A total of 60 female Balb/c mice, aged 6 weeks, were obtained and housed in a sterile rodent facility at the Laboratory Animal Center of Ningbo University. All procedures for experiments with animals followed the rules and regulations set by the Animal Care and Use Committee at Ningbo University.

Ethical approval for the study was granted under Approval No.11781.

The experimental mice were split into six groups, each containing 10 mice. Two groups received a daily oral dose of 25 mg/kg Qipian® for 90 days. The protocol outlined in Figure 1A was followed to establish the asthmatic model. On day 91, three groups of mice—including with one group administrated Qipian®—were underwent sensitization through intraperitoneal injection of a gel solution with ratio of mass of OVA and Al(OH)3 of 20 μg: 2 mg in 0.9% NaCl solution, as previously described, with minor modifications (Figure 1A) (18). Subsequently, 3% OVA in normal saline solution was administered via nasal drip once daily from day 112 to day 116 using a 200 μL pipette. The other three groups were given injections of PBS, administered intraperitoneally and intranasally, as control groups. The dose selection for both Qipian® and the dexamethasone (DEX) was in accordance with our previous work (7). Among the groups, the first group was the PBS control mice. The second and fifth groups received a daily oral dose of 25 mg/kg Qipian® (lasting for 3 months) with the OVA-induced asthma model and PBS control, respectively. The fourth group was the OVA-induced asthma group treated with PBS. The third and sixth groups were injected with 1.0 mg/kg DEX daily from day 112 to day 116 before OVA challenge and PBS control, respectively (Figure 1A). After a 24-h interval, all mice were euthanized with a high concentration of isoflurane (5% in oxygen or medical air) for a minimum of 2 min after cessation of breathing. Samples of blood, lung tissues, and bronchoalveolar lavage fluid (BALF) were collected for further analysis. Meanwhile, mouse feces samples from every group were collected and stored for further 16rDNA sequencing. All fecal samples were snap-frozen and stored at −80°C until DNA extraction. The storage period did not exceed 1 week, and samples underwent no freeze–thaw cycles to ensure microbiome integrity.

The study recorded the asthmatic behavior scores of each mouse based on the number of nasal scratching, sneezing frequency, and degree of wheezing after nasal dripping. The criteria followed previous reports (Table 1) (7, 19).

After the mice were euthanized, bronchoalveolar lavage was performed by instilling 1 mL of PBS into the lungs three times, following established protocols (20). The resulting BALF was collected and subsequently centrifuged. The pellet obtained from centrifugation was resuspended in ice-cold PBS for further analysis. The supernatant was utilized to assess changes in cytokines (IL-4, IL-5, IL-13, IL-10, IL-33, and IFNγ) and immunoglobulins (Igs, IgG, IgA, IgE, and OVA-IgE), while the pellet was employed to determine the proportion of different immune cells. The cell count in each sample was measured using the hematology analyzer XS 500i (Sysmex, Kobe, Japan).

Serum preparation: Blood samples were collected and incubated at 4°C overnight to allow complete clotting. Serum was separated by centrifugation at 12,000 × g for 10 min at 4°C, aliquoted, and stored at −80°C until analysis of immunoglobulin levels (total IgG, IgA, IgE, and OVA-specific IgE). Lung tissue homogenization: For tachykinin extraction, lung tissues were weighed and homogenized on ice in RIPA lysis buffer (1:10 w/v) supplemented with 1× Halt™ Protease Inhibitor Cocktail (Thermo Fisher, #78430) using a TissueLyser II (Qiagen) for 2 min at 25 Hz (7). The homogenates were centrifuged at 10,000× g for 15 min at 4°C. The resulting supernatant was collected, and its total protein concentration was quantified using the Pierce™ BCA Protein Assay Kit (Thermo Fisher, #23225) according to the manufacturer's instructions. Tachykinin [SP, neurokinin A (NKA), neurokinin B (NKB)] levels were then measured in the normalized supernatants (21, 22).

Specific antibodies were applied to 96-well plates and left to incubate overnight at 4°C. Following that, the specimens were placed onto the dishes and left to incubate for a duration of 2 h at a temperature of 37°C. After that, secondary antibodies conjugated with horseradish peroxidase were introduced and left to incubate for an hour. To observe the enzyme's reaction, a solution of tetramethylbenzidine substrate was added and the change in color was identified using a SpectraMax Paradigm multimode reader from Molecular Devices, CA, USA. The plates were washed three times following each step. Prior to quantification, absorbance at 450/595 nm was determined after acidification with 0.18 M H₂SO₄. Details of ELISA kits are provided in Table 2.

The lateral lobe of the left lung were embedded in a sagittal plane orientation in paraffin, sectioned at a thickness of 4–5 µm, and stained with hematoxylin–eosin or periodic acid–Schiff (PAS) according to standard protocols (23). Slides were examined under a light microscope to determine the inflammatory cell infiltrates, as well as mucus hypersecretion. Lung inflammation severity was measured on a scale of 0–3, as described in a previous study (24). Mucus production was assessed visually in 10 randomly chosen microscopic fields, with a scoring system ranging from 0 to 4 based on PAS staining intensity. A score of 0 meant no visible PAS stain, 0.5 indicated a low level of staining, 1 represented about a quarter of the bronchial space filled with mucus, 2 indicated half, 3 indicated three-quarters, and 4 indicated full mucus coverage (7). Each group was scored on a minimum of three non-serial sections per mouse to ensure the stability of the experimental results. All semiquantitative histological analyses (inflammation and mucus scores) were conducted by a researcher blinded to sample identity and treatment group. For each mouse, a minimum of two non-serial lung sections were scored, and the average score per animal (biological replicate, n = 5 mice/group) was used for statistical analysis.

The methodology described here outlines a rigorous process for preparing lung tissues for downstream analysis. Lung tissues were fragmented followed by digestion using Dulbecco's Modified Eagle Medium (DMEM) digestion media to release individual cells. The resulting suspension was passed through a nylon mesh strainer to eliminate debris and aggregates, yielding single-cell suspensions. RBC lysis buffer further ensures purity by removing residual red blood cells. Cells were then washed with DMEM medium supplemented with fetal bovine serum to help maintain cell viability. Detection of alterations in Treg cells and CD103⁺ dendritic cells (DCs) after treatment was made possible by staining using specific biomarkers.

Single-cell suspensions were prepared at a density of 2 × 10⁶ cells/mL. To mitigate non-specific antibody binding, cells were blocked with anti-CD16/32 Fc receptor antibody (eBioscience) for 15 min at 4°C. Surface staining was performed using fluorochrome-conjugated antibodies against I-A/I-E (M5/114.15.2), CD11b (M1/70), CD11c (N418), and CD103 (2E7), as previously outlined (25). For regulatory T cell (Treg) analysis, cells were stained with anti-CD4-FITC and anti-CD25-APC (BD Biosciences), followed by fixation, permeabilization (eBioscience kit), and intracellular staining with anti-Foxp3-PE. Fluorescence-minus-one and isotype controls were included in each experiment to accurately set gates and determine non-specific background. Compensation was applied using single-stain controls. Cells were analyzed on a Beckman Coulter CytoFLEX™ (CytoFLEX). Standardized gating strategies were employed: Cells were gated on live/singlets → lymphocytes (for Tregs) or live/singlets → CD11c + (for DCs) → subsequent markers. For each biological replicate (n = 3 mice/group), data from a minimum of 50,000 live, single-cell events were collected and analyzed using FlowJo v10.8.1 (7).

Fecal samples were collected, immediately flash-frozen in liquid nitrogen, and stored at −80°C until processing. Microbial genomic DNA was extracted using the HiPure Stool DNA Kit (Magen) according to the manufacturer's instructions. Negative controls (sterile water) were included during extraction and library preparation to monitor contamination. The V3–V4 hypervariable regions of the bacterial 16S rRNA gene were amplified with barcoded primers (341F/806R) and sequenced on an Illumina NovaSeq 6000 platform (LC-Bio, China) using 2 × 250 bp paired-end chemistry.

Bioinformatic processing: Raw paired-end reads were quality-filtered using Fastp (v0.23.2) to remove low-quality bases (Q-score <20) and adapters. Amplicon sequence variants (ASVs) were inferred using the DADA2 pipeline in QIIME2 (v2022.8), which includes error correction, chimera removal, and taxonomic assignment against the SILVA 138 database. The final ASV table was rarefied to a depth of 10,000 sequences per sample for all downstream analyses. Alpha diversity (Shannon, Simpson indices) and beta diversity (Bray–Curtis dissimilarity) were calculated (26). Principal coordinates analysis (PCoA) was performed and visualized using the ggplot2 package in R (27). Differential abundance was assessed using linear discriminant analysis effect size (LEfSe) with a logarithmic LDA score threshold of >2.0 to identify biomarker taxa and p-values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) method (28, 29). Venn diagrams were generated to visualize shared ASVs. Predictive functional profiling of the microbiota was performed using Tax4Fun2, mapping 16S rRNA gene sequences to KEGG orthologs. Notably, these functional predictions are inferred from taxonomic data and represent putative metabolic potential.

Statistical analysis: Alpha and beta diversity metrics were calculated in QIIME2. Group differences in beta diversity were tested with PERMANOVA (999 permutations) using the vegan package in R. Differential ASV abundance was determined using the ANCOM-BC package with FDR correction. LEFSe was used for biomarker discovery (LDA score >2.0, FDR-adjusted p < 0.05). Functional potential was predicted from 16S data using Tax4Fun2 (30).

All statistical analyses were performed using GraphPad Prism 7.0 software (La Jolla, CA, USA). The information is presented as averages with the standard error of the mean. Student's t-test was used for data that followed a normal distribution, whereas the Mann–Whitney U test was employed for data that did not follow a normal distribution. Lung inflammation scores and BALF cell counts were analyzed using two-way analysis of variance with Sidak's multiple comparisons test. Statistical significance was determined as a P value of less than 0.05. Selection of the appropriate statistical techniques depended on the characteristics of the data to ensure precise group comparisons and the validity of experimental results.

OVA sensitization and challenge significantly induced airway inflammation, characterized by eosinophil infiltration and mucus accumulation in lung tissues. LTQA significantly alleviated these pathological features, as indicated by reduced inflammatory cell infiltration in H&E-stained sections and decreased mucus production (Figures 1B–D). Histological examination demonstrated that LTQA suppressed airway inflammation to a degree comparable to DEX. Furthermore, cytological analysis of BALF revealed that LTQA markedly reduced the OVA-induced increase in eosinophil count (Figures 1F,J).

The humoral immune response plays a crucial role in the onset and progression of asthma, involving activation of B cells and DCs to produce immunoglobulins (Igs) such as IgE and IgG upon exposure to exogenous antigens. In our model, OVA sensitization and challenge resulted in a significant elevation of total IgE (Figures 2A,E), OVA-specific IgE (Figures 2B,F), and IgG (Figures 2D,H) in both serum and BALF. LTQA significantly reduced the levels of these Igs, indicating its suppressive effect on OVA-induced humoral response (Figures 2A,B,D–F,H). In contrast to previous findings on short-term treatment with Qipian® (7), LTQA showed minimal inhibitory effect on IgA secretion in serum and BALF (Figures 2C). Notably, it increased IgA levels in BALF following OVA challenge (Figure 2G). These findings suggest that LTQA differentially modulates systemic and mucosal humoral immunity.

CD4⁺ T cells, particularly Th2 cells, are essential effector cells that orchestrate the immune response in asthma (7). We therefore examined the impact of LTQA on the secretion of Th1/Th2 cytokines in BALF. The results demonstrated that LTQA effectively mitigated the OVA-induced elevation of Th2 cytokines, including interleukin-4 (IL-4), IL-5, and IL-13 (Figures 3A–C). Conversely, LTQA increased concentrations of Th1 cytokine IFN-γ and the regulatory cytokine IL-10 (Figures 3D,E). IL-5 plays a significant role in airway inflammation by regulating the recruitment, activation, and survival of eosinophils, which release various pro-inflammatory mediators (31). The suppressive effect of LTQA on IL-5 was consistent with the observed reduction in BALF eosinophils (Figure 1F). Furthermore, as the IL-33/ST2 pathway is known to drive type 2 immunity and allergic pathology (32, 33), LTQA also reduced the IL-33 release compared to the OVA model group (Figure 3F).

Regulatory T (Treg) cells, characterized by the expression of CD4, CD25, and Foxp3, are known to be pivotal in the pathogenesis of asthma (34). To investigate the immunomodulatory potential of LTQA, we assessed its impact on CD4⁺ CD25⁺ Foxp3⁺ Treg cells within lung tissue (Figure 4A). In non-asthmatic mice, LTQA treatment was associated with a significant increase in Treg cells compared to the PBS control (Qipian® group: 30.82% ± 4.36% vs. PBS group: 10.75% ± 1.41%, p < 0.01). Conversely, OVA challenge markedly reduced the Treg cell proportion (OVA group: 4.34% ± 0.51% vs. PBS group: 10.75% ± 1.41%, p < 0.001). Notably, LTQA administration to asthmatic mice restored this deficit, resulting in a significantly higher Treg level than in the PBS-treated OVA model (OVA + Qipian® group: 15.05% ± 1.35% vs. OVA + PBS group: 4.34% ± 0.51%, p < 0.05). In contrast, DEX treatment did not significantly alter Treg cell proportions in either healthy or OVA-challenged mice. These findings suggest that the alleviation of airway inflammation by LTQA may be associated with its ability to bolster the Treg cell population.

Furthermore, we examined CD103⁺ DCs, which are critical antigen-presenting cells implicated in priming the differentiation of naïve T cells into Tregs in mice (35, 36). Our data indicate that LTQA treatment significantly increased the frequency of CD11c⁺ I-A/I-E⁺ CD11b⁻ CD103⁺ DCs compared to both PBS and OVA control groups (Figure 4B). Given that this DC subset has been shown to promote the generation of CD4⁺ Foxp3⁺ Treg cells (36), we infer that the concurrent expansion of CD103⁺ DCs and Treg cells observed in our study may be interrelated. This correlative evidence aligns with established literature on the role of CD103⁺ DCs in Treg development. Importantly, CD103⁺ DCs are also known to influence Th1 and Th17 responses in the airways following allergen exposure (37). Thus, the increase in this DC subset may also relate to the observed shift toward a Th1-dominant cytokine profile (Figure 3), contributing to the restoration of Th1/Th2 balance.

Eosinophils can raise airway sensory nerve (ASN) density in asthma (38). When activated by inflammatory mediators from immune cells, ASNs release tachykinins (SP, NKA, and NKB) (39). Our investigation revealed that OVA challenge significantly elevated pulmonary tachykinin levels, an effect that was attenuated by both LTQA and DEX (Figures 5A–C). LTQA pretreatment also inhibited the rise in SP. Concomitantly, both treatments reduced asthma symptom scores (Figure 5D). Collectively, these findings indicate that LTQA's beneficial effects are correlated with reduced tachykinin signaling. This modulation may be a direct effect or, more likely, a secondary consequence of the overall suppression of eosinophilic and Th2 inflammation.

Mounting evidence suggests that the homeostasis and dysbiosis of intestinal flora are associated with protection and pathogenesis in asthmatic patients and mice (40, 41). Some BLs could also protect the host from asthma by promoting timely maturation of the gut microbiome (8). To investigate whether LTQA exhibits its antiasthma activity via modulation of the gut microbiome in the asthma model, bacterial 16S rDNA sequencing was conducted. At the genus level, there were no significant differences observed in the α-diversity of gut microbiota among mice in the six groups (Figures 6A,B) based on the Shannon and Simpson indices. PCoA and analysis of similarities (ANOSIM) of Bray–Curtis distances at the OTU level revealed significant differences in β diversity among the six groups (Figures 6C,D), with distinct bacterial flora distributions observed between the normal mice group and the OVA-induced asthmatic group (Figure 6E, p < 0.05). Among the top 30 genera of the gut microbiota, LTQA increased the abundance of Lactobacillus and Stenotrophomonas and reduced the abundance of Ligilactobacillus and Desulfovibrio, whereas DEX treatment increased the ratio of Ligilactobacillus and Clostridia_UCG−014 and reduced the abundance of Lactobacillus compared to the model group (Figure 6E). Together, these analyses demonstrate an association between LTQA treatment and specific shifts in the gut microbial community, providing correlative support for its proposed systemic effects via the lung–gut axis.

The composition of the gut microbiota was significantly altered by both asthma induction and therapeutic intervention. PCoA revealed clear separation between groups (Figures 7A,B). Differential abundance testing identified specific taxa associated with each condition. Ovalbumin challenge significantly increased the relative abundance of Staphylococcus and decreased Comamonas and certain Clostridiales compared to healthy control (p < 0.05). LTQA treatment significantly modulated the OVA-perturbed microbiota, characterized by an increase in Lactobacillus sp. L-YJ and Streptococcus danieliae and a decrease in several Clostridia-associated taxa. Dexamethasone induced a third unique profile, enriching for Turicibacter and Lachnospiraceae. Furthermore, microbial alpha diversity, measured by ASV richness, was significantly reduced in the asthmatic model (Figure 7C). To infer potential functional consequences, we used Tax4Fun to predict KEGG pathway abundances. The predicted metagenome of OVA-challenged mice was enriched for pathways involved in core metabolic processes (e.g., energy metabolism, transcription), whereas that of LTQA-treated mice was enriched for pathways involving environmental information processing (e.g., membrane transport) (Figure 7D). It is important to note that these are predictions based on 16S rRNA gene sequences. They suggest hypotheses for microbial metabolic shifts that require direct metagenomic or metabolomic validation.