Qin-Shuai Zhang, Li-Li Sha, Hao Zhang ,Yuan-Long Liu ,Min Liu*

1Technology Department, Xiamen Bencao Zhenyuan Pharmaceutical Technology Co., Ltd, Xiamen China; 2General Medicine Department, Rizhao People's Hospital, Rizhao China; 3Department of Traditional Chinese Medicine, Rizhao People's Hospital,Rizhao China; 4Technology Department,Anguo Yaodu Traditional Chinese Medicine Museum Co., Ltd, Anguo, Hebei Province,China; 5Integrated Traditional Chinese and Western Medicine Department, The Fifth People's Hospital of Weifang City, Weifang China.

Abstract Background: Oral administration of indigo naturalis (IN) is used as a complementary and alternative medicine(CAM) regimen for the treatment of myelodysplastic syndromes (MDS).However, its mechanism of action has not been fully elucidated and needs to be further explored.Methods: By searching the traditional Chinese medicine system and analyzing platforms (TCMSP), bioinformatics analysis tool for the molecular mechanism of traditional Chinese medicine (BATMAN-TCM), and Swiss Target Prediction network database, the main active components and potential targets of IN were obtained.Based on this, a component-target network was established by Cytoscape 3.6.1 software.Differentially expressed genes (DGEs) in MDS were obtained from three GEO (Gene Expression Omnibus) gene chips.Then, the protein-protein interaction (PPI) network of DGEs was constructed and analyzed by STRING database and Cytoscape 3.6.1 software.In addition, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) biological enrichment analysis were carried out using REVIGO and KEGG Orthology Based Annotation System (KOBAS) on DGEs, respectively.Identification of IN-MDS compound targets was performed by matching potential targets of active components with disease-related targets.The results of KEGG pathway enrichment analysis were combined with compound targets to screen key targets.In the end, molecular docking was performed by SYBYL-X2.1 to verify the key targets.Results:Nine active components of IN and 439 potential targets of IN were identified by analyzing TCMSP, BATMAN-TCM, and Swiss Target Prediction network databases.Three MDS disease-related gene microarray chips were obtained from the GEO databases: GSE4619, GSE19429, and GSE58831.Through this analysis, 87 DEGs were finally obtained using the Venn diagram.A PPI network of DEGs was then constructed, in which 18 genes were upregulated and 69 genes were downregulated.After the GO enrichment results were de-redundant, the representative GO terms were obtained by using REVIGO semantic similarity measuremen.The KEGG biological pathway analysis using the KOBAS indicated that the Hippo signaling pathway is important in MDS.The Hippo signaling pathway involves four genes: AREG, LEF1, SMAD7, and TCF4.By matching and mapping DEGs with potential targets,six IN-MDS compound targets were obtained: PDE4B, PLAUR, ELANE, NR3C1, AREG, and LEF1.We found that AREG and LEF1 are consistent with the genes involved in the Hippo signaling pathway.Through molecular docking simulation, we found that the indican binds best to AREG and LEF1.Conclusion:Based on the integrated pharmacology model, the material basis of the efficacy and biological molecular mechanism of IN in the treatment of MDS was systematically studied, which provided a novel indication of the CAM regimen for the improvement of MDS management.

Keywords: Indigo Naturalis; Myelodysplastic Syndromes; Molecular Mechanism; Integrated Pharmacological Study.

1 Introduction

Myelodysplastic syndrome (MDS) is a group of refractory, highly heterogeneous malignant hematological diseases with compromised quality and abnormal quantity of blood cells originating from hematopoietic stem cells.It is characterized by abnormal blood cell development and has a high risk of transformation into acute myeloid leukemia (AML)[1].Since 2006, the Food and Drug Administration(FDA) of the United States has not approved any drug for MDS treatment, and the current treatment lose its efficacy after two to three years of initial treatment for most patients due to disease progression.Unfortunately,continuously activated kinase mutations that can be targeted in MDS are very rare, and it is not clear how best to design targeted therapy for MDS-related mutations in genes, such as ASXL1,SF3B1,JAK2,and TET2, which affect signal transduction, DNA methylation, chromatin modification, and RNA splicing[2-5].In addition, MDS clone heterogeneity and clone structure mean that it is currently impossible to know which are early events and which are late events that are only important for the survival of subcloned structures [6].Therefore, there is an urgent need to explore new biomarkers related to the occurrence of MDS and develop new treatment strategies with high selectivity, few side effects, and good curative potential.

As a complementary alternative medicine (CAM),traditional Chinese medicine has a long history and has received increasing attention from researchers in the treatment of malignant hematological diseases [7, 8].The representative is the efficacy of arsenic trioxide(ATO) (derived from the traditional Chinese medicine white arsenic) in the treatment of acute promyelocytic leukemia (APL).In-depth research on this drug had a significant impact on the medical field, which directly brought ATO to the forefront of APL therapy research[9-11].In 2016 and 2018, ATO was approved by the European Union and FDA, respectively, and has become the first-line of treatment for APL.Therefore,many Chinese patients with MDS as well as doctors and researchers are increasingly considering CAM regimens for the treatment of MDS.

Among these regimens, IN is the preferred choice.Its efficacy has been well verified in clinical practice[12, 13].Modern pharmacological studies have shown that IN has multiple biological activities, including enhancement of tumor necrosis factor-induced apoptosis through modulation of the nuclear factorkappa B signaling pathway [14].Other studies have shown that IN can enhance the efficacy of arsenic in APL, improve event-free survival, reduce adverse events and hospitalization duration and be a potential adjuvant drug for the APL treatment [15-17].IN can also selectively enhance the levels of CD4(+)CD25(+) Treg cells, making the host more conducive to inducing immune tolerance, which is important for the treatment of autoimmune diseases [18].IN also has antioxidant[19], antibacterial [20], anti-inflammatory[21], an danti-viral properties [22].Among the many pharmacological effects of IN, its anti-tumor effect has received increasing attention.Therefore, for ensuring clinical efficacy, it is necessary to further study the mechanism of IN in MDS.Since the 1990s,Chinese scholars have pioneered the application of an IN formula for the treatment of MDS and achieved good clinical effects [23].However, the mechanism of action of IN in the treatment of MDS has not been fully elucidated due to the complex components of IN.Furthermore, the traditional pharmacological concept of "one drug, one target, one disease" does not fully adapt to the model of modern biomedical research.

The rapid development of molecular biology provides a new perspective for Chinese medicine research on tumors.With the development of bioinformatics and the accumulation of high-throughput data, drugs, interacting proteins, disease phenotypes,and other data can be integrated via molecular networks.Network construction using computer models enables us to understand complex biological processes involved in tumor occurrence.Moreover, it provides an important reference for tumor intervention and treatment using traditional Chinese medicine.Under such circumstances, new and more effective therapeutic drugs or methods may be produced in the future.

In this study, we used gene chips to obtain MDSrelated DEGs, constructed their PPI network, and performed a biological function analysis.Finally, the effect of IN components on MDS targets was verified by molecular docking.This integrated pharmacology approach based on bioinformatics, molecular big data analysis, and network pharmacology has enabled us to predict the targets of IN and reveal the mechanism of IN intervention in MDS.Additionally, the present study explores the multi-component and multi-target anti-tumor model of traditional Chinese medicine.The technical procedures used in this study are presented in the form of a flowchart Figure 1.

Figure 1: Flowchart of the technical procedures in this study.

2 Materials and methods

2.1 Screening for the active components in Indigo Naturalis.

Based on the absorption, distribution, metabolism,and excretion (ADME) parameters [24], we screened for the active components in IN by searching the pharmacological online analysis database of TCMSP(https://tcmsp-e.com/) [25] and BATMAN--TCM(http://bionet.ncpsb.org.cn/batman-tcm/),respectively[26].The screening criteria were oral bioavailability(OB)≥30% and drug similarity (DL)≥0.18 [27].The corresponding SMILES structures of the active components were obtained by searching the PubChem network database (https://pubchem.ncbi.nlm.nih.gov/)[28].

2.2 Potential targets of IN.

The Swiss Target Prediction network (http://www.swisstargetprediction.ch/) [29], BATMAN-TCM and TCMSP databases were then employed to predict and analyze the active components and potential targets of IN.The search parameters of BATMAN-TCM databaseswere set to Score cutoff ≥ 20, and P value cutoff 0.05.The SMILES structures of the components obtained from PubChem ( https://pubchem.ncbi.nlm.nih.gov/ ) were imported into the Swiss target prediction network database.In addition, potential targets were imported into the UniProt database(https://www.uniprot.org/) [30].The potential target protein names were standardized as gene names, and the selection criterion for popular organisms was set as "human." A dataset of drug-related targets was obtained, and potential targets were identified after duplicate removal.

2.3 Construction of a “components-targets” network(C-T network) of IN.

The C-T network is an interaction network of the active components and their potential targets,which was carried out using the STRING online analysis database(https://string-db.org/cgi/input.pl).STRING is a database of known and predicted PPI.The interactions include direct (physical) and indirect (functional)associations that stem from computational prediction based on knowledge transfer between organisms and interactions aggregated from other (primary) databases[31].The relationships between the components and targets were visualized using Cytoscape software.Cytoscape3.6.1 (https://cytoscape.org/) is an open software application for visualizing, integrating,modeling, and analyzing interactive networks.

2.4 Construction of MDS-related targets database.

Firstly, we searched the MDS gene chips in the GEO database to obtain the microarray data of MDS patients and normal controls, which is a public functional genomics data repository for storing and retrieving data on high-throughput gene expression and genomic hybridization data [32] .We manually searched the GEO database using the key term "Myelodysplastic syndrome" with the following criteria to filter the datasets: (a) studies must use human samples, (b) the sample size must be larger than 60 cases, and (c) data must be obtained from MDS patients and healthy controls.GEO2R was used to analyze the data of patients with MDS and normal controls.GEO2R is an interactive web tool that allows users to compare two or more samples in the GEO series to identify DEGs under different experimental conditions.Herein, the grouping was defined according to the sample type.To obtain results that are more convincing and reduce the single-chip error, our chip was standardized and logarithmically converted.We screened differential gene expression for each chip according to the following criteria:P< 0.05, LogFC≥|1|.DEGs shared by the three chips were visualized in a Venn diagram for further analysis.A Venn diagram was used to display the overlapping area of the element set.The Venn diagram of DEGs was obtained using Venny 2.1 (https://bioinfogp.cnb.csic.es/tools/venny/).Because the GEO database is available to the public,no approval from the local ethics committee was required.

2.5 Construction and analysis of PPI network of DEGs.

The STRING database was used for PPI network analysis of the DEGs.This database is useful for mining and identifying important genes.The relationships between DEGs were visualized using Cytoscape software.

2.6 GO and KEGG Enrichment Analysis of DEGs.

The Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/) was used to obtain GO terms about the DEGs.It is well known that DAVID is a biological information database that integrates biological data and analysis tools to provide systematic and comprehensive biological function annotation information for large-scale gene or protein lists [33].Currently, GO function annotation is divided into three branches:biological processes (BP), cell components (CC), and molecular functions (MF).The REVIGO database(http://revigo.irb.hr/) can further obtain representative GO terms using methods that rely on semantic similarity measurement (a representative subset of the terms using a simple clustering algorithm).In this way, the GO enrichment results are de-redundant,and these remaining results can be visualized in semantic similarity-based scatterplots to help the researcher resolve these semantic relationships and hierarchical relationships [34].Further more, KEGG pathway analysis can visualize the pathways of DEGs to enable a more intuitive and comprehensive understanding of the pathways associated with these genes.The KOBAS (http://kobas.cbi.pku.edu.cn/)online analysis database was used to perform KEGG pathway enrichment analysis of DEGs [35].Version3.0 of KOBAS intelligently determines the priority of related biological pathways.In addition, KOBAS has expanded downstream exploratory visualization capabilities to select and understand rich results.

2.7 Compound targets acquisition and biological process analysis.

Venny 2.1 was used to construct a Venn diagram of the active components potential targets and disease targets to identify their overlapping area.The targets contained in the overlapping area were the compound targets.The Venn diagram more intuitively demonstrated that IN could interfere with MDS through multiple gene targets.Biological process analysis and visualization were performed using BiNGO of the Cytoscape plugin.

2.8.Active components-key targets docking.

Based on compound targets, combined with KOBAS pathway analysis to determine key targets finally.Then,the key targets were carried out to dock with the active components.Notably,molecular docking is a molecular modeling process that predicts the interaction of small molecules with macromolecular structures, such as proteins, and scores their complementary values at the binding sites.It is a dynamic research field with strong practicability in structure-based drug design and elucidation of biochemical pathways.It can also be used to predict the protein binding mode and affinity,which are the main aspects of successful docking experiments [36].We used Tripos Sybyl-x2.0 software for molecular modeling.The targets crystal structure in PDB format were obtained from the RCSB protein database (https://www.RCSB.org/).The mol2 format files for the corresponding active components were obtained from TCMSP and ZINC databae (http://zinc.docking.org/, last modified on January 4, 2019).ZINC retains all available molecules in biologically relevant ready-to-dock formats [37].

3 Result

3.1 Active components of IN and the construction of C-T network.

It is very meaningful to use network pharmacological screening technology to identify the active components of a single traditional Chinese medicine.By searching the TCMSP and BATMAN-TCM databases, we retrieved 29 active components from the IN.Of these components, nine were screened based on the OB and DL of the ADME parametersTable1.Then, following duplicate removal,439 potential targets Supplementary Table 1 were screened using the target of the Swiss Target Prediction network, BATMAN-TCM and TCMSP online databases, and the C-T network was constructed using Cytoscape software Figure 2.

Figure 2: C-T network of IN.Orangish yellow nodes represent core active compounds of IN, green nodes represent potential targets of IN.

Table 1: Active compounds and ADME parameters of IN.

3.2 Identification and PPI network construction of DEGs.

First, three-microarray gene chips were obtained from the GEO database.The volcano plot displays the expression of DEGs Figure 3A.Therein, the GSE58831 dataset (Platform GPL570) consisted of 159 MDS samples and 17 normal controls, the GSE19429 dataset (Platform GPL570) involved 183 MDS samples and 17 normal samples, and the GSE4619 dataset(Platform GPL570) included 55 MDS specimens and 11 normal specimens.Second, the intersections of DEGs were obtained by the Venn analysis Figure 3B,87 DEGsSupplementary Table 2 were obtained from the overlap, including 18 upregulated genes and 69 downregulated genes.Lastly, we used the STRING online database to construct the PPI network of DEGs,which hides the unconnected nodes in the network, and then we used the Cytoscape software to visualize the PPI network.Figure 3C.

Table2 :The KEGG pathway analysis of DEGs

Figure 3:Identification and analysis of DEGs in three datasets (GSE19429,GSE4619, and GSE58831).(a) Volcano plots of the three datasets.(b) Venn diagram.DEGs were selected based on LogF≥|1|and P <0.05, among the three sets (GSE19429,GSE4619 and GSE58831) (c)PPI network of the DEGs.Red nodes represent upregulated genes and light blue nodes represent downregulated genes.Unconnected nodes in the network were hidden.

3.3 GO and KEGG enrichment analysis of DEGs.

The GO and KEGG enrichment analysis were used to comment on the 87 DEGs of MDS.The GO term enrichment analysis (Supplementary Table 3)was carried out through the DAVID and REVIGO online databases.GO analysis terms were divided into three ontologies, namely, CC, BP, and MF.CC showed that the functions of DEGs were mainly enriched in the transcription factor complex and cytoplasm.BP involved multiple signal pathways,such as regulation of transcription from RNA polymerase II promoter,G-protein coupled receptor signaling pathway, negative regulation of cell proliferation, type I interferon signaling pathway, B cell differentiation, negative regulation of cell migration, epidermal growth factor receptor signaling pathway, steroid hormone mediated signaling pathway, humoral immune response, negative regulation of transforming growth factor beta receptor signaling pathway, BMP signaling pathway, and regulation of phosphatidylinositol 3-kinase signaling,et al.For MF, DEGs were mainly enriched in protein binding, transcriptional activator activity, RNA polymerase II core promoter proximal region sequencespecific binding, RNA polymerase II core promoter proximal region sequence-specific DNA binding,steroid hormone receptor activity, collagen binding,beta-catenin binding, transmembrane receptor protein tyrosine kinase adaptor activity, I-SMAD binding, and glucocorticoid receptor binding, et al.

Subsequently,REVIGO was used to delete redundant GO terms and summarize the results as follows:negative regulation of transcription by RNA polymerase II, cellular protein-containing complex localization,cellular response to cytokine stimulus, and B cell differentiation Figure 4A.We found that most GO terms were involved in transcription by RNA polymerase II.Therefore, we hypothesized that the pathogenesis of MDS may be related to transcription by RNA polymerase II, similar to a study by Dominik Beckm,which identified "RNA polymerase II promoters" as the most tightly controlled biological function (38).KEGG pathway analysis of DEGs Supplementary Table 4 was performed using the KOBAS online analysis database,which included 35 KEGG pathways and were visualized by bubble plot Figure 4B.On this basis, we used the corrected P-value to further filter and obtained six KEGG pathways of DEGs, including primary immunodeficiency, Hippo signaling pathway, colorectal cancer, Basal cell carcinoma, Thyroid cancer and B cell receptor signaling.The specific analysis results are showed in Table 2.Moreover, the target-pathway network (T-P network) was built and visualized using Cytoscape software Figure 4C.Increasing evidence suggests that the Hippo signaling pathway may play an important role in MDS.

Figure 4: GO and KEGG enrichment analysesof DEGs.(a) GO analysis scatter plot.Visualization of the significantly associated GO biological processes using REVIGO.The semantic similarity scatterplot shows the cluster representatives after redundancy reduction.The bubble size indicates the frequency of the GO term and the color indicates the log10 value (adjusted P-value).The color legend is shown at the left corner of the panel.(b) Bubble chart of the KOBAS enrichment analysis.Each bubble represents an enriched function, and the size of the bubble from small to large: [0.05,1], [0.01,0.05), [0.001,0.01), [0.0001,0.001), [1e-10,0.0001), [0,1e-10).The color of the bar is same as that of the circular network, which represents different clusters.For each cluster, if there are more than five terms, the top five with the highest enrichment ratio will be displayed.(c) Target-pathway (T-P) networks.The yellow nodes represent the pathway; the green nodes represent the targets.

3.4 Identification and biological process analysis of IN-MDS compound targets.

Based on the above analyses, the potential targets of IN were matched with the disease targets of MDS to obtain the compound targets of IN-MDS Figure 5A.These compound target genes were PDE4B, PLAUR,ELANE, AREG, NR3C1 and LEF1, and they were all downregulated genes.To further study the potential value of these six compound targets, we conducted a biological analysis, and the results are shown in Fig ure 5B.

Figure 5: Identification and biological process analysis of IN-MDS compound targets.(a) Venn diagram of DEGs and the potential targets.The six compound target genes are PDE4B, PLAUR, ELANE,AREG, NR3C1, and LEF1 (b) The biological process analysis of compound target genes via BiNGO.The color depth of nodes represents corrected P-value.The size of node represents the number of genes in the node.

3.5 Components-targets docking analysis.

In recent years, molecular docking has become an important technology in computer-aided drug research.By analyzing the literature and considering the importance of the Hippo signaling pathway in the analysis of the KEGG pathway, we finally determined AREG and LEF1 as key targets.Thus, they were verified by molecular docking with IN active components.In the present study, nine components were scored for docking with two key targets.The scoring results are listed in Table 3.Through the analysis of the results, we found that the top three components were indican, bisindigotin, and betasitosterol.Their scores are greater than 4, which met our screening criteria.To further explore the interaction mechanism between the active components and key targets, three active components with higher scores were simulated for docking.A molecular docking model was subsequently fabricated to visualize the common binding sites.The details of the binding sites are shown in Figure 6.By integrating the molecular docking scoring results and simulating binding sites,using the Surflex-Dock score (total score) as the standard, the indican was found to best bind to AREG and LEF1.

Table 3: Docking scoring results of the nine active components and two key targets of IN.

Figure 6: The protein–ligand of the docking simulation.The three core compounds (indican, bisindigotin and beta-sitosterol) of IN are docked with two key targets (AREG, LEF1).

beta-sitosterol 4.1766 bisindigotin 4.4740 indican 4.7920 indirubin 2.6799 isovitexin 3.3150 LEF1-1cg7 10h-indolo,[3,2-b],quinoline 3.3938 indigo 4.9861 6-(3-oxoindolin-2-ylidene)indolo[2,1-b]quinazolin-12-one 3.7701 Isoindigo 4.2563 beta-sitosterol 6.2232 bisindigotin 5.9469 indican 6.3207 indirubin 4.1057 isovitexin 5.4572

4 Discussion

MDS are myeloid neoplasms, whose main feature is ineffective hematopoiesis leading to cytopenia(anemia, infection, and bleeding) and progression to acute myeloid leukemia in one-third of patients [39].Studies have shown that there is a clear continuum between the myelodysplastic and leukemic phases [40].Although understanding of the pathophysiology of MDS is continuously improving, the overall progress in the treatment has been limited.Due to effectiveness and safety reasons, the application of allogeneic transplantation has been limited to the treatment of MDS.Therefore, more attention should be paid to developing more effective treatment strategies for many patients who are ineligible for transplantation [39].

IN is a traditional Chinese medicine with multiple biological activities, and its anti-tumor effect has been widely recognized by the medical community.Therefore, to further explore the mechanism of action of IN in the treatment of MDS, an integrated pharmacological model was systematically used to explore the active componentsand potential targets.In this study, we identified the active components of IN through TCMSP and BATMAN-TCM databases.Based on the OB and DL parameters, nine active components were identified.Through a literature review, we found that the research on IN mainly focused on indigo, indirubin, and tryptophan, and relatively little research was performed on other components such as beta-sitosterol (BS), bisindigotin, and indican.The integrated pharmacology model to study the efficacy can also be used to optimize the drug design of IN,reduce toxicity and side effects, and make IN safer and more effective in clinical applications.

BS is generally considered a non-toxic compound with anticancer properties.Studies have shown that BS interferes with a variety of cellular signaling pathways,including cell proliferation, apoptosis, invasion,angiogenesis, and metastasis [41].In leukemic cells,BS can anti-inhibit proliferation and induce apoptosis of human leukemia U937 cells by activating caspase-3 and increasing the Bax/Bcl-2 ratio [42].In addition,some studies have shown that polymorphisms in CYP1A1 (polypeptide 1) are involved in the metabolism of carcinogens and anticancer drugs, which may increase the risk of AML in adults, especially when CYP1A1 and GSTT1(glutathione S-transferase )are present simultaneously [43].However, bisindigotin,as an indigoid derivative, is not only an inducer of the CYP1A1 system but also an effective inhibitor of the CYP1A1 enzyme [44].These studies fully illustrate the characteristics of multi-component and multi-targeted Chinese medicine.

The development of high-throughput data provides potential research strategies for guiding the discovery of new active components in traditional Chinese medicine through network pharmacology.Furthermore, it also contributes to the deepening of our understanding of malignant tumors and more comprehensively understand the mechanisms of development of malignant tumors [45].In our current research, based on the above nine active components, we used the TCMSP, BATMAN-TCM, and Swiss Target Prediction network database to obtain a total of 439 IN potential targets and constructed a "component-target" network.Then, we integrated three MDS disease-related GEO microarray gene chips and finally obtained 87 DEGs,among which 18 genes were upregulated and 69 genes were downregulated.Then, GO and KEGG biological analyses were performed.The KEGG pathway analysis indicated six pathways for DEGs, including primary immunodeficiency, colorectal cancer, hippo signaling pathway, basal cell carcinoma,B cell receptor signaling pathway, and thyroid cancer.Further analysis revealed that the Hippo signaling pathway has important research significance in MDS, involving four genes:AREG, LEF1, SMAD7, and TCF4.The Hippo pathway regulates a variety of cellular processes, including cell survival, proliferation, and differentiation [46].Furthermore, it has been shown that its dysregulation is related to cancer development.In addition, it was found that the Hippo pathway is a tumor suppressor pathway;when the components of these regulatory pathways are mutated, it leads to the formation of an overgrowth phenotype [47].Increasing evidence has suggested that the Hippo pathway plays a key role in tumorigenesis by inhibiting cell proliferation, promoting apoptosis,and regulating stem/progenitor cell expansion [48].Therefore, components of the Hippo pathway may be good therapeutic targets in human cancers, such as MDS, which is a group of refractory malignant hematological diseases originating from hematopoietic stem cells.

In the process of matching and mapping DEGs and potential targets, six IN-MDS compound targets were identified: PDE4B, PLAUR, ELANE, NR3C1, AREG,LEF1.The AREG and LEF1 genes are also involved in the Hippo signaling pathway and are determined as key targets for molecular docking.Through the combined results of binding sites and molecular docking, we found that the indican binds best to AREG and LEF1.Metabolic analysis showed that indican, indigo, and indirubin belong to the same indigo pigments [49].Indican (indoxyl-β-D-glucoside) is present in several Chinese herbs, and its major metabolite is indoxyl sulfate (IS) [50].In a study analyzing the effect of oral indican on the pharmacokinetics of methotrexate(MTX), it was found that the main metabolite IS inhibited a variety of anion transporters and increased the systemic exposure of MTX and the mean residence time [51].A recent phase 1b study found that insulinlike growth factor (IGF)-MTX conjugates for highgrade MDS can produce stable diseases and prolong life [52].According to the above research results, it is clear that indican can enhance the efficacy of MTX in MDS treatment.

In a study of prognostic markers for MDS patients,LEF1 was found to be the most significant differentially expressed gene between early and advanced MDS[53].LEF1 expression was significantly related to the survival of MDS patients [54].Marked LEF1 downregulation has been shown to be associated with adverse outcomes in patients with MDS, especially those with advanced MDS [55].This is consistent with the results of our biological analysis.Therefore,targeting the LEF1 gene has important research significance in the prognosis and treatment of MDS.AREG, as another target of indican compound, was found for the first time to belong to MDS-related latent core genes and pathways.Interestingly, we found that AREG is a bifunctional growth-regulating glycoprotein of the epidermal growth factor (EGF)family and is associated with a variety of cancers, such as multiple myeloma (MM) [56] and pancreatic cancer[57].The production of AREG significantly enhances cancer malignancy.Targeting AREG not only reduces the chemoresistance of cancer cells but also avoids programmed cell death 1 ligand (PD-L1)-mediated immune suppression and restores human immunity[58].

5.Conclusion

As discussed above, based on the integrated pharmacology model of network pharmacology,bioinformatics, and computer simulation docking technology, the complex network of IN multicomponent, multi-target, and multi-pathway therapy for MDS was systematically studied.The molecular basis of efficacy and the mechanism of action of IN in the treatment of MDS were further explored,which provided new strategies for improvement in the management of MDS.