LncRNA HOXD-AS1 functions as a ceRNA to promote hepatocellular carcinoma metastasis in zebrash xenograft models

Background: A few studies have shown that long noncoding RNA (lncRNA) HOXD cluster antisense RNA 1 (HOXD-AS1) plays an important role in hepatocellular carcinoma (HCC) metastasis as a competing endogenous RNA (ceRNA), but there is little in vivo evidence. This study aims to explore the zebrash HCC xenograft as an in vivo metastasis model to verify the ceRNA network of HOXD-AS1. Methods: The quantitative reverse transcription PCR (qRT-PCR) assay was used to assess the expression level of HOXD-AS1 in HCC cell lines. Knockdown of HOXD-AS1 or miR-130a-3p was performed by transfecting small interfering RNA (siRNA) or microRNA (miRNA) inhibitor, respectively. The proliferation and invasion of HCC cells in vitro were analyzed by CCK-8 and transwell assays. The growth and metastasis of HCC cells in vivo were assessed by zebrash xenograft models. Results: We veried that HOXD-AS1 was overexpressed in all tested HCC cell lines than the normal hepatic cells. Silence of HOXD-AS1 suppressed cell proliferation and invasion in Hep3B and Huh7 HCC cell lines in vitro. In zebrash xenograft models, knockdown of HOXD-AS1 also reduced the growth and metastasis of the two HCC cells. Moreover, downregulation of miR-130a-3p not only increased the HCC metastasis, but also rescued the metastasis which inhibited by silence of HOXD-AS1 in vitro and in vivo. Conclusions: Our study demonstrates the metastasis role of the HOXD-AS1/miR-130a-3p ceRNA network in HCC cells in vitro and in vivo, and these ndings suggest that zebrash xenograft model could be used for ceRNA mechanism verication in tumor metastasis.

Background Zebra sh xenograft model has been developed rapidly for human cancers in recent years [1][2][3]. So far, kinds of zebra sh cell line derived xenograft (zCDX) models have been established, such as colorectal cancer, melanoma, hepatocellular carcinoma, breast cancer, lung adenocarcinoma, and so on [1][2][3][4][5][6]. Moreover, a few of zebra sh patient derived xenograft (zPDX) models also can be generated successfully [1,2,[7][8][9]. Compared with the mouse xenograft model, the transparent zebra sh larva xenograft can be used to observe the transplanted cancer cells at the cellular level in intact animals, which provides an in vivo strategy to study the tumor growth and metastasis simultaneously, even the ne tumor cell behavior [1,10,11]. Thus, with the help of zebra sh xenograft model, metastasis of tumor cells could be assessed in more precise way in a few days, even for personalized clinical testing of metastasis inhibition.
HCC has an incidence of approximately 800,000 new cases annually, and it is the third most common cause of cancer-related death worldwide [12]. Although kinds of innovative therapeutic strategies are used for HCC treatment, survival rate of HCC patients is still poor [13,14]. It is mainly due to the high rates of recurrence and metastasis after surgical resection, and the metastatic HCC cells to other parts of the body are the prominent cause of cancer-related death [15]. To understand the underling mechanisms of HCC metastasis and develop the corresponding clinical treatment would reduce the rates of cancerrelated death, which depends on good in vivo metastasis models.
LncRNAs are a class of non-coding transcripts (>200 nucleotides) and have been thought to be "noise" previously [16]. In the past few decades, increasing evidences indicate that lncRNA play an important role in diverse physiological and pathological processes [17]. In HCCs, many lncRNAs have been reported which is crucial in tumor growth, metastasis and drug resistance [18]. Several recent studies have demonstrated the ceRNA mechanism of different lncRNAs in HCC metastasis, such as HULC/miR-372/PRKACB, LINC00974/miR-642/KRT19, AGAP2-AS1/miR-16-5p/ANXA11, and so on [19][20][21][22]. Zebra sh xenograft can offer the fast and intuitive metastasis model in cancer research for lncRNA study [23], which prompts us to explore its application in veri cation of ceRNA network in cancer metastasis.
LncRNA HOXD-AS1 (also named HAGLR) has been rst reported as a marker of neuroblastoma progression [24]. It is transcribed from the HOXD cluster on human chromosome 2q31.2 in an antisense manner, and it contains eight exons [25]. HOXD-AS1 has been proved that it is closely associated with the progression of several tumors [26]. Among these tumors, HOXD-AS1 mainly functions as a ceRNA which binding with miRNAs in different cancers, such as HCC, lung cancer, ovarian cancer, glioma, breast cancer, colorectal cancer, cholangiocarcinoma and cervical cancer [27][28][29][30][31][32][33][34][35]. So far, most of these studies lack in vivo metastasis models of HOXD-AS1, and no study offers the in vivo evidence of its ceRNA network. Considering this, we attempt to establish a feasible in vivo strategy by using zebra sh xenograft models.
In this study-, we rst veri ed the high expression of HOXD-AS1 in different HCC cell lines. Next, we assessed the proliferation and invasion by silencing the HOXD-AS1 in Hep3B and Huh7 cell lines both in the cell culture systems and zebra sh xenograft models. Then, we inspected the metastasis role of miR-130a-3p which is competitively bound with HOXD-AS1. Further, we veri ed that miR-130a-3p inhibition could rescue the HCC metastasis which was decreased by HOXD-AS1 knockdown in vitro and in vivo.

Zebra sh husbandry
Adult zebra sh were maintained in a sh auto culture system (Haishen, China) at 28°C, and the light cycle

Cell culture
The human HCC cell lines HepG2, Hep3B, Huh7, and the normal hepatic cell line LO2 were used in this study. All cell lines were obtained from CUNMAI Biotechnologies (Shanghai, China) in 2019 and all cell lines were authenticated by STR test. And all cell lines were tested for mycoplasma contamination by PCR. LO2 cells were cultured in 1640. HepG2, Hep3B, and Huh7 cells were cultured in DMEM. Both media were supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin and then cultured in a humidi ed air atmosphere containing 5% carbon dioxide at 37 ℃.

RNA extraction and qRT-PCR
Total RNA was extracted from cell lines using TRIzol reagent. Total RNA was reverse transcribed to cDNA using 1st Strand cDNA Synthesis SuperMix for qPCR kit (Takara). We performed real-time PCR analyses using SYBR Green Master Mix kit (Takara) following the instructions to detect the expression levels of HOXD-AS1 in different liver cell lines. Data were analyzed based on the ∆∆CT method and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the internal control. The sequences for gene-speci c primers were HOXD-AS1-F: 5´-ATTCGTCTGACTTGGCTCTT-3´and HOXD-AS1-R: 5´-CCTGTTTTGACCTTTTCCTG-3´. The GAPDH sequences were GAPDH-F: 5´-GGGAGCCAAAAGGGTCAT-3 and GAPDH-R: 5´-GAGTCCTTCCACGATACCAA-3´.

Cell proliferation assays
Cell proliferation assays were performed using Cell Counting Kit-8 (CCK-8, DOJINDO, Japan). HOXD-AS1 and NC siRNA were transfected into Hep3B and Huh7 cell lines, then cultured in 6-well plates for 24 hrs. A total of 2×10 3 cells were placed in each well of 96-well plates. Cell proliferation was monitored by measuring the optical density (OD) at 450 nm every 24 hrs according to the manufacturer's instructions, from 0 to 96 hrs, then we analyzed data of each group.

Cell invasion assays
Hep3B and Huh7 were transfected with HOXD-AS1 and NC siRNA. After 24 hrs, transfected cells were plated in 24-well plates with 8-mm-pore size chamber inserts. 2.0×10 4 or 6.0×10 4 cells were plated into the upper chambers, which was diluted with the serum-free culture medium. Then the transwell was put into the lower chamber, which including 800 μL medium with 10% FBS. After 24 hrs, the cells migrated to the bottom surface of the membrane, then the cells were xed with methanol, and stained with 0.1% crystal violet for 30 minutes. The images were taken under the microscope for analysis.
Zebrafish xenograft models Before injection, cancer cells were labeled with uorescent dye CM-DiI (Invitrogen, USA). The detailed protocol was as follows: cells were collected, then washed three times with HBSS. The cells were then labeled with CM-DiI at 37 ℃ for 5 min, following by 15 min at 4 ℃. Next, the unincorporated dye was removed by rinsing three times with HBSS and the labeled cells were prepared for injection. The labeled cancer cells were injected into the perivitelline space (PVS) of 48-hpf (hrs post fertilization) zebra sh larvae. Each zebra sh larvae were mounted in 1.2% low-melting gel, and about 300-400 cells were implanted into PVS by the micro-injector. After injection, the zebra sh larva xenografts were cultured at 34℃ until the end of experiments. At 1 day post injection (dpi), the successfully injected xenografts with similar tumor size were selected for the following experiments.
In vivo imaging and quantitative analysis At 4 dpi, the zebra sh larvae were mounted in 1.2% low-melting gel, and the images were taken by stereotype microscope (MVX10, Olympus, Japan) or confocal microscope using a 20X water-immersion objective lens (Fluoview 3000, Olympus, Japan). The spatial resolution of the images was 1600×1200 (MVX10) or 1024×1024 pixels (Fluoview 3000). The images taken by stereotype microscope were quantitatively analyzed by ImageJ software.

Statistical analysis
Statistical analysis was performed using unpaired Student's t-tests. A level of p<0.05 was considered to be statistically signi cant. Results were displayed as the mean ±SEM.

Results
HOXD-AS1 was highly expressed in HCC cell lines We rst examined the expression level of HOXD-AS1 in three HCC cell lines (HepG2, Hep3B and Huh7) using qRT-PCR via comparisons with the normal human liver cell line LO2. Among these cell lines, we found HOXD-AS1 was highly expressed in all tested HCC cell lines, and the values were 4.5~65.6-fold higher than that of the LO2 cell line (Fig. 1a). To evaluate the function of HOXD-AS1 in HCC progression, we tried to silence the expression of HOXD-AS1 in the three HCC cell lines by transfecting the siRNAs [36]. At 24 hrs later, we collected the transfected cells to analyze the knockdown e ciency of HOXD-AS1 by qRT-PCR. Compared with NC, the e ciency of rst siRNA (si1-HOXD-AS1) was 40.9%, 96.6%, 75.7% in HepG2, Hep3B and Huh7 cells respectively, and the second siRNA (si2-HOXD-AS1) was 3.9%, 41.0% and 45.2%, respectively (Fig. 1b-d). Based on knockdown e ciency data, we chose Hep3B and Huh7 cell lines for further studies.

Knockdown of HOXD-AS1 suppressed growth and metastasis of Hep3B cells in vitro and in vivo
We next examined the cell proliferation in Hep3B cell by CCK-8 assay and found that the viability of Hep3B was signi cantly decreased after transfected with si1-HOXD-AS1, but not si2-HOXD-AS1 (Fig. 2a). Base on it, we only examined the cell invasion by silencing HOXD-AS1 using si1-HOXD-AS1, and we found that the invasion of Hep3B cell was repressed after knockdown of HOXD-AS1 compared with the control (Fig. 2b and Additional le 1: Fig. 1). To determine the roles of HOXD-AS in HCC progression in vivo, we transplanted the Hep3B cells with si1-HOXD-AS1 transfection into zebra sh embryos. These transplanted cells were rstly transfected with si1-HOXD-AS1for 24 hrs, and then were labeled with CM-DiI. About 400 cells were inoculated into the PVS of 2-dpf Tg( i1a:EGFP) zebra sh larvae, which the vascular endothelial cells were labeled by EGFP. At 4 dpi, the yolk and trunk of the zebra sh larva samples were imaged by stereomicroscopy and confocal microscopy. We rst quanti ed the area with CM-DiI positive signals in yolk which represented the cell proliferation, and found knockdown of HOXD-AS1 decreased the growth of Hep3B cells compared with the control (Fig. 2c-e and Additional le 2: Fig. 2). Then we quanti ed the area with CM-DiI positive signals in trunk which represented the metastasis, and we also found silence of HOXD-AS1 decreased the metastasis of Hep3B cells (Fig. 2f-h and Additional le 3: Fig. 3). These results show that HOXD-AS1 plays important role in the progression of Hep3B cells.

Knockdown of HOXD-AS1 suppressed growth and metastasis of Huh7 cells in vitro and in vivo
We did the same experiments in Huh7 cells. Different from Hep3B cells, knockdown of HOXD-AS1 only slightly decreased the proliferation of Huh7 cells at 72-hr post-transfection (Fig. 3a), but the invasion of Hep3B cell was signi cantly repressed ( Fig. 3b and Additional le 4: Fig. 4). In zebra sh xenograft models, the silence of HOXD-AS1 dramatically decreased the growth and metastasis of Huh7 cells (Fig.  3c-h, Additional le 5: Fig. 5 and Additional le 6: Fig. 6). These results not only con rm the oncogenic role of HOXD-AS1 in HCC, but also indicate its critical function in HCC metastasis.

Knockdown of miR-130a-3p increased the metastasis of HCC cells in vitro and in vivo
There are a few studies which reported that HOXD-AS1 functioned as a ceRNA to facilitate HCC metastasis [27,28]. To prove this role, we chose one of its competitive binding miRNA miR-130a-3p for veri cation, because there were lot of evidence that miR-130a-3p speci cally inhibited metastasis in multiple cancer cells [28,31,[37][38][39][40]. We rstly veri ed the metastasis function of miR-130a-3p in cultured Huh7 cells. After knockdown of miR-130a-3p by its inhibitor, we found that the invasion of Huh7 cells was signi cantly increased compared with the control (Fig. 4a and Additional le 7: Fig. 7). We also transplanted the miR-130a-3p inhibitor transfected cells in zebra sh larvae, and the CM-DiI positive cells were also dramatically increased in trunk (Fig .4b and Additional le 8: Fig. 8). These results demonstrate that miR-130a-3p plays an important role in the inhibition of HCC metastasis.
Knockdown of miR-130a-3p rescued the HCC metastasis which was decreased by silence of HOXD-AS1 in vitro and in vivo To investigate whether miR-130a-3p involved in HOXD-AS1-mediated metastasis in HCC cells, we cotransfected HOXD-AS1 siRNA and miR-130a-3p inhibitor and the results of transwell assay showed that the miR-130a-3p inhibitor rescued the invasion which was decreased by HOXD-AS1 knockdown in Huh7 cell ( Fig. 5a and Additional le 9: Fig. 9). To demonstrate it in vivo, we implanted Huh7 cells transfected with si1-HOXD-AS1 and miR-13a-3p inhibitor, and we found knockdown of miR-130a-3p also rescued the metastasis which was inhibited by silence of HOXD-AS1 (Fig. 5b and Additional le 10: Fig. 10). The results indicate that HOXD-AS1 promotes HCC metastasis through competitively binding with miR-130a-3p.

Discussion
HOXD-AS1 has been revealed that it could regulate the proliferation, migration, invasion, apoptosis and cycle progression through different pathways in HCC [27,28,41,42]. Among these studies, HOXD-AS1 has been proved that it can function as a ceRNA by binding with miR-19a, miR-130a-3p and miR-326 [27,28,41]. In this present study, we veri ed that HOXD-AS1 functioned as an oncogenic lncRNA in growth and metastasis in HCC cell lines in vitro and in vivo. We also showed that miR-130a-3p, which could bind with HOXD-AS1, has an inhibition role in HCC metastasis. Moreover, the rescue experiments also indicate HOXD-AS1 might regulate the HCC metastasis via miR-130a-3p. These ndings rstly demonstrate that the zebra sh xenograft model is a feasible model for study the biological function of the ceRNA network in cancer.
Mouse metastasis models are widely used for the study of human cancers. It has been reported that HOXD-AS1 promotes the HCC cells metastasis by using mouse metastasis models [27,41]. At 28 days after tail vein injecting or at 42 days after intra-spleen injecting in the nude mouse, overexpressed HOXD-AS1 in HCCLM3 cells shows increased metastasis in liver and lung [27]. Consistent with these results, our present study shows that knockdown of HOXD-AS1 decreases the metastasis of HCC cells in zebra sh xenograft models. These data suggest that zebra sh xenografts could be a reliable metastasis model for human cancers.
Moreover, zebra sh xenograft models show more advantages in tumor biology. First, we accessed the metastasis role of HOXD-AS1 by zebra sh xenograft in only 4 days after transplantation, whereas mouse metastasis xenograft models require at least 4 weeks. Second, although we mainly focused on metastasis of HCC cells in the present study, we could analyze the HCC proliferation by the same xenograft samples. Third, by combing different transgenic zebra sh lines, we also can study the microenvironment of HCC cells, such as angiogenesis by Tg( i1a:EGFP) transgenic line. In the present study, we did not found any angiogenic differences in our experiments, but we found the growth difference between Hep3B and Huh7 cells which Huh7 cells preferred growing in a more concentrated way than Hep3B (Fig. 2c and Fig. 3c). As more and more oncologists use zebra sh as a model organism, more speci c and more precise xenograft models will be developed.
Recent studies have demonstrated that some of the lncRNAs, as ceRNAs, can regulate the expression levels of miRNA targets to impact the progression of tumors by competitive binding with miRNAs [43,44]. In our study, we veri ed knockdown of miR-130a-3p promoted HCC metastasis. In addition, knockdown of HOXD-AS1 decreased metastasis of HCC cells, and this downregulation could be rescued by miR-130a-3p inhibitor in zebra sh xenograft. These ndings show that zebra sh xenograft could be feasible in vivo model for verifying the ceRNA pathway in human cancers. In total, zebra sh xenografts could offer reliable cancer models, which will be gradually developed in more applications.

Conclusion
In conclusion, our study demonstrates that HOXD-AS1 plays an important role in HCC metastasis in zebra sh xenograft models. Additionally, silencing miR-130a-3p can rescue the HCC metastasis which is decreased by knockdown of HOXD-AS1. Our study suggests that zebra sh xenograft models are the feasible metastasis model for ceRNA research.

Availability of data and materials
The data used to support the ndings of this study are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate