Hepatocyte-specific Ptpn6 deletion promotes hepatic lipid accretion, but reduces NAFLD in diet-induced obesity: Potential role of PPARγ


  • Elaine Xu,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Marie-Pier Forest,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Michael Schwab,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Rita Kohen Avramoglu,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Emmanuelle St-Amand,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Annabelle Z. Caron,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Kerstin Bellmann,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Michaël Shum,

    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    Search for more papers by this author
  • Gregory Voisin,

    1. McGill University and Génome Québec Innovation Centre, Montréal, Québec, Canada
    Search for more papers by this author
  • Marilene Paquet,

    1. Département de Pathologie et de Microbiologie, Faculté de Médecine Vétérinaire, Université de Montréal, St-Hyacinthe, Québec, Canada
    Search for more papers by this author
  • Alain Montoudis,

    1. Department of Nutrition, Centre Hospitalier Universitaire Sainte-Justine, Université de Montréal, Montréal, Québec, Canada
    Search for more papers by this author
  • Emile Lévy,

    1. Department of Nutrition, Centre Hospitalier Universitaire Sainte-Justine, Université de Montréal, Montréal, Québec, Canada
    Search for more papers by this author
  • Katherine A. Siminovitch,

    1. Department of Medicine, University of Toronto, Mount Sinai Hospital Samuel Lunenfeld Research Institute, Toronto, ON, Canada
    Search for more papers by this author
  • Benjamin G. Neel,

    1. Campbell Family Cancer Research Institute, Ontario Cancer Institute, Princess Margaret Hospital and Department of Medical Biophysics, University of Toronto, Ontario, Canada
    Search for more papers by this author
  • Nicole Beauchemin,

    1. Goodman Cancer Research Centre and Departments of Biochemistry, Medicine and Oncology, McGill University, Montréal, Québec, Canada
    Search for more papers by this author
  • André Marette

    Corresponding author
    1. Department of Medicine, Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Ste-Foy, Québec, Canada
    • Address reprint requests to: André Marette, Ph.D., Cardiology Axis of the Institut Universitaire de Cardiologie et de Pneumologie de Québec (Hôpital Laval), Québec, Canada, G1V 4G5. E-mail: andre.marette@criucpq.ulaval.ca; fax: 418-656-4749.

    Search for more papers by this author

  • Potential conflict of interest: Dr. Neel consults for and owns stock in Koltan.

  • Funded by grants from the Canadian Institutes of Health Research (CIHR, FRN# 82817) to A.M., N.B. and K.S. and grant R37 CA49152 to B.G.N. A.M. holds a CIHR/Pfizer Research Chair in the pathogenesis of insulin resistance and cardiovascular diseases. K.S. is supported by a Canada Research Chair and the Sherman Family Chair in Genomic Medicine. B.G.N. is a Canada Research Chair and is also partially supported by the Ontario Ministry of Health and Long-Term Care and the Princess Margaret Hospital Foundation. E.X. received the Canadian Diabetes Association (CDA) doctoral student award and a studentship from the CIHR training program in obesity at Laval University.


Hepatocyte-specific Shp1 knockout mice (Ptpn6H-KO) are protected from hepatic insulin resistance evoked by high-fat diet (HFD) feeding for 8 weeks. Unexpectedly, we report herein that Ptpn6H-KO mice fed an HFD for up to 16 weeks are still protected from insulin resistance, but are more prone to hepatic steatosis, as compared with their HFD-fed Ptpn6f/f counterparts. The livers from HFD-fed Ptpn6H-KO mice displayed 1) augmented lipogenesis, marked by increased expression of several hepatic genes involved in fatty acid biosynthesis, 2) elevated postprandial fatty acid uptake, and 3) significantly reduced lipid export with enhanced degradation of apolipoprotein B (ApoB). Despite more extensive hepatic steatosis, the inflammatory profile of the HFD-fed Ptpn6H-KO liver was similar (8 weeks) or even improved (16 weeks) as compared to their HFD-fed Ptpn6f/f littermates, along with reduced hepatocellular damage as revealed by serum levels of hepatic enzymes. Interestingly, comparative microarray analysis revealed a significant up-regulation of peroxisome proliferator-activated receptor gamma (PPARγ) gene expression, confirmed by quantitative polymerase chain reaction. Elevated PPARγ nuclear activity also was observed and found to be directly regulated by Shp1 in a cell-autonomous manner. Conclusion: These findings highlight a novel role for hepatocyte Shp1 in the regulation of PPARγ and hepatic lipid metabolism. Shp1 deficiency prevents the development of severe hepatic inflammation and hepatocellular damage in steatotic livers, presenting hepatocyte Shp1 as a potential novel mediator of nonalcoholic fatty liver diseases in obesity. (Hepatology 2014;59:1803–1815)


acetyl-CoA carboxylase


Animal Care and Handling Committee


alkaline phosphatase


alanine transaminase


apolipoprotein B


aspartate transaminase


Canadian Council on Animal Care


diacylglycerol acyltransferase


dominant negative


fatty acid binding protein


fatty acid synthase


false discovery rate


farnesoid X receptors


extensor digitorum longus


high density lipoprotein


hepatocellular carcinoma


high-fat diet


3-hydroxy-3-methyl-glutaryl-CoA reductase


interferon gamma


inhibitor of kappa B




intraperitoneal glucose tolerance test


intraperitoneal pyruvate tolerance test


c-Jun N-terminal kinase


liver X receptor


microsomal triglyceride transfer protein


nonalcoholic fatty liver disease


nonalcoholic steatohepatitis


nonesterified fatty acids


nuclear factor kappa-light-chain-enhancer of activated B cells


oral lipid tolerance tests






PE N-methyltransferase


peroxisome proliferator-activated receptor gamma


PPAR response element


protein tyrosine phosphatase


regulated upon activation, normal T-cell expressed and secreted


retinoid X receptor


stearoyl CoA desaturase


standard diet


Src homology 2 domain-containing PTP 1


Src homology 2 domain-containing PTP 2


small hairpin RNA


sterol-regulated element binding protein 1


signal transducer and activator of transcription


type 2 diabetes




tumor necrotic factor alpha


very low density lipoprotein

Nonalcoholic fatty liver diseases (NAFLD) and type 2 diabetes (T2D) are prevalent metabolic disorders associated with obesity. However, despite the usual simultaneous occurrence of hepatic insulin resistance and steatosis in obesity, their relationship has been a constant debate.[1, 2] Indeed, whereas certain interventions alleviating insulin resistance also reduce hepatic steatosis, liver lipid accretion does not always cause insulin resistance.[3-6]

Protein tyrosine phosphatases (PTPs) are key modulators of insulin action and glucose metabolism in liver and peripheral tissues,[7] but their role in lipid metabolism and hepatic steatosis regulation remains poorly understood. Mice lacking PTP-1B specifically in liver are protected from diet-induced insulin resistance[8, 9] and show decreased hepatic triglycerides on a 5-week high-fat diet (HFD), which was not observed after 18-weeks of HFD feeding (10). More recently, it has been demonstrated that mice with liver-specific knockout (KO) of Src homology 2 domain-containing PTP Shp2 (LSHKO) are protected from insulin resistance, and exhibit reduced liver steatosis in association with blunted hepatic inflammation.[10] The mechanisms underlying hepatic lipid metabolism modulation by PTPs remain poorly understood.

We previously showed, using viable motheaten (Ptpn6me-v/me-v or mev) mice, expressing a catalytically impaired Shp1, a key role for this PTP in modulating glucose metabolism in insulin target tissues.[11-13] Interference of Shp1 activity by adenoviral expression of a dominant-negative (DN) Shp1 mutant or small hairpin RNA-mediated gene silencing improved hepatic insulin signaling and glucose tolerance.[11] We also observed increased expression of Shp1 in the liver and other metabolic tissues of HFD-induced obese mice and found that Ptpn6H-KO mice were protected from developing hepatic insulin resistance and disturbed glucose homeostasis.[12]

These data implicate Shp1 as a cell-autonomous regulator of hepatocyte glucose homeostasis and insulin clearance, but its potential role in lipid metabolism has yet to be examined. The major goal of the present study was to determine the function of liver parenchymal Shp1 in the modulation of lipid metabolism and obesity-linked NAFLD. We show that despite marked improvement in hepatic insulin action compared with their wild-type Ptpn6f/f counterparts, HFD-fed Ptpn6H-KO mice exhibited increased hepatic steatosis due to enhanced lipogenesis and fatty acid transport in combination with reduced lipid export. Remarkably, steatotic livers of obese Ptpn6H-KO mice were protected from inflammation and showed less hepatic damage than that seen in their obese Ptpn6f/f counterparts. We further demonstrate that up-regulation of hepatocyte peroxisome proliferator-activated receptor gamma (PPARγ) underlies this protection and that Shp1 regulates PPARγ activity in vitro. These findings unravel a novel role for hepatocyte Shp1 in the regulation of PPARγ and hepatic lipid metabolism, suggesting that Shp1 may participate in the development of obesity-linked NAFLD.

Materials and Methods

Animals and Genotyping

Mice (C57BL/6J background) were housed under controlled temperature (23°C) and a 12-hour light/dark cycle with water and food ad libitum. Two-month-old male Ptpn6H-KO mice[12] were kept either on the standard diet (SD) or transferred to an HFD (54.8% fat by kcal, total 4.8 kcal/g of diet, TD93075, Harlan Teklad) for 8 or 16 weeks. General phenotyping was performed as described in the Supporting Materials and Methods. All studies were approved by the Animal Care and Handling Committee of Laval University under Canadian Council on Animal Care standards.

Hepatic Lipid Extraction and Histopathological Analyses

Total lipid was isolated from ∼0.1 g of frozen liver using chloroform-methanol extraction.[14] Hepatic cholesterol and triglyceride (TG) content was enzymatically determined (Randox Lab kit) from reconstituted lipid extract.[15] Histopathological analyses were performed on hematoxylin and eosin (H&E) liver sections to determine degree of hepatic steatosis and inflammatory grade (see Supporting Materials and Methods), presented as number of animals with the same grade.[16]

Measurement of Hepatic Fatty Acids and Phospholipids

Hepatic content of fatty acids was assessed by gas-liquid chromatography.[17, 18] Tissue phospholipids were separated by thin-layer chromatography using a chloroform:methanol:water:acetic acid (65:25:4:1 by volume) solvent, quantified by phosphorus analysis using the Bartlett method,[19] and referred to as total phospholipids.

Luminex Hepatic Cytokine Detection

Fifty mg of frozen liver was homogenized in 0.8 mL lysis buffer (1% NP-40 in 1× PBS with 1 mM PMSF and 1× dilution of Sigma protease inhibitors, pH 7.4), and 200 μg of proteins were assayed for cytokine detection (Bio-Plex cytokine assay kit) using a Luminex reader (Bio-Rad Lab, Hercules, CA).

Comparative Microarray Analysis

Agilent SurePrint G3 mouse GE 8×60K microarray chips (>60,000 probes interrogating 40,000 transcripts) were used to compare gene expression patterns between 6-hour fasted and 8-week HFD-fed Ptpn6f/f and Ptpn6H-KO liver complementary DNA (cDNA) reverse transcripts. Microarray results were analyzed using FlexArray software. Gene expression was considered significantly altered with P < 0.05 by EB (Wright and Simon) bioinformatics statistical analysis and a cutoff line of 1.1-fold changes. Elimination by false discovery rate (FDR) score was also employed. Further analyses were carried out using the Ingenuity Pathway Analysis (IPA) program. Validation of microarray results was performed by quantitative polymerase chain reaction (Q-PCR).

Isolation of Primary Hepatocytes and Ex Vivo Assays

Livers were perfused in anesthetized mice and primary hepatic cells were isolated, cultured as described,[20] and used for ex vivo assays (see Supporting Materials and Methods).

Statistical Analysis

Statistical analyses were performed by either two-tailed Student t test or two-way analysis of variance (ANOVA) using JMP-8 software (Cary, NC). P values were considered significant if less than 0.05. Standard errors of the mean (SEM) are represented in the graph.


Ptpn6H-KO Mice Maintain Hepatic Insulin Sensitivity Even After Prolonged HFD Feeding

Ptpn6f/f and Ptpn6H-KO mice were randomly distributed for 8 or 16 weeks SD or HFD diet studies. As Ptpn6H-KO mice exhibited protection against HFD-induced hepatic insulin resistance after 8 weeks,[12] we also examined their glucose homeostasis after 16 weeks on the HFD. Fasting glycemia and plasma insulin levels of Ptpn6H-KO mice remained significantly lower than that of Ptpn6f/f animals on both diets for 16 weeks, although still exhibiting the HFD-induced increases in fasting glycemia (Fig. 1A). SD-fed Ptpn6H-KO mice showed a modest but significant improvement in glucose tolerance and glucose-stimulated insulin secretion, assessed by plasma insulin and C-peptide level (Fig. 1B), as described.[12] Although HFD-induced glucose homeostasis was not improved in Ptpn6H-KO mice as compared with Ptpn6f/f controls, the knockout animals were significantly less hyperinsulinemic during the course of intraperitoneal glucose tolerance test (IPGTT), which may be partly explained by a tendency towards reduced insulin secretion as reflected by their C-peptide levels (Fig. 1B). Hepatic gluconeogenesis, assessed by intraperitoneal pyruvate tolerance test (IPPTT), was significantly less in Ptpn6H-KO mice, compared with Ptpn6f/f controls on HFD but not on SD (Fig. 1C). To evaluate potential changes in peripheral insulin sensitivity, insulin-stimulated glucose uptake was measured in isolated extensor digitorum longus muscles. Besides the expected insulin resistance of muscles from HFD-fed animals, no significant differences were observed (Fig. 1D), confirming the selective hepatic phenotype of the obese Ptpn6H-KO mice for glucose metabolism, even after prolonged HFD feeding.

Figure 1.

Glucose homeostasis of Ptpn6f/f and Ptpn6H-KO mice following prolonged diet studies. (A) Fasting glycemia and plasma insulin levels from Ptpn6f/f and Ptpn6H-KO mice at the end of 16 weeks SD or HFD feeding (n = 6-12; ^P < 0.05 diet effect, *P < 0.05 genotype difference). (B) IPGTT blood glucose, plasma insulin, and C-peptide (n = 10-17) and (C) IPPTT blood glucose of Ptpn6f/f and Ptpn6H-KO mice on SD or HFD (n = 10-17); *P < 0.05 genotype difference. (D) Ex vivo glucose (2DG) uptake by skeletal muscle EDL isolated from Ptpn6f/f and Ptpn6H-KO mice fed HFD for 16 weeks (n = 7-12; ^P < 0.05 diet effect).

Augmented Development of Hepatic Steatosis in Ptpn6H-KO Mice

Ptpn6f/f and Ptpn6H-KO mice on HFD for 8 weeks displayed similar energy intake, body weight gain, and fat mass.[12] Comparably increased body weight and fat mass were also observed after more prolonged (16 weeks) HFD consumption (Supporting Table S2). Both Ptpn6H-KO and Ptpn6f/f mice developed hepatic steatosis upon HFD feeding after 8 and 16 weeks (Fig. 2). Unexpectedly, the Ptpn6H-KO livers were more steatotic, as revealed by increased liver weight and hepatic TG levels, with comparable hepatic cholesterol content (Fig. 2A,B). This difference was also evident on H&E-stained liver sections, which showed more lipid droplets in HFD-fed Ptpn6H-KO mice (Fig. 2C,D). Pathological grading also revealed more severe hepatic steatosis in Ptpn6H-KO mice compared with the control Ptpn6f/f mice (Fig. 2E,F).

Figure 2.

Enhanced development of hepatic steatosis in Ptpn6H-KO mice. (A,B) Liver weight, hepatic TG, and total cholesterol in Ptpn6f/f and Ptpn6H-KO mice fed SD or HFD for 8 or 16 weeks after 6 hours fasting (n = 12-13; ^P < 0.05 diet effect, *P < 0.05 genotype difference). (C,D) Representative H&E staining and (E,F) histological analysis of steatosis in five different grades of severity by microscopic assessment of liver sections from the same mice and presented as occurrence of arbitrary sample numbers (n = 12-13); grading system: 1 = minimal, 2 = slight or few, 3 = moderate or several, 4 = marked or many, 5 = severe.

Enhanced Lipogenesis, Increased Fatty Acid Uptake, and Reduced Lipid Export in the Ptpn6H-KO Livers in HFD-Induced Obesity

Exploring the potential mechanisms underlying augmented hepatic steatosis in Ptpn6H-KO mice, we found that DNA-binding activity of sterol-regulated element binding protein 1 (SREBP-1) (and Srebf1 expression, see Fig. 7B) was up-regulated in livers from Ptpn6H-KO compared with Ptpn6f/f mice on HFD for 8 weeks. By contrast, SREBP-2 activity was not affected (Fig. 3A). The levels of the lipogenic enzymes, fatty acid synthase (FAS) and acetyl-CoA carboxylase (ACC) (Fig. 3B), as well as other lipogenic enzymes, such as stearoyl CoA desaturase (SCD) 1 and 3 (data not shown), also were higher in Ptpn6H-KO mice on HFD, due to increased gene expression (Fig. 7B). Consistent with the comparable nuclear SREBP-2 activity and hepatic cholesterol content in Ptpn6H-KO and Ptpn6f/f mice, there was no difference in the levels of the rate-limiting cholesterogenic enzyme 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMGCR) in these mice (Fig. 3A,B). We also assessed ChREBP activity since this lipogenic transcription factor has previously been shown to dissociate hepatic steatosis from insulin resistance in both mouse and human subjects.[21] The hepatic activity of ChREBP was significantly higher in Ptpn6H-KO livers as compared to Ptpn6f/f livers from 16-week HFD-fed mice (Fig. S2).

Figure 3.

Augmented hepatic lipogenesis in HFD-fed Ptpn6H-KO mice. (A) SREBP-1 and -2 nuclear DNA binding activity, expressed as relative units of absorbance normalized by amount of nuclear protein (A/nP); n = 12-13. (B) Hepatic enzymes, FAS, ACC, and HMGCR from liver lysates of 8-week SD- or HFD-fed Ptpn6f/f and Ptpn6H-KO mice fasted 6 hours were detected using immunoblotting and quantified with eEF2 as loading control (n = 14-20). ^P < 0.05 diet effect, *P < 0.05 genotype difference.

To monitor postprandial hepatic fatty acid uptake, we conducted fasting-refeeding studies. These studies revealed a significant reduction in plasma level of nonesterified fatty acids (NEFA) and a drastic surge of total hepatic fatty acid content in HFD-fed Ptpn6H-KO mice during refeeding (Fig. 4A), suggesting increased uptake of postprandial fatty acids into Ptpn6H-KO livers. Accordingly, expression of two main hepatic fatty acid transporters, CD36 and fatty acid binding protein 5 (FABP5), was increased in Ptpn6H-KO livers (Fig. 4B). Furthermore, oleic acid uptake was enhanced in primary hepatocytes from 8-week HFD-fed Ptpn6H-KO mice, compared with control Ptpn6f/f hepatocytes (Fig. 4C). These findings are consistent with greater internalization of lipids from the circulation of HFD-fed Ptpn6H-KO mice during the postprandial state. Consequently, increased accumulation of several hepatic fatty acid species, including oleic acid, was detected in Shp1-deficient livers (Tables S3 and S4). Levels of palmitoleic acid, a lipid hormone (lipokine) mainly secreted by adipose tissue affecting systemic metabolism during fasting state,[22] were increased in Ptpn6H-KO mice (Table S3).

Figure 4.

Increased hepatic fatty acid uptake in HFD-fed Ptpn6H-KO mice. (A) Plasma NEFA level and hepatic fatty acid content of 8-week SD- and HFD-fed Ptpn6f/f and Ptpn6H-KO mice fasted 6 hours and refed 2 hours (n = 7-13); hepatic fatty acid species. (B) Q-PCR validation of Cd36 and Fabp5 gene expression in livers of 8-week HFD-fed Ptpn6f/f and Ptpn6H-KO mice fasted 6 hours (n = 12-13). (C) Ex vivo oleic acid uptake by primary hepatocytes isolated from Ptpn6f/f and Ptpn6H-KO mice fed HFD for 8 weeks, calculated from intracellular content of [14C]-oleic acid isotope count. *P < 0.05 genotype difference.

Although fasting plasma TG levels were comparable between Ptpn6H-KO mice and littermate controls, circulating TG levels during refeeding were blunted in HFD-fed Ptpn6H-KO mice (Fig. 5A). Under the same conditions, cholesterol levels following refeeding were not affected by hepatocyte Shp1 deficiency. We therefore investigated the rate of TG clearance and postprandial hepatic TG export by way of lipoproteins. Oral lipid tolerance tests demonstrated similar TG clearance between Ptpn6H-KO and Ptpn6f/f mice on SD and HFD (Fig. 5B), suggesting that the detected differences in postprandial TG levels might be linked to decreased hepatic lipoprotein secretion. Indeed, HFD-fed Ptpn6H-KO mice displayed lower plasma levels of ApoB100 and ApoB48 in the postprandial state 2 hours after refeeding (Fig. 5C) which, together with unchanged TG clearance, is indicative of reduced hepatic output of ApoB-containing very low density lipoprotein (VLDL) particles. Conversely, plasma ApoA1 levels were not affected (Fig. 5C), suggesting that release of high density lipoprotein (HDL) was normal in Ptpn6H-KO mice. Decreased hepatic ApoB secretion was verified by pulse-chase experiments performed in isolated primary Ptpn6H-KO hepatocytes, which not only contained less but also secreted less newly synthesized ApoB (Fig. 5D,E), due to markedly accelerated ApoB degradation compared with Ptpn6f/f cells (Fig. 5F). Accordingly, diminished levels of hepatic cellular phospholipids (PL) also were observed in Ptpn6H-KO mice (Fig. 5G), accompanied by decreased gene expression of important enzymes for VLDL-assembly (Fig. 5H), namely, diacylglycerol acyltransferase 1 (DGAT1) and DGAT2, as well as a similar tendency in the gene expression of microsomal triglyceride transfer protein (MTTP). Lower expression of the hepatic lipase gene (Lipc) might indicate reduced extrahepatic TG processing between lipoprotein particles, possibly due to a general decrease in TG export. The expression of a key enzyme in the conversion of phosphatidylethanolamine (PE) to phosphatidylcholine (PC), the N-methyltransferase PEMT, which is required for the secretion of VLDL,[23] was increased significantly (Fig. 5H), perhaps in compensation for reduced VLDL release. This increase in PEMT expression did not alter the ratio of hepatic PC to PE, however (data not shown).

Figure 5.

Reduced hepatic lipid export in HFD-fed Ptpn6H-KO mice. (A) Plasma TG and cholesterol at the end of 8-week diet studies with 6 hours fast (n = 13-26/genotype and diet group) or 2 hours refed (n = 6-8; ^P < 0.05 diet effect, *P < 0.05 genotype difference). (B) Plasma TG levels in Ptpn6f/f and Ptpn6H-KO mice during OLTT up to 6 hours postoral administration of pure olive oil at 10 μL/g body weight of the animal (n = 18). (C) Immunoblot analysis of postprandial plasma apolipoproteins ApoB (100 and 48) and ApoA1 using commercial antibodies with albumin as loading control (n = 6-8). (D) Radioisotope exposure of [35S]-methionine-containing ApoB sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) post-IP; (E) densitometry normalized by total protein content relative to basal amount in cell or media at time 0 hours, and (F) calculated ratio of ApoB degradation to secretion for both ApoB100 and 48 during 2 hours from ApoB pulse-chase experiment (n = 4; *P < 0.05 genotype difference). (G) Total hepatic phospholipid (PL) content (n = 12-13) and (H) Q-PCR analysis of hepatic gene expression in mouse liver with Actb as sample control (n = 11-12); *P < 0.05 genotype difference.

Uncoupling of Hepatic Steatosis and Liver Inflammation in Ptpn6H-KO Mice

The dichotomy between improved liver insulin resistance and increased hepatic steatosis in Ptpn6H-KO mice prompted us to assess hepatic inflammation, a well-known driver of the progression from simple hepatic steatosis to nonalcoholic steatohepatitis (NASH) in obesity. We found similar hepatic inflammatory profiles in 8-week HFD-fed Ptpn6H-KO and Ptpn6f/f mice, as revealed by measurements of multiple inflammatory cytokines and chemokines (Fig. 6A). We further assessed hepatic inflammation by evaluating the activation of c-Jun N-terminal kinase (JNK) and nuclear factor kappa B (NFκB) pathways in liver lysates. Hepatic JNK levels and JNK2 phosphorylation were increased by 8-week HFD feeding, but there were no significant differences in these parameters between Ptpn6f/f and Ptpn6H-KO mice (Fig. 6B). NFκB signaling was increased slightly in SD-fed Ptpn6H-KO liver, as indicated by decreased inhibitor of kappa B alpha (IκBα) and increased pSer32 IκBα levels. No additional differences in the NFκB pathway were detected in HFD-fed animals of either genotype (Fig. 6C). Most important, progression of hepatic inflammation after 16 weeks of HFD was reduced in Ptpn6H-KO animals, as shown by lessened hepatic content of interleukin (IL)−3, IL-6, regulated upon activation, normal T-cell expressed and secreted (RANTES), and tumor necrosis factor alpha (TNFα) compared with littermate controls (Fig. 6A). This was associated with lower amount of F4/80-positive cells in liver sections from Ptpn6H-KO mice as compared to Ptpn6f/f liver sections (Fig. 6D). Moreover, there was a reduction in the level of hepatocellular damage typically associated with hepatic steatosis and inflammation, as indicated by hepatic aspartate aminotransferase (AST) and alanine aminotransferase (ALT) serum levels in HFD-fed Ptpn6H-KO animals after 8 weeks and even 16 weeks (Fig. 6E). A moderate but consistent rise in serum ALK levels was observed in the HFD-fed Ptpn6H-KO mice with no alterations in other major hepatic factors (Tables S5 and S6). In short, these data demonstrate that obese mice lacking hepatocyte Shp1 acquired more hepatic steatosis but, at the same time, were protected from the development of more severe NAFLD, as revealed by blunted hepatic inflammation and concurrent lack of hepatocellular damage despite the continuous lipid accumulation upon prolonged HFD feeding.

Figure 6.

Diet-induced hepatic inflammation in Ptpn6H-KO mice is similar to or improved over that in Ptpn6f/f controls. (A) Measurement of hepatic IL-1α, IL-2, IL-3, IL-6, IFNγ, RANTES, and TNFα by Luminex (BioRad) in 200 μg of proteins extracted from frozen liver of Ptpn6f/f and Ptpn6H-KO mice fed HFD for 8 and 16 weeks (n = 12-13; *P < 0.05 genotype difference). (B,C) Immunoblot analysis and quantification from the same liver samples: (B) phosphorylation (pT183/pY185) and total JNK isoforms 2 and 1 with β-actin as loading control (n = 5-13); (C) phosphorylation (pS32 IκBα; pS536 NFκB p65) and total IκBα and NFκB p65, dotted lines on blot separate noncontiguous sections of the same gel with eEF2 as loading control (n = 9-13); ^P < 0.05 diet effect, *P < 0.05 genotype difference. (D) Histological F4/80 staining in liver sections from Ptpn6f/f and Ptpn6H-KO mice fed HFD for 16 weeks (n = 3-5; *P < 0.05 genotype difference). (E) Serum hepatic enzyme markers AST and ALT levels in Ptpn6f/f and Ptpn6H-KO mice fed SD or HFD for 8 and 16 weeks (n = 7-13; ^P < 0.05 diet effect, *P < 0.05 genotype difference).

Figure 7.

Pparγ-expression, PPARγ activity, and transcription of many other genes involved in lipid metabolism are regulated by Shp1. (A) Scheme of microarray analysis methodology and result highlights for lipid metabolism comparing hepatic mRNA expression levels between Ptpn6f/f and Ptpn6H-KO mice fed HFD for 8 weeks. (B) A list of important genes in lipid metabolism with up-regulated (gray) or down-regulated (white) expression detected by the microarray analysis. (C) Validation of Pparγ expression and PPARγ nuclear activity by Q-PCR and DNA binding activity assay, respectively; mouse Actb was used as the transcript level control for Q-PCR, and activity was expressed as relative unit of absorbance normalized by amount of nuclear protein (A/nP); n = 12-13; ^P < 0.05 diet effect, *P < 0.05 genotype difference. (D) Luciferase assays of Flp-In T-Rex 293-cells stably transfected with an empty control or an inducible Shp1-overexpression construct, and cotransfected with pGL3-PPRE (PPRE-luc) and an empty vector or pSV40-PPARγ2. Cells were treated 24 hours before harvesting with 1 μg/mL tetracycline to induce Shp1 expression and during the last 16 hours with 10 μM rosiglitazone where indicated (representative graph from n = 3).

Cell-Autonomous Up-Regulation of PPARγ in Ptpn6H-KO Mice

To further explore the mechanisms underlying the surprising co-occurrence of hepatic lipid accretion and hepatocellular protection in Ptpn6H-KO mice, we analyzed global gene expression in livers of Ptpn6H-KO and Ptpn6f/f livers from mice fed HFD for 8 weeks by using Agilent SurePrint G3 microarrays. Pathway analyses by IPA identified “lipid metabolism” as the most significantly altered pathway in Ptpn6H-KO livers, with 216 associated genes showing altered expression (Fig. S1). Of these, 18 genes regulated by the nuclear receptors LXR/RXR, 26 genes by FXR/RXR, 23 genes involved in fatty acid biosynthesis, and 19 genes specifically marking the development of hepatic steatosis were differentially expressed in Ptpn6H-KO livers (Fig. 7A). Figure 7B shows 27 well-known lipid metabolism genes whose expression was significantly down-regulated (9 genes) or up-regulated (18 genes). The expression of 5 down-regulated and 15 up-regulated genes were validated by Q-PCR to show equal or even more differences (data not shown). The significant increase in Pparγ expression was particularly interesting, as PPARγ is a key mediator of lipid storage and adipogenesis[24] and plays an important role in obesity-induced hepatic steatosis.[25] Since PPARγ also negatively modulates inflammation,[25, 26] we elected to further explore its role in the hepatoprotective phenotype exhibited by Ptpn6H-KO mice.

Both the Pparγ expression (detected by Q-PCR) and PPARγ nuclear activity (measured by DNA-binding assay) were significantly up-regulated in Ptpn6H-KO livers compared with Ptpn6f/f livers obtained from mice fed either SD or HFD for 8 weeks (Fig. 7C). To determine whether Shp1 directly modulates PPARγ, the transcriptional activity of PPARγ was examined with a luciferase reporter assay in Flp-In T-Rex 293-cells carrying either a Shp1-expressing construct or an empty construct as a control (Fig. 7D). These cells were cotransfected with a PPAR response element (PPRE)-luciferase reporter construct and either an empty vector or a plasmid expressing PPARγ2 transcribed from the SV40 promoter. Exogenous expression of PPARγ2 led to a strong increase in luciferase activity in control cells, which was more pronounced in the presence of the PPARγ agonist rosiglitazone. Shp1 expression reversed the PPARγ-dependent activation. Knockdown of Shp1 by siRNA in HEK293-cells cotransfected with the above mentioned PPRE-luciferase reporter and the PPARg2-expressing plasmid increased the PPARg2-mediated luciferase activity both with and without Rosiglitazone treatment (Supporting Fig. 3A,B). Together, these results confirm that Shp1 is a novel, hepatic and cell-autonomous PPARγ regulator.


We have previously shown that Shp1 is increased in liver and in isolated hepatocytes from 8-week HFD-fed obese mice, and that hepatocyte-specific Shp1-deficient mice are protected from obesity-linked hepatic insulin resistance.[12] In the current report, we confirm that Ptpn6H-KO mice continue to display improved glucose homeostasis during prolonged HFD feeding. The unexpected finding of the current study is that Ptpn6H-KO mice are not protected from hepatic steatosis. On the contrary, hepatocyte Shp1 deficiency actually enhances liver lipid accretion, concomitant with improved hepatic glucose metabolism. We suspect that the enhanced insulin action on lipogenic pathways accounts for this unexpected increase of hepatic steatosis in Ptpn6H-KO mice on HFD, as hyperinsulinemia persists in these animals, although lower than in control mice.[12] Consistent with this, the expression of many lipogenic modulators, such as SREBP-1, ACC, and FAS, are up-regulated in Ptpn6H-KO liver, possibly as a consequence of increased insulin receptor signaling and Ceacam1 activation,[12] both of which promote lipogenesis.[27-29] Augmented fatty acid uptake, due to fatty acid transporter CD36 and FABP5 increased expression, could contribute to the augmented lipid accretion in Ptpn6H-KO liver; enhanced insulin action could also account for up-regulation of these transporters.[27, 28, 30] Reduced liver output of TG-rich lipoproteins is suggested by decreased plasma ApoB100 and ApoB48 as well as lower postprandial plasma TG levels in the face of unchanged TG clearance. Presumably, enhanced intracellular insulin-stimulated ApoB degradation underlies this reduction in lipoprotein output, as hepatic ApoB expression and postprandial plasma TG clearance remain unchanged, although this remains to be confirmed by direct in vivo measurements of TG and ApoB secretion rates.

The clear uncoupling between hepatic steatosis and insulin resistance in the Ptpn6H-KO mouse was unexpected, because liver insulin resistance is thought to be a key determinant of NAFLD pathogenesis.[31-34] We previously reported the association of hepatic steatosis with insulin resistance in Ceacam1-deficient livers.[15] Conversely, Ptpn6H-KO mice showed heightened insulin signaling and augmented Ceacam1 activation,[12] yet still exhibit hepatic steatosis on HFD, even more than their insulin-resistant Ptpn6f/f littermates. Hepatic steatosis in these two animal models strongly suggests that hepatic insulin resistance is not absolutely required for the development of hepatic steatosis in diet-induced obesity. The higher ChREBP activity detected in the liver from HFD-fed Ptpn6H-KO mice also supports the dissociation between hepatic steatosis and insulin resistance, as mice overexpressing ChREBP remained insulin-sensitive despite augmented hepatic steatosis.[21] This is consistent with the finding that mice with hepatic-specific deletion of the insulin receptor (LIRKO) do not display hepatic steatosis despite their severe hepatic insulin resistance.[35] This might also explain why subjects with insulin receptor mutations do not develop hepatic steatosis despite being overtly diabetic, or conversely why human subjects with familial hypobetalipoproteinemia present normal insulin sensitivity despite their excessive hepatic TG storage.[36, 37] Although LIRKO mice and patients with insulin receptor mutations are hyperinsulinemic, they fail to develop hepatic steatosis, probably because they lack effective hepatic insulin receptor signaling, thus preventing insulin-promoted FAS-dependent lipid synthesis or ApoB degradation. Notably, blocking ApoB-associated VLDL secretion in animal models or by inactivating mutations in the human APOB gene result in hepatic steatosis without hepatic insulin resistance.[5, 38] Alternatively, ApoB accumulation due to decreased degradation resulting from lipid-induced endoplasmic reticulum (ER) stress might cause hepatic insulin resistance.[39]

Besides insulin signaling, other insulin-independent factors could contribute to the uncoupling of insulin resistance and hepatic steatosis. Our finding of decreased gene expression of DGAT2 is of particular interest, as this key enzyme in TG synthesis promotes hepatic steatosis without concomitant insulin resistance,[40, 41] although these findings were not confirmed in a subsequent study.[42] In the Ptpn6H-KO steatotic livers, DGAT2 could have decreased in an attempt to limit excessive lipid accretion. Altered membrane PL composition links obesity-induced lipid accumulation in the liver to hepatic insulin resistance by inducing ER stress,[43] but we did not observe any alteration in the PC to PE ratio in Ptpn6H-KO mice despite their increased Pemt expression.

Intriguingly, the less severe hepatotoxicity (as revealed by AST and ALT levels) in the face of more profound hepatic lipid accretion correlates with reduced liver inflammation in Ptpn6H-KO mice fed an HFD for 8 or 16 weeks. Concomitantly, Pparγ expression and PPARγ activity were increased in Ptpn6H-KO livers, as was the expression of multiple Pparγ target genes, including Acaca, CD36/FAT, Fabp5, and Scd1. PPARγ is recognized as a key mediator of lipogenesis and hepatic steatosis,[25, 28, 44, 45] while also a positive modulator of insulin sensitivity, providing a less inflammatory environment for lipid storage.[46] Moreover, liver-specific disruption of PPARγ in lipodistrophy or leptin-deficient obese mice has shown that hepatocyte PPARγ contributes to hepatic steatosis but protects against insulin resistance.[47, 48] We therefore tested the hypothesis that PPARγ is directly regulated by Shp1 and demonstrated that this PTP negatively regulates PPARγ-mediated activity of a PPRE-luciferase reporter in HEK293 cells. Shp1 plays a negative role in the regulation of transcriptional processes, such as by dephosphorylation of STAT molecules.[49] However, to the best of our knowledge, Shp1 is the first PTP shown to modulate PPARγ activity in a cell-autonomous manner, although the exact mechanism of this regulation still remains to be clarified. These studies provide strong evidence that hepatocyte Shp1 restrains PPARγ activity in liver, thus providing a likely mechanism for the promotion of hepatic steatosis without provoking inflammation and hepatocellular damage in Ptpn6H-KO liver.

Our data also distinguish Shp1 from Shp2, as the hepatocyte-specific deletion of this closely associated PTP induces spontaneous hepatocellular carcinoma resulting from augmented inflammation and necrosis.[50] On the other hand, a more recent study using liver-specific Shp2 KO mice showed that these mice are protected from both liver steatosis and insulin resistance upon HFD feeding as well as hepatic inflammation, which is possibly a result of reduced adiposity due to increased energy expenditure.[10]

Taken together, our results provide genetic evidence that hepatocyte Shp1 is a novel modulator of hepatic lipid metabolism, particularly in diet-induced obesity. Remarkably, increased hepatic lipid accretion in obese Ptpn6H-KO mice does not result in hepatic insulin resistance, liver inflammation, or hepatocellular damage. Moreover, hepatocyte Shp1 deficiency markedly increases PPARγ activity in liver, which may protect against inflammation and liver damage. In conclusion, these data strongly suggest that Shp1 exhibits dual functions in hepatocytes, promoting insulin resistance and inflammation while restraining PPARγ-dependent hepatic steatosis, both contributing to the development of severe NAFLD in diet-induced obesity.


The authors thank Christine Dion, Jennifer Dumais, Maryse Pitre, Kim Denault, Geneviève Chevrier, Patricia Pelletier, Dr. Philippe St-Pierre, and Eric Fisher for technical assistance. We also thank Drs. Robert M. O'Doherty, Yves Deshaies, Roger McLeod, and Phillip White for valuable opinions and advice.