SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Human patatin-like phospholipase domain-containing 3 (PNPLA3) is associated with increased liver fat content and liver injury. Here, we show that nutritional status regulates PNPLA3 gene expression in the mouse liver. Sterol response element binding protein-1 (SREBP-1) activated PNPLA3 gene transcription via sterol regulatory elements (SREs) mapped to the promoter region. Chromatin immunoprecipitation and electrophoretic mobility shift assays confirmed that SREBP-1 proteins bound to the identified SREs. Furthermore, SREBP-1c mediated the insulin and liver X receptor agonist TO901317-dependent induction of PNPLA3 gene expression in hepatocytes. Adenovirus-mediated overexpression of mouse PNPLA3 increased intracellular triglyceride content in primary hepatocytes, and knockdown of PNPLA3 suppressed the ability of SREBP-1c to stimulate lipid accumulation in hepatocytes. Finally, the overexpression of PNPLA3 in mouse liver increased the serum triglyceride level and impaired glucose tolerance; in contrast, the knockdown of PNPLA3 in db/db mouse liver improved glucose tolerance. Conclusion: Our data suggest that mouse PNPLA3, which is a lipogenic gene directly targeted by SREBP-1, promotes lipogenesis in primary hepatocytes and influences systemic lipid and glucose metabolism. (HEPATOLOGY 2011;)

Nonalcoholic fatty liver disease (NAFLD) represents a spectrum of hepatic abnormalities associated with the accumulation of excess triglycerides in hepatocytes that varies in severity from steatosis to steatohepatitis, hepatic fibrosis, and cirrhosis.1 Both environmental and genetic factors contribute to NAFLD development, but the genetic factors that predispose patients to hepatic steatosis are largely unknown. Recently, several genomewide association studies (GWAS) identified patatin-like phospholipase domain-containing 3 (PNPLA3; alternatively known as adiponutrin) as a key genetic factor conferring susceptibility to NAFLD.2-8 A common genetic variant of PNPLA3 (rs738409 C/G, encoding I148M) is strongly associated with an increased liver fat content and susceptibility to NAFLD. Furthermore, this variant has also been associated with alcohol liver disease,9 increased circulating levels of liver enzymes, and hepatic injury.7, 8, 10, 11

PNPLA3 was originally identified from 3T3-L1 preadipocytes12 and belongs to the PNPLA family of proteins, which have a common C-terminal patatin-like domain.13PNPLA3 is mainly expressed in the adipose tissue of rodents; its expression in white adipose tissue is decreased by fasting and increased by a high-carbohydrate diet.12, 14 In genetically obese fa/fa rats, the PNPLA3 messenger RNA (mRNA) level in adipose tissue is dramatically elevated.12 In contrast to animal models, the PNPLA3 mRNA level in the human liver is much higher than in adipose tissue.15 The PNPLA3 expression level in subcutaneous adipose tissue of obese subjects is significantly increased, compared to lean subjects.16 Interestingly, human PNPLA3 has both triglyceride hydrolase and acetyl coenzyme A (acyl-CoA)-independent transacylase activities in vitro.17 In addition to regulation by nutritional status, PNPLA3 gene expression can also be controlled by insulin, glucose,14, 18, 19 and the liver X receptor (LXR) agonist, TO901317.20

SREBP-1c (sterol regulatory element binding protein-1c) belongs to the SREBP family, which is composed of three members: SREBP-1a, SREBP-1c, and SREBP-2. SREBP-1a and SREBP-1c are produced from a single gene through the use of alternate promoters and different first exons, whereas SREBP-2 is produced from a separate gene.21, 22 SREBP-1c is as an important transcription factor in insulin-mediated lipogenesis signaling in hepatocytes.23 T0901317, a highly selective LXR agonist, powerfully induces the expression of SREBP-1c in the mouse liver.24

Although there is strong evidence implicating PNPLA3 in hepatic steatosis, its precise physiological function in vivo remains unknown. We show that SREBP-1c mediates insulin-dependent induction of PNPLA3 gene expression. PNPLA3 acts as a lipogenic gene to promote triglyceride synthesis in primary hepatocytes. Overexpression of mouse PNPLA3 in the mouse liver impairs glucose tolerance in mice. In contrast, knockdown of PNPLA3 in the db/db mouse liver improves glucose tolerance.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Animal Treatments.

Eight-week-old male C57BL/6J, Leprdb/+ (db/+), and Leprdb/Leprdb (db/db) mice and 12-week-old Lepob/Lepob (ob/ob) mice were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China) and housed and maintained in 12-hour light and dark photoperiods. SREBP-1c−/− or SREBP-1c+/+ mice (3 months old, male, C57BL/6J; 129S6/SvEv mixed background) were kindly provided by Professor Youfei Guan (Department of Physiology and Pathophysiology, Beijing University Health Science Center, Beijing, China). The mouse protocol was approved by the Animal Research Committee in the Institute of Laboratory Animals, Chinese Academy of Medical Sciences and Peking Union Medical College and conformed to criteria outlined in the National Institutes of Health (NIH; Bethesda, MD) Guide for the Care and Use of Laboratory Animals.

Preparation of Expression Plasmids and Recombinant Adenoviruses.

The full-length mouse PNPLA3 gene was amplified by polymerase chain reaction (PCR) from C57BL/6J mouse liver cDNA and cloned into pcDNA3.1 using the following PCR primer pairs: 5′-CGGAGCTTCAAGCACCATGTA TG-3′ (forward) and 5′-CCTGGTAGAGGGGAGCA GGCA-3′ (reverse). Recombinant adenoviruses expressing mature SREBP-1a, SREBP-1c, or PNPLA3 were generated as previously described.25

Isolation and Culture of Mouse Primary Hepatocytes.

To prepare mouse primary hepatocytes, mice were anesthetized with an intraperitoneal (IP) injection of bromethol, and hepatocytes were isolated and cultured as previously reported.26

Real-time Reverse-Transcriptase Polymerase Chain Reaction Analysis.

Quantitative real-time reverse-transcriptase PCR (qRT-PCR) was performed using the SYBR Green I Q-PCR kit (TransGen) on a Bio-Rad IQ5 system. Sequences of primers are shown in Supporting Table 1.

RNA Interference.

Short-hairpin RNA (shRNA)-encoding DNA sequences were synthesized by Invitrogen (Carlsbad, CA) and constructed into adenovirus plasmids, and adenoviruses were prepared according to previously described procedures.25 The small interfering RNA (siRNA) sequences transcribed from these adenoviruses were as follows: control scrambled siRNA (siPNPLA3-scr), 5′-CUUACGCUGAGUACUUCGA-3′ and siRNAs against PNPLA3 (siPNPLA3-1, 5′-GUGUCUGAGUUCCAUUCCA-3′ and siPNPLA3-2, 5′-GGAGAGCUGUGCUAUCAAG-3′).14

Adenoviral Infection and In Vivo Glucose Tolerance Tests.

For in vivo infections, adenoviruses such as adenovirus-containing green fluorescent protein (Ad-GFP) (control), Ad-PNPLA3 (2 × 1010 viral particles), Ad-siPNPLA3-scr (control), and Ad-siPNPLA3-2 (6 × 1010 viral particles) were delivered by tail vein injection into C57BL/6J, db/+, ob/ob, and db/db mice. Five days after injection, mice were injected with glucose gavage (1 g glucose/kg body weight) via IP injection after a 16-hour fasting, and blood-glucose levels were determined using a glucose monitor (OneTouch; LifeScan, Inc., Milpitas, CA). Seven days after injection, mice were sacrificed for metabolic analysis.

Fast Protein Liquid Chromatography Analysis of Serum Lipoprotein Distribution.

Serum from five mice per group was pooled and subjected to fast protein liquid chromatography (FPLC). Serum lipoproteins were separated with a Superose 6 column, and 0.5-mL fractions were collected for the analysis of triglyceride content by assay kits.

Hepatic Very Low Density Lipoprotein Secretion Assay.

In vivo very low density lipoprotein (VLDL) production was measured in overnight-fasted mice injected via tail vein with 500 mg/kg body weight of tyloxapol. Blood samples were collected by tail bleeding at several time points for serum triglyceride measurement.

Statistical Analysis.

Data represent the mean ± standard error (SE) values of three independent duplicate experiments. Statistical analysis was performed with one-way analysis of variance, followed by the Student's t test. P < 0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Nutritional Regulation of PNPLA3, PNPLA2, and PNPLA5 Gene Expression in the Mouse Liver.

To explore the physiological role of PNPLA3 in lipid and glucose metabolism, we first investigated the PNPLA3 gene-expression level in the liver of genetically obese and diabetic mice. Our results indicated that the PNPLA3 mRNA level was increased by approximately 30- and 7.5-fold in the liver of ob/ob and db/db mice, respectively (Fig. 1A). Additionally, SREBP-1c showed a similar expression pattern in ob/ob and db/db mice (Fig. 1A).

thumbnail image

Figure 1. Mouse PNPLA3 gene expression is regulated by nutritional status and dysregulated in genetically obese and diabetic mice. (A) Total RNA was extracted from the livers of C57BL/6J (control) and genetically obese (ob/ob) mice (left panel), as well as from db/+ (control) and diabetic (db/db) mice (right panel). mRNA levels of PNPLA3 and SREBP-1c were quantified by real-time PCR, normalized to β-actin, and expressed relative to liver mRNA levels of control mice. (B) Eight-week-old male C57BL/6J mice were divided into three groups: fed ad libitum (Feeding), fasting for 48 hours (Fasting), and fasting for 48 hours followed by refeeding for 12 hours (Refeeding). Total RNA was extracted from livers, and PNPLA3, SREBP-1c, FAS, PNPLA2, and PNPLA5 mRNA levels were quantified by real-time PCR, normalized to β-actin, and expressed relative to liver mRNA levels of feeding mice. n = 4 per group. *P < 0.05, **P < 0.01.

Download figure to PowerPoint

Next, we studied whether the expression of PNPLA3 and the other two family members, PNPLA2 and PNPLA5, would be regulated in the mouse liver in response to nutritional status. Fasting led to a decrease in mRNA levels of PNPLA3, SREBP-1c, and fatty acid synthetase (FAS) in the liver, and refeeding restored these gene-expression levels. In sharp contrast, PNPLA2 and PNPLA5 exhibited the opposite expression pattern as PNPLA3 in the liver under identical nutritional conditions (Fig. 1B). These results imply that mouse PNPLA3 is more likely involved in the anabolic pathway (i.e., lipogenesis), but less likely involved in the lipolytic pathway.

Transactivation of the PNPLA3 Promoter by SREBP-1.

PNPLA3 exhibited a similar expression pattern as SREBP-1c in response to changes in nutritional status; therefore, we speculated that SREBP-1c may regulate PNPLA3 gene expression. A computer-based screen of the sequence from −992 bp (base pairs) to +145 bp, relative to the transcription start site of the mouse PNPLA3 gene, revealed two putative SREs, located from −65 to −57 and −51 to −44 bp, respectively (Fig. 2A). To investigate the mechanism of PNPLA3 activation, the promoter region of PNPLA3 (−992 to +145 bp) was then cloned and fused to a luciferase reporter gene (p-PNPLA3-992). The transfection of SREBP-1a or SREBP-1c expression plasmid into HepG2 cells caused an approximately three-fold activation of the p-PNPLA3-992 reporter gene (Fig. 2B). Next, a series of truncated segments of the promoter fused to the luciferase gene (p-PNPLA3-170 and p-PNPLA3-43) were transfected into HepG2 cells to map the cis-acting element that conferred the SREBP-1-dependent activation of the luciferase reporter gene. The activation remained unchanged when the promoter region was truncated to −170/+145 bp, which still contained the putative SREs. However, when the reporter gene was further truncated to −43/+145 bp (deleting both putative SREs), the activation was almost completely lost (Fig. 2B). When either putative SRE of the PNPLA3 promoter was mutated, the activation of the mutant reporter gene (p-PNPLA3-mut1 and p-PNPLA3-mut2) by SREBP-1 was modestly suppressed, but not completely abolished. However, when both SREs were mutated, this activation was completely lost (Fig. 2C). These results suggest that both SREs are functional SREBP-1 cis-elements, and that the mutation of either SRE does not completely block the SREBP-1–dependent activation of PNPLA3 gene transcription. We also observed that overexpression of SREBP-1a or SREBP-1c increased PNPLA3 and FAS (positive control) mRNA levels in HepG2 cells, C2C12 myoblasts, and primary cultured hepatocytes (Fig. 3).

thumbnail image

Figure 2. Functional analysis of the promoter region of PNPLA3 in HepG2 cells. (A) Analysis of mouse PNPLA3 promoter sequences revealed two potential sterol regulatory elements (SREs) (SRE1: from −65 to −57 bp; SRE2: from −51 to −44 bp relative to transcription start site). (B) 5′-deletion series of the mouse PNPLA3 promoter fused to the luciferase reporter gene (p-PNPLA3-992, p-PNPLA3-170, and p-PNPLA3-43) were cotransfected into HepG2 cells, together with pcDNA3.1 (control) or SREBP-1 expression plasmids. After 48 hours, cells were harvested, and relative luciferase activity (RLA) was corrected for Renilla luciferase activity and normalized to the control activity. (C) Reporter gene plasmids containing 1.2 kb of the wild-type PNPLA3 promoter (p-PNPLA3) or SRE mutants (p-PNPLA3-mut1, p-PNPLA3-mut2, and p-PNPLA3-mut1+2) were transfected into HepG2 cells, together with pcDNA3.1 (control) or SREBP-1 expression plasmids. After 48 hours, cells were harvested, and the RLA was corrected for Renilla luciferase activity and normalized to the control activity. Construction of reporter gene plasmids, transient transfection, and luciferase assays are described in Supporting Materials and Methods. *P < 0.05.

Download figure to PowerPoint

thumbnail image

Figure 3. Induction of PNPLA3 gene expression by SREBP-1. HepG2 cells (A) and C2C12 myoblasts (B) were transfected with pcDNA3.1 (control) or SREBP-1 expression plasmids. Primary cultured mouse hepatocytes (C) were infected with adenoviruses expressing GFP (Ad-GFP; control) or mature SREBP-1 (Ad-SREBP-1a and Ad-SREBP-1c), and, after 48 hours, cells were harvested for RNA extraction. PNPLA3 and FAS mRNA levels were quantified by real-time PCR, normalized to β-actin, and expressed relative to controls (pcDNA3.1 or Ad-GFP). *P < 0.05, **P < 0.01.

Download figure to PowerPoint

SREBP-1 Binds to SREs Identified in the Mouse PNPLA3 Gene Promoter In Vitro and In Vivo.

To determine whether SREBP-1 proteins physically bound to the PNPLA3 SREs, electrophoretic mobility shift assays (EMSAs) were performed using nuclear extracts and synthetic double-stranded oligonucleotides. The nuclear proteins extracted from HEK-293A cells reacted with the labeled oligonucleotide probes, resulting in shifted bands (Fig. 4A). The addition of a 200-fold excess of an unlabeled probe (i.e., cold probe) as a competitor abolished the formation of the protein-DNA complex. In contrast, the addition of a 200-fold excess of the mutant unlabeled competitor (i.e., mutated probe) did not affect the formation of this complex (Fig. 4A).

thumbnail image

Figure 4. SREBP-1 associates with the PNPLA3 promoter in vitro and in vivo. (A) HEK293A cells were transfected with SREBP-1a (left panel) and SREBP-1c (right panel) expression plasmids. Nuclear proteins were extracted, and EMSAs were performed as described in Supporting Materials and Methods. Biotin-labeled, double-stranded oligonucleotides were incubated with (lanes 2 and 6) or without nuclear extract containing SREBP-1 proteins (lanes 1 and 5). An approximately 200-fold molar excess of unlabeled wild-type (lanes 3 and 7) or mutant (lanes 4 and 8) probe was used as the competitor. DNA-protein complexes are indicated by arrows. NS, no specific shift. (B) Primary cultured mouse hepatocytes were infected with adenoviruses expressing SREBP-1a or 1c. ChIP assays were performed as described in Materials and Methods. Protein-DNA complexes were immunoprecipitated with control mouse IgG (control) or anti-SREBP-1 antibody. The resultant DNA was analyzed by PCR with primers to amplify the promoter region flanking the PNPLA3 SREs. Amplified DNA fragments are indicated by arrows. (C) Untreated primary cultured mouse hepatocytes were used to perform ChIP assays, as described in Panel B.

Download figure to PowerPoint

Additionally, chromatin immunoprecipitation (ChIP) assays were carried out to investigate whether SREBP-1 could be recruited to the PNPLA3 promoter in vivo. First, mouse primary hepatocytes were isolated and infected with adenoviruses expressing SREBP-1a or SREBP-1c. The PNPLA3 promoter fragment containing the SREs could be amplified from the precipitates obtained when using SREBP-1–specific antibody, but not when using normal mouse immunoglobulin G (IgG; negative control) in cells infected with Ad-SREBP-1a or Ad-SREBP-1c (Fig. 4B). We also obtained similar results from primary hepatocytes without treatment with any adenovirus (Fig. 4C). These results suggest that both exogenous and endogenous SREBP-1 proteins can bind to the PNPLA3 gene promoter in vivo.

Insulin and TO901317 Induce PNPLA3 Expression via SREBP-1c.

SREBP-1c is the dominant member of the SREBP protein family in the liver and is selectively stimulated by insulin in the lipid metabolism pathway.21, 22, 27 Incidentally, insulin induces PNPLA3 gene expression14, 18, 19; thus, we hypothesized that SREBP-1c would mediate the stimulatory effect of insulin on PNPLA3. To this end, primary hepatocytes prepared from wild-type and SREBP-1c-null mouse livers were treated with insulin or the LXR agonist, TO901317. Insulin and TO901317 significantly increased SREBP-1c and FAS levels in wild-type hepatocytes, which is consistent with previous findings (Fig. 5A,C).27 An apparent increase in the PNPLA3 mRNA level was also observed in wild-type hepatocytes after insulin or TO901317 treatment (Fig. 5B). In sharp contrast, insulin or TO901317 did not stimulate PNPLA3 or FAS mRNA levels in primary SREBP-1c knockout hepatocytes (Fig. 5B,C). These results clearly demonstrate that SREBP-1c mediates insulin and LXR agonist TO901317-dependent effect on PNPLA3 gene expression in vivo.

thumbnail image

Figure 5. SREBP-1c mediates insulin and LXR-induced PNPLA3 gene expression. Wild-type (WT) or SREBP-1c knockout (KO) primary hepatocytes were isolated and treated with dimethyl sulfoxide (DMSO;control), insulin (100 nM), or T0901317 (5 μM) for 66 hours. Total RNA was extracted, and mRNA levels of PNPLA3 (A), FAS (B), and SREBP-1c (C) were quantified by real-time PCR, normalized to β-actin, and expressed relative to the control cell (DMSO). *P < 0.05, **P < 0.01.

Download figure to PowerPoint

Mouse PNPLA3 Increases Triglyceride Accumulation and Partially Mediates SREBP-1c–Dependent Stimulation of Triglyceride Accumulation in Mouse Primary Hepatocytes.

We next studied whether PNPLA3 would influence triglyceride metabolism in hepatocytes. PNPLA3 was effectively expressed in primary hepatocytes by adenovirus (Fig. 6A), and overexpression of PNPLA3 significantly increased triglyceride accumulation in hepatocytes, as revealed by oil red O staining and triglyceride quantitation (Fig. 6C,D); however, the cholesterol level was not influenced (Fig. 6E). To further study the physiological role of the SREBP-1c-dependent activation of PNPLA3, a recombinant adenovirus, expressing an shRNA targeting PNPLA3, was generated. An efficient siRNA (Ad-siPNPLA3-2) for silencing PNPLA3 expression dramatically decreased endogenous PNPLA3 mRNA in hepatocytes by more than 80%, compared to the control scrambled siRNA (Ad-siPNPLA3-scr) (Fig. 6B; P < 0.01). Our results (Fig. 6C,D) are in accord with a previous study demonstrating that overexpression of mature SREBP-1c in hepatocytes induces lipid accumulation.23 Knockdown of endogenous PNPLA3 by Ad-siPNPLA3-2 in primary hepatocytes did not affect intracellular triglyceride content (data not shown). However, when hepatocytes were coinfected with Ad-SREBP-1c and Ad-siPNPLA3-2, specific knockdown of PNPLA3 suppressed the ability of SREBP-1c to stimulate lipid accumulation (Fig. 6C,D), suggesting that PNPLA3 functions as a downstream target gene of SREBP-1c to mediate the effect of SREBP-1c on lipid accumulation in hepatocytes. These data clearly demonstrate that PNPLA3 is involved in lipogenesis in hepatocytes.

thumbnail image

Figure 6. PNPLA3 influences triglyceride accumulation in mouse primary hepatocytes. (A) Primary mouse hepatocytes were infected with adenoviruses expressing GFP (Ad-GFP; control) or Flag-tagged PNPLA3 (Ad-PNPLA3); proteins were extracted, and Western blot analysis was performed using antibodies specific to Flag and β-actin. (B) Primary hepatocytes were isolated and infected with Ad-siPNPLA3-scr (control), Ad-siPNPLA3-1, or Ad-siPNPLA3-2, as indicated. Total RNA was extracted and PNPLA3 mRNA levels were measured by real-time PCR, normalized to β-actin, and expressed relative to control cells. (C) Primary mouse hepatocytes were infected with Ad-GFP (control), Ad-PNPLA3, or Ad-SREBP-1c in the presence or absence of Ad-siPNPLA3-scr or Ad-siPNPLA3-2, as indicated. After 48 hours, cells were stained with oil red O. (D, E) Primary hepatocytes were treated as described in (C), and cells were washed three times with phosphate-buffered saline before an enzymatic assay was performed to measure intracellular triglyceride (D) and cholesterol (E) contents. Triglyceride and cholesterol levels are shown as mg/g protein. *P < 0.05, **P < 0.01.

Download figure to PowerPoint

Overexpression of PNPLA3 Promotes VLDL Secretion and Impairs Glucose Tolerance.

To determine the in vivo function of PNPLA3 in the liver, adenovirus particles expressing PNPLA3 (Ad-PNPLA3) were injected intravenously into C57BL/6J and db/+ genetic background mice via the tail vein. Western blot results indicated that adenovirus-mediated Flag-tagged PNPLA3 was effectively expressed in the liver (Supporting Fig. 1A). Histological analysis of liver sections (hematoxylin and eosin staining and oil red O staining) indicated that there were no obvious morphological differences between Ad-PNPLA3–infected and Ad-GFP–infected liver of db/+ or C57BL/6J mice (Supporting Fig. 1B). In support of the staining results, overexpression of PNPLA3 in the liver did not influence triglyceride content in the liver of C57BL/6J or db/+ mice (Fig. 7A). However, we observed that overexpression of PNPLA3 led to a significant increase in serum triglyceride content in db/+ background mice. In C57BL/6J background mice, Ad-PNPLA3–infected mice showed a tendency toward a higher serum triglyceride content, compared to Ad-GFP–infected mice, although the difference was not statistically significant (Fig. 7A). Conversely, there was no significant difference in serum or hepatic total cholesterol and free fatty acids between Ad-PNPLA3–infected and Ad-GFP–infected mice (Supporting Tables 2 and 3). To characterize serum lipoprotein composition, FPLC analysis was performed. Elution profiles revealed that overexpression of PNPLA3 in the liver of db/+ and C57BL/6J mice led to an increase in triglyceride content in VLDL particles (Fig. 7B). We also measured hepatic VLDL secretion after injecting tyloxapol and observed that adenovirus-mediated overexpression of PNPLA3 in C57BL/6J and db/+ mouse liver increased lipoprotein production (Fig. 7C), which contributed to the increase in serum triglyceride content. Finally, we explored the hepatic expression of several genes involved in lipoprotein assembly and secretion in Ad-PNPLA3–infected and Ad-GFP–infected mice. Overexpression of PNPLA3 in mouse livers enhanced the expression of microsomal triglyceride transport protein (MTP), which is the rate-limiting enzyme in the VLDL assembly (Fig. 7D).

thumbnail image

Figure 7. Adenovirus-mediated overexpression of PNPLA3 in mouse liver increases serum triglyceride accumulation and impairs glucose tolerance. (A) C57BL/6J and db/+ mice were intravenously injected via the tail vein with Ad-GFP (control) or Ad-Flag-PNPLA3 (2 × 1010 viral particles). Serum (left panel) and hepatic (right panel) triglyceride levels were measured as described in Materials and Methods. (B) C57BL/6J and db/+ mice were treated as described in (A). Serum lipoproteins were separated by FPLC, as described in Materials and Methods. (C) C57BL/6J and db/+ mice were treated as described in (A). Hepatic VLDL secretion was measured as described in Materials and Methods. (D) C57BL/6J and db/+ mice were treated as described in (A). Total RNA was extracted and MTP mRNA levels were measured by real-time PCR, normalized to β-actin, and expressed relative to control (Ad-GFP). (E) C57BL/6J and db/+ mice were treated as described in (A). After 5 days, GTT was performed on these mice after 16-hour fasting using glucose given at 1 g/kg mouse body weight by IP injection. Plasma-glucose levels were measured as described in Materials and Methods. (F) C57BL/6J mice treated as described in (A). After 5 days, mice were fasted overnight and anesthetized with tribromoethanol, followed by injection of 5 U of insulin or saline (as a control) via the inferior vena cava. Five minutes later, animals were sacrificed and liver protein lysates were subjected to Western blot analysis. Results are means ± SE (n = 5 per group): *P < 0.05 versus Ad-GFP treated C57BL/6J or db/+ control mice, **P < 0.01 versus Ad-GFP treated db/+ control mice.

Download figure to PowerPoint

Previous reports demonstrated that a PNPLA3 variant (rs738409) is associated with insulin secretion, insulin sensitivity, and glucose homeostasis.19, 28 Thus, we studied whether the acute overexpression of PNPLA3 by adenovirus in the liver would affect glucose metabolism in vivo. Overexpression of PNPLA3 in the liver of C57BL/6J or db/+ mice did not change the fasting glucose level. However, in IP glucose tolerance tests (GTTs), Ad-PNPLA3–infected db/+ or C57BL/6J mice displayed significantly impaired glucose tolerance, compared with control Ad-GFP–infected mice, as shown by increased blood-glucose levels (Fig. 7E). Next, we studied whether alteration of PNPLA3 expression level in the mouse liver would affect the insulin-signaling pathway, leading to impaired glucose tolerance. Indeed, we observed that overexpression of PNPLA3 impaired insulin-induced Akt phosphorylation (pSer473), but without changes in total Akt level (Fig. 7F; Supporting Fig. 1C).

Knockdown of PNPLA3 in db/db Mouse Liver Improves Glucose Tolerance.

PNPLA3 gene expression in the liver is highly induced in ob/ob or db/db mice; therefore, we next studied whether the knockdown of PNPLA3 in the liver of these mice would affect systemic glucose or lipid metabolism. Ad-siPNPLA3-2 injected into ob/ob or db/db mice via the tail vein dramatically decreased the hepatic PNPLA3 mRNA level by more than 75%, compared to Ad-siPNPLA3-scr, indicating that Ad-siPNPLA3-2 can knock down the native PNPLA3in vivo (Fig. 8A; P < 0.05). Oil red O staining revealed that lipid droplets in hepatocytes from Ad-siPNPLA3-2–infected ob/ob and db/db mice were smaller, compared to control Ad-siPNPLA3-scr–infected mice (Supporting Fig. 2). Hepatic triglyceride content of Ad-siPNPLA3-2–infected ob/ob and db/db mice was lower, compared to control mice, although differences were not statistically significant (Fig. 8B). No significant differences in serum or liver total cholesterol or free fatty acid content were observed between Ad-siPNPLA3-2–infected and Ad-siPNPLA3-scr–infected mice (Supporting Tables 4 and 5). Furthermore, we observed that the knockdown of PNPLA3 in db/db mouse liver improved glucose tolerance, as revealed by GTT, whereas PNPLA3 knockdown in ob/ob mouse liver did not significantly influence glucose tolerance, indicating the difference between mice with different genetic backgrounds (Fig. 8C). Again, we measured Akt phosphorylation (pSer473) level in the livers of Ad-siPNPLA3-2–infected and Ad-siPNPLA3-scr–infected db/db mice. Knockdown of PNPLA3 in db/db mouse liver enhanced insulin-induced Akt activation (Fig. 8D).

thumbnail image

Figure 8. Adenovirus-mediated knockdown of PNPLA3 in mouse liver improves glucose tolerance. (A) ob/ob and db/db mice were intravenously injected via the tail vein with Ad-siPNPLA3-scr (control) or Ad-siPNPLA3-2 (6 × 1010 viral particles). After 7 days, total RNA was extracted from liver, and PNPLA3 mRNA levels were measured by real-time PCR, normalized to β-actin, and expressed relative to control mice. (B) ob/ob or db/db mice were treated as described in (A). Hepatic triglyceride levels were measured as described in Materials and Methods. (C) ob/ob or db/db mice were treated as described in (A). After 5 days, GTT was performed on these mice after 16-hour fasting using glucose given at 1 g/kg mouse body weight by IP injection. Plasma-glucose levels were measured as described in Materials and Methods. (D) db/db mice were treated as described in (A). After 5 days, mice were fasted overnight and anesthetized with tribromoethanol, followed by IP injection of 5 U of insulin or saline (as a control). Ten minutes later, animals were sacrificed and liver protein lysates were subjected to Western blot analysis. Results are means ± SE (n = 4 per group): *P < 0.05 versus Ad-siPNPLA3-scr-treated ob/ob or db/db control mice.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Despite the strong evidence linking PNPLA3 with multiple liver diseases,2-10 the precise physiological function of PNPLA3in vivo remains unclear. PNPLA3 and adipose triglyceride lipase (ATGL) belong to the PNPLA family of proteins and share high homology; however, the regulation of PNPLA3 is opposite that of ATGL, which is the primary lipase responsible for the initial step of triglyceride hydrolysis.29ATGL is up-regulated by fasting and glucocorticoids and decreased by feeding and insulin.14, 30, 31 Instead, regulation of PNPLA3 is reminiscent of that of genes involved in lipogenesis, such as FAS and SCD-1.12, 14 Knockdown of ATGL inhibits lipolysis, whereas PNPLA3 knockdown has little effect.14 Based on these facts, we speculated that PNPLA3 is involved in lipogenesis and is less likely involved in catabolic processes in vivo.

Our results clearly demonstrate that overexpression of mouse PNPLA3 promotes triglyceride accumulation in primary hepatocytes. Indeed, Lake et al. reported that the overexpression of mouse ATGL decreased intracellular triglyceride levels, whereas the overexpression of mouse PNPLA3 showed a trend toward increased triglyceride synthesis in HEK293 cells,31 which is consistent with our data. Although adenovirus-mediated overexpression of mouse PNPLA3 in the livers of C57BL/6J or db/+ mice did not significantly promote hepatic triglyceride content, serum triglyceride levels of Ad-PNPLA3–infected mice were increased. One possible explanation for this result is that PNPLA3-induced triglycerides in the liver may be secreted into serum through an unidentified molecular mechanism, thereby increasing serum triglyceride levels.

Interestingly, He et al. reported that overexpression of the PNPLA3 I148M isoform by adenovirus in mouse liver significantly increased hepatic triglyceride levels, whereas overexpression of wild-type human PNPLA3 did not change hepatic triglyceride content in mice. The investigators concluded that the I148M mutation inactivates triglyceride hydrolase activity of PNPLA3, resulting in hepatic triglyceride accumulation.32 However, this explanation, that PNPLA3 may function as triglyceride hydrolase in vivo, seems unlikely, because overexpression of wild-type PNPLA3 in mouse liver failed to lower hepatic triglyceride content (He et al.32 and our present data), and regulation of PNPLA3 in the liver and adipose tissue by lipogenic stimuli is most consistent with a role in lipogenesis. Basantani et al. also provided strong evidence against a critical role for PNPLA3 in murine triglyceride hydrolysis and concluded that PNPLA3 may directly or indirectly promote triglyceride synthesis, which is enhanced by the I148M mutation.33

Although Huang et al. reported that SREBP-1 regulated PNPLA3 gene expression, they mapped the SREBP-binding element to intron 1 of PNPLA3, located 1,051 bp downstream of the first exon.20 However, our data indicate that SREs in the PNPLA3 gene promoter are crucial for SREBP-1 regulation of PNPLA3, because mutation of SREs in the promoter region completely abolished the activation of PNPLA3 gene transcription by SREBP-1. Furthermore, EMSA and ChIP assays also demonstrated that SREBP-1 proteins bind to SREs in the promoter region of PNPLA3in vitro and in vivo. However, we cannot rule out the possibility that the SRE in intron 1 of PNPLA3 is also functional, although luciferase reporter assays were not performed in their study.20

Two independent studies demonstrated that PNPLA3 knockout in mice had no effect on body weight, adipose mass or development, insulin sensitivity, fatty liver development, or energy homeostasis.33, 34 Possible reasons for this lack of response are as follows. First, PNPLA5, which has similar enzymatic activity to PNPLA3 and is up-regulated in the adipose tissue of PNPLA3-null mice, may partially compensate for the PNPLA3 deficiency. Second, a species-specific difference in PNPLA3 gene function, expression, or regulation may exist; human and mouse PNPLA3 differ primarily at the C-termini. Third, mouse PNPLA3 is normally expressed at very low levels in mouse liver; thus, PNPLA3 deficiency in mice did not noticeably affect fatty liver development. However, this situation is completely different in humans, where PNPLA3 is mainly expressed in the liver.15

Although PNPLA3 deficiency had no significant effect on the overall phenotype, when these mice were fed with a high-fat diet for 15 weeks, blood-glucose levels during GTT in PNPLA3-deficient mice were modestly, but significantly, lower than that in wild-type mice.34 We showed that overexpression of PNPLA3 by adenovirus in the livers of C57BL/6J and db/+ mice inhibited the insulin-signaling pathway, subsequently resulting in impaired glucose tolerance. In contrast, acute PNPLA3 knockdown in db/db mouse liver activated the insulin-signaling pathway and improved glucose tolerance. These data indicate that PNPLA3 is also implicated in glucose homeostasis under certain metabolic conditions.

In summary, our study demonstrates that SREBP-1c mediates the insulin effect on PNPLA3 gene expression, and that PNPLA3 acts as a lipogenic gene to promote triglyceride synthesis in primary hepatocytes. In addition to this involvement in lipid metabolism, PNPLA3 affects systemic glucose homeostasis in vivo.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Dr. Hitoshi Shimano for providing SREBP-1a and SREBP-1c plasmids.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_24402_sm_suppinfo.doc74KSupporting Information
HEP_24402_sm_suppinfofig1.tif11671KSupporting Information Figure 1
HEP_24402_sm_suppinfofig2.tif11513KSupporting Information Figure 2

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.