| چکیده انگلیسی مقاله |
Objective and background A main goal in regard of AML therapy is to tackle two main obstacles ; refractoriness to induction chemotherapy and relapse after remission which may be initiated from a rare population of cells called, Leukemia stem cells (LSCs);Which lead to the rapid clonal expansion of immature myeloblasts. Interestingly, these cells possess a number of stem cell properties such as quiescence and survival that are linked to therapy resistance. Therefore, targeting these cells could eradicate AML. In general, there are three main approaches in targeting LSCs :1) Targeting stem cell related molecular pathways 2) Targeting stem cell specific surface markers 3) Targeting the stem cell microenvironment. In recent studies, it is suggested that applying bioinformatic approaches may have an effective role in identifying precise biomarkers or targets in LSCs. Gene Regulatory Network (GRN) is an advisable tool in choosing specific genes, which may have a high probability in regulating the stemness core in LSCs. GRN is considered as an adequate approach in mapping gene interactions and pinpointing hot spots also known as “hub genes”, that have the most interaction with others. By applying this approach and using a defined set of criteria, we can attain a list of genes and perform further studies on them in wetlab. Methodology Research question: What genes may be considered as key regulators in the stemness and survival pathways of LSCs? A difficult problem is to pinpoint key regulators which result to the invincible presence of LSCs. During the following years many studies have been conducted in identifying precise biomarkers or targets, which can lead to the efficient targeting of these cells. In this study we used GRN; an adequate bioinformatics tool which may be used to attain a list of candidate genes. Therefore, In this study, 5 main steps were preformed: 1) Data collection: a data set of 495 Gene expression profiles (GSE76009) was obtained from GEO database, the following microarray data was generated using Affymetrix Human Genome (HG) U133 Plus 2.0 GeneChips. Data normalization and background correction was conducted using SVA package in R software (3.4.2 version). 2) Differentially expression analysis was performed between two defined groups of LSC+ and LSC- using limma package in R Software with the following cut off; FC>2, Adj.p-val< 0.05. 3) Network reconstruction was conducted using ARACNE algorithm and cytoscape software with two defined cut offs; FC @1.5 and, FC@1.3 .4) Network analysis, in this stage hub genes were attained; the selected genes(hub genes with the most interactions), had a higher DF than the excluded ones. 5) Functional analysis and biological evidences were studied for the candidate genes. Enrichment was preformed on the following genes using Enrichr database and the following genes were confirmed in literature by a defined set of criteria such as; a) the expression level of the candidate genes b) literarture-based functional assays c) self renewal and survival related pathways using signaling pathway databases such as KEGG and PANTHER (Fig1). Figure1. A general workflow of the following steps. Results In result, two networks were attained. The first network was constructed for 534 genes, which consisted a total amount of 31 hub genes (10 DF 25), the second one which was larger was constructed for 925 genes with a total amount of 50 hub genes (20 DF 74). Following the previous step, common genes were obtained between the hub genes (31 hub genes & 50 hub genes) (Fig2). As observed below, 29 genes were found common in both lists. Figure2. The following Venn diagram shows 29 common genes between the two hub gene lists. The following 29 genes were studied through literature and were enriched in enrichR database. In result, 4 genes were enriched (P-val: 3.067e-7, Adj. P-val: 0.00001779) in transcriptional misregulation signaling pathways in AML, leading us to this notion that these candidate genes might play a significant role as key regulators in the stemness core of LSCs. Conclusion The appointed cut offs which were applied in the GRN construction were chosen based upon previous studies. More networks can be constructed using other attained datsets; resulting to more hub genes. It must be also noted that, a practical implication of this data would be to test the following genes in an in vitro AML model to see whether the down regulation of these genes can effect the stemness in LSCs or not. References Hope KJ, Jin L, Dick JE. Acute myeloid leukemia originates from a hierarchy of leukemic stem cell classes that differ in self-renewal capacity. Nat Immunol. 2004;5:738 Li, H., et al., Learning the structure of gene regulatory networks from time series gene expression data. BMC genomics, 2011. 12(5): p.1 Ng, Stanley WK, et al. A 17-gene stemness score for rapid determination of risk in acute leukaemia. Nature 540.7633 (2016): 433-437. Pan, Jia-Qi et al. “lncRNA Co-Expression Network Model for the Prognostic Analysis of Acute Myeloid Leukemia.” International Journal of Molecular Medicine 39.3 (2017): 663–671. PMC. Web. 29 July 2017. |
