The procedures of the animal experiments were approved by the Ethics Committee of Xiangya Hospital of Central South University and followed the guidelines for the Use and Care of Laboratory Animals by International Committees.

Normal liver tissues (n = 10) and NAFLD liver tissues (n = 20) were obtained from Xiangya Hospital of Central South University between August 2019 and September 2020. Liver tissues in NAFLD were obtained from patients with NAFLD. Normal liver tissues were obtained from patients who had been in a car accident. Clinical samples were collected: the participants were aged 18–64 years old and had a body mass index (BMI) ≥ 25 kg/m²; The following were the exclusion criteria: Individuals who reported ongoing or recent alcohol consumption (men’s standard drinks 21/week and women’s standard drinks 14/week); people who had a history of exposure to hepatotoxic drugs, with liver disease or liver cancer; had an active chronic gastrointestinal disorder (e.g., inflammatory bowel disease, ulcerative colitis, celiac disease) or taking any medication or supplement known to affect body composition (Wilson et al., 2016). The present study was approved by the Institutional Ethical Committee of Xiangya Hospital of Central South University (IRB{S}NO. 202104073). Written informed consent was obtained from all participants.

Liver tissues of patients were fixed in 4% paraformaldehyde in PBS overnight, paraffin-embedded, and cut into 5-μm-thick sections. Paraffin sections were deparaffinized and rehydrated. The sections were heated in sodium citrate buffer in a microwave for antigen retrieval, washed with phosphate-buffered saline (PBS) for 5 min (three times) at room temperature, incubated in 3% H₂O₂ at room temperature for 25 min, washed with PBS for 5 min (three times) and blocked and incubated in goat serum for 30 min. Then, the samples were stained with primary antibodies (anti-ARNT, anti-CD36 and anti-PPARα) overnight at 4°C. After that, samples were incubated with secondary antibody (HRP-labeled) for 30 min at 37°C. Finally, freshly prepared diaminobenzidine (DAB) was added for colour development. All the antibodies were obtained from Abcam. The tissues were observed under a fluorescence microscope.

The NAFLD animal model was in line with a previously report (Zhou et al., 2008). Male C57BL/6 mice (8 weeks old, n = 6/group) were from Vital River Laboratory Animal Technology (Beijing, China). Four groups included standard chow diet (SCD; 65% carbohydrate, 10% fat and 25% protein) group, high fat diet (HFD; 35% carbohydrate, 50% fat and 15% protein) group, HFD + sh-NC group and HFD + sh-MALAT1 group. After 12 weeks of HFD treatment, mice were sacrificed for collection of blood samples and liver tissues. The standard feed and high fat feed used in the study were customized by Beijing Keao Xieli Feed Co., Ltd (Beijing, China).

HepG2 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States). The culture condition was Dulbecco’s modified Eagle’s medium (DMEM; Gibico, United States) containing 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, United States) in 5% CO₂ and 37°C. To mimic NAFLD in vitro, HepG2 cells were treated with 1 mM free fatty acids (FFA; oleate: palmitate = 2:1) for 24 h. Subsequently, cells in each group were isolated for RNA/protein, and then the isolated proteins were used for further study.

The shRNA MALAT1 (sh-MALAT1), sh-negative control (sh-NC), miR-206 inhibitor, inhibitor NC, miR-206 mimics, miR-NC, sh-ARNT and sh-PPARα were obtained from GenePharma (Shanghai, China). HepG2 cells transfected with sh-MALAT1, sh-NC, miR-206 inhibitor, inhibitor NC, miR-206 mimics, miR-NC, sh-ARNT or sh-PPARα by using Lipofectamine 3000 (Invitrogen).

The liver tissues of mice were extracted and then fixed with 4% polyformaldehyde in phosphate buffer, after which the samples were dehydrated and embedded with paraffin. Subsequently, the as-prepared paraffin block was sliced into 4 μm slices, and then stained with 5% hematoxylin (10 min) and 1% eosin (5 min). Finally, the samples were observed with a microscopy.

First, the treated frozen sections of liver tissues and HepG2 cells were subjected to 4% paraformaldehyde (10 min). After washing, Oil-Red O staining solution was added in both liver samples and cell samples for 10 min and hematoxylin staining solution was used for nuclear staining. Results of lipid accumulation were inspected under microscope (Carl Zeiss MicroImaging GmbH, Göttingen, Germany).

A TG assay kit (Applygen Technologies Inc., Beijing, China) was used to examine TG level in lysed cells. TG concentration was normalized to total protein concentration which was detected via the BCA method, expressing as mmol/mg protein.

The total RNA of transfected cells and liver tissues were first collected with a TRIzol reagent (Invitrogen). After that, the reverse transcriptions of RNA into cDNA were conducted through a reverse transcriptase kit from Invitrogen Company. The used primers were shown in Table 1. PCR processes were conducted on ABI 7500 Real-Time PCR System. The relative expression levels of RNA normalized to GAPDH or U6 were calculated using the 2^−∆∆Ct method.

Total protein from liver tissues or cells was isolated by using lysis buffer. Then, extracted proteins were first transferred into the buffer solution, then boiled for 5 min at 100°C, and collected with centrifugation. Subsequently, 30 µg protein samples were separated by SDS-PAGE and then transferred to PVDF membrane (Millipore, Merck Shanghai, United States). After incubating with 5% skim milk, the membrane was placed with primary antibodies (anti-PPARα, ab126285; anti-CD36, ab252922; anti-ACC1, ab109368; anti-SCD1, ab236868 at 1:1000, from Abcam; anti-ARNT, PA5-106857 at 1:1000, from Sigma; anti-SREBP1, NB100-2215; anti-FASN, AF5927 at 1:1000, from Novus). After washing, membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (1:3000; Abcam) for 1 h. Lastly, enhanced chemiluminescence system (ECL; Millipore) was used to visualize protein bands and relative densitometry was analyzed using ImageJ software.

BODIPY fatty acid uptake experiment was referred to Khalifeh-Soltani et al. (Garcia et al., 2022). In brief, 1 µM BODIPY-C16 (Invitrogen) was allowed for incubation with HepG2 cells at 37°C for 30 min. Then, surface-associated BODIPY was removed at 4°C with 0.2% BSA in PBS. Finally, the fluorescence of BODIPY was detected via a fluorescence plate reader.

Mouse blood samples were obtained by eye-balls removal method. Using a biochemical auto-analyzer (Fuji Medical System, Tokyo, Japan), total cholesterol (TC), TG, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were detected.

The wild-type/mutated 3′UTRs of ARNT (ARNT-WT/ARNT-MUT) luciferase vectors were formed by using amplified DNA sequences cloned into pmirGLO reporter plasmids (Promega, Madison, WI, United States). And the MALAT1-WT/MALAT1-MUT luciferase reporter vectors constructed as above. HepG2 cells were co-transfected with luciferase vectors and miR-206 mimics or miR-NC with Lipofectamine 3000 transfection reagent (Invitrogen). The luciferase activities were measured after 48 h transfection. The relationship between PPARα promoter and ARNT was also ascertained by luciferase assay as described above.

The RIP assay was conducted by adopted an EZ-Magna RIP kit (Millipore). First, lysis buffer was used to lyse the collected cells, and then magnetic beads with antibody targeting Ago2 or IgG as well as RIP buffer were added into the lysate for incubation overnight at 4 °C. After that, proteinase K was employed to incubate with these magnetic beads, and total RNAs were subsequently isolated for the relative enrichment of MALAT1 and miR-206 analysis via qPCR.

HepG2 cells at the density of 2 × 10⁶ cells/mL were used for this experiment. After adding formaldehyde, the cell suspension was rotated gently for 10 min. Then glycine was incubated for 5 min crosslink. After centrifuging at 1000 × g for 5 min, the pellet was resuspended in nuclear lysis buffer containing protease and phosphatase inhibitor cocktail. The sheared chromatin was obtained using a probe sonicator and proceeded to immunoprecipitation. In brief, primary antibody was added and sample was rotated overnight and Dynabeads was incubated for 2 h. After elution and purification, sample was proceeded with ChIP-qPCR.

All data were presented as mean ± SD and GraphPad Prism 6 software was used for statistical analysis. Student’s t test was used to assess the difference between two groups. One-way ANOVA was adopted when three or more groups were compared. p value <0.05 represented the significant difference. StarBase (http://starbase.sysu.edu.cn/) and miRmap (https://www.mirmap.ezlab.org/app/) were applied to predict the binding sites between miR-206 and MALAT1/ARNT. The interaction of ARNT and PPARα promoter was predicted via JASPAR (http://jaspar.genereg.net/). Meanwhile, three independent experiments were performed in this research.

In order to investigate the role of MALAT1, miR-206 and ARNT in NAFLD, qPCR was performed. As revealed in Figure 1A, the levels of MALAT1 and ARNT in NAFLD tissues were significantly higher, compared with those in normal tissues. In contrast, the level of miR-206 in NAFLD tissues was much lower than that in normal tissues (Figure 1A). In addition, the data of IHC showed that the levels of ARNT and CD36 in NAFLD tissues were significantly up-regulated, while the levels of PPARα were down-regulated (Figure 1B). In vivo model of NAFLD was firstly established, and then the following experiments were carried out. As shown in Figures 1C–F, the levels of TC, TG and the enzyme activities of ALT and AST in serum of mice were up-regulated in the HFD group than those in SCD group. In addition, obvious steatosis and excessive accumulation of lipids were observed in HFD mice (Figure 1G). The levels of MALAT1 and CD36 were markedly increased, while the expression of PPARα was apparently downregulated in HFD mice (Figures 1H,I). Besides, FFA treatment remarkably promoted lipid accumulation and increased the level of TG in HepG2 cells (Figures 1J,K). Meanwhile, the expressions of MALAT1 and CD36 in HepG2 cells were up-regulated by FFA, while the level of PPARα was downregulated (Figures 1L,M). Together, MALAT1 was up-regulated in NAFLD.

The level of MALAT1 was increased by FFA in HepG2 cells, while sh-MALAT1 dramatically reversed FFA-induced increase of MALAT1 expression (Figure 2A). Additionally, knockdown of MALAT1 facilitated the expression of PPARα and downregulated the expression of CD36 (Figures 2B,C). As expected, FFA-induced increase of TG level was restored in the presence of MALAT1 knockdown (Figure 2D). Knockdown of MALAT1 could suppress FFA uptake and lipid accumulation in FFA-treated hepatocytes (Figures 2E,F). It is worth mentioning that FFA-increased lipid synthesis-related proteins (SREBP1, ACC1, SCD1, FASN) were markedly abolished by MALAT1 silencing (Figures 2G,H). Besides, FFA-induced decrease of p-ACC1 and PGC-1α levels was significantly rescued by knockdown of MALAT1, and MALAT1 shRNA notably restored the effect of FFA on the level of PPARγ (Figures 2I,J). These results demonstrated that knockdown of MALAT1 could reverse FFA-induced lipid accumulation via mediation of PPARα/CD36 axis.

Next, we further investigated the mechanisms by which MALAT1 regulates lipid accumulation in hepatocytes. As shown in Figures 3A,B, FFA markedly downregulated the level of miR-206, while increased the expression of ARNT. Furthermore, miR-206 had binding sites with MALAT1/ARNT (Figures 3C,D). Moreover, miR-206 overexpression notably decreased the luciferase activity in MALAT1-WT reporter, while no obvious changes was found in MALAT1-MUT reporter group (Figure 3E). In addition, RIP analysis showed that MALAT1 and miR-206 were notably enriched in Ago2 (Figure 3F), and miR-206 distinctly decreased the luciferase activity of ARNT-WT reporter (Figure 3G). Knockdown of MALAT1 evidently upregulated the level of miR-206 and downregulated the level of ARNT; more importantly, miR-206 downregulation partially reversed the effects of sh-MALAT1 (Figures 3H,I). Therefore, MALAT1 upregulated the expression of ARNT through sponging miR-206.

To further ascertain the regulatory function of MALAT1, we then investigated the effects of miR-206 on lipid accumulation in hepatocytes. As shown in Figure 4A, MALAT1 knockdown could downregulate the expression of ARNT, CD36 and upregulate the expression of PPARα, while the inhibition of miR-206 partially reversed the effects of sh-MALAT1. In addition, miR-206 down-regulation also reversed the effects of MALAT1 silencing on TG level in FFA-treated HepG2 cells (Figure 4B). Furthermore, from BODIPY-labeled fatty acid uptake assay (Figure 4C), the effects of sh-MALAT1 on FFA uptake were partially decreased with the inhibition of miR-206. Via Oil Red O staining, the same trends can be found in lipid accumulation (Figure 4D). Moreover, miR-206 downregulation significantly up-regulated the expression of SREBP1, ACC1, SCD1 and FASN, while MALAT1 knockdown exhibited the opposite (Figure 4E). Furthermore, the effect of MALAT1 shRNA on p-ACC1, PPARγ and PGC-1α levels in FFA-induced cells was significantly reversed by miR-206 inhibitor (Figures 4F,G). All these results proved that miR-206 inhibitor reversed the effect of MALAT1 silencing on FFA-induced lipid accumulation.

Three binding sites between motif of ARNT and PPARα promoter were predicted ( Figures 5A,B). In addition, ChIP results showed that ARNT directly bound to PPARα promoter (containing E1 but not E2 or E3), and luciferase assays proved that E1 (GTACGTGA) was a functional site (Figures 5C,D). Furthermore, ARNT down-regulation markedly upregulated the level of PPARα and downregulated the expression of CD36, while knockdown PPARα reversed these phenomena (Figure 5E). All these results indicated that ARNT could regulate the expression of PPARα via binding to its promoter.

Finally, we investigated the detailed relation between PPARα and MALAT1 in lipid accumulation. As shown in Figure 6A, MALAT1 knockdown significantly upregulated the expression of PPARα and downregulated the level of CD36, and sh-PPARα significantly reversed the effects of sh-MALAT1. In addition, knockdown of PPARα also reversed sh-MALAT1-decreased TG level in FFA-induced cells (Figure 6B). The same trend could be found in BODIPY-labeled fatty acid uptake and Oil Red O staining assays (Figures 6C,D), indicating that PPARα reversed the effect of MALAT1 shRNA on FFA-induced lipid accumulation. Furthermore, this conclusion could be further confirmed by the qPCR, which showed that PPARα blocked MALAT1-induced downregulation of SREBP1, ACC1, SCD1 and FASN level in FFA-treated HepG2 cells (Figure 6E). Collectively, MALAT1 may modulate FFA-induced hepatic lipid accumulation through miR-206/ARNT axis.

As shown in Figures 7A,B, HFD-induced increase of TC and TG contents in mice was notably reversed by knockdown of MALAT1. Consistently, MALAT1 silencing notably rescued the effect of HFD on ALT and AST activity in mice (Figures 7C,D). Moreover, HFD significantly induced the lipid accumulation and inflammatory injury in liver tissues, which was markedly alleviated in the presence of MALAT1 shRNA (Figure 7E). In addition, the expressions of MALAT1, ARNT and CD36 in tissues of mice were notably up-regulated by HFD, while HFD exerted the opposite effect on miR-206 and PPARα levels (Figures 7F,G). Meanwhile, the effect of HFD on MALAT1, ARNT, CD36, miR-206 and PPARα levels was significantly abolished by MALAT1 knockdown (Figures 7F,G). In summary, knockdown of MALAT1 alleviated the progression of NAFLD in vivo.