The L. reuteri strain was obtained from China Center for Type Culture Collection (CCTCC). The strain was cultured in deMan Rogosa Sharpe (MRS) broth at 37°C under anaerobic conditions for 12 to 16 h.

Specific-pathogen-free (SPF) BALB/c female mice (6 weeks old, 16–20 g, SiPeiFu Biotechnology Company, Suzhou, China) were housed in an environmentally controlled animal facility. Mice were maintained on a 12-h light/12-h dark cycle to simulate natural circadian rhythms and preserve physiological homeostasis. The temperature indoors ranged from 20 to 26°C, with humidity maintained between 40 and 70% to prevent environmental stress. This animal study protocol was approved by the Ethics Committee of Huadong Medical Institute of Biotechniques (No.2024NKY012).

The AR mice model was established through ovalbumin (OVA) induction. After a week adaptation period, mice were randomly assigned to three groups (n = 6): the negative control (NC) group, the OVA group, and the L. reuteri (LR) group. In brief, OVA-induced mice were sensitized via intraperitoneal injection with a mixture of 100 μg ovalbumin (OVA; Sigma, St. Louis, MO) and 4 mg aluminum hydroxide (Solarbio, Beijing, China) in 200 μL PBS on days 1, 3, 5, 7, 9, 11, and 13. Subsequently, mice were intranasally challenged with 10% OVA (10 μL per nasal cavity) daily from day 15 to 21. During OVA induction, mice in the LR group received daily oral administration of 0.2 mL L. reuteri suspension containing 1 × 10⁹ colony forming units for 3 weeks. The NC group were gavaged with PBS all the time. After the final OVA challenge, nasal sneezing and nose scratching frequency were recorded for 10 min, concurrent with fecal sample collection.

Mice were randomly assigned to five groups (n = 8): the NC group, the OVA group, and three experimental groups receiving different doses of luteolin (5-LO: 5 mg/kg, 10-LO: 10 mg/kg, 20-LO: 20 mg/kg). To induce an allergic response, mice were intraperitoneally injected with a solution containing 100 μg ovalbumin and 4 mg aluminum hydroxide on days 1, 3, 5, 7, 9, 11, and 13. From day 15 to day 28, mice underwent intranasal stimulation with 10% OVA (Dong et al., 2021). During intranasal OVA stimulation, mice in the LO group was administered intraperitoneally luteolin (Merced, California, United States), dissolved in 0.2 mL normal saline, once daily from day 15 to 28. After the final OVA challenge, nose-scratching frequency was observed and recorded for 10 min.

After sacrificing the mice, nasal mucosal tissues were fixed in 4% paraformaldehyde (Beyotime, Shanghai, China) to preserve morphology. The fixed tissues underwent ethanol dehydration and paraffin embedding. Paraffin blocks were sectioned into 5 μm thickness using a Minux S700A automatic rotary slicer (RWD Life Sciences, Shenzhen, China). Sections were dewaxed in xylene, rehydrated in ethanol, and stained with hematoxylin (Biyuntian, Shanghai, China) for 20–30 s, followed by a distilled water rinse. Excess stain was removed with hydrochloric acid ethanol for 5 s (Pician, Shanghai, China), and sections were briefly treated with ammonia for 5 s (Solarbio, Beijing, China) to enhance nuclear clarity. Following eosin staining (20 s), sections were dehydrated, cleared with xylene, and examined under a CX23 microscope at 400x magnification (Olympus, Tokyo, Japan). Microscopic analysis enabled the observation of cell morphology, structure, and nasal mucosal lesions.

ELISA was performed to quantify OVA-sIgE in serum, along with IL-4, IL-5, IL-13, IFN-γ, IL-17, IL-10, RORγt, and Foxp3 levels in both nasal lavage fluid (NALF) and serum samples (Lapuda, Nanjing, China). Samples were preprocessed for uniformity and stability before being adding to 96-well plates coated with specific antibodies. Biotinlabeled secondary antibodies were applied, followed by further incubation and washing steps. A streptavidin-HRP conjugate was added to bind biotin and introduce HRP enzymatic activity. Addition of a chromogenic substrate initiated an HRP-catalyzed oxidation reaction, generating colored products whose absorbance at 450 nm was measured using a microplate reader (Promega, Madison, WI). Antigen concentrations were determined from a standard curve.

Fecal samples were collected and subjected to DNA extraction using a Fast DNA Spin Kit (MP Biomedicals, Santa Ana, CA, United States). The hypervariable region V3-V4 of the bacterial 16S rRNA gene were amplified with primer pairs 338F: 5′- ACTCCTAC GGGAGGCAGCA-3′ and 806R: 5′- GGACTACHVGGGTWTCT AAT-3′. PCR products were checked on agarose gel and purified through the Omega DNA purification kit (Omega Inc., Norcross, GA, United States). The purified PCR products were collected and the paired ends (2 × 250 bp) was performed on the Illumina Novaseq 6,000 platform (Fang et al., 2022). Identification and removal of primer sequences was process by Cutadapt (Bolger et al., 2014). PE reads obtained from previous steps were assembled by USEARCH and followed by chimera removal using UCHIME. The high-quality reads generated from above steps were used in the following analysis. Sequences with similarity >97% were clustered into the same operational taxonomic unit (OTU) by USEARCH and the OTUs conuts less than 2 in all samples were filtered (Edgar, 2013). Clean reads then were conducted on feature classification to output an ASVs (amplicon sequence variants) by dada2, and the ASVs counts less than 2 in all samples were filtered. Taxonomy annotation of the OTUs was performed based on the Naive Bayes classifier in QIIME2 using the SILVA database with a confidence threshold of 70% (Quast et al., 2013). The Alpha diversity was calculated and displayed by the QIIME2 and R software, respectively. Beta diversity was determined to evaluate the degree of similarity of microbial communities from different samples using QIIME. Principal coordinate analysis (PCoA) and heatmaps were used to analyze the beta diversity. Furthermore, we employed linear discriminant analysis effect size (LEfSe) to test the significant taxonomic difference among group. A logarithmic LDA score of 4.0 was set as the threshold for discriminative features (Segata et al., 2011).

This study employed untargeted metabolic analysis using LC/MS (Mir et al., 2015), which included a Waters Acquity I-Class PLUS liquid chromatograph coupled with a Waters Xevo G2-XS QTOF mass spectrometer. A Waters Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 100 mm) was utilized, with mobile phase A composed of 0.1% formic acid in aqueous solution and mobile phase B composed of 0.1% formic acid in acetonitrile, for both positive and negative ion modes. Data acquisition was performed in MSe mode on the Waters Xevo G2-XS QTOF mass spectrometer controlled by MassLynx V4.2 software. Compound classification and pathway information were obtained from databases such as KEGG, HMDB, and lipid maps. Differential analysis was conducted using grouping information, followed by T-test analysis to determine compound significance (p-value).

For targeted metabolism investigation of flavones in supernatants from L. reuteri cultures, metabolites were extracted and analyzed using a Waters ACQUITY I-Class ultra-performance liquid chromatograph equipped with an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm particle size, Waters). Prior to UHPLC–MS/MS analysis, standard solutions of target compounds were introduced (Liu et al., 2013). Multiple parent ion-daughter ion pairs were optimized for MRM parameters, selecting the most intense ion pair for quantitative analysis and others for qualitative analysis of target compounds.

WB was used to assess p-STAT3 (1:1000, ab68153), and STAT3 (1:1000, ab267373), and MAPK signaling pathway-associated proteins including p38 (1:1000, #9212), p-p38 (1:1000, #4511). Nasal mucosa samples were weighed and homogenized, and proteins were extracted using a total protein extraction kit (Beyotime, Shanghai, China). Protein concentration was determined using the BCA assay (Vazyme, Nanjing, China). The samples were then loaded into a 10% SDS-PAGE gel for separation and subsequently transferred to a polyvinylidene fluoride (PVDF) membrane. The membranes were blocked with 5% skim milk at room temperature for 1 h, then incubated. Primary antibodies, diluted in TBST at recommended concentrations, were incubated with the membranes overnight at 4°C. After washing with TBST, the membranes were incubated with secondary antibodies at room temperature for 1 h. Protein bands were visualized using a detection reagent (Thermo Fisher Scientific, Waltham, MA, United States), and Image J software was used to quantify band intensity.

In this study, measurement data described as means ± standard deviation underwent statistical analyses using GraphPad Prism 9.0 (GraphPad Software, La Jolla, CA). For the comparison of two sets of data, an independent sample t-test was used; whereas for the comparison between multiple groups, a one-way analysis of variance (one-way ANOVA) was performed. In all statistical analyses, a p-value less than 0.05 was used as the criterion for statistically significant differences. * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001.

To assess the impact of L. reuteri strains on the pathological manifestations of AR, a model was established through the induction of OVA. The experimental design was depicted in Figure 1A. The OVA group exhibited significant rhinitis symptoms characterized by frequent nasal scratching and sneezing compared to the NC group. However, these symptoms were reduced in the LR group (Figures 1B,C). Compared to the NC group, the OVA-sIgE serum level in the OVA group exhibited a statistically significant elevation (p < 0.0001). Notably, in the LR group, a marked decrease in the OVA-sIgE serum level was observed (Figures 1D, p < 0.0001). Moreover, in order to assess pathological changes in mice, the nasal mucosal tissues were subjected to HE staining. The OVA treatment resulted in a significant inflammatory cell infiltration in the nasal cavity of mice, compared to the NC group. L. reuteri strain treatments reduced inflammatory cell infiltration and tissue edema compared to the OVA group (Figure 1E).

To assess the immunomodulatory impact of L. reuteri administration in mice, we evaluated whether the treatment with L. reuteri modulates the Th1/Th2 and Treg/Th17 immune responses in the AR mice. Obviously, the OVA group exhibited significantly elevated levels of IL-4, IL-5, IL-13 (indicators of Th2 cell activity) in both serum and NALF compared to the NC group (p < 0.0001). In the LR group, these cytokine levels were significantly reduced (Figures 2A–F). In the OVA group, the level of IFN-γ (an indicator of Th1 cell activity) was significantly reduced both in the serum and NALF compared to the NC group, but was not significantly increased in the LR group (Figures 2G,H). The level of IL-17 (an indicator of Th17 cell activity) was significantly increased both in the serum and NALF in the OVA group compared to the NC group and significantly decreased in the serum in the LR group (Figures 2I,J). As depicted in Figures 2K,L, the level of IL-10 (an indicator of Treg cell activity) in the OVA group was significantly decreased compared to the NC group (p < 0.01). Furthermore, a marked elevation of IL-10 was observed in the LR group, reaching a highly significant level (p < 0.0001).

To explore the differences of gut microbiota among the NC, OVA, and LR groups, a comprehensive analysis was conducted using 16S rRNA sequencing analysis on mouse fecal samples. The alpha diversity was assessed using the Shannon index to evaluate species diversity within each group. Compared to the NC group, the Shannon index was reduced in the OVA group. However, there was a notable increase in the Shannon index in the LR group (Figure 3A). The beta diversity analysis was employed to compare the differences in microbial community composition and structure. The PCoA revealed significant differences in microbial community composition between the NC and OVA groups. Additionally, microbial composition also varied between the OVA and LR groups (Figure 3B,). Figure 3C showed the composition of species abundance between the groups at the genus level. The LefSe further evaluated the differences in species abundance between groups to identify statistically significant biomarkers. We identified differential bacteria, such as Ligilactobacillus, Blautia, Parabacteroides, and Akkermansia (Figure 3D). Furthermore, the correlation heatmap was created to assess the relationship between different bacterial genera and changes in Th1 and Th2 cytokine levels between the OVA and LR groups. We found that Prevotellaceae_UCG_001 and Odoribacter were positively correlated with IL-4, IL-5, IL-13 cytokines, whereas a positive correlation between the Ligilactobacillus and IFN-γ cytokines (Figure 3E). The analysis of the relative abundance of individual bacterial genera revealed that supplementation with L. reuteri significantly increased the abundance of Ligilactobacillus (Figures 3F, p < 0.01). This increased abundance of Ligilactobacillus in the LR group was strongly positively associated with Th1 cytokines. In contrast, the genus-level abundance of Prevotellaceae_UCG_001 and Odoribacter was decreased in the LR group (Figures 3G,H). The increased abundance of these bacteria in the OVA group was strongly positively associated with Th2 cytokines. However, the abundance of Akkermansia increased almost exclusively in the LR group (Figure 3I).

To further explore the effective substances of L. reuteri treatment for alleviating AR symptoms, we conducted an untargeted metabolic analysis on mouse fecal samples. The principal component analysis (PCA) was performed to assess the variation between different groups. As shown in Figure 4A, the metabolite profiles of the NC and OVA groups were significantly different, and there was also a notable variation between the OVA and LR groups. Through orthogonal projections to latent structures-discriminant analysis (OPLS-DA), the LR group was identified as a valid model for screening differential metabolites compared to the OVA group (Figure 4B). We created the hierarchical clustering heatmap of the differential metabolites and observed clear differences in flavonoid compounds between the OVA and LR groups, including luteolin, vaccarin, daidzein, pelargonidin, and senegenin (Figure 4C). The abundance of the differential metabolite luteolin between the LR and OVA groups was significantly increased in the LR group (Figures 4D, p < 0.001). Additionally, we performed a pure culture of L. reuteri and conducted flavonoids-targeted metabolic analysis on the supernatant of L. reuteri culture medium, revealing the production of a small amount of luteolin (Figure 4E).

Figure 5A illustrated the effects of LO treatment in a mouse model of AR. The mice in the OVA group exhibited a significantly higher frequency of nose scratching compared to the NC group. However, the frequency of nose scratching was reduced in the 10-LO and 20-LO groups (Figure 5B, p < 0.01). Additionally, mice exposed to OVA showed a significantly elevated serum level of OVA-sIgE in the OVA group. Luteolin treatment effectively ameliorated OVA-induced upregulation of OVA-sIgE, with the most pronounced reduction observed in the 20-LO group (Figure 5C, p < 0.0001). HE staining revealed that OVA treatment caused epithelial cell shedding and increased inflammatory cell infiltration in the nasal mucosa, effects which were mitigated by LO treatment (Figure 5D).

In the OVA group, we observed a significant increase in the concentrations of IL-4, IL-5, and IL-13 in both NALF and serum compared to the NC group (p < 0.0001). In contrast, the levels of these cytokines in the LO-5, LO-10, and LO-20 groups were markedly lower than those in the OVA group (Figures 6A–F). The serum level of IFN-γ was not significantly elevated in the LO-5 and LO-10 groups, but was significantly higher in the LO-20 group compared to the OVA group (Figure 6G). Additionally, the level of IFN-γ in NALF was significantly increased in the 10-LO and LO-20 groups (Figure 6H). Th2 cytokine levels decreased significantly with increasing doses of luteolin, whereas Th1 cytokines increased notably at higher doses. Notably, in the OVA group, IL-17 levels were significantly elevated in both serum and NALF compared to the NC group. This increase was significantly reduced in the LO-20 group (Figures 6I,J). Moreover, IL-10 levels in serum and NALF were significantly higher in the LO-20 group (Figures 6K,L).

RORγt is a major transcription factor for IL-17, while Foxp3 is a key marker of Treg cells. To determine if RORγt promotes Th17 polarization in AR mice, we measured its concentration in NALF by ELISA. In the OVA group, Foxp3 levels were significantly reduced in NALF, whereas they were increased in the 20-LO group (Figure 7A). RORγt levels in the OVA groups were notably higher compared to the NC group, and the increase was significantly reversed in the 20-LO group (Figure 7B). To further investigate the mechanisms, we assessed the expression of the Th17-related STAT3 signaling pathway by measuring p-STAT3 and STAT3 levels. The western blotting analysis revealed that p-STAT3 protein expression in the nasal mucosa was significantly higher in the OVA group compared to the NC group (Figure 7C). LO treatment effectively reduced p-STAT3 protein expression (Figures 7D,E). The p-STAT3/STAT3 ratios in the 20-LO group showed a significant decrease (Figure 7F). This suggests that the OVA-induced AR model presented an activated STAT3 signaling pathway. Next, we examined the inhibitory effects of luteolin on the MAPK signaling pathway activation. Western blot analysis was used to investigate protein expression within the MAPK signaling pathway, which is involved in the inflammatory response. Phosphorylation of p38 was significantly higher in the nasal mucosa of the OVA group compared to the NC group and was notably reduced in the 20-LO group (Figures 7G–I). Additionally, the p-p38/p38 ratios in the 20-LO group showed a significant decrease (Figure 7J).