Overexpression of lncRNA ITGB2-AS1 Predicts Adverse Prognosis in Acute Myeloid Leukemia

Ying-li Zhou School of Medicine, Jiangsu University, Zhenjiang, Jiangsu, China Zi-jun Xu Zhenjiang Clinical Research Center of Hematology, Zhenjiang, Jiangsu, People’s Republic of China Dong-ming Yao Zhenjiang Clinical Research Center of Hematology, Zhenjiang, Jiangsu, People’s Republic of China Jing-dong Zhou Department of Hematology, A liated People’s Hospital of Jiangsu University, Zhenjiang, Jiangsu, People’s Republic of China Tingjuan Zhang Department of Hematology, A liated People’s Hospital of Jiangsu University, Zhenjiang, Jiangsu, People’s Republic of China Ji-chun Ma Zhenjiang Clinical Research Center of Hematology, Zhenjiang, Jiangsu, People’s Republic of China Jiang Lin Zhenjiang Clinical Research Center of Hematology, Zhenjiang, Jiangsu, People’s Republic of China Jun Qian (  qianjun0007@hotmail.com ) Zhenjiang First People's Hospital https://orcid.org/0000-0002-2649-8121


Introduction
Acute myeloid leukemia (AML) is a highly heterogeneous disease in cytogenetics and molecular biology, which is characterized by poor differentiation and uncontrolled proliferation of immature myeloid progenitor cells [1]. At present, genetic abnormalities such as chromosomal aberrations and gene mutations are considered as the most powerful prognostic information [2]. However, AML patients with moderate cytogenetic risk perform differently in terms of chemotherapy consolidation, which means that more new markers involved in leukemogenesis and molecular strati cation need to be further improved in order to better classify risks and ultimately nd better treatments [3].
With the application of next-generation sequencing technology, thousands of LncRNA have been discovered in many solid tumors. Recent evidence suggests that these LncRNAs play a crucial role in gene regulation, thereby affecting various aspects of cell homeostasis, including proliferation, survival, migration, or genome stability [4]. LncRNA ITGB2-AS1, up-regulated in pancreatic cancer, breast cancer, osteosarcoma, and ovarian cancer, plays an important role in promoting the proliferation, invasion, migration and metastasis of cancer cells, and is associated with poor prognosis [5][6][7][8][9][10]. ITGB2, as a gene on the ITGB2-AS1 complementary chain, has also been found to be involved in tumor adhesion, invasion, angiogenesis and speci c immune response [11]. In addition, beta 2-integrin-derived signaling has been revealed to induce survival and proliferation of neonatal AML cells by activating the Syk/STAT signaling axis [12]. However, the expression and clinical signi cance of ITGB2-AS1 and ITGB2 remain unknown in AML. Herein, this study was aimed to explore the expression pattern and clinical impact of ITGB-AS1 and ITGB2 in the context of known molecular prognosticators in AML, and found that ITGB2-AS1 can serve as a biomarker for prognosis prediction.

Karyotype and gene mutation detection
By conventional R-banding method, karyotype was analyzed at the time of initial diagnosis. Risk classi cation based on the karyotype ndings has been done as previously described [13][14][15][16][17]. Mutations in C-KIT, NPM1, DNMT3A, N/K-RAS, and U2AF1 were detected by high-resolution melting analysis, whereas FLT3-ITD and CEBPA mutations were detected by direct DNA sequencing.

Statistical and bioinformatics analyses
SPSS software version 20.0 (IBM Corporation, Armonk, NY, USA) was used to carry out the statistical analysis. Receiver operating characteristic (ROC) curve and area under the ROC were applied to assess the value of ITGB2-AS1 and ITGB2's expression. Besides, Pearson's chi-squared analysis was conducted to determine the difference of categorical variables between ITGB2-AS1 high group and ITGB2-AS1 low group. Through Kaplan-Meier method and Cox regression analysis, the effect of ITGB2-AS1 expression on prognosis was analyzed. Logistic regression analysis was used to identify the independent risk factors on complete remission (CR). In all tests, P<0.05 was de ned as statistically signi cant. R script was used for plotting gene volcano maps co-expressed with ITGB2-AS1. Details of bioinformatics and receiver operating characteristic (ROC) curve were shown in Additional Methods.

Expression pattern of ITGB2-AS1 and ITGB2 in AML
By using the GEPIA data (http://gepia.cancer-pku.cn/detail.php), we found that the expression of ITGB2-AS1 was signi cantly increased in AML patients compared with normal BM samples (Fig. 1a, P<0.001). A similar result was also found in another dataset (GSE63270; Fig. 1b, P<0.001). In order to con rm the results, we further analyzed the expression of ITGB2-AS1 in our cohort of 109 AML patients and 31controls. ITGB2-AS1 expression was consistently up-regulated in whole cohort AML, Non-M3-AML and CN-AML compared with controls ( Fig. 1c; P<0.0003, =0.001, and <0.0001, respectively). Next, we identi ed the positive correlation between ITGB2-AS1 and 37 genes from 18107 genes using cBioPortal (Fig. 3a, R>0.7, P<0.05). Among them, ITGB2 attracted our attention because of its special position on chromosome. We analyzed the expression of ITGB2 in the GSE63270 dataset ( Figure. 2b, P<0.001), the online website GEPIA (Fig. 2a, P<0.001) as well as our cohorts (Fig. 2c, P=0.0009, =0.0014, and <0.0001, respectively), and found that ITGB2 was also upregulated in AML. Furthermore, the positive correlation was con rmed between ITGB2 and IGB2-AS1 expression in our cohort (Fig. 3b).

Association between ITGB2-AS1 expression and clinical characteristics
109 AML patients of our cohort were divided into two subgroups (ITGB2-AS1 high and ITGB2-AS1 low ) according to the median level of ITGB2-AS1 transcript. The comparison of clinical/laboratory characteristics between the two subgroups was shown in Table1. No signi cant differences were observed in peripheral blood counts, BM blasts, FAB classi cation, cytogenetics, and common gene mutations (P>0.05). However, ITGB2-AS1 high patients were older than those ITGB2-AS1 low patients (P=0.023). Moreover, CR rate was signi cantly lower in ITGB2-AS1 high patients than in ITGB2-AS1 low patients (P=0.005).

Association between ITGB2 expression and clinical characteristics
The whole cohort of AML patients was also divided into two subgroups according to the median level of ITGB2 transcript (ITGB2 high and ITGB2 low ) (Table2). Consistent with the results of ITGB2-AS1, signi cant differences in age and CR rate were also revealed between the two subgroups (P=0.015 and =0.02, respectively).
3.4 Effect of ITGB2-AS1 and ITGB2 expression on chemotherapy response in AML Among our cohort, 88 patients had available follow-up data. We found that ITGB2-AS1 high patients had a lower CR rate (P=0.005, Table 1). Additionally, clinical characteristics in patients with and without CR were further compared. Signi cant differences were found in ITGB2-AS1 expression, age, WBCs, platelets, BM blast, and risk group (P<0.05, Table 3). However, there was no signi cant difference in the expression of ITGB2 in patients with and without CR (P=0.065, Table 3). Logistic regression analysis including the most predictive factors was further performed, which revealed that ITGB2-AS1 expression was an independent risk factor affecting CR in whole-cohort AML patients (P=0.027, Table4).

Association between ITGB2 expression and prognosis in AML patients
Similar results were shown in AML patient with ITGB2 over expression, though the high expression of ITGB2 was not an independent prognostic risk factor (P=0.589, Table5), it tended to indicate poor prognosis. By using the OncoLnc, we found that ITGB2 high patients in TCGA dataset had signi cantly reduced OS (Fig. 5a, P<0.010), which was con rmed in our whole patient cohort (Fig. 5b, P=0.020), but not in non-M3 AML (Fig. 5c, P=0.106) and CN-AML (Fig. 5d, P=0.094). In addition, the expression of ITGB2 in AML patients had the trend affecting on LFS (Fig. 5e, P=0.078).

Discussion
Over the past decades, the importance of non-coding RNA has received increasing attention [18].
Numerous studies have found that lncRNAs play important roles in the proliferation, differentiation and apoptosis of cells [19][20][21][22][23]. For example, the expression of PVT1 can induce apoptosis and necrosis of AML cells by downregulating the expression of c-Myc [24]. Recent studies have shown that lncRNA UCA1 and CRNDE also play an important role in the proliferation and differentiation of AML cells [25,26]. The study of Zhang et al. showed that HOTAIRM1 affects the differentiation and maturation of myeloid cells by regulating the expression level of the annexin gene, and downregulation of HOTAIRM1 expression will prevent all-trans retinoic acid (ATRA)-induced granulocyte differentiation [27].Therefore, gaining insight into the role of lncRNA in AML may provide opportunities for early diagnosis and therapeutic targeting of AML.
The roles of ITGB2-AS1 in tumorigenesis has just been explored in a few solid tumors [5][6][7][8][9][10]. Initially, ITGB2-AS1 was found overexpressed while its promoter was highly methylated in pancreatic cancer [5]. At a similar time, Liu et al identi ed that ITGB2-AS1 is upregulated and associated with poor survival in breast cancer [6]. Their further studies disclosed that ITGB2-AS1 could induce ITGB2 expression in breast cancer cells and then promote the migration and invasion. Then, upregulation of ITGB2-AS1 and prognostic relevance was discovered in osteosarcoma, ovarian cancer and pancreatic cancer [7][8][9].
As far as we know, this is the rst study on ITGB2-AS1 in leukemia. We found that and ITGB2-AS1 was signi cantly up-regulated in AML patients compared to controls. Moreover, we revealed that ITGB2-AS1 overexpression may have an adverse impact on chemotherapy response, which was con rmed by the lower CR rate. Furthermore, we also con rmed the adverse effect of ITGB2-AS1 overexpression on survival. All these results indicate that ITGB2-AS1 can add additional prognostic information by stratifying molecularly de ned patients into more homogeneous groups and help to select better treatment strategies. Without doubt, prospective studies are needed to con rm the prognostic prediction of ITGB2-AS1 overexpression before it can be clinically used in AML.
Integrin family has been shown to be involved in leukemogenesis [28,29]. Our previous studies disclosed the clinical relevance of two members of integrin family, ITGA2 and ITGBL1 [30,31]. In this study, we demonstrated for the rst time that ITGB2 is also overexpressed in AML patients. In addition, we also found the positive correlation between ITGA2-AS1 with ITGB2 in AML. The effects of ITGB2 overexpression on chemotherapy response and survival were found by univariate analysis, but not by multivariate analysis. More cases should be investigated to reveal the signi cance of ITGB2 aberration in AML. Further functional studies of ITGB2 and ITGB2-AS1 in leukemia are also needed.
In conclusion, our results indicate that ITGB2-AS1 is overexpressed in AML and is an independent poor prognostic factor in AML. Furthermore, ITGB2 expression is also upregulated in AML and is associated with ITGB2-AS1 expression.

Consent for publication
Written informed consent was obtained from all enrolled individuals before their participation.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.     AML, acute myeloid leukemia; CR, complete remission; OR, odds ratio; WBC, white blood cell.  AML, acute myeloid leukemia; HR, hazard ratio; OS, overall survival; WBC, white blood cell.