Abstract

Background

   The prevalence of breast cancer (BRCA), which is common among women, is
   on the rise. This study applied network pharmacology to explore the
   potential mechanism of action of herba sarcandrae in BRCA and construct
   a prognostic signature composed of inflammation-related genes.

Methods

   The active ingredients of herba sarcandrae were screened using the
   SymMap, TCMID, and TCMSP platforms, and the molecular targets were
   determined in the UniProt database. The “drug-active compound-potential
   target” network was established with Cytoscape 3.7.2. The molecular
   targets were subjected to disease ontology, gene ontology (GO), and
   Kyoto Encyclopedia of Genes (KEGG) analyses. AutoDock software was used
   for molecular docking. Differentially expressed genes (DEGs) related to
   inflammation were obtained from the BRCA Cancer Genome Atlas (TCGA)
   database. In the training cohort, the univariate Cox regression model
   was applied to preliminarily screen prognostic genes. A multigene
   signature was built by the least absolute shrinkage and selection
   operator (LASSO) regression model, followed by validation through
   Kaplan‒Meier, Cox, and receiver operating characteristic (ROC)
   analyses.

Results

   Forty-one active compounds were identified, and 265 therapeutic targets
   for herba sarcandrae were predicted. GO enrichment results revealed
   significant enrichment of biological processes, such as response to
   xenobiotic stimuli, response to nutrient levels, and response to
   lipopolysaccharide. KEGG analysis revealed significant enrichment of
   pathways such as AGE-RAGE and chemical carcinogenesis receptor
   activation signaling pathways. In addition, the herbs Marc-Andre and
   rutin were shown to mediate BRCA cell proliferation and apoptosis via
   the interferon regulatory factor 1 (IRF1)/signal transducer and
   activator of transcription 3 (STAT3)/programmed death-ligand 1 (PD-L1)
   pathway. Sixteen inflammatory signatures, including BST2, GPR132,
   IL12B, IL18, IL1R1, IL2RB, IRF1, and others, were constructed, and the
   risk score was found to be a strong independent prognostic factor for
   overall survival in BRCA patients. The 16-inflammation signature was
   associated with several clinical features (age, clinical stage, T, and
   N classifications) and could reflect immune cell infiltration in tumor
   microenvironments with different immune cells.

Conclusions

   Herba sarcandrae and rutin were shown to mediate BRCA cell
   proliferation and apoptosis via the IRF1/STAT3/PD-L1 pathway, and the
   16-member inflammatory signature might be a novel biomarker for
   predicting BRCA patient prognosis, providing more accurate guidance for
   clinical treatment prognosis evaluation and having important reference
   value for individualized treatment selection.

   Keywords: Herba sarcandrae, Rutin, IRF1, Inflammation, Prognosis, BRCA

Highlights

     * •
       Herba sarcandrae and Rutin inhibited BRCA cell proliferation and
       apoptosis.
     * •
       Herba sarcandrae and Rutin mediated the IRF1/STAT3/PD-L1 pathway.
     * •
       16-inflammation signature may be a marker for predicting the
       prognosis of BRCA.

1. Introduction

   Breast cancer (BRCA) is a leading cancer affecting females worldwide,
   carrying a significant socioeconomic burden [[39]1,[40]2]. Although the
   prognosis of many patients with BRCA may improve compared with that of
   patients with other solid tumors receiving multiple treatments,
   including radical surgery, chemotherapy, radiotherapy, and targeted
   therapy, some patients with BRCA achieve adverse outcomes [[41]3]. The
   complexity of the molecular mechanisms that regulate tumorigenesis and
   progression determines the heterogeneity of BRCA, and the choice of
   treatment regimen and disease prognosis is subject to this
   heterogeneity [[42]4]. Molecular advances have allowed a deeper
   comprehension of the cellular pathways that control BRCA development,
   facilitating advances in identifying diagnostic markers and developing
   new therapeutic strategies [[43]5]. The identification of new drugs for
   BRCA and the establishment of tools with accurate prognostic predictors
   for patients with BRCA are essential for guiding clinical management.

   TCM herba sarcandrae is a dried whole grass of grass corals belonging
   to the Chloranthaceae family and is commonly used to treat a variety of
   diseases [[44]6]. Multiple compounds, such as coumarins, phenolic acid
   flavonoids, and sesquiterpenes, have been isolated and identified from
   herba sarcandrae [[45]7]. Pharmacological studies have confirmed that
   herba sarcandrae extracts have anti-thrombocytopenia, anti-tumor,
   anti-inflammatory, and other effects [[46]8]. Due to the complexity of
   the chemical composition of plant extracts and the synergistic effects
   of various chemical components, as well as the fact that various
   chemical components interact with many targets, it is difficult to
   understand the molecular mechanisms by which they act on certain
   molecular targets. In particular, its antitumor effect has been
   clinically confirmed [[47]9]. For instance, uvangoletin extracted from
   herba sarcandrae can induce HepG2 cell apoptosis and autophagy,
   resulting in the inhibition of HepG2 cell proliferation and metastasis
   through the protein kinase B/mammalian target of rapamycin,
   mitogen-activated protein kinase (MAPK), and transforming growth
   factor-beta/SMAD family member 2 signaling pathways [[48]10].
   Rosmarinic acid extracted from herba sarcandrae repressed
   MDA-MB-231 cell proliferation, migration and facilitated apoptosis,
   which may be associated with B-cell lymphoma-2 (Bcl-2) and Bax
   (Bcl2-associated X) protein levels [[49]11]. Ethyl acetate obtained
   from herba sarcandrae induced cell cycle arrest and increased the
   Bax/Bcl-2 ratio in leukemic HL-60 cells [[50]12]. Although herba
   sarcandrae has been revealed to have anti-tumor effects on BRCA, the
   active ingredients and molecular mechanisms of herba sarcandrae against
   tumors are unclear and need to be further studied.

   Biomarkers based on inflammatory responses or signaling pathways are
   expected to predict survival and guide the design of personalized
   therapies for BRCA patients [[51]13]. This is primarily due to the
   effect of inflammation on the tumor phenotype and clinical treatment
   efficacy [[52]14]. Inflammation has become one of the main markers of
   cancer progression [[53]15] and plays key roles in tumor initiation and
   outcome [[54]16,[55]17]. However, in the field of clinical management,
   there are still knowledge gaps regarding the ability of
   inflammation-specific prognostic features to predict long-term survival
   in patients with BRCA.

   Therefore, we used network pharmacology methods to identify gene
   targets and pathways involved in the interaction between herba
   sarcandrae and BRCA and to investigate the effect of herba sarcandrae
   and its active ingredient rutin on BRCA cell malignancy. By utilizing
   bioinformatics technology to mine and obtain RNA and clinical data from
   the BRCA-TCGA database, differentially expressed genes (DEGs) related
   to inflammation were screened, and prognostic models were constructed
   to compare significant differences in survival between the high-risk
   and low-risk groups.

2. Materials and methods

2.1. Screening of active ingredients and prediction of targets in herba
sarcandrae

   Multiple drug-related databases for symptom mapping (SymMap,
   [56]http://www.symmap.org/), the Traditional Chinese Medicine
   integrative database (TCMID, [57]http://www.megabionet.org/tcmid/), and
   Traditional Chinese Medicine Systems Pharmacology (TCMSP,
   [58]https://old.tcmsp-e.com/tcmsp) were used to retrieve the relevant
   targets of the active ingredients of herba sarcandrae, and the targets
   were converted into corresponding genes with the help of the UniProt
   database ([59]https://www.uniprot.org/).

2.2. Construction of the “drug-active compound-potential target (Dacpt)”
network

   The Dacpt network was established by Cytoscape 3.7.2
   ([60]https://cytoscape.org/). For this network, both active compounds
   and potential targets of herba sarcandrae were included. Node colors
   and sizes were fashioned based on degree values.

2.3. Disease ontology (DO), gene ontology (GO), and Kyoto Encyclopedia of
Genes and Genomes (KEGG) enrichment analyses

   DO, GO, and KEGG enrichment analyses of the relevant targets of the
   active ingredients of herba sarcandrae were carried out by the “cluster
   Profiler” package (P < 0.05). The diseases associated with the targets
   of herba sarcandrae active ingredients were identified through DO
   analysis. The biological processes, molecular functions, and cellular
   components related to the targets of herba sarcandrae active
   ingredients were identified through GO analysis. In addition, KEGG
   analysis identified the signaling pathways enriched by active
   ingredient potential targets of herba sarcandrae.

2.4. Screening of differentially expressed genes (DEGs) linked to the
inflammatory response in BRCA

   Clinical and transcriptome expression data associated with BRCA were
   downloaded from the Gene Expression Omnibus (GEO) and The Cancer Genome
   Atlas (TCGA) ([61]https://portal.gdc.cancer.gov/) databases. Clinical
   data were collected and summarized using the “limma” package in R
   software to construct an expression matrix related to the inflammatory
   response. The screening of DEGs was performed according to a false
   discovery rate (FDR) < 0.05 and a |log[2]-fold change (FC) |≥0.585.
   Kaplan‒Meier plots were generated for overall survival (OS) analysis.

2.5. Docking and verification of potential active ingredients with core
target molecules

   Mol*2 files were downloaded from the PubMed database. The PDB format of
   the core protein domains was downloaded from the PDB database
   ([62]http://www.rcsb.org/). The LibDock module of Discovery Studio 2019
   was used to complete the virtual screening as follows: water molecules
   were removed, and DS2019 was applied to minimize protein and ligand
   energy. For the correct ionization n and tautomeric states of amino
   acid residues, all nonpolar hydrogens are merged (removed), and partial
   atomic charges are assigned using the Gasteiger–Marsili method.
   Molecular docking was further performed to evaluate the possible
   binding mode between the herba sarcandrae component and the core
   protein binding site. The three compounds with the highest LibDockScore
   values were selected for subsequent binding energy and RMSD
   calculations as follows: PyMOL software was used to dehydrate and
   dephosphate the protein. AutoDockTools 1.5.7 software was used to
   convert the PDB format of the top 6 most active drug components and the
   core protein-encoding gene files into pdbqt format. A Vina script was
   used to calculate the molecular binding energy and display the
   molecular docking results. The molecular docking results of the
   ligand‒receptor complex were displayed in 3D and 2D to evaluate the
   reliability of the bioinformatics analysis.

   Data on the underlying active components of rutin were retrieved from
   the TCMSP database. The 3D structures of TRF1 docking targets were
   determined with the Worldwide Protein Data Bank (PDB) database
   ([63]https://www.rcsb.org/). The PDB ID is [64]4CRL. The docking
   results were visualized using AutoDock Tools in conjunction with PyMol
   software.

2.6. Cell culture

   The mouse BRCA cell line 4T1 (CL-0007, Procell, Wuhan, China) and the
   human BRCA cell line MDA-MB-231 (human) (CL-0150, Procell) were
   cultured in Roswell Park Memorial Institute (RPMI)-1640 medium
   (Procell) and Leibovitz's L-15 medium (Procell) supplemented with 10 %
   fetal bovine serum (FBS; Thermo Fisher Scientific, Pittsburgh, PA, USA)
   and 1 % penicillin‒streptomycin (ScienCell, San Diego, California,
   USA), respectively, in a humid incubator at 37 °C and 5 % CO[2].

2.7. Cell counting Kit-8 (CCK-8)

   4T1 (mouse) and MDA-MB-231 (human) cell lines were seeded and treated
   with concentrations of Epimedin C (#HY-N0260, MCE, USA), rutin (#N1833,
   APExBIO, USA), or β-sitosterol (#N1688, APExBIO) (0, 10, 50, 100, 500,
   500 μM) for 48 h. Ten microliters of CCK-8 reagent (#G4103, Servicebio,
   Wuhan, China) were added. The absorbance at 450 nm was measured using a
   microplate reader (Thermo).

2.8. Flow cytometry assay

   BRCA cell lines were seeded in 96-well plates and treated with rutin
   (0, 10, 50, 100, 500, or 500 μM) for 48 h. Assays for apoptosis were
   performed using an Annexin-V-FITC and propidium iodide (PI)
   double-staining kit (#100-101-100, GOONIE, Guangzhou, China) based on
   the manufacturer's instructions. Ultimately, the detection of stained
   cells was accomplished by flow cytometry (Miltenyi BiotecTM FACS Quant
   10; Milteny Biotech, Germany). The extracted data were analyzed using
   FlowJo software (Tree Star, San Carlos, CA).

2.9. Coimmunoprecipitation (Co-IP)

   BRCA cell lines treated with or without rutin (50 μM) were lysed in
   RIPA buffer. Lysates and corresponding antibodies were incubated
   overnight (4 °C). Then, magnetic beads (MCE, HY-K0208, treated with 3 %
   BSA before use) were incubated with protein-antibody complexes
   overnight at 4 °C. The beads were washed 5 times with precooled wash
   buffer, followed by the addition of 5 × sample buffer and cooking at
   95 °C for 10 min. The supernatant was extracted by centrifugation
   (12,000 rpm, 15 min, 4 °C). The relevant proteins were analyzed by
   western blotting.

2.10. Western blotting

   Total protein was extracted using a protein extraction kit (#P0013C,
   RIPA Lysis Buffer, Beyotime). The protein concentration was measured
   using the BCA method (Biosharp BL521A). Equal amounts of proteins
   (30 μg) were separated and transferred to PVDF membranes. Following
   encapsulation in 5 % skim milk, the PVDF membranes were incubated with
   the following primary antibodies: anti-IRF1 (1:1000, #a7692, ABclonal,
   Wuhan, China), anti-phospho-(Ser/Thr) Phe (1:1000, Cell Signaling
   Technology, #9631), anti-STAT3 (1:1000, #10253-2-ap, Proteintech,
   Wuhan, China), anti-phospho-STAT3-Y705 (1:200, #AP0070, ABclonal),
   anti-PD-L1 (1:1000, #A1645, ABclonal), and GAPDH (1:50000, #60004-1-Ig,
   Proteintech). Following incubation with secondary antibodies
   (#BL003A/#BL001A, Biosharp, Beijing, China), an enhanced
   chemiluminescence (ECL) system was used to visualize the protein bands,
   followed by quantification with ImageJ software.

2.11. Tumor-bearing nude mouse model construction

   Briefly, 15 female BALB/C nude mice (3–4 weeks old) (Gempharmatech,
   Jiangsu, China) were utilized (5 mice/group). The 4T1 cell suspension
   (100 μL, 5 × 10^7/mL) was subcutaneously implanted in the right armpit
   of nude mice. After tumorigenesis (0.5–0.7 cm^3), the mice were
   administered Zhongjiefeng (STATE MEDICAL PERMITMENT No. Z20090321,
   5 mg/kg, gavage, once a day) and rutin (500 mg/kg, intraperitoneal
   injection, once a day). Tumor size was calculated by the following
   formula: volume (mm^3) = length × width^2/2. Twenty days later, the
   tumors were removed after the animals had been euthanized, followed by
   measurement of the tumor weight and processing of the tumor tissues for
   subsequent analyses. The study was approved by the Ethics Committee of
   Guangzhou Forevergen Medical Laboratory Animal Center (approval number
   IACUC-AEWC-F2303012).

2.12. Immunohistochemical (IHC) analysis

   The excised tumor samples were embedded and fixed. Sections were
   microwaved in Retrieval Solution (Amyjet, Wuhan, China) at pH 6.0 or pH
   9.0 for 7 min, followed by incubation in 3 % H[2]O[2] for 10 min.
   Nonspecifically bound sites were then blocked with 2.5 % normal horse
   serum (Vector Biolabs, Burlingame, CA, USA). The sections were
   incubated with an antibody against Ki67 (#ab279653, Abcam, USA). After
   incubation with a secondary antibody (#ab125913, Abcam) for 30 min, the
   sections were stained with 3,3′-diaminobenzidine solution (Sangon
   Biotech, Shanghai, China) and hematoxylin (Sangon Biotech).
   Representative areas were viewed under a light microscope (Nikon 90i,
   Tokyo, Japan).

2.13. Immunofluorescence (IF)

   The sections were routinely dewaxed and hydrated, followed by the
   addition of 3 % H[2]O[2] to remove endogenous oxidative enzymes. After
   antigen retrieval by boiling in 0.01 M sodium citrate buffer solution
   (pH 6.0) (Sangon Biotech) and sealing with 5 % BSA, sections (5 μM)
   were incubated overnight with antibodies against CD4 (#A0363, 1:1000,
   ABclonal), CD8 (#A11856, 1:1000, ABclonal), or PD-L1 (#A1645, 1:1000,
   ABclonal) at 4 °C. After washing with PBS three times, the sections
   were incubated for 2 h with a secondary antibody (ZF-0516/ZF-0512,
   Zhongshan Jinqiao Technology Co., Ltd., Beijing, China). After staining
   with 4’,6-diamino-2-phenylindole (DAPI), the results were observed
   using a fluorescence microscope (M152-N, Mingmei, Shandong, China).

2.14. Statistical analysis

   The experimental data were analyzed using GraphPad Prism 8.0. The
   values from three independent experiments are presented as the
   mean ± standard deviation (SD). Significant differences between the two
   groups were identified by t-tests. Comparisons involving multiple
   groups were performed using a one-way analysis of variance (ANOVA). A
   criterion of p < 0.05 was established for statistical significance.

3. Results

3.1. Corresponding targets of the active ingredients of herba sarcandrae

   To investigate the relationship between the active ingredients of herba
   sarcandrae and the corresponding targets in BRCA, the regulatory
   network of herba sarcandrae-active ingredient targets was first
   explored through multiple databases. Briefly, the relevant targets of
   the active ingredient of herba sarcandrae were retrieved through
   multiple drug-related databases, such as SymMap, TCMID, and TCMSP, and
   the gene names of the matching targets were corrected via the UniProt
   database ([65]https://www.uniprot.org/). A total of 41 active
   ingredients of herba sarcandrae were identified, and 265 potential
   targets of these active ingredients were identified. A protein-protein
   interaction (PPI) network of herba sarcandrae-active ingredient-targets
   was constructed with the help of Cytoscape 3.7.2 software ([66]Fig.
   1A). We then conducted DO/GO/KEGG enrichment analysis on potential
   targets of herba sarcandrae active ingredients. DO analysis found that
   a total of 67 genes were enriched in BRCA, including
   RELA/CCNA2/ESR1/PTGS2/PPARG/MAPK14/GSK3B/CCND1/ESR2 ([67]Fig. 1B). The
   potential targets of herba sarcandrae active ingredients were mainly
   enriched in response to xenobiotic stimulus, response to nutrient
   levels, and so on ([68]Fig. 1C). KEGG analysis presented that the
   potential active ingredient targets of herba sarcandrae were mainly
   enriched in the AGE-RAGE signaling pathway in diabetic complications,
   lipid and atherosclerosis, fluid shear stress and atherosclerosis, and
   other pathways ([69]Fig. 1D).

Fig. 1.

   [70]Fig. 1
   [71]Open in a new tab

   Composition-target network of herba sarcandrae. (A) Regulatory network
   of herba sarcandrae-active ingredient-targets generated using Cytoscape
   (https://cytoscape.org/; ver. 3.7.2). (B) Disease Ontology (DO)
   analysis of the diseases associated with the targets of the active
   ingredients of herba sarcandrae. (C) Gene Ontology (GO) analysis of the
   biological processes, molecular functions, and cellular components
   related to the targets of the active ingredients of herba sarcandrae.
   (D) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis
   of the signaling pathways enriched by the potential targets of the
   active ingredients of herba sarcandrae.

3.2. DEGs related to the inflammatory response in BRCA

   To analyze breast cancer-associated inflammatory genes, we downloaded
   clinical transcriptome expression data associated with BRCA from the
   TCGA database and used the R software “limma” package to collate and
   summarize the downloaded data to further derive the expression matrix
   associated with the inflammatory response (a total of 200 inflammatory
   response-associated genes were included). DEGs related to the
   inflammatory response in BRCA were screened (FDR value < 0.05,
   |log[2]FC| ≥ 0.585). Fifty-five upregulated mRNAs and 46 downregulated
   mRNAs were screened, and volcano heatmaps and expression matrix
   heatmaps of these DEGs were generated, as shown in [72]Fig. 2A and B.
   In addition, we conducted univariate analysis on these DEGs and found
   that high expression of IL1R1 and KCNMB2, as well as low expression of
   BST2, CCL17, CCL5, CD48, CXCL6, CXCL9, GPR132, IL12B, and IL18, was
   strongly linked to BRCA prognosis ([73]Fig. 2C). The aforementioned
   group of inflammatory response genes that affect outcomes for BRCA
   patients was used for survival prognosis analysis using TCGA-BRCA or
   GEO-BRCA clinical data. The results showed that the BST2, CXCL6,
   GPR132, IL12B, IL18, IL1R1, IL2RB, IRF1, IRF7, KCNMB2, LCK, RGS1, SELL,
   TACR1, TAPBP, and TNFRSF1B genes had significant impacts on the
   survival and prognosis of patients with BRCA ([74]Fig. S1A-P). Among
   them, BRCA patients with high expression of BST2, CXCL6, GPR132, IL12B,
   IL18, IL2RB, IRF1, IRF7, LCK, RGS1, SELL, TACR1, TAPBP, or TNFRSF1B had
   a better prognosis than those with low expression of these genes. In
   contrast, BRCA patients with higher expression of IL1R1 or KCNMB2 had a
   worse prognosis.

Fig. 2.

   [75]Fig. 2
   [76]Open in a new tab

   Identification and annotation of DEGs associated with the inflammatory
   response in breast cancer (BRCA) (A) Volcano plot of DEGs related to
   the inflammatory response in BRCA based on data from The Cancer Genome
   Atlas (TCGA). (B) Heatmap of DEGs related to the inflammatory response
   in BRCA. Blue represents the normal sample; pink represents the tumor
   sample. (C) Forest plots from univariate Cox regression analysis of
   TCGA data.

3.3. Construction of risk score models associated with the BRCA inflammatory
response

   To explore the association of immune-inflammation-related DEGs with
   survival, we constructed prognostic models for the above 16 INF genes.
   According to the risk score, all patients in the TCGA cohort were
   clustered into high-risk (525 patients) and low-risk (526 patients)
   groups, and survival was significantly lower in BRCA patients in the
   high-risk subgroup ([77]Fig. 3A). Similarly, overall survival was
   significantly reduced in BRCA patients in the high-risk group according
   to the GEO database ([78]Fig. 3B). However, patients suffering from
   BRCA within the high-risk group had considerably shorter survival than
   those within the low-risk group ([79]Fig. 3C). Univariate Cox
   regression model analysis revealed that age, tumor stage, T and N
   classification, and risk score may influence overall survival in
   patients with BRCA ([80]Fig. 3D). The factors that may affect the
   prognosis of patients with BRCA were incorporated into a multivariate
   Cox regression model, and the results showed that age, tumor stage, and
   risk score were found to be independent prognostic factors for BRCA
   ([81]Fig. 3E). Subsequently, we performed ROC analysis to evaluate the
   performance of a model containing 16 INF genes in predicting prognosis.
   The prognostic prediction of the 16-INF gene risk score model for BRCA
   patients had an AUC of 0.639, which was superior to that of age and T
   and N classifications ([82]Fig. 3F). In addition, we constructed ROC
   curves for the 1-, 3-, and 5-year periods, and the results showed that
   the AUC of these three groups was greater than 0.6, suggesting that the
   model containing 16 INF genes had good accuracy in predicting survival
   rates within 3 years for BRCA patients ([83]Fig. 3G).

Fig. 3.

   [84]Fig. 3
   [85]Open in a new tab

   Construction and validation of a BRCA inflammatory response-related
   risk scoring model. (A–B) OS of patients in the TCGA-BRCA (A) and
   GEO-BRCA (B) cohorts. (C) PFS of patients in the TCGA-BRCA cohort.
   (D–E) Univariate (D) and multivariate (E) analyses were conducted using
   the 16-INF gene signature and clinical covariates based on overall
   survival in the TCGA-BRCA cohort. (F) ROC analysis of the specificity
   and sensitivity of prognosis prediction by the 16-interferon (INF) gene
   risk score, age, T and N classifications, and tumor stage in the
   TCGA-BRCA cohort in terms of overall survival. (G) Validation of the
   prognostic value of the survival-dependent receiver operating
   characteristic (ROS) curve at 1, 2, and 3 years in the TCGA cohort.

3.4. Correlations of the risk score models with clinical features and immune
cell functions

   To investigate whether this risk model is associated with the clinical
   features of patients with BRCA, the Wilcoxon rank sum test was carried
   out. BRCA patients with older age, later TNM stage, and higher tumor
   grade were observed in the high-risk group ([86]Fig. 4A–D). The
   function of the 16-INF gene (overexpression) risk model in the immune
   response was described. The correlation between the risk score and
   infiltrating immune cells was analyzed, and the results showed that the
   16-INF gene (overexpression) risk model did not significantly differ
   among the six types of immune infiltration, but it was mainly enriched
   in the C1 and C3 subtypes ([87]Fig. 5A). The study inferred the
   proportion of immune cells in 22 types of infiltrating immune cells
   from standardized gene expression data using the CIBERSORT algorithm
   with 1000 permutations. The low-risk group had higher proportions of T
   cells CD8^+, B-cell memory, activated T cells CD4^+ memory, T cells
   follicular helper, regulatory T cells, activated NK cells, monocytes,
   and M1 macrophages than did the high-risk group, but the high-risk
   group had significantly greater proportions of M2 macrophages, resting
   NK cells, M0 macrophages, and resting mast cells than did the low-risk
   group ([88]Fig. 5B). Moreover, its immune functions were mainly
   concentrated in check-point, cytolytic_activity, APC_co_inhibition,
   MHC_class_I, APC_co_stimulation, CCR, HLA, inflammation-promoting, and
   so on ([89]Fig. 5C).

Fig. 4.

   [90]Fig. 4
   [91]Open in a new tab

   Relationships between the risk model and clinical features. (A–D)
   Analysis of the relationships between the 16-INF gene risk score and
   clinical features such as age, T and N classifications, and tumor stage
   in the TCGA-BRCA cohort.

Fig. 5.

   [92]Fig. 5
   [93]Open in a new tab

   Relationships between the risk model and immune response. (A) Analysis
   of the relationship between the 16-INF gene risk score and six types of
   infiltrating immune cells. (B) Analysis of the degree of immunocyte
   infiltration in the 22 immunocyte subunits in the high-risk and
   low-risk groups. (C) Differences in immune-related functions between
   the high- and low-risk groups. (*p < 0.05; **p < 0.01; ***p < 0.001).

3.5. GO and KEGG enrichment analyses of prognostic risk genes

   To determine the relevance of the BRCA inflammatory response gene
   cluster to immune cell function, the Bioconductor package and
   clusterProfiler package in R were used to perform GO analysis and KEGG
   pathway enrichment analysis of 16 INF genes and DEGs associated with
   prognostic risk genes. GO analysis revealed that the enriched
   biological processes of these genes included positive regulation of
   leukocyte activation, positive regulation of cell activation, and
   immunoglobulin production. The components were mainly enriched in
   circulating, immunoglobulin complex, immunoglobulin complex, and the
   external side of the plasma membrane. The enriched molecular functions
   included C–C chemokine binding, immunoglobulin receptor binding, and
   antigen binding. Furthermore, they focused mainly on the T-cell
   receptor signaling pathway, Th17 cell differentiation, viral protein
   interaction with cytokine and cytokine receptors, hematopoietic cell
   lineage, primary immunodeficiency, cytokine‒cytokine receptor
   interaction, and other pathways. In addition, we performed PPI and
   topology analyses, the network diagram of which is shown in [94]Fig.
   6A–D.

Fig. 6.

   [95]Fig. 6
   [96]Open in a new tab

   The core gene network of inflammation genes related to BRCA. (A) The
   core targets of PPIs. (B) The distribution network of degree core
   targets. (C) The distribution network of stress core targets. (D) The
   distribution network of betweenness core targets. (E) Molecular docking
   results of the core protein IRF1 with the active compound rutin.

3.6. Core protein screening and molecular docking

   To screen the most core drug-sensitive proteins, the 41 active
   ingredients of herba sarcandrae mined from the results were molecularly
   docked to identify potential drugs targeting the core protein IRF1
   (mutual mapping and matching between herba sarcandrae targets and
   BRCA-INF prognostic risk genes). The pdb structure of the core protein
   IRF1 was downloaded from the PDB (PDB ID: [97]4CRL). We downloaded the
   sdf files of 41 active ingredients from the PubChem database and docked
   all of them with the core protein IRF1. Based on the LibDockScore
   (LibDockScore) values, we found that IRF1 might bind strongly to
   epimedin C, lignoceric acid, docosanoate, pelargonidin-3-rhamnosyl
   glucose, Engelitin, sitoglucide, beta-sitosterol, sitosterol, and
   rutin, suggesting that these compounds may be drugs that target IRF1.
   We further used AutoDock-Vina to dock drugs to more accurately identify
   drugs that bind to the IRF1 domain. Epimedin C, lignoceric acid,
   Docosanoate, pelargonidin-3-rhamnosyl glucoside, Engelitin,
   Sitogluside, beta-sitosterol, sitosterol, rutin, and 9 other active
   ingredients were selected as drug candidates for the IRF1 active site,
   and their affinities were ≤-5.0 kcal/mol. Additionally, the RMSD values
   were <2.00, except for those of rutin and the core protein IRF1,
   indicating that the docking model was reliable. Combining the LibDock
   score, binding energy, and RMSD values, the docking complex formed by
   Sitogluside and the core protein IRF1 was the best ([98]Fig. 6E).
   Therefore, the above 9 active ingredients may be used as drugs that
   target the IRF1 protein.

3.7. Rutin mediated proliferation, apoptosis and the IRF1/STAT3/PD-L1 pathway
in BRCA cells

   To analyze the effects of the traditional Chinese medicine monomers
   epimedin C, rutin, and β-sitosterol on BRCA cell lines, we treated BRCA
   cells with concentrations of sitogluside, epimedin C, rutin, and
   β-sitosterol (0, 10, 50, 100, 500 μM) for 48 h, respectively. The CCK-8
   assay results also showed that the effect of Sitogluside treatment was
   not very significant ([99]Fig. S2). CCK-8 assays further showed that
   the monomers epimedin C, rutin, and β-sitosterol at concentrations
   greater than or equal to 50 μM could significantly repress the activity
   of BRCA cells, and rutin could inhibit BRCA cell activity at a
   concentration of 10 μM ([100]Fig. 7A). Subsequently, rutin was selected
   for further analysis. Moreover, rutin facilitated BRCA cell apoptosis
   in a concentration-dependent manner ([101]Fig. 7B). Western blotting
   data showed that rutin markedly upregulated IRF1 in mouse 4T1 and human
   MDA-231 cells, and the expression level of IRF1 gradually increased
   with increasing rutin concentration ([102]Fig. 7C). Co-IP revealed that
   rutin (50 μM) weakened the phosphorylation of Phe in BRCA cells
   ([103]Fig. 7D). We further analyzed the effect of rutin on the
   expression of the immune-associated genes STAT3, p-STAT3, and PD-L1,
   which are located downstream of IRF1. Western blotting revealed that
   rutin markedly reduced p-STAT3 and PD-L1 protein levels in BRCA cells
   ([104]Fig. 7E). Overall, rutin mediated proliferation, apoptosis, and
   the IRF1/STAT3/PD-L1 pathway in BRCA cells.

Fig. 7.

   [105]Fig. 7
   [106]Open in a new tab

   Effects of rutin on proliferation, apoptosis, and the IRF1/STAT3/PD-L1
   pathway in BRCA cells. (A) CCK-8 analysis of the effects of different
   concentrations of epimedin C, rutin, or β-sitosterol (0, 10, 50, 100,
   500 μM) on BRCA cell viability. (B) Flow cytometry analysis of the
   effect of rutin (50 μM) on the apoptosis of BRCA cells. (C) Mouse 4T1
   and human MDA-231 cells were treated with 0, 10, 100, or 500 μM rutin
   for 48 h, and the protein level of IRF1 was detected by western
   blotting. Image was cropped. (D) Co-IP analysis of the interaction
   between rutin and IRF1. Image was cropped. (E) Western blot analysis of
   p-STAT3/STAT3/PD-L1 protein levels in MDA-MB-231 cells treated with or
   without rutin. Image was cropped. (*p < 0.05; **p < 0.01;
   ***p < 0.001).

3.8. Herba sarcandrae and rutin decreased tumor growth in a nude mouse
tumorigenic model

   To investigate the effect of herba sarcandrae and rutin on BRCA growth
   in vivo, we constructed a nude mouse tumorigenesis model with 4T1
   cells. The results showed that the size of the tumors formed by 4T1
   cells was smaller in the herba sarcandrae and rutin groups than in the
   control group ([107]Fig. 8A–C). IHC staining revealed fewer
   ki67-positive cells in tumor samples from the herba sarcandrae and
   rutin groups than in tumor samples from the control group ([108]Fig.
   8D). IF staining revealed an increase in CD4 and CD8 expression and a
   decrease in PD-L1 expression in tumor samples derived from SMN- and
   rutin-treated mice ([109]Fig. 8E). Furthermore, herba sarcandrae and
   rutin treatment decreased the level of IRF1 protein phosphorylation in
   mouse tumors, as evidenced by co-IP ([110]Fig. 8F). In addition, herba
   sarcandrae and rutin treatment decreased p-STAT3, STAT3, and PD-L1
   protein levels in mouse tumors ([111]Fig. 8G). Overall, herba
   sarcandrae and rutin decreased BRCA growth in a nude mouse tumorigenic
   model by blocking the IRF1/STAT3/PD-L1 pathway.

Fig. 8.

   [112]Fig. 8
   [113]Open in a new tab

   Antitumor effects of herba sarcandrae and rutin in a nude mouse tumor
   model. (A and B) Representative images of nude tumor model mice and
   tumors from different groups (n = 5 per group). (C) Time course of
   tumor growth in each group at the indicated time points. (D)
   Representative Ki-67 staining of tumors from different groups. (E) IF
   detection of CD4, CD8, and PD-L1 protein levels in tumors from
   different groups. (F) Co-IP analysis of the interaction between herba
   sarcandrae or rutin and IRF1. Image was cropped. (G) Protein levels of
   p-STAT3, STAT3, and PD-L1 in tumors from different groups. Image was
   cropped. (*p < 0.05; **p < 0.01; ***p < 0.001).

4. Discussion

   Molecular docking is a theoretical simulation method that mainly
   studies the interactions between molecules and predicts their binding
   mode and affinity [[114]18]. This approach plays a crucial role in the
   field of computer-aided drug research. The principle of spatial
   matching and energy matching aims to find the best combination of drug
   molecules and biological molecules to understand the mechanism of
   action of drugs [[115]19]. In cancer research, molecular docking helps
   to discover new anticancer drugs or drug candidates, and the
   therapeutic effect of drugs on cancer can be predicted by simulating
   the interaction between drugs and cancer-related target proteins.
   Network pharmacology is an emerging method that combines computer
   science and medicine to explain the mechanism of action of drugs on
   diseases [[116]20]. This method can be used to systematically study and
   identify bioactive compounds in traditional Chinese medicine
   formulations and visualize their multi-target and multi-pathway
   mechanisms [[117]21]. In cancer research, network pharmacology is
   helpful for comprehensively understanding the role of drugs in complex
   biological systems and for predicting and analyzing the pharmacological
   effects and potential mechanisms of drugs [[118]22]. Combining these
   two technologies gives full play to their advantages [[119]23].
   Molecular docking can provide detailed molecular interaction data for
   network pharmacology to more accurately construct interaction networks
   between drugs and cancer-related target proteins. Network pharmacology
   can analyze these interactions at the system level and reveal the
   multitarget and multi-pathway mechanisms of the action of drugs in
   cancer treatment. This combination helps to discover new anticancer
   drugs, optimize the therapeutic effect of existing drugs, and provide
   strong support for personalized cancer treatment. The combination of
   these two techniques has been used in multiple cancer studies, such as
   studies on colorectal cancer [[120]24], leukemia [[121]25], lung cancer
   [[122]26], osteosarcoma [[123]27], and breast cancer [[124]28].

   Many studies have shown that the volatile oil and extract of herba
   sarcandrae have certain inhibitory effects on cancer cells, such as
   leukemia and lung cancer cells [[125]8]. In addition, the total
   flavonoid glycoside of herba sarcandrae also has a strong killing
   effect on Aili ascites-derived cancer cells in vitro, and the
   inhibition rate of cancer cells after intraperitoneal administration
   can reach 70.8 % [[126]29]. These findings suggest that herba
   sarcandrae may play a vital role in cancer treatment, helping to slow
   the growth of tumor cells and improve the quality of life of cancer
   patients. In this study, the network of the
   “component-target-disease-pathway” pathway in the herba Sarcandrae was
   fully characterized through integrated pharmacological methods.
   Subsequently, by analyzing high-throughput data, DEGs related to
   inflammation were screened from the BRCA-TCGA database, and biomarkers
   and risk prediction genes related to BRCA inflammation were identified.
   Finally, molecular docking technology predicted the affinity between
   small drug molecules and targets, revealing the interactions between
   drug molecules and targets, elucidating their structure-activity
   relationships and validating the herba Sarcandrae's ability to modulate
   the immunosuppression of the BRCA process through IRF in breast cancer
   cells and tumor animal models.

   The research and development model of network pharmacology based on
   multiple components and multiple targets is more in line with the
   complex and diverse components and pharmacological effects of TCM. This
   study identified a total of 41 active drug molecules of herba
   sarcandrae and predicted 265 therapeutic targets for herba sarcandrae
   through integrated pharmacological methods. DO analysis revealed that a
   total of 67 genes were enriched in BRCA among 265 potential targets,
   including RELA, CCNA2, ESR1, PTGS2, PPARG, MAPK14, GSK3B, CCND1, and
   ESR2. GO analysis revealed that 265 potential targets were mainly
   enriched in response to xenobiotic stimuli, response to oxygen levels,
   response to molecules of bacterial origin, response to
   lipopolysaccharide, response to nutrient levels, etc. Moreover, 265
   potential targets were mainly enriched in pathways such as the AGE-RAGE
   signaling pathway in diabetic applications and chemical carcinogenesis
   receptor activation. These results indicated that herba sarcandrae
   could be used for the clinical treatment of tumors, metabolic diseases,
   and inflammatory diseases, which was consistent with modern research
   results and in line with the application experience of herba sarcandrae
   [[127]8,[128]30]. These results indicated that the multitarget drug
   herbsa sarcandrae can simultaneously stimulate multiple targets in the
   disease signaling pathway, providing a new avenue for BRCA clinical
   treatment.

   In this study, 101 DEGs associated with BRCA inflammation, including 55
   upregulated genes and 46 downregulated genes, were identified by
   analyzing TCGA-BRCA data. Univariate analysis and bioinformatics
   analysis revealed that the effects of the BST2 [[129]31], CXCL6
   [[130]32], GPR132 [[131]33], IL12B [[132]34], IL18 [[133]35], IL1R1
   [[134]36], IL2RB [[135]37], IRF1 [[136]38], IRF7 [[137]39], KCNMB2
   [[138]40], LCK [[139]41], RGS1 [[140]42], SELL, TACR1, TAPBP and
   TNFRSF1B [[141]43] genes on the prognosis of patients with BRCA were
   significant, and a search of the literature revealed that some of the
   16 genes are therapeutic targets and risk predictors for BRCA
   [[142]44,[143]45]. A prognostic model for the 16 inflammatory genes was
   constructed, and the predictive ability of this prognostic model was
   verified, highlighting the prognostic survival curves and plots
   containing 16 gene signatures and clinicopathological parameters
   showing great prognostic efficacy, which might allow clinicians to make
   judgments about outcomes for individual patients. Overall, high
   expression of IL1R1 and KCNMB2 and low expression of BST2, CCL17, CCL5,
   CD48, CXCL6, CXCL9, GPR132, IL12B, and IL18 are related to BRCA patient
   survival and prognosis. However, subsequent experimental verification
   of these genes is still needed to identify potential clinical
   therapeutic targets.

   We also conducted a functional analysis of 16 gene signatures. These
   genes were mainly related to biological processes such as positive
   regulation of leukocyte activation, positive regulation of cell
   activation, and immunoglobulin production. These pathways were mainly
   related to cytokine receptor interactions, lineage, primary immunity,
   and hematopoietic cell viral protein interactions with cytokines and
   cytokine receptors. Through the above results, we easily found that the
   prognostic model was mainly related to the immune response, as
   inflammation is significantly associated with the human immune response
   [[144]46]. TILs are a mixture of proinflammatory immune cells and
   immunosuppressive cells [[145]47]. Interstitial TILs are more important
   for the prognosis of patients with BRCA, as they are not affected by
   tumor nest density or growth patterns [[146]48]. TILs can clear tumor
   cells, recognize tumor antigens, and induce antitumor immune responses
   [[147]49]. High levels of TILs can predict favorable prognosis for BRCA
   patients [[148]50] and are strongly associated with high pathological
   complete response rates in BRCA patients receiving neoadjuvant therapy
   [[149]51,[150]52]. This study analyzed the effect of this prognostic
   model on TILs. Analysis confirmed that low-risk patients often had
   greater numbers of TILs and stronger immune pathway activity. High
   levels of TILs suggest that patients have a better prognosis [[151]50],
   so the low-risk group in this model has a better prognosis.

   IRF1 is the most important transcription factor in the interferon (IFN)
   pathway. IRF1 is induced by type I IFN (IFN-α, IFN-β), type II IFN
   (IFN-γ), IL-1, IL-6, and TNF-α [[152]53]. IRF1, a transcription factor,
   can activate a series of target genes, stimulate the production of
   inflammatory factors, promote immune cell differentiation and
   maturation, inhibit cell proliferation, and induce cell apoptosis. IRF1
   plays an essential role during the EMT process in BRCA [[153]38].
   Several studies have associated the transcription factor IRF1 with
   tumor-suppressive activities [[154]54,[155]55]. SPOP mutants are
   deprived of the capability to degrade IRF1 and upregulate PD-L1
   expression, thus accelerating tumor immune escape in endometrial cancer
   [[156]56]. CD8^+ T cells in IRF1-deficient tumors suppress enhanced
   cytotoxicity in vivo. Tumor cells lacking IRF1 lose the ability to
   upregulate PD-L1 expression both in vivo and in vitro and are more
   susceptible to T-cell-mediated killing [[157]54]. PD-L1 is upregulated
   in many cancers due to exogenous cellular stress. Studies have shown
   that PD-L1 is upregulated by STAT and IRF1 signaling in response to DNA
   double-strand breaks (DSBSs) [[158]55]. This study conducted molecular
   docking on 41 active components of herba sarcandrae and identified 9
   active components, namely, epimedin C, lignoceric acid, docosanoate,
   pelargonidin-3-rhamnosyl glucoside, Engelitin, sitoglucide,
   beta-sitosterol, sitosterol, and rutin, as candidate drugs targeting
   IRF1. Functionally, the monomers epimedin C, rutin, and β-sitosterol
   inhibited the activity of BRCA cells in a concentration-dependent
   manner. Subsequently, rutin was selected for further analysis.
   Moreover, rutin also promoted the apoptosis of BRCA cells in a
   concentration-dependent manner. Tumorigenesis in nude mice verified the
   tumor-inhibiting effect of herba sarcandrae and rutin on BRCA.
   Moreover, rutin weakened the phosphorylation of Phe in BRCA cells.
   Whether the docking site of the active ingredient of herba sarcandrea
   on IRF1 is the key site of the downstream cascade of phosphorylation
   remains to be further proven. In addition, rutin reduced p-STAT3 and
   PD-L1 protein levels in BRCA cells and subcutaneous tumors. These
   results showed that herba sarcandrae and rutin mediated cell
   proliferation and apoptosis via the IRF1/STAT3/PD-L1 pathway in BRCA.

   The novelty of this study was that it was the first to explore and
   verify the mechanism of action of herba sarcandrae in the treatment of
   BRCA using emerging network pharmacology and that the potential of a
   16-level inflammatory signature to predict prognosis in patients with
   BRCA has also been identified. In addition, the disadvantage of this
   study was that the different molecular types of BRCA had not been
   studied, which could be further analyzed in the future.

   In summary, based on network pharmacology and molecular docking, the
   inhibitory effects of herba sarcandrae and rutin on BRCA might be
   mediated by the IRF1/STAT3/PD-L1 pathway through the key target IRF-1.
   The 16-level inflammatory signature could serve as a novel biomarker
   for predicting BRCA patient prognosis, providing more accurate guidance
   for clinical treatment prognosis evaluation and having important
   reference value for individualized treatment selection. However, breast
   cancer contains multiple subtypes, and whether this study applies to
   other subtypes of breast cancer is a major limitation of this study.

Ethical statements

   The study was approved by the Ethics Committee of Guangzhou Forevergen
   Medical Laboratory Animal Center (approval number IACUC-AEWC-F2303012).

Funding

   Our work supported by the Guangdong Basic and Applied Basic Research
   Foundation [Grant number 2023A1515140028 and 2021A1515220130] and the
   Dongping talent plan of Foshan Fosun Chancheng Hospital [Grant number
   047].

Data availability

   Data included in article/supp. material/referenced in article.

CRediT authorship contribution statement

   Jie Yuan: Writing – review & editing, Writing – original draft,
   Validation, Project administration, Methodology, Investigation, Funding
   acquisition, Formal analysis, Data curation, Conceptualization. Minxia
   Lin: Writing – review & editing, Writing – original draft, Validation,
   Methodology, Investigation, Formal analysis, Data curation,
   Conceptualization. Shaohua Yang: Writing – review & editing,
   Visualization, Validation, Software, Resources, Methodology,
   Investigation, Data curation, Conceptualization. Hao Yin: Writing –
   review & editing, Visualization, Validation, Software, Resources,
   Methodology, Investigation, Data curation, Conceptualization. Shaoyong
   Ouyang: Writing – review & editing, Visualization, Validation,
   Software, Resources, Methodology, Investigation, Data curation,
   Conceptualization. Hong Xie: Writing – review & editing, Visualization,
   Validation, Software, Resources, Methodology, Investigation, Data
   curation, Conceptualization. Hongmei Tang: Writing – review & editing,
   Visualization, Validation, Software, Resources, Methodology,
   Investigation, Data curation, Conceptualization. Xiaowei Ou: Writing –
   review & editing, Visualization, Validation, Software, Resources,
   Methodology, Investigation, Data curation, Conceptualization. Zhiqiang
   Zeng: Writing – review & editing, Validation, Project administration,
   Methodology, Investigation, Formal analysis, Data curation,
   Conceptualization.

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Abbreviation

   BRCA
          Breast cancer

   TCM
          Traditional Chinese Medicine;

   TCMID
          Traditional Chinese Medicine integrative database

   TCMSP
          Traditional Chinese Medicine Systems Pharmacology

   GO
          Gene ontology

   DO
          Disease ontology

   KEGG
          Kyoto encyclopedia of genes

   CCK-8
          Cell counting kit-8

   IHC
          Immunohistochemical

   IF
          Immunofluorescence

   DEGs
          Differentially expressed genes

   TCGA
          The cancer genome atlas

   GEO
          Gene Expression Omnibus

   Co-IP
          Co-immunoprecipitation

   SD
          Standard deviation

   PD-L1
          Programmed death-ligand 1

   STAT3
          Signal transducer and activator of transcription 3

   IRF1
          Interferon regulatory factor 1

Footnotes

   ^Appendix A

   Supplementary data to this article can be found online at
   [159]https://doi.org/10.1016/j.heliyon.2024.e31137.

Contributor Information

   Jie Yuan, Email: candlewjy@163.com.

   Zhiqiang Zeng, Email: zengzhiqiang@fsccyy.com.

Appendix A. Supplementary data

   The following are the Supplementary data to this article.
   Multimedia component 1
   [160]mmc1.csv^ (294.2KB, csv)
   Multimedia component 2
   [161]mmc2.docx^ (5.1MB, docx)
   Multimedia component 3
   [162]mmc3.pdf^ (195.6KB, pdf)

References