Anti-p-p38 MAPK (Thr180/Tyr182) antibody, anti-p38 MAPK antibody, and FITC-anti-F4/80 antibody were from Cell Signaling Technology. Anti-p-ERBB2/HER2 (phospho Y877) antibody, anti-ERBB2/HER2 antibody, anti-GAPDH antibody, anti-CD206 antibody, and Alexa Fluor 647-anti-ERBB2 (phospho Y877) antibody were from Abcam. A macrophage isolation kit (peritoneum) was purchased from Miltenyi Biotec.

Male wild-type (WT) C57BL/6J mice aged 6–8 weeks were purchased from Beijing Huafukang Biotechnology Co., Ltd. All mice were subsequently bred and settled under specific pathogen-free (SPF) conditions at Tsinghua University. In our study, we anesthetized WT mice with an intraperitoneal injection of tribromoethanol (10 mg/kg). After midline laparotomy, for cecum ligation and puncture (CLP) groups, the cecum was exposed and ligated immediately below the ileocecal valve, causing intestinal obstruction. After two punctures with an 18G needle, the cecum was placed back and the abdominal wall was sealed. Reduning group and CLP+Reduning group were administered with Reduning (50 or 100 mg/kg) by intraperitoneal injection once per day for 7 days before the operation. For the sham group, the cecum was exposed but did nothing, and then the abdominal wall was closed. BAL fluid (BALF) was collected by rinsing the alveolar cavity with 1 ml precooled PBS, and it was stored at −20°C. The right lungs were embedded in 10% formalin prepared for morphology examination. For all these groups, normal saline was injected intraperitoneally just after the operation to simulate clinical fluid replenishment. At 21 days after the operation, mice were sacrificed, and the inferior lobe of the right lung was used as the study sample. All study designs were supervised and approved by the Ethics Committee of Beijing Tsinghua Changgung Hospital (protocol code NCT05095324).

After fixation in 4% paraformaldehyde, the lung tissues were dehydrated, embedded in paraffin, and sliced for Masson staining. The morphological structure, inflammatory infiltration, and degree of fibrosis of the lung tissues were evaluated under a microscope after the sections were stained. The Ashcroft score for pulmonary fibrosis was performed on the lung histopathological sections of each mouse, and the scoring criteria were as follows (Carl-Mcgrath et al., 2008): 0: normal lung; 1–2: slight thickening of alveoli and bronchial walls; 3–4: moderate thickening of alveolar and bronchial walls, but no evident destruction of lung structure; 5–6: more pulmonary fibrosis with specific lung structure destruction, fiber bundle formation, and small fiber aggregation; 7: severe pulmonary structure distortion, extensive fibrosis, and honeycomb lung; and 8: pulmonary fibrosis.

IHC was performed following the manufacturers’ protocols. Sections were backstained, dehydrated, removed, and mounted with modified methylethoxylin. All regiments included appropriate positive and negative controls. The percentage of F4/80 protein positive in the lung tissue was derived from three separate data.

After the sections were dewaxed and hydrated, lx EDTA antigen repair solution (PH 9.0) was used for antigen repair. After PBS cleaning, endogenous peroxidase block-agent was added to the tissue sections to block the activity of endogenous peroxidase. Goat serum was added to the tissue sections for sealing and incubated in a wet box at room temperature for 30 min. After that, primary antibody diluted with sterile PBS (a-SMA antibody concentration 1:500) was added to the covered tissue sections and incubated at 4°C overnight. An appropriate amount of immunofluorescent secondary antibody was added and was inoculated for 1 h against light at room temperature. Finally, the nucleus was stained with DAPI dye solution, and the tablet was sealed with an anti-fluorescent quench agent.

Cells from mice BALF were collected and harvested after centrifugation at 1,000 rpm for 10 min at 4°C and suspended in precooled sterile PBS containing 2% FBS. Mouse alveolar macrophages (AMs) were isolated using a macrophage isolation kit (peritoneum) following the manufacturer’s instructions. Then, Western blot was applied to detect the expression level of the indicated proteins.

AMs were cultured with Dulbecco’s modified Eagle medium (DMEM) and high glucose (Gibco, United States) under standard conditions (37°C, 5% CO₂) in the laboratory of Tsinghua Changgung Hospital. The experiments were exhibited after two passages. Cells were incubated with LPS (100 ng/ml). After the indicated length of time, cell culture supernatant and cells were collected. The supernatants were analyzed by enzyme-linked immunosorbent assay (ELISA) to evaluate pro-inflammatory and anti-inflammatory factors. Then, Western blot were applied to detect the expression level of the indicated proteins.

ERBB2 proteins were evaluated by Western blotting. Total protein was extracted according to the standard protein extraction protocol. The protein concentration was quantified by a BCA protein assay kit. The sample was electrophoretic in a 10% SDS-PAGE gel and transferred to a polyvinylidene fluoride film. The film is then sealed overnight with 5% milk at 4°C. The membrane was rinsed three times with Tris-buffered saline Tween-20 (TBST) and incubated with secondary antibodies at room temperature for 40 min. After washing with TBST again, the protein signal was observed, and the membrane signal was displayed by an enhanced chemiluminescence Western blot detection system. The staining intensity of the strip was measured using ImageJ software. The antibodies used in our study are described in the previous “reagents” section. Protein expression levels were defined as gray values normalized to the steward gene GAPDH and expressed as multiples of the control. All experiments were performed three times and three times independently.

Data were expressed as the mean ± SD. The differences between variables were compared using one-way ANOVA or a two-tailed Student’s t-test using SPSS software (version 22.0; United States). p < 0.05 was considered statistically significant. Moreover, Prism 8 (GraphPad Software, United States) was responsible for visualizing.

In order to better understand the intricate relationship among Reduning, active components, and corresponding targets, a network diagram of TCM, compound, and target was constructed based on the included TCM, active components, and target sites. The network was imported into Cytoscape 3.7.2 software to draw the network diagram, and the results are shown in Figure 2. Topological analysis selected the top five compounds by degree as key compounds, as shown in Table 2.

In order to further explore which pathways are mainly affected by the shared targets of diseases and TCM, we conducted GO and KEGG enrichment analysis on the shared targets. A total of 1,634 GO entries were enriched. There were 217 cell components (CCs), which mainly located membrane rafts, membrane microdomains, neuronal cell bodies, synaptic membranes, postsynaptic membranes, and components of presynaptic membranes. There were 3,477 biological processes (BP), which mainly regulated and participated in the processes of positive regulation of kinase activity, positive regulation of MAPK cascade, calcium homeostasis, peptidyl-serine modification, peptidyl-serine phosphorylation, peptidyl-tyrosine phosphorylation, peptidyl-tyrosine modification, and protein self-phosphorylation. There were 426 molecular functions (MFs), mainly manifested in protein serine/threonine/tyrosine kinase activity, transmembrane receptor protein tyrosine kinase activity, nuclear receptor active ligand-activated transcription factor activity, and non-transmembrane protein tyrosine kinase activity. The top eight enrichment items with p-value were selected, and according to the p-value, Q-value of each item, and the number of genes enriched on it, the ggplot2 package of the R language was used to draw the bubble map of GO analysis. The abscissa represents the number of targets, the left represents BP, CC, and MF, and the color represents the p-value. The smaller the p-value, the more red the color is; the larger the p-value, the more blue the target is, as depicted in Figure 3A.

A total of 191 KEGG pathways were enriched. The top 20 enrichment items with p-value were selected, and according to the p-value, Q-value of each item, and the number of genes enriched on it, the ggplot2 package of the R language was used for visual analysis of the results. The abscissa represents the number of genes enriched, the left represents the pathway name, and the color represents the p-value. The smaller the p-value, the more red the color, and the larger the p-value, the more blue the color. As shown in Figure 3B, shared targets were mainly enriched in the MAPK signaling pathway, the calcium signaling pathway, the AMP signaling pathway, the sphingolipid signaling pathway, the apoptosis pathway, the neurotrophic factor signaling pathway, the AGE-RAGE signaling pathway, the HIF-1 signaling pathway, and the PD-1 checkpoint in diabetic complications.

In order to clarify the interaction between active components and targets and to predict the binding mode and affinity of complexes, AutoDock software was used to dock the aforementioned five key components and five targets. The compound names, molecular weights, and 3D structures of the active ingredients were determined from the PubChem database, and then, the corresponding 3D structures of the active ingredients were downloaded from the RCSB PDB database (http://www.rcsb.org/). AutoDock software was used to prepare the ligands and proteins required for molecular docking, remove water molecules, hydrogenate, modify amino acids, optimize energy, adjust force field parameters for the crystal structure of the target protein, and then meet the low-energy conformation of the ligand structure. Finally, the five key target structures and five key active components were used for molecular docking, and the affinity (kcal/mol) value represents the binding ability of the two. Finally, Discovery Studio and PyMol software were used to analyze and observe the docking results. The key active components stigmasterol, kaempferol, quercetin, beta-sitosterol, and corymbosin were used to perform molecular docking verification with the key targets SRC, TNF, CASP3, GAPDH, and ERBB2. The results showed that all 25 groups had good binding activity. The top five results were selected and analyzed. Table 3 and Figure 4 depict these relationships.

In order to verify whether the binding of the three groups of complexes with ERBB2 is stable, MD was applied to explore the stability and kinetic characteristics of the complexes in aqueous solution. The atomic root mean square deviation (RMSD) was used to measure the stability of the system, and the RMSD fluctuation range was small in the three groups of dynamics simulations, which could quickly stabilize to 0.2 nm after the beginning of the simulation (Figure 5A). RMSF can reflect the fluctuation in the simulation process, and the results show that the three groups of simulations share similar binding patterns, and their flexibility ranges are basically the same (Figure 5B). The radius of gyration (Rg) was used to evaluate the closeness of the architecture. The fluctuation amplitude simulated by the three groups was consistent with RMSD. As shown in Figure 5C, the stigmasterol complex with ERBB2 (group I) was slightly larger in magnitude than the beta-sitosterol complex with ERBB2 (group II) and the quercetin complex with ERBB2 (group III). SASA was used to explore the solvent-accessible surface of the protein surface, and it was found that the SASA of the protein in the range of 0–50 ns steadily decreased and the protein gradually contracted (Figure 5D). The number of HBNUM between protein ligands during simulation was further calculated. The first group was stable at one to two contacts, the second group was stable at 0–1 contacts, and the third group was fluctuating at one to three contacts (Figure 5E). Further analysis of the corresponding binding modes revealed that the first two stable binding postures were small molecules firmly adsorbed in the hydrophobic cavities around the B-sheet and A-helix (Figures 5F,H). Figure 5G illustrates the stable binding of small molecules to VAL-25, ASN-302, and TYR-303 of the ERBB2 (HER2) protein in the third group of the simulation system.

As shown in Figures 6A,B, in the the CLP group, blue collagen fibers were deposited in the lung tissue, showing diffuse distribution, collapse of alveolar structure, evident thickening of the alveolar septum, and abnormal deposition of extracellular matrix. The CLP+RDN (50/100 mg/kg) group showed scattered collagen fiber deposition in the lung tissue, which was lighter than the CLP group. In the sham group, the CLP+RDN (25 mg/kg) group, and the RDN group, no notable collagen fiber deposition was observed in the lung tissue (p > 0.05). As shown in Figures 6C,D, the expression level of α-SMA protein in the lung tissue was observed by IHC. The expression of α-SMA in the lung tissue of the CLP group was significantly higher than that of the sham group. There was a small amount of α-SMA expression in the lung tissue of the CLP+RDN (50/100 mg/kg) group, which was lighter than the CLP group. In the sham group, there was no notable α-SMA expression in the lung tissues of the CLP+RDN (25 mg/kg) and the RDN groups, compared with the sham group (p > 0.05).

To confirm the cell type to which Reduning applied, IF was used to investigate the expression and localization of p-ERBB2 in each group. As shown in 7A, p-ERBB2 was co-located with F4/80. As shown in Figure 7B, compared with the sham group, p-ERBB2 was significantly enhanced in the CLP group (p < 0.01). In comparison with the CLP group, p-ERBB2 was sharply reduced after RDN intervention (p < 0.01).

To further detect the underling mechanism of RDN, AMs were isolated from the lung tissue. We detected the ERBB2-p38 MAPK signaling pathway-related proteins in each group. As illustrated in Figures 8B,C, the expression levels of p-ERBB2/ERBB2 and p-p38/p38 MAPK which were enhanced in the CLP group were significantly dampened by Reduning in the CLP+Reduning group (p < 0.001). There is no significant difference between the Reduning group and the sham group (p > 0.05).

To clarify the impact of RDN on the polarization of AMs, flow cytometry was used to measure the expression of CD206. As shown in Figure 9, compared with the sham group, CD206 in the CLP group was evidently promoted. Reduning could alleviate the expression level of CD206. Also, the cell apoptosis proportions were 0.47 ± 0.16, 28.63 ± 1.53, and 0.51 ± 0.17 for the sham group, the CLP group, and the CLP+Reduning group, respectively (p < 0.001).