ARID1B/SUB1‐activated lncRNA HOXA‐AS2 drives the malignant behaviour of hepatoblastoma through regulation of HOXA3

Abstract It has been becoming increasingly evident that long non‐coding RNAs (lncRNAs) play important roles in various human cancers. However, the biological processes and clinical significance of most lncRNAs in hepatoblastoma (HB) remain unclear. In our previous study, genome‐wide analysis with a lncRNA microarray found that lncRNA HOXA‐AS2 was up‐regulated in HB. Stable transfected cell lines with HOXA‐AS2 knockdown or overexpression were constructed in HepG2 and Huh6 cells, respectively. Our data revealed knockdown of HOXA‐AS2 increased cell apoptosis and inhibited cell proliferation, migration and invasion in HB. Up‐regulation of HOXA‐AS2 promoted HB malignant biological behaviours. Mechanistic investigations indicated that HOXA‐AS2 was modulated by chromatin remodelling factor ARID1B and transcription co‐activator SUB1, thereby protecting HOXA3 from degradation. Therefore, HOXA‐AS2 positively regulates HOXA3, which might partly demonstrate the involvement of HOXA3 in HOXA‐AS2‐mediated HB carcinogenesis. In conclusion, HOXA‐AS2 is significantly overexpressed in HB and the ARID1B/HOXA‐AS2/HOXA3 axis plays a critical role in HB tumorigenesis and development. These results might provide a potential new target for HB diagnosis and therapy.


| INTRODUC TI ON
Hepatoblastoma (HB), which usually originates from immature liver precursor cells, is the most common liver tumour in children, accounting for more than 65% of childhood liver malignancies and about 1% of all childhood tumours. [1][2][3][4] Despite the increasing diversity of treatments, the prognosis of some patients with poor classification remains unsatisfactory. [5][6][7][8] Hence, further understanding of the development of HB is needed to improve diagnosis, prevention and treatment. However, the etiology of HB is unclear; thus, it is necessary to explore the mechanisms underlying the pathogenesis of HB.
Long non-coding RNAs (lncRNAs) are generally defined as RNA transcripts longer than 200 nt that do not code for proteins but are ubiquitously expressed in mammalian genomes and participate in a variety of biological processes, such as chromatin remodelling and transcription regulation, among others. 9,10 It is reported that some lncRNAs are closely linked to a variety of tumour-associated biological processes, such as tumour initiation and progression. [11][12][13] Given the importance of lncRNAs in cancer, in the current study, we focused on lncRNA HOXA-AS2, which was highly expressed in our previous research by lncRNA microarray analysis. HOXA-AS2 is 1048 nt long and locates on chromosome 7 situates between HOXA3 and HOXA4 in the HOXA cluster. HOXA-AS2, which mainly promotes cell proliferation, but also interacts with the enhancer of zeste homolog 2 of polycomb repressive complex to repress gene expression, is previously shown to promote the development of cancers of the liver and stomach. [14][15][16][17] However, the biological functions of HOXA-AS2 in HB have not yet been reported. In our previous report, lncRNA screening shows that HOXA-AS2 is up-regulated in HB, and in vitro experiments have shown that HOXA-AS2 is associated with multiple malignant biological behaviours of proliferation, apoptosis and invasion. 15,18 The role of HOXA-AS2 may be involved in the stability of HOXA3 or the function of other oncogene products that bind to it. 17 These findings indicate that HOXA-AS2 is an oncogene in HB tumorigenesis and may be a potential biomarker for HB diagnosis and therapy.

| Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was extracted with TRIzol reagent (Takara Bio, Inc, Shiga, Japan), and 1 μg was used to synthesize complementary DNA (cDNA) in a 20-μL system using the cDNA Reverse Transcription Kit (Takara).
The cDNA was diluted by fivefold and then subjected to qRT-PCR analysis, which was performed with SYBR Green PCR Master Mix

| Cell apoptosis detection
Cells were digested with trypsin without EDTA; then, it was col-

| Cell migration and invasion assay
The cell migration and invasion assays were performed with 24-well culture plates with 8μm micropore inserts. The cells were subjected to starvation treatment before the assays. For the migration assay,

| Western blot analysis
Cell proteins were extracted with radioimmunoprecipitation assay buffer (Beyotime Institute of Biotechnology). After blocking with 8% skim milk, the membranes were incubated with the primary antibodies at 4°C overnight. After washing three times with Tris-

| CRISPR CAS9
Guide RNAs were cloned into the LentiCRISPR v2 vector (a gift

| Immunohistochemical analysis
Immunohistochemical analysis of the HB tissue samples was performed with a standard procedure. Tissue sections were incubated with primary antibodies against ARID1B (Abcam, ab57461) and HOXA3 (Abcam, ab230879) at 4°C overnight.

| Statistical analysis
All assays were repeated three times. All results are presented as the mean ± standard deviation (SD). Groups were compared using Student's t test. A probability (P) value of < .05 was considered statistically significant. All statistical analyses were performed with Prism software (version 5.0; GraphPad Software, Inc, San Diego, CA, USA).

| HOXA-AS2 was significantly up-regulated in HB tissues and cell lines
Our previous study found that HOXA-AS2 was up-regulated in HB.
To determine whether HOXA-AS2 was involved in tumorigenesis, HOXA-AS2 expression was quantified in HB tissues. The qRT-PCR results showed that HOXA-AS2 expression was obviously upregulated in HB tissues, as compared with matched adjacent nontumour tissues ( Figure 1A). Increased HOXA-AS2 expression was also observed in two HB cell lines, as compared to normal L02 liver cells ( Figure 1B). Meanwhile, the expression of HOXA3 was estimated via an adjacent gene of antisense lncRNA HOXA-AS2, which found that the expression of HOXA3 was positively associated with HOXA-AS2 levels in HB cell lines ( Figure 1C). Nuclear and cytosolic separation indicated HOXA-AS2 was mainly located in nucleus than cytosol ( Figure 1D).

| HOXA-AS2 was an apoptosis repressor in HB
To investigate the function of HOXA-AS2, stable transfected cell lines (HepG2 and Huh6) were constructed for HOXA-AS2 knockdown and overexpression, respectively. qRT-PCR was used to detect the efficiency of these cell lines ( Figure 1D

| HOXA-AS2 knockdown inhibited tumour growth and colony formation
The CCK-8 assay was performed to study the effect of HOXA-AS2

| HOXA-AS2 up-regulated the expression of HOXA3 by forming RNA-RNA dimers with HOXA3 mRNA
HOXA-AS2 has a 94-bp reverse complementation region with the 5′ untranslated region of HOXA3, which may form RNA-RNA dimers and increase the stability of HOXA3 ( Figure 5A). The results of qRT-PCR and western blot showed that knockdown of HOXA-AS2 decreased HOXA3 expression, while overexpression of HOXA-AS2 increased HOXA3 expression ( Figure 5B,C). The results of a ribonuclease protection assay indicated that RNA-RNA dimers inhibited the degradation of HOXA3 mRNA via RNase ( Figure 5D). The mRNA stability test indicated that overexpression of HOXA-AS2 reduced the degradation of HOXA3 ( Figure 5E).

| HOXA3 was associated with tumour migration and invasion, but not proliferation
To further investigate the function of HOXA3 in HB, stable trans-

HOXA-AS2 and HOXA3
To study the mechanism of HOXA-AS2 and HOXA3 in HB, an RNA pull-down assay for HOXA-AS2 ( Figure S1B) and immunoprecipitation of HOXA3 ( Figure S1C, Figure 6A) were performed. Some gene products related to liver tumours, such as SFN, TRIM29 and ENO1, were found to bind to HOXA-AS2 and HOXA3.

| ARID1B and SUB1 co-regulated the transcription of HOXA-AS2
With ( Figure 6I,J). Consistent with these findings, immunohistochemistry results found that ARID1B was down-regulated in HB, while HOXA3 was up-regulated, as compared with adjacent normal liver tissues ( Figure 6K).

| Knocking down HOXA-AS2 inhibited tumour growth in vivo
The tumour formation of nude mice showed that the tumour volume and weight were significantly lower than that of the Immunofluorescence experiments on tumour tissues formed in nude mice showed that Ki67, a marker for tumour proliferation, and HOXA3 were significantly decreased in the HOXA-AS2 knockdown group than in the control group ( Figure 7D). Based on all our research results, a mechanistic diagram centred on HOXA-AS2 was created ( Figure 7E).

| D ISCUSS I ON
The results of the present study demonstrated that up-regulation of HOXA-AS2 had greatly contributed to cell growth, migration and invasion in HB. Also, HOXA-AS2 was highly expressed in HB tissues and HB cell lines, as compared to normal liver tissues and cell lines.
Moreover, knockdown of HOXA-AS2 increased cell apoptosis and Furthermore, HOXA-AS2 was found to bind with HOXA3 mRNA to promote HOXA3 stability. Hence, HOXA-AS2 may be considered as a potential marker for the diagnosis and treatment of HB.
HOXA-AS2 has a positive regulatory effect on HOXA3, which could partly explain the involvement of HOXA3 in HOXA-AS2-mediated However, knockdown or overexpression HOXA3 could affect cell F I G U R E 7 HOXA-AS2 regulated HB cell proliferation in vivo. A, The down-regulation of HOXA-AS2 inhibited the tumorigenesis ability of HB cells in mice, and nude mice were randomly divided into down-regulation of HOXA-AS2 group and control group (n = 7); B, the tumour volume of the nude mice in HOXA-AS2 down-regulation group and control group; C, the tumour weight of the nude mice in HOXA-AS2 down-regulation group and control group; D, immunofluorescence for Ki67 and HOXA3, and the scale bars were 200 μm; E, molecular mechanism diagram we envision. **P < .01; ***P < .001 migration and invasion, but not cell proliferation. We assumed that increased cell proliferation induced by HOXA-AS2 might be caused by the TRIM29, ENO1, SFN and LAP3 proteins or other signalling pathways, which will be the direction of our subsequent research.

| CON CLUS ION
In conclusion, this study elucidated the function of HOXA-AS2 in HB. ARID1B/SUB1 activated HOXA-AS2 expression and facilitated HOXA3 stability via interactions with TRIM29, ENO1 and SFN.
These results provide references to reveal potential molecular mechanisms of HB for the future development of new diagnostic and therapeutic targets. However, other functions of HOXA-AS2 in HB must be further elucidated.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.
Consent for publication: All the authors give their consent for publication.

E TH I C S S TATEM ENT
The studies involving human tissues and animal study were reviewed and approved by The Ethics Committee of Childrens' Hospital of Fudan University.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Availability of supporting data: All data generated in this study are included in the manuscript.