as described previously^19,49–52. The fresh feces from each group were mixed and homogenized with sterile PBS for ultimate concentration as 50 mg feces/mL⁵³. Next, the samples were dispensed as required for the test (0.5 g/sample), centrifuged and the supernatant collected, centrifuged and and the supernatant was collected. For transplantation, the mice (1 week after gestation) were administered daily oral antibiotics (100 mg/kg vancomycin, neomycin sulfate, and 200 mg/kg ampicillin, metronidazole). The whole process was performed for 5 successive days to clear the gut microbiota⁵⁰. After the antibiotic treatment, the subject mice were administered fecal supernatant (300 µL) orally for 3 consecutive days, followed every 2 days for 19 days⁵⁰. Animal experimental procedures were conducted according to the Guidelines for the care and use of laboratory animals and obtained approval from the Northwest A&F University Animal Ethics Committee (Approval No. 202106A124).

Contents of PS-MPs in tissues and feces

Various tissues were collected from mice and fixed in 4% paraformaldehyde. Samples were subsequently embedded in paraffin and stained. Detection of fluorescence signals of MPs in tissues by inverted fluorescence microscopy (Olympus, Japan). Quantitative analysis was performed as previously described (details are provided in the Supporting Information)^54,55.

Immunochemical and histologic analysis

Mammary and intestinal sections were stained with PAS and HE and then subjected to fluorescence microscopy (Leica, Germany). Histological scoring was performed according to a previously described method⁵⁶. The double-blind method was used to calculate the number of PAS-positive cells (PAS + ) in every crypt. During IHC staining, the tissue sections were exposed to the relevant secondary antibody for 30 min at room temperature. Images were obtained using fluorescent microscopy (Olympus, Japan). Positive areas were measured using the Image Pro program.

Enzyme-linked immunosorbent assay (ELISA)

Follow-up ELISAs were performed according to the manufacturers’ (Servicebio, China) instructions, using the following reagents: LPS (Cat#: SEB526Ge), IL-1β (Cat#: GEM0002), IL-6 (Cat#: GEM0001), TNF-α (Cat#: GEM0004), MPO (Cat#: EK2133/2-01). The OD values were recorded with a multi-functional microplate reader (BioTek, USA), and the concentration of the samples was computed with the standard curve⁵⁷.

PS-MP observation in the colon and mammary gland

For SEM (SU3900, Japan), the samples were observed directly after gold spraying. For TEM, the samples were dehydrated, embedded in resin, sectioned using an ultrathin microtome, and colored with lead citrate and uranium acetate. The tissues were then observed for ultrastructure by TEM (HT7800, Japan).

SCFA concentration analysis

The fresh fecal tissues were collected and stored immediately at -80 °C. SCFA concentrations in the samples were measured using an ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) system (Vanquish™ Flex UHPLC-TSQ Altis™, Thermo Scientific Corp., Germany). Specific descriptions of the methodology can be found in the Supplementary Information.

RNA sequencing analysis

Total RNA from mammary tissue was collected with the TRIzol kit (Invitrogen, USA). rRNA was then removed using a conventional kit (Illumina, CA, USA), and transcriptome libraries were constructed. The obtained libraries were generated with an Illumina HiSeqTM 2500 system (Illumina, CA, US). The sequencing data were analyzed by Genedenovo Biotechnology Co., Ltd.

16S rRNA sequencing

Fecal genomic DNA was obtained with a kit (TianGen, Beijing) according to the manufacturer’s recommendations. The concentration and purity of these extracts were determined by 1% agarose gel electrophoresis. Purified amplicons were sequenced using the NovaSeq platform (Illumina, USA) following the manufacturer’s recommendations. PCoA analysis aimed to show microbial multi-dimensional results, and LEfSe was applied to explain the differences in bacteria between various groups.

Quantitative RT‑PCR (RT-qPCR)

The total RNA was obtained from breast and colon tissues with the TRIzol kit (Invitrogen, USA). Total RNA from the samples was then back-transcribed into cDNA using an appropriate kit (Yeasen, Shanghai, China). Transcription products were expanded by TB Green® Premix Ex Taq™ II (Takara, Dalian, China). β-actin was taken as an internal reference gene, and the 2^-ΔΔCt method was applied to analyze the relative expression of each gene. The sequences of the primers used are presented in Table S1.

Detection by western blotting

As described previously, total protein of the mammary gland and colon tissues was obtained using the RIPA buffer (Solarbio, China)⁵⁸. Proteins were separated by SDS-polyacrylamide gel electrophoresis and bound to nitrocellulose membranes (Millipore, MA, USA). After blocking with 5% skim milk powder, immunoblotting was performed. The specific primary antibodies used were as follows: ZO-1 (Wanleibio, WL03419), P-NF-κB (Wanleibio, WL02169), β-actin (Wanleibio, WL01372), TLR4 (Sigma, SAB5700798), NF-κB (Cell Signaling Technology, Cat#8242), MyD88 (Cell Signaling Technology, Cat#4283), Claudin-3 (Abways, CY1015), Occludin (Abways, CY5997), GPX4(Abcam, Cat# ab125066), ACSL4 (Abcam, Cat# ab155282), SLC7A11 (Abcam, Cat# ab307601), FTH1 (Abcam, Cat# ab75973).

Statistical analysis

All data are expressed as means ± standard error mean (SEM). Statistical analysis was performed using GraphPad Prism (9.0, CA, USA). T-tests were used to compare data between two groups, and the one-way ANOVA was used for comparisons between multiple groups. A p-value < 0.05 indicated statistical significance.

Author contributions

Writing – original draft: Z.W. Visualization: S.W. Conceptualization: S.L. Investigation: Z.W. Methodology: F.L. Resources: Q.B. Funding acquisition and Project administration: X.A. All authors read and approved the published version of the manuscript.

Data availability

Data are available on request from the corresponding author.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41538-025-00517-5.