Long Noncoding RNA HOXA-AS2 Functions as an Oncogene by Binding to EZH2 and Suppressing LATS2 in AML

The expression of HOXA-AS2 in AML cell line was detected by qRT-PCR. AML cases from the public database (GEPIA) were also included in this study. Cell counting kit-8 (CCK-8) assay, ow cytometry, immunouorescence and Western blot were used to detect the role of HOXA-AS2 in AML cells. Luciferase reporter gene detection, RIP, RNA pull-down and RNA-ChIP detection were used to demonstrate the molecular biological mechanism of HOXA-AS2 in AML. b cycle of transfected NB4 and THP-1 cells. c Western blot analysis of cyclin D3, cyclin A2, P-rb, and CDK4 after si-HOXA-AS2 and si-LATS2 co-transfection in NB4 and THP-1 cells. β-actin protein was used as an internal control. d The level of ki67 protein in NB4 and THP-1 cells co-transfection with si-HOXA-AS2 and si-LATS2 was detected by immunouorescence. e The levels of CD11b and CD14 protein in NB4 and THP-1 cells co-transfection with si-HOXA-AS2 and si-LATS2 were detected by ow cytometry analysis. f Western blot analysis of CD11b and CD14 after si-HOXA-AS2 and si-LATS2 co-transfection in NB4 and THP-1 cells. β-actin protein was used as an internal control. ∗ p < 0.05, ∗∗ p < 0.01, compared with the NC siRNA group. #p < 0.05, ##p < 0.01, compared with the si-HOXA-AS2 group.


Background
The two main features of AML are uncontrolled malignant proliferation and blocked differentiation, accounting for 30% of leukemia related deaths [8,27]. With the progress of science and technology, although the treatment of leukemia has made great progress, but the cure of the disease is still unable to solve, the ve-year survival rate of patients is still only 25% [1,3,6,13,26]. Although AML patients after treatment, the condition has been greatly improved, but the prognosis of many patients is still not ideal.
The risk of relapse and drug resistance will still occur during chemotherapy, which may become the dilemma of AML treatment [4,10].
According to the results of human genome research, only a small part of human genes can encode protein, while the remaining large part of human genome genes (~ 97%) are tightly transcribed into noncoding RNA [9,22,24]. As a new noncoding RNA molecule, the abnormal expression of lncRNA in tumor tissue has attracted extensive attention of researchers [23,28]. Many studies have shown that lncRNA is likely to be a carcinogen gene [18,19] or a tumor suppressor gene [18] in the occurrence and development of tumors, and has increasingly become a biomarker of tumor diagnosis, treatment or prognosis, including AML [15,35]. Therefore, it is very important to identify more new cancer related lncRNAs and explore their biological functions and molecular mechanisms in order to explore new therapeutic methods of AML.
As a 1048 1048-bp lncRNA, HOXA-AS2 has been shown to enhance gastric cancer proliferation through epigenetic erythrocyte sedimentation rate (ESR) silencing the expression of p21, plk3 and DDIT3 [36,39]. However, little is known about the role of HOXA-AS2 in the pathogenesis and development of AML. Our results showed that HOXA-AS2 was up-regulated in AML cells. TCGA database also showed that HOXA-AS2 expression was also up-regulated in AML tissues and was associated with poor prognosis in AML patients. The results of functional test showed that the silencing of HOXA-AS2 not only inhibited the proliferation and induced differentiation of AML cells, but also reduced the tumorigenicity of nude mice. We further proved that HOXA-AS2 acts as a modular scaffold for histone modi ed complexes by binding with EZH2 to silence LATS2 expression. In conclusion, our results suggest that HOXA-AS2-EZH2-LATS2 axis may provide a new strategy for the diagnosis and treatment of AML.

Materials And Methods
Cell Lines and Culture NB4 and THP-1 cell lines were provided by Shanghai Gecko gene in 1640 medium containing 10% fetal bovine serum (FBS) (Gibco, United States) containing penicillin and streptomycin.

Rna Extraction And Qrt-pcr Assays
Trizol (Invitrogen, USA) was used to extract total RNA from tissues and cell lines, and then reverse transcripted using Primer-Script One Step RTPCR kit (TaKaRa, China). Finally, SYBR premix dimming eraser Kit (TaKaRa, China) was used for real-time RT-PCR detection. GAPDH was used as a standardized control. All measurements were repeated three times. The primer sequences are shown in Table 1. Rna Pull-down Assay The nucleoprotein was incubated with Biotin labeled HOXA-AS2 truncated probe and Invitrogen overnight. Finally, the protein was recovered and detected by Western blot.
Chromatin Immunoprecipitation (chip) ChIP analysis was carried out according to the manufacturer of EZ-ChIP kit (Upstate Biotechnology, Lake Placid, NY, USA). The nal results were determined by RT-qPCR.

Fluorescence In Situ Hybridization ( sh)
Fluorescence in situ hybridization (FISH) was used to detect the location of HOXA-AS2 in the cells. The experimental method was similar to that reported before [11]. Fluorescence imaging was analyzed by laser scanning confocal microscope.

Tumor Xenografts
The NCG male nude mice provided from Nanjing Institute of model animals grew adaptively in the SPF environment (Anhui Medical University) for one week, and the transfected cells (LV-shHOXA-AS2 and LV-shNC) were inoculated subcutaneously. Eight weeks after inoculation, the cells were killed.

Immunohistochemistry
Tumor tissue was xed in 4% paraformaldehyde and embedded in para n. After sectioning, it was incubated with CD11b antibody (Bioss, China), CD14 antibody (Bioss, China), Ki67 antibody (Bioss, China) and LATS2 antibody (CST, USA) overnight at 4 ° C. Then they were incubated with the second antibody and stained with diaminobenzidine. Photographs were taken with a light microscope.
Nitroblue Tetrazolium (nbt) Assay NB4(1 × 10 5 cells/ml) was inoculated in a 6-well plate and treated with ATPR. A 10 µl aliquot of NBT solution, composed of 10 mg/ml NBT (Sigma-Aldrich) and 2 ug/ml PMA (Sigma-Aldrich), was added to each well, and then cells were incubated for 30 min at 37 °C. Then using light microscopy to analysis.

Statistical Analysis
Data were expressed as mean ± standard deviation. One-way analysis of variance (ANOVA) and Duncan test were used for comparison between multiple groups. If the p-value was less than 0.05, the difference is considered signi cant. The experimental results represent at least three independent experiments.

HOXA-AS2 levels were increased in AML
In order to explore the role of HOXA-AS2 in AML, we rst used GEPIA database to analyze the expression of HOXA-AS2 in AML patients. The results showed that the expression of HOXA-AS2 in AML tissues (left bar) was signi cantly higher than that in normal tissues (right bar) ( Fig. 1a) (http://gepia.cancerpku.cn/detail.php). Therefore, Kaplan Meier survival analysis showed that the overall survival rate of AML patients with high expression of HOXA-AS2 (n = 53) was signi cantly lower than that of patients with low expression of HOXA-AS2 (n = 53) (Fig. 1b). qRT-PCR results showed that HOXA-AS2 was highly expressed in many leukemia cell lines (KG-1, NB4, U937, HL-60, and THP-1 cells) compared to the normal human monocytes (Fig. 1c). Based on the above results, we conducted a series of in vitro and in vitro studies to determine the function of HOXA-AS2 in AML.

Hoxa-as2 Regulates Aml Cell Proliferation
In order to further explore the biological function of HOXA-AS2 in AML cells, we rst constructed the siRNA and overexpression vector of HOXA-AS2. The e ciency of silencing and overexpression was veri ed by qRT-PCR (Fig. 1d, e). CCK8 results showed that silencing HOXA-AS2 could inhibit the proliferation of AML cells, but overexpression of HOXA-AS2 in AML cells had the opposite effect ( Fig. 2a,  b). Then ow cytometry and Western blot were used to demonstrate the above results. The results of ow cytometry showed that inhibition of HOXA-AS2 expression could induce G0 / G1 cycle arrest of AML cells, thus inhibiting the proliferation of AML cells, while overexpression of HOXA-AS2 had the opposite effect in AML cells (Fig. 2c, d). Overexpression of HOXA-AS2 in AML cells can increase the expression of G0/G1 marker protein (cyclin D3, cyclin A2, P-Rb and CDK4) (Fig. 2e). In contrast, our results also showed that silencing HOXA-AS2 could induce the decrease of G0/G1 marker protein (Fig. 2f). In order to further con rm the effect of HOXA-AS2 on the proliferation of AML cells, we used immuno uorescence to analyze the expression of Ki67 in the transfected AML cells (Fig. 2g, h). Western blot showed that silencing HOXA-AS2 inhibited Ki67 expression compared with that in control tumors (Fig. 2i). These results demonstrate that HOXA-AS2 expression levels are associated with AML proliferation.

Hoxa-as2 Regulates Aml Cell Differentiation
Our results have demonstrated that HOXA-AS2 can regulate the proliferation of AML. As we all know, AML is mainly caused by blocked differentiation. We used a series of experiments to prove the effect of HOXA-AS2 on differentiation. Many studies have shown that CD11b and CD14 are classic differentiation markers of leukemia. Flow cytometry showed that knockdown of HOXA-AS2 could induce the differentiation of AML cells (Fig. 3b). On the contrary, overexpression of HOXA-AS2 decreased the proportion of differentiated cells in AML cells compared with control cells (Fig. 3a). Subsequently, Western blotting showed that silencing HOXA-AS2 promoted the expression of CD11b and CD14 in AML cells (Fig. 3d), whereas overexpression of HOXA-AS2 resulted in decreased expression of CD11b and CD14 (Fig. 3c). The reduction of NBT showed that HOXA-AS2 knockdown signi cantly increased the NBT reduction in AML cells compared with NC group (Fig. 3e). In addition, shRNA knockdown of HOXA-AS2 could promote the expression of CD11b and CD14 (Fig. 3f). In conclusion, these data suggest that HOXA-AS2 can not only regulate the proliferation of AML, but also regulate the differentiation of AML.

Hoxa-as2 Regulated Lats2 Transcription In Aml Cells
In order to study the role of LATS2 in AML, we next explore the relationship between HOXA-AS2 and LATS2. As shown in Fig. 4a, silencing HOXA-AS2 signi cantly increased LATS2 protein levels compared with negative controls. On the contrary, overexpression of HOXA-AS2 signi cantly down regulated the expression of LATS2 (Fig. 4b). The results of gene level showed that inhibition of HOXA-AS2 could promote the expression of LATS2 mRNA (Fig. 4c), and overexpression of HOXA-AS2 could inhibit the expression of LATS2 mRNA in AML cells (Fig. 4d). To further elucidate the speci c mechanism of HOXA-AS2-regulating LATS2 expression, we used a human LATS2 luciferase reporter gene plasmid for the luciferase reporter gene analysis. Knocking down of HOXA-AS2 can increase LATS2 promoter activity in AML cells (Fig. 4e). Overexpression of HOXA-AS2 can also decrease LATS2 promoter activity in AML cells (Fig. 4f). These results suggest that the transcription of LATS2 in AML cells can be regulated by HOXA-AS2.

HOXA-AS2 regulated transcriptional expression of LATS2 through recruitment of EZH2 in AML cells
The expression of HOXA-AS2 was further con rmed by FISH. The nucleus was labeled with DAPI and HOXA-AS2 was labeled with Cy3 and 18S rRNA (cytoplasmic positive). The results showed that HOXA-AS2 was located in the nuclei of AML cells (Fig. 5a). The next CHIP analysis results show that overexpression of HOXA-AS2 can improve its binding ability to EZH2 (Fig. 5b). Many literatures have shown that EZH2 may be a key regulator of AML, so we used EZH2 siRNA to silence EZH2 expression. RT-qPCR and western blot results showed that the expression of EZH2 and LATS2 was negatively correlated, and silencing EZH2 could promote the expression of LATS2 ( Fig. 5c and d). Subsequently, RNA immunoprecipitation (RIP) analysis showed that EZH2 protein could bind to HOXA-AS2 (Fig. 5e). The direct interaction between EZH2 and HOXA-AS2 was further demonstrated by applying the total protein ( Fig. 5f) to the HOXA-AS2 pull-down assay. Taken together, we concluded that HOXA-AS2 regulated transcriptional expression of LATS2 through recruitment of EZH2 in AML cells.
Moreover, the western blot results indicate that co-transfection increase the downregulated expression of cyclin D3, cyclin A2, P-Rb and CDK4 induced by the knockdown of HOXA-AS2 (Fig. 6d). Flow cytometry analysis showed that the decreased LATS2 expression reversed the differentiation induced by HOXA-AS2 silencing (Fig. 6e). Moreover, the western blot results indicate that co-transfection decrease the upregulated expression of CD11b and CD14 induced by the knockdown of HOXA-AS2 (Fig. 6f). These results indicated that the effect of HOXA-AS2 on AML partially involves targeting LATS2.

Knockdown Of Hoxa-as2 Inhibits Nsclc Tumorigenesis In Vivo
We have studied the in vitro study of HOXA-AS2 in AML cells, and then the biological function of HOXA-AS2 in vivo. To further study the carcinogenic effect of HOXA-AS2 in AML, the control cells and shRNA HOXA-AS2 NB4 cells were injected into nude mice. Tumor volume was measured every week after injection, and the tumor tissue was removed 4 weeks later. The results showed that compared with the control group, silencing HOXA-AS2 signi cantly inhibited tumor growth ( Fig. 7a and b). The tumor weight of HOXA-AS2 knockout group was signi cantly lower than that of control group (Fig. 7c). Moreover, RT-PCR analysis showed that the expression level of HOXA-AS2 in tumor was lower than that after shRNA transfection (Fig. 7d). In addition, immunohistochemistry and Western blot showed that HOXA-AS2 upregulated LATS2 protein ( Fig. 7e and f). In conclusion, these data suggest that HOXA-AS2 may affect the proliferation and differentiation of AML tumors in vivo, as well as in vitro experiments.

Discussion
With the continuous improvement and progress of experimental technology, more and more new lncRNAs have been discovered by RNA sequencing technology, which play a main role in the occurrence and development of diseases and malignant progress [2,25,37]. Many studies have shown that many lncRNAs express differently in AML and can regulate their functions [11,12,21,30]. However, the function and molecular mechanism of many lncRNAs associated with AML are still unclear.
HOXA-AS2, as a new 1048 bp lncRNA, was rst reported to act as an inhibitor of apoptosis in NB4 promyelocytic leukemia cells treated with all trans retinoic acid [39]. It has been reported many times in many types of malignant tumors. Studies have shown that the high expression of HOXA-AS2 may be related to various biological processes of malignant tumors, such as apoptosis, invasion, migration, proliferation and so on [7,32,34,38]. Our results indicate that HOXA-AS2 is highly expressed in AML tissues and cell lines. The predicted results of the database website (GEPIA) showed that the high expression of HOXA-AS2 was related to the survival of AML patients, and the high expression of HOXA-AS2 had a shorter survival time. In addition, silencing HOXA-AS2 in AML cells inhibited cell proliferation and induced differentiation. Overexpression of HOXA-AS2 has the opposite biological function. In vivo experiments also show that knockout of HOXA-AS2 can inhibit tumor growth.
As a highly conserved histone methyltransferase, EZH2 is a subunit of PRC2, which plays a regulatory role mainly by triggering H3K27me3 trimethylation. It can inhibit the translation of many target genes and regulate cell cycle regulation, aging, cell proliferation, differentiation, apoptosis and tumorigenesis [20,29]. There are many reports that lncRNA can bind to EZH2, and then regulate the expression of downstream genes to exert its biological functions [5,33]. Our recent results show that ATPR can upregulate the expression of lncRNA NR-104098 and inhibit the differentiation and proliferation of AML induced by ATPR [11]. More in-depth mechanism studies showed that lncRNA NR-104098 played a role by enhancing the binding of E2F1 to EZH2 promoter [11]. Our results indicate that HOXA-AS2 can directly bind to EZH2 and inhibit the proliferation and induce differentiation of AML by regulating the expression of LATS2.
As a con rmed tumor suppressor [31], LATS2 can play an anti-tumor role through different signaling pathways [14,16,17]. Our ChIP results con rm that EZH2 can directly bind to LATS2 promoter region in AML to regulate its expression. In addition, further rescue experiments showed that HOXA-AS2 played its biological function through LATS2. Silence of LATS2 can reverse this behavior.

Conclusions
In conclusion, the up-regulation of HOXA-AS2 may be associated with the negative prognostic factors of AML patients. This study rst reported the high expression of HOXA-AS2 in AML tissues and cells.
Silencing HOXA-AS2 can inhibit the tumorigenesis of nude mice, inhibit its proliferation and induce its differentiation in AML cells. Moreover, HOXA-AS2 inhibited the expression of LATS2 by binding with EZH2. Our results suggest that HOXA-AS2 may be a new target for AML therapy, and provide new hope for the diagnosis and treatment of AML targeting lncRNA.     Upregulation of HOXA-AS2 inhibited LATS2 transcription in AML cells. a, b The LATS2 protein level was determined by western blot in NB4 and THP-1 cells transfected with HOXA-AS2 plasmid or transfected with si-HOXA-AS2#1, si-HOXA-AS2#2. c, d The LATS2 mRNA level was determined by qRT-PCR in NB4

Figure 5
HOXA-AS2 downregulated transcriptional expression of LATS2 through recruitment of EZH2 in AML cells.
a HOXA-AS2 distribution in NB4 and THP-1 cells was detected by FISH. b HOXA-AS2 plasmid was transfected into NB4 and THP-1 cells, and chrome immunoprecipitations were performed by using speci c anti-EZH2 antibodies. c NB4 and THP-1 cells transfected with EZH2 siRNAs or the control siRNA for 72 h were collected, and EZH2 and LATS2 protein levels were detected by Western-blot assay. d EZH2 siRNAs (EZH2-1, 2) or the control siRNA were transfected into NB4 and THP-1 cells for 48 h, EZH2 and LATS2 mRNA levels were then assessed by RT-qPCR. e RNA immunoprecipitations were performed in NB4 and THP-1 cells, and the relative quantities of HOXA-AS2 were detected by RT-qPCR assay, normalized to the input groups. IgG and EZH2 represented for the groups coprecipitation with IgG protein and anti-EZH2 antibody respectively. f Total proteins were extracted from NB4 and THP-1 cells, and then HOXA-AS2 pull-down assay was performed. The EZH2 protein levels were evaluated by Western-blot. HOXA-AS2 probe represented the biotin-labeled HOXA-AS2 probe group and control stood for the oligo probe group.

Figure 6
Silencing LATS2 potentially involves the oncogenic function of HOXA-AS2. a CCK8 assay was performed to determine the cell viability for si-HOXA-AS2 and si-LATS2 co-transfected NB4 and THP-1 cells. b The cell cycle of transfected NB4 and THP-1 cells. c Western blot analysis of cyclin D3, cyclin A2, P-rb, and CDK4 after si-HOXA-AS2 and si-LATS2 co-transfection in NB4 and THP-1 cells. β-actin protein was used as an internal control. d The level of ki67 protein in NB4 and THP-1 cells co-transfection with si-HOXA-AS2 and si-LATS2 was detected by immuno uorescence. e The levels of CD11b and CD14 protein in NB4 and THP-1 cells co-transfection with si-HOXA-AS2 and si-LATS2 were detected by ow cytometry analysis. f Western blot analysis of CD11b and CD14 after si-HOXA-AS2 and si-LATS2 co-transfection in NB4 and THP-1 cells. β-actin protein was used as an internal control. * p < 0.05, * * p < 0.01, compared with the NC siRNA group. #p < 0.05, ##p < 0.01, compared with the si-HOXA-AS2 group.

Figure 7
Knockdown of HOXA-AS2 inhibits AML tumorigenesis in vivo. a The total number of tumors after removal from the mice. b The tumor volumes were calculated every week after inoculation. c The tumor weights after the tumors were harvested. d qRT-PCR analyses indicated that the HOXA-AS2 expression was signi cantly increased in vivo. e Then, protein was extracted from the tumor tissues, and expressions of LATS2 protein was detected by Western blot. f Immunohistochemistry analysis of LATS2 was obtained from tumors. β-actin served as a loading control. * p < 0.05, * * p < 0.01, compared with the NC shRNA group.