Gut dysbiosis induces the development of mastitis through a reduction in host anti-inflammatory enzyme activity by endotoxemia

The experimental design has been shown in Fig. 1. We first investigated the ruminal microbiota profiles in healthy and mastitis cows (Fig. 1A). To indicate the role of the ruminal microbiota in mastitis pathogenesis, RMT from healthy (H-RMT) and mastitis cows (M-RMT) to mice was performed (Fig. 1B). According to the differences identified between the healthy and mastitis groups in cows and RMT mice, we next investigated the role of ruminal dysbiosis-derived low-grade endotoxemia in the pathogenesis of mastitis in mice (Fig. 1C). Finally, we studied whether impaired host anti-inflammatory enzyme by ruminal dysbiosis aggravates pathogen-induced mastitis in mice (Fig. 1D).

Mastitis was diagnosed according to macroscopical inflammatory changes (e.g., redness, edema, fever, and pain) and somatic cell counts (SCC, SCC > 5 × 10⁵ cells/mL) [33]. Healthy cows were defined as no mammary injury and SCC less than 2 × 10⁵ cells/mL [34]. A total of 16 cows (4–6 years) were employed for the dairy sample collection, including 8 healthy and 8 mastitis cows from Qingzhou city, Shandong province, China. All cows were fed a standard diet consisting of hay and grains to meet their energy requirements [8, 33]. All healthy and mastitis cows were not exposed to antibiotics for 6 months before sample collection. The serum, milk, and ruminal samples were harvested aseptically and stored at − 80 °C until determination of systemic inflammation, mastitis, and the ruminal microbiota, respectively.

A total of 160 specific pathogen-free (SPF) grade BABL/c mice (22–24 g, 120 females and 40 males) were purchased from Liaoning Changsheng Biotechnology Co., Ltd. (Benxi, China). All mice were fed with sufficient standard diet and water under SPF-grade feeding conditions with 12 h of light daily. The mice were further mated at a ratio of three females to one male until pregnancy was confirmed, and then the male mice were separated. The RMT model was established according to previously described methods [22, 23]. In brief, ruminal samples from each group, including the healthy and mastitis groups, were mixed and homogenized with sterile PBS at a final concentration of 50 mg/mL. The prepared homogenate was centrifuged at 100×g and 4 °C for 2 min, and then the supernatant was collected and dispensed based on experimental demand. Mice were transplanted 1 week after pregnancy. These mice were treated with antibiotic (200 mg/kg of ampicillin (Sigma Aldrich, USA), metronidazole (Sigma Aldrich, USA), and neomycin (Sigma Aldrich, USA) and 100 mg/kg vancomycin (Sigma Aldrich, USA)) daily through oral gavage for 5 days to eliminate commensal microbiota. Next, antibiotics were removed for 1 day, and mice were incubated daily with 300 μL/mouse ruminal microbiota for three consecutive days and then twice a week for 3 weeks.

To investigate the role of low-grade LPS on the development of mastitis, mice were treated with LPS (2 μg/mouse or 20 μg/mouse daily, n = 5/group, E. coli 0111:B4, Sigma Aldrich, USA) intraperitoneally for 10 consecutive days from the first day of delivery according to previous studies [24, 32]. For treatment with cIAP and zanamivir, cIAP (20 U/mL, n = 6/group, Biolegend, CA, USA) or zanamivir (0.5 mg/mL, n = 6/group, Yuanyebio, Shanghai, China) was supplemented in the drinking water immediately following the initial LPS treatment and continued for the duration of the study as indicated [30].

A mouse mastitis model was established by treatment with E. coli according to our previous description [20]. Briefly, E. coli were cultured in LB medium for 12 h to reach the mid-logarithmic phase. RMT-treated mice or conventional lactating mice 7 days after parturition were then separated from the offspring for three hours and injected with E. coli (10⁷ CFU/mL) through the milk ducts using a 100-μL syringe with a 30-gauge blunt needle. To investigate the protective effects of host ALP on E. coli-induced mastitis in mice, lactating mice 7 days after parturition were pretreated with cIAP (75 U/kg, Biolegend, CA, USA) and zanamivir (500 μg/kg, Yuanyebio, Shanghai, China) for 2 h followed by E. coli infection. Twenty-four hours after the E. coli injection, the mammary gland tissues were harvested and stored at − 80 °C until the detection of inflammatory parameters.

All tissue samples used for histological analysis, including the mammary glands, ileum, colon, and liver tissues from RMT mice and the mammary glands from LPS- and E. coli-treated mice, were fixed with 4% paraformaldehyde for more than 48 h. The fixed samples were then embedded in paraffin and sliced into 5-μm slices (five slices per mouse). The prepared sections were further stained with hematoxylin and eosin (H&E) and assessed with an optical microscope (Olympus, Tokyo, Japan). The histological score was determined as previously described [20].

The mammary gland tissues from RMT-, LPS-, and E. coli-treated mice were weighed and homogenized (0.1 mg/mL) using PBS. The prepared homogenates were centrifuged at 12,000×g and 4 °C for 10 min, and the supernatants were collected for tumor necrosis factor (TNF)-α and interleukin (IL)-1β detection by using TNF-α and IL-1β ELISA kits (Biolegend, CA, USA).

The mammary samples from RMT-, LPS-, and E. coli-treated mice were prepared to generate a 10% homogenate using the provided MPO buffer, and MPO activity was detected according to the manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, China).

To investigate the effects of different treatments on E. coli-induced mastitis, the mammary E. coli load was detected as previously described [35]. In brief, mammary samples were harvested aseptically and homogenized at a final concentration of 0.1 mg/mL using sterile PBS. The prepared homogenates were subjected to tenfold serial dilutions until the desired concentrations were reached, and then the diluted homogenates were plated on MacConkey agar (Hopebio, Qindao, China). The number of E. coli was counted, and the absolute quantities were calculated based on the corresponding dilutions after 24 h of growth.

Serum LPS levels from cows and RMT- and LPS-treated mice were detected as previously described [23]. In brief, the sera were collected under pyrogen-free conditions, and serum LPS levels were detected using a LPS ELISA kit according to the manufacturer’s instructions (Lanpaibio, Shanghai, China). Serum ALP levels from cows and RMT-treated mice were measured with an ALP assay kit according to the manufacturer’s instructions (Nanjing Jiancheng Bioengineering Institute, China). Serum cytokines in cows and RMT-treated mice were detected using bovine TNF-α and IL-1β ELISA kits (Lanpaibio, Shanghai, China) and mouse TNF-α and IL-1β ELISA kits (Biolegend, CA, USA).

To test for major mastitis pathogens, including Staphylococcus aureus (S. aureus), Streptococcus, and E. coli, in the milk of healthy and mastitis cows in this study, 50-μL milk samples were plated on tryptone soya broth agar (TSB, Hopebio, Qindao, China), blood agar (Hopebio, Qindao, China), and MacConkey agar (Hopebio, Qindao, China), respectively. After 24 h of incubation, the numbers of bacterial colony-forming units (CFUs) were counted.

A routine blood test was performed to detect the changes in blood immune cells in RMT-treated mice. Blood samples from RMT mice were collected into vacuum tubes with an anticoagulant (ethylenediamine tetraacetic acid), and then blood immune cells were detected with an automated hemocyte analyzer according to the manufacturer’s instructions (HB-7510, Sinnowa, China). Serum biochemical tests for cows and RMT-treated mice were performed as previously described [36]. Briefly, blood samples from cows and RMT-treated mice were collected into vacuum tubes with a coagulation activator (thrombin). Biochemical analysis was performed with an automatic analyzer (CS-T240, Dirui Industrial Co., Ltd., China) using the commercial veterinary kit (DiAvTest, Russia).

Fecal samples from the RMT treatment groups were weighed and prepared as 10% homogenates using PBS and then centrifuged at 3000×g and 4 °C for 10 min. The supernatants were collected, and fecal lipocalin-2 levels were detected by ELISA according to the manufacturer’s instructions (Lanpaibio, Shanghai, China).

Paraffin sections of the colon were prepared as mentioned above. The prepared paraffin sections were dewaxed with xylene and serial alcohol solutions as previously described [28]. The sections were further stained using an Alcian Blue Stain Kit according to the manufacturer’s instructions (Solarbio, Beijing, China).

All samples used for immunochemistry were prepared into 5-μm sections and dewaxed as mentioned above, and immunochemistry was performed according to our previous study [20, 37]. Briefly, the dewaxed colon sections were subjected to antigen retrieval using sodium citrate and washed with PBS. The prepared sections were further incubated with endogenous peroxidase blockers (SAP (Mouse/Rabbit) IHC Kit, Maixin Biotechnology, Fuzhou, China) for 30 min and normal nonimmune goat serum (SAP (Mouse/Rabbit) IHC Kit, Maixin Biotechnology, Fuzhou, China) for 30 min at room temperature. The sections were further incubated with mucin-2 antibody (1:200, diluted with 5% goat serum, Affinity Biosciences, USA) overnight at 4 °C, and then the secondary antibody (goat-anti rabbit IgG) was added for 30 min of incubation at room temperature. Next, the sections were incubated with horseradish peroxidase (HRP) (SAP (Mouse/Rabbit) IHC Kit, Maixin Biotechnology, Fuzhou, China) for 20 min at room temperature and washed with PBS. After treatment with DAB color-developing fluid (Maixin Biotechnology, Fuzhou, China), the sections were stained with hematoxylin for 3 min and then differentiated with hydrochloric acid alcohol and ammonium hydroxide. After dehydration, the sections were sealed with neutral resin.

Ruminal samples from cows (n = 16, 8 cows from each group of the healthy and mastitis groups) and fecal samples from RMT-treated mice (n = 21, 7 mice from each of the control, H-RMT, and M-RMT groups) were used for bacterial DNA extraction and microbial analysis. Total genomic DNA from the ruminal and fecal samples was extracted using the CTAB method. DNA concentration and purity were monitored on 1% agarose gels. DNA was diluted to 1 ng/μL using sterile water. The 16S rRNA genes of the V4 region were amplified using a 515F-806R primer with the barcode. All PCRs were carried out with 15 μL of Phusion® High-Fidelity PCR Master Mix (New England Biolabs), 2 μM forward and reverse primers, and approximately 10 ng template of DNA. Thermal cycling consisted of initial denaturation at 98 °C for 1 min, followed by 30 cycles of denaturation at 98 °C for 10 s, annealing at 50 °C for 30 s, and elongation at 72 °C for 30 s. Finally, the samples were incubated at 72 °C for 5 min. The same volume of 1X loading buffer (containing SYB green) was mixed with the PCR products, and electrophoresis was performed on a 2% agarose gel for detection. Then, the PCR products were purified with a Qiagen Gel Extraction Kit (Qiagen, Germany). Sequencing libraries were generated using the TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) following the manufacturer’s recommendations, and index codes were added. Library quality was assessed on a Qubit@ 2.0 Fluorometer (Thermo Scientific) and Agilent Bioanalyzer 2100 system. Finally, the library was sequenced on an Illumina NovaSeq platform, and 250-bp paired-end reads were generated.

Alpha diversity was performed to analyze the complexity of taxa diversity, including observed species, Chao1, ace, and PD_whole tree index. Principal coordinate analysis (PCoA) was performed to identify the microbial structures in the ruminal samples and fecal samples of the RMT mice. Linear discriminant analysis effect size (LEfSe) was performed to identify bacterial taxa that were differentially enriched in the different cows and RMT mice treatment groups [38]. Spearman correlation analysis was performed to analyze the relationship between different ruminal or gut microbial taxa and host parameters using Wekemo Bioincloud (https://www.bioincloud.tech↗). Tax4Fun analysis was performed to identify bacterial functions that were enriched in cows and RMT mice [39].

Total RNA was extracted from colonic tissues from RMT-treated mice (n = 7 per group) using TRIzol (Invitrogen, Carlsbad, CA, USA), and quantitative RT-PCR was performed as described previously [20]. In brief, total RNA was extracted and prepared for cDNA. Furthermore, quantitative RT-PCR was performed using TransStart Tip Green qPCR SuperMix (TransGen Biotech, Beijing, China) and then FastStart Universal SYBR Green Master Mix (ROX) (Roche, Switzerland, Basel) in a Step One Plus apparatus (Applied Biosystems, Foster City, CA, USA). The 2^−ΔΔCt method was utilized to determine the gene expression. GAPDH was used as an endogenous control, and all data were normalized to the control group. The primers used in this study are detailed in Additional file 1: Table S1.

Total protein extraction from mammary and colonic tissues and western blotting were performed according to previous studies [19, 20]. In brief, mammary and colonic samples were weighed and homogenized using tissue protein extract (Thermo Fisher Scientific, USA). After centrifugation at 12,000×g and 4 °C for 10 min, the supernatants were harvested, and protein concentration determination was determined using a BCA Protein Assay Kit (Thermo Fisher Scientific, USA). SDS-PAGE (12% or 15%) was performed to separate the targeted proteins according to molecular size. The targeted proteins were transferred to 0. 45-μm PVDF membranes and treated with 5% skim milk for 3 h at room temperature. Furthermore, specific primary antibodies, including p-p65, p-65, p-IκB, IκB, STING, p-TBK1, TBK1, p-IRF3, IRF3, occludin, ZO-1, and claudin-3 from Affinity Biosciences (OH, USA); TLR4, cGAS, NLRP3, ASC, and IL-1β from Cell Signaling Technology (Boston, USA); and β-actin from Immnoway Biotechnology (USA), were added for incubation at 4 °C overnight. After incubating with secondary antibodies (goat anti-rabbit IgG or rabbit anti-mouse IgG, Immnoway Biotechnology, USA) for 2 h at room temperature, the proteins were detected using an ECL plus western blotting detection system.

GraphPad Prism 8 (San Diego, CA, USA) was used for statistical analysis, and the data are expressed as the mean ± SD or represented as a boxplot. Significant differences between the two groups were evaluated by the Mann-Whitney U test (nonparametric) or two-tailed unpaired Student’s t test (parametric). For comparisons among more than two groups, a one-way analysis of variance (ANOVA) and a post hoc Tukey test were performed. For all statistical tests, p < 0.05 indicates significance. Other specific statistical analyses are mentioned in the figure legends.

To study the relevance between ruminal dysbiosis and mastitis in cows, 8 healthy cows (H) and 8 cows with clinical mastitis (M) were employed. Mastitis was diagnosed by increased milk SCC (Fig. 2A). Pathogen tests showed that mastitis cows had a higher load of E. coli in their mammary glands, but no significant differences in other pathogens were found (data not shown). We next performed serum biochemical tests and found that mastitis cows had changed liver function parameters and reduced serum ALP levels (Fig. S1). Furthermore, we confirmed that the M group had increased systemic inflammatory responses, as evidenced by increased contents of the serum proinflammatory cytokines TNF-α and IL-1β and LPS levels (Fig. 2B–D). These results indicated that mastitis cows had increased systemic inflammation.

We further investigated whether these increased systemic inflammation and mastitis parameters are related to the ruminal microbiota. Principal coordinate analysis (PCoA) showed a separation in the ruminal microbiota structure between the H and M groups based on unweighted UniFrac distance (R² = 0.11395, p = 0.002, Fig. 2E). The M group had an increase in the observed species in the ruminal microbiota compared with the H group (Fig. 2F). Other alpha diversity indices, including Chao1 (Fig. 2G), ace, and PD_whole tree index were markedly increased in the M group compared with the H group (Fig. S2A-B). At the phylum level, Proteobacteria was enriched in the M group and Bacteroidetes was reduced (Fig. 2H). At the genus level, the ruminal microbiota composition between the H and M groups was significantly different (Fig. S2C). We further performed LEfSe analysis to identify the different bacterial taxa that were enriched in the H and M groups. The results showed that four bacterial genera, Moraxella, Saccharofermentans, Eubacterium_nodatum_group, and Rikennellaceae_RC9_gut_group, were enriched in the M group compared with the H group (Fig. 2I). Prevotella was the only genus depleted in the M group compared with the H group (Fig. 2I). Furthermore, we showed that the M group had different bacterial functions compared with the H group (Fig. S2D-E). Tax4Fun analysis showed a marked separation of the metabolism pathways between the H and M groups with 26 pathways enriched in the M group, and 9 pathways were depleted (Fig. S2F). Several increases in potentially pathogenic bacterial pathways and aerobic respiration-associated pathways were detected in the M group, including nitrogen respiration, and plant and human pathogens (Fig. S2F). Additionally, the association between different key genera and systemic inflammatory parameters was determined by Spearman correlation analysis. The results showed that microbes enriched in the M group, including Moraxella, Saccharofermentans, Eubacterium_nodatum_group, and Rikennellaceae_RC9_gut_group, were significantly positively correlated with most proinflammatory parameters and negatively correlated with serum ALP levels (Fig. 2J). In contrast, the correlation coefficient for Prevotella enriched in the H group was the opposite (Fig. 2J). These results suggest that cows with mastitis had distinct ruminal microbiota compositions and functions, which are associated with systemic inflammation and mastitis.

To investigate the causative relationship between ruminal dysbiosis and mastitis, RMT from healthy (H-RMT) and mastitis cows (M-RMT) to recipient mice was performed (Fig. 3A). After 26 days of colonization, we found that M-RMT mice had significant mammary damage compared with that of control or H-RMT mice (Fig. 3B, C). To confirm this observation, we next showed that M-RMT increased the production of the proinflammatory cytokines TNF-α and IL-1β in the mammary gland compared with that of the control and H-RMT groups (Fig. 3D, E). In addition, increased MPO activity, a specific marker for neutrophils, was detected in M-RMT mice compared with control or H-FMT mice (Fig. 3F). Moreover, M-RMT impaired the blood-milk barrier by reducing the tight junction (TJ) proteins ZO-1, occludin, and claudin-3 compared with the control and H-RMT groups (Fig. 3G, H). Of note, no significant differences were observed between control and H-FMT mice (Fig. 3A–H), which indicates that ruminal dysbiosis can induce mastitis symptoms in mice.

An increase in inflammatory cytokine production occurred due to the activation of proinflammatory pathways, such as NF-κB and NLRP3, which have been reported to play central roles in the pathogenesis of mastitis [19, 20]. Indeed, we found that M-RMT mice had higher NLRP3 and NF-κB activation than control or H-RMT mice by increasing NLRP3, ASC, IL-1β, p-p65, and p-IκB protein levels (Fig. 4A–G). Mounting evidence has recently shown that cyclic GMP-AMP synthase (cGAS)-mediated stimulator of interferon genes (STING) activation is a key regulator for the activation of the NF-κB and NLRP3 pathways [40–42], in which cGAS recognizes cytoplasmic DNA, such as mitochondrial DNA (mtDNA) caused by cellular stress or pathogen invasion to activate STING, leading to increased phosphorylation of TANK-binding kinase 1 (TBK1) and interferon regulatory factor 3 (IRF3) [42]. We next studied the expression of the cGAS-STING signature and showed that M-RMT mice had increased protein levels of cGAS, STING, p-TBK1, and p-IRF3 compared with those of control and H-FMT mice (Fig. 4H, J–M). Exposure to LPS has been reported to induce mtDNA synthesis through Toll-like receptor (TLR)-4 [42], which is also involved in the activation of the NF-κB and NLRP3 pathways and mastitis pathogenesis. As expected, we found that M-RMT increased the mammary TLR4 level compared with that of the control and H-FMT groups (Fig. 4H, I). These results suggest that the TLR4-cGAS-STING-NF-κB/NLRP3 pathways participate in M-RMT-induced mastitis in mice.

We next investigated the effects of RMT on systemic and mucosal inflammation. We first determined the significant histological changes in the ileum, colon, and liver in the M-RMT group compared with the control and H-FMT groups (Fig. 5A, B). Consistently, M-FMT markedly increased the expression of the proinflammatory gene, including TNF-α, IL-1β, IL-6, and CCL2, in the colon compared with the control and H-RMT groups (Fig. 5C). In addition, we found that M-RMT increased the lipocalin-2 level, a marker of mucosal inflammation [24], in the feces compared with those in control and H-RMT mice (Fig. 5D). We next showed that M-RMT, but not H-RMT, increased the monocyte and lymphocyte proportions in the blood (Fig. S3A), which was accompanied by distinct serum biochemical characteristics (Fig. S3B). Moreover, we found that M-RMT mice had increased serum TNF-α and IL-1β levels compared with those in control and H-RMT mice (Fig. 5E, F). In addition to these observations, M-RMT also increased serum LPS levels compared with the control and H-RMT groups (Fig. 5G). These results indicated that RMT from mastitis cows induced mucosal and systemic inflammation in mice.

Increased mucosal inflammation is known to destroy the gut barrier [23, 28], which orchestrates the first line of defense to protect the host against commensal or pathogenic microorganisms. We showed that M-RMT reduced the number of colonic goblet cells, which are responsible for the production of mucin [43], compared with the control and H-RMT groups (Fig. 5H), Consistently, M-RMT mice had decreased mucin-2 levels in the colon compared with control and H-RMT mice (Fig. 5H, I). Furthermore, we showed a significant decrease in the expression of genes related to TJs, such as Tip1, occludin, and claudin-3, in M-RMT mice compared with control and H-RMT mice (Fig. 5J). Likewise, M-RMT reduced ZO-1, occludin, and claudin-3 protein levels in the colon compared with the control and H-RMT groups (Fig. 5K, L). These data indicate that RMT from mastitis cows to recipient mice induced gut barrier disruption, resulting in endotoxemia and systemic immune balance.

We next investigated the changes in the gut microbiota of recipient mice. PCoA identified a marked separation in the gut microbiota compositions between the H-RMT and M-RMT groups based on the unweighted UniFrac distance (R² = 0.13452, p = 0.025, Fig. 6A). Similar to the donor cows, mice in the M-RMT group had an increase in the observed species in the gut microbiota compared with those in the H-RMT group (Fig. 6B). Other alpha-diversity indices, including Chao1 (Fig. 6C), PD_whole tree and ace index (Fig. S4A-B), were also markedly increased in the M-RMT group compared with the H-RMT group. At the phylum level, M-RMT increased Proteobacteria abundance and reduced Bacteroidetes abundance (Fig. 6D), which was similar to the differences between the H and M groups (Fig. 2H). Additionally, M-RMT reduced the abundance of Firmicutes in the gut microbiota of recipient mice (Fig. 6D). The difference in the microbiota composition between the H-RMT and M-RMT groups at the genus level was also significantly different (Fig. 6E). LEfSe showed that four bacterial genera, Escherichia_Shigella, Neisseria, Agathobacter, and Atopostipes, were enriched in the M-RMT group compared with the H-RMT group (Fig. 6F), while Ruminococcus_gnavus_group was depleted in the M-RMT group compared with the H-RMT group (Fig. 6F). Consistent with the donor, we found that the M-RMT group had distinct bacterial functions compared with the H-RMT group, as 23 pathways were enriched in the M-RMT group, and 12 were depleted (Fig. S4C-E). Of note, the M-RMT group was enriched in several potentially pathogenic bacterial pathways and aerobic respiration-associated pathways that were also detected in the M group (Fig. S2F). Moreover, Spearman correlation analysis showed that microbes enriched in the M-RMT group, including Escherichia_Shigella, Neisseria, Agathobacter, and Atopostipes, were significantly correlated with most of the proinflammatory parameters in the mammary gland, serum, and mucosa (Fig. 6G). These results indicate that M-RMT induced the differences in gut dysbiosis in recipient mice, which may be responsible for the development of systemic inflammation and mastitis.

Considering the increase in serum LPS in the M group cows and M-RMT recipient mice, as well as the enriched gram-negative opportunistic pathogens, Moraxella, Escherichia_Shigella, Neisseria, and Agathobacter in the M or M-RMT group, we then investigated the role of gut dysbiosis-derived low-grade endotoxemia in the pathogenesis of mastitis. Lactating mice were recurrently exposed to different doses of LPS intraperitoneally to mirror low-grade endotoxemia and serum LPS increased in a dose-dependent manner (Fig. 7A). Histological analysis showed that recurrent LPS exposure induced significant mammary injury compared with control mice (Fig. 7B, C). Consistently, recurrent LPS exposure increased the inflammatory marker levels, including mammary TNF-α, IL-1β, and MPO activity (Fig. 7D–F). In addition, recurrent exposure to LPS decreased the levels of the mammary TJ proteins ZO-1, occludin, and claudin-3 (Fig. 7G–J), indicating impaired blood-milk barrier integrity. Moreover, recurrent LPS exposure activated the TLR4-cGAS-STING-NF-κB/NLRP3 pathways, as evidenced by increased TLR4, cGAS, STING, p-TBK1, p-IRF3, p-p65, p-IκB, NLRP3, ASC, and IL-1β contents, in a dose-dependent manner in the mammary gland compared with the control group (Fig. 7G, K–T). These results suggest that gut dysbiosis-induced low-grade endotoxemia can trigger the development of mastitis in mice.

Single exposure to low-dose LPS was not enough to induce marked inflammatory responses, since the host had developed an orchestrated defense mechanism for LPS detoxification, such as ALP, which is widely distributed in the gut, liver, and mammary glands [30–32, 44]. Given the reduction in ALP in mastitis cows and M-RMT-recipient mice (Figs. S1 and S3B), we next investigated whether gut dysbiosis-derived low-grade systemic endotoxemia facilitates the development of mastitis by degrading ALP. We first showed that serum LPS levels were negatively correlated with serum ALP levels in cows and recipient mice (Fig. 8A, B). To confirm the role of ALP in protecting the host against systemic endotoxemia-induced mastitis, cIAP was applied during LPS treatment. Indeed, treatment with cIAP alleviated systemic endotoxemia-induced mammary injury and proinflammatory parameters compared with those in the LPS group (Fig. 8C–G). Additionally, treatment with cIAP improved the LPS-induced decrease in TJ proteins (Fig. 8H–K) and inhibited the activation of the TLR4-cGAS-STING-NF-κB/NLRP3 pathways caused by recurrent LPS exposure (Fig. 8H, L–R). These results suggest that host ALP reduction facilitates low-grade systemic endotoxemia-triggered mastitis in mice.

We next studied how gut dysbiosis-derived LPS reduces host ALP. TLR4-mediated sialylation through neuraminidase (Neu) has been shown to degrade ALP [30–32]. We therefore detected the gene expression of mammary Neu1-4 in RMT recipient mice and found that M-RMT mice had increased Neu3 levels in the mammary gland compared with control and H-RMT mice (Fig. 9A). Consistently, LPS treatment also increased Neu3 level in the mammary gland compared with the control group (Fig. 9B). We therefore investigated the role of Neu3 in the development of mastitis by administering zanamivir, a specific Neu inhibitor [30, 32], to mice. The results showed that zanamivir treatment ameliorated systemic endotoxemia-induced mastitis (Fig. 9C–G). Additionally, zanamivir treatment restored the expression of TJ proteins and reduced the activation of the TLR4-cGAS-STING-NF-κB/NLRP3 pathways caused by recurrent LPS exposure (Fig. 9H–R). These data indicate that low-grade systemic endotoxemia derived from gut dysbiosis can induce mastitis through the depletion of host anti-inflammatory ALP by the LPS-TLR4-Neu signature.

Considering the increased E. coli burden detected in mastitis cows in this study, we next studied whether this increase was attributed to gut dysbiosis. As expected, we found that E. coli-treated M-RMT mice developed more severe mastitis than E. coli-treated control and H-RMT mice (Fig. 10A, B). Similarly, E. coli-treated M-RMT mice had a higher E. coli burden in the mammary glands than E. coli-treated control and H-RMT mice (Fig. 10C). Given the protective role of ALP in recurrent LPS exposure-induced mastitis, we then studied whether a reduction in host ALP caused by increased Neu aggravates E. coli-induced mastitis. Indeed, treatment with ALP and zanamivir ameliorated E. coli-induced mastitis, as shown by improved mammary injury and reduced E. coli load in the mammary glands compared with the E. coli treatment group (Fig. 10D–F). Moreover, ALP and zanamivir reduced mammary proinflammatory markers caused by E. coli (Fig. 10G–I). These results suggest that gut dysbiosis aggravates E. coli-induced mastitis by reducing ALP via increasing Neu in mice.

The animal experiments were approved by Jilin University (Changchun, China).

Not applicable.

The authors declare that they have no competing interests.

All data generated during the current study are included in this article (and its supplementary information files). The 16S rRNA gene sequencing data in the present study are available in the NCBI Sequence Read Archive (SRA) repository under accession number PRJNA815921.