Long non-coding RNA NEAT1-modulated abnormal lipolysis via ATGL drives hepatocellular carcinoma proliferation

From 2008 to 2012, archival HCC tissues were obtained from patients at the First Affiliated Hospital of Harbin Medical University. Informed consent was obtained from each patient prior to biopsy or surgery, and ethical approval for the use of human subjects was obtained from the Research Ethics Committee of the First Affiliated Hospital of Harbin Medical University. Detailed characteristics of patients were summarized in Additional file 1: Table S1 and Table S2.

The L02 immortalized liver cell line and 293 T was purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Science, China. HepG2, Huh7, SK-Hep-1, and HCCLM3 were purchased from the American Type Culture Collection (Manassas, VA, USA). Huh7-luciferase-transfected and HCCLM3-luciferase-transfected cells were purchased from Berthold Technologies. All cell lines were cultured in DMEM supplemented with 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin.

Human Lenti-shNEAT1-GFP, Lenti-shPNPLA2-GFP, Lenti-PNPLA2-GFP, Lenti-vector-GFP control and Lenti-sh-control-GFP were designed and purchased from GeneChem Technologies (Shanghai, China). Transfection was performed according to standard procedures. The shRNA sequences used in this study are listed as follow:

Small interfering RNA (siRNA), si-control, miR-124-3p mimic, miR-124-3p miR-124-3p inhibitor, negative control were purchased from Ribobio (Guangzhou, China). For transfection, miR-124-3p mimic, miR-124-3p inhibitor, negative control, siRNA or si-control in Lipofectamine 2000 (Invitrogen) was transfected into cells according to the manufacturer’s instructions. The siRNA sequences are listed in Additional file 2: Table S3.

For the Cell Counting Kit-8 assay (CCK-8), cells were seeded at a density of 2500–4000 cells/well in 96-well plates. In vitro cell proliferation was assessed by Cell Counting Kit-8 (Dojindo) according to the manufacturer’s instructions. For the cell growth assays, HCC cells were seeded at a density of 0.5 × 10⁴ per well. The number of viable cells was determined at different timepoints. For the Colony formation assay, cells were seeded in 6-well plates at a density of 500–800 cells/well and cultured for 14 days. Colonies were stained with 0.5% Crystal Violet for 10 min and counted.

DAG and FFA contents were determined with a DAG ELISA Kit (BlueGene Biotech, E01D0010) and Free Fatty Acid Quantitation Kit (Sigma, MAK044) respectively, following the manufacturers’ protocols.

Western blot analysis was performed as previously described [14]. In brief, whole cell or tissue extracts were prepared using RIPA buffer. After electrophoresis, proteins were electroeluted at 120 Volts onto a polyvinylidenedifluoride (PVDF) membrane (Invitrogen). Indicated primary antibodies were used. Protein bands were visualized by an enhanced chemiluminescence assay kit (Super Signal Pierce Bio-technology). The following antibodies against ATGL/PNPLA2(Cayman, 10006409), PPARα(Abcam, ab8934), MAGL(Abcam, ab24701), HSL(Abcam,ab45422), p53(Santa Cruz, SC126) and p21(Abcam, ab109520), Bax (Abcam, ab32503), β-Actin (Cell Signaling Technology) were used. Western blot analyses were repeated at least three times.

Expression of Ki-67 and ATGL was evaluated using an immunohistochemical (IHC) method described previously [14].

1,3-Dilinoleoyl-rac-glycerol (Sigma, D9508) was dissolved in fresh dimethyl sulfoxide (DMSO) for stock solution at 50 mM (or 50 mg/ml for the in vivo study). Similarly, oleic acid (Sigma, O1008) was dissolved in fresh DMSO to 50 mM (or 50 mg/ml for the in vivo study). 1,3-Dilinoleoyl-rac-glycerol and oleic acid were mixed at a 1:1 ratio to prepare the DAG+FFA mixture. Working solution was added with pre-set DAG and FFA concentrations by mixing common serum-free medium proportionately. Final concentrations were set at 8 μM, 16 μM and 32 μM. Atglistatin (Sigma, SML1075) was dissolved in DMSO for stock solution at 10 mM. Final working concentrations of 40 μM Atglistatin was used in all experiments. Nutlin-3a(Sigma, SML0580) was dissolved in DMSO for stock solution at 5 mg/ml. Final working concentrations of 10 μM Nutlin-3a was used in experiments.

Fluorescent In situ hybridization (FISH) was performed with a Fluorescent In Situ Hybridization Kit (RiboBio, Guangzhou, China), following the manufacturers’ protocols. The CY3 labeled miR-124-3p probe were designed and purchased from GenePharma Technologies (Shanghai, China). The sequence was: 5’-GGCAUUCACCGCGUGCCUUA-3′.

The mutation detection of the p53 gene was carried out by amplification of exons 2–11 from genomic DNA with 7 pairs of primers. Primer sequences are listed in Additional file 2: Table S5. Ex TaqDNA Polymerase used in PCR amplification was purchased from Takara, Japan. Amplification was done following the manufacturers’ protocols. PCR products were sequenced using an ABI3730XL DNA sequencer (Applied Biosystems). Profiles of p53 mutation were showed in Additional file 2: Table S6.

Male BALB/c nude mice (4–6 weeks old) were obtained from the experimental animal center of the Shanghai Institute for Biological Sciences (SIBS) and housed under standard conditions and care according to the institutional guidelines for animal care. All animal experiments were approved by the Institutional Animal Care and Use Committee of Harbin Medical University. To establish an orthotopic HCC mouse model, 4 × 10⁶ cells in 100 μL of phosphate-buffered saline were subcutaneously injected into the flanks of nude mice. After 1–2 weeks, the subcutaneous tumors were resected and diced into 1 mm³ cubes, which were then implanted into the left lobes of the livers of the nude mice. A DAG and FFA mixture was used in vivo experiment. DAG and FFA solution was prepared in 20% DMSO and 15% Tween 80 in 0.9% saline. After 1 week of implantation mice were injected 20 mg/kg DAG+FFA intraperitoneally for 5 weeks (5 days per week). Mice were imaged by the bioluminescence IVIS Imaging System weekly and mice were then sacrificed.

The luciferase reporter assay was performed according to the method described previously [16].

Statistical analysis was performed with the GraphPad Prism software package (v. 4.02; San Diego, CA, USA) or SPSS 16.0 software (Chicago, IL, USA). Student’s t-test or one-way ANOVA was applied to determine the significance between groups. Statistical analyses between different treatments, in different cell cohorts or at different time points were performed using two-way ANOVA with the Bonferroni’s correction. Overall survival (OS) was compared with the Kaplan-Meier method, and the significance was determined by the log-rank test. Correlations were calculated using Spearman rank-order coefficients. Values were expressed as mean ± SD values. Statistically significant was concluded at *P < 0.05, **P < 0.01, ***P < 0.001; NS represents no statistically significant.

The intracellular lipolytic pathway is currently known to be mainly regulated by three enzymes: namely, ATGL, HSL and MAGL [5]. Thus, we first elucidated whether these three enzymes were aberrantly expressed in five pairs of HCC tissues and adjacent non-tumorous liver tissues. The results indicated that ATGL mRNA was more aberrantly expressed in HCC tissues than HSL and MAGL mRNA (Additional file 3: Figure S1A). Next, we evaluated ATGL expression in 40 pairs of HCC tissues. Quantitative real-time PCR (qRT-PCR) revealed that mRNA levels of ATGL were higher in HCC tissues than in their adjacent non-tumorous liver tissues (Fig. 1a). Moreover, we found that ATGL expression was positively correlated with tumor size. However, it was not correlated with HBV infection (Additional file 1: Table S1). We also evaluated the expression of ATGL by western blot in clinical HCC tissues. The expression of ATGL was higher in HCC tissues than in their adjacent non-tumorous liver tissues (Fig. 1b). In addition, we evaluated the expression of ATGL by immunohistochemical (IHC) staining in clinical HCC tissues using tissue microarrays and found that 76.11% (86/113) were positive in HCC tissues compared with 12.20% (5/41) in normal liver tissues (Fig. 1c). Moreover, ATGL expression was negatively correlated with patient survival in 40 HCC tissues (Fig. 1d). To elucidate the effect of ATGL on HCC carcinogenesis, we analyzed ATGL protein and mRNA levels in a panel of HCC cell lines. The results indicated that ATGL protein and mRNA levels increased progressively from healthy liver cells to HCC cells with low growth potential and, finally, to HCC cells with high growth potential (Fig. 1e, f).

ATGL has been reported to initiate the process of TAG metabolism by hydrolyzing TAG into DAG and FFA [17]. To confirm this process in HCC cells, we generated lentiviral constructs expressing ATGL-GFP and sh-ATGL-GFP to infect HCC cells. Western blot results confirmed remarkable ATGL overexpression (Additional file 3: Figure S1B). Simultaneously, Compared with the control shRNA, ATGL expression was notably decreased by lenti-sh-ATGL#2 as shown in Additional file 3: Figure S1B. Thus, sh-ATGL-#2 was chosen for further experiments. Our data indicated that overexpression of ATGL increased intracellular FFA and DAG levels in HCC cells (Additional file 3: Figure S1C). Similarly, ATGL knockdown or treatment with Atglistatin (selective inhibitor of ATGL) reduced in intracellular FFA and DAG levels (Additional file 3: Figure S1D). In addition, we found that higher levels of DAG and FFA accumulated in HCC tissues compared with in the corresponding peritumoral tissues (Fig. 1g). These experimental results suggested that ATGL was highly expressed and maintained elevated levels of DAG and FFA in human HCC tissues. This indicated that ATGL and its products, DAG and FFA, may play important roles in HCC development.

We next assess the effects of ATGL on the growth of HCC cells. According to growth curves and CCK-8 assay results, the up-regulation of ATGL significantly promoted Huh7 and HepG2 cell growth (FIig.2a and Additional file 4: Figure S2A). Colony formation assay also revealed that up-regulating ATGL expression resulted in more visible colonies compared with control in Huh7 and HepG2 cell lines (Fig. 2b and Additional file 4: Figure S2B and C).

Next we explored whether DAG and FFA facilitates HCC cell growth. CCK-8 assay results showed that treatment with DAG and FFA separately facilitated HCC cell growth in a dose-dependent manner (Additional file 5: Figure S3A and B). Simultaneously, treatment with16μM DAG together with 16 μM FFA (16 μM DAG+ FFA) in HCC cells resulted in a synergistic effect to facilitate HCC cell growth (Additional file 5: Figure S3C).

Considering the important roles of DAG and FFA in the promotion of HCC cell growth, we hypothesized that the pro-tumorigenic effects of ATGL were mediated by DAG+FFA. We reasoned that, if the pro-tumorigenic effects of ATGL were mediated by DAG+FFA, then the impaired pathogenicity of sh-ATGL cancer cells might be rescued by treatment with exogenous sources of DAG+FFA. To test this hypothesis, we treated sh-ATGL HCC cells with 16 μM DAG+FFA. According to growth curves and CCK-8 assay results, sh-ATGL significantly inhibited HCC cell growth and this effect was completely rescued by treatment with 16 μM DAG +FFA (Fig. 2c and Additional file 4: Figure S2D). Colony formation assay further confirmed this process (Fig. 2d and Additional file 4: Figure S2E and F).

To determine whether these results were reproducible in vivo, we constructed an orthotopic xenografts HCC model in nude mice. Consistent with the results from in vitro assays, larger HCC tumors were observed in Huh7 overexpressing tumors compared with Huh7 control tumors (Fig. 2e-g). There were more Ki-67 positive cells in Huh7 overexpressing tumors compared with Huh7 control tumors (Fig. 2h and Additional file 6: Figure S4A). To explore the role of DAG and FFA on tumor growth in vivo, we injected 20 mg /kg DAG+FFA intraperitoneally daily for 5 weeks. The impaired tumor growth rate of HCCLM3 sh-ATGL tumors was completely rescued in mice treated with DAG+FFA. Notably, treatment with DAG+FFA significantly promoted the growth of HCCLM3 sh-control tumors (Fig. 2i-k). Furthermore, the IHC results indicated that Ki-67 was significantly decreased in sh-ATGL tumors and this effect was completely rescued in mice treatment with DAG +FFA (Fig. 2l and Additional file 6: Figure S4B). In addition, levels of DAG and FFA in mice orthotopic HCC tissues were higher in the Huh7 overexpressing tumors than in the control tumors. In contrast, the silencing of ATGL expression had the opposite effects (Fig. 2m). In summary, these results indicated that ATGL promotes HCC cell growth by maintaining elevated levels of DAG and FFA in vitro and in vivo.

Recent publications outline a regulatory role for LncRNA in lipid metabolism [18]. Thus, we speculated that lncRNAs might modulate ATGL expression. To test this hypothesis, we screened for and identified six lncRNAs co-expressed with ATGL in liver cancer tissues using the online software tool Co-LncRNA (http://​www.​bio-bigdata.​com/​Co-LncRNA/​) [19]. Then, we knocked down the expression of these lncRNAs using siRNA to examine their effect on ATGL in HCC cells. The transfection efficiencies were detected by qRT-PCR (Additional file 7: Figure S5A). Our results demonstrated DANCR knockdown up-regulated the expression of ATGL and that NEAT1 knockdown reduced the expression of ATGL in SK-Hep-1 cells (Fig. 3a). Given that resent studies has demonstrated that DANCR promotes HCC progression [20], and our study demonstrated that ATGL also promoted HCC cell growth. Therefore we pay attention to NEAT1. Next, we assessed NEAT1 levels in HCC cells with different growth potentials and found that NEAT1 RNA levels increased progressively from healthy liver cells to HCC cells with low growth potential and, finally, to HCC cells with high growth potential (Fig. 3b). Further studies revealed that knockdown of NEAT1 by lentiviral transfection down-regulated ATGL mRNA and protein levels in SK-Hep-1 and HCCLM3 cells (Fig. 3c). The transfection efficiency was detected by qRT-PCR (Additional file 7: Figure S5B). However, we did not observe alterations in MAGL and HSL at mRNA or protein levels (Additional file 8: Figure S6.A and B).

Next, we investigated the effect of NEAT1 on lipolysis in hepatoma cells. Reductions in intracellular FFA and DAG levels were observed following NEAT1 knockdown, however overexpression of ATGL blocks this phenomenon. Simultaneously, knockdown of NEAT1 with knockdown of ATGL (or treatment with Atglistatin) resulted in reduced DAG and FFA levels in HCCLM3 and SK-Hep-1 cells, however the effects were less than additive (Fig. 3d and e).

To confirm NEAT1 was differentially expressed in HCC tissues, 40 pairs of HCC tissues and matching para-cancer tissues were used to evaluate NEAT1 expression. As expected the expression of NEAT1 was higher in HCC tissues than in the corresponding para-cancer liver tissues (Fig. 3f). Furthermore, IHC staining results indicated that ATGL was significantly higher in HCC tissues with relatively high NEAT1 expression compared to that in tissues with relatively low NEAT1 expression (Fig. 3g). Moreover, we found that NEAT1 expression was positively correlated with tumor size. However, there was no correlation with HBV infection (Additional file 1: Table S2). In addition, NEAT1 expression was found to be positively correlated with ATGL, DAG and FFA levels in HCC tissues (Fig. 3h-j). In summary, our results revealed that the NEAT1 modulates ATGL expression in HCC cells and disrupts the lipolysis of hepatoma cells via ATGL.

Next we evaluated the role of NEAT1 in HCC. According to growth curves and CCK-8 assay results, we found that NEAT1 knockdown significantly inhibited the proliferation of HCC cells. This inhibitory effect was reversed by the overexpression of ATGL (Fig. 4a and b). Colony formation assay further confirmed this effect (Fig. 4c). Given that recent publications indicated that NEAT1 is transcriptionally regulated by the tumor suppressor p53(TP53) [21, 22], the survival effect of NEAT1 may alter across different cancers. We also tested the TP53 gene mutations in 40 HCC tissues by PCR and direct sequencing. Given that exons 2-11 are coding exons of the p53 gene, we analyzed the sequence of p53 exons 2–11 for mutations. P53 mutations were found in 35% HCC tissues(14/40). The p53 mutations were located in exons 4-9 (Additional file 2: Table S6). Consistent with the published reports, our results showed that the expression of NEAT1 was higher in TP53 wild-type tissues than in the TP-53 mutant liver tissues (Additional file 9: Figure S7A). Further, treatment with 10 μM Nutlin-3a(TP53 stabilizer) for 24 h resulted in higher NEAT1 levels in TP53-wild-type Hep-G2 and SK-hep-1 cells but not in TP53 mutant Huh7 (p.Tyr220Cys) and HCCLM3 (p.Arg249Ser) cells (Additional file 9: Figure S7B). These results indicated that NEAT1 is a TP53 target gene in HCC. Given that a recent publication reported that NEAT1 expression can in turn prevent accumulation of TP53 [21]. We tested TP53 expression in NEAT1 knockdown cells. Consistent with the previous report, TP53 was upregulated following NEAT1 knockdown (Additional file 9: Figure S7C). To confirm this phenomenon, we also evaluated the expression of p21 and Bax, the target gene of p53. The results indicated that p21 and Bax was upregulated following NEAT1 knockdown in HCC cell (Additional file 9: Figure S7D). These results indicated that NEAT1 is a TP53 target gene in HCC and plays an oncogenic role in HCC. Further, ATGL is responsible for NEAT1-mediated HCC cell growth in vitro.

Considering that the pro-tumorigenic effects of ATGL were mediated by DAG and FFA, we hypothesized that the effect of NEAT1 on HCC cell growth was also mediated by DAG and FFA. We found that NEAT1 knockdown significantly inhibited the proliferation of HCC cells. This inhibitory effect was reversed by treatment with 16 μM DAG+FFA. Notably, treatment with DAG+FFA significantly promoted HCC cell growth (Fig. 4d-f).

Thus, we concluded ATGL and its products, DAG and FFA, are responsible for NEAT1-mediated HCC cell growth. Abnormal lipolysis regulated by NEAT1-ATGL axis promotes HCC cell growth in vitro.

To better understand the hepatocarcinogenesis role of NEAT1-regulated abnormal lipolysis, we established orthotopic xenografts HCC model in nude mice. The results revealed that the inhibition of NEAT1 expression led to smaller HCC tumors than control tumors, but this effect was blocked by co-transfection of sh-NEAT1 + ATGL. Additionally, sh-NEAT1 + ATGL tumors were smaller compared with ATGL overexpression tumors (Fig. 5a-c). Further, the IHC results indicated that Ki-67 and ATGL was significantly decreased in sh-NEAT1 group and this effect was rescued in mice co-transfection of sh-NEAT1 + ATGL (Fig. 5d). We then determined whether increased DAG+FFA delivery could rectify the tumor growth defect observed in sh-NEAT1 cells in vitro. Orthotopic xenografts mice were injected with 20 mg/kg DAG+FFA intraperitoneally daily for 5 weeks. The impaired tumor growth rate of sh-NEAT1 tumors was completely rescued in mice injected with DAG+FFA. Notably, treatment with DAG+FFA significantly promoted HCCLM3 sh-control tumor growth (Fig. 5e-g). Ki-67 IHC staining further confirmed this effect (Fig. 5h). In addition, levels of DAG and FFA in mice with orthotopic HCC tissues were lower in the sh-NEAT1 tumors than in the control tumors, however, this effect was reversed by co-transfection of sh-NEAT1 + ATGL (Fig. 5i and j). Taken together, these results indicate that NEAT1-regulated abnormal lipolysis contributes to HCC growth, and that this process is mediated by ATGL and its product, DAG+FFA, in vivo.

Several studies have confirmed that microRNAs are important target of lncRNAs [23, 24]. To identify whether any microRNAs represent potential targets that can be bind to ATGL and NEAT1, we used three online software tools (targetscan: http://​www.​targetscan.​org/​; microrna.​org: http://​www.​microrna.​org/​; Starbase: http://​starbase.​sysu.​edu.​cn/​.) to identify potential binding sites. We identified four microRNAs (hsa-miR-103a-3p, hsa-miR-214-3p, hsa-miR-124-3p and hsa-miR-107) that exhibited potential to bind to both NEAT1 and ATGL. MiR-124-3p, which had the highest alignment score/Pct score is reported to have anti-cancer effects, therefore, we focused our investigation on this miRNA. The predictive information between miR-124-3p and the binding sites in the NEAT1/ATGL 3′-UTRs was illustrated in Fig. 6a. We further analysed the minimum free energy value of the hybrid between miR-124-3p and the binding site on the NEAT1/ATGL 3′-UTRs. The minimum free energy values were − 14.6 kcal/mol and − 17.5 kcal/mol, respectively, which are within the range of genuine miRNA-target pairs. Given the fact that miRNAs can target the nuclear lncRNA [25, 26], we first explored the subcellular distribution of miR-124-3p by FISH. The results showed that miR-124-3p was present both in the nucleus and cytoplasm (Additional file 10: Figure.S8A). Next we explored that whether NEAT1 is able to regulate miR-124-3p expression. The results indicated that NEAT1 knockdown led to a significant increase in miR-124-3p expression in HCC cells as assayed by qRT-PCR (Fig. 6b). In addition, a dual-luciferase reporter assay demonstrated that miR-124-3p mimics decreased the luciferase activities of NEAT1-WT but failed to influence the mutant, suggesting that miR-124-3p is able to directly bind to the NEAT1-WT target sites in 293 T cells (Fig. 6c). These results prove that NEAT1 regulate miR-124-3p expression in a sequence-specific and binding-dependent manner. Next, we overexpressed miR-124-3p using a miR-124-3p mimic. Western blot and qRT-PCR results indicated that exogenous miR-124-3p significantly reduced ATGL expression in HCCLM3 and SK-Hep-1 cells (Fig. 6d). Our previous results demonstrated that knockdown of NEAT1 down-regulated ATGL expression, and we found that this suppression can be attenuated in HCC cells by inhibiting miR-124-3p (Fig. 6e). The transfection efficiencies of miR-124-3p were detected by qRT-PCR (Additional file 11: Figure S9A). In addition, a dual-luciferase reporter assay demonstrated that exogenous miR-124-3p reduced the luciferase activity of ATGL-WT but did not affect the luciferase activity of ATGL-MUT in 293 T cells (Fig. 6f). To further prove that miR-124-3p is involved in the cross-regulation between NEAT1 and ATGL, We performed a rescue experiment with a dual-luciferase reporter assay using ATGL with mutant miR-124-3p binding sites. The results showed that depletion of NEAT1 in 293 T cells inhibited the luciferase activity of ATGL-WT but not ATGL-MUT. Further, inhibition of miR-124-3p reversed this decrease in luciferase activity for ATGL-WT, but not for ATGL-MUT (Additional file 12: Figure S10A). Correlation analyses revealed that expression levels of NEAT1 and ATGL were inversely associated with that of miR-124-3p in 40 clinical HCC samples (Fig. 6g). These results support the idea that NEAT1 regulates ATGL expression through miR-124-3p.

To confirm the effect of miR-124-3p on lipolysis, we overexpressed miR-124-3p using miR-124-3p mimics in HCC cell lines. Our results indicated that both DAG and FFA levels were decreased in HCC cells (Additional file 13: Figure S11A and B). We also evaluated NEAT1 and miR-124-3p expression in five pairs of HCC and matched non-tumor tissues by qRT-PCR. NEAT1 was upregulated and miR-124-3p was downregulated in the five HCC tissues compared with the matched non-tumor tissues (Additional file 14: Figure S12A).

Considering that cancer cells have been reported to increase fatty acid oxidation (FAO) for cell survival due to compromised glucose uptake and ATGL-PPARα signaling have been reported to mediate FAO [27, 28], we hypothesized that NEAT1 might serve as a means to mediate FAO through PPARα. To test this hypothesis, we first tested whether ATGL mediated PPARα in HCC cells. Our results showed that overexpression of ATGL significantly up-regulated PPARα expression in HCC cells (Fig. 7a and b). As various studies using a variety of biochemical techniques have firmly corroborated the direct physical association between fatty acids and PPARα and have thus established fatty acids as bona fide PPARα ligands [29], we hypothesized that DAG or FFA may active PPARα expression. To examine whether DAG+FFA regulates PPARα expression, we treated HCC cells with 16 μM and 32 μM DAG+FFA for 36 h. The resulting data demonstrated that DAG + FFA up-regulates the expression of PPARα in Huh7 and HCCLM3 cells in a dose dependent manner (Fig. 7c). In addition, our results showed that knockdown of NEAT1 reduced PPARα levels, however, treatment with miR-124-3p inhibitor (or overexpression of ATGL/treatment with 16 μM DAG+FFA) blocked this process (Fig. 7d). The transfection efficiencies of miR-124-3p were detected by qRT-PCR (Additional file 11: Figure S9B). PPARα is a known oncogene in HCC [30]. Thus, we concluded that NEAT1 mediates HCC cell growth via miR-124-3p/ATGL/DAG+FFA/PPARα signaling. Moreover, NEAT1 may mediate FAO via PPARα, though this effect requires further study in the future.

Our results indicated NEAT1 expression was negatively correlated with patient survival (Fig. 8a). It was previously shown that a combination of molecular markers can improve the prediction of patient prognosis [31]. We investigated whether this was the case for NEAT1 and ATGL in HCC patients. Indeed, patients with tumors exhibiting high levels of NEAT1 and high levels of ATGL had lower survival (Fig. 8b). Thus, combining NEAT1 and ATGL expression improves the prediction of patient outcome.