African Green Monkey Kidney Cells (Marc-145 cells) were cultured in Dulbecco’s Modified Eagle’s Medium [DMEM (Gibco, Waltham, MA, USA)], supplemented with 10% fetal bovine serum (FBS); Highly pathogenic type 2 PRRSV (PRRSV-2) strain HuN4-F112 (a Marc-145-adaptive strain) was purchased from Jilin Zhengye Biological Company. Ma-Xing-Shi-Gan-San, a TCM formula composed of Mahuang, Kuxingren, Shigao, and Gancao, was stored in our laboratory and diluted in DMEM.

The cytotoxicity of MXSGS was assessed by Cell Counting Kit-8 [CCK-8 (Beyotime, Shanghai, China)]. Marc-145 cells were first seeded into 96-well plates at 1 × 10⁵ cells/mL. Simultaneously, a blank control group was established using only a maintenance medium (DMEM with 2% FBS) without any cells. The MXSGS was serially diluted two-fold in the maintenance medium and then incubated with Marc-145 cells in 96-well plates for 72 h. A negative control group, which consisted solely of the maintenance medium, was also included. After 72 h, 10 μL of CCK-8 reagent was added to each well and incubated at 37°C for 1 h; cytotoxicity was subsequently assessed using the Epoch 2 Microplate Spectrophotometer (BioTek, Winooski, VT, USA).

Marc-145 cells were inoculated in six-well plates with PRRSV at a multiplicity of infection (MOI) of 0.1 for 1 h, followed by three washes with phosphate-buffered saline (PBS). Subsequently, maintenance medium containing different concentrations of MXSGS (2.25, 4.50, 9.00, and 18.00 mg/mL) were added to the respective wells for co-incubation for 72 h. A negative control group was established without PRRSV inoculation, while a positive control group was infected with PRRSV and supplemented only with maintenance medium. After 72 h, two independent double-blind researchers assessed the cytopathic effect (CPE) using an inverted microscope (Motic, Xiamen, China).

To investigate the inhibitory effect of MXSGS on PRRSV, we extracted total RNA using TRIzol Reagents (Invitrogen, Carlsbad, CA, USA), and the RNA from each sample was subsequently reverse transcribed into cDNA using a reverse transcription kit (ABclone, Woburn, MA, USA). The cDNA was subjected to RT-qPCR analysis employing the SYBR Green Kit (ABclone, Woburn, MA, USA), with β-actin selected as the housekeeping gene. The expression levels of PRRSV ORF7 mRNA in different treatment groups were analyzed using the 2^-ΔΔCt method. Additionally, the ORF7 mRNA copy number was quantified by absolute quantitative PCR methods (Table 1). The RT-qPCR assay was performed on the qTOWER³ real-time PCR system (Analytik Jena, Jena, Germany).

Marc-145 cells seeded in 96-well plates were infected with serial 10-fold dilutions of samples in eight replicates, incubated at 37°C with 5% CO2 for 5 days; meanwhile, a negative control group was established without viral exposure. After incubation, the cytopathic effects in each well were assessed to calculate PRRSV viral titer based on the Reed-Muench method (Pizzi, 1950). The results were expressed as 50% tissue culture infective dose (TCID₅₀)/mL.

The TCMSP database¹ (Ru et al., 2014) was utilized to extract the constituents of the Mahuang, Kuxingren, and Gancao. The criteria for screening in TCMSP were set as oral bioavailability (OB) > 30% and drug-likeness (DL) > 0.18 (Chang et al., 2024). Each constituent’s simplified molecular linear input specification (SMILES) numbers were collected through the PubChem database.² These SMILES were employed to screen for active ingredients on the SwissADME platform,³ with a gastrointestinal absorption score classified as “high” and a drug-likeness rating of at least 3 “Yes” responses. Subsequently, the SwissTargetPrediction platform⁴ was used to predict the targets corresponding to the active ingredient of Mahuang, Kuxingren, and Gancao. Since Shigao is not included in the TCMSP database, calcium (Ca), magnesium (Mg), iron (Fe), potassium (K), and sodium (Na) were selected as primary components based on previous literature (Liu et al., 2024), and their targets were collected in the STITCH 5.0 database.⁵ Finally, targets associated with Mahuang, Kuxingren, Gancao, and Shigao were combined while qualifying species as “Sus scrofa” for screening; target names were standardized using the UniProt database.⁶

To identify the PRRSV disease targets, the keyword “porcine reproductive and respiratory syndrome” was queried in the GeneCards,⁷ OMIM⁸ and PharmGKB database.⁹ These databases illuminate the relationship between targets and diseases from various perspectives. Then, disease targets were consolidated, and the species was qualified as “sus scrofa” to filter the validated targets and standardize the target nomenclature through the UniProt database (accessed on 15 September 2024).

The targets of the active ingredients in MXSGS were compared with the disease targets associated with PRRSV to identify common targets, which were defined as direct therapeutic targets acting on PRRSV with MXSGS. The common targets were utilized to construct a target network via the String database¹⁰ (Doncheva et al., 2019), with species restricted to “Sus scrofa.” Subsequently, a protein–protein interaction (PPI) network was generated using Cytoscape version 3.7.1 (Shannon et al., 2003). Topological analysis of the intersected targets was conducted based on UnDir degree, betweenness centrality, and closeness centrality to screen core potential targets for MXSGS treatment of PRRSV.

Subsequently, the active ingredients and common potential targets were imported into Cytoscape version 3.7.1 to construct the “MXSGS ingredient-PRRSV-targets network,” filtering for key active ingredients of MXSGS based on UnDir degree by Analyze Network plug-in.

The core targets of MXSGS against PRRSV, identified from the 2.9 analysis, were input into the DAVID database¹¹ (Dennis et al., 2003), with the species set to “Sus scrofa.” Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed. The top 20 GO enrichment results and KEGG pathways were selected for visualization using an online bioinformatics tool.¹² Then, the KEGG analysis results were imported into Cytoscape version 3.7.1 for topological analysis to construct the “MXSGS-PRRSV-targets-pathways network.”

The top 6 potential core targets were identified from the PPI network and the “MXSGS-PRRSV-targets-pathways network, “respectively. The active ingredients with the highest 5 UnDir degrees were selected from the “MXSGS ingredient-PRRSV-targets network.” All screened targets and ingredients were utilized for molecular docking validation. The 3D structures of ALB (PDB ID: 4LA0), PPARG (PDB ID: 9CK0), CASP3 (PDB ID: 1RE1), STAT3 (PDB ID: 6NJS), TGFB1 (PDB ID: 6UJA), JAK2 (PDB ID: 7F7W), TLR4 (PDB ID: 2Z66), PRKACA (PDB ID: 7Y1G), and PRKACB (Prediction with AlphaFold) were downloaded from the PDB database.¹³ Subsequently, water molecules and small molecule ligands were removed from core target proteins by PyMOL version 2.6.0. Charge balancing and hydrogenation were performed with AutoDock Tools version 1.5.7 software, after which the processed protein targets were converted into .pdbqt format files.

Meanwhile, the .mol2 structure of the active ingredient was downloaded through the TCSMP database (accessed on 30 September 2024) and converted into the .pdbqt format files using AutoDock Tools 1.5.7 software. Next, molecular docking was conducted with AutoDock Vina software to obtain binding energy affinity. Specifically, affinity < −4.25 kcal/mol indicates the presence of binding activity between ligand and target; affinity < −5.0  kcal/mol implies good binding activity; affinity < −7.0 kcal/mol suggests strong docking activity (Ji et al., 2023).

Finally, the optimal conformations of the key active ingredient in MXSGS and the core target protein were visualized using PyMOL version 2.6.0.

GraphPad Prism 10.1.2 software (GraphPad Software, La Jolla, CA, USA) was used for data analysis. Differences between groups were assessed using the two-tailed Student’s t-test, one-way ANOVA, or two-way ANOVA analysis.

Firstly, we determined the safety concentration of MXSGS on Marc-145 cells using the CCK-8 assay (Figure 1). The results indicated that MXSGS did not significantly reduce cell viability at concentrations below 18.00 mg/mL.

Then, MXSGS demonstrated a significant dose-dependent reduction in PRRSV ORF7 mRNA levels in Marc-145 cells. For instance, there was a 655-fold and a 2 × 10⁴-fold decrease in PRRSV mRNA levels with 9.00 mg/mL MXSGS from 48 h. p. i. to 72 h. p. i., respectively (Figure 2A). Similarly, MXSGS significantly reduced PRRSV titers in a concentration-dependent manner; specifically, the group treated with 9.00 mg/mL MXSGS exhibited reductions of 10^3.36-fold and 10^3.94-fold in PRRSV titers after incubation for 48 h and 72 h, respectively, compared to the control group (Figure 2B).

Additionally, we investigated whether MXSGS could mitigate the cytopathic effect (CPE) induced by PRRSV in Marc-145 cells. The CPE in the virus group was evident, characterized by rounded, detached cells that aggregated into clumps. In contrast, the CPE was significantly alleviated in the MXSGS treatment groups (2.25, 4.50, 9.00, and 18.00 mg/mL), demonstrating a dose-dependent relationship (Figure 2C). These results indicate that MXSGS can effectively inhibit PRRSV infection in Marc-145 cells.

To further investigate the impact of MXSGS administration on PRRSV, three groups with different treatment stages were established (Figure 3A). Firstly, the PRRSV ORF7 mRNA level was reduced by nearly half in the prophylactic administration group (Figure 3B). The viral titer was significantly decreased only in the high-dose MXSGS group (18.00 mg/mL), while no significant reduction in PRRSV titers was observed in the middle and low-dose groups (9.00 and 4.50 mg/mL) (Figure 3E). Secondly, all doses of MXSGS administered in the direct virucidal group exhibited no significant inhibition of PRRSV mRNA expression levels and viral titers (Figures 3C,F). Finally, the therapeutic administration group demonstrated reductions of 3.26-fold, 15.57-fold, and 126.84-fold in PRRSV mRNA expression levels (Figure 3D), along with cuts of 1.12-fold, 10^3.2-fold, and 10^4.34-fold in viral titers (Figure 3G) for low-dose, medium-dose, and high-dose groups of MXSGS, respectively. Hence, both prophylactic and therapeutic administration modes inhibited PRRSV proliferation. Notably, the therapeutic administration group exhibited the most pronounced anti-PRRSV effects.

To investigate the mechanism of MXSGS against PRRSV, we examined its effects on each phase of the viral life cycle: attachment, internalization, replication, and release. Firstly, pre-cooled Marc-145 cells were inoculated with PRRSV (MOI = 1) after adding MXSGS (9.00 mg/mL) and incubated at 4°C for 1 h. The results from RT-qPCR indicated that PRRSV attachment was reduced by 0.31-fold in the MXSGS group (Figure 4A). Secondly, pre-cooled cells were infected with PRRSV for 1 h at 4°C before being treated with MXSGS (9.00 mg/mL) for an additional 3 h at 37°C; this treatment resulted in a 1.95-fold reduction in the amount of PRRSV internalized into Marc-145 cells (Figure 4B). Thirdly, PRRSV-infected Marc-145 cells were switched to incubation with MXSGS (9.00 mg/mL) at 6 h. p. i., leading to reductions in PRRSV ORF7 mRNA levels by factors of 3.65-fold, 5.26-fold, and 6.10-fold at 7, 8, and 9 h. p. i., respectively (Figure 4C). Finally, after switching to incubation with MXSGS (9.00 mg/mL) at 18 h. p. i., the copy numbers of PRRSV ORF7 mRNA in cell supernatants decreased by 26.10-fold, 17.27-fold, and 16.73-fold after drug intervals of 20 min, 40 min, and 60 min, respectively. In summary, MXSGS inhibited all phases of the PRRSV life cycle, including attachment, internalization, replication, and release.

We initially screened 82 ingredients (Supplementary Table S1) from MXSGS, identified 204 UniProt-validated targets associated with MXSGS (Supplementary Table S2), and obtained 634 disease-related targets after merging and removing duplicates, which included 606, 111, and 8 PRRSV targets sourced from the GeneCards, PharmGKB, and OMIM databases, respectively (Supplementary Table S3). Then, we selected 140 targets at the intersection of MXSGS and PRRSV targets (Supplementary Table S4; Figure 5A), constructing a Protein–Protein Interaction (PPI) network using the STRING database (Figure 5B) and Cytoscape version 3.7.1. The PPI network was then subjected to topological analysis via the Centiscape version 2.2 plug-in, which identified 20 core targets; among these, the top 6 included ALB (UnDir value = 56), PPARG (UnDir value = 30), CASP3 (UnDir value = 27), STAT3 (UnDir value = 26), TGFB1 (UnDir value = 25), and PRKACA (UnDir value = 22) (Figure 5C). Additionally, we constructed the “MXSGS ingredient-PRRSV-targets network” (Figure 5D), which illustrates that MXSGS comprises multiple components, with each ingredient corresponding to various targets. Subsequently, the top 5 active small molecules within the MXSGS ingredients were identified based on their UnDir degree: Calycosin (UnDir value = 29), Odoratin (UnDir value = 29), Glyzaglabrin (UnDir value = 28), 7,2′,4′-trihydroxy-5-methoxy-3-arylcoumarin (UnDir value = 27), and Eriodictyol (UnDir value = 26).

Subsequently, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were conducted on the 20 core targets identified from the PPI network analysis. The GO analysis yielded a total of 52 entries (p<0.05), comprising 38 related to biological processes (BP), 5 of molecular functions (MF), and 9 associated with cellular components (CC). The top 20 GO enrichment results were visualized (Figure 6A), indicating that MXSGS could inhibit PRRSV through Toll-like receptor signaling pathway, typical NF-κB signaling, positive regulation of interleukin-6 production, inflammatory response, viral defense response, and intrinsic immune response, etc. In addition, KEGG enrichment analysis revealed a total of 61 pathways (p < 0.05). The results of the top 20 KEGG pathways (Figure 6B) showed that the effection of MXSGS in anti-PRRSV primarily enriched in pathways in cancer, the AGE-RAGE signaling pathway, associated with diabetic complications, Th17 cell differentiation, coronavirus disease-pneumococcal pneumonia, etc.

Utilizing these selected KEGG pathways, we constructed the “MXSGS-PRRSV-targets-pathways network” (Figure 6C). Through topological analysis, 6 key targets including STAT3 (UnDir value = 14), JAK2 (UnDir value = 12), TLR4 (UnDir value = 11), PRKACB (UnDir value = 11), PRKACA (UnDir value = 11), and TGFB1 (UnDir value = 10) were identified. In summary, a total of 9 core targets of MXSGS against PRRSV, ALB, PPARG, CASP3, STAT3, TGFB1, JAK2, TLR4, PRKACA, and PRKACB were obtained from the above two networks for molecular docking analysis.

The top 5 active ingredients of MXSGS (acting as ligands) and a total of 9 core targets (serving as receptors) were utilized for molecular docking analysis, where all binding energies were lower than −6.0 kcal/mol, indicating an excellent molecular docking effect (Figure 6D). Lower binding energy values suggest more stable molecular ligand and protein receptor conformations. Among these, 10 ligand-receptor pairs exhibited high binding affinity with binding energies lower than −8.5 kcal/mol; we visualized these optimal conformations using PyMOL version 2.6.0 (Figure 7).

Interaction forces were observed with at least 1 hydrogen bond in all 10 ligand-receptor pairs. For instance, Odoratin exhibited a favorable interaction with the JAK2 protein, characterized by an absolute binding energy value of −8.8 kcal/mol and the formation of 6 hydrogen bonds at the docking site, with an average bond length of 2.7 Å. The connection position situated at the JAK2 protein base acid residues, namely LYS-581A, GLN-626A, LYS-630A, SER-633A, and SER-698A (Figure 7A). The binding energy between Glyzaglabrin and PRKACA was −9.4 kcal/mol, forming 4 hydrogen bonds with an average bond length of 2.31 Å. The connection position was identified at the amino acid residues GLU-121A, VAL-123A, GLU-127A, and THR-183A (Figure 7B). Additionally, 7,2′,4′-trihydroxy-5-methoxy-3-arylcoumarin interacting with PRKACA noted binding energy of −8.6 kcal/mol while generating 7 hydrogen bonds at the docking position with an average bond length of 2.5 Å; this connection involved amino acid residues LEU-49A, THR-51A, GLU-121A, VAL-123A, GLU-127A, and THR-183A (Figure 7F).