Study on the Mechanism of Qing-Fei-Shen-Shi Decoction on Asthma Based on Integrated 16S rRNA Sequencing and Untargeted Metabolomics

Qing-Fei-Shen-Shi decoction (QFSS) consists of Prunus armeniaca L., Gypsum Fibrosum, Smilax glabra Roxb., Coix lacryma-jobi L., Benincasa hispida (Thunb.) Cogn., Plantago asiatica L., Pyrrosia lingua (Thunb.) Farw., Houttuynia cordata Thunb., Fritillaria thunbergii Miq., Cicadae Periostracum, and Glycyrrhizae Radix Et Rhizoma Praeparata Cum Melle. QFSS shows significant clinical efficacy in the treatment of asthma. However, the specific mechanism of QFSS on asthma remains unclear. Recently, multiomics techniques are widely used in elucidating the mechanisms of Chinese herbal formulas. The use of multiomics techniques can better illuminate the multicomponents and multitargets of Chinese herbal formulas. In this study, ovalbumin (OVA) was first employed to induce an asthmatic mouse model, followed by a gavage of QFSS. First, we evaluated the therapeutic effects of QFSS on the asthmatic model mice. Second, we investigated the mechanism of QFSS in treating asthma by using an integrated 16S rRNA sequencing technology and untargeted metabolomics. Our results showed that QFSS treatment ameliorated asthma in mice. In addition, QFSS treatment affected the relative abundances of gut microbiota including Lactobacillus, Dubosiella, Lachnospiraceae_NK4A136_group, and Helicobacter. Untargeted metabolomics results showed that QFSS treatment regulated the metabolites such as 2-(acetylamino)-3-[4-(acetylamino) phenyl] acrylic acid, D-raffinose, LysoPC (15 : 1), methyl 10-undecenoate, PE (18 : 1/20 : 4), and D-glucose6-phosphate. These metabolites are associated with arginine and proline metabolism, arginine biosynthesis, pyrimidine metabolism, and glycerophospholipid metabolism. Correlation analysis indicated that arginine and proline metabolism and pyrimidine metabolism metabolic pathways were identified as the common metabolic pathways of 16s rRNA sequencing and untargeted metabolomics. In conclusion, our results showed that QFSS could ameliorate asthma in mice. The possible mechanism of QFSS on asthma may be associated with regulating the gut microbiota and arginine and proline metabolism and pyrimidine metabolism. Our study may be useful for researchers to study the integrative mechanisms of Chinese herbal formulas based on modulating gut microbiota and metabolism.


Introduction
Asthma is a global common chronic infammatory airway disease, which is characterized by increased mucus secretion in the airways mediated by multiple cells, as well as infammatory factors, airfow obstruction, and airway remodeling. Te clinical symptoms usually manifest as wheezing, coughing, and dyspnea. Globally, approximately 300 million people sufer from asthma [1]. In China, epidemiological surveys have shown that the overall prevalence of asthma in people over 20 years of age is 4.2% [2]. Asthma treatment regimens mainly include bronchodilators (such as beta-2 agonists and aminophylline), antiallergic infammatory drugs (such as glucocorticoids and sodium cromoglicate), and immunomodulators. However, most therapeutic drugs have certain side efects, such as nausea and vomiting, when intravenous theophylline is used [3] and cardiovascular side efects when β-adrenergic agonists [4] are used. In addition to the potential side efects, the antiasthmatic medications in current clinical treatments also impose signifcant fnancial strain on patients. Terefore, the development of safe and efective medicines to relieve asthma has become a lucrative research topic.
Traditional Chinese medicine (TCM) has accumulated thousands of years of clinical experience in the treatment of asthma. A clinical randomized multicenter trial found that Ping-Chuan-Yi-Qi (PCYQ) granules signifcantly improved peak expiratory fow rate (PEFR) and reduced serum cytokine levels in asthmatic patients [5]. A meta-analysis showed that the combination of TCM with conventional treatment improved the clinical symptoms of asthma patients with a better safety profle compared to the conventional treatments [6]. Wang et al. summarized the research progress of TCM in the treatment of asthma and discussed in detail the studies of extremely efective components of TCM in regulating immune imbalance in asthma patients [7]. Elucidating the mechanism of TCM for asthma can help the development of new antiasthmatic drugs and promote the modernization of TCM at the same time.
Gut microbiota is a general term for the microbiota that colonize in the intestines. Gut microbiota are diverse, and mainly include Firmicutes, Bacteroides, Proteobacteria, and Actinomycetota, and each phylum is distributed in diferent parts of the intestine in diferent proportions. Tese bacteria play an important role in maintaining the dynamic equilibrium between the internal and external environments. Te concept of the "gut-lung axis" emphasizes the interaction between microorganisms indicated by the epithelium of the gastrointestinal and respiratory tracts [8]. Recent studies have shown that gut microbiota play a key role in the pathogenesis of asthma [9]. Bifdobacteria, Akkermansia, and Faecalibacterium are less abundant and Candida and Rhodotorula are more abundant in the gut of children with asthma compared to that in normal subjects, and this dysbiosis may further aggravate the immune imbalance [10]. In addition, the long-term use of hormones and antibiotics in asthma patients can also aggravate gut dysbiosis, which is not conducive to the recovery of the disease [11]. Regulating gut microbiota may provide new therapeutic ideas for treating asthma. Recuperating lung decoction increases the levels of benefcial bacteria such as Lactobacillus and Bifdobacterium [12]. Gu-Ben-Fang-Xiao decoction can promote T regulatory cell (Treg) diferentiation and thus relieve asthma by regulating gut microbiota [13]. Tuo-Min-Ding-Chuan decoction can afect phenylalanine, ascorbate, and aldarate metabolism by regulating the abundance of Desulfovibrio, Butyricimonas, Prevotella, and other microbiota in the gastrointestinal tract, and glutathione metabolism to treat asthma [14].
Qing-Fei-Shen-Shi decoction (QFSS), which consists of Prunus armeniaca L., Gypsum Fibrosum, Smilax glabra Roxb., Coix lacryma-jobi L., Benincasa hispida (Tunb.) Cogn., Plantago asiatica L., Pyrrosia lingua (Tunb.) Farw., Houttuynia cordata Tunb., Fritillaria thunbergii Miq., Cicadae Periostracum, and Glycyrrhizae Radix Et Rhizoma Praeparata Cum Melle, has been used for the treatment of acute attack asthma in the clinic for more than 30 years. A clinical study showed that QFSS signifcantly improved respiratory function and reduced the frequency of attacks in patients with asthma [15,16]. However, the specifc mechanism of QFSS in the treatment of asthma remains unclear. In this study, ovalbumin (OVA) was frst employed to induce an asthmatic mouse model, followed by a gavage of QFSS. First, we evaluated the therapeutic efects of QFSS on asthmatic mice. Second, we investigated the mechanism of QFSS in treating asthma by integrating 16s rRNA sequencing technology and untargeted metabolomics.

Reagents.
Te detailed information of reagents has been presented in supplementary materials.

Experimental Animals.
Sixty adult male specifcally pathogen-free (SPF) grade BALB/c mice weighing 20 g ± 2 g were provided by Beijing Huafukang Co. Ltd., with Certifcate of Approval No. being SCXK (Beijing) 2019-0008. Te mice were housed in a SPF-grade clean environment, with 5 mice per cage at a temperature of 22°C ± 2°C, a humidity of 50% ± 15%, and a day/night time of 12 h/12 h, and free access to food and water was provided. All animal experiments were performed in accordance with the National Institutes of Health (NIH) guidelines for the care and use of laboratory animals, and all procedures were approved by the Animal Medicine and Animal Protection Ethics Committee of Qingdao University (approval no. QDU-AEC-202282).  Te 60 mice were randomly divided into the control group, model group, DXM  group, LD-QFSS group, MD-QFSS group, and the HD-QFSS  group. After 1 week of acclimatization feeding, mice from  model, DXM, LD-QFSS, MD-QFSS, and HD-QFSS groups received OVA treatment to induce asthma. Meanwhile, mice in the control group were intraperitoneally injected with 0.2 mL saline on days 0, 7, and 14, and received nebulized inhalation of saline for 30 min, once every 2 days for 4 consecutive weeks starting from day 21. Starting from day 21, mice in the control and model groups were given 0.2 mL saline once per day via intragastric administration. Te mice in the DXM group, LD-QFSS group, MD-QFSS group, and the HD-QFSS group were given DXM 2 mg/kg [18], QFSS 13.5 g/kg, QFSS 27 g/kg, and QFSS 54 g/kg once per day for 4 weeks was administered via intragastric administration, respectively ( Figure 1). Clinically, the dosage of QFSS for adult patients (body weight: 70 kg) was 210 g (the total raw materials) once per day. Te dosage for mice was calculated by following the classical pharmacological formula. MD-QFSS (human equivalent dose) = 210 g (the total raw materials)/70 kg (human weight) × 9 (conversion coefcient). Te low dose and high dose were 0.5× and 2× of the middle dose, respectively.

Enhanced Pause (Penh)
Testing. After 4 weeks of QFSS intervention, mice were anesthetized with 1% sodium pentobarbital. Te trachea was fully exposed, and a catheter was inserted after the condition of the mice was stable. One end of the catheter was connected to the trachea and the other end was connected to the EMKA Small Animal Lung Function Test System Ventilator. Te mice in the control group were frst exposed to phosphate-bufered saline (PBS) nebulizer to obtain Penh baseline values, by direct administration of acetylcholine aerosol via the ventilator. Each mouse was then exposed to diferent concentrations of acetylcholine (6.25, 12.5, 25, and 50 g/L) and Penh values were recorded for 3 min at each dose.

Serum and Bronchoalveolar Lavage Fluid (BALF)
Collection. After 4 weeks of QFSS intervention, mice were anesthetized and blood was collected from the abdominal aorta using a syringe, and the collected blood was centrifuged at 3,000 rpm/min for 15 min to collect the serum. Ten, the mice were sacrifced, and the thoracic cavity was opened. Te cervical trachea was exposed layer by layer, and a Lanz incision was performed on the trachea. Te mouse lavage needle was inserted into the lower end of the right main bronchus, and the trachea and mouse lavage needle were ligated by surgical sutures. Ten, the hilum of the left lung was frmly ligated using surgical sutures to ensure that the left lung was in an airtight position. Using a syringe, 1 mL of saline was withdrawn and connected to the lavage needle ligated to the cervical trachea, where saline was slowly injected into the right lung of mice. Ten, the saline was gently withdrawn back to obtain BALF after staying in the alveoli for 15-30 s. Saline was closely observed for exudation during the lavage. Tis injection and withdraw procedures were repeated 3 times, and a total of approximately 2.5 mL of BALF was recovered from each mouse.

Lung Wet-to-Dry (W/D) Weight Ratio.
Te left lung of the mouse that was not lavaged was retained. Te wet mass (W) of the left lung was measured, followed by drying in a constant temperature oven at 80°C for 48 h until the lung weight no longer decreased. At this point, the lung was weighed as the dry weight (D), and the lung W/D ratio was calculated.
2.9. BALF Total Cell Count, Eosinophil Count, and the Total Protein Concentration Assay. Te BALF was centrifuged at 4°C for 10 min at 3,000 rpm/min, and the cell precipitate and supernatant were collected separately. Te total number of cells in the BALF precipitate was enumerated using a cell counting plate. Te eosinophil count in BALF was enumerated using Wright's-Giemsa staining. Te total protein concentration in BALF supernatant was measured using the bicinchoninic acid (BCA) total protein concentration assay kit.

Pathological Staining of Lung Tissues.
After 24 h of fnal stimulation, the mice were sacrifced and the lung tissues of mice from each group were collected, followed by fxation in the formalin solution. Ten, the fxed tissues were embedded in parafn, cut into 3 μm sections, and routinely stained with hematoxylin and eosin (HE) and Masson. Te histopathological changes of the lung tissues of each group were observed under a light microscope, and the infammation levels of the lung tissues in HE staining were scored based on the previous study [19]. Te Masson staining was also scored as described previously [19].

Lung Tissue Biochemical
Tests. 50 mg of lung tissues were weighed and 450 μL of normal saline was added. After that, the solution underwent ultrasonic homogenization before centrifugation at 4°C for 10 min at 3,000 rpm/min. Lung tissue homogenates were prepared by collecting the supernatants of homogenized tissues. Te BCA kit was used to normalize the protein levels in tissue homogenates, which were then used to detect the activities of superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), and malondialdehyde (MDA) levels according to the instructions of the kit. analysis. After the mice were anesthetized and sacrifced, the cecum was isolated by opening the abdominal cavity, and the contents of the end of the cecum were obtained and placed in sterile Eppendorf (EP) tubes and stored instantly in a lowtemperature freezer at −80°C. Under aseptic conditions, 200 mg of fecal samples were weighed and added to 2 mL EP tubes, and genomic DNA was extracted from each sample according to the operating instructions of the Fecal DNA Extraction Kit. Te quality and concentration of the extracted genomic DNA were measured using a Termo Nanodrop 2000 spectrophotometer and 1% agarose gel. DNA was diluted to 1 ng/µL with sterile water, depending on the concentration.

Polymerase Chain Reaction (PCR) Amplifcation and
Sequencing of 16S rRNA Gene. Based on the characteristics of the amplifed 16S region, PCR amplifcation of the hypervariable region of 16S rDNAV3-V4 was performed with the amplifcation primers 338F (5′-ACTCCTACGGGA GGCAGCAG-3′) and 806R (5′-GGAC-TACHVGGGTWTCTAAT-3′). After performing 2% agarose gel electrophoresis to quantify the amplifed products, the Illumina NovaSeq platform was selected for sequencing, and a 250 bp paired-end sequence was obtained. After the raw data were obtained by sequencing, the raw data were assembled and fltered to obtain the clean data. Ten, the entire clean data of all samples were clustered using UParse software (UParsev7.0.1001) and the sequences were clustered into operational taxonomic units (OTUs) with 97% agreement.

Sequencing Data Processing and Analysis.
Diversity analysis was performed on the sequencing data results. Shannon and Simpon indices were used to assess the gut microbiota alpha diversity of each group. Te results of beta diversity analysis were presented by principal coordinate analysis plot (PCoA) analysis, an unconstrained data downscaling analysis that presents similarities and diferences in community composition across sample groups. Te Wilcoxon rank-sum test was used to test for diferences between groups in diversity indices. In addition, OTU was used for species classifcation of OTU by comparing it to the database, and the species abundance of each group of samples was analyzed at the phylum level and genus level. Finally, the analysis of the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States database (PICRUSt) was used to predict the relevant gene pathways that may be afected by each group of diferential microbiotas.

Metabolomics Analysis.
Six animals were randomly selected in control, model, and HD-QFSS groups for the untargeted metabolomics analysis. A 100 mg of lung tissue was added to 500 μL of 80% methanol solution, vortexed and shaken, and left to stand in an ice bath for 5 min. Te tissue was centrifuged at 15, 000g and 4°C for 20 min. Ten, the supernatant was collected and diluted with water to obtain a 53% methanol level. After that, the supernatant was collected and centrifuged for 20 min at 15, 000g and 4°C to obtain a tissue homogenate for untargeted metabolomic analysis using liquid chromatography-mass spectrometry (LC-MS). All samples were obtained in equal amounts individually and mixed; these were used as the quality control (QC) samples. Te specifc chromatographic and mass spectrometric conditions and data processing and analysis were performed according to our previously published article [20,21] (supplementary material (available here)). Spearman's correlation analysis was used to fnd the correlation between therapeutic indicators, diferential metabolites, and changed gut microbiota.

Statistical
Processing. Statistics and analysis were performed using SPSS 20.0 statistical software, and data were expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) followed by Tukey's post-hoc test was used for comparison among groups. A diference of P < 0.05 was considered to be statistically signifcant.

Terapeutic Efects of QFSS on the Asthma Model Mice.
Te Penh values in the model group were signifcantly higher than those in the control group after administering interventions of diferent concentrations of Ach (P < 0.01, respectively). Compared to the model group, Penh values were signifcantly lower in the DXM and HD-QFSS groups after 6.25 mg/mL Ach intervention (P < 0.01 and P < 0.05, respectively); Penh values signifcantly decreased in the DXM, MD-QFSS, and HD-QFSS groups after the 12.5 mg/ mL, 25 mg/mL, and 50 mg/mL Ach interventions (all P < 0.01) ( Table 1). Te lung tissue W/D ratio was signifcantly higher in asthmatic mice compared to that in the control group (P < 0.01). Te ratios were signifcantly lower in the DXM, MD-QFSS, and HD-QFSS treatment groups compared to that in the model group (P < 0.01, P < 0.05, and P < 0.01, respectively; Figure 2(a)). Te total protein content of BALF supernatant was signifcantly increased in the model group compared to that in the control group (P < 0.01), and signifcantly decreased in the DXM, MD-QFSS, and HD-QFSS groups compared to that in the model group (all P < 0.01; Figure 2(b)). Te total cell count in BALF was signifcantly increased in the model group compared to that in the control group (P < 0.01), and the DXM, MD-QFSS, and HD-QFSS groups signifcantly decreased the total cell count in BALF compared to the model group (all P < 0.01; Figure 2(c)). More eosinophils were found in the BALF of the model group compared to that in the control group (P < 0.01), and lower eosinophil counts were found in the BALF of the DXM, MD-QFSS, and HD-QFSS groups compared to that in the model group (all P < 0.01; Figure 2(d)). In addition, serum IgE levels were elevated in the model group compared to those in the control group (P < 0.01), but serum IgE levels were signifcantly lower in the DXM, LD-QFSS, MD-QFSS, and HD-QFSS groups compared to those in the model group (P < 0.01, P < 0.05, P < 0.01, and P < 0.01, respectively; Figure 2(e)).
HE staining showed no pathological changes in the lung tissue of the control group. Disorderly bronchial epithelial cell arrangement, bronchial smooth muscle thickening, and a large number of infammatory cells could be observed in the model group, while DXM and QFSS treatment improved pathological changes in the lungs (Figure 3(a)). Similarly, infammation scores were signifcantly higher in the model group than those of the control group (P < 0.01), and infammation scores were lower in the DXM, MD-QFSS, and HD-QFSS groups than in the model group (P < 0.01, respectively; Figure 3(c)). Masson staining showed a signifcant increase in the collagen deposition in the lungs of the asthmatic model mice compared with mice in the control group. QFSS and DXM treatment reduced the collagen deposition in the lungs of mice infected with asthma ( Figure 3(b)). Likewise, the Masson staining score was higher in the model group than that of the control group (P < 0.01), and the Masson scores were lower in the DXM, MD-QFSS, and HD-QFSS groups (P < 0.01, respectively; Figure 3(d)).

Efects of QFSS on Infammation and Oxidative Stress in
Asthmatic Mice. We measured cytokine levels in BALF to see if QFSS afected the level of infammation in the asthma model mice. In addition, MDA levels, SOD, and GSH-Px activities were measured in lung tissue homogenates to assess the efect of QFSS on oxidative stress. ELISA results showed that the levels of IL-4, IL-5, and IL-13 were increased as expected in the model group compared to those in the control group (P < 0.01), while these cytokines were signifcantly reduced in the DXM group compared to those in      and MDA levels increased in the model group (all P < 0.01), whereas SOD and GSH-Px activities increased and MDA levels decreased in the QFSS high-dose group compared to those in the model group (P < 0.01, P < 0.05, and P < 0.01, respectively; Table 2).
Te abovementioned results showed that the asthma model was established successfully, and that QFSS had a therapeutic efect on asthma, which was most signifcant at a high dose. Terefore, the HD-QFSS group was selected for the subsequent gut microbiota and lung metabolite study.

Efect of QFSS on Gut Microbiota of Asthmatic Mice.
To investigate the changes of QFSS on the gut microbiota of asthmatic mice, we analyzed the fecal microbiota of diferent groups of mice by using 16S rRNA high-throughput sequencing. Te dilution curve in each sample tended to be fat after the number of sequences increased to 10,000. Few novel OTUs can be identifed after the number of sequences increased to 40,000, indicating that the depth of sequencing is reasonable and appropriate (Figure 4(a)). Te abundance and diversity of microbial communities within the samples were analyzed using alpha diversity (Shannon's index and Simpson's index). Te results showed that there were no signifcant diferences in the Shannon index and the Simpson index in each group (Figures 4(b) and 4(c)). Next, we analyzed the composition of microbial communities of diferent samples using beta diversity and evaluated them by PCoA and clustering analysis. Te PCoA results showed that the sample points in the model group were completely separated from the control group, and the sample points in the HD-QFSS group were closer to those in the control group (Figure 4(d)). Clustering analysis also showed similar results (Figure 4(e)).
Te composition of the gut microbiota in each group of samples at the phylum level is shown in Figure 4(f ), with Firmicutes and Bacteroidetes as the dominant taxa. Te Firmicutes/Bacteroidetes (F to B) ratio was signifcantly higher in the model group than that in the control group (P < 0.01), while the F-to-B ratio was signifcantly lower in the HD-QFSS group than that in the model group (P < 0.05, Figure 4(g)). At the genus level, the relative abundance of Lachnospiraceae_NK4A136_group (P < 0.05) and Helicobacter (P < 0.05) was signifcantly higher in the asthmatic mouse model than that in the control group, and the relative abundance of Lactobacillus (P < 0.05), Alloprevotella (P < 0.05), Bacteroidota (P < 0.05), and Dubosiella (P < 0.05) were signifcantly decreased. QFSS signifcantly increased the relative abundance of Lactobacillus (P < 0.01) and Dubosiella (P < 0.05), and decreased the relative abundance of Lachnospiraceae_NK4A136_group (P < 0.05) and Helicobacter (P < 0.05) compared to the model group (Figure 4(h)).
Te functional changes of intestinal microbiota between the control and model groups and between the model and HD-QFSS groups were predicted by PICRUSt analysis. Diferential metabolic pathways (P < 0.05) with the top ten proportions were listed in Figures 4(i) and 4(j). Te common pathways (control vs. model and model vs. HD-QFSS) included glutathione metabolism, pyrimidine metabolism, amino acid metabolism, glycine, serine, threonine metabolism, and arginine and proline metabolism.

Efect of QFSS on Lung Metabolite Levels in Asthmatic
Mice. We further used untargeted metabolomic analysis to explore the efects of QFSS on pulmonary metabolites in asthmatic mice. Te principal component analysis (PCA) score plot showed that the control group could be well diferentiated from the model group, and the model group was well diferentiated from the HD-QFSS group ( Figure 5(a)). QC samples were clustered in PCA plots, indicating a good stability of the metabolomics analysis system [22]. Te partial least squares discriminant analysis (PLS-DA) model was constructed to assess the explanatory power (R 2 ) and predictive power (Q 2 ) of the group and to identify the diferential metabolites. Compared to the control group, R 2 Y � 0.97 and Q 2 Y � 0.56 in the model group ( Figure 5 Te following two criteria were used to screen for diferential metabolites: P < 0.05 and VIP > 1.0 (control vs. model groups, or model vs. HD-QFSS groups). A total of 43 diferential metabolites were screened (Table 3). Compared to the control group, metabolites such as methyl 10-undecenoate, Lfucose, and L-ornithine were signifcantly higher in the model group, and metabolites such as 1-oleoyl-sn-glycero-3-phosphocholine, 2-deoxyuridine, and 3-amino-4 methylpentanoic acid and other metabolites were signifcantly decreased. Compared to the model group, metabolites such as 2-(acetylamino)-3-[4-(acetylamino) phenyl] acrylic acid, D-rafnose, and LysoPC (15 :1) were signifcantly higher in the HD-QFSS group, and metabolites such as methyl 10-undecenoate, PE (18 :1/20 : 4), D-glucose6-phosphate, and other metabolites were signifcantly lower in the HD-QFSS group. In addition, we used MetaboAnalyst software to analyze the metabolic pathways of the 43 diferential metabolites obtained (FC > 1.2 or FC < 0.8). Compared to the control group, the metabolic pathways that were altered in the model group mainly included arginine and proline metabolism, arginine biosynthesis, starch and sucrose metabolism, pyrimidine metabolism, and glycerophospholipid metabolism; while the metabolic pathways afected by QFSS mainly included arginine and proline metabolism, arginine biosynthesis, pyrimidine metabolism, and glycerophospholipid metabolism ( Figure 5(f)). Te common pathways between PICRUSt analysis of 16S rRNA sequencing and pathway analysis of untargeted metabolomics included arginine and proline metabolism and pyrimidine metabolism pathways, indicating that QFSS may regulate these metabolic pathways to treat asthma by afecting the gut microbiota. Terefore, these pathways were discussed in detail.

Spearmen's Analysis of the Correlations between the Results of 16S rRNA Sequencing, Terapeutic Efects, and
Untargeted Metabolomics. We further used Spearmen's analysis in order to study the relationship of diferential gut microbiota with the therapeutic indicators and changed metabolites. As shown in Figure 6(a), Lachnospir-aceae_NK4A136_group, Helicobacter, Alloprevotella, Bacteroidota, and Dubosiella exhibited signifcant correlations with most of the therapeutic indicators. Besides, Lachno-spiraceae_NK4A136_group, Helicobacter, Dubosiella, and Lactobacillus exhibited signifcant correlations with many diferential metabolites (Figure 6(b)).

Discussion
In the present study, we constructed an asthma model mouse using OVA. Te results showed that the model group mice had reduced lung function and lung histopathological examination revealed a large number of infammatory cell infltrates in lung tissue. A previous study has shown that the W/D ratio of the lung was increased in asthmatic model mice [23]. Te increase in lung W/D ratio could refect the increase in lung permeability, and the lung permeability is impaired during the progression of asthma. In addition, the lung tissue W/D ratio of mice in the model group was increased, and the  Figure 6: Spearmen's correlation analysis (heatmap). (a, b) Correlations between therapeutic indicators and gut microbiota (a) and between untargeted metabolomics and gut microbiota (b). Color coding scale indicates the correlation coefcient from heatmap, and the deeper red or blue indicates the higher absolute of the value. * P < 0.05; △ P < 0.01. number of total BALF cells, eosinophils, and total protein were increased, indicating increased lung tissue permeability and increased serum IgE levels. Tese results were consistent with the pathological manifestations of asthma [23], suggesting that the model was successfully constructed. QFSS intervention signifcantly improved lung function, reduced lung tissue permeability, and downregulated serum IgE levels in asthmatic mice. In addition, we selected DXM as a positive control drug, which is a hormonal drug commonly used in the treatment of asthma [24]. Te results showed that high-dose QFSS did not difer signifcantly from the DXM in improving lung function and reducing lung tissue permeability in asthmatic mice, and these results confrmed the therapeutic efects of QFSS in asthma.
Airway infammation and oxidative stress are important pathological manifestations of asthma, and we next investigated the efects of QFSS on airway infammation and oxidative stress in asthmatic mice. Te results showed that the QFSS intervention reduced IL-4, IL-5, and IL-13 levels and increased IFN-c levels in BALF of asthmatic mice. T helper cell (T) 1/T2 balance is important for maintaining the immune homeostasis of the body, and T1/T2 imbalance is an important mechanism for the development of airway infammation in asthma. A study has found that there were reduced levels of T1 in asthma, which reduces the levels of certain T1-secreted cytokines including IFN-c, and elevated levels of T2, which can produce large amounts of IL-4, IL-5, and IL-13. Tese secreted cytokines can exacerbate the occurrence and progression of allergic reactions and promote T2 diferentiation, thereby inhibiting T1 diferentiation [25]. Decreasing the T2-associated cytokines IL-4, IL-5, and IL-13, and increasing the T1 cytokine IFN-c may help to alleviate the infammatory response in asthma.
Our results also showed that QFSS increased SOD and GSH-Px activities and decreased MDA levels in the lung tissue of asthmatic mice, suggesting that QFSS inhibits oxidative stress in asthmatic mice. Oxidative stress is also an important pathological response in asthma as the accumulation of considerable amounts of reactive oxygen species (ROS) occur in the airways during the onset of asthma and these ROS exacerbate lipid peroxidation reactions and cause respiratory epithelial cell damage [26]. SOD and GSH-Px are important antioxidant enzymes that promote ROS scavenging. MDA is an end product of lipid peroxidation and its elevated level can refect the severity of oxidative stress [27].
In this study, 16S rRNA high-throughput sequencing technology was used to study the efect of QFSS on the structure and composition of gut microbiota in asthmatic mice. Alpha diversity of gut microbiota refers to the diversity of microbiota within a specifc region or ecosystem, and is a comprehensive indicator of the richness and homogeneity of the microbiota. Our results show no signifcant diferences in the alpha diversity of gut microbiota in each group, indicating that the total numbers of OTUs were similar in each group. Terefore, we measured the beta diversity of gut microbiota in each group. Te beta diversity of mouse gut microbiota was analyzed by PCoA and clustering analysis. Te overall structure and composition of the gut microbiota of asthmatic mice changed signifcantly, QFSS could afect the beta diversity of the gut microbiota of asthmatic mice, and the beta diversity of the gut microbiota of asthmatic mice was more similar to the control group after QFSS intervention. Te results of the analysis of the relative abundance of gut microbiota showed that QFSS could decrease the high levels of the F-to-B ratio caused by asthma. Changes in the F-to-B ratio are intimately associated with many diseases and metabolic disorders and infammatory responses can be alleviated by decreasing the F-to-B ratio [28].
In addition, the relative abundance of Lachnospir-aceae_NK4A136_group and Helicobacter was signifcantly increased and the relative abundance of Lactobacillus, Alloprevotella, Bacteroidota, and Dubosiella was signifcantly decreased in the mouse model of asthma; QFSS could signifcantly increase the relative abundance of Lactobacillus and Dubosiella, and decrease the relative abundance of Lachnospiraceae_NK4A136_group and Helicobacter. In a previous study, Lachnospiraceae_NK4A136_group was signifcantly and positively correlated with IgE and IL-33 [29]. Helicobacter is a Gram-negative spiral bacterium that is closely associated with many gastrointestinal diseases [30]. However, there is controversy over the efect of Helicobacter in asthma. Some studies have reported a degree of the protective efect of Helicobacter against asthma; however, other studies have pointed out that there is no negative association between Helicobacter and asthma [31]. Terefore, the interaction between Helicobacter and asthma requires further in-depth study. Lactobacillus, a natural microorganism with immunomodulatory abilities, has been shown to alleviate respiratory diseases such as asthma in several animal studies and clinical trials [32]. Highthroughput sequencing studies of the gut microbiota of asthmatic patients have revealed a lower abundance of Alloprevotella in treated patients [33]. Bacteroidota in the gut metabolizes polysaccharides and oligosaccharides to provide nutrients and vitamins to the host and other gut microbiota. However, when Bacteroides colonize other sites, it has the potential to become opportunistic pathogens [34]. In a study on allergic asthmatic mice, it was found that Faecalibacterium prausnitzii may alleviate pathological changes in asthmatic mice by regulating the production of Dubosiella and short-chain fatty acids in gut microbiota [35]. At the same time, our correlation analysis showed that Helicobacter and Lachnospiraceae NK4A136 groups were positively correlated with the vast majority of asthma pathological indicators and proinfammatory factors, and negatively correlated with SOD, GSH-Px, and IFN-c. In contrast, Alloprevotella, Bacteroidota, and Dubosiella behaved in contrast to the aforementioned microbiota, but failed to show statistical signifcance in eosinophil counts and IFN-c. Lactobacillus showed a signifcant negative correlation only with the W/D ratio and eosinophil count. Notably, the Lachnospiraceae NK4A136 group showed an extremely wide range of signifcant associations, suggesting that we can focus on studying the deeper relationship between this group and asthma.
Untargeted metabolomics of lung homogenates showed that QFSS afects arginine and proline metabolism, arginine Evidence-Based Complementary and Alternative Medicine biosynthesis, glycerophospholipid metabolism, and pyrimidine metabolism in asthmatic mice. After correlating the diferential metabolic pathways obtained from metabolomics with those deduced from 16s rRNA sequencing, arginine and proline metabolism, and pyrimidine metabolism metabolic pathways were identifed as their common pathways, suggesting that QFSS may infuence metabolic pathways through the regulation of gut microbiota, thus exerting the efect of asthma treatment.

Arginine and Proline
Metabolism. In our study, we found that 4-guanidinobutyric acid levels decreased in asthmatic mice, while both L-ornithine and proline levels increased. Te 4-guanidinobutyric acid levels increased and both Lornithine and proline levels decreased signifcantly after treatment with QFSS. Both 4-guanidinobutyric acid and Lornithine are downstream metabolites of arginine. Studies have shown that arginine metabolism plays an important role in asthma, which is related to nitric oxide (NO) metabolism in arginine biosynthesis, and high exhaled NO is one of the representative features in most asthmatic patients [36]. Coincidentally, L-ornithine is also one of the important metabolic intermediates in arginine biosynthesis and Lornithine is also a precursor of several asthma-related metabolites, and its derivative proline, a precursor of collagen, is associated with airway remodeling [37], while the derivative spermine increases airway sensitivity to methacholine [38]. Te 4-guanidinobutyric acid has been shown to have anti-H. pylori efects [39], which is consistent with the results of our correlation analysis. However, the results did not show sufcient statistical signifcance between 4guanidinobutyric acid and gut microbiota. QFSS may afect 4-guanidinobutyric acid through other pathways, which still need to be further studied. Moreover, Helicobacter and Lachnospiraceae NK4A136 groups were positively correlated with L-ornithine and proline, while Alloprevotella and Bacteroidota were negatively correlated with L-ornithine.

Pyrimidine Metabolism.
Our study found that lung levels of 2-deoxyuridine, uracil, and uridine were signifcantly decreased in mice with asthma models. All these metabolites were signifcantly elevated after QFSS treatment. Many experiments have demonstrated that uridine has antiinfammatory efects and is able to suppress the classical characteristics of asthma airway infammation [40,41]. Uracil can be interconverted with 2-deoxyuridine or uridine. However, there are very few studies on uracil and 2deoxyuridine in asthma and we failed to fnd valuable results. Hence, further studies are needed to explore the possible potential relationship. We found negative correlations between Helicobacter and Lachnospiraceae NK4A136 group and uracil and uridine, while positive correlations were present between Lactobacillus and uracil and uridine and 2-deoxyuridine. Tere was also a positive correlation between Bacteroidota and uridine.

Conclusion and Future Prospective
Recently, multiomics techniques are widely used in elucidating the mechanisms of Chinese herbal formulas. Te use of multiomics techniques can better illuminate the multicomponents and multitargets of Chinese herbal formulas [42]. Our results showed that QFSS could ameliorate asthma in mice. Te possible mechanism of QFSS on asthma may be associated with regulating gut microbiota and arginine and proline metabolism and pyrimidine metabolism. Our study may be useful for researchers to study the integrative mechanisms of Chinese herbal formulas based on modulating gut microbiota and metabolism.
In future, the fecal transplantation and the gut microbiota depletion model can be used to further validate the detailed mechanisms of QFSS on asthma based on modulating gut microbiota and host metabolism. Besides, our future study will deeply evaluate the efects of QFSS on the expression of relative metabolic enzymes, in order to deeply elucidate the metabolic regulatory mechanism of QFSS. Moreover, studies showed that the changes of metabolism in immune cells (such as macrophages, neutrophils, and lymphocytes) are important immune mechanism of asthma [43]. Te efects of QFSS on metabolism in immune cells can also be studied in our future study.
Tere are also some limitations of our study. Only highdose QFSS treatment mice were selected for the gut microbiota and metabolomics analysis in our study. Further study can be carried out to evaluate the efects of diferent dosages of QFSS on gut microbiota and metabolites in the asthmatic model mice.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Disclosure
Haibo Hu, Guojing Zhao, and Kun Wang share the frst authorship.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
Haibo Hu, Guojing Zhao, and Kun Wang conducted the experiments and wrote the manuscript. Ping Han and Fengchan Wang provided experimental help. Haiyan Ye, Na Liu, and Peixia Zhou analyzed the data and interpreted the results. Xuechao Lu and Zhaoshan Zhou provided ideas, resources, and technical guidance for the whole work. Huantian Cui supervised the experiments. All the authors contributed to the article and approved the submitted version. 16 Evidence-Based Complementary and Alternative Medicine