Report
Hepatic Diacylglycerol-Associated Protein Kinase Cε Translocation Links Hepatic Steatosis to Hepatic Insulin Resistance in Humans Graphical Abstract
Authors Kasper W. ter Horst, Pim W. Gilijamse, Ruth I. Versteeg, ..., Kitt F. Petersen, Gerald I. Shulman, Mireille J. Serlie
Correspondence
[email protected]
In Brief ter Horst et al. show that simple hepatic steatosis in the context of NAFLD does not fully explain hepatic insulin resistance in obese humans. Impaired hepatic insulin action was characterized by accumulation of diacylglycerol subspecies in the cytosol of hepatocytes and translocation of PKCε from the cytosol to the membrane.
Highlights d
The presence of hepatic steatosis is associated with insulin resistance
d
Intrahepatic triglycerides are not sufficient for hepatic insulin resistance
d
Diacylglycerol in hepatic cytosol predicts insulin inhibition of glucose production
d
Diacylglycerol-associated insulin resistance is characterized by PKCε translocation
ter Horst et al., 2017, Cell Reports 19, 1997–2004 June 6, 2017 ª 2017 The Author(s). http://dx.doi.org/10.1016/j.celrep.2017.05.035
Cell Reports
Report Hepatic Diacylglycerol-Associated Protein Kinase Cε Translocation Links Hepatic Steatosis to Hepatic Insulin Resistance in Humans Kasper W. ter Horst,1 Pim W. Gilijamse,1 Ruth I. Versteeg,1 Mariette T. Ackermans,2 Aart J. Nederveen,3 Susanne E. la Fleur,1,2,11 Johannes A. Romijn,4 Max Nieuwdorp,5,6,7 Dongyan Zhang,8 Varman T. Samuel,9 Daniel F. Vatner,9 Kitt F. Petersen,9 Gerald I. Shulman,8,9,10 and Mireille J. Serlie1,12,* 1Department
of Endocrinology and Metabolism of Clinical Chemistry, Laboratory of Endocrinology 3Department of Radiology 4Department of Medicine 5Department of Vascular Medicine Academic Medical Center, 1105AZ Amsterdam, the Netherlands 6Department of Internal Medicine 7Institute for Cardiovascular Research VU University Medical Center, 1081HV Amsterdam, the Netherlands 8Howard Hughes Medical Institute, Yale University, New Haven, CT 06519, USA 9Department of Medicine 10Department of Cellular and Molecular Physiology Yale University School of Medicine, New Haven, CT 06520, USA 11Metabolism and Reward Group, Netherlands Institute for Neuroscience, 1105BA Amsterdam, the Netherlands 12Lead Contact *Correspondence:
[email protected] http://dx.doi.org/10.1016/j.celrep.2017.05.035 2Department
SUMMARY
INTRODUCTION
Hepatic lipid accumulation has been implicated in the development of insulin resistance, but translational evidence in humans is limited. We investigated the relationship between liver fat and tissue-specific insulin sensitivity in 133 obese subjects. Although the presence of hepatic steatosis in obese subjects was associated with hepatic, adipose tissue, and peripheral insulin resistance, we found that intrahepatic triglycerides were not strictly sufficient or essential for hepatic insulin resistance. Thus, to examine the molecular mechanisms that link hepatic steatosis to hepatic insulin resistance, we comprehensively analyzed liver biopsies from a subset of 29 subjects. Here, hepatic cytosolic diacylglycerol content, but not hepatic ceramide content, was increased in subjects with hepatic insulin resistance. Moreover, cytosolic diacylglycerols were strongly associated with hepatic PKCε activation, as reflected by PKCε translocation to the plasma membrane. These results demonstrate the relevance of hepatic diacylglycerol-induced PKCε activation in the pathogenesis of NAFLD-associated hepatic insulin resistance in humans.
Insulin resistance and non-alcoholic fatty liver disease (NAFLD) are common and consequential complications of obesity (Grundy, 2000). Ectopic triglyceride accumulation in the liver has clearly been linked to non-alcoholic steatohepatitis (NASH), liver cirrhosis, and hepatocellular carcinoma (Wong et al., 2014), but it is less clear whether there is a causal relationship between intrahepatic triglyceride (IHTG) accumulation and hepatic insulin resistance (Farese et al., 2012; Shulman, 2000). Elevated IHTG content is associated with hepatic and/or muscle insulin resistance in most, but not in all, human studies (Bugianesi et al., 2005; Deivanayagam et al., 2008; Gastaldelli et al., 2007; Korenblat et al., 2008; Kotronen et al., 2008; Petersen et al., 2002, 2005; Sanyal et al., 2001; Seppa¨la¨-Lindroos et al., 2002). Moreover, in addition to triglycerides, it is now established that other (possibly more bioactive) lipid species may also accumulate in hepatocytes in the context of NAFLD. In several animal models, buildup of diacylglycerol (DAG) has been linked to hepatic insulin resistance through activation of protein kinase Cε (PKCε), which may directly interfere with hepatic insulin signaling (Jornayvaz et al., 2011; Samuel et al., 2004, 2007; Samuel and Shulman, 2016; Xia et al., 2015). In this regard, PKCε has recently been shown to phosphorylate insulin receptor (INSR) threonine1160, which led to inhibition of INSR kinase activity (Petersen et al., 2016). Furthermore, mice with a threonineto-alanine mutation at the homologous residue threonine1150 (InsrT1150A mice) were protected from high-fat-diet-induced
Cell Reports 19, 1997–2004, June 6, 2017 ª 2017 The Author(s). 1997 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Table 1. Characteristics of Study Participants According to Extent of Hepatic Steatosis, n = 133 No Steatosis (n = 52) Female sex
Mild Steatosis (n = 41)
Severe Steatosis (n = 40)
p
27 (51)
14 (34)
18 (45)
0.229 0.634
Age (years)
49 ± 2
51 ± 2
49 ± 2
BMI (kg/m2)
39 ± 1
41 ± 1
41 ± 1
0.442
Body fat (%)
45 ± 1
45 ± 2
46 ± 1
0.695 0.301
Fasting glucose (mmol/L)
5.2 ± 0.1
5.2 ± 0.1
5.4 ± 0.1
Triglycerides (mmol/L)
1.3 ± 0.1
1.6 ± 0.1
1.7 ± 0.2a
0.039
Cholesterol (mmol/L)
4.9 ± 0.1
4.9 ± 0.2
5.0 ± 0.2
0.671
High-density lipoprotein (mmol/L)
1.2 ± 0.0
1.1 ± 0.1
1.1 ± 0.0
0.287
Low-density lipoprotein (mmol/L)
3.0 ± 0.1
3.1 ± 0.2
3.2 ± 0.1
0.620
Alanine aminotransferase (U/L)
26 ± 2
38 ± 3b
40 ± 3c
Aspartate aminotransferase (U/L)
22 ± 1
30 ± 3
32 ± 3b
0.004
Gamma-glutamyl transpeptidase (U/L)
34 ± 4
42 ± 7
42 ± 4
0.320
C-reactive protein (mg/L)
4.9 ± 0.9
6.0 ± 1.0
7.6 ± 1.3
0.197
1,862 ± 48
2,000 ± 72
1,931 ± 54
Resting energy expenditure (kcal/day) IHTG (%)
2.9 ± 0.2
10.4 ± 0.4c
23.3 ± 1.0,c,d
99% enriched; Cambridge Isotopes). Basal rates of EGP and lipolysis were determined after 2 hr of tracer equilibration. Next, hepatic insulin sensitivity (expressed as suppression of basal EGP by insulin) and adipose tissue insulin sensitivity (expressed as suppression of circulating FA levels or suppression of glycerol Ra by insulin) were assessed after 2 hr of low-dose insulin infusion (Actrapid 20 mU $ m 2 $ min 1; Novo Nordisk Farma). Finally, peripheral insulin sensitivity (expressed as insulin-stimulated Rd of glucose) was assessed after 2 hr of high-dose insulin infusion (60 mU $ m–2 $ min–1). During hyperinsulinemia, plasma glucose was maintained at 5.0 mmol/L by frequent bedside monitoring and variable infusion of exogenous glucose (enriched with 1% [6,6-2H2] glucose). Liver Biopsies A subset of subjects (n = 29) underwent bariatric surgery shortly ( 15%, respectively. Fatty acids (FA), intrahepatic triglycerides (IHTG). a
Groups were compared by 1-way ANOVA with Bonferroni correction for multiple testing.
* p < 0.05 vs no steatosis. ** p < 0.01 vs no steatosis. *** p < 0.001 vs no steatosis.
1
ter Horst et al Supplemental Information
Figure S1. Additional Metabolic Flux Data. Related to Figure 1. (A) The Hepatic Insulin Sensitivity Index was reduced in subjects with hepatic steatosis, but did not differ between subjects with mild or severe steatosis. The index was calculated as 100 / (fasting plasma insulin x basal endogenous glucose production). (B) Expressing peripheral insulin sensitivity as percentage increase relative to basal glucose uptake yielded similar results.
2
ter Horst et al Supplemental Information
(C) Expressing peripheral insulin sensitivity relative to lean body mass yielded similar results. (D) Insulin clearance during low-dose insulin infusion (step 1 of the clamp) did not differ between subjects with varying extent of steatosis. Insulin clearance was calculated as insulin infusion rate / plasma insulin concentration during the clamp. (E) Insulin clearance during high-dose insulin infusion (step 2) did not differ between subjects with varying extent of steatosis. Data are mean ± SEM (n = 133). Rate of disappearance (Rd), fat-free mass (FFM). ** p < 0.01 vs no steatosis. *** p < 0.001 vs no steatosis.
3
ter Horst et al Supplemental Information
Figure S2. Hepatic DAG Accumulation in the Cytosolic Fraction, but Not Total DAG Content, Membrane DAG Content, or Ceramide Content, Predicted Hepatic Insulin Sensitivity in Obese Subjects. Related to Figure 3. (A) Hepatic ceramide content and insulin suppression of EGP. (B) Total hepatic DAG content and insulin suppression of EGP. (C) Hepatic DAG content in the cytosolic fraction and insulin suppression of EGP. (D) Hepatic DAG content in the membrane fraction and insulin suppression of EGP. Lines were fitted by least squares method (n = 25-29). Diacylglycerol (DAG), endogenous glucose production (EGP). 4
ter Horst et al Supplemental Information
Table S2. Correlations Between Individual Hepatic DAG Species and Insulin Sensitivity of EGP Suppression (n = 29). Related to Figure 3. Cytosolic fraction Species r p Abundance ra [mean ± SEM (nmol/g)] C18:1-C16:0 -0.44 0.019 222.6 ± 33.9 0.01 C20:4-C20:5 -0.12 0.551 17.1 ± 2.3 0.41 C16:0-C20:4 -0.36 0.061 0.7 ± 0.1 0.39 C18:0-C20:4 -0.27 0.172 0.6 ± 0.1 0.41 C18:0-C18:0 -0.47 0.012 15.2 ± 2.1 -0.12 C18:2-C18:0 -0.33 0.083 7.9 ± 1.2 0.34 C18:1-C18:0 -0.42 0.027 27.6 ± 4.2 0.06 C16:0-C16:0 -0.47 0.013 73.5 ± 12.0 -0.21 C18:0-C16:0 -0.45 0.015 23.5 ± 3.5 -0.10 C18:1-C18:1 -0.41 0.030 326.4 ± 51.2 0.17 C18:1-C18:2 -0.33 0.091 59.6 ± 9.5 0.34 C18:2-C18:2 -0.23 0.233 9.8 ± 1.6 0.38 C16:0-C18:2 -0.33 0.085 79.0 ± 12.4 0.29 Highly abundant species that were also related to hepatic insulin Diacylglycerol (DAG), endogenous glucose production (EGP). a
a
Membrane fraction p Abundance [mean ± SEM (nmol/g)] 0.948 215.2 ± 14.6 0.028 84.7 ± 6.9 0.041 1.6 ± 0.1 0.032 2.0 ± 0.1 0.559 15.3 ± 0.9 0.078 12.9 ± 0.8 0.760 32.8 ± 2.4 0.296 71.0 ± 4.5 0.630 20.6 ± 1.3 0.377 377.4 ± 34.2 0.076 65.8 ± 6.8 0.044 10.4 ± 1.1 0.130 100.9 ± 5.9 sensitivity are highlighted.
Correlations were evaluated by Pearson’s correlation coefficient.
5
ter Horst et al Supplemental Information
Table S3. Correlations Between Individual Hepatic Ceramide Species and Insulin Sensitivity of EGP Suppression (n = 25). Related to Figure 3. Species
ra
Abundance [mean ± SEM (nmol/g)] C16:0 -0.27 0.187 58.8 ± 2.1 C18:0 -0.24 0.254 16.6 ± 0.6 C20:0 -0.28 0.180 43.8 ± 6.6 C22:0 -0.27 0.190 65.1 ± 2.4 C24:1 -0.16 0.447 40.9 ± 1.2 C24:0 -0.06 0.779 52.4 ± 2.1 Endogenous glucose production (EGP). a
p
Correlations were evaluated by Pearson’s correlation coefficient.
6
ter Horst et al Supplemental Information
Table S4. Characteristics of Subjects with Normal Hepatic Insulin Sensitivity or with Hepatic Insulin Resistance, Who Had Liver Biopsies Taken during Bariatric Surgery. Related to Figure 3. Sensitive Resistant pa (n = 10) (n = 8) Female sex 6 (60) 3 (38) 0.343 Age (years) 51 ± 2 45 ± 4 0.135 2 Body mass index (kg/m ) 41 ± 1 47 ± 3 0.055 Body fat content (%) 45 ± 1 47 ± 3 0.603 Fasting glucose (mmol/l) 4.8 ± 0.1 5.5 ± 0.3 0.013 Fasting insulin (pmol/l) 115 ± 20 183 ± 20 0.029 -1 -1 Basal EGP (µmol∙kg ∙min ) 6.9 ± 0.3 7.2 ± 0.3 0.518 Insulin suppression of EGP (%) 87 ± 2 56 ± 1 < 0.001 IHTG content (%) 9.2 ± 3.0 14.2 ± 2.6 0.209 Data are count (%) or mean ± SEM. Hepatic insulin sensitivity and resistance were defined as insulin suppression of EGP > 80% and < 65%, respectively. Endogenous glucose production (EGP), intrahepatic triglyceride (IHTG). a
Groups were compared by 2-tailed independent samples t test.
7
ter Horst et al Supplemental Information
Figure S3. Release of FA from Lipolysis during Hyperinsulinemic Conditions Predicts Hepatic DAG Accumulation. Related to Figure 3. (A) The rate of lipolysis during hyperinsulinemic-euglycemic clamps was assessed by tracerdilution of [1,1,2,3,3-2H5]glycerol. (B) Plasma FA concentrations during hyperinsulinemia. Data were fitted by least squares method (n = 28). Diacylglycerol (DAG), fatty acids (FA).
8
