Comprehensive Expression and Prognosis Analyses of ITGB1 in AML

a dangerous type of The emergence of multidrug resistance and the and of we have that the bone marrow microenvironment was closely related to the poor prognosis of AML. the underlying mechanisms are still far fully By utilizing the bioinformatics we screened out integrin as the hub gene, which is associated with the bone marrow microenvironment mediated changes of AML cells, with expression downloaded from Biotechnology Information-Gene Expression


Abstract
Background Acute myeloid leukemia (AML) is a dangerous type of leukemia. The emergence of multidrug resistance (MDR) and recurrence limits the prognosis and survival of patients. In recent studies, we have known that the bone marrow microenvironment was closely related to the poor prognosis of AML. However, the underlying mechanisms are still far from fully understood. By utilizing the bioinformatics analysis, we screened out integrin β1 (ITGB1) as the hub gene, which is associated with the bone marrow microenvironment mediated changes of AML cells, with expression pro le GSE73157 downloaded from National Center for Biotechnology Information-Gene Expression Omnibus (NCBI-GEO) database.
Methods R studio software was used to screen out candidate hub genes and further visualize the differential expression. R package "limma" was to nd out differentially expressed genes (DEGs). Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway were conducted by R package "cluster Pro ler". Furthermore, protein-protein interaction (PPI) network was also performed by online tool STRING and software Cytoscape. Last but not least, online tool PrognoScan and GEPIA was utilized for the evaluation of clinical signi cance of the selected hub gene. P and Cox p value <0.05 was considered to be statistical signi cance.

Results
ITGB1 was ltrated as the only hub gene in this pro le. We found that patients with high expression of ITGB1 had signi cantly longer overall survival (OS) than those with low expression (COX p value= 0.016730). Besides, the expression of the ITGB1 gene in AML patients is lower than that in normal people signi cantly (p value<0.01).

Conclusion
We identi ed ITGB1 as a key gene in the bone marrow microenvironment mediated poor prognosis in AML. The down-regulated expression of ITGB1 was related to AML patients' poor outcome. ITGB1 may be a potential marker for predicting and guiding AML treatment.

Background
Leukemia is a malignant clonal disease of the hematopoietic system, which accounts for 4% of all cancers [1]. Among them, acute myeloid leukemia (AML) is a common and dangerous type of leukemia. Currently, chemotherapy is still the most important option for the treatment of AML. However, the emergence of multidrug resistance (MDR) and recurrence greatly limits the effects of chemotherapy drugs. Therefore, the research on MDR of AML cells has always been a research hotspot in hematology.
In recent years, studies have found that the bone marrow microenvironment not only plays an important role in the hematopoietic process, but is closely related to the occurrence and development of tumors.
The bone marrow microenvironment contains multiple cell types, such as osteoblasts, mesenchymal stem cells, endothelial cells, neurons and so on. In normal humans, the bone marrow microenvironment affects hematopoietic stem cells mostly through the regulation of blood oxygen partial pressure and the secretion of cytokines [2]. However, in cancer patients, the occurrence of MDR and recurrence is inseparable from the role of bone marrow microenvironment. Leukemia and leukemia stem cells (LSCs) interact with the bone marrow microenvironment in a different way from hematopoietic stem cells to affect the development of leukemia [3]. LSCs can hide in the bone marrow niche to escape the killing effect of chemotherapy. As a result, the bone marrow niche has also become a key part of the recurrence and MDR of leukemia [4][5][6]. According to recent researches, we have known that MDR mediated by the bone marrow microenvironment might stem from the following several mechanisms: 1. The in uence of cytokines [7,8]; 2. Changes in adhesion of leukemia cells [9,10]; 3. The regulation of expression of MDRrelated genes [11]; 4. The adjustment of cell cycle and metabolism [12]. In addition, many factors can also be contributed to the recurrence of leukemia, such as the secretion of cytokines or the changes with certain pathways.
As the emergence and rapid development of high-throughput platforms and microarray, numerous molecular heterogeneity on AML has been acknowledged. In order to nd potential key biomarker of AML related to the bone marrow microenvironment-mediated poor prognosis, through the use of biological information technology and the analysis of high-throughput sequencing chip data, we screen out the differentially expressed genes (DEGs) of AML cells cultured alone and co-cultured with stromal cells, which were all treated with different concentrations of Arsenic trioxide (ATO). Furthermore, we pick up ITGB1 as a hub gene that can be considered as the predictor of prognosis. We systematically revel the occurrence and development of AML on the level of molecule and provide a potential guidance for targeted therapy.
Materials And Methods 2.1 Analysis and screening of differential expressed genes (DEGs) The GEO database (https://www.ncbi.nlm.nih.gov/geo/) of the National Center for Biotechnology Information was used to select and download the gene express pro le GSE73157 [13]. This pro le was generated on the platform of GPL17077 (Agilent-039494 SurePrint G3 Human GE v2 8*60K Microarray 039381). In order to reduce false positive results, the online analysis tool GEO2R in the GEO database was used to process the data pro le. The screen criteria were adjusted p value < 0.05 and the use of Benjamini-Hochberg. Then, R package "limma" was used to clarity the DEGs between NB4 cultured alone and co-cultured with stromal cells after the treatment of ATO. The screen criteria were adjusted p value < 0.05 and [log 2 FoldChange (log 2 FC)] > 1 [14]. All genes were visualized by volcanic maps and top 50 dramatically differentially expressed genes were selected to draw a heatmap by R package "ggplot2" [15].
2.2 Functional enrichment analysis and pathway analysis of DEGs R package "clusterPro ler" was used to conduct Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis in order to preform functional enrichment and pathway analysis of DEGs [16]. Biological process (BP), molecular function (MF) and cellular component (CC) are included in GO enrichment. Analysis results were illustrated as gures by R package "GOplot" [17]. A p value less than 0.05 was considered statistically signi cant. Also, the ClueGo plug-ins of Cytoscape software 3.8.1 was utilized to display the relationship between pathways [18]. A p value less than 0.05 was considered statistically signi cant.

Screening candidate genes through protein-protein interaction (PPI) network
The PPI prediction website STRING (https://string-db.org) constructed protein interaction network [19]. And Cytoscape software 3.8.0 was used to visualize the interactions of DEGs [18]. Further, we used "cytoHubba" to degree the interaction of hub-gene clustering by the method of Maxima Clique Centrality (MCC). Overlapped genes were separated by online toll Bioinformatics & Evolutionary Genomics (http://bioinformatics.psb.ugent.be/webtools/Venn/) and were displayed by Wayne diagram.

Selecting candidate genes and conducting survival analysis
In order to evaluate the prognostic value of candidate genes in AML patients, we utilized the PrognoScan database (http://dna00.bio.kyutech.ac.jp/PrognoScan/) and online tool Gene Expression Pro ling Interaction Analysis (http://gepia.cancer-pku.cn/index.html) to conduct the survival analysis [20,21]. The results are displayed with hazard ratio (HR) and Cox p value from a Log-rank test. Cox p value < 0.05 was considered statistically signi cant. Besides, according to online tool Gene Expression Pro ling Interaction Analysis (GEPIA), we compared the expression of candidate genes in AML patients and normal people. The screen criteria were adjusted p value < 0.01 and [log 2 FoldChange (log 2 FC)] > 1

Analysis and screening of differential expressed genes (DEGs)
The goal of this study is to screen out the hub gene of AML related to the bone marrow microenvironment-mediated MDR (Figure 1). Perform statistical analysis on GEO database, we found that the data set GSE73157contains 32080 DEGs in AML cultured alone groups and co-cultured groups (All the following are referred to as alone groups and co-culture groups, respectively). By using R package "limma", DEGs were classi ed as up-regulated genes (marked as red plots) and down-regulated genes (marked as green plots) (Figure 2a). Besides, a heatmap illustrated top 50 DEGs (Figure 2b).

Screening candidate genes through protein-protein interaction (PPI) network
In order to further understand the DEGs, we displayed the protein-protein interaction (PPI) network including all the DEGs in the two selected pathways mentioned above in STRING (Figure 4a, b). Then, cytoHubba plug-ins in Cytoscape was also used to screen out top 30 candidate hub genes of each pathway according nodes rank (Figure 4c, d), and all the top hub genes has been listed in  Subsequently, the correlation between hub genes and prognosis of AML patients was analyzed through PrognoScan and GEPIA online survival website. The results showed that except for ITGB1 gene, as the result of no survival signi cance, the other genes were removed. We found that patients with high expression of ITGB1 had signi cantly longer overall survival (OS) than those with low expression (COX p value= 0.016730) ( gure 5a, b). Besides, through the online website GEPIA, we performed a visual analysis of the expression of hub gene ITGB1 in the TCGA database in AML patients and normal people, showing that the expression of the ITGB1 gene in AML patients is lower than that in normal people signi cantly (p value<0.01) (Figure 5c).

Discussion
With the elucidation of molecular mechanisms and the emergence of targeted therapies, the treatment of AML has made great progress, and the overall survival rate of patients has been signi cantly improved. However, the resistance of leukemia cells and the recurrence after treatment are still two major problems that plague us. We now believe that the protective effect of leukemia stem cells (LSCs) mediated by the bone marrow microenvironment is also an important factor for the refractory and recurrence of leukemia.
Recent studies have shown that there is a complex interaction between leukemia cells and the bone marrow microenvironment. The two induce and support each other, which together lead to the occurrence and development of the disease.
In this study, we analyzed 32080 DEGs between cultured alone and co-cultured with stomal cells AML cells based on GSE73157 dataset. Then, we used GO enrichment analysis and KEGG pathway analysis to detect the dataset to explore the interaction between DEGs. GO analysis showed that these genes mainly reacted with the process of angiogenesis, collagen-containing extracellular matrix production and cytokine binding, suggesting that these DEGs may be involved in improving the ability of tumor cell to migrate and invade, promoting metastasis, recurrence and multiple drug resistance of cancer [22,23]. KEGG analysis illustrated that 120 DEGs mainly affect the AML cells through PI3K-Akt signaling pathway.
As we all known, phosphatidylinositol 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/Akt/mTOR) signaling pathway is over activated in hematological malignancies, and the development of inhibitors of PI3K/Akt/mTOR signaling pathway has become a common concern for hematology researchers in recent years [24]. Excessive activation of this pathway may promote the tumor proliferation and weaken the immune monitoring of tumor cells [25]. At the same time, we utilized the online tool STRING to display the functional interaction between proteins to discover the underlying mechanism of occurrence of development of AML. And then using the software Cytoscape 3.8.1 to analyze the protein interaction network diagram, we screened out FN1, THBS1, ITGB1, THBS2 as 4 candidate hub genes. The above 4 candidate genes not only interact closely with other genes, but may also determine the functions of other genes. After further survival analysis of the 4 selected genes, we found that only ITGB1 gene is statistically signi cant. We found that the expression of ITGB1 is negatively correlated with the prognosis of AML. Furthermore, compared with normal people, the expression of ITGB1 is signi cantly lower in AML patients. That is, ITGB1 may be a key gene for evaluating the prognosis and conducting targeted therapy of AML patients.
A huge amount of evidences show that integrin family molecules are involved in the occurrence and development of tumors, and play an important role in regulating tumor cell proliferation, invasion and metastasis and other biological behaviors [26]. Integrin β1 (ITGB1) is a relatively important part of molecules in the cell adhesion molecule family, and is the main receptor that mediates the interaction between cells and extracellular matrix. This saying is consistent with our above analysis results: the bone marrow microenvironment may affect AML cells by regulating cell adhesion molecules and extracellular matrix. At present, our understanding of ITGB1 is limited to the primary stage, and its biological functions are not yet very clear. In the direction of solid tumors, there are certain studies on the direction of ITGB1, but there are few reports on the research of ITGB1 of hematological tumors in the world. In summary, ITGB1 may be a potential marker for predicting and guiding the treatment of AML, which requires further attention and research.

Conclusion
In summary, ITGB1 was a hub gene related to the bone microenvironment mediated poor prognosis in AML. The expression of ITGB1 was negatively correlated with prognosis of AML. This study used bioinformatics methods to analyze the relevant data of AML researches to obtain potential markers for predicting and guiding AML treatment, which provides a theoretical basis for a deeper understanding of the occurrence and development of AML, and also give the follow-up animal experiments and clinical trials research directions. Authors' contributions XYZ, NJ and BAC were responsible for con rming the topic, collecting the data and analyzing article. XYZ designed and wrote the paper, edited the gures, legends and tables. All authors read and approved the nal manuscript.