3.1 The expectorant effect of FF was related with gut microbiota
The tracheal phenol red excretion test is a simple and effective method to investigate the effect of drugs on tracheobronchial secretion [12]. To assess the effect of FF on promoting tracheal secretion, phenol red output was examined after intraperitoneal injection of phenol red solution. As shown in Fig. 2B, compared with the normal controls, the FF and POS treatment could significantly increase the phenol red secretion, which indicating that both FF and POS showed the expectorant effect.
In order to determine whether the expectorant effect of FF was related with gut microbiota, the mice pretreated with antibiotics cocktail for depleting gut microbiota were subjected to the expectorant test. It was obvious that there was no significant difference between the antibiotics pretreated mice (ABFF group) and CON group for the phenol red secretion. Thus, the expectorant effect of FF was attenuated when the gut microbiota was suppressed by the antibiotics.
3.2 FF treatment could modulate the gut metabolome and the effect could also be attenuated by antibiotics
The fecal metabolome can reflect the complex biosystem within the gastrointestinal tract and the interaction with resident microbiota, host and drugs [13]. Thus, we further evaluated the gut metabolome by untargeted metabolomics based on UHPLC-Q TOF MS. PCA was performed to visualize the clustering of samples (Fig. 2C-D), and the results showed that CON, FF and ABFF groups were separated obviously. In addition, all the QC samples fell within the 2 SD’s region and 95% confidence interval, which indicated that the analytical system showed satisfactory stability (Fig. 2E-F). Based on the criteria of t-test (P < 0.05) and fold changes (FC ≥ 1.2 or ≤ 0.8), the differential metabolites between CON and FF were screened out. Compared with the control mice, 16 and 12 metabolites were significantly up-regulated and down-regulated after FF treatment, respectively (Table S1). Among them, the regulatory effect of 20 metabolites could be attenuated after the AB pretreatment, and these metabolites might be related with the expectorant effect of FF.
3.3 Fecal microbiota transplants could promote the tracheal phenol red output and modulate the gut metabolome
To further confirm the expectorant effect of FF was related with the gut microbiota, the fecal microbiota transplants (FMT) solution created from expelled stool from control mice and FF pretreated mice were transferred to antibiotics pretreated mice. As shown in Fig. 3B, compared with the CFMT group, the amount of phenol red in the AB group did not change significantly, so the AB treatment did not interfere with the phenol red excretion experiment. However, the amount of phenol red in the FFMT group was significantly higher (P < 0.05), which indicating that the FFMT solution exhibited expectorant effect.
The fecal metabolome of CFMT and the FFMT were also compared by the metabolomics. Based on the criteria [t-test (P < 0.05), FC ≥ 1.2 or ≤ 0.8], 27 differential metabolites were determined (Table S2). Venn diagrams (Fig. 3C) indicated that 13 metabolites were common to FF vs CON and FFMT vs CFMT. Pathway enrichment analysis of these metabolites revealed 10 metabolic pathways (P < 0.05, Fig. 3D), such as arginine biosynthesis, histidine metabolism, aminoacyl-tRNA biosynthesis, beta-alanine metabolism, arginine and proline metabolism.
3.4 The CQA metabolites QA and DCA also exerted expectorant effect
In order to determine whether the live bacterial or the metabolites in the fecal microbiota transplants solution was responsible for the expectorant effect, the FFMT solution was heated by autoclaving to prepare HFMT solution. Compared with CFMT solution, HFMT solution could also promote the tracheal phenol red output (P < 0.05), but the effect was weaker than FFMT solution (Fig. 3B). Thus, the metabolites in the HFMT solution might also be related with the expectorant effect of FF. Chemical analysis by UHPLC-Q TOF MS showed that CA, QA and DCA were detected as the major metabolites in the HFMT solution (Table S3), and all of them were metabolized by the gut microflora from CQAs in the FF.
In order to investigate the role of these metabolites in the expectorant effect of FF, the mice were orally administrated with CA, QA and DCA for 2 weeks. The results showed that DCA and QA could significantly promote the production of tracheal phenol red in mice (Fig. 4A), while CA did not exert the expectorant effect. Therefore, the expectorant effect of FF was related to DCA and QA. As DCA and QA were transformed by the gut microbiota, the expectorant effect of FF was attenuated greatly when the gut microbiota was suppressed by the antibiotics. It has been reported that gut microbiota can hydrolyze diCQAs into monocaffeoylquinic acid and caffeic acid [14, 15]. As shown in Fig. 4B, CA, DCA, and QA were reduced after antibiotics pretreatment, which were in agreement with the decrease of expectorant effect of FF.
The regulatory effect of DCA and QA on the fecal metabolome was further investigated by the metabolomic approach. Based on the criteria of t-test (P < 0.05) and fold changes (FC ≥ 1.2 or ≤ 0.8), 18 differential metabolites were determined between the DCA and the CON group (Table S4), while 43 differential metabolites were determined between the QA and the CON group (Table S5). Pathway enrichment analysis showed that the arginine metabolism was both enriched for the QA and DCA (Fig. 4C).
3.5 Arginine supplementation exhibited expectorant effect and showed similar effect on the lung metabolome as FF
According to the enrichment pathway analysis, arginine biosynthesis was enriched in the expectorant effect of FF, DCA, and QA. It has been reported that arginine supplementation could alleviate the symptoms of asthma, COPD, cystic fibrosis, bronchopulmonary dysplasia and pulmonary hypertension [16]. Thus, arginine supplementation experiment was carried out to determine whether arginine showed the expectorant effect. As shown in Fig. 5A, oral arginine could promote the tracheal phenol red output (P < 0.05). Thus, arginine also showed expectorant effect, however, the effect was weaker than FF.
The regulatory effect of arginine and FF on the lung metabolome was further compared by the metabolomics. A total of 34 differential metabolites were found between CON and FF (Table S6), including 20 upregulated and 14 downregulated. A total of 37 differential metabolites were found between CON and Arg (Table S7), including 19 upregulated and 18 down regulated. There were 19 common metabolites both regulated by FF and Arg (Fig. 5B). For the metabolites relating with arginine biosynthesis, it was obviously that the oral administration of FF could increase the arginine and citrulline in the lung tissue, while decrease the ornithine.
In addition, FF could increase the phospholipids in the lung tissue, including PI 16:0_20:4, PC 6:0_18:4, PC 16:0_18:2, PG 16:0_16:0, PG 16:0_18:2, PG 16:0_16:1, PG 16:0_18:2, PC 16:0_22:2, PC 16:0_16:0 (Table S6). These phospholipids have been reported to be served as the components of pulmonary surfactant (PS). It was interesting that the oral arginine could also significantly regulate some of these phospholipids, such as PI 16:0_20:4, PC 16:0_16:0, PC 6:0_18:4, PG 16:0_18:2 (Fig. 5C).
3.6 QA and DCA showed similar effect on the lung metabolome as FF
The regulatory effect of QA and DCA on the lung metabolome was also investigated. The differential metabolites for QA included 20 upregulated and 7 downregulated metabolites (Table S8). The differential metabolites for DCA included 26 upregulated and 4 downregulated metabolites (Table S9).
The distribution of metabolites which are unique and common among three comparisons (FF vs CON, DCA vs CON and QA vs CON) was shown in the Venn diagram (Fig. 6A). The results indicate that 20 out of 34 differential metabolites were unique for FF vs CON, 9 out of 30 were unique for DCA vs CON, and 8 out of 30 were unique for QA vs CON. In addition, there were 11 differential metabolites common between FF vs CON and DCA vs CON, 12 commons between FF vs CON and QA vs CON. A total of 9 common metabolites were among the three comparisons, which including xanthine, arginine, citrulline, phenylalanine, cis-5,8,11,14,17-eicosapentaenoic acid (EPA), PG 16:0_18:2, Stearate, PG 16:0_16:1, PC 16:0_16:0.
Further enrichment analysis (Fig. 6B) of differential compounds involved in FF vs CON, DCA vs CON and QA vs CON showed that 8 metabolic pathways were co-enriched, including: phenylalanine metabolism, arginine biosynthesis, aminoacyl-tRNA biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis, biosynthesis of unsaturated fatty acids, purine metabolism, alanine, aspartate and glutamate metabolism, arginine and proline metabolism. For the phospholipids in the lung tissue, DCA and QA could also increase the abundance of PG 16:0_18:2, PG 16:0_18:1, PG 16:0_16:1, PC 18:0_18:1 and PC 16:0_16:0. Thus, DCA and QA showed similar effect on the lung metabolome as FF.