LncRNA FAM83H-AS1 Promotes Aerobic Glycolysis and Tumor Progression by Regulating miR-4684-5p/ZBTB38 Axis in Esophageal Squamous Cell Carcinoma

Background: Dysregulation of lncRNAs is implicated in esophageal squamous cell carcinoma (ESCC) progression; However, the precise function of lncRNA FAM83H-AS1 in ESCC remains unknown. Methods: FAM83H-AS1, miR-4684-5p and ZBTB38 mRNA expressions were detected via qRT-PCR. ZBTB38, GLUT1 and LDH-A protein expressions were tested via Western blot. Cell proliferation, migration and invasion were evaluated via CCK-8 and transwell assay, respectively. A nude mouse xenograft model was used to investigate the role of FAM83H-AS1 in xenograft ESCC growth. The metabolic shift in ESCC cells was examined via glycolysis analysis. The interaction between FAM83H-AS1, miR-4684-5p and ZBTB38 was analyzed via computational algorithms, RNA pull-down, RIP and dual luciferase reporter assay. Results: We found that FAM83H-AS1 was upmodulated in ESCC cell lines. FAM83H-AS1 knockdown hampered ESCC cell proliferation, migration, invasion and aerobic glycolysis, while FAM83H-AS1 overexpression demonstrated the opposite effects. FAM83H-AS1 knockdown also delayed the tumor growth in vivo. Moreover, FAM83H-AS1 interacted with miR-4684-5p/ZBTB38 axis in ESCC cells. ZBTB38 overexpression or miR-4684-5p inhibition partially reversed the inhibitory effect of FAM83H-AS1 knockdown on cell migration, invasion and aerobic glycolysis in ESCC cells. Conclusion: Our present results indicate FAM83H-AS1 accelerated aerobic glycolysis and tumorigenesis of ESCC by sponging miR-4684-5p and triggering the expression of ZBTB38, providing new insights into mechanism of ESCC progression and therapeutic strategy.

negative breast cancer (TNBC) tissues and cells, and overexpression of FAM83H-AS1 promoted TNBC cell proliferation, migration and invasion [11]. However, up to data, the function and related mechanism of lncRNA FAM83H-AS1 has not been systematically studied in ESCC. Therefore, we here provided strong evidences to support the promoting role of FAM83H-AS1 in aerobic glycolysis and tumor progression of ESCC.
LncRNAs can function as miRNA sponges for the miRNA-dependent gene silencing in ESCC [5].
However, the influence of miR-4684-5p on tumors has not been reported so far. It is noteworthy that as a transcription factor, ZBTB38 is involved in human neuroblastoma cell regulation, proliferation and apoptosis, whereas knockdown of ZBTB38 induces neuroblastoma cell death potentially [12]. ZBTB38 may promote cell migration and invasion of bladder cancer cells via up-modulation of Wnt/β-catenin pathway [13]. The above ndings suggest that ZBTB38 may function as an oncogene in some tumors.
Nevertheless, it is not clear whether FAM83H-AS1 can play a role in ESCC by modulating miR-4684-5p and/or ZBTB38.
The key role of the aerobic glycolysis in tumors was elucidated in the past decade [14,15]. Enhanced glycolysis is one of the characteristics of ESCC and other tumors as the tumor cells rely on the glycolytic pathway for their energy needs [16,17]. This characteristic of cell metabolism has been successfully exploited for cancer diagnosis and therapy [18]. Many proteins, such as lactate dehydrogenase-A (LDH-A) and glucose transporter type 1 (GLUT1), control the irreversible steps in glycolysis catalyze reactions, and are considered as key enzymes of aerobic glycolysis [19,20]. However, how lncRNAs modulate glycolysis by interacting with the key enzymes of aerobic glycolysis is still poorly understood.
In the present study, the effects of FAM83H-AS1 on glycolysis in ESCC cell lines and role of FAM83H-AS1/miR-4684-5p/ZBTB38 axis was systematically investigated in ESCC. Our study suggests a novel role and mechanism that FAM83H-AS1/miR-4684-5p/ZBTB38 axis contributed to glucose metabolism and tumor progression of ESCC.

Cell lines and cell culture
The human ESCC cell lines EC9706, EC109, KYSE30 and KYSE150, and a human esophageal squamous epithelial cell line Het-1A were obtained from the Chinese Academy of Sciences Committee on Type Culture Collection Cell Bank (Shanghai, China). The cells were cultured in Roswell Park Memorial Institute 1640 medium (RPMI-1640; Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich, St. Louis, MO, USA) at 37℃ in a 5% CO 2 atmosphere. RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR) Total cellular RNA was extracted using TRIzol reagent (Invitrogen, Waltham, CA, USA). For miRNA expression analysis, qRT-PCR was carried out using the TaqMan MicroRNA Reverse Transcription kit and TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA). The relative quanti cation of miR-4684-5p was performed using the 2 -△△Ct method, with U6 used as an internal control. For FAM83H-AS1 and ZBTB38 expression analysis, qRT-PCR was performed using the TaqMan High-Capacity cDNA Reverse Transcription Kit and TaqMan Fast PCR Master Mix (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. The relative quanti cation of FAM83H-AS1 and ZBTB38 was performed using the 2 -△△Ct method, with GAPDH used as an internal control. The reactions were performed independently in triplicate, and the primer sequences are listed in Table 1. The RNA expression levels were analyzed as described previously [13,21].

Name
Primer sequence  Cell migration and invasion assay ESCC cell migration was evaluated using a transwell migration assay. 5 × 10 4 ESCC cells/well were resuspended in 250 μl RPMI 1640 supplemented with 1% FBS and plated onto uncoated 8-µm transwell lter inserts in 24-well plates in triplicate. The lower chambers contained 500 μl of RPMI 1640 supplemented with 15% FBS as a chemoattractant. The non-migratory cells in the upper chamber were removed with a cotton swab after incubation for 16 h, while cells on the bottom side were xed in 100% methanol and stained with 0.5 µg/ml 4′,6-diamidino-2-phenylindole (DAPI) for 5 min. Then stained cells were counted using a uorescence microscope (Eclipse 80i; Nikon Corporation, Tokyo, Japan) in ve random elds.
A transwell invasion assay was performed to assess cell invasion ability. ESCC cells were seeded into the upper chamber of Matrigel-coated inserts containing 250 μl RPMI 1640 containing 1% FBS. RPMI 1640 containing 15% FBS was added to the lower chamber as a chemoattractant. After incubation for 48 h, cells in the upper chamber were removed, and cells on the bottom side of chamber were xed in 70% ethanol and stained with 0.1% crystal violet for 30 min. The stained cells were lysed with 200 μl of lysis reagent. Finally, 100 μl of lysate was taken to a 96-well plate, and the absorbance was measured at 560 nm by a microplate reader (550; Bio-Rad, USA).
Glucose uptake and lactate production assay ESCC cells were cultured in glucose-free RPMI 1640 for 16 h, and then incubated with high-glucose RPMI 1640 under normoxic conditions for an additional 24 h. Culture supernatants were then collected for intracellular glucose levels measure using a uorescence-based glucose assay kit (BioVision, Milpitas, California, USA) according to the manufacturer's instructions. Lactate levels were measured using a lactate oxidase-based colorimetric assay read at 540 nm according to the manufacturer's instructions (Beyotime, Wuxi, China) and normalized to cell number.

Bioinformatics analysis
The potential target miRNAs of FAM83H-AS1 were predicted via computational algorithms, including TargetScan (http://www.targetscan.org/vert_72/) and miRDB (http://mirdb.org/). The highest-ranked predicted potential target of FAM83H-AS1 was miR-4684-5p. To identify potential genes targeted by miR-4684-5p, we also used the TargetScan and miRDB. From the list of target genes obtained, all genes likely to contribute to ESCC progression were extracted.

RIP assay
RIP was performed using a Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, MA, USA) following the manufacturer's instructions. In brief, 1 × 10 7 ESCC cells transfected with miR-4684-5p mimic and control mimic were lysed in the RIP lysis buffer containing a protease inhibitor cocktail. Next, the ESCC cell supernatant was incubated with the RIP buffer containing a magnetic bead conjugated with antibodies against human AGO2 or the control normal mouse IgG. Then the protein and DNA in the RIP complex were removed using RNase-free DNase I and Proteinase K. Then, the immunoprecipitated RNA complex was isolated and subjected to qRT-PCR to detect the enrichment of FAM83H-AS1 or ZBTB38.

In vivo xenograft experiments
The animal experiments were approved by the Animal Care and Use Committee of Taizhou University School of Medicine and performed in accordance with the relevant guidelines and regulations of the committee. BALB/c athymic nude mice (female, 4 weeks old) were purchased from Shanghai Laboratory Animal Center (Shanghai, China) and randomly divided into three groups (Blank control, si-FAM83H-AS1 and si-NC; n = 5 each). A total of 1 × 10 7 EC9706 cells (trancfected with si-FAM83H-AS1 or si-NC, untreated cells as blank control, respectively) were subcutaneously injected into one ank of each nude mouse. Tumor volume was calculated using the formula (Length×Width 2 )/2. Tumors were harvested and weighed after 24 d.

Statistical analysis
All data are presented as the mean ± standard deviation (SD). The statistical analyses were performed using SPSS 18.0 software (IBM, New York, USA). Differences between groups were analyzed using Student's t-test (two groups) or one-way ANOVA (multiple groups). P<0.05 was considered statistically signi cant.

FAM83H-AS1 facilitated the aerobic glycolysis in ESCC cells
To explore whether FAM83H-AS1 in uences the malignant phenotype of ESCC cells by modulating the aerobic glycolysis, the effects of FAM83H-AS1 on glucose uptake and lactate production were identi ed by glycolysis analysis. As shown in Fig.3a-d, knockdown of FAM83H-AS1 signi cantly suppressed the glucose uptake (Fig.3a) and lactate production (Fig.3b) in EC9706 cells, while overexpression of FAM83H-AS1 showed the opposite effects in EC109 cells (Fig.3c,d).
qRT-PCR and Western blotting analysis were performed to detect the expression of aerobic glycolysis associated proteins in ESCC cells. The results revealed that FAM83H-AS1 knockdown decreased GLUT1 and LDH-A mRNA and protein expression in EC9706 cells (Fig.3e-g). In contrast, FAM83H-AS1 overexpression led to increased GLUT1 and LDH-A mRNA and protein expression in EC109 cells (Fig.3h-j).

Discussion
The role of lncRNAs in ESCC formation and progression has gradually emerged [5]. LncRNA FAM83H-AS1 has been reported to play an oncogenic role in tumors by modulating cell proliferation, migration and invasion [11,[21][22][23][24]. However, the role and mechanism of FAM83H-AS1 contributing to ESCC formation and progression has not been elucidated. In the present study, we rst found that FAM83H-AS1 expression was signi cantly higher than that of human esophageal squamous epithelial cell line Het-1A (Fig.1a), indicating that FAM83H-AS1 might contribute to the malignant nature of ESCC cells. The present mechanism experiments revealed the interaction between FAM83H-AS1, miR-4684-5p and ZBTB38 in ESCC progression (Fig.4,5). Then gain-and loss-of-function ( Fig.1-3) and rescue experiments (Fig.7) veri ed a promoting effect of FAM83H-AS1 on ESCC cell migration, invasion and aerobic glycolysis by modulating miR-4684-5p/ZBTB38 axis .
Previous studies suggest that aerobic glycolysis is implicated in the malignant behaviors of ESCC [25][26][27]. Enhanced aerobic glycolysis is a striking feature of ESCC which relays upon the glycolytic pathway for their energy needs [25][26][27]. Moreover, increasing data have revealed that lncRNAs modulate tumor invasion and metastasis by affecting the aerobic glycolysis [28]. Thus, as important and extensive regulators of the aerobic glycolysis, lncRNAs might be suitable candidates for ESCC diagnosis and treatment. Here, we found that overexpression of FAM83H-AS1 not only accelerated the cell migration and invasion of ESCC cells (Fig.2), but also facilitated glucose uptake and lactate production (Fig.3), suggesting that the acquisition of metastastic potential of ESCC induced by FAM83H-AS1 might bene t from glucose metabolism. Therefore, we hypothesized that FAM83H-AS1 may in uence the malignant phenotype of ESCC partly by modulating the aerobic glycolysis. A number of enzymes are involved in the aerobic glycolysis process, including GLUT1 and LDH-A [19,29]. LDH-A can accelerate the invasion and proliferation of pituitary adenoma through the upmodulation of GLUT1 [30]. Thus, we determined the GLUT1 and LDH-A expressions following overexpression or downmodulation of FAM83H-AS1, and found that knockdown or overexpression of FAM83H-AS1 decreased or increased the expressions of GLUT1 and LDH-A (Fig.3), as well as the downmodulation or upmodulation of glucose uptake and lactate production levels, indicating that FAM83H-AS1 may affect aerobic glycolysis process of ESCC partly by modulating GLUT1 and LDH-A.
The competing endogenous RNAs (ceRNAs) hypothesis proposes a novel regulatory network involving lncRNAs, miRNAs, circular RNAs (circRNAs) and pseudogenes, and such a network is of great signi cance in tumor formation and progression [31]. Accumulating evidence demonstrated that lncRNAs exerted biological roles in various tumors by functioning as ceRNAs for sponging miRNA to modulating miRNAs' target gene expression [32]. LncRNAs can block the repression of miRNA on its target gene by competitively binding to the miRNA, thus serving as tumor promoters or tumor suppressors [33]. Thus, we supposed that FAM83H-AS1 may exert an oncogenic role in ESCC by functioning as a ceRNA for sponging miRNAs. By use of informatics analysis tools, we screened a few miRNAs sharing binding sites with lncRNA FAM83H-AS1, and found the highest-ranked predicted potential target of FAM83H-AS1 was miR-4684-5p (Table 2, Fig.4a,b). Through luciferase activity, RNA pull-down and RIP assays, the binding relationship between FAM83H-AS1 and miR-4684-5p was identi ed in ESCC cells (Fig.4c-e). Therefore, miR-4684-5p was chosen for our object, and we speculated that FAM83H-AS1 might serve as a ceRNA to absorb miR-4684-5p.
Since the function of miRNAs are realized by modulating their target genes' mRNA expression [34], we here conducted the bioinformatics analysis and veri cation experiments (Table 3, Fig.5), and showed that ZBTB38 was the potential downstream target of miR-4684-5p. ZBTB38 is a transcriptional activator that belongs to the zinc nger protein family and contains the typical BTB (for Broad-Complex, Tramtrack and Bric a brac) domains [12]. As a transcription factor, ZBTB38 is involved in cell regulation, proliferation and apoptosis, whereas functional de ciency of ZBTB38 induces the human neuroblastoma cell death [12]. Moreover, ZBTB38 play an oncogenic role in neuroblastoma and bladder cancer [12,13]. Based on these studies, we hypothesized that FAM83H-AS1 may be partially required for ZBTB38 to exert its oncogenic effect in ESCC. To address this point, the present rescue experiment showed that restoration of ZBTB38 partly blocked FAM83H-AS1 knockdown-induced suppression of cell migration, invasion and glycolysis in ESCC cells (Fig.7). Hence, the results suggested that FAM83H-AS1 may promote cell migration, invasion and glycolysis of ESCC cells by modulating ZBTB38.
Our present results supported the notion that FAM83H-AS1 acts as a molecular sponge for miR-4684-5p to modulate the expression level of ZBTB38. Therefore, the above ndings suggested the possibility that FAM83H-AS1/miR-4684-5p/ZBTB38 may be a novel mechanism responsible for ESCC progression.

Conclusion
We highlight that FAM83H-AS1 acts as a ceRNA by sponging miR-4684-5p to promote ZBTB38 expression, thereby facilitating the aerobic glycolysis, tumor invasion and metastasis of ESCC (Fig.8     Bioinformatics analysis tools suggested that miR-4684-5p might share the binding sites with FAM83H-AS1. c Luciferase reporter assay indicated the molecular level combination of miR-4684-5p and FAM83H-AS1. d FAM83H-AS1 was pulled down by biotinylated miR-4684-5p-WT. e AGO2-RIP followed by qPCR to detect FAM83H-AS1 enrichment level after miR-4684-5p overexpression via miR-4684-5p mimics. *P<0.05, and **P<0.01 vs control. Bioinformatics analysis tools suggested the associability within miR-4684-5p and ZBTB38. c Luciferase reporter assay indicated the covalent targeting of miR-4684-5p with ZBTB38 mRNA 3′-UTR in EC9706 cells. d RIP assays using antibodies against AGO2 or IgG were performed in cellular lysates from EC9706 cells, and qRT-PCR demonstrated the relative enrichment of ZBTB38 in EC9706 cells transfected with miR-4684-5p or control mimics. *P<0.05, and **P<0.01 vs control.  Effects of miR-4684-5p inhibition or ZBTB38 overexpression on the suppression of cell migration, invasion and glycolysis induced by FAM83H-AS1 knockdown in EC9706 cells. The transwell assay was performed to detect the invasion (a) and migration (b). Glucose uptake and lactate production assay was performed to detect the glucose uptake (c) and lactate production (d). *P< 0.05, **P<0.01 vs si-NC; #P<0.05, ##P<0.01 vs vector control.