Identification of key genes related to dexamethasone-resistance in acute lymphoblastic leukemia

Drug resistance is the main cause of poor chemotherapy response in acute leukemia. Despite the extensive use of dexamethasone(DEX) in the treatment of acute lymphoblastic leukemia for many years, the mechanisms of dexamethasone – resistance has not been fully understood. We choose GSE94302 from GEO database aiming to identify key genes that contribute to the DEX resistance in acute lymphoblastic leukemia. Differentially expressed gene(DEGs) are selected by using GEO2R tools. A total of 837 DEGs were picked out, including 472 up-regulated and 365 down-regulated DEGs. All the DEGs were underwent gene ontology(GO) analysis and Kyoto Encyclopedia of Gene and Genome(KEGG) pathway analysis. In addition, the DEGs-encoded protein-protein interaction (PPI) was screened by using Cytoscape and Search Tool for the Retrieval of Interacting Genes(STRING). Total 20 genes were found as key genes related to DEX resistance with high degree of connectivity, including CDK1, PCNA, CCNB1, MYC, KPNA2, AURKA, NDC80, HSPA4, KIF11, UBE2C, PIK3CG, CD44, CD19, STAT1, DDX41, LYN, BCR, CD48, JAK1 and ITGB1. They could be used as biomarkers to identify the DEX-resistant acute lymphoblastic leukemia.


Introduction
Acute lymphoblastic leukemia(ALL) is a heterogeneous hematologic malignant disorder characterized by the proliferation of immature lymphoid cells in the bone marrow, peripheral blood, and other organs. It is the most common childhood cancer and a major cause of death of adult leukemia [1] .The treatment of ALL comprises supportive therapy, anti-leukemia chemo-therapy, immunotherapy and other target therapy. The anti-leukemia chemo-therapy regime includes the administration of a glucocorticoid (prednisone, prednisolone, or dexamethasone), vincristine, and at least one other agent (usually asparaginase and anthracycline) [2] . With the use of intensive chemotherapy regimen, the complete remission rates of adult ALL were 85 to 90%, but the long-term survival rates are only 30 to 50% [3] . Compared to the childhood ALL, the adult ALL patients have poorer prognosis, and the resistance to chemotherapeutic drugs, including dexamethasone(DEX) -resistance, is one of the most important cause. The mechanism of DEX-resistance was not fully understood and needed further investigation [4] .
Recently, gene expression profiling has been proven to be valuable in the pathogenesis, diagnosis and prognosis of many tumors including ALL [5] . Molecular analysis of the data from some public databases is a promising method to explore the disease related biomarkers. This article is aimed to identify the key genes and pathways relating to the DEX -resistant ALL patients, and it will be helpful to investigate the mechanisms of DEX resistance.

Microarray Data
GSE94302 from Gene Expression Omnibus(GEO) was chosen for this research, which is a public and freely available database (https://www.ncbi.nlm.nih.gov/geo/) [6] .
GSE94302 is based on the following platform: GPL10558 (Illumina HumanHT-

Gene Ontology and Pathway Enrichment Analysis
The Gene Ontology(GO) concept was intended to use a common vocabulary to annotate the homologous gene and protein sequences in multiple organisms, thus we could query and retrieve genes and proteins based on their shared biology [7] . GO analysis and pathways analysis were carried out by KEGG PATHWAYS(http://www.kegg.jp/) [8] and DAVID(https://david.ncifcrf.gov/) [9] with p＜0.05 as the cut-off criterion.

Analysis
The online database STRING(https://string-db.org/) was used to develop DEGsencoded proteins and protein-protein interaction network. Search Tool for the Retrieval of Interacting Genes (STRING) collects experimental and predictable information associated with the interactions of protein pairs in a given cell context via calculating the combined score of PPIs [10] . The DEGs were mapped into STRING, and the confidence score≥0.4, maximum number of interactors = 0 are set as the cut-off criterions. In addition, Cytoscape software [11] was used to screen the modules of PPI network.

Identification of DEGs and key genes
To identify the genes may be related to the dexamethasone sensitivity, DEGs between sensitive and resistant samples are screened using the GEO2R online analysis tool with p value ＜ 0.05 and |logFC|＞ 1. A total of 837 DEGs were identified, including 472 up-regulated and 365 down-regulated genes (Fig 1).
Besides, 10 key genes in each group with high degree of connectivity were picked out (Table1). X-axis: -log p value, y-axis: -log|FC(fold change)|.DEGs were marked as blue dot.

GO function and KEGG pathway enrichment analysis
In order to understanding the DEGs well, function and pathway enrichment analysis were conducted. DEGs were divided into two groups according to the criterion whether genes are up-regulated or down-regulated. All the up-regulated genes were put into DAVID software, the functions of up-regulated genes are screened (Fig 2   A). The results showed that the up-regulated genes were particularly enriched in   A represents the functional annotation of up-regulated genes, and B represents the down-regulated genes.

PPI network and module analysis
Using the STRING online database and Cytoscape, the PPI network of the key genes from two groups(Table1) were made (shown in Fig 3).  Fig 3: The protein -protein interaction network of DEGs.

A: B:
A: The protein-protein interaction network of top 10 hub gene in up-regulated DEGs.

Discussion
Chemotherapy is the main treatment for pediatric and adult ALL, the therapeutic regimen contains induction, consolidation and maintenance therapy phrases [12] . The induction therapy agents typically include glucocorticoid (prednisone or dexamethasone), vincristine and asparaginase, with or without anthracycline.
According to previous studies, remission-induction treatment can eradicate the initial leukemic cell burden and restore the normal hematopoiesis (get complete remission, CR) in 96-99% of children and 78-92% of adult ALL [13] , but there were only 30% to 50% adult ALL patients achieve clinical cure [3,14] . The drug resistance is the main cause of the poor prognosis of adult ALL. Therefore, early prediction of the resistant-drug had an essential role in guiding ALL therapy.
Dexamethasone -resistance is the common cause of the poor prognosis of adult ALL.
This study was focused on identifying the key genes related to dexamethasoneresistance in ALL, The GSE94302 from GEO database was chosen in the study. proven to be closely related to tumor development and drug resistance [15] .Cohen Y et al. found that CCNB1 was shown to be down-regulated in multiple myeloma and acute myeloid leukemia cells [16] . In this study, we found that high expression of CCNB1 lead to ALL cells sensitive to dexamethasone treatment. CDK1(cyclin dependent kinase 1) is a member of the Serine/Threonine protein kinase family. Previous studies showed that CDK1 was related to many types of tumors growth and the patients survival [17][18][19][20][21] .
Based on our analysis, CDK1 was identified as the key gene in up-regulated DEGs, and GO annotations related to nucleoplasm, Ub1 conjunction cytoskeleton and cell division.
The association of CDK1 between drug resistance hasn't been well explained. Therefore, our results could be helpful to explore its role in drug resistance deeply. CCNB1 and CDK1 have been found to regulated drug sensitiveness not only through cell cycle pathway, but p53 pathway according to our analysis. Previous study also found that PCNA was associated with cancers development [22,23] and patients prognosis [24,25] .
This study found that PCNA could serve as a biomarker indicating dexamethasone sensitivity through cell cycle pathway based on our analysis. MYC gene was first discovered in Burkitt lymphoma patients, it has been proven to play key role in the development of many tumors [26][27][28] . In our study, the level of MYC expression was binding. Drug's performance can be enhanced or reduced by protein binding [29] . Recent research has found KPNA2 is significantly associated with tumor differentiation, tumor depth, lymph node metastasis, venous invasion, recurrence and clinical response through L-type amino acid transporter 1 [30] . In addition, KPNA2 has been proven to be a potential marker of prognosis and therapeutic sensitivity in colorectal cancer patients [31] . But its mechanism needs further investigation. Our study also indicated that KPNA2 could be a biomarker for ALL prognosis. AURKA plays an important role in the development of pediatric glioblastoma (pGBM), and its inhibitor is considered as the effective therapy of pGBM [32] . AURKA, is a type of Aurora kinases. Recently, several Au0072ora kinase inhibitors were being investigated as novel anticancer therapy in breast cancer [33] . The importance of AURKA in ALL has not been studied deeply, our analysis will provide a new angle for further investigation. NDC80 is overexpressed in a variety of human cancers [34] , its GO analysis suggests it is functional in protein binding, cell cycle, nucleus and cell division. The value of NDC80 in leukemia has not been reported in previous studies. HSPA4, KIF11 and UBE2C all have relationships with many carcinomas [35][36][37] , but their roles in leukemia need to be further studied.
The down-regulated genes are dominantly related to Epstein-Barr virus infection, platelet activation, hematopoietic cell lineage and fc gamma R-mediated phagocytosis, for example, PIK3CG, LYN, CD19 and CD44. PIK3CG encodes a protein that belongs to the pi3/pi4-kinase family of proteins. The product of it is an enzyme which is thought to play a pivotal role in the regulation of cytotoxicity in NK cells. PIK3CG has been proven contributed to B cell development and maintenance, transformation, and proliferation [38] . The finding in this study indicted that PIK3CG regulated DEX sensitiveness through platelet activation, regulation of actin cytoskeleton, Epstein-Barr virus infection, and fc gamma R-mediated phagocytosis pathway. LYN has been a biomarker in many cancers [39,40] . In this study, LYN also could be a sign on the response to dexamethasone. CD19 is located on the surface of B lineage cells, except for plasma cells and on follicular dendritic cells. It has been used to diagnose cancers that arise from this type of cell -notably B-cell lymphomas [41] . CD19 has also been proven to be a useful treatment target of B cell malignance. CD19 may regulate drug resistance in ALL by hematopoietic cell lineage and Epstein-Barr virus infection pathway. CD44 participates in a wide variety of cellular functions including lymphocyte activation, recirculation and homing, hematopoiesisand tumor metastasis. CD44 expression is an indicative marker for effector-memory T-cells. The high levels of the adhesion molecule CD44 on leukemic cells were essential to generate leukemia [42] . CD44 was considered as cancer stem cell-like marker [43] . However, the value upon drug resistance of CD44 in neoplasms remains controversial [44] . Therefore, it is necessary to identify the exact role of it in cancers.
STAT1, DDX41, BCR, CD48, JAK1 and ITGB1 are related to phosphoprotein according to the GO annotation. STAT1 is a member of the STAT protein family. In mammals, the JAK/STAT pathway is the principal signaling mechanism for a wide array of cytokines and growth factors. STAT1 and JAK1 participate in the biological process via participating in type I interferons binding. STAT1 is a key regulatory gene in autoimmune diseases and metastatic melanoma [45,46] . Studies found that JAK/STAT pathway genes may play roles in lymphomagenesis, but they still need further investigation [47][48][49] . It has been shown that germline mutations in DDX41 gene in several leukemia families, and DDX41 mutation could develop neoplasia through acquisition of another somatic mutation. The recognition of DDX41 mutated cases may have implications for surveillance, assessment of prognosis, and the design of targeted therapies [50] . Breakpoint cluster region(BCR) is at the chromosome 9 breakpoint.
Although the BCR-ABL fusion protein has been extensively studied, the function of the normal BCR gene product is not clear. In our analysis, BCR may take part in DEX resistance, but its specific mechanism is unclear. CD48 is a B-lymphocyte activation marker, expressed on all peripheral blood lymphocytes (PBL) including T cells, B cells, Null cell and thymocytes [51][52][53] . CD48 is being investigated among other markers in research on disease markers, and our study will be useful for research. ITGB1(integrin beta1) is associated with adverse prognosis of prostate cancer through regulating Caveolin-1 (CAV1), CAV1 was over-expressed in prostate cancer and predicts adverse prognosis [54] . ITGB1 was indicated to regulated drug resistance via platelet activation and regulation of actin cytoskeleton pathway according to our study, the relationships between ITGB1 and leukemia need further research.
However, the finding here should be taken prudently.