Genetic Chinese Medicine Network Pharmacology Research, Quercetin Improves Drug Resistance in Acute Myeloid Leukemia by Targeting CXCL10


 Background: Acute myeloid leukemia (AML) is the most common acute leukemia in adults and is a highly heterogeneous and fatal disease. At present, the main method of treatment of AML is chemotherapy, but patients who relapse often develop resistance and are not sensitive to chemotherapy. Chinese medicine network pharmacology can provide new ideas about improving AML resistance.Methods: The gene expression data of relapsed drug-resistant AML and primary AML are from Gene Expression Omnibus (GEO) database. Based on the network pharmacology of traditional Chinese medicine, the effective components and target genes of Jiedu Huayu Decoction were analyzed. we performed Kyoto Encyclopedia of Genes and Genomes (KEGG), Gene Ontology (GO) analyses, Protein-protein interaction (PPI) network and construction on overlapping genes. We will perform prognostic analysis and gene correlation analysis of overlapping genes in GEPIA. The binding energy between the differential gene and the active ingredient of the drug was studied by molecular docking.Results: We found that quercetin, the active ingredient in Jiedu Huayu Decoction, can target CXCL10, thereby improving AML resistance.Conclusions: In this study, we found that quercetin improves drug resistance in acute myeloid leukemia by targeting CXCL10 based on the GEO database and the network pharmacology study of Chinese medicine.


Introduction
Acute myeloid leukemia (Acute myeloid leukemia, AML) is a type of malignant hematological disease characterized by rapid progression of the disease and poor prognosis [1]. According to the latest cancer statistics in 2020, there have been 11,180 deaths and 19,940 new cases of AML [2]. With the improvement of chemotherapy regimens, hematopoietic stem cell transplantation, gene targeted therapy, biological immunotherapy, traditional Chinese medicine and other treatment methods, some progress has been made in the treatment of AML. However, due to severe side effects and multi-drug resistance, about twothirds of the patients are prone to relapse after rst-line treatment [3,4]. After recurrence, chemotherapy has poor effect and poor prognosis, and there is no good treatment plan for the patients after recurrence [5,6]. Once AML patients relapse, the tumor cells are easy to develop resistance to chemotherapeutic drugs, and it is easy to develop multi-drug resistance to a variety of chemotherapeutic drugs with different structures and functions, and then it is di cult to cure [7][8][9]. Therefore, there is an urgent need to develop new drugs to improve AML resistance.
Traditional Chinese Medicine (TCM) has been used in Asia for more than 2,000 years. Due to its outstanding e cacy, abundant resources and low toxicity, it has gradually been accepted by non-Chinese medicine practitioners. In the clinical treatment of cancer, many traditional Chinese medicine formulations are used alone or as adjuvants for conventional chemotherapy [10,11]. Studies have shown that Chinese medicine combined with chemotherapy can increase the remission rate of AML patients and improve the resistance of AML patients [12,13]. The jiedu huayu recipe medicine is composed of Indigo Naturalis, Pseudobulbus Cremastrae Seu Pleiones, Paris polyphylla, Polygonum cuspidatum, Curcuma Zedoaria, Ligusticum chuanxiong Hort, Salvia Miltiorrhiza and Psoraleae Fructus. Studies have shown that this prescription has an important effect on improving drug resistance in leukemia patients [14]. In traditional Chinese medicine formulations, a variety of herbal ingredients and biologically active ingredients target multiple receptors and produce synergistic or antagonistic effects [15]. However, its drug components are numerous and drug interactions are complex, which brings certain di culties to the mechanism of its anti-tumor effect. Chinese medicine network pharmacology is a new research eld that combines pharmacology and pharmacodynamics. It is based on systems biology, multi-directional pharmacology and high-throughput analysis, analyzes the connection between drugs, targets and diseases by building a biological network, and has been widely used in Chinese medicine to treat cancer [16,17]. Analyzing the ingredients of Jiedu Huayu Recipe through network pharmacology, we discovered a hydroxy avonoid called quercetin, which is the active ingredient of Polygonum cuspidatum and its chemical name is 3,3,4,5,7 -Pentahydroxy avone [18]. Anti-in ammatory, antioxidant and anticancer activities are some of the mainly described quercetin mechanisms of action [19][20][21]. And it has been found that quercetin can improve the remission rate of AML chemotherapy and induce apoptosis of HL60 cells [22,23]. Studies have found that quercetin can improve the drug resistance of prostate cancer, breast cancer, gastric cancer and other cancers [24][25][26], but its mechanism of improving drug resistance in AML is less.
In this study, we searched for the bioactive ingredients and targets of the drug through the network pharmacology of traditional Chinese medicine and the GEO database to screen AML relapsed resistance genes, and nally found that quercetin targets CXCL10 and cooperates with HMOX1 and IRF1 improve AML resistance.

Microarray data
We extracted gene expression (GSE66525) pro ling data from the Gene Expression Omnibus (GEO) database at the National Center for Biotechnology Information. The AML-associated dataset GSE66525 submitted by Wieser R based on the GPL11532 platform was obtained from the GEO database and includes 11 AML primary samples and 11 AML recurrence normal samples.

Identi cation of differentially expressed genes (DEGs)
The limma package is a core component of Bioconductor, an R-based open-source software development project in statistical genomics [27]. The package is designed in such a way that, after initial preprocessing and normalization, the same analysis pipeline is used for data from all technologies. For the data from GEO, the R package limma was applied to perform analysis to identify the DEGs. The "R" software was applied to construct heat maps and Volcano map, and the regions in which the differential genes were mainly concentrated were highlighted.
Identi cation of Biologically active ingredients in jiedu huayu recipe All components of the eight Chinese medicinal herbs in jiedu huayu recipe (Indigo Naturalis, Pseudobulbus Cremastrae Seu Pleiones, Paris polyphylla, Polygonum cuspidatum, Curcuma Zedoaria, Ligusticum chuanxiong Hort, Salvia Miltiorrhiza and Psoraleae Fructus) were retrieved from the traditional Chinese medicine systems pharmacology (TCMSP) database (http://tcmspw.com/) [28]. In drug absorption, distribution, metabolism, and excretion (ADME) processes, oral bioavailability (OB) is one of the most signi cant pharmacokinetic parameters [29]. As a qualitative concept applied in drug design to estimate the druggability of a molecule [30], the drug-likeness (DL) index is useful for rapid screening of active substances. The active components were the ltered by combining oral bioavailability (OB) ≥30% and drug-likeness (DL) index ≥0.18 as suggested by the TCMSP database.

Prediction of Drug Targets for Biologically active ingredients in jiedu huayu recipe
The protein targets of the active substances in jiedu huayu recipe were retrieved from the TCMSP database and the traditional Chinese medicine integrated database (TCMID, http://www.megabionet.org/tcmid/).

Screening of overlapping genes between DEGs and drug targets of biologically active ingredients
The target gene of the active ingredient of the drug and the differential gene in the GEO database are analyzed by Perl software, and the overlapping part of the two differential genes is de ned as a common differential gene, and its expression level in the GEO database is stored for subsequent analysis.

Functional enrichment analyses for overlapping genes
The Kyoto Encyclopedia of Genes and Genomes (KEGG) is a database resource for understanding highlevel functions and utilities of the biological system from molecular-level information. The Gene Ontology (GO) could be used to perform enrichment analysis. We used DAVID (https://david.ncifcrf.gov/) to make KEGG pathway analysis and GO enrichment analysis.
Protein-protein interaction (PPI) network construction PPI analysis is used to search core genes and gene modules related to carcinogenesis. In this study, PPI network analysis of the overlapping genes were performed using the search tool for the Retrieval of Interacting Genes (STRING) database.
Prognostic Analysis and Gene Correlation Analysisin GEPIA Gene Expression Pro ling Interactive Analysis (GEPIA), a web-based tool to deliver fast and customizable functionalities based on TCGA and GTEx data, provides key interactive and customizable functions including differential expression analysis, pro ling plotting, correlation analysis, patient survival analysis, similar gene detection and dimensionality reduction analysis [31]. We used GEPIA to analyze the prognosis of 17 overlapping genes. P-values <0.05 were considered statistically signi cant. Then we used prognostic-related genes and 17 overlapping genes for genetic correlation analysis. The correlation of gene expression was evaluated by Spearman's correlation and statistical signi cance, and the strength of the correlation was determined using R 0.5.

Molecular docking
Molecular docking is a key tool in structural molecular biology and computer-assisted drug design [32]. We use molecular docking to predict the combination of active ingredients of drugs and differential genes. First of all, we prepared 3D structure of targets and compounds which were mined from Protein Data Bank (PDB) (https://www.rcsb.org/) and PubChem (https://pubchem.ncbi.nlm.nih.gov/) databases [33,34]. Then, the AutoDock tool was applied to perform a molecular docking [35]. Finally, PyMOL 2.3.2 software was used for visual processing to check the binding status of ligands and receptor binding sites [36].

Identi cation of differentially expressed genes in primary and relapsed AML patients
For the AML-associated dataset GSE66525 from GEO, the R package limma was applied to perform analysis to identify the DEGs. We screened a total of 458 differential genes, of which 133 genes were upregulated in patients with relapsed AML, and 326 genes were down-regulated in patients with recurrent AML. Heat map and volcano map show differential gene expression (Fig. 1A, 1B).

Network pharmacological analysis of Chinese medicine effective ingredients and targets of Jiedu Huayu Recipe
We rst screened the effective ingredients of eight Chinese herbal medicines(Indigo Naturalis, Pseudobulbus Cremastrae Seu Pleiones, Paris polyphylla, Polygonum cuspidatum, Curcuma Zedoaria, Ligusticum chuanxiong Hort, Salvia Miltiorrhiza and Psoraleae Fructus) in the Jiedu Huayu Recipe in the TCMSP database by setting the conditions of OB ≥ 30% and DL ≥ 0.18, and then searched for the target genes corresponding to each active ingredient. We found the active ingredients of six kinds of traditional Chinese medicines in the TCMSP database, but did not nd the active ingredients of Paris polyphylla and Psoraleae Fructus. When searching for the target genes of the active ingredients, we found ve of the target genes of traditional Chinese medicine, but did not nd the target of the active ingredients of Pseudobulbus Cremastrae Seu Pleiones. We organize the drugs, active ingredients, and target genes into Table 1.

Screening and analysis of overlapping genes between DEGs and bioactive ingredient drug targets
We use Perl software to analyze the overlapping genes of the target gene of the active ingredient of the drug and the differential gene in the GEO database. We screened 17 overlapping genes. Compared with primary patients with AML, the expression levels of overlapping genes in relapsed patients are sorted into Table 2. And we show the active ingredients of drugs corresponding to overlapping genes ( Fig. 2A). We found that the 17 overlapping genes screened were basically consistent with the target genes of quercetin, the active ingredient of Polygonum cuspidatum (Table 3, Fig. 2B). We speculate that quercetin has a certain therapeutic effect on patients with relapsed resistant AML.  Fig. 2C, 2D). KEGG pathway analysis found that overlapping genes are mainly enriched in AGE-RAGE signaling pathway in diabetic complications, HIF-1 signaling pathway, Fluid shear stress and atherosclerosis, In uenza A, Proteoglycans in cancer, MicroRNAs in cancer, TNF signaling pathway, Leukocyte transendothelial migration, Osteoclast differentiation and Natural killer cell mediated cytotoxicity. Finally, GO analysis was performed with overlapping genes. Regarding the biological processes (BP), the overlapping genes were signi cantly enriched in response to lipopolysaccharide, response to molecule of bacterial origin, regulation of angiogenesis, response to oxygen levels, regulation of vasculature development, regulation of response to wounding, response to nutrient, negative regulation of cell migration, negative regulation of cell motility and epithelial cell migration (Fig. 2E). The molecular function (MF) enriched by the overlapping genes included ligand-dependent nuclear receptor transcription coactivator activity, peptidase regulator activity, transcription coactivator activity, long-chain fatty acid binding, RNA polymerase II transcription factor activity, sequence-speci c transcription regulatory region DNA binding, transcription factor activity, sequence-speci c DNA binding transcription factor recruiting, fatty acid binding, transcription coactivator activity, transcription factor recruiting, peptidase activator activity, nuclear receptor activity (Fig. 2F). In the cellular component (CC), the analysis revealed that enrichment mainly occurred at membrane raft, membrane microdomain, membrane region, caveola, plasma membrane raft (Fig. 2G).

Correlation analysis and prognostic analysis of overlapping genes in GEPIA
We used the GEPIA database to analyze whether 17 overlapping genes are related to the prognosis of patients with AML recurrence (Supplementary Fig. 1). P value < 0.05 was considered statistically signi cant. We found that the CXCL10 gene is associated with the prognosis of patients with relapsed AML (Fig. 3A). We found that CXCL10 is the target of quercetin, and we speculate that quercetin can improve the prognosis of patients with relapsed drug-resistant AML. Compared with primary AML patients, CXCL10 expression is low in relapsed patients, and low CXCL10 expression is an independent predictor of poor prognosis in relapsed patients. Then we used CXCL10 to analyze the correlation with 17 overlapping genes, and found that HMOX1, IRF1 and CXCL10 have a high correlation (R > 0.5) (Fig. 3B,  3C).

Molecular docking
The structure of CXCL10 and quercetin were introduced into AutoDockTools for molecular docking. In general, the lower the binding free energy, the more stable the binding between the ligand and protein receptor. According to the results of molecular docking, we found that quercetin has a strong binding activity with CXCL10 (Fig. 4).

Discussion
Acute myeloid leukemia (AML) is the most common acute leukemia in adults and is a highly heterogeneous and fatal disease [37]. At present, the main method of treatment of AML is chemotherapy, but patients who relapse often develop resistance and are not sensitive to chemotherapy. Finding new therapeutic drugs and targets is the key to improving drug resistance. In this study, we analyzed the active ingredients and targets of the drug in Jieduhuayu Decoction based on the network pharmacology of Chinese medicine, and analyzed the differential genes between patients with relapsed drug-resistant AML and patients with primary AML through the GEO dataset to nd methods to improve AML resistance.
This study is the rst bioinformatics study to report the relationship between TCM network pharmacology and AML relapsed drug resistance differential genes. Through the differential gene screening of Chinese medicine network pharmacology and GEO database, we nally found 17 coincident genes, and found that 16 of these 17 coincident genes are the target genes of quercetin. By analyzing 17 overlapping genes, we found that CXCL10 is related to the prognosis of AML patients with relapsed resistance. Low expression of CXCL10 in patients with relapsed resistance indicates a poor prognosis. And through the correlation analysis of 17 overlapping genes, we found that HMOX1, IRF1 and CXCL10 have a good correlation. Therefore, we believe that quercetin affects the prognosis of patients with AML relapsed resistance by targeting CXCL10.
CXC chemokine ligand 10 (CXCL10), also known as interferon-γ-inducible protein 10 (IP-10), is one of the CXC chemokine superfamily, which can regulate immune response, angiogenesis, Apoptosis and other effects are related to the occurrence, development, treatment, and prognosis of various tumors [38,39]. Studies have shown that CXCL10 is related to cancer prognosis. Breast cancer patients with high CXCL10 expression have a higher survival rate than breast cancer patients with low CXCL10 expression. CXCL10 can be used as a bene cial predictor of tamoxifen in the treatment of breast cancer [40]. Research on rectal cancer shows that the high expression of CXCL10 has high treatment sensitivity, and CXCL10 is related to the e cacy and prognosis of rectal cancer [41]. Studies on tongue cancer have shown that the expression of CXCL10 has a signi cant correlation with the e cacy of radiotherapy for tongue cancer. The higher the expression of CXCL10, the better the e cacy, which can be used as an evaluation index for the prognosis and overall survival rate of radiotherapy for tongue cancer [42]. Combined with previous studies and the results of our bio-analysis, we have reason to believe that targeting CXCL10 with quercetin can improve the prognosis of AML patients. Heme oxygenase-1 (HMOX1), as a key protein in oxidative stress, is related to the degradation of heme into biliverdin, iron and carbon monoxide [43].
Studies have found that HMOX1 is one of the most important mechanisms leading to AML chemotherapy resistance, and HMOX1 can inhibit cell apoptosis by activating the JNK/c-JUN signaling pathway [44]. Studies have found that HMOX1 can regulate the expression of CXCL10 in colon cancer [45]. Interferon regulatory factor (IRF), an important type of transcriptional regulator, has a variety of functions in the regulation of gene expression during in ammation, immune response, cell proliferation, cell cycle progression, T cell differentiation and DNA damage [46]. Studies have shown that the up-regulation of IRF-1 in cancer foci is considered to be a way to improve the prognosis and reduce the resistance to immunotherapy; and it has been found that IRF-1 response genes can reduce the risk of AML [47]. Some studies have shown that IRF1 plays a role in the corneal innate immune response by regulating the expression of CXCL10 [48]. Combined with our research ndings, HMOX1, IRF1 and CXCL10 have a high correlation in AML patients. We speculate that HMOX1 and IRF1 may participate in CXCL10 to improve the prognosis of AML patients with relapsed drug resistance. At the same time, we also found that HMOX1, IRF1 and CXCL10 are all target genes of quercetin.
This study had several limitations: (1) the relatively small number of cases evaluated; (2) The pathway of the study has not been validated in vivo and in vitro experiments.

Conclusion
In summary, through a comprehensive analysis of gene expression and Chinese medicine network pharmacology, we have identi ed multiple abnormally expressed genes and pathways that may be related to AML relapse and drug resistance.