PNPLA3 I148M mediates the regulatory effect of NF‐kB on inflammation in PA‐treated HepG2 cells

Abstract Both PNPLA3 I148M and hepatic inflammation are associated with nonalcoholic fatty liver disease (NAFLD) progression. This study aimed to elucidate whether PNPLA3 I148M is involved in NF‐kB‐related inflammation regulation in NAFLD. HepG2 cells homozygous for the PNPLA3 I148M mutation were used. The human PNPLA3 promoter sequence was screened for NF‐kB binding sites using the MATCH and PATCH tools. NF‐kB‐mediated transcriptional regulation of the PNPLA3 gene was assessed by luciferase reporter assay, EMSA and ChIP‐qPCR. Wild‐type (I148I) and mutant (M148M) PNPLA3 were overexpressed using stable lentivirus‐mediated transfection. The pCMV vector and siRNA were transiently transfected into cells to direct NF‐kB overexpression and PNPLA3 silencing, respectively. A putative NF‐kB binding site in the human PNPLA3 promoter was shown to be necessary for basal and NF‐kB‐driven transcriptional activation of PNPLA3 and protein/DNA complex formation. Supershift analysis demonstrated a protein/DNA complex specifically containing the NF‐kB p65 and p50 subunits. ChIP‐qPCR confirmed the endogenous binding of NF‐kB to the human PNPLA3 promoter in response to NF‐kB overexpression and palmitic acid (PA) challenge. The silencing of PNPLA3 blocked the overexpression of NF‐kB or PA‐induced TNF‐α up‐regulation. Moreover, mutant PNPLA3 overexpression prevented NF‐kB inhibitor–induced down‐regulation of TNF‐α expression in PA‐treated HepG2 cells. Finally, the overexpression of mutant but not wild‐type PNPLA3 increased TNF‐α expression and activated the ER stress–mediated and NF‐kB‐independent inflammatory IRE‐1α/JNK/c‐Jun pathway. Human PNPLA3 was shown to be a target of NF‐kB, and PNPLA3 I148M mediated the regulatory effect of NF‐kB on inflammation in PA‐treated HepG2 cells, most likely via the IRE‐1α/JNK/c‐Jun ER stress pathway.


| INTRODUC TI ON
Nonalcoholic fatty liver disease (NAFLD) is a clinicopathological syndrome characterized by diffuse ballooning fatty degeneration of hepatocytes and fat storage and excludes effects caused by alcohol and other known harmful factors to the liver. 1 The histological spectrum of NAFLD ranges from simple steatosis to steatohepatitis, hepatic fibrosis, cirrhosis and even hepatocellular carcinoma. 1,2 The histologic progression of NAFLD involves two steps of liver injury (the two-hit theory): The first hit is mainly intrahepatic fat accumulation caused by insulin resistance, and the second hit is mainly liver injury and steatohepatitis caused by inflammation. 3 Genome-wide association studies (GWAS) have shown that the PNPLA3 gene, which is mainly expressed in the liver and fat, is susceptible to NAFLD. 4 The function of the PNPLA3 gene remains unclear, although in vitro studies have shown that the PNPLA3 protein has triacylglycerol (TG) hydrolase, 5 lysophosphatidyl acyltransferase (LPAAT) 6 and calcium-independent phospholipase A2 (iPLA2) activities. 7 The PNPLA3 I148M (rs738409) polymorphism is associated with not only liver fat content but also hepatocyte steatohepatitis, fibrosis and cirrhosis, 8 suggesting that PNPLA3 I148M plays a key role in the progression of NAFLD. As the two-hit theory of NAFLD claims that inflammation contributes to the progression of NAFLD and more inflammatory infiltration and liver damage were found in NAFLD patients carrying PNPLA3 I148M than those carrying wild-type genotype, 9 we speculated that PNPLA3 I148M is closely related to liver inflammation. Furthermore, NF-kB is the most important transcription factor regulating inflammation and involved in the pathogenesis of NAFLD progression. 10 Therefore, we are very interested in the involvement of the PNPLA3 gene in NF-kB signalling, which links inflammatory responses in NAFLD. In this study, we show that PNPLA3 is a target gene regulated by NF-kB and that the PNPLA3 M148M protein participates in regulating the palmitic acid (PA)induced inflammatory response in HepG2 cells carrying a homozygous PNPLA3 148M genotype. These findings may provide a new means to elucidate the role of PNPLA3 I148M in NAFLD.

| Cell culture and treatment
Human hepatocellular carcinoma (HepG2) cells obtained from cell bank of CAS were cultured in MEM/EBSS medium (SH30024.01B, Hyclone) supplemented with 10% foetal bovine serum (10099-141, Gibco). HepG2 cells at 60%-70% confluence were starved in serumfree medium overnight before treatment. To induce FFA overloading, the cells were treated with 500 mmol/L palmitic acid (PA, P5585, Sigma-Aldrich) for certain times. Control cells were incubated with the same medium containing the same amount of solvent (BSA) used to dissolve the PA. For treatment experiments, the cells were pretreated with 10 μmol/L PDCT (P8765, Sigma-Aldrich) for 6 hours before transfection or 10 μmol/L BAY11-7082 (B5556, Sigma-Aldrich) for 1 hour before PA treatment.

| Transient transfection and Dual-luciferase reporter assay
The
HepG2 cells were treated with 2 μg siRNA using X-tremeGENE siRNA Transfection Reagent (04476493001, Roche) in a serum-free medium for 6 hours, and then, the medium was supplemented with serum and maintained in culture for 48 hours. Cells were then either lysed to use for Western blotting or were evaluated for mRNA expression.

| Electrophoretic Mobility Shift Assay (EMSA)
Nuclear extracts were prepared from HepG2 cells using Nuclear  and data from NF-kB IP and control IP were presented as enrichment relative to input DNA. ChIP-qPCR was repeated twice to confirm the reproducibility of results. The quality of chromatin enzymatically sheared was assessed using agarose gel electrophoresis ( Figure   S1A). Dissolution curve of the primers is presented in Figure S1B.

| Isolation of total RNA extraction and analysis by qPCR
Total RNA was extracted from HepG2 cells using TRIzol reagent

| Western blotting and ELISA
Total cell lysates were extracted from HepG2 cells using cell lysis buffer (PH 6.8, 50 mmol/L Tris, 1% SDS, protease inhibitor). Nuclear protein was extracted using NucBuster™ Protein Extraction Kit  Table S2.

| Statistical analysis
Data represent the mean ± standard deviation (SD) values of three independent duplicate experiments. Statistical analysis was performed using one-way ANOVA analysis of variance followed by Student's t test. A value of P < .05 was considered to be statistically significant.

| NF-kB is involved in regulating human PNPLA3 expression
To clarify the relationship between NF-kB and PNPLA3 expression, we detected the expression of PNPLA3 protein and mRNA in HepG2 cells transiently overexpressing NF-kB p65. As shown in Figure

| Identification of a NF-kB binding site −357/−366 bp upstream of the translation start site of the human PNPLA3 gene
The Matrix Search for Transcription Factor Binding Sites (MATCH) 11 and Pattern Search for Transcription Factor Binding Sites (PATCH) programs were used to search for NF-kB binding sites up to 5.0 kilobases upstream of the human PNPLA3 promoter. The TRANSFAC 9.4 database 12 was used to construct specific binding site weight matrices for prediction. We found a putative NF-kB binding site located −357 bp to −366 bp upstream of the translation start site of human PNPLA3 (Figure 2A) that conforms well to the NF-Kb binding consensus sequence (GGG RNN YYC C, where R is A or G, and Y is C), although DNA sequence alignment analysis showed this sequence was not conserved with mouse, rat and chicken ( Figure S3).
The pGL3-WT, PGL3-Mutant and pGL3-Basic reporter constructs were transiently transfected into HepG2 cells, and basal transcriptional activity was measured 24 hours after cotransfection by dual-luciferase reporter assay. As shown in Figure

| NF-kB, rather than SREBP-1c, regulated PNPLA3 expression in long-term PA-treated HepG2 cells
Given that SREBP-1c is a known PNPLA3 target gene that can also be activated by PA, it is first necessary to clarify changes in SREBP-1c and NF-kB expression over time during PA intervention to avoid confounding effects of SREBP-1c. Expression of nucleoprotein NF-kB and SREBP-1c was detected before (0 hours) and after 6, 12 and 24 hours of PA treatment. As shown in Figure 4A, the nucleo-

| PA increases the endogenous binding of NF-kB to the human PNPLA3 promoter
The endogenous interaction between NF-kB and the PNPLA3 promoter was detected by ChIP-qPCR in HepG2 cells treated with PA F I G U R E 1 PNPLA3 expression was regulated by NF-kB in HepG2 cells. HepG2 cells were transfected with blank pCMV (pCMV-Mock) or pCMV-p65 with or without pretreatment of NF-kB inhibitor PDTC for 6 h, and then, the protein expression of PNPLA3 and nuclear NF-kB (A), and mRNA expression of PNPLA3 and TNF-α (C) were detected using Western blotting and real-time PCR, respectively. HepG2 cells were transfected with pCMV-Mock or pCMV-p65 with or without pre-transfection of pCMV-IkBαM, and then, the protein expression of PNPLA3 and nuclear NF-kB (B), and mRNA expression of PNPLA3 and TNF-α (D) were detected using Western blotting and real-time PCR, respectively. The results of real-time PCR are presented as relative mRNA levels from three independent experiments normalized to the mock transfected control. #P < .05 compared with pCMV-Mock, *P < .05 compared with pCMV-p65 for 24 hours. Cells overexpressing NF-kB due to pCMV NF-kB p65 plasmid transfection were used as a positive control. As shown in Figure 4D, the putative NF-kB target region was enriched by almost 16-fold (P < .05) over the negative control following PA treatment and almost 60-fold when NF-kB was overexpressed (P < .05). These findings suggested that PA increases the endogenous binding of F I G U R E 2 NF-kB transactivated human PNPLA3 promoter through a putative NF-kB binding site. A, Human PNPLA3 promoter upstream of the 5′UTR. The putative NF-kB binding sites are highlighted with boxes; SREBP-1c and NFY binding sites are underlined with a thick line. The ATG translation start codon where the A is numbered with 1 is indicated in bold and boxed. B, Relative luciferase activity of different PNPLA3 promoter-reporter constructs. HepG2 cells were transfected with pGL3-WT, PGL3-Mutant and pGL3-Basic reporter constructs for 24 h to measure the relative luciferase activities of PNPLA3 promoter by dual-luciferase assays. Relative luciferase activity was corrected for Renilla luciferase activity and normalized to the pGL3-Basic activity from three independent experiments. #P < .05 compared with pGL3-Basic, *P < .05 compared with pGL3-WT. C, NF-kB-driven relative luciferase activities from different PNPLA3 promoter-reporter constructs in HepG2 cells. Each PNPLA3 promoter-reporter construct was transiently cotransfected with pCMV-Mock or pCMV-p65 into HepG2 cells. Cell lysates were collected 24 h post-cotransfection, and dual-luciferase assays were performed. Relative luciferase activity was corrected for Renilla luciferase activity and normalized to the pCMV-Mock activity from three independent experiments. # P < .05 compared with pCMV-Mock. A grey-filled rectangle represents the putative NF-kB binding site, and a cross represents the mutation of the binding element NF-kB and the PNPLA3 promoter and that PNPLA3 is regulated directly by NF-kB in vivo.

| PNPLA3 mediated the inflammatory regulation by NF-kB in PA-treated HepG2 cells
To determine the role of PNPLA3 in NF-kB-mediated inflamma-

| PNPLA3 M148M, but not wild-type PNPLA3, regulated TNF-α expression and activated the ER stress-associated IRE-1α/JNK/c-Jun pathway
As PNPLA3 I148M is a well-known risk genotype for NAFLD, while

| D ISCUSS I ON
The PNPLA3 I148M is susceptible to NAFLD 4 and NAFLD progression, 8  with PA for 24 hours, PNPLA3 was transcriptionally regulated by NF-kB rather than SREBP-1c, and mutant PNPLA3 M148M, but not wild-type PNPLA3 I148I, was involved in PA-induced regulation of the inflammatory factor TNF-α. This study is the first to report the mechanism by which NF-kB regulates PNPLA3 and the role of PNPLA3 I148M in regulating inflammation in NAFLD.
Although PNPLA3 is the most studied gene susceptible to NAFLD, a functional study using knockout mice did not find any abnormal metabolic phenotype. 13,14 Moreover, liver-specific PNPLA3 cell line, which is widely used to study liver diseases, was found to carry homozygous PNPLA3 M148M genotype. 17 This characteristic makes HepG2 cells a natural mutation model in which to study the function of human PNPLA3 I148M and its role in NAFLD. Some researchers even consider HepG2 cells to be an ideal in vitro model for studying PNPLA3. 17 SREBP-1c is a transcription factor that regulates lipid synthesis and can also be activated by PA in vitro. Our previous study showed the direct transcriptional regulation of human PNPLA3 by SREBP-1c. 19 Therefore, in addition to the effect of NF-kB on PNPLA3 expression, the effect of SREBP-1c on PNPLA3 expression after PA treatment should be considered. In this study, the nuclear expression of SREBP-1c and NF-kB varied with PA treatment time.
The pattern of changes in SREBP-1c and NF-kB expression was consistent with a report from Nagaya et al 21 showing that inflammation persisted, and hepatic lipid deposition and SREBP-1c expression decreased as NASH progressed. Based on our results, 24 hours was chosen as the PA treatment time because the interaction between NF-kB and PNPLA3 was significant at this time, and nuclear SREBP-1c expression was close to its baseline level.
GWAS found that people carrying PNPLA3 I148M are susceptible to the presence and progression of NAFLD, 8 but the mechanism of this susceptibility is still unknown. Whether increased susceptibility is attribute to a functional gain due to increased LAPPT activity 6 or a functional deficiency due to inhibited TG lipase activity 5,22 remains controversial. Additionally, the mechanism by which PNPLA3 I148M promotes NAFLD progression could not be explained by its LAPPT or TG lipase activities. This study in a PA-induced NAFLD HepG2 cell model is the first to report that PNPLA3 I148M is regulated by NF-kB and involved in the regulation of TNF-α expression.
TNF-α is a proinflammatory cytokine that contributes to the second hit in NASH pathogenesis. 23  In summary, as shown in Figure 7, we found that the human PNPLA3 gene is a target of NF-kB and contains an NF-kB-binding site −357 bp to −366 bp upstream the PNPLA3 translation start site. In addition, PNPLA3 I148M was shown to be transcriptionally activated by NF-kB to increase TNF-α expression through the ER stress IRE-1a/JNK/c-Jun pathway in HepG2 cells treated long-term with PA.

F I G U R E 6
Overexpression of PNPLA3 M148M activated ER stress signal IRE-1α-JNK-c-Jun inflammatory pathway. A, PNPLA3 M148M overexpression increased TNF-α expression in a NF-kB independent way. HepG2 cells were stably infected with lentiviral PNPLA3 M148M (LV-148M), lentiviral PNPLA3 I148I (LV-148I) and mock lentivirus (LV-Mock), respectively. Nucleoprotein and cytoplasmic protein were extracted after infection to measure protein expressions of nuclear NF-kB, PNPLA3 and TNF-α by Western blotting (left panel). RNA was extracted to measure PNPLA3 and TNF-α mRNA levels by real-time PCR (right panel). The real-time PCR results are presented as the mean ± SD from three independent experiments. #P < .05 compared with LV-Mock; *P < .05 compared with LV-148M. B, PNPLA3 M148M but not I148I overexpression activated IRE-1α-JNK-c-Jun pathway. HepG2 cells were stably transfected with LV-148M, LV-148I and LV-Mock, respectively. Cytoplasmic protein was extracted to measure protein expressions of IRE-1α, total and phosphorylation JNK1/2, and c-Jun by Western blotting (left panel). The relative band intensities (right panel) in Western blots (n = 2) were determined using ImageJ and normalized to β-actin. Statistical significance was performed by one-way ANOVA. Data are presented as the mean ± SD. *P < .05 compared with LV-Mock F I G U R E 7 Scheme of PNPLA3 I148M-related inflammatory signalling in HepG2 treated with PA. PNPLA3 gene is transcriptionally upregulated by NF-kB during long-term PA treatment. The distribution of wild-type and mutant PNPLA3 proteins is different: I148M mutant proteins are distributed in lipid droplets, while wild-type proteins are distributed in the cytoplasm. PNPLA3 I148M protein then activates the IRE1a signalling of ER stress, followed by phosphorylating JNK1/2 and up-regulating c-Jun expression, which finally up-regulates c-Jundependent expression of inflammatory cytokines, such as TNF-α