LINC01705 derived from adipocyte exosomes regulates hepatocyte lipid accumulation via an miR‐552‐3p/LXR axis

ABSTRACT

Aims/Introduction

High glucose increases the accumulation of lipid droplets in hepatocytes, which eventually results in nonalcoholic fatty liver disease in patients with diabetes. However, the specific mechanism or communication between adipocyte and hepatocyte lipid metabolism is still ambiguous.

Materials and Methods

In this study, exosomes released from human adipocytes were isolated and identified by their morphology, size, and marker proteins by using transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and western blotting (WB). Gene expression was detected by qRT‐PCR and WB. Lipid accumulation was determined by oil red O staining and analyses of total cholesterol (TC) and triglyceride (TG) content.

Results

Our results showed that co‐culture of HepG2 cells with adipocytes under high glucose conditions stimulated lipid deposition and LINC01705 expression in the HepG2 cells. Exosomes extracted from adipocytes cultured under high glucose conditions had higher levels of LINC01705 than exosomes extracted from adipocytes cultured under normal glucose conditions. Moreover, LINC01705 expression was also elevated in exosomes extracted from diabetes patients when compared with exosomes isolated from normal volunteers, and exosomes from patients who had diabetes complicated with fatty liver (DCFL) had the highest levels of LINC01705 expression. Treatment of HepG2 cells with exosomes extracted from high glucose‐stimulated adipocytes promoted lipid deposition and LINC01705 expression in HepG2 cells. Further experiments showed that overexpression of LINC01705 promoted HepG2 lipid metabolism, while inhibition of LINC01705 had the opposite effect. Mechanistically, LINC01705 competitively bound to miR‐552‐3p, and treatment with miR‐552‐3p inhibitor reversed the effects induced by LINC01705 knockdown. Moreover, miR‐552‐3p was found to regulate the transcription activity of LXRα and thereby modulate lipid metabolism‐related gene expression.

Conclusions

When taken together, our findings showed that high glucose increased the LINC01705 levels in adipocyte exosomes, and thereby improved HepG2 lipid accumulation via an miR‐552‐3p/LXR axis.

High glucose‐induced adipocyte‐derived exosomes improve hepatocyte lipid accumulation. Exosomal LINC01705 carried by adipocytes could be absorbed by hepatocytes, and LINC01705 affects the lipid metabolism of hepatocytes by targeting miR‐552‐3p/LXR axis.

INTRODUCTION

Liver is one of the main organs that controls metabolic homeostasis, which plays a key role in regulating blood glucose and lipid levels ¹ , ² . Hepatocytes absorb glycerol and fatty acids in the blood, and synthesize triglycerides and cholesterol ³ . When liver injury occurs, blood glucose levels and lipid metabolism become disrupted, and the risk for developing diabetes increases ⁴ . On the other hand, liver lipid metabolism is also associated with diabetes. Approximately 66% of obese people and 50% of patients with diabetes suffer from non‐alcoholic fatty liver disease (NAFLD) ⁵ . Evidence suggests that diets with a high sugar content increase the accumulation of lipid droplets within hepatocytes, which results in nonalcoholic fatty liver disease or hepatic steatosis ⁶ , ⁷ , ⁸ . Therefore, it is of great importance to explore the mechanism of hepatic lipid metabolism disorder induced by diabetes in order to find methods for treating the disease.

Exosomes are single‐membrane, secreted vesicles with diameters of 30–200 nm. An exosome has the same topology as a cell and encloses a variety of proteins, cytokines, lipids, nucleic acids (RNAs), and glycoconjugates ⁹ . Exosomes participate in cell–cell communication by transferring their contents into different cells. Therefore, exosomes have emerged as important mediators of disease progression or therapy ¹⁰ , ¹¹ , ¹² . Research has shown that exosomes exert multiple effects on type 2 diabetes mellitus. For example, mesenchymal stem cell‐derived exosomes improve hepatic glucose levels and lipid metabolism in type 2 diabetes mellitus rats ¹³ , reverse peripheral insulin resistance, and relieve β‐cell destruction ¹⁴ . Furthermore, adipocyte‐derived exosomes may promote breast cancer aggressiveness in patients with type 2 diabetes mellitus ¹⁵ , and intestinal exosomes induced by eating a high fat diet can be taken up by macrophages and hepatocytes, leading to the development of insulin resistance ¹⁶ .

LncRNAs are non‐coding RNA molecules >200 nt in length, and are reported to participate in multiple biological processes by interacting with miRNA molecules to regulate gene expression at the epigenetic, transcriptional, and post‐translational levels ¹⁷ . In recent years, exosomes containing long noncoding RNAs (lncRNAs) have been shown to regulate communications between cells during the pathological process of diabetes ¹⁸ . LncRNA SNHG7, which is released from mesenchymal stem cell exosomes, was found to suppress the endothelial–mesenchymal transition process and also tube formation by human retinal microvascular endothelial cells in diabetic retinopathy via an miR‐34a‐5p/XBP1 axis ¹⁹ . Clinical research found that altered levels of MALAT1 and H19 derived from serum or serum exosomes are associated with type 2 diabetes ²⁰ . LINC01705 is a less studied lncRNA, but its levels were reported to be elevated in the adipose tissue of type 2 diabetes‐obese patients ²¹ . LncExpDB dataset showed that LINC01705 was highly expressed in fat tissue and preadipocytes (mesenchymal stem cells of adipose tissue), but was expressed at low levels in liver tissue.

The liver X receptor (LXR) is a nuclear receptor that forms a heterodimer by binding to RXR, and subsequently controls cholesterol homeostasis and lipogenesis by directly or indirectly regulating lipid metabolism‐related molecules ²² such as sterol regulatory element binding transcription factor 1 (SREBP1) ²³ , ²⁴ , fatty acid synthase (FASN) ²⁵ , and stearoyl‐CoA desaturase (SCD) ²⁶ . SREBPs are thought to drive transcriptional programs that increase cellular cholesterol synthesis and import ²⁷ . FASN is a key lipogenic enzyme, which catalyzes the terminal steps in de novo biogenesis of fatty acids ²⁸ . Stearoyl‐CoA desaturase is an enzymatic node which functions in the conversion of saturated fatty acids into monounsaturated fatty acids ²⁹ .

As mentioned above, LINC01705 is overexpressed in type 2 diabetes‐obese patients. However, its roles in diabetes and liver lipid accumulation have never been studied. Furthermore, it is unknown whether LINC01705 participates in hepatic lipid metabolism disorder induced by diabetes. Therefore, this research focused on exploring its function and mechanism.

METHODS

Cell lineage and cell culture

Human adipose‐derived mesenchymal stem cells (ADSCs), human adipocyte cells, and HepG2 cells were purchased from Procell (Wuhan, China). The ADSCs and adipocytes were cultured in specialized medium (CM‐H202 and CM‐H112, respectively, Procell), and the HepG2 cells were cultured in MEM medium supplemented with 10% FBS (PM150410, Procell). All cells were cultured at 37°C in a 5% CO₂ atmosphere. Next, the cells were stimulated with high glucose (25 mM) for 24 h, and normal glucose (5 mM) was used as a control. The co‐culture of HepG2 cells with adipocytes was performed by using Transwell inserts (Millipore, Burlington, MA, USA); during which, HepG2 cells were seeded into the lower chambers and adipocytes were seeded into the upper chambers.

Patient samples

Serum samples were collected from normal volunteers, diabetes patients, as well as from patients who had diabetes complicated with fatty liver (DCFL; 10 samples per group). The levels of LDL‐C, total cholesterol, and triglyceride in the serum samples were detected by ELISA. Moreover, exosomes in serum were collected and their expression of LINC01705 was determined by qRT‐PCR.

Exosome isolation and characterization

The supernatants of cultured adipocytes were collected to extract exosomes. As described in a previous study ³⁰ , the supernatants were first centrifuged at 2,000 g for 10 min at 4°C to remove cell debris, and subsequently centrifuged at 110,000 g for 70 min at 4°C to isolate the exosomes. The pellets were resuspended in 1 mL of PBS, and a 0.22 μm filter membrane was used for filter sterilization. Next, each sample was centrifuged again at 110,000 g for 70 min and washed once; after which the exosome pellets were collected in 200 μL of PBS. The size and concentration of the exosomes were measured using nanoparticle tracking analysis (NTA; Wayen Biotechnologies, Shanghai, China). Exosome morphology was examined by electron microscopy, and the purity and characteristics of the exosomes were analyzed by western blotting (WB) based on the expression of exosome markers (CD63, CD81, and TSG101).

Transmission electron microscopy (TEM)

The exosomes were separated by trypsin digestion and then fixed with 2% glutaraldehyde at 4°C for 2 days. Next, the samples were treated with 1% osmate for 30 min, followed by dehydration with a gradient ethanol series (50%, 70%, 80%, and 100%). The samples were then soaked in a 100% acetone/Epon 812 (Shell Chemical Co, Houston, Texas) solution, and ultrathin sections (60 nm) were prepared. Next, the sections were stained with 5% uranium acetate for 30–60 min, followed by staining with lead citrate for 10 min. Finally, the samples were observed under a Tecnai‐10 transmission electron microscope (Philips, Amsterdam, The Netherlands).

Real‐time quantitative polymerase chain reaction (qRT‐PCR)

Trizol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract the total RNA from tissues or cells as described in the manufacturer's instructions. cDNA was prepared by using oligo (dT) 20 and Superscript II reverse transcriptase (Invitrogen). Finally, PCR was performed using SYBR green PCR Master Mix (Qiagen, Germany), and the mRNA/miRNA expression ratios were normalized to those of GAPDH/U6 expression, respectively. The primers used for qRT‐PCR are shown in Table 1. The relative expression levels of mRNA for miRNAs were calculated using the 2^−ΔΔCt method.

Table 1.

Primer pairs used for quantitative RT‐PCR analysis

Gene ID	Sequence (5′‐3′)
GAPDH F	TGTTCGTCATGGGTGTGAAC
GAPDH R	ATGGCATGGACTGTGGTCAT
LINC01705 F	GAGGTCACAAAGTGGCTTTGG
LINC01705 R	GGCTCATTCATTGTTCACTCTACTC
SREBP1 F	CTGACCGACATCGAAGGTG
SREBP1 R	CAGGGAAGTCACTGTCTTGG
FASN F	CTTCGAGGTGTCAGAGAACG
FASN R	CCCAGTTATCCTTCCACAGC
SCD F	TGTCTCTGCTACACTTGGGA
SCD R	ATGAGCTCCTGCTGTTATGC
U6 F	CTCGCTTCGGCAGCACA
U6 R	AACGCTTCACGAATTTGCGT
miR‐552‐3p RT	CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTTGTCTA
miR‐552‐3p F	ACACTCCAGCTGGGAACAGGTGACTGGTTAG
miR‐552‐3p R	AACAGGTGACTGGTTAGACAA

F, forward primer; R, reverse primer; RT, reverse transcription primer.

Western blotting (WB)

Total proteins were extracted from exosomes or cells, and the protein concentration in each extract was measured using the bicinchoninic acid protein assay (BCA; Pierce, Rockford, IL, USA). Next, a 40 μg sample of total protein from each extract was separated by Tris‐glycine SDS‐PAGE (4%–20%), and the protein bands were transferred onto PVDF membranes (Sigma‐Aldrich, St Louis, MO, USA), which were subsequently blocked with non‐fat milk. The membranes were then incubated for 8–10 h with the following primary antibodies: anti‐CD63 (GTX17441, GeneTex, Irvine, CA, USA), anti‐CD81 (ab109201, Abcam, Cambridge, MA, USA), anti‐TSG101 (ab125011, Abcam), anti‐SREBP1 (SAB2102992, Sigma), anti‐FASN (ab99359, Abcam), anti‐SCD (ab236868, Abcam), and anti‐β‐actin (ab8227, Abcam), followed by incubation with a secondary antibody conjugated with horseradish peroxidase. The immunostained protein bands were then visualized using an enhanced chemiluminescence kit (ECL; Pierce).

Cell transfection

Plasmids containing short interfering RNA targeting LINC01705 (si‐LINC01705) or LINC01705 overexpression fragments (OE‐LINC01705) were used to achieve the knockdown or overexpression of LINC01705, and empty plasmids (vectors) served as controls (GenePharma, Shanghai, China). An miR‐552‐3p inhibitor (GenePharma) was used to knockdown miR‐552‐3p expression. Lipofectamine®3000 reagent (Invitrogen) was used for all transfections. The cells were transfected for 48 h and then collected for subsequent analysis.

Measurements of intracellular triglyceride (TG) and total cholesterol (TC)

The levels of intracellular triglyceride and total cholesterol in FFA (free fatty acids)‐exposed HepG2 cells were detected by using two assay kits (BC0625 and BC1985, respectively, Solarbio, Beijing, China) and measured with an automatic biochemical analyzer (Olympus Corporation, Tokyo, Japan).

Oil red O staining

After treatment with different conditions, the HepG2 cells were first incubated for 24 h with 0.5 mmol/L FFA; after which, they were washed twice with phosphate‐buffered saline (PBS), fixed for 20 min with 4% formaldehyde, and thereafter rewashed three times with PBS. The cells were then stained for 15 min with oil red O (D027‐1‐1, Nanjing Jiancheng Bioengineering Institute, Nanjing, China), and then washed three times with PBS. The resultant lipid droplets were observed and photographed under an electron microscope (Olympus).

Dual‐luciferase reporter assay

The LINC01705 sequence containing the miR‐552‐3p binding site (UCACCUGU) was fused onto the 3′UTR of a psicheck2 luciferase reporter vector (Promega, Madison, WI, USA) to produce a wild‐type reporter vector (wild type). Site‐directed mutagenesis was performed to produce a mutant of the miR‐552‐3p binding site in the LINC01705 sequence (AGUGGACA), which was subsequently inserted into a luciferase reporter vector to produce mutant‐type reporter vectors (mutant type). Next, Lipofectamine®3000 was used to co‐transfect HepG2 cells with the miR‐552‐3p mimics and reporter vectors for 48 h; after which, luciferase activity was measured using a Dual‐Luciferase Reporter System (Promega).

Detection of LXRa transcription

The LXRa binding element (LXRE) was cloned into the psicheck2 luciferase reporter vector. Next, the cells were co‐transfected with si‐LINC01705 and the miR‐552‐3p inhibitor along with the reporter vector, and luciferase activity was measured using the dual‐luciferase reporter system.

RNA fluorescence in situ hybridization (FISH)

A Specific Alexa Fluor 647‐labeled miR‐552‐3p probe (5′‐UUGUCUAACCAGUCACCUGUU‐3′) was obtained from Servicebio Technology (Wuhan, China). After being digested with proteinase K, the fixed HepG2 cells were pre‐hybridized with a hybridization solution and then incubated overnight at 42°C with the miR‐552‐3p probe in hybridization buffer. After being blocked with rabbit serum, the cells were incubated for 50 min at 37°C with mouse anti‐DIG‐HRP; after which, they were incubated for 5 min with Cy3‐TSA reagent at room temperature in the dark. Cell nuclei were stained with DAPI for 5 min at room temperature. MiR‐552‐3p displayed a red florescence. The cells were then photographed under a fluorescence microscope (Olympus).

RNA immunoprecipitation (RIP) assay

HepG2 cells (1 × 10⁷) were lysed with RIP lysis buffer (Millipore, Billerica, MA, USA). The cell lysates were then incubated with RIP immunoprecipitation buffer containing magnetic beads coated with human anti‐LXRa or negative control mouse IgG (ab176323 and ab190475, Abcam). Next, the samples were incubated with proteinase K. At the end of the incubation period, the immunoprecipitated RNA was isolated and analyzed by qRT‐PCR to determine the level of miR‐552‐3p.

RNA pull‐down

RNA pull‐down assays were conducted as described previously ³⁰ , ³¹ . In short, 1 × 10⁷ cells were washed in ice‐cold phosphate‐buffered saline, lysed in 500 μL co‐IP buffer, and then incubated with 3 μg of biotin‐labeled miR‐552‐3p probe to bind endogenous miR‐552‐3p at room temperature for 2 h. Next, 50 μL of washed Streptavidin C1 magnetic beads (Invitrogen) was added to each binding reaction and incubated at room temperature for 50 min; after which, the beads were carefully washed with co‐IP buffer. Finally, the RNA‐binding protein LXRa was detected by western blotting.

Electrophoretic mobility shift assay (EMSA) and shift western blotting

A specific probe for LXRE was synthesized by Generay Biotech (Shanghai, China) and labeled with digoxigenin‐11‐ddUTP. A 5 mg extract of nuclear protein was incubated with the probe and binding buffer, along with the miR‐552‐3p mimics or miR‐552‐3p‐MUT. Next, the DNA binding complexes were separated by electrophoresis on a non‐denaturing polyacrylamide gel, and transferred onto a nylon membrane. Finally, CSPD was used to examine the chemiluminescence.

Statistical analysis

All data were analyzed using IBM SPSS Statistics for Windows, Version 20 software (IBM Corp., Armonk, NY) and the results are expressed as a mean value ± standard deviation (SD). All data were analyzed for normality by using the Shapiro–Wilk test. Comparisons of normally distributed data between two groups were performed using the Student's t‐test, and differences between normally distributed data obtained from more than two groups were analyzed by one‐way anova, followed by the Dunnett T3 test. The Whitney U test was used to compare non‐parametric data. A value of P < 0.05 was considered to be statistically significant.

RESULTS

Lipid metabolism in HepG2 cells was affected by adipocyte co‐culture

To explore the role of LINC01705 in diabetes and liver lipid accumulation, we first examined its expression in cultured HepG2 cells, adipose‐derived mesenchymal stem cells (ADSCs), and human adipocyte cells, respectively, that were cultured under standard culture conditions. The results showed that adipocytes had the highest levels of LINC0175 expression, and HepG2 cells had the lowest levels of LINC0175 expression (Figure 1a). We next used a Transwell system to establish a co‐culture cell model for use in investigating the effect of adipocytes on HepG2 cell lipid accumulation. Meanwhile, we also added high glucose (HG) to the model in order to mimic the diabetes condition. Subsequent oil red staining revealed that either high glucose treatment or co‐culture with adipocytes could effectively induce ectopic lipid deposition in HepG2 cells. Moreover, high‐glucose treatment further increased the lipid deposition in HepG2 cells that were co‐cultured with adipocytes (Figure 1b). Similarly, either high‐glucose treatment or co‐culture with human adipocytes increased the contents of triglyceride and total cholesterol in the HepG2 cells, and high‐glucose treatment further increased the contents of triglyceride and total cholesterol in HepG2 cells that had been co‐cultured with adipocytes (Figure 1c). Next, the total RNA and total proteins of HepG2 cells in each group were extracted and analyzed for the expression of genes related to lipid metabolism. Those results showed that high‐glucose treatment and co‐culture with adipocytes significantly increased the expression of SREBP1, FASN, and SCD. Furthermore, high‐glucose treatment increased SREBP1, FASN, and SCD expression in HepG2 cells that had been co‐cultured with adipocytes (Figure 1d). Our data also showed that LINC01705 expression in HepG2 cells was increased by high‐glucose treatment or co‐culture with adipocytes. Under conditions of co‐culture with adipocytes, high‐glucose treatment further increased LINC01705 expression in HepG2 cells (Figure 1e).

Figure 1.

Lipid metabolism in HepG2 cells was affected by co‐culture with adipocytes. (a) LINC01705 mRNA expression in HepG2 cells, ADSCs, and adipocytes was detected by qRT‐PCR. (b) Representative images of oil red O staining in different groups of HepG2 cells cultured under FFA‐induced steatosis conditions. (c) The triglyceride and total cholesterol contents of HepG2 cells treated with FFA. (d) The levels of SREBP1, FASN, and SCD mRNA and protein in HepG2 cells were detected by qRT‐PCR and western blotting. (e) LINC01705 expression in HepG2 cells was detected by qRT‐PCR. Ctrl, normal HepG2; HG, HG‐treated HepG2; Co‐culture, HepG2 co‐cultured with adipocytes; HG + Co‐culture, HepG2 cells co‐cultured with adipocytes in HG medium. *P < 0.05, **P < 0.01, ***P < 0.001, vs HepG2 or Ctrl group; ###P < 0.001, ADSC vs adipocyte or HG vs HG + Co‐culture; $$$P < 0.001, Co‐culture vs HG + Co‐culture.

Identification of exosomes and LINC01705 expression in exosomes

Next, we explored how co‐culture with adipocytes regulates LINC01705 expression in HepG2 cells. Exosomes can transfer their contents, including lncRNA molecules, into cells. We therefore speculated that adipocytes might affect HepG2 cells via the release of exosomes. To confirm that hypothesis, we extracted adipocyte‐derived exosomes and identified them by analyzing their morphology, size, and marker proteins by use of TEM, NTA, and WB, respectively. Exosomes are hollow spherical microvesicles with a diameter of approximately 100 nm. Our TEM and NTA analyses showed that the sizes of the extracted exosomes were in agreement with that specification (Figure 2a,b). Furthermore, the levels of exosome biomarker proteins (CD63, CD81, and TSG101) were greater in the microvesicles than in the cell lysates (Figure 2c). These findings indicated that the exosomes had been successfully isolated. Next, the levels of LINC01705 in exosomes extracted from normal glucose‐treated adipocytes (Ctrl) or high glucose‐treated adipocytes (HG) were detected by qRT‐PCR. Those results showed that high glucose increased the amount of LINC01705 carried by exosomes (Figure 2d). Moreover, we also examined LINC01705 expression in exosomes extracted from the serum of normal volunteers, diabetes patients, and patients who had diabetes complicated with fatty liver (DCFL). Those results showed that the levels of LDL‐C, total cholesterol, and triglyceride were all elevated in the serum of diabetes patients, when compared with normal volunteers, and were even further elevated in DCFL serum (Figure 2e). Moreover, LINC01705 expression was also higher in exosomes from the diabetes group, when compared with the normal group, and was highest in the DCFL group (Figure 2f).

Figure 2.

Identification of exosomes and LINC01705 expression in exosomes. (a) Representative TEM image of exosomes. Scale bar = 100 μm. (b) The size of adipocyte‐derived exosomes was measured by DLS. (c) Representative western blots for CD63, CD81, and TSG101 protein expression in adipocyte lysates or exosomes. (d) The levels of LINC01705 expression in exosomes extracted from normal glucose‐treated and high glucose‐treated adipocytes were detected by qRT‐PCR. (e) The levels of LDL‐C, total cholesterol, and triglyceride in serum obtained from normal volunteers, diabetes patients, and patients who had diabetes complicated with fatty liver (DCFL) were detected by ELISA. (f) LINC01705 expression in human samples was detected by qRT‐PCR. *P < 0.05, ***P < 0.001, vs Ctrl or Normal; ###P < 0.001, vs Diabetes.

The adipocyte‐secreted exosomes affected HepG2 cell lipid metabolism under conditions of high glucose

To confirm that adipocytes affected HepG2 cell lipid metabolism via the release of exosomes, the HepG2 cells were treated with exosomes derived from normal glucose‐stimulated adipocytes or high glucose‐stimulated adipocytes; after which, lipid metabolism‐related indicators and LINC01705 expression in HepG2 cells were detected. Oil red staining indicated that exosomes derived from adipocytes significantly increased cellular lipid deposition when compared with the untreated group (Ctrl group), and exosomes derived from high‐glucose‐stimulated adipocytes dramatically increased cellular lipid deposition when compared with the adipocyte exosomes group (Figure 3a). Similarly, exosomes derived from adipocytes significantly increased the triglyceride and total cholesterol contents of HepG2 cells when compared with the untreated group of HepG2 cells; while exosomes derived from high‐glucose‐stimulated adipocytes even further increased the triglyceride and total cholesterol levels in HepG2 cells when compared with the adipocyte exosomes group (Figure 3b). Next, the levels of SREBP1, FASN, and SCD expression were detected by qRT‐PCR and western blotting. Those data showed that treatment with adipocyte exosomes significantly increased the levels of SREBP1, FASN, and SCD expression in HepG2 cells when compared with the untreated group, and exosomes derived from high‐glucose‐stimulated adipocytes had an even greater effect (Figure 3c,d). Finally, LINC01705 expression in HepG2 cells was found to be increased in the adipocyte exosome treated group, and exosomes derived from high‐glucose‐stimulated adipocytes even further increased LINC01705 expression (Figure 3e). These results demonstrated that high‐glucose‐stimulated adipocyte exosomes affect hepatocyte lipid metabolism. In addition, our data also suggested that LINC01705 carried by exosomes could be absorbed by HepG2 cells, and might be involved in hepatocyte lipid metabolism.

Figure 3.

Adipocyte‐secreted exosomes produced under HG conditions affected lipid metabolism in HepG2 cells (a) Representative images of oil red O staining of HepG2 cells in different groups under conditions of HG and FFA treatment. (b) The triglyceride and total cholesterol contents of HepG2 cells cultured under HG conditions. (c) The levels of SREBP1, FASN, and SCD mRNA in HepG2 cells were detected by qRT‐PCR. (d) The levels of SREBP1, FASN, and SCD protein in HepG2 cells were detected by western blotting. (e) LINC01705 expression in HepG2 cells was detected by qRT‐PCR. Ctrl, normal HepG2; Exosome, HepG2 treated with exosomes derived from normal glucose‐stimulated adipocytes; HG‐Exosome, HepG2 cells treated with exosomes derived from high glucose‐stimulated adipocytes. **P < 0.01, ***P < 0.001, vs Ctrl group; #P < 0.05, ##P < 0.01, ###P < 0.001, Exosome vs HG‐Exosome.

LINC01705 affected hepatocyte lipid metabolism and targeted miR‐552‐3p

To explore the mechanism underlying the role of LINC01705 in hepatocyte lipid metabolism, LINC01705 expression was manipulated by transfection with si‐RNA and a LINC01705 recombinant plasmid. The efficiencies of overexpression and knockdown of LINC01705 in HepG2 cells were confirmed by qRT‐PCR (Figure 4e). Oil red staining indicated that LINC01705 overexpression significantly increased cellular lipid deposition; while LINC01705 knockdown dramatically decreased cellular lipid deposition when compared with a Scramble group (Figure 4a). Similarly, LINC01705 overexpression increased the cellular triglyceride and total cholesterol contents, while LINC01705 knockdown decreased the triglyceride and total cholesterol contents (Figure 4b). The levels of SREBP1, FASN, SCD, in HepG2 cells were also increased by LINC01705 overexpression, and decreased by LINC01705 knockdown (Figure 4c,d).

Figure 4.

LINC01705 affected HepG2 lipid metabolism and targeted miR‐552‐3p. HepG2 cells were transfected with LINC01705 siRNA or OE‐LINC01705 plasmids to achieve the knockdown or overexpression of LINC01705. (a) Representative images showing oil red O stained HepG2 mouse hepatocytes in different groups under conditions of HG and FFA treatment. (b) The triglyceride and total cholesterol contents of HepG2 cells under conditions of HG and FFA treatment. (c) The levels of SREBP1, FASN, and SCD mRNA in HepG2 cells were detected by qRT‐PCR. (d) The levels of SREBP1, FASN, and SCD protein in HepG2 cells were detected by western blotting. (e) LINC01705 expression in HepG2 cells was detected by qRT‐PCR. (f) The binding site between LINC01705 and miR‐552‐3p. (g) Wild‐type and mutant‐type LINC01705 luciferase reporter vectors were constructed and co‐transfected into HepG2 cells along with miR‐552‐3p mimics. Next, luciferase activity was examined. (h) The levels of miR‐552‐3p expression in HepG2 cells as detected by qRT‐PCR. (i) The levels of miR‐552‐3p in HepG2 cells as detected by FISH assays. ***P < 0.001, vs Vector group; ##P < 0.01, ###P < 0.001, Scramble vs si‐LINC01705.

Considering that miRNAs can mediate crosstalk between lncRNA and mRNA ³² , ³³ it was predicted by the online prediction tool LncRNASNP2 (http://bioinfo.life.hust.edu.cn/lncRNASNP/#!/) that LINC01705 might target miR‐552‐3p (Figure 4f). Therefore, we constructed wild‐type and mutant‐type LINC01705 luciferase reporter vectors and co‐transfected them into HepG2 cells along with miR‐552‐3p mimics, and then examined luciferase activity. Figure 4g shows that transfection with miR‐552‐3p mimics suppressed the luciferase activity of the wt‐LINC01705 luciferase plasmid, while the miR‐552‐3p mimics failed to change the luciferase activity of the mut‐LINC01705 luciferase plasmid. Furthermore, LINC01705 overexpression significantly inhibited miR‐552‐3p expression; while LINC01705 knockdown dramatically increased miR‐552‐3p expression (Figure 4h). Representative images of FISH assays also confirmed that LINC01705 overexpression significantly reduced the levels of miR‐552‐3p in HepG2 cells, while LINC01705 knockdown increased those levels (Figure 4i). Thus, the relationship between LINC01705 and miR‐552‐3p was confirmed.

The dynamic effects of LINC01705 and miR‐552‐3p on HepG2 cell lipid metabolism

After confirming the relationship between LINC01705 and miR‐552‐3p, we investigated the dynamic effects of the LINC01705/miR‐552‐3p axis on hepatocyte lipid metabolism. LINC01705 si‐RNA and the miR‐552‐3p inhibitor were co‐transfected into HepG2 cells, and the levels of LINC01705 and miR‐552‐3p expression were detected. Those results showed that LINC01705 knockdown significantly increased miR‐552‐3p levels, and also attenuated the decreases in miR‐552‐3p levels induced by the miR‐552‐3p inhibitor; however, the miR‐552‐3p inhibitor did not alter LINC01705 expression (Figure 5a,b). Representative images of oil red staining suggested that the miR‐552‐3p inhibitor increased cellular lipid deposition; while LINC01705 knockdown reversed the effect induced by the miR‐552‐3p inhibitor (Figure 5c). Similarly, miR‐552‐3p inhibition increased the triglyceride and total cholesterol contents of HepG2 cells, while LINC01705 knockdown produced the opposite effect (Figure 5d). The levels of SREBP1, FASN, and SCD in HepG2 cells were increased by the miR‐552‐3p inhibitor, whereas LINC01705 knockdown decreased the SREBP1, FASN, and SCD expression induced by the miR‐552‐3p inhibitor (Figure 5e,f). We then further explored the mechanism by which LINC01705/miR‐552‐3p regulated HepG2 lipid metabolism. FISH assays showed that miR‐552‐3p was located in cytoplasm and nucleus, which meant that miR‐552‐3p might regulate gene transcription. It is reported that miR‐552‐3p regulates hepatocyte lipid metabolism by binding to LXR and affecting its transcriptional activity, which is one of the main regulators of lipid metabolism related‐genes ³⁴ . We then conducted RIP assays and RNA‐pull down studies to confirm the binding between miR‐552‐3p and LXRa. Those results indicated that miR‐552‐3p bound to the LXRa protein (Figure 5g,h). In addition, LINC01705 overexpression significantly reduced the binding of miR‐552‐3p to LXRa protein (Figure 5h). Next, the transcriptional activity of LXRa was detected. The results showed that si‐LINC01705 inhibited the transcriptional activity of LXRa, while the miR‐552‐3p inhibitor increased the transcriptional activity LXRa (Figure 5i). It was reported that LXRa binds to the LXRE promoter and that miR‐552‐3p has a sequence similar to that of LXRE (Figure 5j). Therefore, miR‐552‐3p might bind to LXRa and block its binding to LXRE. To confirm that hypothesis, an EMSA assay was performed. The results showed that miR‐552‐3p mimics could weaken the binding between LXRa and the LXRE probe, while the miR‐552‐3p‐MUT could not (Figure 5k).

Figure 5.

The dynamic effects of LINC01705 and miR‐552‐3p on HepG2 lipid metabolism. HepG2 hepatocytes were transfected with LINC01705 siRNA and the miR‐552‐3p inhibitor. (a) LINC01705 and miR‐552‐3p expression were detected by qRT‐PCR. (b) The levels of miR‐552‐3p were detected by FISH assays. (c) Oil red O staining detected the formation of lipid droplets in HepG2 cells cultured under conditions of HG and FFA treatment. (d) The triglyceride and total cholesterol contents in HepG2 cells cultured under conditions of HG and FFA treatment. (e) The levels of SREBP1, FASN, and SCD mRNA in HepG2 cells were detected. (f) The levels of SREBP1, FASN, and SCD protein in HepG2 cells were detected. (g) Lysates of HepG2 cells were mixed with miR‐552‐3p or miR‐552‐3p antisense, followed by western blotting for detection of LXRa protein levels. (h) HepG2 cells were subjected to immunoprecipitation (IP) with antibodies against human LXRa after being transfected with the OE‐LINC01705 plasmid, and then analyzed by qRT‐PCR using primers specific for miR‐552‐3p. (i) Transcriptional activity of LXRa was detected by a dual‐luciferase reporter assay. (j) The binding motif of LXR (LXRE) as shown in the JASPAR database, and the sequences of miR‐552‐3p and the miR‐552‐3p‐MUT. (k) EMSA assays confirmed that miR‐552‐3p blocks the binding of LXRa with DR4. *P < 0.05, **P < 0.01, ***P < 0.001, vs Ctrl group; ###P < 0.001, si‐LINC01705 vs si‐LINC01705 + inhibitor.

DISCUSSION

A large number of patients with type 2 diabetes mellitus eventually develop NAFLD (non‐alcoholic fatty liver disease), and drugs used for treatment of type 2 diabetes mellitus, such as glitazones, show some degree of efficacy in the treatment of NASH (non‐alcoholic steatohepatitis) ³⁵ , which suggests interactions between the liver and type 2 diabetes mellitus. Previous studies showed that diets high in sugar increase the accumulation of lipid droplets in hepatocytes, which results in nonalcoholic fatty liver disease or hepatic steatosis ⁶ , ⁷ , ⁸ . However, the detailed mechanism of interaction between hepatocytes and adipocytes remains ambiguous. Because exosomes containing lncRNA have been reported to regulate cellular communication, we explored whether human adipocytes might affect hepatocytes via exosomes. We first found that co‐culture of HepG2 cells with adipocytes under high glucose conditions stimulated lipid deposition and LINC01705 expression in HepG2 cells. Next, we extracted exosomes from adipocytes treated with normal glucose or high glucose, and found that LINC01705 expression was increased in exosomes derived from the high adipocytes. Moreover, LINC01705 expression was elevated in exosomes extracted from diabetes patients when compared with exosomes obtained from normal volunteers, and exosomes obtained from patients with diabetes complicated by fatty liver (DCFL) had the highest levels of LINC01705. In addition, treatment of HepG2 cells with exosomes extracted from high‐glucose‐stimulated adipocytes also increased lipid deposition and LINC01705 expression in HepG2 cells, suggesting that adipocytes can regulate HepG2 lipid metabolism via exosomes and LINC01705 might play a role in that regulation.

Exosomes are microvesicles secreted by cells, and they carry proteins, lipids, and nucleic acids ³⁴ . Exosomes obtained from adipose tissue have been reported to have multiple functions. Exosomes derived from adipocyte progenitor cells can efficiently restore impaired diabetic wound healing by delivering lncH19 to injured tissue ³⁶ . Small Nucleolar RNA Host Gene 9 (SNHG9), delivered by adipocyte‐derived exosomes, can prevent endothelial dysfunction in obese patients by suppressing inflammation and apoptosis ³⁷ . Exosomal H19, derived from cholangiocytes, promotes hepatic stellate cell activation and cholestatic liver fibrosis ³⁸ . In our study, we isolated exosomes produced by normal adipocytes or HG‐treated adipocytes and then investigated the effects of those exosomes on hepatocyte lipid metabolism. The results showed that exosomes derived from normal treated adipocytes effectively induced hepatocyte lipid deposition, and exosomes from HG‐treated adipocytes had an even stronger effect.

LINC01705 is a rarely studied lncRNA, and it has never been studied in diabetes. In a lncRNA‐miRNA‐mRNA network study, LINC01705 levels were found to be elevated in the adipose tissue of type 2 diabetes‐obese patients, but the role and mechanism of LINC0175 have not been further explored ²¹ . In our study, we found that exosomes obtained from adipocytes cultured in a high glucose environment had higher LINC01705 levels than exosomes derived from adipocytes cultured in a normal glucose environment. Moreover, the exosomes were assimilated by HepG2 cells, and LINC01705 expression in the HepG2 cells was promoted. To examine the role of LINC01705, get‐of‐function and loss‐of‐function studies were performed using HepG2 cells. The results showed that HepG2 cells in the OE‐LINC01705 group had greater amounts of lipid deposition, while the HepG2 cells in the si‐LINC01705 group had lower amounts of lipid deposition. We next explored the mechanism of LINC01705. LINC01705 was predicted to target miR‐552‐3p, and that targeting was confirmed by a dual‐luciferase reporter assay. Furthermore, the miR‐552‐3p inhibitor reversed the effect of si‐LINC01705 on HepG2 lipid deposition.

The protective effect of miR‐552‐3p in diabetes has been studied previously. Ma et al. reported that miR‐552‐3p could inhibit PCSK9 expression, increase LDL uptake in HepG2 cells, and decrease serum LDL levels in mice fed a high fat diet ⁴⁰ . In another study, miR‐552‐3p was found to protect rat INS‐1E insulinoma cells from glucolipotoxicity ¹⁷ . Our results showed that elevated miR‐552‐3p levels suppressed lipid deposition in HepG2 cells and inhibition of miR‐552‐3p produced the opposite effect, which also verified the protective effect of miR‐552‐3p. In addition, miR‐552‐3p was reported to inhibit the transcriptional activity of the LXR (the liver X receptor), and thereby regulates the expression of molecules related to lipid metabolism ⁴¹ . Our study indicated the expression of lipid metabolism‐related genes (SREBP1, FASN, and SCD) was also regulated by LINC01705/miR‐552‐3p. Further experiments showed that the transcriptional activity of LXRa was inhibited by si‐LINC01705 and was promoted by the miR‐552‐3p inhibitor. The combination of miR‐552‐3p and LXRa was also confirmed by RNA pull‐down studies and RIP assays. LXR belongs to the nuclear receptor superfamily, the members of which usually share similar DNA motifs ³⁹ . Some members of the nuclear receptor superfamily are shown in Figure S1, and most of them have been reported to participate in hepatocyte lipid accumulation, such as ESR1 ⁴² , HNF4A ⁴³ , NR1D1 ⁴⁴ , NR2C2 ⁴⁵ , and RORA ⁴⁶ . Therefore, miR‐552‐3p may also regulate lipid molecules by a mode similar to that of LXR.

In summary, exosomes derived from adipocytes cultured in a high glucose environment were found to improve hepatocyte lipid accumulation. Exosomal LINC01705 carried by adipocytes can be absorbed by hepatocytes, and LINC01705 affects lipid metabolism in hepatocytes by targeting the miR‐552‐3p/LXR axis. Our results suggest that targeting the exosome formation process in adipocytes, LINC01705 or the miR‐552‐3p/LXR axis might be a novel method for reducing hepatocyte lipid accumulation and treating nonalcoholic fatty liver disease or hepatic steatosis in patients with type 2 diabetes mellitus.

DISCLOSURE

The author declares no competing financial interests.

Approval of the research protocol: N/A.

Informed consent: N/A.

Approval date of Registry and the Registration No. of the study/trial: N/A.

Animal Studies: N/A.

ETHICAL STATEMENT

This research is approved by the Ethical Committee of Jiangxi Provincial People's Hospital (KT2020001), and informed consents were obtained from all volunteers.

Supporting information

Figure S1 | Binding motifs of other members of the nuclear receptor superfamily.

DATA AVAILABILITY STATEMENT

The analyzed data sets generated during the present study are available from the corresponding author on reasonable request.